2-fold at these two time points were excluded from analysis .\",\n \"The nls package in R was used to fit the following equation for each gene:y=C+log21-e-kx-t0 , where x is the labelling time in hours ( using 9999 h as the labeling time for the steady-state data point ) , y is log2 value of the normalized read count ( level of labeled RNA ) , and t0 is the time offset .\",\n \"k is the decay constant , which was used to determine the half-life t1/2 using , t1/2=ln2k , and C is a coefficient determined by both k and the synthesis rate S , such that , C=log2Sk .\",\n \"To fit t0 for each condition , the value of t0 was varied from 0 . 05 to 0 . 7 , with an interval of 0 . 05 , and the value that gave the smallest mean square-loss of y when fitting to data from all genes was used .\",\n \"When fitting to results for each gene , C was bound by ( 0 , Inf ) , and k was bound by ( 0 , Inf ) .\",\n \"C was initialized with max ( y ) , and k was initialized with 0 . 23 .\",\n \"Genes without a converged fit ( 27 of 9644 , 17 of 9646 , and 12 of 9728 in analysis of the siControl , siPABPC1 , and siPABPC1&4 samples , respectively ) were omitted from down-stream analyses .\",\n \"mRNA half-life values are reported in Figure 4—figure supplement 3—source data\",\n \"1 . Graphs were generated and statistical analyses were performed using R ( R Development Core Team , 2013 ) .\",\n \"Statistical parameters including the value of n , statistical test , and statistical significance ( p value ) are reported in the figures or their legends .\",\n \"No statistical methods were used to predetermine sample size .\",\n \"Statistical tests for correlations ( between nonoverlapping dependent groups ) were performed based on a published method ( Silver and Hittner , 2004 ) using R package cocor ( Diedenhofen and Musch , 2015 ) .\",\n \"For luciferase assays of injected oocytes , each replicate refers to a group of 7–10 oocytes .\",\n \"For 35S labeling of injected oocytes , each replicate refers to a group of 5–8 oocytes .\",\n \"For the puromycin-based translation assay , each replicate refers to a separate transfection experiment .\",\n \"Raw and processed data from sequencing were deposited in the GEO database ( GSE166544 ) .\",\n \"TAIL-seq and PAL-seq analyses were performed using a custom script written in Python 2 . 7 and available at https://github . com/coffeebond/PAL-seq , ( copy archived at swh:1:rev:b0aa1ba99cc89ce2080069b50cd441a7718b7b03 , Xiang , 2021a , Xiang , 2021b ) .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["In animal oocytes and early embryos , mRNA poly ( A ) -tail length strongly influences translational efficiency ( TE ) , but later in development this coupling between tail length and TE disappears .","Here , we elucidate how this coupling is first established and why it disappears .","Overexpressing cytoplasmic poly ( A ) -binding protein ( PABPC ) in Xenopus oocytes specifically improved translation of short-tailed mRNAs , thereby diminishing coupling between tail length and TE .","Thus , strong coupling requires limiting PABPC , implying that in coupled systems longer-tail mRNAs better compete for limiting PABPC .","In addition to expressing excess PABPC , post-embryonic mammalian cell lines had two other properties that prevented strong coupling: terminal-uridylation-dependent destabilization of mRNAs lacking bound PABPC , and a regulatory regime wherein PABPC contributes minimally to TE .","Thus , these results revealed three fundamental mechanistic requirements for coupling and defined the context-dependent functions for PABPC , which promotes TE but not mRNA stability in coupled systems and mRNA stability but not TE in uncoupled systems ."],"string":"[\n \"In animal oocytes and early embryos , mRNA poly ( A ) -tail length strongly influences translational efficiency ( TE ) , but later in development this coupling between tail length and TE disappears .\",\n \"Here , we elucidate how this coupling is first established and why it disappears .\",\n \"Overexpressing cytoplasmic poly ( A ) -binding protein ( PABPC ) in Xenopus oocytes specifically improved translation of short-tailed mRNAs , thereby diminishing coupling between tail length and TE .\",\n \"Thus , strong coupling requires limiting PABPC , implying that in coupled systems longer-tail mRNAs better compete for limiting PABPC .\",\n \"In addition to expressing excess PABPC , post-embryonic mammalian cell lines had two other properties that prevented strong coupling: terminal-uridylation-dependent destabilization of mRNAs lacking bound PABPC , and a regulatory regime wherein PABPC contributes minimally to TE .\",\n \"Thus , these results revealed three fundamental mechanistic requirements for coupling and defined the context-dependent functions for PABPC , which promotes TE but not mRNA stability in coupled systems and mRNA stability but not TE in uncoupled systems .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Cells are microscopic biological factories that are constantly creating new proteins .","To do so , a cell must first convert its master genetic blueprint , the DNA , into strands of messenger RNA or mRNA .","These strands are subsequently translated to make proteins .","Cells have two ways to adjust the number of proteins they generate so they do not produce too many or too few: by changing how many mRNA molecules are available for translation , and by regulating how efficiently they translate these mRNA molecules into proteins .","In animals , both unfertilized eggs and early-stage embryos lack the ability to create or destroy mRNAs , and consequently cannot adjust the number of mRNA molecules available for translation .","These cells can therefore only regulate how efficiently each mRNA is translated .","They do this by changing the length of the so-called poly ( A ) tail at the end of each mRNA molecule , which is made up of a long stretch of repeating adenosine nucleotides .","The mRNAs with longer poly ( A ) tails are translated more efficiently than those with shorter poly ( A ) tails .","However , this difference disappears in older embryos , when both long and short poly ( A ) tails are translated with equal efficiency , and it is largely unknown why .","To find out more , Xiang and Bartel studied frog eggs , and discovered that artificially raising levels of a protein that binds poly ( A ) tails , also known as PABPC , improved the translation of short-tailed mRNAs to create a situation in which both short- and long-tailed mRNAs were translated with near-equal efficiency .","This suggested that short- and long-tailed mRNAs compete for limited amounts of the translation-enhancing PABPC , and that long-tailed mRNAs are better at it than short-tailed mRNAs .","Further investigation revealed that eggs also had to establish the right conditions for PABPC to enhance translation and had to protect mRNAs not associated with PABPC from being destroyed before they could be translated .","Overall , Xiang and Bartel found that in eggs and early embryos , PABPC and poly ( A ) tails enhanced the translation of mRNAs but did not influence their stability , whereas later in development , they enhanced mRNA stability but not translation .","This research provides new insights into how protein production is controlled at different stages of animal development , from unfertilized eggs to older embryos .","Understanding how this process is regulated during normal development is crucial for gaining insights into how it can become dysfunctional and cause disease .","These findings may therefore have important implications for research into areas such as infertility , reproductive medicine and rare genetic diseases ."],"string":"[\n \"Cells are microscopic biological factories that are constantly creating new proteins .\",\n \"To do so , a cell must first convert its master genetic blueprint , the DNA , into strands of messenger RNA or mRNA .\",\n \"These strands are subsequently translated to make proteins .\",\n \"Cells have two ways to adjust the number of proteins they generate so they do not produce too many or too few: by changing how many mRNA molecules are available for translation , and by regulating how efficiently they translate these mRNA molecules into proteins .\",\n \"In animals , both unfertilized eggs and early-stage embryos lack the ability to create or destroy mRNAs , and consequently cannot adjust the number of mRNA molecules available for translation .\",\n \"These cells can therefore only regulate how efficiently each mRNA is translated .\",\n \"They do this by changing the length of the so-called poly ( A ) tail at the end of each mRNA molecule , which is made up of a long stretch of repeating adenosine nucleotides .\",\n \"The mRNAs with longer poly ( A ) tails are translated more efficiently than those with shorter poly ( A ) tails .\",\n \"However , this difference disappears in older embryos , when both long and short poly ( A ) tails are translated with equal efficiency , and it is largely unknown why .\",\n \"To find out more , Xiang and Bartel studied frog eggs , and discovered that artificially raising levels of a protein that binds poly ( A ) tails , also known as PABPC , improved the translation of short-tailed mRNAs to create a situation in which both short- and long-tailed mRNAs were translated with near-equal efficiency .\",\n \"This suggested that short- and long-tailed mRNAs compete for limited amounts of the translation-enhancing PABPC , and that long-tailed mRNAs are better at it than short-tailed mRNAs .\",\n \"Further investigation revealed that eggs also had to establish the right conditions for PABPC to enhance translation and had to protect mRNAs not associated with PABPC from being destroyed before they could be translated .\",\n \"Overall , Xiang and Bartel found that in eggs and early embryos , PABPC and poly ( A ) tails enhanced the translation of mRNAs but did not influence their stability , whereas later in development , they enhanced mRNA stability but not translation .\",\n \"This research provides new insights into how protein production is controlled at different stages of animal development , from unfertilized eggs to older embryos .\",\n \"Understanding how this process is regulated during normal development is crucial for gaining insights into how it can become dysfunctional and cause disease .\",\n \"These findings may therefore have important implications for research into areas such as infertility , reproductive medicine and rare genetic diseases .\"\n]","isConversational":false},"year":{"kind":"string","value":"2021"}}},{"rowIdx":84,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["neuroscience"],"string":"[\n \"neuroscience\"\n]","isConversational":false},"title":{"kind":"string","value":"Synaptic organization of the Drosophila antennal lobe and its regulation by the Teneurins"},"id":{"kind":"string","value":"elife-03726-v1"},"sections":{"kind":"list like","value":[["Understanding how neural circuits process information requires knowing not only the neuroanatomical connectivity and electrophysiological properties of individual neurons , but also the organization of synapses between specific neuronal types .","Knowledge of these basic organizational principles , as well as the mechanisms for how they are achieved in a complex circuit , represents a critical goal for neuroscience .","Historically , neuromuscular junctions have been model synapses for mechanistic studies because of their ease of access ( Keshishian et al . , 1996; Sanes and Lichtman , 1999 ) .","However , neuromuscular circuits display less complexity compared to neural circuits in the central nervous system ( CNS ) .","Serial-section electron microscopy ( EM ) has offered remarkable resolution in studying CNS synapses in the context of circuits , including the entire Caenorhabditis elegans nervous system ( White et al . , 1986 ) , the mouse olfactory bulb ( Hinds and Hinds , 1976a , 1976b ) , retina ( Briggman et al . , 2011; Helmstaedter et al . , 2013 ) , and visual cortex ( Bock et al . , 2011 ) , and the Drosophila visual system ( Meinertzhagen and O'Neil , 1991; Takemura et al . , 2013 ) .","However , these approaches are labor intensive and not easily amenable to analysis following genetic perturbation , at varying developmental stages , or with intact samples .","Indeed , even for well-known synaptic organizers like Neurexin and Neuroligin ( Chia et al . , 2013 ) , studies investigating their function were rarely carried out in an intact system with a resolution sufficient to reveal their detailed functions in synaptogenesis between defined neuronal types .","As such , there is a need to examine CNS synapses in the intact brain , at the level of light microscopy in a system with sufficient complexity to reflect features of the CNS , but tractable enough to dissect connectivity and function using genetic tools .","Current work in Drosophila has sought to fulfill this need ( Kremer et al . , 2010; Christiansen et al . , 2011; Berger-Muller et al . , 2013; Chen et al . , 2014 ) .","The Drosophila olfactory system contains complex circuits that transform chemosensory input to the regulation of behavioral output .","Olfactory information is received by first-order olfactory receptor neurons ( ORNs ) and is transmitted by ORN axons to the antennal lobe , the first olfactory processing center in the brain .","There , the information is conveyed to the dendrites of second-order projection neurons ( PNs ) through class-specific ORN–PN synapses .","PN axons carry signals to higher-order brain centers ( Vosshall and Stocker , 2007 ) .","The local interneurons ( LNs ) innervate the antennal lobe extensively with their dendrites ( Chou et al . , 2010 ) and are involved in transforming ORN input to PN output ( Wilson , 2013 ) .","These three groups of neurons have extensive interconnectivity and indeed , the antennal lobe has emerged as a model neural circuit for investigating the general principles of information processing ( Wang , 2012; Wilson , 2013; Twick et al . , 2014 ) .","While many antennal lobe neurons have been examined electrophysiologically ( Wilson , 2013 ) , there has been little analysis of their synaptic architecture: how many connections are made , where they are made in relation to other neurons in the circuit , and what molecules are required for their organization .","Electron microscopy has been used to probe different genetic conditions ( Acebes and Ferrus , 2001; Acebes et al . , 2011 ) but could not definitively assign glomerular identity or cell type for individual synapses , highlighting the need for structural analysis at the level of light microscopy in three dimensions .","Knowledge about synaptic organization is critical to understand how sensory information is received , processed , and relayed by this circuit .","Thus , the fly olfactory circuit represents an excellent opportunity to assess synaptic organization in an intact system at the resolution of individual , defined neuronal classes , linking structure to physiology and function .","Here , we utilized an approach where presynaptic active zone and postsynaptic neurotransmitter receptor proteins were tagged with fluorescent molecules to highlight and quantitatively describe the synaptic organization of the Drosophila antennal lobe .","We identified distinct rules that govern the number , density , and spatial organization of olfactory synapses for multiple classes of neurons .","To understand how these rules are implemented at the molecular level , we investigated the function of the Teneurins , evolutionarily conserved type II transmembrane proteins recently shown to regulate synaptic partner matching and neuromuscular synaptic organization ( Hong et al . , 2012; Mosca et al . , 2012 ) .","We show that presynaptic Ten-a and postsynaptic Ten-m regulate central synapse number and active zone structure .","These represent the first in vivo functions of Teneurins in organizing CNS synapses , which could contribute to their association with brain disorders such as bipolar disorder ."],["To visualize synapses in vivo in genetically identifiable cells , we explored strategies to label active zones , sites of presynaptic neurotransmitter release ( Zhai and Bellen , 2004 ) .","In Drosophila , the primary unit of active zone architecture is an electron dense structure called the T-bar , a specialized formation that promotes vesicle docking and release ( Wichmann and Sigrist , 2010 ) present at peripheral ( Jia et al . , 1993 ) and central synapses ( Meinertzhagen and O'Neil , 1991; Prokop and Meinertzhagen , 2006 ) .","The major structural components of the T-bar include Bruchpilot ( Brp ) and Rim-binding protein ( Wagh et al . , 2006; Liu et al . , 2011 ) , with Brp encoding the ‘table-top’ of the T-bar and Rim-binding protein comprising the ‘pedestal’ ( Figure 1A ) .","Widely used monoclonal antibodies ( nc82 ) against Brp ( Blanco Redondo et al . , 2013 ) label synaptic neuropil and discrete puncta corresponding to active zones ( Wagh et al . , 2006 ) .","However , these antibodies cannot distinguish synapses formed by a given class of neurons in complex CNS circuits .","Therefore , to visualize active zones in genetically identifiable cells , we sought to label Brp with the GAL4/UAS system ( Brand and Perrimon , 1993 ) . 10 . 7554/eLife . 03726 . 003Figure 1 . Measuring CNS synapses using Bruchpilot-Short .","( A ) Diagram comparing full-length Bruchpilot and Bruchpilot-Short .","Numbers denote amino acids and orange represents coiled coil regions ( after Wagh et al . , 2006 ) .","The positions of Brp and Brp-Short in relation to known active zone proteins are also shown in a simplified format ( after Liu et al . , 2011 ) .","( B–D )","High magnification confocal stacks of a single DA1 glomerulus with putative synapses labeled by Brp-Short ( B ) , neurites labeled by mCD8-GFP ( C ) , and the merge ( D ) .","Insets show higher magnification of a single optical section to demonstrate the punctate nature of Brp-Short-mStraw .","( E–G )","Screenshots of three-dimensional renderings show conversion of Brp-Short into ‘Spots’ ( E ) and mCD8-GFP into a volumetric surface rendering ( F ) .","For ( F ) and ( G ) , the transparency of the surface rendering is at 60% to highlight internal structure and make Brp-Short puncta visible .","( H ) Quantification of Brp-Short puncta ( red , left axis ) and neurite volume ( black , right axis ) in males and females for five ORN classes .","Statistical comparisons between males and females of a single genotype were done by student's t test .","Significance across all genotypes for density was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .","***p < 0 . 001 .","In all cases , n ≥ 9 brains , 18 antennal lobes for each genotype .","( I ) Quantification of synaptic density ( Brp-Short puncta/volume of neurites ) in males ( blue ) and females ( pink ) for five ORN classes .","Despite considerable differences in Brp-Short puncta number , all classes of ORNs display nearly identical synaptic densities .","Scale bar = 5 µm for all images . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00310 . 7554/eLife . 03726 . 004Figure 1—figure supplement 1 . Validation of Brp-Short .","( A ) Representative high magnification confocal single sections of Or67d-positive ORNs innervating the DA1 glomerulus expressing Brp-Short-mStraw and DSyd1-EGFP and stained with antibodies against mStraw ( red ) and GFP ( green ) .","Note the significant overlap between Brp-Short and DSyd-1 , a known presynaptic marker .","( B ) Representative high magnification confocal single sections of Or67d-positive ORNs innervating the DA1 glomerulus expressing Brp-Short-mStraw and mCD8-GFP in a control animal .","( C ) Representative high magnification confocal single sections of Or67d-positive ORNs innervating the DA1 glomerulus expressing Brp-Short-mStraw , mCD8-GFP and an RNAi transgene against Brp .","As endogenous Brp has been impaired , Brp-Short-mStraw is greatly diminished at the synapse .","Scale bar = 5 μm .","( D ) A single electron dense region as imaged by immunoelectron microscopy in an animal expressing Brp-Short-EGFP in all ORNs .","5-nm Nanogold dots ( arrowheads ) decorate electron-dense structures characteristic of active zones following immunostaining against the GFP epitope , suggesting that Brp-Short localizes to endogenous active zones .","Scale bar = 25 nm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00410 . 7554/eLife . 03726 . 005Figure 1—figure supplement 2 . Additional examples of the Brp-Short assay . Representative high magnification confocal z-stack images of ORNs innervating the VA1d ( Or88a-GAL4; [A] ) , VA1lm ( Or47b-GAL4; [B] ) , DL4 ( AM29-GAL4; [C] ) , and DM6 ( AM29-GAL4; [D] ) glomeruli expressing Brp-Short-mStraw and stained with antibodies to mStraw ( red ) and N-Cadherin ( blue ) .","In all cases , puncta are clearly visible; these glomeruli were quantified in Figure 1H .","Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 005 Full-length Brp is a 1740 amino acid protein that can aggregate upon ectopic overexpression ( Wagh et al . , 2006 ) , limiting its use as a labeling strategy .","However , Bruchpilot-Short ( Brp-Short ) , a nonfunctional 754-residue portion of Brp ( Figure 1A ) , localizes to endogenous sites of full-length Brp ( Schmid et al . , 2008; Fouquet et al . , 2009 ) without deleterious effects .","Thus , under the control of a binary expression system , Brp-Short can act as a reporter of endogenous active zones without disrupting morphology or function ( Schmid et al . , 2008; Fouquet et al . , 2009; Kremer et al . , 2010 ) .","This strategy has been employed at neuromuscular synapses ( Schmid et al . , 2008; Fouquet et al . , 2009 ) , in the mushroom body ( Kremer et al . , 2010; Christiansen et al . , 2011 ) , and in the visual system ( Berger-Muller et al . , 2013 ) .","We utilized Brp-Short to quantitatively measure the number of active zones , a key element of understanding synaptic organization .","We expressed an mStrawberry-tagged version ( Brp-Short-mStraw: Owald et al . , 2010 ) in Or67d ORNs that innervate the DA1 glomerulus .","Coexpression with a known presynaptic protein , DSyd-1 , revealed extensive colocalization ( Figure 1—figure supplement 1 , panel A ) , suggesting Brp-Short-mStraw correctly localizes to synapses .","Further , RNAi against a portion of endogenous Brp not encoded by the Brp-Short transgene in those ORNs resulted in a loss of Brp-Short signal at ORN presynapses ( Figure 1—figure supplement 1 , panels B–C ) , though Brp-Short was still detectable in the cell body .","This supports the hypothesis that Brp-Short requires endogenous Brp for proper localization to synapses; alternatively , Brp-Short could require endogenous Brp for proper trafficking to axon terminals .","Finally , to confirm the veracity of the synaptic signal , we performed immuno-electron microscopy ( immuno-EM ) using antibodies to the GFP-tag ( the mStraw tag is not amenable to immuno-EM ) of a Brp-Short-GFP transgene ( Schmid et al . , 2008 ) expressed in all ORNs via pebbled-GAL4 ( Sweeney et al . , 2007 ) .","With pebbled-GAL4 , Brp-Short expression and localization was similar to that observed for ORN-class-specific GAL4 drivers ( data not shown ) .","While immuno-EM cannot be utilized as a quantitative strategy for determining the number of active zones ( as the conditions needed to preserve the GFP epitope necessarily reduce the resolution of the technique to clearly discern all synaptic densities ) , it can be used to verify that Brp-Short localizes to electron-dense regions that are consistent with synaptic active zones .","Indeed , 5-nm gold particles were observed to decorate putative active zone profiles ( Figure 1—figure supplement 1 , panel D ) in ultrathin sections of fly antennal lobes .","90 . 1% of Nano-Gold dots colocalized with synaptic densities ( 675 5-nm Nano-Gold particles out of 749 particles observed ) .","Combined , these experiments demonstrate that Brp-Short localization is consistent with active zones in the fly CNS .","To determine the number of active zones produced by ORN axons that project to a given glomerulus , we expressed Brp-Short-mStraw in various ORN classes ( Figure 1B–D , Figure 1—figure supplement","2 ) together with UAS-mCD8-GFP ( Lee and Luo , 1999 ) as a general neurite label .","Single optical sections ( Figure 1B–D , insets ) revealed the punctate nature of the Brp-Short-mStraw signal .","To quantify the number of Brp-Short puncta in these ORNs , we imaged individual glomeruli using high magnification confocal microscopy followed by image deconvolution and 3-D rendering ( Video 1; see ‘Materials and methods’ ) .","Image processing allowed us to calculate the number of Brp-Short puncta ( Figure 1B , E ) as ‘Spots’ and the neurite volume by surface rendering of the membrane marker mCD8-GFP ( Figure 1C , F ) .","The numbers of Brp-Short puncta exhibited low variability for multiple classes of ORNs ( Figure 1H , Figure 1—figure supplement 2 ) , suggesting that neuronal classes form stereotyped numbers of active zones across animals .","While extensive EM reconstruction is not available for the antennal lobe ( and thus , absolute counts of olfactory synapses in each glomerulus for comparison ) , the recent use of labeled Brp in the Drosophila visual system indicated that one Brp punctum is equivalent to one active zone ( Chen et al . , 2014 ) .","Therefore , it is likely that the Brp-Short label largely reflects actual active zone number . 10 . 7554/eLife . 03726 . 006Video 1 . Three-dimensional rendering of a Brp-Short and mCD8-GFP in a DA1 glomerulus .","( related to Figure","1 ) Animation of a three-dimensional rendering of the glomerulus shown in Figure 1B–D .","Each channel is highlighted and the transformation to either Spots ( for Brp-Short puncta ) or a surface rendering to measure neurite volume ( for mCD8-GFP staining ) is shown .","High magnification of the rendered glomerulus reveals the distribution of Brp-Short puncta within the glomerulus . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 006 In examining five classes of ORNs , we found that ORN synapse number , as indicated by the number of Brp-Short puncta , scaled with ORN axon volume ( Figure 1H ) , which itself scaled with glomerular size .","Moreover , the assay revealed sex-specific differences in synapse number for two glomeruli , DA1 and VA1lm , that coincide with known sex-specific differences in glomerular volume ( Stockinger et al . , 2005 ) .","Remarkably , despite differences in synapse number and ORN neurite volume , all observed ORN classes had a constant synaptic density of 0 . 56 Brp-Short puncta per µm3 ( Figure 1I ) .","These data suggested the existence of an organizational mechanism that controls synaptic density .","In all , our results validated the use of Brp-Short as a quantifiable presynaptic active zone label in central neurons and revealed quantitative parameters that govern the synaptic contacts of ORNs: a stereotyped number of connections for each neuronal class from animal to animal and a constant synaptic density regardless of ORN class .","In the olfactory system , individual ORNs of a particular class ( as denoted by their odorant receptor expression ) project to an identifiable glomerulus ( Vosshall and Stocker , 2007 ) .","In terms of their projection pattern ( Jefferis and Hummel , 2006 ) and physiological function ( Wilson , 2013 ) , individual ORNs can largely be treated equivalently .","However , it is unknown whether each neuron contributes an equal number of synapses to the aggregate , stereotyped average or whether there is heterogeneity in synapse number at the level of individual neurons .","Different results would suggest distinct rules for how individual neurons determine how many connections to form .","To address this , we utilized MARCM ( Lee and Luo , 1999 ) to label the synapses and membranes of individual ( or small subsets of ) ORNs that innervate two glomeruli , DL4 and DM6 .","Most frequently , we obtained 2–3 cell clones ( Figure 2A–B ) , but we observed occurrences of single-cell clones ( Figure 2—figure supplement 1 , panels A–B ) and large clones ( Figure 2—figure supplement 1 , panels C–D ) that encompassed the entire glomerulus .","Single neuron clones were readily discernible in confocal stacks; cell numbers could not be precisely assigned to larger than 2–3 cell clones , however , due to axon fasciculation .","For all clones ( n = 100 for DL4 , 99 for DM6 ) except the largest 7 for DL4 and 9 for DM6 , we quantified the number of Brp-Short puncta and the volume of the mCD8-GFP-labeled neurites as above ( Figure 1 ) , and plotted the distribution of puncta number across all clones in histograms .","For both DL4 ( Figure 2C ) and DM6 ( Figure 2D ) , we observed that as clone size increased , the number of Brp-Short puncta appeared to increase in discrete steps of a similar size , suggesting that each individual neuron contributes an equal number of synapses to the glomerular average .","If each neuron contributed a widely different number of synapses to the average , we would have observed a more continuous distribution . 10 . 7554/eLife . 03726 . 007Figure 2 . Clonal analysis of ORN synapse number . Representative high magnification confocal stacks of small clones of DL4 ORNs ( A ) and DM6 ORNs ( B ) expressing Brp-Short-mStraw ( red ) and mCD8-GFP ( green ) .","( C ) Frequency histogram of Brp-Short puncta in DL4 clones ( n = 93 clones from 64 animals ) .","( D ) Frequency histogram of Brp-Short puncta in DM6 clones ( n = 90 clones from 64 animals ) .","Both graphs exhibit notable periodicity , suggesting that each neuron contributes approximately the same number of synapses .","Colored dashed lines indicate Gaussian fits for individual peaks .","Solid black lines represent the best-fit sum of Gaussian relationship for each dataset .","( E ) Quantification of Brp-Short puncta from identified single neuron clones to DL4 ( n = 16 clones ) or DM6 ( n = 17 clones ) .","( F ) Quantification of synapse density from all clones .","In both DL4 and DM6 , the clonal average density is identical to the class average density .","In ( E ) and ( F ) , data represent mean ± SEM .","Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00710 . 7554/eLife . 03726 . 008Figure 2—source data 1 . Table of Gaussian curve-fitting data . Table of statistical values for three potential sum of Gaussian fits each for the DL4 and DM6 MARCM datasets . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00810 . 7554/eLife . 03726 . 009Figure 2—source data 2 . Table of resampling approach . Table of simulation data for three distributions of the DL4 and DM6 MARCM data sets and the probabilities of obtaining simulated data more evenly distributed than observed . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00910 . 7554/eLife . 03726 . 010Figure 2—figure supplement 1 . Additional DL4 and DM6 ORN MARCM clones . Representative high magnification confocal stacks of single cell clones of DL4 ORNs ( A ) and DM6 ORNs ( B ) and large clones of DL4 ORNs ( C ) and DM6 ORNs ( D ) expressing Brp-Short-mStraw and mCD8-GFP and stained for antibodies against mStraw ( red ) , GFP ( green ) , and N-Cadherin ( blue ) .","As previously observed ( Joo et al . , 2013 ) , single cell ORN clones have a low signal-to-noise ratio while discerning larger clones is more evident .","In A–B , the glomeruli are outlined ( dashed line ) .","Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 010 To statistically analyze the distributions for DL4 and DM6 ORNs , we determined their best-fit relationships using the first four peaks for both data sets , which had the most observations .","Each individual peak could be reliably fit using a single Gaussian relationship ( r2 > 0 . 98 , Figure 2C–D , colored dashed lines ) , and the entire set of peaks could be fit by a sum of Gaussians ( see Figure 2—source data 1; ‘Materials and methods’ ) .","The DL4 relationship was best fit by a 4-term sum of Gaussians ( Figure 2C , solid black line ) with a value of 19 . 13 for the apex of the first peak , corresponding to the ‘single cell’ contribution of active zone puncta .","This is in close agreement with our original data of 19 . 75 ± 0 . 68 puncta for single cell clones ( Figure 2E ) .","The DL6 relationship was best fit by a 5-term sum of Gaussians ( Figure 2D , solid black line ) , with a value of 28 . 62 active zone puncta as the ‘single cell’ contribution , which is again in strong agreement with our original data of 29 . 06 ± 0 . 95 puncta for single cell clones ( Figure 2E ) .","Thus , in both cases , a sum of Gaussians can accurately represent the observed peaks and predict single ‘quantum’ contributions similar to those observed for single cell clones .","These ‘quantum’ values are consistent with the ‘step size’ of Brp-Short puncta measured in clones with increasing size ( Figure 2C–E ) .","Further , in comparing DL4 and DM6 , the difference in the number of puncta per single cell clone ( Figure 2E ) is consistent with the numerical relationship between the aggregate glomerular averages ( Figure 1H ) .","All observed clones also displayed a nearly identical synaptic density ( Figure 2F ) regardless of their size , consistent with measurements in whole glomeruli .","To further test whether each ORN contributes the same number of synapses to the aggregate , we examined the spacing of the peaks .","If the peaks were perfectly evenly spaced , the standard deviation in the distances ( S ) between each peak would be 0 .","Using the individual Gaussian relationships ( colored dashed lines ) , we calculated an S value of 0 . 235 for DL4 and 2 . 26 for DM6 .","While visual assessment suggested uniformity , we used these values to more rigorously assess this .","We used a resampling approach and calculated an S value for each simulation to test the null hypothesis that there is no periodic relationship between the peaks .","A lower S value than that which was observed means that the simulation resulted in peaks that were more evenly spaced .","Figure 2—source data 2 details these results .","For distributions of DL4 or DM6 , there was a low probability ( <0 . 00048 or <0 . 05 , respectively ) of seeing peaks as evenly distributed as that observed .","Thus , we reject the null hypothesis of no quantal relationship for both DL4 and DM6 .","In all , this approach demonstrates that the peaks have equal spacing in between , further supporting our assertion that each ORN contributes an equivalent number of active zones to the aggregate and that each class of ORNs has a characteristic ‘step size’ consistent with its aggregate average .","To better understand CNS synaptic organization , it is important to know not only how many connections are formed , but also where they form .","Brp-Short allowed us to examine the spatial organization of active zones of any specific neuronal class within individual glomeruli .","As an approximation of the complement of active zones for the DA1 glomerulus , we examined its component ORNs , PNs , and LNs .","Using GAL4 drivers for the ORNs ( Or67d-GAL4 ) , PNs ( Mz19-GAL4 ) , and an LN line that includes most ipsilateral LNs ( NP3056-GAL4 ) , all of which innervate the DA1 glomerulus , we examined the localization of their synapses .","For all classes , Brp-Short puncta colocalized with endogenous Brp staining ( Figure 3—figure supplement 1 , panels A–C ) recognized by an antibody whose epitope is not contained within Brp-Short ( Wagh et al . , 2006 ) , demonstrating that Brp-Short recognizes endogenous active zones .","These three classes of neurons , when combined , provide 3342 ± 99 Brp-Short puncta per glomerulus .","When the endogenous number of active zones is quantified using an antibody to endogenous Brp using the same image-processing strategy , DA1 contains 3849 ± 80 active zones ( Figure 3—figure supplement 1D ) .","This suggests that the majority of synapses in the DA1 glomerulus are indeed made by Or67d-GAL4+ ORNs , Mz19-GAL4+ PNs , and NP3056-GAL4+ LNs; the discrepancy may be contributed by synapses made by additional neurons , such as other classes of local interneurons ( Chou et al . , 2010 ) , atypical projection neurons ( Lai et al . , 2008 ) , or potentially peptidergic neurons ( Ignell et al . , 2009; Carlsson et al . , 2010 ) , or may represent a slight underestimation by the Brp-Short assay .","Using single optical sections within the DA1 glomerulus , we observed two levels of organization: ( 1 ) the location of neurites and Brp-Short puncta with respect to the glomerulus ( the subglomerular distribution ) , and ( 2 ) Brp-Short puncta with respect to their own neurites for each class of neurons ( the subcellular distribution ) .","Of the three classes of neurons observed , ORNs contribute the majority of the Brp-Short puncta ( 1632 ± 22 ) and were widely distributed within the glomerulus ( Figure 3A , D ) , though with distinct voids ( Figure 3D , asterisk ) where the glomerulus lacked ORN neurites and Brp-Short puncta .","PNs were the second major contributors of Brp-Short puncta within the glomerulus ( 1182 ± 24 ) .","These connections likely synapse onto PN and LN dendrites , as shown by electrophysiology ( Kazama and Wilson , 2009 ) , and are analogous to the dendrodendritic synapses between secondary dendrites of mitral cells and olfactory bulb interneurons in the vertebrate olfactory bulb ( Urban and Arevian , 2009 ) .","These Brp-Short puncta showed a more even distribution throughout the glomerulus ( Figure 3B , E ) , with smaller voids than those observed for ORNs despite reduced overall density .","Finally , local interneurons ( LNs ) contribute the fewest Brp-Short puncta ( 528 ± 54 ) of the three classes .","The LNs accessed by the NP3056-GAL4 line make up the majority of ipsilateral LNs including most or all pan-glomerular projections ( Chou et al . , 2010 ) , but may not be representative of all classes of LNs .","Within the glomerulus , LN distribution was the most restricted ( Figure 3C , F ) .","When examined at equivalent optical sections , these projections mostly filled the gaps in the ORN projection pattern , consistent with previously observed results for neurites ( Hummel and Zipursky , 2004 ) . 10 . 7554/eLife . 03726 . 011Figure 3 . Antennal lobe neurons display distinct subglomerular architecture .","( A–C )","Single , high-magnification confocal optical sections through the center of the DA1 glomerulus in three animals where ORNs ( A ) , PNs ( B ) , or LNs ( C ) are expressing Brp-Short .","A different color denotes the synapses of each class .","( D ) Single , high magnification confocal optical sections through the center of the DA1 glomerulus where ORNs express Brp-Short-mStraw ( red ) and mCD8-GFP ( green ) .","Brp-Short clusters ( arrowheads ) are visible , along with regions of neurite devoid of Brp-Short puncta ( arrow ) and voids where ORNs do not project ( asterisk ) .","( E ) Single , high magnification confocal optical sections through the center of the DA1 glomerulus where PNs are expressing Brp-Short-mStraw ( red ) and mCD8-GFP ( green ) .","Regions of PN neurite without Brp-Short puncta are visible ( arrow ) as well as small voids lacking Brp-Short puncta and neurite projections ( asterisk ) .","( F ) Single , high magnification confocal optical sections through the center of the DA1 glomerulus where LNs are expressing Brp-Short-mStraw ( red ) and mCD8-GFP ( green ) .","Brp-Short puncta are largely distributed evenly throughout neurites .","Images in panels D–F ( from three animals ) demonstrate distinct patterns of synapse localization and distribution within the glomerulus .","Dashed lines denote the glomerulus in ( C ) and ( F ) as defined by N-Cad staining ( not shown ) .","Scale bar = 5 μm .","( G–I )","Cumulative frequency histograms of the nearest neighbor distance between Brp-Short puncta in ORNs ( G ) , PNs ( H ) , and LNs ( I ) .","The average is indicated ( μ ) and the Cluster % of puncta with an NND ≤ 0 . 75 μm .","Gray traces represent individual glomeruli .","Red traces represent the aggregate average .","In all cases , n ≥ 10 animals , 19 antennal lobes , and 400 ( I ) , 800 ( H ) , or 1400 ( G ) individual puncta . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 01110 . 7554/eLife . 03726 . 012Figure 3—figure supplement 1 . Comparison of Brp-Short to endogenous Brp .","( A ) Representative high magnification confocal z-stack of the DA1 glomerulus in animals with an Or67d-mCD8-GFP transgene to mark its borders and stained with antibodies against GFP ( red ) and endogenous Brp ( green ) .","( B–D )","Representative high magnification confocal single sections of Or67d-positive ORNs ( B ) , Mz19-positive PNs ( C ) , and NP3056-positive LNs ( D ) innervating the DA1 glomerulus , each expressing Brp-Short-mStraw and stained with antibodies against mStraw ( red ) and endogenous Brp ( green ) .","Each image was taken from different animals at a roughly equivalent Z-plane .","In all cases , Brp-Short-mStraw puncta are coincident with endogenous Brp puncta .","Endogenous Brp puncta that are not Brp-Short-mStraw positive are likely from other neurons not expressing the Brp-Short-mStraw transgene .","Dashed lines indicate the glomerular boundary as defined by N-Cad staining ( not shown ) .","( D ) Quantification of Brp-Short puncta from ORNs ( green ) , PNs ( blue ) , and LNs ( magenta ) , the sum of ORNs + PNs + LNs , and the number of endogenous Brp puncta from the DA1 glomerulus ( black ) .","In all cases , data represent mean ± SEM and n ≥ 7 animals , 14 antennal lobes .","Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 012 With respect to their own neurites , Brp-Short puncta in ORNs were not evenly distributed: distinct sections of neurite lack Brp-Short puncta ( Figure 3D , arrow ) while other sections had large clusters ( Figure 3D , arrowheads ) of Brp-Short puncta ( Figure 3D ) .","To quantify this distribution , we conducted a nearest neighbor distance ( NND ) analysis for Brp-Short puncta in each class of neurons , which has been informative in understanding spatial relationships between synapses ( Bleckert et al . , 2013 ) .","We found that all three types of neurons had distinct mean NND values ( Figure 3G–I ) : ORNs had the closest puncta and LNs the furthest with PNs falling in between the two .","To quantify the clusters observed , we defined clustered puncta as having an NND of 0 . 75 μm or less ( an NND of 0 . 6 μm corresponds to directly touching ) .","For ORNs , these large clusters amounted to 17% of puncta that fell within this range ( Figure 3G ) .","PNs lacked these large clusters , but had some smaller clusters , as well as extended regions of neuronal membrane devoid of Brp-Short puncta ( Figure 3E ) .","Indeed , only 7 . 9% of PN puncta were clustered by this metric ( Figure 3H ) .","Within LN neurites , Brp-Short puncta also displayed significant clustering ( 15%; Figure 3I ) .","However , unlike ORNs and PNs , no extended regions of membrane were observed without synaptic puncta .","This suggests a more even distribution for the remaining puncta , which is supported by the higher value for NND ( Figure 3I ) and consistent with previous imaging using a synaptic vesicle marker ( Chou et al . , 2010 ) .","Overall , these results show distinct synaptic architecture , with respect to the glomerulus and the neurites themselves , made by three different types of olfactory neurons .","Moreover , distinct parameters can be defined for each class in terms of how puncta are clustered and the minimum distance between puncta .","Though some distances may be underestimated , these facets may represent additional rules that govern synaptic organization in the antennal lobe .","Having utilized Brp-Short to determine aspects of presynaptic organization in olfactory neurons ( Figures 1–3 ) , we next sought to investigate the molecular mechanisms that may underlie these rules .","We began with genetic perturbations of candidate molecules , the Teneurins .","Teneurins are type II transmembrane proteins involved in neuronal development ( Young and Leamey , 2009 ) and are disrupted in human neurological disorders ( Aldahmesh et al . , 2012; Psychiatric GWAS Consortium Bipolar Disorder Working Group , 2011 ) .","Recently , we demonstrated that the Drosophila Teneurins , Ten-m and Ten-a , regulate synaptic partner matching in the olfactory and neuromuscular systems via homophilic interaction between elevated Teneurin levels ( Hong et al . , 2012; Mosca et al . , 2012 ) .","Basal levels of the Teneurins interact heterophilically at NMJ synapses to organize connections ( Mosca et al . , 2012 ) .","However , due to the absence of techniques for quantitatively examining central synapses in vivo , a role for the Teneurins in organizing central synapses has not been established .","Due to the similarity in partner matching mechanisms between central and peripheral nervous systems , and the fact that both Ten-a and Ten-m are expressed at basal levels in all glomeruli ( Hong et al . , 2012 ) , we hypothesized that the CNS may use Teneurins similarly to the PNS for synaptic organization .","Specifically , we hypothesized that olfactory neurons would utilize basal levels of Ten-a and Ten-m to control synaptic number .","To test this hypothesis , we used Brp-Short to quantify active zone number in Or47b-positive ORNs innervating the VA1lm glomerulus .","Ten-a is expressed at a basal level in this glomerulus , and therefore does not directly affect synaptic partner matching ( Hong et al . , 2012 ) .","Comparing wild-type ( Figure 4A ) and ten-a null mutant flies ( Figure 4B ) , we found that Or47b ORNs in the ten-a mutant ( Figure 4B ) had 23% fewer Brp-Short puncta ( Figure 4D ) .","Notably , loss of ten-a did not alter ORN axon volume ( Figure 4E ) , despite affecting glomerular morphology ( Figure 4B ) , which is due to the loss of ten-a in neurons of the nearby DA1 glomerulus ( Hong et al . , 2012 ) .","As such , synaptic density was similarly reduced ( Figure 4F ) .","To ensure that changes in glomerular morphology were not related to changes in synapse number , we also examined ORNs that innervate the DL4 glomerulus ( Figure 4—figure supplement 1 ) .","In ten-a mutants , glomerular morphology was unchanged in 87% of animals examined , but all animals displayed a similar reduction in synapse number and density without affecting the axon volume ( Figure 4—figure supplement 1 , panels C–E ) .","Interestingly , at the NMJ , Teneurin perturbations reduced active zone and synaptic bouton numbers , but did not alter active zone density ( TM , unpublished observations ) .","However , in ORNs , neurite volume was unchanged , so active zone density was reduced , suggesting a difference in the specific consequence of Teneurin perturbation in the CNS compared with the NMJ . 10 . 7554/eLife . 03726 . 013Figure 4 . Presynaptic Ten-a is required for proper ORN synapse number and density . Representative high-magnification confocal stacks of Brp-Short puncta in VA1lm ORNs stained with antibodies against Brp-Short-mStraw ( red ) and N-Cadherin ( blue ) in control ( A ) , ten-a null mutants ( B ) and ten-a null mutants where a Ten-a transgene is expressed in this specific class of ORNs using Or47b-GAL4 ( C ) .","( D ) Quantification of Brp-Short puncta in the above conditions as well as following presynaptic RNAi against ten-a , presynaptic RNAi against ten-m , or in ten-a null mutants where a Ten-a transgene is expressed in 2/3 of the PNs , including those innervated by Or47b-positive ORNs .","( E ) Quantification of ORN neurite volume in the same genotypes .","n . s . = not significant .","( F ) Quantification of ORN synaptic density in the same genotypes .","Significance was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .","**p < 0 . 01 , ***p < 0 . 001 .","Unless otherwise noted , significance is compared to control .","In all cases , the ten-a mutant and ten-a ORN RNAi phenotypes are statistically indistinguishable and , in the ten-a mutant , presynaptic Ten-a expression rescues the observed phenotypes .","The disrupted glomerular shape is due to partner matching errors in the ten-a mutant ( Hong et al . , 2012 ) , which is not rescued due to the late onset of Or47b-GAL4 relative to the partner matching process .","Dashed lines denote the glomeruli as delineated by Brp-Short label .","In all cases , data represent mean ± SEM and n ≥ 7 animals , 14 antennal lobes .","Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 01310 . 7554/eLife . 03726 . 014Figure 4—figure supplement 1 . Representative images for additional genetic manipulations from Figure 4 . Representative high-magnification confocal stacks of Brp-Short puncta in VA1lm ORNs stained with antibodies against Brp-Short-mStraw ( red ) and N-Cadherin ( blue ) in control ( A ) , flies expressing ten-m RNAi in Or47b-positive ORNs ( B ) , flies expressing ten-a RNAi in Or47b-positive ORNs ( C ) , and ten-a null mutants where a Ten-a transgene is expressed in 2/3 of the PNs of the antennal lobe using GH146-QF ( D ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 01410 . 7554/eLife . 03726 . 015Figure 4—figure supplement 2 . ten-a phenotypes in DL4 , a glomerulus with no morphology defects .","( A ) Representative high magnification confocal stacks of ORNs innervating the DL4 glomerulus in control animals expressing Brp-Short-mStraw and mCD8-GFP and stained with antibodies to mStraw ( red ) , GFP ( green ) , and N-Cadherin ( blue ) .","( B ) Representative high magnification confocal z-stacks of ORNs innervating the DL4 glomerulus in ten-a null mutant animals expressing Brp-Short-mStraw and mCD8-GFP and stained with antibodies to mStraw ( red ) , GFP ( green ) , and N-Cadherin ( blue ) .","Morphology is unaffected but synapse number is reduced .","( C ) Quantification of Brp-Short-mStraw puncta in control and ten-a mutant animals .","( D ) Quantification of ORN volume in control and ten-a mutant animals .","( E ) Quantification of Brp-Short-mStraw puncta density in control and ten-a mutant animals .","In all cases , data represent mean ± SEM and n ≥ 7 animals , 14 antennal lobes .","***p < 0 . 0001 .","Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 015 As ten-a mutants lack Ten-a in ORNs and PNs , we next sought to determine where Ten-a was required to regulate normal synapse number using tissue-specific rescue and RNAi experiments .","Restoring Ten-a expression to the Or47b ORNs in the ten-a mutant ( Figure 4C ) , but not to the VA1lm PNs that are innervated by Or47b ORNs , rescued the reduction in synapse number and density ( Figure 4D , F ) .","When Ten-a was restored to Or47b-positive ORNs , defects in glomerular morphology persisted despite rescue of synaptic phenotypes ( compare Figure 4C–A ) , as partner matching occurs earlier than synaptogenesis ( and before Or47b-GAL4 turns on ) .","Further consistent with a presynaptic role for Ten-a , knockdown of Ten-a in ORNs by RNAi also reduced the number and density of synapses similarly to the mutant ( Figure 4D , F ) .","Here , glomerular morphology was unaffected as VA1lm is already a ‘Ten-a low’ glomerulus and does not directly require Ten-a for partner matching .","Interestingly , knockdown of ten-m in the Or47b ORNs did not result in any synaptic phenotype ( Figure 4D–F ) , demonstrating specificity for ten-a .","Thus , Ten-a is required presynaptically for proper synapse number and density , likely in a partially redundant fashion with additional molecules .","Previous data and our Brp-Short analyses suggest that the assay is largely reflective of actual synapse number in neurons .","However , we sought to confirm these results using a method independent of confocal microscopy or the Brp-Short label .","We compared synapse number in ultrathin EM sections of wild-type and ten-a mutant flies ( Figure 5A–B ) .","Quantification of all synapses in a glomerulus ( and thus a direct comparison to the Brp-Short assay ) would require serial reconstruction of that entire glomerulus by EM .","However , if the Brp-Short assay accurately reflects changes in synapse number following genetic perturbation , we reasoned that this should be evident in the change in the number of T-bar profiles in individual sections at the EM level .","Indeed , quantification of T-bar profiles showed a 27% reduction in the number of T-bars ( Figure 5I ) .","This is consistent with the results from the Brp-Short assay ( Figure 4D ) , which shows a 23% reduction .","The slight difference may reflect an under-report of the phenotype by our confocal-level analysis or may reflect the fact that the EM analysis does not differentiate between ORN , PN , and LN presynapses while the Brp-Short assay does .","The largely similar magnitude of synapse reduction in EM and confocal studies supports the validity of the Brp-Short assay , and further substantiates that Ten-a is required for proper synapse number in the CNS . 10 . 7554/eLife . 03726 . 016Figure 5 . Ultrastructural analysis of active zones in ten-a mutants . Representative transmission electron microscopy ( TEM ) images of active zones in control ( A ) and ten-a null mutant antennal lobes ( B ) .","T-bar profiles are visible in both genotypes , though reduced in number in the ten-a mutant .","Scale bar = 500 nm .","( C ) High magnification TEM of a single active zone in a control animal .","Note the normal T-bar morphology .","( D–H )","High magnification TEM images of active zones in ten-a mutants .","In some cases , normal ( D ) morphology is observed .","In the majority of cases , defects including multiple T-bars ( E ) , misshapen T-bars ( F ) , detached T-bars ( G ) , and continuous electron dense material ( H ) are evident .","Scale bar = 100 nm .","( I ) Quantification of T-bars per unit perimeter of measured antennal lobe terminals .","All terminals were measured , including ORNs , PNs , and LNs .","ten-a mutants display a 27% reduction from control animals .","Data represent mean ± SEM .","( J ) Distribution of T-bar defects as a percentage of the total T-bars .","Here , n = 1103 T-bar per unit perimeter measurements from three animals for control and n = 847 T-bar per unit perimeter measurements from three animals for ten-a . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 016 At neuromuscular synapses , Teneurins are required for proper T-bar structure .","When Teneurins are absent , T-bars display defects consistent with failed synaptogenesis ( Mosca et al . , 2012 ) .","To determine whether Ten-a was also required in olfactory neurons for T-bar structure , we examined individual T-bar morphology in ultrathin sections of wild-type and ten-a antennal lobes .","In wild-type animals , T-bar morphology was nearly invariant ( Figure 5C ) , with infrequent occurrences of misshapen or multiple T-bars ( Figure 5J ) .","In ten-a mutants , 44% of T-bar profiles appeared normal ( Figure 5D , J ) .","The remaining 56% displayed notable structural defects including continuous , multiple T-bars ( Figure 5E ) , misshapen T-bars ( Figure 5F ) , and T-bars that were detached from the synaptic membrane ( Figure 5G ) .","These defects resembled those observed at ten-a mutant NMJs ( Mosca et al . , 2012 ) .","We also observed continuous , electron-dense material resembling undivided T-bars ( Figure 5H ) that was not found previously .","Thus , Ten-a is required not just for the normal number of active zones , but also for their proper structure in ways both consistent with and beyond that of previous studies .","Specifically , ten-a may be required for the maturation of synapses , as misshapen and undivided T-bars resemble immature synapses ( Owald et al . , 2010 ) .","Teneurins can regulate cytoskeletal architecture ( Nunes et al . , 2005; Morck et al . , 2010; Suzuki et al . , 2014 ) and function at the Drosophila NMJ by organizing spectrin , a cytoskeletal component ( Mosca et al . , 2012 ) .","However , it is unknown whether this function of the Teneurins is conserved in the CNS .","To further probe the mechanism by which Teneurins regulate central synapse number and to determine whether cytoskeletal regulation is involved , we compared levels of spectrin staining in the antennal lobe between wild-type and ten-a mutant flies .","We found that spectrin immunofluorescence in the antennal lobe was reduced in the ten-a mutant compared to wild-type ( Figure 6A–B ) while N-Cadherin , a marker of synaptic neuropil , was unaffected ( Figure 6F ) .","This was observed for α- and β-spectrin ( Figure 6F , Figure 6—figure supplement 1 , panels A–D ) , which function as a heterotetramer ( Machnicka et al . , 2014 ) . 10 . 7554/eLife . 03726 . 017Figure 6 . Ten-a and spectrin function together for normal synapse number .","( A–B )","Representative single confocal optical sections taken at equivalent positions of antennal lobes stained for α-spectrin and N-Cadherin in control ( A ) and ten-a mutant ( B ) adults .","In ten-a mutants , α-spectrin staining is reduced compared to control .","The disrupted morphology of the antennal lobe itself is due to partner matching errors evident in the ten-a mutant ( Hong et al . , 2012 ) .","Scale bar = 20 μm .","( C–E )","Representative confocal Z-stack images of the ORNs in the VA1lm glomerulus expressing Brp-Short in control animals ( C ) , or animals expressing dsRNA against α-spectrin ( D ) or β-spectrin ( E ) , and stained with antibodies against Brp-Short ( red ) and N-Cadherin ( blue ) .","( F ) Quantification of α-spectrin ( green ) , β-spectrin ( red ) , and N-Cadherin ( blue ) immunofluorescence in control and ten-a mutants .","***p < 0 . 001 .","( G ) Quantification of Brp-Short puncta in the noted genotypes .","In all cases , similar reductions in puncta number are observed .","Moreover , the genetic perturbations do not enhance each other , suggesting function in the same pathway .","Significance was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .","***p < 0 . 001 ( compared with control ) .","In all cases , data represent mean ± SEM and n ≥ 10 animals , 19 antennal lobes .","Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 01710 . 7554/eLife . 03726 . 018Figure 6—figure supplement 1 . ten-a regulates spectrin levels and spectrin regulates synapse number . Representative confocal single optical sections of control ( A ) and ten-a mutant ( B ) antennal lobes stained with antibodies to β-spectrin ( red ) and N-Cadherin ( blue ) .","Note the reduction ( quantified in Figure 6F ) of staining in the ten-a mutant .","Scale bar = 20 μm .","( C–E )","Representative high magnification confocal z-stacks of Or67d ORNs expressing Brp-Short-mStraw in control animals ( C ) and animals co-expressing dsRNA against α-spectrin ( D ) or β-spectrin ( E ) and stained with antibodies against mStraw ( red ) and N-Cadherin ( blue ) .","Glomerular morphology is unaffected by spectrin dsRNA but synapse number is reduced .","Scale bar = 5 μm .","( F ) Quantification of Brp-Short puncta in the genotypes from C–E .","***p < 0 . 0001 .","In all cases , data represent mean ± SEM and n ≥ 10 animals , 19 antennal lobes . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 018 To test whether proper spectrin level is required for synapse number , we used ORN-specific knockdown of either α- or β-spectrin and measured Brp-Short puncta .","In both cases ( Figure 6D–E ) , RNAi in Or47b-positive ORNs resulted in a 17–18% reduction in Brp-Short puncta ( Figure 6G ) .","This effect , though slightly weaker , phenocopied the reduction seen in the ten-a mutant .","We observed a similar effect in another ORN class , the Or67d-positive neurons ( Figure 6—figure supplement 1 ) , further supporting a role for spectrin in regulating olfactory synapse number .","Moreover , the ten-a phenotype was not enhanced by concurrent RNAi knockdown of α-spectrin in ORNs ( Figure 6G ) , suggesting that Ten-a and α-spectrin function in the same pathway to regulate olfactory synapse number .","A functional synapse consists of a presynaptic neurotransmitter release site and a postsynaptic neurotransmitter receptor cluster ( Figure 1A ) .","Therefore , critical parameters of synaptic organization within a circuit not only include the location and number of presynaptic active zones , but also postsynaptic receptor clusters .","Therefore , we also examined the organization of postsynapses .","Given that ORNs are cholinergic ( Wilson , 2013 ) , an ideal labeling strategy would image postsynaptic acetylcholine receptors .","We selected the Dα7 acetylcholine receptor subunit because it is endogenously expressed in the antennal lobe ( Fayyazuddin et al . , 2006 ) and it has been used to examine organization in the mushroom body , a higher olfactory center ( Leiss et al . , 2009a , 2009b; Kremer et al . , 2010; Christiansen et al . , 2011 ) .","We used a GFP-tagged Dα7 transgene ( Leiss et al . , 2009b ) under the control of the GAL4/UAS system ( Figure 7A ) to visualize postsynapses in vivo .","Expression of Dα7-GFP in PNs revealed distinct puncta , possibly corresponding to acetylcholine receptor ( AChR ) clusters ( Figure 7B ) .","These puncta were apposed to endogenous Brp puncta , as revealed by nc82 staining ( Supplement 1 to Figure 7 ) , consistent with these puncta representing bona fide synapses .","To examine AChR clusters in PNs , we co-expressed Dα7-GFP with mtdT as a general neurite label ( Figure 7B ) .","As such , the approach is analogous to our Brp-Short assay and yielded similar results , enabling a quantitative assessment of the number ( Figure 7C ) and density ( Figure 7D ) of AChR clusters . 10 . 7554/eLife . 03726 . 019Figure 7 . Measuring postsynaptic acetylcholine receptor clusters with Dα7-GFP .","( A ) Diagram of the GFP-tagged Dα7 acetylcholine receptor subunit used for AChR visualization .","( B ) High magnification confocal z-stack images of PNs in the DA1 and VA1d glomeruli expressing Dα7-GFP and mtdT , stained with antibodies against GFP ( green ) , mtdT ( red ) , and endogenous Brp ( blue ) .","Patches of mtdT labeling represent ascending PN axons within the plane of the glomerulus .","Insets show a high magnification single optical section demonstrating the punctate nature of Dα7-GFP .","( C ) Representative high magnification single optical sections of Mz19-positive PNs in the DA1 glomerulus expressing Dα7-GFP and stained with antibodies against GFP ( green ) and endogenous Brp ( red ) .","The majority of GFP-positive puncta are apposed to or colocalized with endogenous Brp ( yellow ) , consistent with their association with bona fide active zones .","Insets show high magnification of a single optical section where asterisks denote apposed Dα7 puncta .","Brp puncta without apposition likely belong to synapses not labeled by the PN-GAL4 driver .","Scale bar = 5 μm ( 2 μm for inset ) .","( D ) Representative high magnification single optical sections of Mz19-positive PNs and Or67d-positive ORNs in the DA1 glomerulus expressing Dα7-GFP in the PNs and Brp-Short-mStraw in the ORNs using two binary expression systems .","Most ORN active zones ( labeled by Brp-Short , red ) are apposed to or colocalized ( yellow ) with PN Dα7 puncta ( green ) , further supporting that these puncta label bona fide synaptic contacts .","Insets show high magnification of a single optical section where asterisks denote apposed Brp-Short::Dα7 pairs .","Brp puncta without apposition likely belong to ORN synapses with neurons other than PNs and Dα7 puncta without apposition likely correspond to PN postsynaptic sites apposed to synaptic contacts from neurons other than ORNs .","( E ) Quantification of Dα7 AChR puncta ( green , left axis ) and neurite volume ( black , right axis ) in the DA1 and VA1d glomeruli of both male and female adult flies .","Statistical comparisons between males and females of a single genotype were done by student's t test .","***p < 0 . 001 .","( F ) Quantification of AChR density in male ( blue ) and female ( pink ) adults based on the data from ( C ) .","Significance was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .","*p < 0 . 05 .","n . s . = not significant .","In all cases , data represent mean ± SEM and n ≥ 11 animals , 21 antennal lobes .","Scale bar = 5 µm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 019 As we are limited to genetically accessible PN subsets , we focused on identifying organizational parameters in the PNs that innervate the DA1 and VA1d glomeruli via the Mz19-GAL4 driver ( Ito et al . , 1998 ) .","As with ORN presynapses , the assay revealed that the number of AChR puncta scales with glomerular size ( Figure 7E ) .","Further , known sex-specific differences in DA1 , as seen in glomerular volume ( Stockinger et al . , 2005 ) and in ORN synapses ( Figure 1 ) , were also observed ( Figure 7E ) .","The differences between the Brp-Short and AChR assays for the DA1 ( 1632 ± 22 for Brp-Short vs 2007 ± 48 for Dα7 ) and VA1d ( 1107 ± 18 for Brp-Short vs 1762 ± 38 for Dα7 ) glomeruli may reflect the fact that the Brp-Short assay does not distinguish ORN synapses onto PNs and LNs , and the Dα7-GFP assay does not distinguish synapses from ORNs and LNs onto PNs .","As these values are less than twofold different , this is consistent with the majority of synapses labeled being ORN to PN synapses ( Kazama and Wilson , 2009; Wilson , 2013 ) .","The similarity between the numbers of endogenous Brp and Brp-Short puncta suggests that Brp-Short is a more accurate estimator of absolute synapse number .","The larger number of AChRs detected in each glomerulus may reflect an overestimation associated with full-length Dα7 overexpression or that these are postsynaptic not just to cholinergic ORNs , but also other excitatory neurons such as local interneurons ( Chou et al . , 2010 ) or PN-PN chemical synapses ( Ng et al . , 2002; Wilson et al . , 2004 ) .","Calculation of AChR puncta density in PNs revealed subtle but significant differences across different glomeruli .","In the VA1d glomerulus , the densities were identical between males and females ( Figure 7F ) .","However , these were different from AChR puncta densities in the DA1 glomerulus .","There was a modest but significant difference between both male and female AChR densities in DA1 and between both DA1 AChR densities and the shared VA1d AChR density ( Figure 7F ) .","Unlike ORNs , where the Brp-Short density was identical across different classes of neurons , PNs can have different densities between distinct glomeruli and even between sexes for the same glomerulus .","Thus , the parameters that govern presynaptic density may differ from those that govern postsynaptic density in the same glomerulus .","To further examine if the Teneurins regulate postsynaptic acetylcholine receptor number and density , we utilized the Dα7-GFP assay to determine the effect of Teneurin perturbation on AChR puncta number .","We examined PNs in DA1 and VA1d ( Figure 8A ) and counted the AChR puncta of both glomeruli together as one measurement , as partner matching defects following Teneurin perturbation make it difficult to differentiate between the two glomeruli ( Hong et al . , 2012 ) .","In ten-a mutants ( Figure 8B ) , the number of AChR clusters in these glomeruli was decreased by 23% , compared to wild type ( Figure 8D ) .","This is consistent with results from the Brp-Short assay .","Moreover , PN neurite volume was unaffected ( Figure 8E ) , so AChR puncta density was similarly reduced ( Figure 8F ) .","Thus , two independent assays , both pre- and postsynaptic , show the same phenotypes , demonstrating a clear effect of ten-a loss on synapse organization in olfactory neurons . 10 . 7554/eLife . 03726 . 020Figure 8 . Teneurins function to regulate acetylcholine receptor number .","( A–C )","Representative high magnification confocal z-stack images of DA1 and VA1d PNs expressing Dα7-GFP in control animals ( A ) , ten-a mutants ( B ) , or animals expressing RNAi against ten-m in the same PNs ( C ) , and stained with antibodies against GFP ( green ) and N-Cadherin ( blue ) .","DA1 and VA1d glomeruli are outlined together ( white dashed line ) and were determined using the accompanying mtdT label ( not shown ) .","Due to partner matching defects in the ten-a mutant that prevent clear delineation between the two glomeruli , the combined count was compared across genotypes .","( D ) Quantification of Dα7-GFP puncta in the noted genotypes .","( E ) Quantification of PN neurite volume in all genotypes .","( F ) Quantification of AChR density in all genotypes .","All observed phenotypes are statistically indistinguishable .","Significance was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .","***p < 0 . 001 in comparison with control .","n . s . = not significant .","In all cases , data represent mean ± SEM and n ≥ 8 animals , 16 antennal lobes .","Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 020 At the NMJ , presynaptic Ten-a functions largely in a transsynaptic , heterophilic complex with postsynaptic Ten-m to regulate synapse organization ( Mosca et al . , 2012 ) .","Ten-a functions presynaptically in ORNs to ensure proper synapse number ( Figure 4 ) .","Thus , we hypothesized that the loss of Ten-m in postsynaptic PNs should result in a similar phenotype .","As the ten-m mutant is larval lethal ( Zheng et al . , 2011 ) , we expressed a previously validated transgenic RNAi line against ten-m ( Hong et al . , 2012; Mosca et al . , 2012 ) in Mz19 PNs and quantified AChR puncta number using the Dα7 assay ( Figure 8C ) .","ten-m knockdown phenocopied the ten-a phenotype ( Figure 8D ) .","As above , PN neurite volume was unaffected ( Figure 8E ) , leading to a concomitant decrease in AChR puncta density that also phenocopied the ten-a phenotype ( Figure 8F ) .","Further , this reduction was not enhanced by knocking down ten-m in PNs of a ten-a null mutant ( Figure 8D–F ) , suggesting that the two function in the same genetic pathway .","Though genetic evidence suggests that Ten-m functions with Ten-a transsynaptically to regulate synapse number , it could conceivably regulate AChR number independently of how Ten-a regulates active zone number .","Thus , we sought to more explicitly test a transsynaptic role of Ten-m in regulating proper active zone number .","Due to the availability of genetic access to both the presynaptic ORNs and the postsynaptic PNs , we examined the DA1 glomerulus .","Further , DA1 is a ten-m low glomerulus ( Hong et al . , 2012 ) , allowing us to specifically test the role of the basal level of Ten-m .","Using the GAL4 and QF systems , we simultaneously labeled Or67d-positive ORNs with QUAS Brp-Short-mStraw using Or67d-QF ( Liang et al . , 2013 ) and expressed transgenic RNAi against ten-m in DA1 PNs using Mz19-GAL4 ( Figure 9A ) .","Control animals displayed a similar number of Brp-Short puncta as the UAS construct ( Figure 9B , D ) , demonstrating the validity of this reagent .","Postsynaptic ten-m knockdown , however , reduced the number of presynaptic Brp-Short puncta by 20% ( Figure 9C , D ) .","This reduction is similar to the proportion by which ten-m RNAi reduced AChR number ( Figure 8D ) and also to the proportion by which loss of ten-a reduced Brp-Short number ( Figure 4D ) .","This suggests that Ten-m is the postsynaptic partner of Ten-a in regulating presynaptic active zone number , and that the presynaptic Ten-a/postsynaptic Ten-m interaction that regulates NMJ synapses is conserved in olfactory synapses . 10 . 7554/eLife . 03726 . 021Figure 9 . Ten-m regulates presynaptic active zone number transsynaptically .","( A ) Diagram of the antennal lobe with the DA1 glomerulus outlined showing the experimental design .","Or67d-positive ORN axon terminals that innervate DA1 are outlined in black; their active zones are labeled by Or67d-QF-driven QUAS-Brp-Short-mStraw .","Mz19-GAL4-positive PNs that project dendrites to DA1 express UAS-RNAi against ten-m in experimental animals .","( B–C )","Representative high magnification confocal z-stack images of DA1 ORNs expressing Brp-Short in control animals ( B ) or in animals concurrently expressing ten-m RNAi in the DA1 and VA1d PNs ( C ) and stained with antibodies against mStraw ( red ) and N-Cadherin ( blue ) .","( D ) Quantification of Brp-Short puncta in the noted genotypes .","Controls 1 and 2 represent Brp-Sh puncta assayed by GAL4/UAS ( as in Figure","1 ) and QF/QUAS binary system binary system , respectively , and are not significantly different , as assayed by student's t test ( n . s . ) .","Significance of ten-m PN RNAi was assayed by student's t test between it and Control 2 .","***p < 0 . 001 .","In all cases , data represent mean ± SEM and n ≥ 5 animals , 10 antennal lobes .","Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 021"],["In Drosophila , previous approaches to studying central synapses used tagged synaptic vesicle proteins ( Ito et al . , 1998; Robinson et al . , 2002 ) to reveal putative synaptic sites .","While consistent with synapses , they could also stain non-synaptic , trafficking vesicles .","We utilized a structural active zone component , Brp ( Wagh et al . , 2006 ) .","By expressing a truncated Brp transgene , Brp-Short ( Schmid et al . , 2008 ) , using the GAL4/UAS system , this approach can label synapses in any neurons with genetic access .","Brp-Short expression requires endogenous Brp for proper localization ( Figure 1—figure supplement 1 ) and does not alter function or morphology ( Fouquet et al . , 2009; Kremer et al . , 2010 ) .","Thus , it accurately reports endogenous active zones .","Recently , an elegant technique , STaR , was developed , which tags an additional , BAC-sourced copy of Brp with an epitope tag whose expression is conditional upon FLP-recombinase-mediated excision of an intervening stop codon ( Chen et al . , 2014 ) .","An important advantage of STaR is that Brp expression is controlled by its endogenous promoter , thus guarding against mislocalization of Brp or perturbation of synaptic function due to overexpression .","A caveat of Brp-Short is that it is controlled by GAL4 and thus the levels may be different from the endogenous level .","While Brp-Short overexpression does not interfere with synaptic function , care must be taken not to overexpress it to a level that saturates the active zone localization machinery when utilized in new cell types .","The advantages of Brp-Short over STaR are","( a ) that it does not require a cell type-specific FLP transgene , which is not as widely available as GAL4 lines , or a BAC-bearing copy of Brp , and","( b ) that it can be co-expressed with UAS-transgenes for rescue or RNAi for perturbation experiments , although STaR can also achieve this aspect by using a cell-type-specific GAL4 and an extra UAS-FLP transgene .","To examine putative postsynaptic acetylcholine receptor clusters , we use a GFP-tagged subunit , Dα7 , to study cholinergic synapses in the antennal lobe .","This transgene has been used to examine synaptic organization ( Leiss et al . , 2009a , 2009b; Kremer et al . , 2010; Christiansen et al . , 2011 ) .","Though false positives can be associated with full-length protein overexpression , our observation of endogenous Brp puncta apposed to these Dα7-GFP puncta suggests that these receptors are properly localized to endogenous synapses .","This assay complements the Brp-Short presynaptic assay and can be adapted for other tagged postsynaptic receptor transgenes .","Though considerable advances have been made in our understanding of the wiring specificity ( Hong and Luo , 2014 ) and physiological properties ( Wilson , 2013 ) of the Drosophila olfactory circuit in the antennal lobe , little is known about its synaptic organization .","As this system has emerged as a model neural network , a detailed mapping of synaptic organizational principles is integral towards advancing the study of circuit dynamics .","Indeed , the juxtaposition of distinct types of synapses between multiple neuronal classes in the antennal lobe provides a model to study complex synaptic interactions compared to a neuromuscular synapse that features only two synaptic partners: the motoneuron and the muscle .","We utilized the Brp-Short and Dα7 assays to probe how synapses in ORNs , PNs , and LNs are organized in the antennal lobe with respect to their number , density , and location .","This work offers the first comprehensive information on ( 1 ) the number of active zones made by each ORN within a glomerulus , ( 2 ) the stereotypy of synapse numbers between those individual ORNs , ( 3 ) the prominence of PN presynaptic inputs within a glomerulus , suggesting a robust feedback mechanism , and ( 4 ) the relatively small contribution of LN active zones to the antennal lobe circuit .","Our analyses suggest distinct rules that govern the synaptic organization of antennal lobe neurons .","ORNs , the primary input neurons of the olfactory system , are diverse in their olfactory receptor expression , ligand specificity , and glomerular targeting specificity ( Liang and Luo , 2010 ) ; we now show that they also differ in the absolute synapse number ( Figure 1 ) .","However , despite such differences , all five classes we examined ( DA1 , VA1d , VA1lm , DL4 , and DM6 ORNs ) have identical synaptic densities , suggesting that this represents a general rule for other ORN classes .","This may further suggest that the primary job of the ORN is to convey information from the environment into the system as faithfully as possible , that all information is treated equally at this level , and that weighting computations occur downstream in the brain ( Masse et al . , 2009; Wilson , 2013 ) .","Indeed , our analyses indicate that each ORN makes an equivalent , discrete number of synapses within a given glomerulus with little variation ( Figure 2 ) , further supporting this hypothesis .","Interestingly , the density of postsynaptic receptors differs between glomeruli and even within the same glomerulus between sexes ( Figure 7 ) .","This variation can be due to technical caveats , such as Mz19-GAL4 does not label all PNs that innervate DA1 and VA1d ( Jefferis et al . , 2001; Liang et al . , 2013 ) , or that Dα7-GFP clusters do not reflect the absolute number of AChRs .","As the relative numbers still show these differences , an interesting possibility suggested by these results is that postsynaptic PN AChRs already reflect a transformed olfactory representation compared to output synapses of ORNs .","The difference in PN AChR density as compared to the constant density of ORN active zones suggests that different classes of PNs may modulate how information is received by regulating the number of acetylcholine receptors .","This can thus contribute to the transformation of olfactory representation by antennal lobe neurons ( Wilson , 2013 ) .","Fine-scale analysis of synapse localization within projections of ORNs , LNs , and PNs ( Figure 3 ) suggested that these three types of neurons differ in their subglomerular organization .","While occupying the vast majority of territory throughout the entire glomerulus , ORN processes and synapses leave distinct voids .","A significant proportion of LNs form synapses in these voids , likely to other LNs or PNs .","However , there is also overlap between LN and ORN processes and synapses , consistent with physiologically characterized ORN → LN and LN → ORN synapses ( Olsen et al . , 2007 , 2010; Root et al . , 2007; Kazama and Wilson , 2008 , 2009; Olsen and Wilson , 2008; Root et al . , 2008; Huang et al . , 2010; Yaksi and Wilson , 2010; Root et al . , 2011; Wilson , 2011 , 2013 ) .","Within their respective neurites , ORNs and PNs display uneven distributions and synaptic clusters to varying degrees while active zones in LNs are more evenly distributed throughout their processes .","These characterizations contribute to the growing repertoire of studies seeking to understand synaptic organization in Drosophila olfactory circuits ( Murthy et al . , 2008; Caron et al . , 2013; Fisek and Wilson , 2014 ) , adding synapse-level imaging to physiological techniques .","Finally , the existence of synaptic organization parameters detailing number and location suggests molecular mechanisms designed to enforce those rules , both at cellular and circuit levels .","Indeed , such analysis represents an integral part of neuronal circuit analysis , as recent work on retinal neurons has shown that accounting for synapse position is a critical aspect of modeling connectivity ( Kim et al . , 2014 ) .","Our data demonstrate that the Teneurins , a family of transsynaptic adhesion molecules , regulate one of these synaptic organizational paradigms: synapse number .","Beyond molecules like RPTPs ( Takahashi and Craig , 2013 ) and Wnts ( Park and Shen , 2012 ) , there is little known conservation between the organization mechanisms of central synapses and the NMJ .","In vertebrates , previous studies have identified a number of synaptogenic signaling and cell adhesion molecules in the CNS ( Gerrow and El-Husseini , 2006; Missler et al . , 2012 ) , but in many cases , their roles at the NMJ are either minimal or unknown .","Likewise , the roles of pathways including Rapsyn , Dok7 , MuSK , and Tid1 , which are well established at the NMJ ( Wu et al . , 2010 ) , have no well-established roles at CNS synapses .","Among the identified central synaptogenic molecules , no master controller of synapses , like Agrin , has been discovered ( Shen and Scheiffele , 2010 ) .","In the mammalian CNS , synaptic adhesion molecules like Neurexin and Neuroligin have demonstrated organizational roles ( Sudhof , 2008 ) but their role ( if any ) at the NMJ is largely unknown .","In Drosophila , considerable work has been done at the NMJ to understand synapse formation and organization ( Menon et al . , 2013 ) .","For example , neuromuscular Neurexin and Neuroligin regulate synaptic development and assembly ( Li et al . , 2007; Banovic et al . , 2010; Sun et al . , 2011; Chen et al . , 2012; Mosca et al . , 2012; Owald et al . , 2012 ) , but remain ( as with many other identified molecules ) largely untested in the CNS due to the absence of techniques for doing so .","Our approaches have now enabled such an examination , uncovering a strongly conserved synaptic organization function of the Teneurins from PNS to CNS .","Recent work has shown that these evolutionarily conserved proteins are involved in synaptic partner matching between neurons in the Drosophila olfactory system and between muscles and motoneurons at the NMJ ( Hong et al . , 2012; Mosca et al . , 2012 ) .","As there are marked differences between central and peripheral synapses ( like the NMJ ) , it is further unclear whether the mechanisms would be conserved or if they would be wholly different .","Here we find , as assayed by Brp-Short ( Figure 4 ) , ultrastructural ( Figure 5 ) , and Dα7 ( Figure 8 ) analyses , that Teneurins in olfactory neurons are required for normal synapse number .","Teneurin perturbation also reduces synaptic density , a parameter that is highly invariable for ORNs under normal conditions ( Figure 1I ) .","We determined that presynaptic Ten-a in ORNs ( Figure 4 ) likely functions with postsynaptic Ten-m in PNs ( Figures 8 , 9 ) to regulate levels of the spectrin cytoskeleton ( Figure 6 ) .","Our evidence is consistent with spectrin and the Teneurins functioning in the same genetic pathway to regulate synapse organization and density .","The perturbation of active zone number in ORNs by knocking down ten-m in PNs further suggests that Teneurins regulate synapse number and organization in the CNS via a transsynaptic mechanism .","This highlights further conservation between central and peripheral synapse organization in the use of Teneurins .","There are , however , some differences between the CNS and the PNS regarding the Teneurins .","While presynaptic ten-m has a minor role in synaptic organization at the NMJ ( Mosca et al . , 2012 ) , our data suggests the lack of such a role in the CNS ( Figure 4D ) .","Thus , these different systems may also use the Teneurins differently .","Mammalian Teneurins organize the visual system ( Leamey et al . , 2007; Suzuki et al . , 2012; Antinucci et al . , 2013; Young et al . , 2013 ) and Ten-2 can serve as a ligand for Latrophilin and localize to synapses in cultured neurons ( Silva et al . , 2011; Boucard et al . , 2014 ) .","However , as proper synaptic function is impaired in many neuropsychiatric disorders ( Guilmatre et al . , 2014 ) and human Teneurin-4 is associated with increased susceptibility to bipolar disorder ( Psychiatric GWAS Consortium Bipolar Disorder Working Group , 2011 ) , understanding how the Teneurins regulate central synapses is a question with clinical relevance .","Studies in vivo will be important to determine how mammalian Teneurins regulate synaptic organization and whether the different Teneurins can have specific roles at synapses .","In summary , our results demonstrate a role for the Teneurins in regulating the number of central synapses and highlight mechanistic conservation between peripheral and central synapse formation .","Moreover , the fact that ORNs can be mistargeted but still have the correct number of synapses suggests that target choice and synapse organization can be biologically separable , even when they employ the same molecules ."],["The following GAL4 driver lines were used to restrict expression to individual classes of neurons: AM29-GAL4 ( DL4 and DM6 ORNs; Endo et al . , 2007 ) , Or47b-GAL4 ( VA1lm ORNs; Vosshall et al . , 2000 ) , Or67d-GAL4 ( DA1 ORNs; Kurtovic et al . , 2007 ) , Or88a-GAL4 ( VA1d ORNs; Vosshall et al . , 2000 ) , Mz19-GAL4 ( DA1 , DC3 , and VA1d PNs; Jefferis et al . , 2004 ) , Pebbled-GAL4 ( all ORNs; Sweeney et al . , 2007 ) , LN5 ( NP3056 ) -GAL4 ( ∼50 pan-glomerular LNs; Chou et al . , 2010 ) .","The GH146-QF line ( Potter et al . , 2010 ) was used to genetically access 2/3 of all PNs .","The following transgenes were used for labeling or genetic manipulation: UAS-Brp-Short-mStraw ( Fouquet et al . , 2009 ) , UAS-Brp-Short-EGFP ( Schmid et al . , 2008 ) , UAS-Dα7-GFP ( Leiss et al . , 2009b ) , UAS-3xHA-mtdT ( Potter et al . , 2010 ) , UAS-mCD8-GFP ( Lee and Luo , 1999 ) , UAS-DSyd1-EGFP ( Owald et al . , 2010 ) , UAS-α-spectrin-dsRNA ( Pielage et al . , 2005 ) , UAS-β-spectrin-dsRNA ( Pielage et al . , 2005 ) , UAS-Ten-a ( Mosca et al . , 2012 ) , UAS-Brp-IR-104422 ( Vienna Drosophila RNAi Center ) , UAS-Dcr2 ( Dietzl et al . , 2007 ) , UAS-ten-a-RNAi-V32482 , UAS-ten-m-RNAi-V51173 ( Hong et al . , 2012; Mosca et al . , 2012 ) , QUAS-mCD8-GFP ( Potter et al . , 2010 ) , QUAS-Ten-a ( Hong et al . , 2012 ) , and QUAS-Brp-Short-mStraw ( see below ) .","The Df ( X ) ten-a line was used as a ten-a null mutant ( Hong et al . , 2012 ) .","Brp-D3-mStraw ( Brp-Short-mStraw ) was amplified by PCR from the pTWmStraw-Brp-D3 vector ( Owald et al . , 2010; a kind gift of Stephan Sigrist ) using the forward primer 5ʹ-CACCATGGGAACTAGTGAC-3ʹ and the reverse primer 5ʹ-CTAGCTTACGTCACG-3ʹ with a CACC appended to the 5ʹ end of the forward primer for subcloning into the Gateway ( Invitrogen , Carlsbad , CA ) pENTR/D-TOPO vector .","Following a BP reaction to produce pENTR-Brp-Short-mStraw , this plasmid was recombined into the destination vector pQUAST-attB-Gateway ( Mosca et al . , 2012; a kind gift of Vincenzo Favaloro ) using LR clonase .","The resulting pQUAST-attB-Brp-Short-mStraw vector was transformed into the ΦC31 landing site 86Fb on the third chromosome using standard methods .","hsFLP MARCM analyses were conducted as described ( Lee and Luo , 1999; Joo et al . , 2013 ) with the following modification: to obtain small DL4 and DM6 ORN clones , animals were heat-shocked between 48 hr and 72 hr after egg-laying at 37°C for 20 min .","Adult brains were dissected at 10 days post-eclosion and fixed , stained , and processed as previously described ( Wu and Luo , 2006 ) .","The following antibodies were used: rat anti-N-Cadherin [DN-EX#8; 1:40 , Developmental Studies Hybridoma Bank ( DSHB ) ] , mouse anti-Brp ( NC82; 1:40 , DSHB ) , rabbit anti-dsRed ( Living Colors DsRed Polyclonal Antibody , 1:250 , Clontech ) , chicken anti-GFP ( GFP-1020; 1:1000 , Aves Lab ) , mouse anti-α-spectrin ( 3A9; 1:50 , DSHB ) , rabbit anti-β-spectrin ( 1:500 , gift of R Dubreuil ) .","Alexa488- , Alexa568- , and Alexa647-conjugated secondary antibodies were used at a concentration of 1:250 ( Invitrogen , Carlsbad , CA ) .","Samples were mounted in SlowFade Gold Antifade medium ( Life Technologies , Grand Island , NY ) by bridge mount under #1 . 5 cover slips .","Images were obtained on a Zeiss LSM510 Meta confocal laser-scanning microscope ( Carl Zeiss , Oberkochen , Germany ) using a 63× 1 . 4 NA PlanApo lens at an optical zoom of 3× .","Images were centered on the glomerulus in question and the Z-boundaries set according to the confines of the synaptic label ( Brp-Short-mStraw or Dα7-GFP ) .","System gain was set to minimize saturation but maintain a high signal-to-noise ratio .","Selected raw image stacks are available at http://web . stanford . edu/group/luolab/ALsynapse . shtml .","Raw image stacks for quantification were first deconvolved using AutoQuant X3 software ( Media Cybernetics , Rockville , MD ) on a custom-built computer ( Digital Storm , Fremont , CA ) .","Specifically , the images were deconvolved using 10 iterations with settings of medium noise , and a blind point spread function .","Following deconvolution , images were visually inspected to ensure that striping , ringing , or discontinuity artifacts were not introduced .","Further , each image was visually examined to ensure that the borders of puncta were sharpened by the deconvolution and that the membrane staining remained continuous .","Following , images were imported into Imaris 7 . 4 . 1 ( Bitplane AG , South Windsor , CT ) for quantification .","Though deconvolution improved the visual quality of the images , it was dispensable for accurate quantification of Brp-Short or Dα7 puncta .","The ‘Spots’ function was used to quantify Brp-Short or Dα7 puncta in Imaris .","A region of interest was drawn around the glomerulus being quantified .","For individually labeled glomeruli , the standard Imaris box was sufficient to mark the region of interest .","Where multiple glomeruli were labeled and a single glomerulus was quantified , a freehand region of interest was drawn in each section using the neuropil staining as a guideline and a 3D region extrapolated by Imaris .","The spot size ( puncta diameter ) was set at 0 . 6 μm for Brp-Short and 0 . 8 μm for Dα7 , as determined by direct measurement of images and previous studies ( Kremer et al . , 2010 ) .","Background subtraction was selected .","Following automatic detection of puncta , the threshold was manually adjusted ( usually within 5% among different samples ) with visual inspection to ensure that all puncta were identified , minimal background staining or noise was recognized , and that individual puncta were not double-counted .","In all , the resultant count was recorded as the number of puncta .","The accuracy was assessed by comparing puncta detection figures to glomeruli that were counted by hand following 3D projection .","In all cases , there was less than 5% difference , supporting the accuracy of the semi-automatic quantification method .","In male Or67d-GAL4 samples that did not express the synaptic label ( normally have ∼1500 puncta ) , but were processed identically using primary and secondary antibodies , the number of puncta recognized by this method was <30 , suggesting that the ‘false positive’ rate was low .","To ensure that user-introduced noise in threshold adjustment did not introduce significant bias , each image was quantified a second time , at a later time , without access to the prior quantification .","If the difference exceeded 5% , the sample was discarded .","Further , if the averages of an entire genotype were significantly different from each other ( as determined by student's t test ) , the entire cohort was discarded and the experiment repeated .","For the nearest neighbor distance ( NND ) analysis , the ‘DistMin’ was calculated in Imaris using the ‘Spots to Spots Closest Distance XTension’ plug-in for Matlab .","Cumulative frequency histograms were compiled in Prism 6 . 04 ( GraphPad Software , San Diego , CA ) .","As each puncta is ∼0 . 6 μm in diameter , the minimum NND should be 0 . 6 .","Indeed , few measurements were seen below 0 . 6 .","Thus , ‘clustering’ was defined as puncta with an NND of less than 25% of the diameter .","The ‘% Cluster’ was calculated by expressing the number of puncta with an NND value ≤0 . 75 divided by the total number of puncta .","To obtain neurite volume based on mCD8-GFP or mtdT labeling , the ‘Surfaces’ function was used .","Regions of interest were drawn as above .","Background subtraction was set to 3 μm and the smoothing to 0 . 2 μm; these settings optimally reflected actual neurite labeling .","Automatic detection with a 10e5 threshold was then used to remove background , detecting only the glomerulus in question .","For smaller objects ( MARCM clones ) , a 10e1 threshold was used and only the specific objects that directly corresponded to mCD8-GFP labeling were selected and counted .","Volume was then calculated by Imaris .","Density was then expressed as the number of Brp-Short puncta divided by the neurite volume .","For images involving a neurite co-stain , the Brp-Short-mStraw channel was masked in Imaris so that only the puncta within the neurite were identified and counted ( not background puncta resulting from secondary antibodies , or image noise ) .","Images were processed using Photoshop CS4 ( Adobe , San Jose , CA ) and levels or contrast adjusted identically for all cases of comparison .","For cases of immunofluorescence comparison , brains were stained in parallel , imaged on the same slide in the same session , and processed identically .","Statistical and graphical analyses were completed in Excel ( Microsoft , Redmond , WA ) , Prism 6 . 04 ( GraphPad Software , San Diego , CA ) , and Matlab ( Mathworks , Natick , MA ) .","Adult brains were dissected in 0 . 1 M cacodylate buffer and fixed at 4°C for 3 hr in 2 . 5% paraformaldehyde , 5% glutaraldehyde ( vol/vol ) , and 0 . 06% picric acid ( vol/vol ) in 0 . 1 M cacodylate buffer .","They were then washed three times for 20 min each in 0 . 1 M cacodylate buffer on ice and post-fixed with 2% osmium tetroxide for 1 hr at 4°C .","Dehydration , infiltration with resin , and staining with uranyl acetate were all done according to standard protocols .","Antennal lobe regions were identified in thick section following 1% toluidine blue in 1% sodium borate staining and ultrathin sections taken from those regions .","Sections were imaged on a JEM-1400 ( JEOL , Tokyo , Japan ) transmission electron microscope at X3 , 000 to X20 , 000 magnification .","For immuno-EM , adult brains were dissected in 0 . 1 M sodium phosphate buffer and fixed in modified Zamboni's fixative ( 4% paraformaldehyde , 1 . 6% glutaraldehyde , 0 . 2% saturated picric acid in 0 . 1 M sodium phosphate buffer at pH 7 . 4 ) as described ( Kolodziejczyk et al . , 2008 ) .","Brains were then washed in PBS , processed similarly as above , and sectioned .","Sections were then blocked in PBST containing 0 . 05% Tween-20 , 0 . 5% ( wt/vol ) ovalbumin , and 0 . 5% ( wt/vol ) BSA .","Sections were incubated overnight at 4°C in rabbit anti-GFP at 1:300 ( MBL International , Woburn , MA ) .","Following , grids were washed in PBST and incubated in 5 nm gold-conjugated goat anti-rabbit secondary antibody ( Ted Pella , Redding , CA ) at 1:50 for 1 hr at RT .","Grids were then washed in PBST , incubated in 8% glutaraldehyde for 15 min , washed again in water , and contrast stained with 2% uranyl acetate followed by lead citrate .","Ultrastructural analysis was done as previously described ( Watts et al . , 2004; Mosca et al . , 2012 ) and T-bar defects quantified using consecutive sections .","For pairwise comparisons , student's t test was used to assess statistical significance .","For multiple comparisons , a one-way ANOVA was done and corrected for multiple comparisons with posthoc Tukey's multiple comparisons tests done between all genotypes .","Analysis was done in Prism 6 . 0 . 4 ( Graphpad Software , San Diego , CA ) .","For synapse contribution by individual ORNs , GraphPad Prism was used to fit individual Gaussian relationships to each separate peak for the ORN MARCM data ( Figure 2C–D , colored dashed lines ) .","The Curve Fitting Tool in Matlab ( Mathworks , Natick , MA ) was used to determine the best-fit curves that corresponded to the aggregate data set ( Figure 2C–D , solid black line ) .","In each case , multiple strong candidate relationships that recognized the first four peaks were identified using r2 and root mean square error ( RMSE ) values .","To determine which candidate best fit the data , they were then compared using two posthoc tests: corrected Akaike's Information Criteria ( AICc ) and the F-test .","AICc is advantageous as it penalizes distributions with more parameters , avoiding potential over-fitting , and is useful with binned data .","The F-test was used to test specific null hypotheses between the models .","To examine the distribution of the individual peaks ( i . e . , the ‘quantal’ nature ) , the standard deviation in the distances between the peaks ( S ) was calculated .","To determine the probability of such an S value occurring , the aggregate DL4 and DM6 data were individually fit to a Poisson distribution or a Gaussian distribution using Matlab .","Matlab was then used to generate 50 , 000 simulations selecting four peaks that fit those distributions .","Also , a uniform probability distribution was used to select four peaks between 0 and the upper bound of the DL4 and DM6 data sets , respectively .","In all , 300 , 000 simulations were conducted , and the number of events resulting in an S value less than that observed used to determine a probability of obtaining a more ordered set of four peaks than the observed data .","In some cases , UAS-mCD8-GFP or UAS-3xHA-mtdT was included in the genotype ( but not shown in the image ) to ensure an equivalent number of constructs between control and mutant or rescue conditions .","Figure 1: ( B–D ) w; UAS-Brp-Short-mStraw/UAS-mCD8-GFP; Or67d-GAL4 [Nr-1]/+; + .","( H–I )","DL4 and DM6 = w; AM29-GAL4/UAS-Brp-Short-mStraw , UAS-mCD8-GFP; +; + .","DA1 = w; UAS-Brp-Short-mStraw/UAS-mCD8-GFP; Or67d-GAL4 [Nr-1]/+; + .","VA1d = w , Or88a-GAL4/+ or Y; UAS-mCD8-GFP , UAS-Brp-Short-mStraw/+; +; + .","VA1lm = w; Or47b-GAL4/UAS-Brp-Short-mStraw , UAS-mCD8-GFP; +; + .","Figure 1—figure supplement 1: ( A ) w; UAS-Brp-Short-mStraw/+; Or67d-GAL4 [Nr-1]/UAS-DSyd1-EGFP; + .","( B ) w , UAS-Dcr2; UAS-mCD8-GFP/UAS-Brp-Short-mStraw; Or67d-GAL4 [Nr-1]/+; + .","( C ) w; UAS-mCD8-GFP , UAS-Brp-Short-mStraw/UAS-Brp-IR-v104422; Or67d-GAL4 [Nr-1]/+; + .","( D ) w , pebbled-GAL4/+; +; UAS-Brp-Short-GFP/+; + .","Figure 1—figure supplement 2: ( A ) w , Or88a-GAL4/+ or Y; UAS-Brp-Short-mStraw/+; +; + .","( B ) w; Or47b-GAL4/UAS-Brp-Short-mStraw; +; + .","( C and D ) w; AM29-GAL4/UAS-Brp-Short-mStraw; +; + .","Figure 2: ( A–F ) hsFLP122 , UAS-mCD8-GFP; AM29-GAL4/UAS-Brp-Short-mStraw; FRT 2A/FRT2A , tubP-GAL80; + .","Figure 2—figure supplement 1: ( A–L ) hsFLP122 , UAS-mCD8-GFP; AM29-GAL4/UAS-Brp-Short-mStraw; FRT 2A/FRT2A , tubP-GAL80; + .","Figure 3: ( A ) w; UAS-Brp-Short-mStraw/+; Or67d-GAL4 [Nr-1]/+; + .","( B ) w; Mz19-GAL4/UAS-Brp-Short-mStraw; +; + .","( C ) w; UAS-Brp-Short-mStraw/+; NP3056-GAL4/+; + .","( D , G ) w; UAS-Brp-Short-mStraw/UAS-mCD8-GFP; Or67d-GAL4 [Nr-1]/+; + .","( E , H ) w; Mz19-GAL4 , UAS-mCD8-GFP/UAS-Brp-Short-mStraw; +; + .","( F , I ) w; UAS-Brp-Short-mStraw/UAS-mCD8-GFP; NP3056-GAL4/+; + .","Figure 3—figure supplement 1: ( A ) w; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/+; Or67d-GAL4 [Nr-1]/+; + .","( B ) w; Mz19-GAL4/UAS-Brp-Short-mStraw , UAS-mCD8-GFP; +; + .","( C ) w; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/+; NP3056-GAL4/+; + .","Figure 4: ( A ) w; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-mCD8-GFP/+; + .","( B ) w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; +; + .","( C ) w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-Ten-a/+; + .","( D–F )","Control as in ( A ) .","ten-a as in ( B ) .","ten-a ORN RNAi = w , UAS-Dcr2; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-ten-a-IR-v32482/+; + .","ten-a + Ten-a ORN as in C . ten-m ORN RNAi = w , UAS-Dcr2; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-ten-m-RNAi-V51173/+; + .","ten-a + Ten-a PN = w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; GH146-QF , QUAS-mCD8-GFP/QUAS-Ten-a; + .","Figure 4—figure supplement 1: ( A ) w; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-mCD8-GFP/+; + .","( B ) w , UAS-Dcr2; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-ten-m-RNAi-V51173/+; + .","( C ) w , UAS-Dcr2; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-ten-a-IR-v32482/+; + .","( D ) w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; GH146-QF , QUAS-mCD8-GFP/QUAS-Ten-a; + .","Figure 4—figure supplement 2: ( A–C ) w; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/AM29-GAL4; +; + .","( D–F ) w , Df ( X ) ten-a; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/AM29-GAL4; +; + .","Figure 5: Control = Canton S . ten-a = Df ( X ) ten-a; +; +; + .","Figure 6: ( A ) Canton S . ( B ) Df ( X ) ten-a; +; +; + .","( C ) w; Or47b-GAL4 , UAS-Brp-Short-mStraw/UAS-mCD8-GFP; UAS-mCD8-GFP/+; + .","( D ) w; Or47b-GAL4 , UAS-Brp-Short-mStraw/UAS-α-spectrin-dsRNA; UAS-α-spectrin-dsRNA/+; + .","( E ) w; Or47b-GAL4 , UAS-Brp-Short-mStraw/UAS-β-spectrin-dsRNA; UAS-β-spectrin-dsRNA/+; + .","( F ) Control as in ( A ) and ten-a as in ( B ) .","( G ) Control as in ( C ) , α-spec ORN RNAi as in ( D ) , β-spec ORN RNAi as in ( E ) , ten-a = w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; +; + .","ten-a + α-spec RNAi = w , Df ( X ) ten-a; Or47b-GAL4 , UAS-Brp-Short-mStraw/UAS-α-spectrin-dsRNA; UAS-α-spectrin-dsRNA/+; + .","Figure 6—figure supplement 1: ( A ) Canton S . ( B ) Df ( X ) ten-a; +; +; + .","( C ) w; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/+; Or67d-GAL4 [Nr-1] , UAS-mCD8-GFP/+; + .","( D ) w; UAS-Brp-Short-mStraw/UAS-α-spectrin-dsRNA; Or67d-GAL4 [Nr-1]/UAS-α-spectrin-dsRNA; + .","( E ) w; UAS-Brp-Short-mStraw/UAS-β-spectrin-dsRNA; Or67d-GAL4 [Nr-1]/UAS-β-spectrin-dsRNA; + .","( F ) Genotypes as in ( C ) , ( D ) , and ( E ) .","Figure 7: ( B , E–F ) w; Mz19-GAL4 , UAS-3xHA-mtdT/+; UAS-Dα7-EGFP/+; + .","( C ) w; Mz19-GAL4/+; UAS-Dα7-GFP/+; + .","( D ) w , Or67d-QF; Mz19-GAL4/+; UAS-Dα7-GFP/QUAS-Brp-Short-mStraw; + .","Figure 8: ( A ) w , UAS-Dcr2; Mz19-GAL4 , UAS-3xHA-mtdT/+; UAS-Dα7-GFP/+; + .","( B ) w , Df ( X ) ten-a; Mz19-GAL4 , UAS-3xHA-mtdT/+; UAS-Dα7-GFP/+; + .","( C ) w , UAS-Dcr2; Mz19-GAL4 , UAS-3xHA-mtdT/+ , UAS-Dα7-GFP/UAS-ten-m-IR; + .","( D–F )","Control as in ( A ) .","ten-a as in ( B ) .","ten-m PN RNAi as in C . ten-a −/− + ten-m PN RNAi = w , Df ( X ) ten-a; Mz19-GAL4 , UAS-3xHA-mtdT/UAS-Dcr2; UAS-Dα7-GFP/UAS-ten-m-IR; + .","Figure 9: ( B ) w , Or67d-QF; +; QUAS-Brp-Short-mStraw/+; + .","( C ) w , Or67d-QF; Mz19-GAL4/+; QUAS-Brp-Short-mStraw/UAS-ten-m-RNAi-V51173; + .","( D ) Control ( UAS ) = w; UAS-Brp-Short-mStraw/+; Or67d-GAL4 [Nr-1]/+; + .","Other genotypes as in B–C ."]],"string":"[\n [\n \"Understanding how neural circuits process information requires knowing not only the neuroanatomical connectivity and electrophysiological properties of individual neurons , but also the organization of synapses between specific neuronal types .\",\n \"Knowledge of these basic organizational principles , as well as the mechanisms for how they are achieved in a complex circuit , represents a critical goal for neuroscience .\",\n \"Historically , neuromuscular junctions have been model synapses for mechanistic studies because of their ease of access ( Keshishian et al . , 1996; Sanes and Lichtman , 1999 ) .\",\n \"However , neuromuscular circuits display less complexity compared to neural circuits in the central nervous system ( CNS ) .\",\n \"Serial-section electron microscopy ( EM ) has offered remarkable resolution in studying CNS synapses in the context of circuits , including the entire Caenorhabditis elegans nervous system ( White et al . , 1986 ) , the mouse olfactory bulb ( Hinds and Hinds , 1976a , 1976b ) , retina ( Briggman et al . , 2011; Helmstaedter et al . , 2013 ) , and visual cortex ( Bock et al . , 2011 ) , and the Drosophila visual system ( Meinertzhagen and O'Neil , 1991; Takemura et al . , 2013 ) .\",\n \"However , these approaches are labor intensive and not easily amenable to analysis following genetic perturbation , at varying developmental stages , or with intact samples .\",\n \"Indeed , even for well-known synaptic organizers like Neurexin and Neuroligin ( Chia et al . , 2013 ) , studies investigating their function were rarely carried out in an intact system with a resolution sufficient to reveal their detailed functions in synaptogenesis between defined neuronal types .\",\n \"As such , there is a need to examine CNS synapses in the intact brain , at the level of light microscopy in a system with sufficient complexity to reflect features of the CNS , but tractable enough to dissect connectivity and function using genetic tools .\",\n \"Current work in Drosophila has sought to fulfill this need ( Kremer et al . , 2010; Christiansen et al . , 2011; Berger-Muller et al . , 2013; Chen et al . , 2014 ) .\",\n \"The Drosophila olfactory system contains complex circuits that transform chemosensory input to the regulation of behavioral output .\",\n \"Olfactory information is received by first-order olfactory receptor neurons ( ORNs ) and is transmitted by ORN axons to the antennal lobe , the first olfactory processing center in the brain .\",\n \"There , the information is conveyed to the dendrites of second-order projection neurons ( PNs ) through class-specific ORN–PN synapses .\",\n \"PN axons carry signals to higher-order brain centers ( Vosshall and Stocker , 2007 ) .\",\n \"The local interneurons ( LNs ) innervate the antennal lobe extensively with their dendrites ( Chou et al . , 2010 ) and are involved in transforming ORN input to PN output ( Wilson , 2013 ) .\",\n \"These three groups of neurons have extensive interconnectivity and indeed , the antennal lobe has emerged as a model neural circuit for investigating the general principles of information processing ( Wang , 2012; Wilson , 2013; Twick et al . , 2014 ) .\",\n \"While many antennal lobe neurons have been examined electrophysiologically ( Wilson , 2013 ) , there has been little analysis of their synaptic architecture: how many connections are made , where they are made in relation to other neurons in the circuit , and what molecules are required for their organization .\",\n \"Electron microscopy has been used to probe different genetic conditions ( Acebes and Ferrus , 2001; Acebes et al . , 2011 ) but could not definitively assign glomerular identity or cell type for individual synapses , highlighting the need for structural analysis at the level of light microscopy in three dimensions .\",\n \"Knowledge about synaptic organization is critical to understand how sensory information is received , processed , and relayed by this circuit .\",\n \"Thus , the fly olfactory circuit represents an excellent opportunity to assess synaptic organization in an intact system at the resolution of individual , defined neuronal classes , linking structure to physiology and function .\",\n \"Here , we utilized an approach where presynaptic active zone and postsynaptic neurotransmitter receptor proteins were tagged with fluorescent molecules to highlight and quantitatively describe the synaptic organization of the Drosophila antennal lobe .\",\n \"We identified distinct rules that govern the number , density , and spatial organization of olfactory synapses for multiple classes of neurons .\",\n \"To understand how these rules are implemented at the molecular level , we investigated the function of the Teneurins , evolutionarily conserved type II transmembrane proteins recently shown to regulate synaptic partner matching and neuromuscular synaptic organization ( Hong et al . , 2012; Mosca et al . , 2012 ) .\",\n \"We show that presynaptic Ten-a and postsynaptic Ten-m regulate central synapse number and active zone structure .\",\n \"These represent the first in vivo functions of Teneurins in organizing CNS synapses , which could contribute to their association with brain disorders such as bipolar disorder .\"\n ],\n [\n \"To visualize synapses in vivo in genetically identifiable cells , we explored strategies to label active zones , sites of presynaptic neurotransmitter release ( Zhai and Bellen , 2004 ) .\",\n \"In Drosophila , the primary unit of active zone architecture is an electron dense structure called the T-bar , a specialized formation that promotes vesicle docking and release ( Wichmann and Sigrist , 2010 ) present at peripheral ( Jia et al . , 1993 ) and central synapses ( Meinertzhagen and O'Neil , 1991; Prokop and Meinertzhagen , 2006 ) .\",\n \"The major structural components of the T-bar include Bruchpilot ( Brp ) and Rim-binding protein ( Wagh et al . , 2006; Liu et al . , 2011 ) , with Brp encoding the ‘table-top’ of the T-bar and Rim-binding protein comprising the ‘pedestal’ ( Figure 1A ) .\",\n \"Widely used monoclonal antibodies ( nc82 ) against Brp ( Blanco Redondo et al . , 2013 ) label synaptic neuropil and discrete puncta corresponding to active zones ( Wagh et al . , 2006 ) .\",\n \"However , these antibodies cannot distinguish synapses formed by a given class of neurons in complex CNS circuits .\",\n \"Therefore , to visualize active zones in genetically identifiable cells , we sought to label Brp with the GAL4/UAS system ( Brand and Perrimon , 1993 ) . 10 . 7554/eLife . 03726 . 003Figure 1 . Measuring CNS synapses using Bruchpilot-Short .\",\n \"( A ) Diagram comparing full-length Bruchpilot and Bruchpilot-Short .\",\n \"Numbers denote amino acids and orange represents coiled coil regions ( after Wagh et al . , 2006 ) .\",\n \"The positions of Brp and Brp-Short in relation to known active zone proteins are also shown in a simplified format ( after Liu et al . , 2011 ) .\",\n \"( B–D )\",\n \"High magnification confocal stacks of a single DA1 glomerulus with putative synapses labeled by Brp-Short ( B ) , neurites labeled by mCD8-GFP ( C ) , and the merge ( D ) .\",\n \"Insets show higher magnification of a single optical section to demonstrate the punctate nature of Brp-Short-mStraw .\",\n \"( E–G )\",\n \"Screenshots of three-dimensional renderings show conversion of Brp-Short into ‘Spots’ ( E ) and mCD8-GFP into a volumetric surface rendering ( F ) .\",\n \"For ( F ) and ( G ) , the transparency of the surface rendering is at 60% to highlight internal structure and make Brp-Short puncta visible .\",\n \"( H ) Quantification of Brp-Short puncta ( red , left axis ) and neurite volume ( black , right axis ) in males and females for five ORN classes .\",\n \"Statistical comparisons between males and females of a single genotype were done by student's t test .\",\n \"Significance across all genotypes for density was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .\",\n \"***p < 0 . 001 .\",\n \"In all cases , n ≥ 9 brains , 18 antennal lobes for each genotype .\",\n \"( I ) Quantification of synaptic density ( Brp-Short puncta/volume of neurites ) in males ( blue ) and females ( pink ) for five ORN classes .\",\n \"Despite considerable differences in Brp-Short puncta number , all classes of ORNs display nearly identical synaptic densities .\",\n \"Scale bar = 5 µm for all images . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00310 . 7554/eLife . 03726 . 004Figure 1—figure supplement 1 . Validation of Brp-Short .\",\n \"( A ) Representative high magnification confocal single sections of Or67d-positive ORNs innervating the DA1 glomerulus expressing Brp-Short-mStraw and DSyd1-EGFP and stained with antibodies against mStraw ( red ) and GFP ( green ) .\",\n \"Note the significant overlap between Brp-Short and DSyd-1 , a known presynaptic marker .\",\n \"( B ) Representative high magnification confocal single sections of Or67d-positive ORNs innervating the DA1 glomerulus expressing Brp-Short-mStraw and mCD8-GFP in a control animal .\",\n \"( C ) Representative high magnification confocal single sections of Or67d-positive ORNs innervating the DA1 glomerulus expressing Brp-Short-mStraw , mCD8-GFP and an RNAi transgene against Brp .\",\n \"As endogenous Brp has been impaired , Brp-Short-mStraw is greatly diminished at the synapse .\",\n \"Scale bar = 5 μm .\",\n \"( D ) A single electron dense region as imaged by immunoelectron microscopy in an animal expressing Brp-Short-EGFP in all ORNs .\",\n \"5-nm Nanogold dots ( arrowheads ) decorate electron-dense structures characteristic of active zones following immunostaining against the GFP epitope , suggesting that Brp-Short localizes to endogenous active zones .\",\n \"Scale bar = 25 nm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00410 . 7554/eLife . 03726 . 005Figure 1—figure supplement 2 . Additional examples of the Brp-Short assay . Representative high magnification confocal z-stack images of ORNs innervating the VA1d ( Or88a-GAL4; [A] ) , VA1lm ( Or47b-GAL4; [B] ) , DL4 ( AM29-GAL4; [C] ) , and DM6 ( AM29-GAL4; [D] ) glomeruli expressing Brp-Short-mStraw and stained with antibodies to mStraw ( red ) and N-Cadherin ( blue ) .\",\n \"In all cases , puncta are clearly visible; these glomeruli were quantified in Figure 1H .\",\n \"Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 005 Full-length Brp is a 1740 amino acid protein that can aggregate upon ectopic overexpression ( Wagh et al . , 2006 ) , limiting its use as a labeling strategy .\",\n \"However , Bruchpilot-Short ( Brp-Short ) , a nonfunctional 754-residue portion of Brp ( Figure 1A ) , localizes to endogenous sites of full-length Brp ( Schmid et al . , 2008; Fouquet et al . , 2009 ) without deleterious effects .\",\n \"Thus , under the control of a binary expression system , Brp-Short can act as a reporter of endogenous active zones without disrupting morphology or function ( Schmid et al . , 2008; Fouquet et al . , 2009; Kremer et al . , 2010 ) .\",\n \"This strategy has been employed at neuromuscular synapses ( Schmid et al . , 2008; Fouquet et al . , 2009 ) , in the mushroom body ( Kremer et al . , 2010; Christiansen et al . , 2011 ) , and in the visual system ( Berger-Muller et al . , 2013 ) .\",\n \"We utilized Brp-Short to quantitatively measure the number of active zones , a key element of understanding synaptic organization .\",\n \"We expressed an mStrawberry-tagged version ( Brp-Short-mStraw: Owald et al . , 2010 ) in Or67d ORNs that innervate the DA1 glomerulus .\",\n \"Coexpression with a known presynaptic protein , DSyd-1 , revealed extensive colocalization ( Figure 1—figure supplement 1 , panel A ) , suggesting Brp-Short-mStraw correctly localizes to synapses .\",\n \"Further , RNAi against a portion of endogenous Brp not encoded by the Brp-Short transgene in those ORNs resulted in a loss of Brp-Short signal at ORN presynapses ( Figure 1—figure supplement 1 , panels B–C ) , though Brp-Short was still detectable in the cell body .\",\n \"This supports the hypothesis that Brp-Short requires endogenous Brp for proper localization to synapses; alternatively , Brp-Short could require endogenous Brp for proper trafficking to axon terminals .\",\n \"Finally , to confirm the veracity of the synaptic signal , we performed immuno-electron microscopy ( immuno-EM ) using antibodies to the GFP-tag ( the mStraw tag is not amenable to immuno-EM ) of a Brp-Short-GFP transgene ( Schmid et al . , 2008 ) expressed in all ORNs via pebbled-GAL4 ( Sweeney et al . , 2007 ) .\",\n \"With pebbled-GAL4 , Brp-Short expression and localization was similar to that observed for ORN-class-specific GAL4 drivers ( data not shown ) .\",\n \"While immuno-EM cannot be utilized as a quantitative strategy for determining the number of active zones ( as the conditions needed to preserve the GFP epitope necessarily reduce the resolution of the technique to clearly discern all synaptic densities ) , it can be used to verify that Brp-Short localizes to electron-dense regions that are consistent with synaptic active zones .\",\n \"Indeed , 5-nm gold particles were observed to decorate putative active zone profiles ( Figure 1—figure supplement 1 , panel D ) in ultrathin sections of fly antennal lobes .\",\n \"90 . 1% of Nano-Gold dots colocalized with synaptic densities ( 675 5-nm Nano-Gold particles out of 749 particles observed ) .\",\n \"Combined , these experiments demonstrate that Brp-Short localization is consistent with active zones in the fly CNS .\",\n \"To determine the number of active zones produced by ORN axons that project to a given glomerulus , we expressed Brp-Short-mStraw in various ORN classes ( Figure 1B–D , Figure 1—figure supplement\",\n \"2 ) together with UAS-mCD8-GFP ( Lee and Luo , 1999 ) as a general neurite label .\",\n \"Single optical sections ( Figure 1B–D , insets ) revealed the punctate nature of the Brp-Short-mStraw signal .\",\n \"To quantify the number of Brp-Short puncta in these ORNs , we imaged individual glomeruli using high magnification confocal microscopy followed by image deconvolution and 3-D rendering ( Video 1; see ‘Materials and methods’ ) .\",\n \"Image processing allowed us to calculate the number of Brp-Short puncta ( Figure 1B , E ) as ‘Spots’ and the neurite volume by surface rendering of the membrane marker mCD8-GFP ( Figure 1C , F ) .\",\n \"The numbers of Brp-Short puncta exhibited low variability for multiple classes of ORNs ( Figure 1H , Figure 1—figure supplement 2 ) , suggesting that neuronal classes form stereotyped numbers of active zones across animals .\",\n \"While extensive EM reconstruction is not available for the antennal lobe ( and thus , absolute counts of olfactory synapses in each glomerulus for comparison ) , the recent use of labeled Brp in the Drosophila visual system indicated that one Brp punctum is equivalent to one active zone ( Chen et al . , 2014 ) .\",\n \"Therefore , it is likely that the Brp-Short label largely reflects actual active zone number . 10 . 7554/eLife . 03726 . 006Video 1 . Three-dimensional rendering of a Brp-Short and mCD8-GFP in a DA1 glomerulus .\",\n \"( related to Figure\",\n \"1 ) Animation of a three-dimensional rendering of the glomerulus shown in Figure 1B–D .\",\n \"Each channel is highlighted and the transformation to either Spots ( for Brp-Short puncta ) or a surface rendering to measure neurite volume ( for mCD8-GFP staining ) is shown .\",\n \"High magnification of the rendered glomerulus reveals the distribution of Brp-Short puncta within the glomerulus . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 006 In examining five classes of ORNs , we found that ORN synapse number , as indicated by the number of Brp-Short puncta , scaled with ORN axon volume ( Figure 1H ) , which itself scaled with glomerular size .\",\n \"Moreover , the assay revealed sex-specific differences in synapse number for two glomeruli , DA1 and VA1lm , that coincide with known sex-specific differences in glomerular volume ( Stockinger et al . , 2005 ) .\",\n \"Remarkably , despite differences in synapse number and ORN neurite volume , all observed ORN classes had a constant synaptic density of 0 . 56 Brp-Short puncta per µm3 ( Figure 1I ) .\",\n \"These data suggested the existence of an organizational mechanism that controls synaptic density .\",\n \"In all , our results validated the use of Brp-Short as a quantifiable presynaptic active zone label in central neurons and revealed quantitative parameters that govern the synaptic contacts of ORNs: a stereotyped number of connections for each neuronal class from animal to animal and a constant synaptic density regardless of ORN class .\",\n \"In the olfactory system , individual ORNs of a particular class ( as denoted by their odorant receptor expression ) project to an identifiable glomerulus ( Vosshall and Stocker , 2007 ) .\",\n \"In terms of their projection pattern ( Jefferis and Hummel , 2006 ) and physiological function ( Wilson , 2013 ) , individual ORNs can largely be treated equivalently .\",\n \"However , it is unknown whether each neuron contributes an equal number of synapses to the aggregate , stereotyped average or whether there is heterogeneity in synapse number at the level of individual neurons .\",\n \"Different results would suggest distinct rules for how individual neurons determine how many connections to form .\",\n \"To address this , we utilized MARCM ( Lee and Luo , 1999 ) to label the synapses and membranes of individual ( or small subsets of ) ORNs that innervate two glomeruli , DL4 and DM6 .\",\n \"Most frequently , we obtained 2–3 cell clones ( Figure 2A–B ) , but we observed occurrences of single-cell clones ( Figure 2—figure supplement 1 , panels A–B ) and large clones ( Figure 2—figure supplement 1 , panels C–D ) that encompassed the entire glomerulus .\",\n \"Single neuron clones were readily discernible in confocal stacks; cell numbers could not be precisely assigned to larger than 2–3 cell clones , however , due to axon fasciculation .\",\n \"For all clones ( n = 100 for DL4 , 99 for DM6 ) except the largest 7 for DL4 and 9 for DM6 , we quantified the number of Brp-Short puncta and the volume of the mCD8-GFP-labeled neurites as above ( Figure 1 ) , and plotted the distribution of puncta number across all clones in histograms .\",\n \"For both DL4 ( Figure 2C ) and DM6 ( Figure 2D ) , we observed that as clone size increased , the number of Brp-Short puncta appeared to increase in discrete steps of a similar size , suggesting that each individual neuron contributes an equal number of synapses to the glomerular average .\",\n \"If each neuron contributed a widely different number of synapses to the average , we would have observed a more continuous distribution . 10 . 7554/eLife . 03726 . 007Figure 2 . Clonal analysis of ORN synapse number . Representative high magnification confocal stacks of small clones of DL4 ORNs ( A ) and DM6 ORNs ( B ) expressing Brp-Short-mStraw ( red ) and mCD8-GFP ( green ) .\",\n \"( C ) Frequency histogram of Brp-Short puncta in DL4 clones ( n = 93 clones from 64 animals ) .\",\n \"( D ) Frequency histogram of Brp-Short puncta in DM6 clones ( n = 90 clones from 64 animals ) .\",\n \"Both graphs exhibit notable periodicity , suggesting that each neuron contributes approximately the same number of synapses .\",\n \"Colored dashed lines indicate Gaussian fits for individual peaks .\",\n \"Solid black lines represent the best-fit sum of Gaussian relationship for each dataset .\",\n \"( E ) Quantification of Brp-Short puncta from identified single neuron clones to DL4 ( n = 16 clones ) or DM6 ( n = 17 clones ) .\",\n \"( F ) Quantification of synapse density from all clones .\",\n \"In both DL4 and DM6 , the clonal average density is identical to the class average density .\",\n \"In ( E ) and ( F ) , data represent mean ± SEM .\",\n \"Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00710 . 7554/eLife . 03726 . 008Figure 2—source data 1 . Table of Gaussian curve-fitting data . Table of statistical values for three potential sum of Gaussian fits each for the DL4 and DM6 MARCM datasets . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00810 . 7554/eLife . 03726 . 009Figure 2—source data 2 . Table of resampling approach . Table of simulation data for three distributions of the DL4 and DM6 MARCM data sets and the probabilities of obtaining simulated data more evenly distributed than observed . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 00910 . 7554/eLife . 03726 . 010Figure 2—figure supplement 1 . Additional DL4 and DM6 ORN MARCM clones . Representative high magnification confocal stacks of single cell clones of DL4 ORNs ( A ) and DM6 ORNs ( B ) and large clones of DL4 ORNs ( C ) and DM6 ORNs ( D ) expressing Brp-Short-mStraw and mCD8-GFP and stained for antibodies against mStraw ( red ) , GFP ( green ) , and N-Cadherin ( blue ) .\",\n \"As previously observed ( Joo et al . , 2013 ) , single cell ORN clones have a low signal-to-noise ratio while discerning larger clones is more evident .\",\n \"In A–B , the glomeruli are outlined ( dashed line ) .\",\n \"Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 010 To statistically analyze the distributions for DL4 and DM6 ORNs , we determined their best-fit relationships using the first four peaks for both data sets , which had the most observations .\",\n \"Each individual peak could be reliably fit using a single Gaussian relationship ( r2 > 0 . 98 , Figure 2C–D , colored dashed lines ) , and the entire set of peaks could be fit by a sum of Gaussians ( see Figure 2—source data 1; ‘Materials and methods’ ) .\",\n \"The DL4 relationship was best fit by a 4-term sum of Gaussians ( Figure 2C , solid black line ) with a value of 19 . 13 for the apex of the first peak , corresponding to the ‘single cell’ contribution of active zone puncta .\",\n \"This is in close agreement with our original data of 19 . 75 ± 0 . 68 puncta for single cell clones ( Figure 2E ) .\",\n \"The DL6 relationship was best fit by a 5-term sum of Gaussians ( Figure 2D , solid black line ) , with a value of 28 . 62 active zone puncta as the ‘single cell’ contribution , which is again in strong agreement with our original data of 29 . 06 ± 0 . 95 puncta for single cell clones ( Figure 2E ) .\",\n \"Thus , in both cases , a sum of Gaussians can accurately represent the observed peaks and predict single ‘quantum’ contributions similar to those observed for single cell clones .\",\n \"These ‘quantum’ values are consistent with the ‘step size’ of Brp-Short puncta measured in clones with increasing size ( Figure 2C–E ) .\",\n \"Further , in comparing DL4 and DM6 , the difference in the number of puncta per single cell clone ( Figure 2E ) is consistent with the numerical relationship between the aggregate glomerular averages ( Figure 1H ) .\",\n \"All observed clones also displayed a nearly identical synaptic density ( Figure 2F ) regardless of their size , consistent with measurements in whole glomeruli .\",\n \"To further test whether each ORN contributes the same number of synapses to the aggregate , we examined the spacing of the peaks .\",\n \"If the peaks were perfectly evenly spaced , the standard deviation in the distances ( S ) between each peak would be 0 .\",\n \"Using the individual Gaussian relationships ( colored dashed lines ) , we calculated an S value of 0 . 235 for DL4 and 2 . 26 for DM6 .\",\n \"While visual assessment suggested uniformity , we used these values to more rigorously assess this .\",\n \"We used a resampling approach and calculated an S value for each simulation to test the null hypothesis that there is no periodic relationship between the peaks .\",\n \"A lower S value than that which was observed means that the simulation resulted in peaks that were more evenly spaced .\",\n \"Figure 2—source data 2 details these results .\",\n \"For distributions of DL4 or DM6 , there was a low probability ( <0 . 00048 or <0 . 05 , respectively ) of seeing peaks as evenly distributed as that observed .\",\n \"Thus , we reject the null hypothesis of no quantal relationship for both DL4 and DM6 .\",\n \"In all , this approach demonstrates that the peaks have equal spacing in between , further supporting our assertion that each ORN contributes an equivalent number of active zones to the aggregate and that each class of ORNs has a characteristic ‘step size’ consistent with its aggregate average .\",\n \"To better understand CNS synaptic organization , it is important to know not only how many connections are formed , but also where they form .\",\n \"Brp-Short allowed us to examine the spatial organization of active zones of any specific neuronal class within individual glomeruli .\",\n \"As an approximation of the complement of active zones for the DA1 glomerulus , we examined its component ORNs , PNs , and LNs .\",\n \"Using GAL4 drivers for the ORNs ( Or67d-GAL4 ) , PNs ( Mz19-GAL4 ) , and an LN line that includes most ipsilateral LNs ( NP3056-GAL4 ) , all of which innervate the DA1 glomerulus , we examined the localization of their synapses .\",\n \"For all classes , Brp-Short puncta colocalized with endogenous Brp staining ( Figure 3—figure supplement 1 , panels A–C ) recognized by an antibody whose epitope is not contained within Brp-Short ( Wagh et al . , 2006 ) , demonstrating that Brp-Short recognizes endogenous active zones .\",\n \"These three classes of neurons , when combined , provide 3342 ± 99 Brp-Short puncta per glomerulus .\",\n \"When the endogenous number of active zones is quantified using an antibody to endogenous Brp using the same image-processing strategy , DA1 contains 3849 ± 80 active zones ( Figure 3—figure supplement 1D ) .\",\n \"This suggests that the majority of synapses in the DA1 glomerulus are indeed made by Or67d-GAL4+ ORNs , Mz19-GAL4+ PNs , and NP3056-GAL4+ LNs; the discrepancy may be contributed by synapses made by additional neurons , such as other classes of local interneurons ( Chou et al . , 2010 ) , atypical projection neurons ( Lai et al . , 2008 ) , or potentially peptidergic neurons ( Ignell et al . , 2009; Carlsson et al . , 2010 ) , or may represent a slight underestimation by the Brp-Short assay .\",\n \"Using single optical sections within the DA1 glomerulus , we observed two levels of organization: ( 1 ) the location of neurites and Brp-Short puncta with respect to the glomerulus ( the subglomerular distribution ) , and ( 2 ) Brp-Short puncta with respect to their own neurites for each class of neurons ( the subcellular distribution ) .\",\n \"Of the three classes of neurons observed , ORNs contribute the majority of the Brp-Short puncta ( 1632 ± 22 ) and were widely distributed within the glomerulus ( Figure 3A , D ) , though with distinct voids ( Figure 3D , asterisk ) where the glomerulus lacked ORN neurites and Brp-Short puncta .\",\n \"PNs were the second major contributors of Brp-Short puncta within the glomerulus ( 1182 ± 24 ) .\",\n \"These connections likely synapse onto PN and LN dendrites , as shown by electrophysiology ( Kazama and Wilson , 2009 ) , and are analogous to the dendrodendritic synapses between secondary dendrites of mitral cells and olfactory bulb interneurons in the vertebrate olfactory bulb ( Urban and Arevian , 2009 ) .\",\n \"These Brp-Short puncta showed a more even distribution throughout the glomerulus ( Figure 3B , E ) , with smaller voids than those observed for ORNs despite reduced overall density .\",\n \"Finally , local interneurons ( LNs ) contribute the fewest Brp-Short puncta ( 528 ± 54 ) of the three classes .\",\n \"The LNs accessed by the NP3056-GAL4 line make up the majority of ipsilateral LNs including most or all pan-glomerular projections ( Chou et al . , 2010 ) , but may not be representative of all classes of LNs .\",\n \"Within the glomerulus , LN distribution was the most restricted ( Figure 3C , F ) .\",\n \"When examined at equivalent optical sections , these projections mostly filled the gaps in the ORN projection pattern , consistent with previously observed results for neurites ( Hummel and Zipursky , 2004 ) . 10 . 7554/eLife . 03726 . 011Figure 3 . Antennal lobe neurons display distinct subglomerular architecture .\",\n \"( A–C )\",\n \"Single , high-magnification confocal optical sections through the center of the DA1 glomerulus in three animals where ORNs ( A ) , PNs ( B ) , or LNs ( C ) are expressing Brp-Short .\",\n \"A different color denotes the synapses of each class .\",\n \"( D ) Single , high magnification confocal optical sections through the center of the DA1 glomerulus where ORNs express Brp-Short-mStraw ( red ) and mCD8-GFP ( green ) .\",\n \"Brp-Short clusters ( arrowheads ) are visible , along with regions of neurite devoid of Brp-Short puncta ( arrow ) and voids where ORNs do not project ( asterisk ) .\",\n \"( E ) Single , high magnification confocal optical sections through the center of the DA1 glomerulus where PNs are expressing Brp-Short-mStraw ( red ) and mCD8-GFP ( green ) .\",\n \"Regions of PN neurite without Brp-Short puncta are visible ( arrow ) as well as small voids lacking Brp-Short puncta and neurite projections ( asterisk ) .\",\n \"( F ) Single , high magnification confocal optical sections through the center of the DA1 glomerulus where LNs are expressing Brp-Short-mStraw ( red ) and mCD8-GFP ( green ) .\",\n \"Brp-Short puncta are largely distributed evenly throughout neurites .\",\n \"Images in panels D–F ( from three animals ) demonstrate distinct patterns of synapse localization and distribution within the glomerulus .\",\n \"Dashed lines denote the glomerulus in ( C ) and ( F ) as defined by N-Cad staining ( not shown ) .\",\n \"Scale bar = 5 μm .\",\n \"( G–I )\",\n \"Cumulative frequency histograms of the nearest neighbor distance between Brp-Short puncta in ORNs ( G ) , PNs ( H ) , and LNs ( I ) .\",\n \"The average is indicated ( μ ) and the Cluster % of puncta with an NND ≤ 0 . 75 μm .\",\n \"Gray traces represent individual glomeruli .\",\n \"Red traces represent the aggregate average .\",\n \"In all cases , n ≥ 10 animals , 19 antennal lobes , and 400 ( I ) , 800 ( H ) , or 1400 ( G ) individual puncta . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 01110 . 7554/eLife . 03726 . 012Figure 3—figure supplement 1 . Comparison of Brp-Short to endogenous Brp .\",\n \"( A ) Representative high magnification confocal z-stack of the DA1 glomerulus in animals with an Or67d-mCD8-GFP transgene to mark its borders and stained with antibodies against GFP ( red ) and endogenous Brp ( green ) .\",\n \"( B–D )\",\n \"Representative high magnification confocal single sections of Or67d-positive ORNs ( B ) , Mz19-positive PNs ( C ) , and NP3056-positive LNs ( D ) innervating the DA1 glomerulus , each expressing Brp-Short-mStraw and stained with antibodies against mStraw ( red ) and endogenous Brp ( green ) .\",\n \"Each image was taken from different animals at a roughly equivalent Z-plane .\",\n \"In all cases , Brp-Short-mStraw puncta are coincident with endogenous Brp puncta .\",\n \"Endogenous Brp puncta that are not Brp-Short-mStraw positive are likely from other neurons not expressing the Brp-Short-mStraw transgene .\",\n \"Dashed lines indicate the glomerular boundary as defined by N-Cad staining ( not shown ) .\",\n \"( D ) Quantification of Brp-Short puncta from ORNs ( green ) , PNs ( blue ) , and LNs ( magenta ) , the sum of ORNs + PNs + LNs , and the number of endogenous Brp puncta from the DA1 glomerulus ( black ) .\",\n \"In all cases , data represent mean ± SEM and n ≥ 7 animals , 14 antennal lobes .\",\n \"Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 012 With respect to their own neurites , Brp-Short puncta in ORNs were not evenly distributed: distinct sections of neurite lack Brp-Short puncta ( Figure 3D , arrow ) while other sections had large clusters ( Figure 3D , arrowheads ) of Brp-Short puncta ( Figure 3D ) .\",\n \"To quantify this distribution , we conducted a nearest neighbor distance ( NND ) analysis for Brp-Short puncta in each class of neurons , which has been informative in understanding spatial relationships between synapses ( Bleckert et al . , 2013 ) .\",\n \"We found that all three types of neurons had distinct mean NND values ( Figure 3G–I ) : ORNs had the closest puncta and LNs the furthest with PNs falling in between the two .\",\n \"To quantify the clusters observed , we defined clustered puncta as having an NND of 0 . 75 μm or less ( an NND of 0 . 6 μm corresponds to directly touching ) .\",\n \"For ORNs , these large clusters amounted to 17% of puncta that fell within this range ( Figure 3G ) .\",\n \"PNs lacked these large clusters , but had some smaller clusters , as well as extended regions of neuronal membrane devoid of Brp-Short puncta ( Figure 3E ) .\",\n \"Indeed , only 7 . 9% of PN puncta were clustered by this metric ( Figure 3H ) .\",\n \"Within LN neurites , Brp-Short puncta also displayed significant clustering ( 15%; Figure 3I ) .\",\n \"However , unlike ORNs and PNs , no extended regions of membrane were observed without synaptic puncta .\",\n \"This suggests a more even distribution for the remaining puncta , which is supported by the higher value for NND ( Figure 3I ) and consistent with previous imaging using a synaptic vesicle marker ( Chou et al . , 2010 ) .\",\n \"Overall , these results show distinct synaptic architecture , with respect to the glomerulus and the neurites themselves , made by three different types of olfactory neurons .\",\n \"Moreover , distinct parameters can be defined for each class in terms of how puncta are clustered and the minimum distance between puncta .\",\n \"Though some distances may be underestimated , these facets may represent additional rules that govern synaptic organization in the antennal lobe .\",\n \"Having utilized Brp-Short to determine aspects of presynaptic organization in olfactory neurons ( Figures 1–3 ) , we next sought to investigate the molecular mechanisms that may underlie these rules .\",\n \"We began with genetic perturbations of candidate molecules , the Teneurins .\",\n \"Teneurins are type II transmembrane proteins involved in neuronal development ( Young and Leamey , 2009 ) and are disrupted in human neurological disorders ( Aldahmesh et al . , 2012; Psychiatric GWAS Consortium Bipolar Disorder Working Group , 2011 ) .\",\n \"Recently , we demonstrated that the Drosophila Teneurins , Ten-m and Ten-a , regulate synaptic partner matching in the olfactory and neuromuscular systems via homophilic interaction between elevated Teneurin levels ( Hong et al . , 2012; Mosca et al . , 2012 ) .\",\n \"Basal levels of the Teneurins interact heterophilically at NMJ synapses to organize connections ( Mosca et al . , 2012 ) .\",\n \"However , due to the absence of techniques for quantitatively examining central synapses in vivo , a role for the Teneurins in organizing central synapses has not been established .\",\n \"Due to the similarity in partner matching mechanisms between central and peripheral nervous systems , and the fact that both Ten-a and Ten-m are expressed at basal levels in all glomeruli ( Hong et al . , 2012 ) , we hypothesized that the CNS may use Teneurins similarly to the PNS for synaptic organization .\",\n \"Specifically , we hypothesized that olfactory neurons would utilize basal levels of Ten-a and Ten-m to control synaptic number .\",\n \"To test this hypothesis , we used Brp-Short to quantify active zone number in Or47b-positive ORNs innervating the VA1lm glomerulus .\",\n \"Ten-a is expressed at a basal level in this glomerulus , and therefore does not directly affect synaptic partner matching ( Hong et al . , 2012 ) .\",\n \"Comparing wild-type ( Figure 4A ) and ten-a null mutant flies ( Figure 4B ) , we found that Or47b ORNs in the ten-a mutant ( Figure 4B ) had 23% fewer Brp-Short puncta ( Figure 4D ) .\",\n \"Notably , loss of ten-a did not alter ORN axon volume ( Figure 4E ) , despite affecting glomerular morphology ( Figure 4B ) , which is due to the loss of ten-a in neurons of the nearby DA1 glomerulus ( Hong et al . , 2012 ) .\",\n \"As such , synaptic density was similarly reduced ( Figure 4F ) .\",\n \"To ensure that changes in glomerular morphology were not related to changes in synapse number , we also examined ORNs that innervate the DL4 glomerulus ( Figure 4—figure supplement 1 ) .\",\n \"In ten-a mutants , glomerular morphology was unchanged in 87% of animals examined , but all animals displayed a similar reduction in synapse number and density without affecting the axon volume ( Figure 4—figure supplement 1 , panels C–E ) .\",\n \"Interestingly , at the NMJ , Teneurin perturbations reduced active zone and synaptic bouton numbers , but did not alter active zone density ( TM , unpublished observations ) .\",\n \"However , in ORNs , neurite volume was unchanged , so active zone density was reduced , suggesting a difference in the specific consequence of Teneurin perturbation in the CNS compared with the NMJ . 10 . 7554/eLife . 03726 . 013Figure 4 . Presynaptic Ten-a is required for proper ORN synapse number and density . Representative high-magnification confocal stacks of Brp-Short puncta in VA1lm ORNs stained with antibodies against Brp-Short-mStraw ( red ) and N-Cadherin ( blue ) in control ( A ) , ten-a null mutants ( B ) and ten-a null mutants where a Ten-a transgene is expressed in this specific class of ORNs using Or47b-GAL4 ( C ) .\",\n \"( D ) Quantification of Brp-Short puncta in the above conditions as well as following presynaptic RNAi against ten-a , presynaptic RNAi against ten-m , or in ten-a null mutants where a Ten-a transgene is expressed in 2/3 of the PNs , including those innervated by Or47b-positive ORNs .\",\n \"( E ) Quantification of ORN neurite volume in the same genotypes .\",\n \"n . s . = not significant .\",\n \"( F ) Quantification of ORN synaptic density in the same genotypes .\",\n \"Significance was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .\",\n \"**p < 0 . 01 , ***p < 0 . 001 .\",\n \"Unless otherwise noted , significance is compared to control .\",\n \"In all cases , the ten-a mutant and ten-a ORN RNAi phenotypes are statistically indistinguishable and , in the ten-a mutant , presynaptic Ten-a expression rescues the observed phenotypes .\",\n \"The disrupted glomerular shape is due to partner matching errors in the ten-a mutant ( Hong et al . , 2012 ) , which is not rescued due to the late onset of Or47b-GAL4 relative to the partner matching process .\",\n \"Dashed lines denote the glomeruli as delineated by Brp-Short label .\",\n \"In all cases , data represent mean ± SEM and n ≥ 7 animals , 14 antennal lobes .\",\n \"Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 01310 . 7554/eLife . 03726 . 014Figure 4—figure supplement 1 . Representative images for additional genetic manipulations from Figure 4 . Representative high-magnification confocal stacks of Brp-Short puncta in VA1lm ORNs stained with antibodies against Brp-Short-mStraw ( red ) and N-Cadherin ( blue ) in control ( A ) , flies expressing ten-m RNAi in Or47b-positive ORNs ( B ) , flies expressing ten-a RNAi in Or47b-positive ORNs ( C ) , and ten-a null mutants where a Ten-a transgene is expressed in 2/3 of the PNs of the antennal lobe using GH146-QF ( D ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 01410 . 7554/eLife . 03726 . 015Figure 4—figure supplement 2 . ten-a phenotypes in DL4 , a glomerulus with no morphology defects .\",\n \"( A ) Representative high magnification confocal stacks of ORNs innervating the DL4 glomerulus in control animals expressing Brp-Short-mStraw and mCD8-GFP and stained with antibodies to mStraw ( red ) , GFP ( green ) , and N-Cadherin ( blue ) .\",\n \"( B ) Representative high magnification confocal z-stacks of ORNs innervating the DL4 glomerulus in ten-a null mutant animals expressing Brp-Short-mStraw and mCD8-GFP and stained with antibodies to mStraw ( red ) , GFP ( green ) , and N-Cadherin ( blue ) .\",\n \"Morphology is unaffected but synapse number is reduced .\",\n \"( C ) Quantification of Brp-Short-mStraw puncta in control and ten-a mutant animals .\",\n \"( D ) Quantification of ORN volume in control and ten-a mutant animals .\",\n \"( E ) Quantification of Brp-Short-mStraw puncta density in control and ten-a mutant animals .\",\n \"In all cases , data represent mean ± SEM and n ≥ 7 animals , 14 antennal lobes .\",\n \"***p < 0 . 0001 .\",\n \"Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 015 As ten-a mutants lack Ten-a in ORNs and PNs , we next sought to determine where Ten-a was required to regulate normal synapse number using tissue-specific rescue and RNAi experiments .\",\n \"Restoring Ten-a expression to the Or47b ORNs in the ten-a mutant ( Figure 4C ) , but not to the VA1lm PNs that are innervated by Or47b ORNs , rescued the reduction in synapse number and density ( Figure 4D , F ) .\",\n \"When Ten-a was restored to Or47b-positive ORNs , defects in glomerular morphology persisted despite rescue of synaptic phenotypes ( compare Figure 4C–A ) , as partner matching occurs earlier than synaptogenesis ( and before Or47b-GAL4 turns on ) .\",\n \"Further consistent with a presynaptic role for Ten-a , knockdown of Ten-a in ORNs by RNAi also reduced the number and density of synapses similarly to the mutant ( Figure 4D , F ) .\",\n \"Here , glomerular morphology was unaffected as VA1lm is already a ‘Ten-a low’ glomerulus and does not directly require Ten-a for partner matching .\",\n \"Interestingly , knockdown of ten-m in the Or47b ORNs did not result in any synaptic phenotype ( Figure 4D–F ) , demonstrating specificity for ten-a .\",\n \"Thus , Ten-a is required presynaptically for proper synapse number and density , likely in a partially redundant fashion with additional molecules .\",\n \"Previous data and our Brp-Short analyses suggest that the assay is largely reflective of actual synapse number in neurons .\",\n \"However , we sought to confirm these results using a method independent of confocal microscopy or the Brp-Short label .\",\n \"We compared synapse number in ultrathin EM sections of wild-type and ten-a mutant flies ( Figure 5A–B ) .\",\n \"Quantification of all synapses in a glomerulus ( and thus a direct comparison to the Brp-Short assay ) would require serial reconstruction of that entire glomerulus by EM .\",\n \"However , if the Brp-Short assay accurately reflects changes in synapse number following genetic perturbation , we reasoned that this should be evident in the change in the number of T-bar profiles in individual sections at the EM level .\",\n \"Indeed , quantification of T-bar profiles showed a 27% reduction in the number of T-bars ( Figure 5I ) .\",\n \"This is consistent with the results from the Brp-Short assay ( Figure 4D ) , which shows a 23% reduction .\",\n \"The slight difference may reflect an under-report of the phenotype by our confocal-level analysis or may reflect the fact that the EM analysis does not differentiate between ORN , PN , and LN presynapses while the Brp-Short assay does .\",\n \"The largely similar magnitude of synapse reduction in EM and confocal studies supports the validity of the Brp-Short assay , and further substantiates that Ten-a is required for proper synapse number in the CNS . 10 . 7554/eLife . 03726 . 016Figure 5 . Ultrastructural analysis of active zones in ten-a mutants . Representative transmission electron microscopy ( TEM ) images of active zones in control ( A ) and ten-a null mutant antennal lobes ( B ) .\",\n \"T-bar profiles are visible in both genotypes , though reduced in number in the ten-a mutant .\",\n \"Scale bar = 500 nm .\",\n \"( C ) High magnification TEM of a single active zone in a control animal .\",\n \"Note the normal T-bar morphology .\",\n \"( D–H )\",\n \"High magnification TEM images of active zones in ten-a mutants .\",\n \"In some cases , normal ( D ) morphology is observed .\",\n \"In the majority of cases , defects including multiple T-bars ( E ) , misshapen T-bars ( F ) , detached T-bars ( G ) , and continuous electron dense material ( H ) are evident .\",\n \"Scale bar = 100 nm .\",\n \"( I ) Quantification of T-bars per unit perimeter of measured antennal lobe terminals .\",\n \"All terminals were measured , including ORNs , PNs , and LNs .\",\n \"ten-a mutants display a 27% reduction from control animals .\",\n \"Data represent mean ± SEM .\",\n \"( J ) Distribution of T-bar defects as a percentage of the total T-bars .\",\n \"Here , n = 1103 T-bar per unit perimeter measurements from three animals for control and n = 847 T-bar per unit perimeter measurements from three animals for ten-a . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 016 At neuromuscular synapses , Teneurins are required for proper T-bar structure .\",\n \"When Teneurins are absent , T-bars display defects consistent with failed synaptogenesis ( Mosca et al . , 2012 ) .\",\n \"To determine whether Ten-a was also required in olfactory neurons for T-bar structure , we examined individual T-bar morphology in ultrathin sections of wild-type and ten-a antennal lobes .\",\n \"In wild-type animals , T-bar morphology was nearly invariant ( Figure 5C ) , with infrequent occurrences of misshapen or multiple T-bars ( Figure 5J ) .\",\n \"In ten-a mutants , 44% of T-bar profiles appeared normal ( Figure 5D , J ) .\",\n \"The remaining 56% displayed notable structural defects including continuous , multiple T-bars ( Figure 5E ) , misshapen T-bars ( Figure 5F ) , and T-bars that were detached from the synaptic membrane ( Figure 5G ) .\",\n \"These defects resembled those observed at ten-a mutant NMJs ( Mosca et al . , 2012 ) .\",\n \"We also observed continuous , electron-dense material resembling undivided T-bars ( Figure 5H ) that was not found previously .\",\n \"Thus , Ten-a is required not just for the normal number of active zones , but also for their proper structure in ways both consistent with and beyond that of previous studies .\",\n \"Specifically , ten-a may be required for the maturation of synapses , as misshapen and undivided T-bars resemble immature synapses ( Owald et al . , 2010 ) .\",\n \"Teneurins can regulate cytoskeletal architecture ( Nunes et al . , 2005; Morck et al . , 2010; Suzuki et al . , 2014 ) and function at the Drosophila NMJ by organizing spectrin , a cytoskeletal component ( Mosca et al . , 2012 ) .\",\n \"However , it is unknown whether this function of the Teneurins is conserved in the CNS .\",\n \"To further probe the mechanism by which Teneurins regulate central synapse number and to determine whether cytoskeletal regulation is involved , we compared levels of spectrin staining in the antennal lobe between wild-type and ten-a mutant flies .\",\n \"We found that spectrin immunofluorescence in the antennal lobe was reduced in the ten-a mutant compared to wild-type ( Figure 6A–B ) while N-Cadherin , a marker of synaptic neuropil , was unaffected ( Figure 6F ) .\",\n \"This was observed for α- and β-spectrin ( Figure 6F , Figure 6—figure supplement 1 , panels A–D ) , which function as a heterotetramer ( Machnicka et al . , 2014 ) . 10 . 7554/eLife . 03726 . 017Figure 6 . Ten-a and spectrin function together for normal synapse number .\",\n \"( A–B )\",\n \"Representative single confocal optical sections taken at equivalent positions of antennal lobes stained for α-spectrin and N-Cadherin in control ( A ) and ten-a mutant ( B ) adults .\",\n \"In ten-a mutants , α-spectrin staining is reduced compared to control .\",\n \"The disrupted morphology of the antennal lobe itself is due to partner matching errors evident in the ten-a mutant ( Hong et al . , 2012 ) .\",\n \"Scale bar = 20 μm .\",\n \"( C–E )\",\n \"Representative confocal Z-stack images of the ORNs in the VA1lm glomerulus expressing Brp-Short in control animals ( C ) , or animals expressing dsRNA against α-spectrin ( D ) or β-spectrin ( E ) , and stained with antibodies against Brp-Short ( red ) and N-Cadherin ( blue ) .\",\n \"( F ) Quantification of α-spectrin ( green ) , β-spectrin ( red ) , and N-Cadherin ( blue ) immunofluorescence in control and ten-a mutants .\",\n \"***p < 0 . 001 .\",\n \"( G ) Quantification of Brp-Short puncta in the noted genotypes .\",\n \"In all cases , similar reductions in puncta number are observed .\",\n \"Moreover , the genetic perturbations do not enhance each other , suggesting function in the same pathway .\",\n \"Significance was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .\",\n \"***p < 0 . 001 ( compared with control ) .\",\n \"In all cases , data represent mean ± SEM and n ≥ 10 animals , 19 antennal lobes .\",\n \"Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 01710 . 7554/eLife . 03726 . 018Figure 6—figure supplement 1 . ten-a regulates spectrin levels and spectrin regulates synapse number . Representative confocal single optical sections of control ( A ) and ten-a mutant ( B ) antennal lobes stained with antibodies to β-spectrin ( red ) and N-Cadherin ( blue ) .\",\n \"Note the reduction ( quantified in Figure 6F ) of staining in the ten-a mutant .\",\n \"Scale bar = 20 μm .\",\n \"( C–E )\",\n \"Representative high magnification confocal z-stacks of Or67d ORNs expressing Brp-Short-mStraw in control animals ( C ) and animals co-expressing dsRNA against α-spectrin ( D ) or β-spectrin ( E ) and stained with antibodies against mStraw ( red ) and N-Cadherin ( blue ) .\",\n \"Glomerular morphology is unaffected by spectrin dsRNA but synapse number is reduced .\",\n \"Scale bar = 5 μm .\",\n \"( F ) Quantification of Brp-Short puncta in the genotypes from C–E .\",\n \"***p < 0 . 0001 .\",\n \"In all cases , data represent mean ± SEM and n ≥ 10 animals , 19 antennal lobes . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 018 To test whether proper spectrin level is required for synapse number , we used ORN-specific knockdown of either α- or β-spectrin and measured Brp-Short puncta .\",\n \"In both cases ( Figure 6D–E ) , RNAi in Or47b-positive ORNs resulted in a 17–18% reduction in Brp-Short puncta ( Figure 6G ) .\",\n \"This effect , though slightly weaker , phenocopied the reduction seen in the ten-a mutant .\",\n \"We observed a similar effect in another ORN class , the Or67d-positive neurons ( Figure 6—figure supplement 1 ) , further supporting a role for spectrin in regulating olfactory synapse number .\",\n \"Moreover , the ten-a phenotype was not enhanced by concurrent RNAi knockdown of α-spectrin in ORNs ( Figure 6G ) , suggesting that Ten-a and α-spectrin function in the same pathway to regulate olfactory synapse number .\",\n \"A functional synapse consists of a presynaptic neurotransmitter release site and a postsynaptic neurotransmitter receptor cluster ( Figure 1A ) .\",\n \"Therefore , critical parameters of synaptic organization within a circuit not only include the location and number of presynaptic active zones , but also postsynaptic receptor clusters .\",\n \"Therefore , we also examined the organization of postsynapses .\",\n \"Given that ORNs are cholinergic ( Wilson , 2013 ) , an ideal labeling strategy would image postsynaptic acetylcholine receptors .\",\n \"We selected the Dα7 acetylcholine receptor subunit because it is endogenously expressed in the antennal lobe ( Fayyazuddin et al . , 2006 ) and it has been used to examine organization in the mushroom body , a higher olfactory center ( Leiss et al . , 2009a , 2009b; Kremer et al . , 2010; Christiansen et al . , 2011 ) .\",\n \"We used a GFP-tagged Dα7 transgene ( Leiss et al . , 2009b ) under the control of the GAL4/UAS system ( Figure 7A ) to visualize postsynapses in vivo .\",\n \"Expression of Dα7-GFP in PNs revealed distinct puncta , possibly corresponding to acetylcholine receptor ( AChR ) clusters ( Figure 7B ) .\",\n \"These puncta were apposed to endogenous Brp puncta , as revealed by nc82 staining ( Supplement 1 to Figure 7 ) , consistent with these puncta representing bona fide synapses .\",\n \"To examine AChR clusters in PNs , we co-expressed Dα7-GFP with mtdT as a general neurite label ( Figure 7B ) .\",\n \"As such , the approach is analogous to our Brp-Short assay and yielded similar results , enabling a quantitative assessment of the number ( Figure 7C ) and density ( Figure 7D ) of AChR clusters . 10 . 7554/eLife . 03726 . 019Figure 7 . Measuring postsynaptic acetylcholine receptor clusters with Dα7-GFP .\",\n \"( A ) Diagram of the GFP-tagged Dα7 acetylcholine receptor subunit used for AChR visualization .\",\n \"( B ) High magnification confocal z-stack images of PNs in the DA1 and VA1d glomeruli expressing Dα7-GFP and mtdT , stained with antibodies against GFP ( green ) , mtdT ( red ) , and endogenous Brp ( blue ) .\",\n \"Patches of mtdT labeling represent ascending PN axons within the plane of the glomerulus .\",\n \"Insets show a high magnification single optical section demonstrating the punctate nature of Dα7-GFP .\",\n \"( C ) Representative high magnification single optical sections of Mz19-positive PNs in the DA1 glomerulus expressing Dα7-GFP and stained with antibodies against GFP ( green ) and endogenous Brp ( red ) .\",\n \"The majority of GFP-positive puncta are apposed to or colocalized with endogenous Brp ( yellow ) , consistent with their association with bona fide active zones .\",\n \"Insets show high magnification of a single optical section where asterisks denote apposed Dα7 puncta .\",\n \"Brp puncta without apposition likely belong to synapses not labeled by the PN-GAL4 driver .\",\n \"Scale bar = 5 μm ( 2 μm for inset ) .\",\n \"( D ) Representative high magnification single optical sections of Mz19-positive PNs and Or67d-positive ORNs in the DA1 glomerulus expressing Dα7-GFP in the PNs and Brp-Short-mStraw in the ORNs using two binary expression systems .\",\n \"Most ORN active zones ( labeled by Brp-Short , red ) are apposed to or colocalized ( yellow ) with PN Dα7 puncta ( green ) , further supporting that these puncta label bona fide synaptic contacts .\",\n \"Insets show high magnification of a single optical section where asterisks denote apposed Brp-Short::Dα7 pairs .\",\n \"Brp puncta without apposition likely belong to ORN synapses with neurons other than PNs and Dα7 puncta without apposition likely correspond to PN postsynaptic sites apposed to synaptic contacts from neurons other than ORNs .\",\n \"( E ) Quantification of Dα7 AChR puncta ( green , left axis ) and neurite volume ( black , right axis ) in the DA1 and VA1d glomeruli of both male and female adult flies .\",\n \"Statistical comparisons between males and females of a single genotype were done by student's t test .\",\n \"***p < 0 . 001 .\",\n \"( F ) Quantification of AChR density in male ( blue ) and female ( pink ) adults based on the data from ( C ) .\",\n \"Significance was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .\",\n \"*p < 0 . 05 .\",\n \"n . s . = not significant .\",\n \"In all cases , data represent mean ± SEM and n ≥ 11 animals , 21 antennal lobes .\",\n \"Scale bar = 5 µm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 019 As we are limited to genetically accessible PN subsets , we focused on identifying organizational parameters in the PNs that innervate the DA1 and VA1d glomeruli via the Mz19-GAL4 driver ( Ito et al . , 1998 ) .\",\n \"As with ORN presynapses , the assay revealed that the number of AChR puncta scales with glomerular size ( Figure 7E ) .\",\n \"Further , known sex-specific differences in DA1 , as seen in glomerular volume ( Stockinger et al . , 2005 ) and in ORN synapses ( Figure 1 ) , were also observed ( Figure 7E ) .\",\n \"The differences between the Brp-Short and AChR assays for the DA1 ( 1632 ± 22 for Brp-Short vs 2007 ± 48 for Dα7 ) and VA1d ( 1107 ± 18 for Brp-Short vs 1762 ± 38 for Dα7 ) glomeruli may reflect the fact that the Brp-Short assay does not distinguish ORN synapses onto PNs and LNs , and the Dα7-GFP assay does not distinguish synapses from ORNs and LNs onto PNs .\",\n \"As these values are less than twofold different , this is consistent with the majority of synapses labeled being ORN to PN synapses ( Kazama and Wilson , 2009; Wilson , 2013 ) .\",\n \"The similarity between the numbers of endogenous Brp and Brp-Short puncta suggests that Brp-Short is a more accurate estimator of absolute synapse number .\",\n \"The larger number of AChRs detected in each glomerulus may reflect an overestimation associated with full-length Dα7 overexpression or that these are postsynaptic not just to cholinergic ORNs , but also other excitatory neurons such as local interneurons ( Chou et al . , 2010 ) or PN-PN chemical synapses ( Ng et al . , 2002; Wilson et al . , 2004 ) .\",\n \"Calculation of AChR puncta density in PNs revealed subtle but significant differences across different glomeruli .\",\n \"In the VA1d glomerulus , the densities were identical between males and females ( Figure 7F ) .\",\n \"However , these were different from AChR puncta densities in the DA1 glomerulus .\",\n \"There was a modest but significant difference between both male and female AChR densities in DA1 and between both DA1 AChR densities and the shared VA1d AChR density ( Figure 7F ) .\",\n \"Unlike ORNs , where the Brp-Short density was identical across different classes of neurons , PNs can have different densities between distinct glomeruli and even between sexes for the same glomerulus .\",\n \"Thus , the parameters that govern presynaptic density may differ from those that govern postsynaptic density in the same glomerulus .\",\n \"To further examine if the Teneurins regulate postsynaptic acetylcholine receptor number and density , we utilized the Dα7-GFP assay to determine the effect of Teneurin perturbation on AChR puncta number .\",\n \"We examined PNs in DA1 and VA1d ( Figure 8A ) and counted the AChR puncta of both glomeruli together as one measurement , as partner matching defects following Teneurin perturbation make it difficult to differentiate between the two glomeruli ( Hong et al . , 2012 ) .\",\n \"In ten-a mutants ( Figure 8B ) , the number of AChR clusters in these glomeruli was decreased by 23% , compared to wild type ( Figure 8D ) .\",\n \"This is consistent with results from the Brp-Short assay .\",\n \"Moreover , PN neurite volume was unaffected ( Figure 8E ) , so AChR puncta density was similarly reduced ( Figure 8F ) .\",\n \"Thus , two independent assays , both pre- and postsynaptic , show the same phenotypes , demonstrating a clear effect of ten-a loss on synapse organization in olfactory neurons . 10 . 7554/eLife . 03726 . 020Figure 8 . Teneurins function to regulate acetylcholine receptor number .\",\n \"( A–C )\",\n \"Representative high magnification confocal z-stack images of DA1 and VA1d PNs expressing Dα7-GFP in control animals ( A ) , ten-a mutants ( B ) , or animals expressing RNAi against ten-m in the same PNs ( C ) , and stained with antibodies against GFP ( green ) and N-Cadherin ( blue ) .\",\n \"DA1 and VA1d glomeruli are outlined together ( white dashed line ) and were determined using the accompanying mtdT label ( not shown ) .\",\n \"Due to partner matching defects in the ten-a mutant that prevent clear delineation between the two glomeruli , the combined count was compared across genotypes .\",\n \"( D ) Quantification of Dα7-GFP puncta in the noted genotypes .\",\n \"( E ) Quantification of PN neurite volume in all genotypes .\",\n \"( F ) Quantification of AChR density in all genotypes .\",\n \"All observed phenotypes are statistically indistinguishable .\",\n \"Significance was assessed with a one-way ANOVA and corrected for multiple comparisons by a posthoc Tukey's multiple comparisons test .\",\n \"***p < 0 . 001 in comparison with control .\",\n \"n . s . = not significant .\",\n \"In all cases , data represent mean ± SEM and n ≥ 8 animals , 16 antennal lobes .\",\n \"Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 020 At the NMJ , presynaptic Ten-a functions largely in a transsynaptic , heterophilic complex with postsynaptic Ten-m to regulate synapse organization ( Mosca et al . , 2012 ) .\",\n \"Ten-a functions presynaptically in ORNs to ensure proper synapse number ( Figure 4 ) .\",\n \"Thus , we hypothesized that the loss of Ten-m in postsynaptic PNs should result in a similar phenotype .\",\n \"As the ten-m mutant is larval lethal ( Zheng et al . , 2011 ) , we expressed a previously validated transgenic RNAi line against ten-m ( Hong et al . , 2012; Mosca et al . , 2012 ) in Mz19 PNs and quantified AChR puncta number using the Dα7 assay ( Figure 8C ) .\",\n \"ten-m knockdown phenocopied the ten-a phenotype ( Figure 8D ) .\",\n \"As above , PN neurite volume was unaffected ( Figure 8E ) , leading to a concomitant decrease in AChR puncta density that also phenocopied the ten-a phenotype ( Figure 8F ) .\",\n \"Further , this reduction was not enhanced by knocking down ten-m in PNs of a ten-a null mutant ( Figure 8D–F ) , suggesting that the two function in the same genetic pathway .\",\n \"Though genetic evidence suggests that Ten-m functions with Ten-a transsynaptically to regulate synapse number , it could conceivably regulate AChR number independently of how Ten-a regulates active zone number .\",\n \"Thus , we sought to more explicitly test a transsynaptic role of Ten-m in regulating proper active zone number .\",\n \"Due to the availability of genetic access to both the presynaptic ORNs and the postsynaptic PNs , we examined the DA1 glomerulus .\",\n \"Further , DA1 is a ten-m low glomerulus ( Hong et al . , 2012 ) , allowing us to specifically test the role of the basal level of Ten-m .\",\n \"Using the GAL4 and QF systems , we simultaneously labeled Or67d-positive ORNs with QUAS Brp-Short-mStraw using Or67d-QF ( Liang et al . , 2013 ) and expressed transgenic RNAi against ten-m in DA1 PNs using Mz19-GAL4 ( Figure 9A ) .\",\n \"Control animals displayed a similar number of Brp-Short puncta as the UAS construct ( Figure 9B , D ) , demonstrating the validity of this reagent .\",\n \"Postsynaptic ten-m knockdown , however , reduced the number of presynaptic Brp-Short puncta by 20% ( Figure 9C , D ) .\",\n \"This reduction is similar to the proportion by which ten-m RNAi reduced AChR number ( Figure 8D ) and also to the proportion by which loss of ten-a reduced Brp-Short number ( Figure 4D ) .\",\n \"This suggests that Ten-m is the postsynaptic partner of Ten-a in regulating presynaptic active zone number , and that the presynaptic Ten-a/postsynaptic Ten-m interaction that regulates NMJ synapses is conserved in olfactory synapses . 10 . 7554/eLife . 03726 . 021Figure 9 . Ten-m regulates presynaptic active zone number transsynaptically .\",\n \"( A ) Diagram of the antennal lobe with the DA1 glomerulus outlined showing the experimental design .\",\n \"Or67d-positive ORN axon terminals that innervate DA1 are outlined in black; their active zones are labeled by Or67d-QF-driven QUAS-Brp-Short-mStraw .\",\n \"Mz19-GAL4-positive PNs that project dendrites to DA1 express UAS-RNAi against ten-m in experimental animals .\",\n \"( B–C )\",\n \"Representative high magnification confocal z-stack images of DA1 ORNs expressing Brp-Short in control animals ( B ) or in animals concurrently expressing ten-m RNAi in the DA1 and VA1d PNs ( C ) and stained with antibodies against mStraw ( red ) and N-Cadherin ( blue ) .\",\n \"( D ) Quantification of Brp-Short puncta in the noted genotypes .\",\n \"Controls 1 and 2 represent Brp-Sh puncta assayed by GAL4/UAS ( as in Figure\",\n \"1 ) and QF/QUAS binary system binary system , respectively , and are not significantly different , as assayed by student's t test ( n . s . ) .\",\n \"Significance of ten-m PN RNAi was assayed by student's t test between it and Control 2 .\",\n \"***p < 0 . 001 .\",\n \"In all cases , data represent mean ± SEM and n ≥ 5 animals , 10 antennal lobes .\",\n \"Scale bar = 5 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 03726 . 021\"\n ],\n [\n \"In Drosophila , previous approaches to studying central synapses used tagged synaptic vesicle proteins ( Ito et al . , 1998; Robinson et al . , 2002 ) to reveal putative synaptic sites .\",\n \"While consistent with synapses , they could also stain non-synaptic , trafficking vesicles .\",\n \"We utilized a structural active zone component , Brp ( Wagh et al . , 2006 ) .\",\n \"By expressing a truncated Brp transgene , Brp-Short ( Schmid et al . , 2008 ) , using the GAL4/UAS system , this approach can label synapses in any neurons with genetic access .\",\n \"Brp-Short expression requires endogenous Brp for proper localization ( Figure 1—figure supplement 1 ) and does not alter function or morphology ( Fouquet et al . , 2009; Kremer et al . , 2010 ) .\",\n \"Thus , it accurately reports endogenous active zones .\",\n \"Recently , an elegant technique , STaR , was developed , which tags an additional , BAC-sourced copy of Brp with an epitope tag whose expression is conditional upon FLP-recombinase-mediated excision of an intervening stop codon ( Chen et al . , 2014 ) .\",\n \"An important advantage of STaR is that Brp expression is controlled by its endogenous promoter , thus guarding against mislocalization of Brp or perturbation of synaptic function due to overexpression .\",\n \"A caveat of Brp-Short is that it is controlled by GAL4 and thus the levels may be different from the endogenous level .\",\n \"While Brp-Short overexpression does not interfere with synaptic function , care must be taken not to overexpress it to a level that saturates the active zone localization machinery when utilized in new cell types .\",\n \"The advantages of Brp-Short over STaR are\",\n \"( a ) that it does not require a cell type-specific FLP transgene , which is not as widely available as GAL4 lines , or a BAC-bearing copy of Brp , and\",\n \"( b ) that it can be co-expressed with UAS-transgenes for rescue or RNAi for perturbation experiments , although STaR can also achieve this aspect by using a cell-type-specific GAL4 and an extra UAS-FLP transgene .\",\n \"To examine putative postsynaptic acetylcholine receptor clusters , we use a GFP-tagged subunit , Dα7 , to study cholinergic synapses in the antennal lobe .\",\n \"This transgene has been used to examine synaptic organization ( Leiss et al . , 2009a , 2009b; Kremer et al . , 2010; Christiansen et al . , 2011 ) .\",\n \"Though false positives can be associated with full-length protein overexpression , our observation of endogenous Brp puncta apposed to these Dα7-GFP puncta suggests that these receptors are properly localized to endogenous synapses .\",\n \"This assay complements the Brp-Short presynaptic assay and can be adapted for other tagged postsynaptic receptor transgenes .\",\n \"Though considerable advances have been made in our understanding of the wiring specificity ( Hong and Luo , 2014 ) and physiological properties ( Wilson , 2013 ) of the Drosophila olfactory circuit in the antennal lobe , little is known about its synaptic organization .\",\n \"As this system has emerged as a model neural network , a detailed mapping of synaptic organizational principles is integral towards advancing the study of circuit dynamics .\",\n \"Indeed , the juxtaposition of distinct types of synapses between multiple neuronal classes in the antennal lobe provides a model to study complex synaptic interactions compared to a neuromuscular synapse that features only two synaptic partners: the motoneuron and the muscle .\",\n \"We utilized the Brp-Short and Dα7 assays to probe how synapses in ORNs , PNs , and LNs are organized in the antennal lobe with respect to their number , density , and location .\",\n \"This work offers the first comprehensive information on ( 1 ) the number of active zones made by each ORN within a glomerulus , ( 2 ) the stereotypy of synapse numbers between those individual ORNs , ( 3 ) the prominence of PN presynaptic inputs within a glomerulus , suggesting a robust feedback mechanism , and ( 4 ) the relatively small contribution of LN active zones to the antennal lobe circuit .\",\n \"Our analyses suggest distinct rules that govern the synaptic organization of antennal lobe neurons .\",\n \"ORNs , the primary input neurons of the olfactory system , are diverse in their olfactory receptor expression , ligand specificity , and glomerular targeting specificity ( Liang and Luo , 2010 ) ; we now show that they also differ in the absolute synapse number ( Figure 1 ) .\",\n \"However , despite such differences , all five classes we examined ( DA1 , VA1d , VA1lm , DL4 , and DM6 ORNs ) have identical synaptic densities , suggesting that this represents a general rule for other ORN classes .\",\n \"This may further suggest that the primary job of the ORN is to convey information from the environment into the system as faithfully as possible , that all information is treated equally at this level , and that weighting computations occur downstream in the brain ( Masse et al . , 2009; Wilson , 2013 ) .\",\n \"Indeed , our analyses indicate that each ORN makes an equivalent , discrete number of synapses within a given glomerulus with little variation ( Figure 2 ) , further supporting this hypothesis .\",\n \"Interestingly , the density of postsynaptic receptors differs between glomeruli and even within the same glomerulus between sexes ( Figure 7 ) .\",\n \"This variation can be due to technical caveats , such as Mz19-GAL4 does not label all PNs that innervate DA1 and VA1d ( Jefferis et al . , 2001; Liang et al . , 2013 ) , or that Dα7-GFP clusters do not reflect the absolute number of AChRs .\",\n \"As the relative numbers still show these differences , an interesting possibility suggested by these results is that postsynaptic PN AChRs already reflect a transformed olfactory representation compared to output synapses of ORNs .\",\n \"The difference in PN AChR density as compared to the constant density of ORN active zones suggests that different classes of PNs may modulate how information is received by regulating the number of acetylcholine receptors .\",\n \"This can thus contribute to the transformation of olfactory representation by antennal lobe neurons ( Wilson , 2013 ) .\",\n \"Fine-scale analysis of synapse localization within projections of ORNs , LNs , and PNs ( Figure 3 ) suggested that these three types of neurons differ in their subglomerular organization .\",\n \"While occupying the vast majority of territory throughout the entire glomerulus , ORN processes and synapses leave distinct voids .\",\n \"A significant proportion of LNs form synapses in these voids , likely to other LNs or PNs .\",\n \"However , there is also overlap between LN and ORN processes and synapses , consistent with physiologically characterized ORN → LN and LN → ORN synapses ( Olsen et al . , 2007 , 2010; Root et al . , 2007; Kazama and Wilson , 2008 , 2009; Olsen and Wilson , 2008; Root et al . , 2008; Huang et al . , 2010; Yaksi and Wilson , 2010; Root et al . , 2011; Wilson , 2011 , 2013 ) .\",\n \"Within their respective neurites , ORNs and PNs display uneven distributions and synaptic clusters to varying degrees while active zones in LNs are more evenly distributed throughout their processes .\",\n \"These characterizations contribute to the growing repertoire of studies seeking to understand synaptic organization in Drosophila olfactory circuits ( Murthy et al . , 2008; Caron et al . , 2013; Fisek and Wilson , 2014 ) , adding synapse-level imaging to physiological techniques .\",\n \"Finally , the existence of synaptic organization parameters detailing number and location suggests molecular mechanisms designed to enforce those rules , both at cellular and circuit levels .\",\n \"Indeed , such analysis represents an integral part of neuronal circuit analysis , as recent work on retinal neurons has shown that accounting for synapse position is a critical aspect of modeling connectivity ( Kim et al . , 2014 ) .\",\n \"Our data demonstrate that the Teneurins , a family of transsynaptic adhesion molecules , regulate one of these synaptic organizational paradigms: synapse number .\",\n \"Beyond molecules like RPTPs ( Takahashi and Craig , 2013 ) and Wnts ( Park and Shen , 2012 ) , there is little known conservation between the organization mechanisms of central synapses and the NMJ .\",\n \"In vertebrates , previous studies have identified a number of synaptogenic signaling and cell adhesion molecules in the CNS ( Gerrow and El-Husseini , 2006; Missler et al . , 2012 ) , but in many cases , their roles at the NMJ are either minimal or unknown .\",\n \"Likewise , the roles of pathways including Rapsyn , Dok7 , MuSK , and Tid1 , which are well established at the NMJ ( Wu et al . , 2010 ) , have no well-established roles at CNS synapses .\",\n \"Among the identified central synaptogenic molecules , no master controller of synapses , like Agrin , has been discovered ( Shen and Scheiffele , 2010 ) .\",\n \"In the mammalian CNS , synaptic adhesion molecules like Neurexin and Neuroligin have demonstrated organizational roles ( Sudhof , 2008 ) but their role ( if any ) at the NMJ is largely unknown .\",\n \"In Drosophila , considerable work has been done at the NMJ to understand synapse formation and organization ( Menon et al . , 2013 ) .\",\n \"For example , neuromuscular Neurexin and Neuroligin regulate synaptic development and assembly ( Li et al . , 2007; Banovic et al . , 2010; Sun et al . , 2011; Chen et al . , 2012; Mosca et al . , 2012; Owald et al . , 2012 ) , but remain ( as with many other identified molecules ) largely untested in the CNS due to the absence of techniques for doing so .\",\n \"Our approaches have now enabled such an examination , uncovering a strongly conserved synaptic organization function of the Teneurins from PNS to CNS .\",\n \"Recent work has shown that these evolutionarily conserved proteins are involved in synaptic partner matching between neurons in the Drosophila olfactory system and between muscles and motoneurons at the NMJ ( Hong et al . , 2012; Mosca et al . , 2012 ) .\",\n \"As there are marked differences between central and peripheral synapses ( like the NMJ ) , it is further unclear whether the mechanisms would be conserved or if they would be wholly different .\",\n \"Here we find , as assayed by Brp-Short ( Figure 4 ) , ultrastructural ( Figure 5 ) , and Dα7 ( Figure 8 ) analyses , that Teneurins in olfactory neurons are required for normal synapse number .\",\n \"Teneurin perturbation also reduces synaptic density , a parameter that is highly invariable for ORNs under normal conditions ( Figure 1I ) .\",\n \"We determined that presynaptic Ten-a in ORNs ( Figure 4 ) likely functions with postsynaptic Ten-m in PNs ( Figures 8 , 9 ) to regulate levels of the spectrin cytoskeleton ( Figure 6 ) .\",\n \"Our evidence is consistent with spectrin and the Teneurins functioning in the same genetic pathway to regulate synapse organization and density .\",\n \"The perturbation of active zone number in ORNs by knocking down ten-m in PNs further suggests that Teneurins regulate synapse number and organization in the CNS via a transsynaptic mechanism .\",\n \"This highlights further conservation between central and peripheral synapse organization in the use of Teneurins .\",\n \"There are , however , some differences between the CNS and the PNS regarding the Teneurins .\",\n \"While presynaptic ten-m has a minor role in synaptic organization at the NMJ ( Mosca et al . , 2012 ) , our data suggests the lack of such a role in the CNS ( Figure 4D ) .\",\n \"Thus , these different systems may also use the Teneurins differently .\",\n \"Mammalian Teneurins organize the visual system ( Leamey et al . , 2007; Suzuki et al . , 2012; Antinucci et al . , 2013; Young et al . , 2013 ) and Ten-2 can serve as a ligand for Latrophilin and localize to synapses in cultured neurons ( Silva et al . , 2011; Boucard et al . , 2014 ) .\",\n \"However , as proper synaptic function is impaired in many neuropsychiatric disorders ( Guilmatre et al . , 2014 ) and human Teneurin-4 is associated with increased susceptibility to bipolar disorder ( Psychiatric GWAS Consortium Bipolar Disorder Working Group , 2011 ) , understanding how the Teneurins regulate central synapses is a question with clinical relevance .\",\n \"Studies in vivo will be important to determine how mammalian Teneurins regulate synaptic organization and whether the different Teneurins can have specific roles at synapses .\",\n \"In summary , our results demonstrate a role for the Teneurins in regulating the number of central synapses and highlight mechanistic conservation between peripheral and central synapse formation .\",\n \"Moreover , the fact that ORNs can be mistargeted but still have the correct number of synapses suggests that target choice and synapse organization can be biologically separable , even when they employ the same molecules .\"\n ],\n [\n \"The following GAL4 driver lines were used to restrict expression to individual classes of neurons: AM29-GAL4 ( DL4 and DM6 ORNs; Endo et al . , 2007 ) , Or47b-GAL4 ( VA1lm ORNs; Vosshall et al . , 2000 ) , Or67d-GAL4 ( DA1 ORNs; Kurtovic et al . , 2007 ) , Or88a-GAL4 ( VA1d ORNs; Vosshall et al . , 2000 ) , Mz19-GAL4 ( DA1 , DC3 , and VA1d PNs; Jefferis et al . , 2004 ) , Pebbled-GAL4 ( all ORNs; Sweeney et al . , 2007 ) , LN5 ( NP3056 ) -GAL4 ( ∼50 pan-glomerular LNs; Chou et al . , 2010 ) .\",\n \"The GH146-QF line ( Potter et al . , 2010 ) was used to genetically access 2/3 of all PNs .\",\n \"The following transgenes were used for labeling or genetic manipulation: UAS-Brp-Short-mStraw ( Fouquet et al . , 2009 ) , UAS-Brp-Short-EGFP ( Schmid et al . , 2008 ) , UAS-Dα7-GFP ( Leiss et al . , 2009b ) , UAS-3xHA-mtdT ( Potter et al . , 2010 ) , UAS-mCD8-GFP ( Lee and Luo , 1999 ) , UAS-DSyd1-EGFP ( Owald et al . , 2010 ) , UAS-α-spectrin-dsRNA ( Pielage et al . , 2005 ) , UAS-β-spectrin-dsRNA ( Pielage et al . , 2005 ) , UAS-Ten-a ( Mosca et al . , 2012 ) , UAS-Brp-IR-104422 ( Vienna Drosophila RNAi Center ) , UAS-Dcr2 ( Dietzl et al . , 2007 ) , UAS-ten-a-RNAi-V32482 , UAS-ten-m-RNAi-V51173 ( Hong et al . , 2012; Mosca et al . , 2012 ) , QUAS-mCD8-GFP ( Potter et al . , 2010 ) , QUAS-Ten-a ( Hong et al . , 2012 ) , and QUAS-Brp-Short-mStraw ( see below ) .\",\n \"The Df ( X ) ten-a line was used as a ten-a null mutant ( Hong et al . , 2012 ) .\",\n \"Brp-D3-mStraw ( Brp-Short-mStraw ) was amplified by PCR from the pTWmStraw-Brp-D3 vector ( Owald et al . , 2010; a kind gift of Stephan Sigrist ) using the forward primer 5ʹ-CACCATGGGAACTAGTGAC-3ʹ and the reverse primer 5ʹ-CTAGCTTACGTCACG-3ʹ with a CACC appended to the 5ʹ end of the forward primer for subcloning into the Gateway ( Invitrogen , Carlsbad , CA ) pENTR/D-TOPO vector .\",\n \"Following a BP reaction to produce pENTR-Brp-Short-mStraw , this plasmid was recombined into the destination vector pQUAST-attB-Gateway ( Mosca et al . , 2012; a kind gift of Vincenzo Favaloro ) using LR clonase .\",\n \"The resulting pQUAST-attB-Brp-Short-mStraw vector was transformed into the ΦC31 landing site 86Fb on the third chromosome using standard methods .\",\n \"hsFLP MARCM analyses were conducted as described ( Lee and Luo , 1999; Joo et al . , 2013 ) with the following modification: to obtain small DL4 and DM6 ORN clones , animals were heat-shocked between 48 hr and 72 hr after egg-laying at 37°C for 20 min .\",\n \"Adult brains were dissected at 10 days post-eclosion and fixed , stained , and processed as previously described ( Wu and Luo , 2006 ) .\",\n \"The following antibodies were used: rat anti-N-Cadherin [DN-EX#8; 1:40 , Developmental Studies Hybridoma Bank ( DSHB ) ] , mouse anti-Brp ( NC82; 1:40 , DSHB ) , rabbit anti-dsRed ( Living Colors DsRed Polyclonal Antibody , 1:250 , Clontech ) , chicken anti-GFP ( GFP-1020; 1:1000 , Aves Lab ) , mouse anti-α-spectrin ( 3A9; 1:50 , DSHB ) , rabbit anti-β-spectrin ( 1:500 , gift of R Dubreuil ) .\",\n \"Alexa488- , Alexa568- , and Alexa647-conjugated secondary antibodies were used at a concentration of 1:250 ( Invitrogen , Carlsbad , CA ) .\",\n \"Samples were mounted in SlowFade Gold Antifade medium ( Life Technologies , Grand Island , NY ) by bridge mount under #1 . 5 cover slips .\",\n \"Images were obtained on a Zeiss LSM510 Meta confocal laser-scanning microscope ( Carl Zeiss , Oberkochen , Germany ) using a 63× 1 . 4 NA PlanApo lens at an optical zoom of 3× .\",\n \"Images were centered on the glomerulus in question and the Z-boundaries set according to the confines of the synaptic label ( Brp-Short-mStraw or Dα7-GFP ) .\",\n \"System gain was set to minimize saturation but maintain a high signal-to-noise ratio .\",\n \"Selected raw image stacks are available at http://web . stanford . edu/group/luolab/ALsynapse . shtml .\",\n \"Raw image stacks for quantification were first deconvolved using AutoQuant X3 software ( Media Cybernetics , Rockville , MD ) on a custom-built computer ( Digital Storm , Fremont , CA ) .\",\n \"Specifically , the images were deconvolved using 10 iterations with settings of medium noise , and a blind point spread function .\",\n \"Following deconvolution , images were visually inspected to ensure that striping , ringing , or discontinuity artifacts were not introduced .\",\n \"Further , each image was visually examined to ensure that the borders of puncta were sharpened by the deconvolution and that the membrane staining remained continuous .\",\n \"Following , images were imported into Imaris 7 . 4 . 1 ( Bitplane AG , South Windsor , CT ) for quantification .\",\n \"Though deconvolution improved the visual quality of the images , it was dispensable for accurate quantification of Brp-Short or Dα7 puncta .\",\n \"The ‘Spots’ function was used to quantify Brp-Short or Dα7 puncta in Imaris .\",\n \"A region of interest was drawn around the glomerulus being quantified .\",\n \"For individually labeled glomeruli , the standard Imaris box was sufficient to mark the region of interest .\",\n \"Where multiple glomeruli were labeled and a single glomerulus was quantified , a freehand region of interest was drawn in each section using the neuropil staining as a guideline and a 3D region extrapolated by Imaris .\",\n \"The spot size ( puncta diameter ) was set at 0 . 6 μm for Brp-Short and 0 . 8 μm for Dα7 , as determined by direct measurement of images and previous studies ( Kremer et al . , 2010 ) .\",\n \"Background subtraction was selected .\",\n \"Following automatic detection of puncta , the threshold was manually adjusted ( usually within 5% among different samples ) with visual inspection to ensure that all puncta were identified , minimal background staining or noise was recognized , and that individual puncta were not double-counted .\",\n \"In all , the resultant count was recorded as the number of puncta .\",\n \"The accuracy was assessed by comparing puncta detection figures to glomeruli that were counted by hand following 3D projection .\",\n \"In all cases , there was less than 5% difference , supporting the accuracy of the semi-automatic quantification method .\",\n \"In male Or67d-GAL4 samples that did not express the synaptic label ( normally have ∼1500 puncta ) , but were processed identically using primary and secondary antibodies , the number of puncta recognized by this method was <30 , suggesting that the ‘false positive’ rate was low .\",\n \"To ensure that user-introduced noise in threshold adjustment did not introduce significant bias , each image was quantified a second time , at a later time , without access to the prior quantification .\",\n \"If the difference exceeded 5% , the sample was discarded .\",\n \"Further , if the averages of an entire genotype were significantly different from each other ( as determined by student's t test ) , the entire cohort was discarded and the experiment repeated .\",\n \"For the nearest neighbor distance ( NND ) analysis , the ‘DistMin’ was calculated in Imaris using the ‘Spots to Spots Closest Distance XTension’ plug-in for Matlab .\",\n \"Cumulative frequency histograms were compiled in Prism 6 . 04 ( GraphPad Software , San Diego , CA ) .\",\n \"As each puncta is ∼0 . 6 μm in diameter , the minimum NND should be 0 . 6 .\",\n \"Indeed , few measurements were seen below 0 . 6 .\",\n \"Thus , ‘clustering’ was defined as puncta with an NND of less than 25% of the diameter .\",\n \"The ‘% Cluster’ was calculated by expressing the number of puncta with an NND value ≤0 . 75 divided by the total number of puncta .\",\n \"To obtain neurite volume based on mCD8-GFP or mtdT labeling , the ‘Surfaces’ function was used .\",\n \"Regions of interest were drawn as above .\",\n \"Background subtraction was set to 3 μm and the smoothing to 0 . 2 μm; these settings optimally reflected actual neurite labeling .\",\n \"Automatic detection with a 10e5 threshold was then used to remove background , detecting only the glomerulus in question .\",\n \"For smaller objects ( MARCM clones ) , a 10e1 threshold was used and only the specific objects that directly corresponded to mCD8-GFP labeling were selected and counted .\",\n \"Volume was then calculated by Imaris .\",\n \"Density was then expressed as the number of Brp-Short puncta divided by the neurite volume .\",\n \"For images involving a neurite co-stain , the Brp-Short-mStraw channel was masked in Imaris so that only the puncta within the neurite were identified and counted ( not background puncta resulting from secondary antibodies , or image noise ) .\",\n \"Images were processed using Photoshop CS4 ( Adobe , San Jose , CA ) and levels or contrast adjusted identically for all cases of comparison .\",\n \"For cases of immunofluorescence comparison , brains were stained in parallel , imaged on the same slide in the same session , and processed identically .\",\n \"Statistical and graphical analyses were completed in Excel ( Microsoft , Redmond , WA ) , Prism 6 . 04 ( GraphPad Software , San Diego , CA ) , and Matlab ( Mathworks , Natick , MA ) .\",\n \"Adult brains were dissected in 0 . 1 M cacodylate buffer and fixed at 4°C for 3 hr in 2 . 5% paraformaldehyde , 5% glutaraldehyde ( vol/vol ) , and 0 . 06% picric acid ( vol/vol ) in 0 . 1 M cacodylate buffer .\",\n \"They were then washed three times for 20 min each in 0 . 1 M cacodylate buffer on ice and post-fixed with 2% osmium tetroxide for 1 hr at 4°C .\",\n \"Dehydration , infiltration with resin , and staining with uranyl acetate were all done according to standard protocols .\",\n \"Antennal lobe regions were identified in thick section following 1% toluidine blue in 1% sodium borate staining and ultrathin sections taken from those regions .\",\n \"Sections were imaged on a JEM-1400 ( JEOL , Tokyo , Japan ) transmission electron microscope at X3 , 000 to X20 , 000 magnification .\",\n \"For immuno-EM , adult brains were dissected in 0 . 1 M sodium phosphate buffer and fixed in modified Zamboni's fixative ( 4% paraformaldehyde , 1 . 6% glutaraldehyde , 0 . 2% saturated picric acid in 0 . 1 M sodium phosphate buffer at pH 7 . 4 ) as described ( Kolodziejczyk et al . , 2008 ) .\",\n \"Brains were then washed in PBS , processed similarly as above , and sectioned .\",\n \"Sections were then blocked in PBST containing 0 . 05% Tween-20 , 0 . 5% ( wt/vol ) ovalbumin , and 0 . 5% ( wt/vol ) BSA .\",\n \"Sections were incubated overnight at 4°C in rabbit anti-GFP at 1:300 ( MBL International , Woburn , MA ) .\",\n \"Following , grids were washed in PBST and incubated in 5 nm gold-conjugated goat anti-rabbit secondary antibody ( Ted Pella , Redding , CA ) at 1:50 for 1 hr at RT .\",\n \"Grids were then washed in PBST , incubated in 8% glutaraldehyde for 15 min , washed again in water , and contrast stained with 2% uranyl acetate followed by lead citrate .\",\n \"Ultrastructural analysis was done as previously described ( Watts et al . , 2004; Mosca et al . , 2012 ) and T-bar defects quantified using consecutive sections .\",\n \"For pairwise comparisons , student's t test was used to assess statistical significance .\",\n \"For multiple comparisons , a one-way ANOVA was done and corrected for multiple comparisons with posthoc Tukey's multiple comparisons tests done between all genotypes .\",\n \"Analysis was done in Prism 6 . 0 . 4 ( Graphpad Software , San Diego , CA ) .\",\n \"For synapse contribution by individual ORNs , GraphPad Prism was used to fit individual Gaussian relationships to each separate peak for the ORN MARCM data ( Figure 2C–D , colored dashed lines ) .\",\n \"The Curve Fitting Tool in Matlab ( Mathworks , Natick , MA ) was used to determine the best-fit curves that corresponded to the aggregate data set ( Figure 2C–D , solid black line ) .\",\n \"In each case , multiple strong candidate relationships that recognized the first four peaks were identified using r2 and root mean square error ( RMSE ) values .\",\n \"To determine which candidate best fit the data , they were then compared using two posthoc tests: corrected Akaike's Information Criteria ( AICc ) and the F-test .\",\n \"AICc is advantageous as it penalizes distributions with more parameters , avoiding potential over-fitting , and is useful with binned data .\",\n \"The F-test was used to test specific null hypotheses between the models .\",\n \"To examine the distribution of the individual peaks ( i . e . , the ‘quantal’ nature ) , the standard deviation in the distances between the peaks ( S ) was calculated .\",\n \"To determine the probability of such an S value occurring , the aggregate DL4 and DM6 data were individually fit to a Poisson distribution or a Gaussian distribution using Matlab .\",\n \"Matlab was then used to generate 50 , 000 simulations selecting four peaks that fit those distributions .\",\n \"Also , a uniform probability distribution was used to select four peaks between 0 and the upper bound of the DL4 and DM6 data sets , respectively .\",\n \"In all , 300 , 000 simulations were conducted , and the number of events resulting in an S value less than that observed used to determine a probability of obtaining a more ordered set of four peaks than the observed data .\",\n \"In some cases , UAS-mCD8-GFP or UAS-3xHA-mtdT was included in the genotype ( but not shown in the image ) to ensure an equivalent number of constructs between control and mutant or rescue conditions .\",\n \"Figure 1: ( B–D ) w; UAS-Brp-Short-mStraw/UAS-mCD8-GFP; Or67d-GAL4 [Nr-1]/+; + .\",\n \"( H–I )\",\n \"DL4 and DM6 = w; AM29-GAL4/UAS-Brp-Short-mStraw , UAS-mCD8-GFP; +; + .\",\n \"DA1 = w; UAS-Brp-Short-mStraw/UAS-mCD8-GFP; Or67d-GAL4 [Nr-1]/+; + .\",\n \"VA1d = w , Or88a-GAL4/+ or Y; UAS-mCD8-GFP , UAS-Brp-Short-mStraw/+; +; + .\",\n \"VA1lm = w; Or47b-GAL4/UAS-Brp-Short-mStraw , UAS-mCD8-GFP; +; + .\",\n \"Figure 1—figure supplement 1: ( A ) w; UAS-Brp-Short-mStraw/+; Or67d-GAL4 [Nr-1]/UAS-DSyd1-EGFP; + .\",\n \"( B ) w , UAS-Dcr2; UAS-mCD8-GFP/UAS-Brp-Short-mStraw; Or67d-GAL4 [Nr-1]/+; + .\",\n \"( C ) w; UAS-mCD8-GFP , UAS-Brp-Short-mStraw/UAS-Brp-IR-v104422; Or67d-GAL4 [Nr-1]/+; + .\",\n \"( D ) w , pebbled-GAL4/+; +; UAS-Brp-Short-GFP/+; + .\",\n \"Figure 1—figure supplement 2: ( A ) w , Or88a-GAL4/+ or Y; UAS-Brp-Short-mStraw/+; +; + .\",\n \"( B ) w; Or47b-GAL4/UAS-Brp-Short-mStraw; +; + .\",\n \"( C and D ) w; AM29-GAL4/UAS-Brp-Short-mStraw; +; + .\",\n \"Figure 2: ( A–F ) hsFLP122 , UAS-mCD8-GFP; AM29-GAL4/UAS-Brp-Short-mStraw; FRT 2A/FRT2A , tubP-GAL80; + .\",\n \"Figure 2—figure supplement 1: ( A–L ) hsFLP122 , UAS-mCD8-GFP; AM29-GAL4/UAS-Brp-Short-mStraw; FRT 2A/FRT2A , tubP-GAL80; + .\",\n \"Figure 3: ( A ) w; UAS-Brp-Short-mStraw/+; Or67d-GAL4 [Nr-1]/+; + .\",\n \"( B ) w; Mz19-GAL4/UAS-Brp-Short-mStraw; +; + .\",\n \"( C ) w; UAS-Brp-Short-mStraw/+; NP3056-GAL4/+; + .\",\n \"( D , G ) w; UAS-Brp-Short-mStraw/UAS-mCD8-GFP; Or67d-GAL4 [Nr-1]/+; + .\",\n \"( E , H ) w; Mz19-GAL4 , UAS-mCD8-GFP/UAS-Brp-Short-mStraw; +; + .\",\n \"( F , I ) w; UAS-Brp-Short-mStraw/UAS-mCD8-GFP; NP3056-GAL4/+; + .\",\n \"Figure 3—figure supplement 1: ( A ) w; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/+; Or67d-GAL4 [Nr-1]/+; + .\",\n \"( B ) w; Mz19-GAL4/UAS-Brp-Short-mStraw , UAS-mCD8-GFP; +; + .\",\n \"( C ) w; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/+; NP3056-GAL4/+; + .\",\n \"Figure 4: ( A ) w; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-mCD8-GFP/+; + .\",\n \"( B ) w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; +; + .\",\n \"( C ) w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-Ten-a/+; + .\",\n \"( D–F )\",\n \"Control as in ( A ) .\",\n \"ten-a as in ( B ) .\",\n \"ten-a ORN RNAi = w , UAS-Dcr2; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-ten-a-IR-v32482/+; + .\",\n \"ten-a + Ten-a ORN as in C . ten-m ORN RNAi = w , UAS-Dcr2; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-ten-m-RNAi-V51173/+; + .\",\n \"ten-a + Ten-a PN = w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; GH146-QF , QUAS-mCD8-GFP/QUAS-Ten-a; + .\",\n \"Figure 4—figure supplement 1: ( A ) w; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-mCD8-GFP/+; + .\",\n \"( B ) w , UAS-Dcr2; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-ten-m-RNAi-V51173/+; + .\",\n \"( C ) w , UAS-Dcr2; Or47b-GAL4/UAS-Brp-Short-mStraw; UAS-ten-a-IR-v32482/+; + .\",\n \"( D ) w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; GH146-QF , QUAS-mCD8-GFP/QUAS-Ten-a; + .\",\n \"Figure 4—figure supplement 2: ( A–C ) w; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/AM29-GAL4; +; + .\",\n \"( D–F ) w , Df ( X ) ten-a; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/AM29-GAL4; +; + .\",\n \"Figure 5: Control = Canton S . ten-a = Df ( X ) ten-a; +; +; + .\",\n \"Figure 6: ( A ) Canton S . ( B ) Df ( X ) ten-a; +; +; + .\",\n \"( C ) w; Or47b-GAL4 , UAS-Brp-Short-mStraw/UAS-mCD8-GFP; UAS-mCD8-GFP/+; + .\",\n \"( D ) w; Or47b-GAL4 , UAS-Brp-Short-mStraw/UAS-α-spectrin-dsRNA; UAS-α-spectrin-dsRNA/+; + .\",\n \"( E ) w; Or47b-GAL4 , UAS-Brp-Short-mStraw/UAS-β-spectrin-dsRNA; UAS-β-spectrin-dsRNA/+; + .\",\n \"( F ) Control as in ( A ) and ten-a as in ( B ) .\",\n \"( G ) Control as in ( C ) , α-spec ORN RNAi as in ( D ) , β-spec ORN RNAi as in ( E ) , ten-a = w , Df ( X ) ten-a; Or47b-GAL4/UAS-Brp-Short-mStraw; +; + .\",\n \"ten-a + α-spec RNAi = w , Df ( X ) ten-a; Or47b-GAL4 , UAS-Brp-Short-mStraw/UAS-α-spectrin-dsRNA; UAS-α-spectrin-dsRNA/+; + .\",\n \"Figure 6—figure supplement 1: ( A ) Canton S . ( B ) Df ( X ) ten-a; +; +; + .\",\n \"( C ) w; UAS-Brp-Short-mStraw , UAS-mCD8-GFP/+; Or67d-GAL4 [Nr-1] , UAS-mCD8-GFP/+; + .\",\n \"( D ) w; UAS-Brp-Short-mStraw/UAS-α-spectrin-dsRNA; Or67d-GAL4 [Nr-1]/UAS-α-spectrin-dsRNA; + .\",\n \"( E ) w; UAS-Brp-Short-mStraw/UAS-β-spectrin-dsRNA; Or67d-GAL4 [Nr-1]/UAS-β-spectrin-dsRNA; + .\",\n \"( F ) Genotypes as in ( C ) , ( D ) , and ( E ) .\",\n \"Figure 7: ( B , E–F ) w; Mz19-GAL4 , UAS-3xHA-mtdT/+; UAS-Dα7-EGFP/+; + .\",\n \"( C ) w; Mz19-GAL4/+; UAS-Dα7-GFP/+; + .\",\n \"( D ) w , Or67d-QF; Mz19-GAL4/+; UAS-Dα7-GFP/QUAS-Brp-Short-mStraw; + .\",\n \"Figure 8: ( A ) w , UAS-Dcr2; Mz19-GAL4 , UAS-3xHA-mtdT/+; UAS-Dα7-GFP/+; + .\",\n \"( B ) w , Df ( X ) ten-a; Mz19-GAL4 , UAS-3xHA-mtdT/+; UAS-Dα7-GFP/+; + .\",\n \"( C ) w , UAS-Dcr2; Mz19-GAL4 , UAS-3xHA-mtdT/+ , UAS-Dα7-GFP/UAS-ten-m-IR; + .\",\n \"( D–F )\",\n \"Control as in ( A ) .\",\n \"ten-a as in ( B ) .\",\n \"ten-m PN RNAi as in C . ten-a −/− + ten-m PN RNAi = w , Df ( X ) ten-a; Mz19-GAL4 , UAS-3xHA-mtdT/UAS-Dcr2; UAS-Dα7-GFP/UAS-ten-m-IR; + .\",\n \"Figure 9: ( B ) w , Or67d-QF; +; QUAS-Brp-Short-mStraw/+; + .\",\n \"( C ) w , Or67d-QF; Mz19-GAL4/+; QUAS-Brp-Short-mStraw/UAS-ten-m-RNAi-V51173; + .\",\n \"( D ) Control ( UAS ) = w; UAS-Brp-Short-mStraw/+; Or67d-GAL4 [Nr-1]/+; + .\",\n \"Other genotypes as in B–C .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Understanding information flow through neuronal circuits requires knowledge of their synaptic organization .","In this study , we utilized fluorescent pre- and postsynaptic markers to map synaptic organization in the Drosophila antennal lobe , the first olfactory processing center .","Olfactory receptor neurons ( ORNs ) produce a constant synaptic density across different glomeruli .","Each ORN within a class contributes nearly identical active zone number .","Active zones from ORNs , projection neurons ( PNs ) , and local interneurons have distinct subglomerular and subcellular distributions .","The correct number of ORN active zones and PN acetylcholine receptor clusters requires the Teneurins , conserved transmembrane proteins involved in neuromuscular synapse organization and synaptic partner matching .","Ten-a acts in ORNs to organize presynaptic active zones via the spectrin cytoskeleton .","Ten-m acts in PNs autonomously to regulate acetylcholine receptor cluster number and transsynaptically to regulate ORN active zone number .","These studies advanced our ability to assess synaptic architecture in complex CNS circuits and their underlying molecular mechanisms ."],"string":"[\n \"Understanding information flow through neuronal circuits requires knowledge of their synaptic organization .\",\n \"In this study , we utilized fluorescent pre- and postsynaptic markers to map synaptic organization in the Drosophila antennal lobe , the first olfactory processing center .\",\n \"Olfactory receptor neurons ( ORNs ) produce a constant synaptic density across different glomeruli .\",\n \"Each ORN within a class contributes nearly identical active zone number .\",\n \"Active zones from ORNs , projection neurons ( PNs ) , and local interneurons have distinct subglomerular and subcellular distributions .\",\n \"The correct number of ORN active zones and PN acetylcholine receptor clusters requires the Teneurins , conserved transmembrane proteins involved in neuromuscular synapse organization and synaptic partner matching .\",\n \"Ten-a acts in ORNs to organize presynaptic active zones via the spectrin cytoskeleton .\",\n \"Ten-m acts in PNs autonomously to regulate acetylcholine receptor cluster number and transsynaptically to regulate ORN active zone number .\",\n \"These studies advanced our ability to assess synaptic architecture in complex CNS circuits and their underlying molecular mechanisms .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Just as progress in science relies on researchers communicating their findings to other people working in their field , our bodies rely on neurons being able to communicate with other neurons .","This is where structures called synapses come in: synapses allow signals to be passed from one neuron to another .","Neurons and synapses process information by forming circuits in the brain , but relatively little is known about how synapses develop or how they are organized within circuits .","Mosca and Luo have now examined a neural circuit in the fruit fly ( Drosophila ) that receives sensory information about smells in the environment , and then converts this information to signals which can be understood by other parts of the brain .","This particular circuit has previously been identified as a good model of how the brain processes information .","Mosca and Luo found that the synapses in this circuit were organized according to specific ‘rules’ that determined factors such as the quantity and location of synapses at different points in the circuit .","Additionally , it was found that the successful development of synapses required the involvement of two members of a family of proteins called the Teneurins: this family of proteins is involved in a variety of neurodevelopmental processes .","Teneurins have been implicated in bipolar disorder , and malfunctioning synapses are thought to be associated with a number of other mental health conditions , so the results of Mosca and Luo could lead to a better understanding of these conditions ."],"string":"[\n \"Just as progress in science relies on researchers communicating their findings to other people working in their field , our bodies rely on neurons being able to communicate with other neurons .\",\n \"This is where structures called synapses come in: synapses allow signals to be passed from one neuron to another .\",\n \"Neurons and synapses process information by forming circuits in the brain , but relatively little is known about how synapses develop or how they are organized within circuits .\",\n \"Mosca and Luo have now examined a neural circuit in the fruit fly ( Drosophila ) that receives sensory information about smells in the environment , and then converts this information to signals which can be understood by other parts of the brain .\",\n \"This particular circuit has previously been identified as a good model of how the brain processes information .\",\n \"Mosca and Luo found that the synapses in this circuit were organized according to specific ‘rules’ that determined factors such as the quantity and location of synapses at different points in the circuit .\",\n \"Additionally , it was found that the successful development of synapses required the involvement of two members of a family of proteins called the Teneurins: this family of proteins is involved in a variety of neurodevelopmental processes .\",\n \"Teneurins have been implicated in bipolar disorder , and malfunctioning synapses are thought to be associated with a number of other mental health conditions , so the results of Mosca and Luo could lead to a better understanding of these conditions .\"\n]","isConversational":false},"year":{"kind":"string","value":"2014"}}},{"rowIdx":85,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["neuroscience"],"string":"[\n \"neuroscience\"\n]","isConversational":false},"title":{"kind":"string","value":"Decoupling sensory from decisional choice biases in perceptual decision making"},"id":{"kind":"string","value":"elife-43994-v2"},"sections":{"kind":"list like","value":[["You ask a friend about the tilt of a canvas that is perfectly horizontal and she says that the top right corner is up .","What causes her inaccuracy ?","One possibility is that her sensory representation is biased and she perceives the canvas tilted .","Another possibility , however , is that under uncertainty about the orientation of the canvas , she prefers to choose up over down .","This situation exemplifies a major problem in the study of perception: perceptual decisions not only depend on the sensory evidence , but also on decisional components ( Green and Swets , 1966; Gold and Ding , 2013 ) .","When a decision maker needs to select an action from a set of alternatives , the tendency to choose one alternative over the others is known as choice bias ( Gold and Ding , 2013 ) .","Choice biases occur in perceptual tasks even in simple scenarios like the one described above , in which the stimuli carry similar levels of signal relative to a neutral point ( Newsome and Paré , 1988; Mareschal and Clifford , 2012; Jazayeri and Movshon , 2007; Milner et al . , 1992; Tadin et al . , 2003 ) .","Choice biases are extensively studied in relation to how current decisions are influenced by previous decisions , what is known as choice history biases ( Abrahamyan et al . , 2016; Akaishi et al . , 2014; Fründ et al . , 2014; Urai et al . , 2017; Fischer and Whitney , 2014; Braun et al . , 2018; St John-Saaltink et al . , 2016; Fritsche et al . , 2017; Hermoso-Mendizabal et al . , 2019 ) .","Depending on the perceptual scenario , people tend to repeat their previous choice ( Akaishi et al . , 2014; Braun et al . , 2018 ) , alternate between choices ( Fritsche et al . , 2017 ) or idiosyncratically repeat or alternate ( Urai et al . , 2017; Abrahamyan et al . , 2016 ) .","Whether choice history biases reflect a bias in the sensory evidence or in the decision process is under debate ( Akaishi et al . , 2014; Fischer and Whitney , 2014; St John-Saaltink et al . , 2016; Fritsche et al . , 2017 ) .","Unlike choice history biases , some choice biases are related to the overall idiosyncratic tendency to choose one alternative over the others , not conditioned by the previous choices .","We will refer to these as global choice biases .","The existence of these biases is acknowledged ( Gold and Ding , 2013; Kingdom and Prins , 2016; Morgan et al . , 2012; García-Pérez and Alcalá-Quintana , 2013; Peters et al . , 2016 ) and they are included in current models of perceptual decision making ( Abrahamyan et al . , 2016; Akaishi et al . , 2014; Urai et al . , 2017; Braun et al . , 2018; Hermoso-Mendizabal et al . , 2019 ) , but whether they reflect sensory or decisional processes has not been , to our knowledge , assessed ( we searched in Google Scholar several times , last one on February 2019 , the following keywords: choice biases , response biases , motor biases , perceptual biases and sensory biases; we identified the relevant articles and searched within the references cited; we also tracked the articles that cited the relevant articles ) .","The problem to identify their origin is that , in standard perceptual paradigms , the tendency to choose one alternative more often is consistent with a biased sensory representation , but also with a bias in the decision process ( García-Pérez and Alcalá-Quintana , 2013; Gold and Ding , 2013 ) .","Here , to disentangle the contribution of sensory and decisional processes to global choice bias , we measured choice behavior in a simple common perceptual discrimination task with two symmetric alternatives , but under two different task instructions .","The instructions varied the response mapping between perception and the category of the alternatives ."],["On a given trial of the symmetric task ( the model for the asymmetric task is described in Supplementary Models ) a grating with orientation θi that could be clockwise or counterclockwise , is presented and the participant decides whether its orientation is consistent with clockwise acw or counterclockwise accw orientation .","A simple standard signal detection theory ( SDT ) model assumes that: ( 1 ) the sensory evidence s associated to the stimulus is a random variable normally distributed with variance σ2 centered on μcw when θi is clockwise and centered on μccw when θi is counterclockwise; ( 2 ) to make choices , the participant sets a criterion β , computes the log likelihood ratio logΛ of s under the hypothesis that μ=μcw and the hypothesis that μ=μccw ( 1 ) logΛ ( s ) =log ( p ( s|μ=μcw ) p ( s|μ=μccw ) ) =log ( ( 2πσ2 ) −1/2e− ( s−μcw ) 22σ2 ( 2πσ2 ) −1/2e− ( s−μccw ) 22σ2 ) and chooses the action acw when logΛ ( s ) >β .","In standard SDT models the origin and scale of the sensory evidence is arbitrary ( Wickens , 2001; Green and Swets , 1966 ) .","A typical choice is ( Wickens , 2001; Green and Swets , 1966 ) ( 2 ) μcw=d′2μccw=−d′2σ=1which results in logΛ ( s ) =d′s .","That is , one can use directly the evidence as a decision variable and choose acw when s>c=β/d′ ( Wickens , 2001; Green and Swets , 1966 ) .","The probability of choosing clockwise when θi is clockwise is ( 3 ) P ( a=acw;μ=μcw ) =12π∫c∞e− ( s−μcw ) 22ds=Φ ( μcw−c ) where Φ is the standard cumulative normal function .","To relate this probability to the magnitude of the stimulus , this standard SDT could be extended to include how the stimulus is transduced into sensory evidence ( García-Pérez and Alcalá-Quintana , 2013; Schneider and Bavelier , 2003; García-Pérez and Peli , 2014; Schneider and Komlos , 2008 ) .","Adding the transduction provides a meaningful origin and scale to the sensory evidence .","Assuming that the transduction is linear ( 4 ) μcw=aθi+bthe probability of choosing clockwise as a function of θi becomes ( 5 ) P ( a=acw;θi ) =Φ ( aθi+b−c ) =Φ ( θi− ( δD−δS ) σ ) Where ( 6 ) δS=ba=sensorybiasδD=ca=decisionalbiasσ=1awhich corresponds to a psychometric function centered on δD−δS with slope given by σ .","Consider the trials in which the grating is presented in orientations around the horizontal orientation and the reference is on the right ( the reasoning that follows also holds for vertical orientations ) .","If the orientation of the grating can take M different values , each value is presented n times and ki is the number of times that the participant pressed the down key ( clockwise ) , then the probability density function that defines statistical model for these trials is ( 7 ) f ( k1 , … , kM;δS , δD , σ ) =∏i=1M ( nki ) Φ ( θi− ( δD−δS ) σ ) ki ( 1−Φ ( θi− ( δD−δS ) σ ) ) n−ki Given that f depends on δD−δS , it is not possible to distinguish sensory from decisional biases ( Witt et al . , 2015; Schneider and Komlos , 2008; García-Pérez and Alcalá-Quintana , 2013 ) .","Consider now the trials for the left reference .","A criterion c associated to a bias to choose responses consistent with clockwise orientation for the right reference corresponds to a criterion −c for the left reference .","The statistical model that incorporates responses for the two locations of the reference is ( 8 ) f ( k1 , … , k2M;δS , δD , σ ) =∏i=1M ( nki ) ( 1−ΦR ) n−ki∏i=M+12M ( nki ) ΦLki ( 1−ΦL ) n−kiwhere ( 9 ) ΦR=Φ ( θi− ( δD−δS ) σ ) ΦL=Φ ( θi− ( −δD−δS ) σ ) and the index i larger than M refers to the trials for the left reference .","Now , by fitting two psychometric functions conjointly , δD and δs are not degenerated and could be estimated .","This is the exact model that we fitted for 17 of the 34 groups ( 17 participants that discriminate orientations around the horizontal and the vertical orientation in Experiment 1 ) .","For the rest of the groups , we also fitted this basic model , but adding some extra features that significantly improved the quality of the fit ( using likelihood ratio tests , see Materials and methods ) .","The first feature is related to the variability of the sensory evidence σ .","The basic model considers that this variability does not depend on where the participants imagined the reference , which results in the two psychometric functions having the same slope .","We found this to be the case in 28 of the 34 fits , but in six groups we found that considering different variabilities ( σ1 and σ2 ) for the two locations of the reference was better .","It might be that orienting attention to different parts of the visual field might change the uncertainty in the sensory evidence for some participants .","The other feature that we added consisted in adding lapses , which are responses that are incorrect independently of the sensory evidence ( Kingdom and Prins , 2016; Gold and Ding , 2013 ) .","This might occur , for example , when the participant misses the stimulus because of blinking or a loss of attention .","To incorporate lapses into the basic model , ΦR need to be replaced by λ1+ ( 1−λ1−λ2 ) ΦR and ΦL by λ1′+ ( 1−λ1′−λ2′ ) ΦL ( Kingdom and Prins , 2016 ) .","From the 34 groups , adding lapses improved the fit in 14 cases .","From those , three fits required the four lapse parameters , 4 fits required two lapse parameters ( one for each psychometric function , λ1=λ2 ; λ1′=λ2′ ) and 7 fits required one lapse rate ( λ1=λ2=λ1′=λ2′ ) .","Importantly , including a lapse rate parameter λ∗ placed on the right asymptote for the right reference ( 10 ) λ1+λ∗+ ( 1−λ1−λ1−λ∗ ) ΦRand placed on the left asymptote for the left reference ( 11 ) λ2+ ( 1−λ2−λ2−λ∗ ) ΦLdid not improve significantly the fit for any group , which indicates that decisional biases cannot be explained by a tendency of participants to select one alternative completely independently of the sensory evidence .","Perceptual tasks with two symmetric alternatives have been also modeled using a high threshold model , called the indecision model ( García-Pérez and Alcalá-Quintana , 2013; García-Pérez and Peli , 2014 ) .","This model is similar to the SDT model described , but divides the sensory axis in three regions delimited by c1 and c2 .","When the sensory evidence is lower than c1 the participant chooses accw and when is larger than c2 , chooses acw .","Critically , when the sensory evidence lies between c1 and c2—called interval of uncertainty—the model assumes that the observer is undecided and guesses the response ( choosing accw with probability ξ ) .","The probability of choosing clockwise when θi is clockwise and the reference is on the right is ( 12 ) P ( a=acw;μ=μcw ) =12π∫c2∞e− ( s−μcw ) 22ds+ξ12π∫c1c2e− ( s−μcw ) 22ds=Φ ( μcw−c2 ) +ξ ( Φ ( c2−μcw ) −Φ ( c1−μcw ) ) For the left reference , assuming that c1 and c2 do not change , ξ should be replaced by 1−ξ .","We fitted this model to the 19 groups with significant decisional biases and found that for all of them ( except for participant eight for the vertical condition in Experiment","1 ) the SDT model was better using the Akaike information criterion ( Burnham and Anderson , 2004 ) .","Given that the indecision model has two parameters more than the SDT model ( c is replaced by c1 and c2 and ξ is introduced ) , we also compared the SDT model to a simplified version of the indecision model with symmetric boundaries c1=−c2 ( García-Pérez and Alcalá-Quintana , 2013; García-Pérez and Peli , 2014 ) .","In this case , the SDT model was better for all the groups ."],["We showed that , in a simple discrimination task with two symmetric alternatives , most people exhibit idiosyncratic global choice biases .","Changing the stimulus-response mapping and testing a task with two asymmetric alternatives , we found that these biases reflect biases in the sensory evidence and in the decisional process .","Specifically , about half of the participants showed biases consistent with only a sensory origin and about half of the participant biases with a contribution of sensory and decisional biases .","We also found that a very few participants exhibited biases consistent with only a decisional origin .","Across participants , the magnitude of the sensory bias was about three times as larger as the magnitude of the decisional bias .","Our findings suggest that if a person has a bias to report that a perfectly horizontal canvas is tilted towards a given direction , it is likely that the largest contribution to the bias has a sensory origin .","To our knowledge , these idiosyncratic sensory biases have not been linked to asymmetries in the neural representation of the stimulus .","Possible asymmetries include a difference in the number of neurons tuned to clockwise and counterclockwise orientation if the decoding is of the population-vector type or a difference in the gain of neurons tuned to clockwise and counterclockwise orientation for more general decoding schemes ( Dayan and Abbott , 2001; Schwartz et al . , 2007 ) .","More recently , it has been proposed that the asymmetries could also emerge from the dynamics of competing neural networks ( Lebovich et al . , 2018 ) .","Importantly , for these asymmetries to bias perception , the downstream decoding mechanisms should have not been calibrated with the environmental property to compensate the asymmetries ( decoding ambiguity; Schwartz et al . , 2007 ) .","An SDT model , in which the sensory biases were included as an intercept term in a linear transduction of the stimuli ( García-Pérez and Alcalá-Quintana , 2013; Schneider and Bavelier , 2003; García-Pérez and Peli , 2014; Schneider and Komlos , 2008 ) and decisional biases correspond to a shift in the criterion , fitted well the choice behavior of the participants .","A model in which participants have a tendency to choose one alternative independently of the magnitude of the sensory evidence did not fit well the data; it rather predicts an asymmetry on the lapse rate across locations of the reference that was not observed .","Finally , a model ( García-Pérez and Alcalá-Quintana , 2013; García-Pérez and Peli , 2014 ) in which participants guess when the sensory evidence lies within some uncertainty range was not parsimonious: first , to fit the data for the participants with pure sensory biases , the model needs that participants , when uncertain , guess the two alternatives equally often; second , for the participants with decisional biases , we found that the SDT model provided a better fit .","Perceptual discrimination with two symmetric alternatives are often regarded as Type 1 tasks , tasks for which the responses could be designated as correct or incorrect ( Kingdom and Prins , 2016 ) .","If a stimulus has positive signal ( for example , rightward motion ) relative to a neutral point ( no net motion ) , but the decision-maker chooses the alternative consistent with negative signal ( leftward motion ) , the response is considered an error .","In this case , a psychometric function of the proportion of correct responses as a function of the unsigned signal is often fitted , and precision is estimated as the amount of signal that is required to reach a certain level of correct responses .","Our results , however , suggest that in some cases a positive signal might be perceived consistently as a negative signal ( sensory bias ) .","Consequently , it might be inappropriate to consider these responses as errors and , in case feedback is given , provide a negative reward .","Our results suggest , thus , that perceptual discrimination tasks with two symmetric alternatives might be better regarded as Type 2 tasks , tasks for which there are not correct and incorrect responses ( Kingdom and Prins , 2016; Gold and Ding , 2013 ) .","Accordingly , the psychometric function that should be fitted is the proportion of times that a category was selected ( for example , rightward motion ) as a function of the signed signal , and precision should be estimated using the slope of the psychometric function ( Gold and Ding , 2013 ) .","Discrimination tasks with two symmetric alternatives are commonly used to assess how perception is affected by contextual cues ( Carrasco et al . , 2004; Schwartz et al . , 2007 ) , but in some scenarios it is unclear whether the context influences perception or biases decisions ( Carrasco , 2011; García-Pérez and Alcalá-Quintana , 2013; Morgan et al . , 2012; Störmer et al . , 2009; Schneider , 2011; Anton-Erxleben et al . , 2010; Jogan and Stocker , 2012; Mather and Sharman , 2015 ) .","Our results indicate that , even when the symmetry of the task is not broken by the context , half of the participants exhibit decisional biases .","This suggests that to reliably estimate sensory biases in the presence of contextual cues , it might be better to use manipulations of the stimulus-response mapping like the one that we used or tasks for which the potential contribution of decisional biases is reduced or eliminated ( Jogan and Stocker , 2012; Patten and Clifford , 2015; Morgan , 2014 ) .","The perceptual discrimination task with two symmetric alternatives that we have used is also known as the method of single stimuli ( Morgan et al . , 2012 ) In this method , the decision-maker is asked to categorize the signal presented against an internal absolute reference that corresponds to a natural neutral point that the decision-maker possibly has learnt during her life ( Morgan et al . , 2012 ) —verticality or horizontality in our case .","The sensory biases that we have measured are deviations from this internal reference .","In the non-symmetric version of this task , the Yes-No task ( Green and Swets , 1966; Kingdom and Prins , 2016; Wickens , 2001 ) , the decision-maker decides , for example , whether a dim light is present or absent .","For this task , decision-makers have also idiosyncratic tendencies to report that the signal is present or absent ( Green and Swets , 1966 ) .","Given that , in this task , there is no natural neutral point , it is harder to assess whether these tendencies reflect differences in the strength that each decision-maker perceives or differences in how conservatively they report that the signal is present or absent ( Jin and Glickfeld , 2018 ) .","Another popular task to assess perception is the two alternative forced choice task ( Green and Swets , 1966; Wickens , 2001; Kingdom and Prins , 2016 ) .","In this task , two stimuli are presented in a random order—for example , a vertical grating and a grating slightly tilted clockwise—and the decision-maker decides in which interval the grating was more tilted clockwise .","This task assesses the precision to discriminate similar orientations , but cannot measure sensory biases from verticality: a bias of the decision-maker to perceive orientations clockwise will affect the stimuli presented in the two intervals in the same direction without affecting the selection of the interval with the more tilted stimulus .","To facilitate that participants internalize two different associations between perception and the responses , we asked them to compare the orientation of the stimulus to a reference imagined in two different locations .","Given that the reference was not physically present , we think , however , that it is likely that participants were performing judgments of absolute orientation relative to an internal point of horizontality ( or verticality ) .","As we found evidence that the sensory biases were consistent with a global perceptual rotation , if participants were performing judgments of relative orientation between the stimulus and the reference—instead of judgments of absolute orientation—then sensory biases should not be expected , as both the stimulus and the reference should be biased in the same direction .","We found a good agreement between the sensory biases estimated from the symmetric and the asymmetric task .","We also found a good agreement between tasks when the biases were estimated taking into account only one localization of the reference , and thus for which it was not possible to assess the contribution of sensory and decisional biases ( see legend for Figure 3A ) .","This is expected given the large contribution of sensory biases to the total magnitude of the biases .","This agreement across tasks contrasts with the disagreement found between the analogous tasks in the temporal domain .","In the temporal domain , the analogous to the symmetric task is the temporal order judgment .","Given , for example , an auditory and a visual event presented at some asynchrony , the task is to report whether the auditory or the visual event is perceived first .","The location parameter of the psychometric fit for the proportion of ‘auditory perceived first’ responses as a function of the asynchrony provides a measure of the bias .","The asymmetric task is the simultaneity judgment , in which is necessary to report whether the auditory and the visual event were perceived simultaneously .","In this case , the bias is estimated as the asynchrony for which the proportion of simultaneous responses is maximum .","It has been shown that the biases estimated from these two tasks do not agree ( Love et al . , 2013; Linares and Holcombe , 2014; van Eijk et al . , 2008 ) , which suggests that decision making for time perception might be more affected by decisional biases ( Linares and Holcombe , 2014; Shore et al . , 2001; Schneider and Bavelier , 2003 ) ."],["The study was approved by the ethical committee of the University of Barcelona and followed the requirements of the Helsinki convention .","The participants , who did not know the hypothesis of the experiments , provided written consent to perform the experiments .","Twenty-one participants were recruited for Experiment 1 and sixteen for Experiment 2 .","Stimuli were generated using PsychoPy ( Peirce , 2007 ) , displayed on a Sony G520 CRT screen ( 40 cm width and 30 cm height; 60 Hz refresh rate ) and viewed from a distance of 57 cm in a dark room .","They consisted in a Gabor patch ( standard deviation ( sd ) of the Gaussian envelope: 1 . 33 degrees of visual angle ( dva ) ; maximum luminance: 79 cd/m2 ) and a red Gaussian blob ( sd: 0 . 1 dva; maximum luminance: 19 cd/m2 ) on top of it that participants were asked to fixate during the experiment .","Stimuli were presented against a uniform circular grey background ( diameter: 25 dva; luminance: 43 cd/m2 ) that was displayed in a black background ( luminance: 0 . 2 cd/m2 ) .","The verticality of the Gabor was calibrated using a pendulum .","Before the experiments , to facilitate the understanding of the instructions , a reference ( a green gaussian luminance profile; sd: 0 . 1 dva; maximum luminance: 29 cd/m2 ) was displayed for 5 to 10 trials at the corresponding cardinal location at a distance of 6 dva from the center of the fixation point .","The data and the code to do the statistical analysis and create the figures is available at https://github . com/danilinares/2018LinaresAguilarLopezmoliner ( Linares , 2019; copy archived at https://github . com/elifesciences-publications/2018LinaresAguilarLopezmoliner ) .","The model fitting , goodness of fit and model selection ( likelihood ratio test and Akaike information criterion ) was performed using the R package quickspy ( Linares and López-Moliner , 2016 ) , which under the development version allows fitting several psychometric functions conjointly ."]],"string":"[\n [\n \"You ask a friend about the tilt of a canvas that is perfectly horizontal and she says that the top right corner is up .\",\n \"What causes her inaccuracy ?\",\n \"One possibility is that her sensory representation is biased and she perceives the canvas tilted .\",\n \"Another possibility , however , is that under uncertainty about the orientation of the canvas , she prefers to choose up over down .\",\n \"This situation exemplifies a major problem in the study of perception: perceptual decisions not only depend on the sensory evidence , but also on decisional components ( Green and Swets , 1966; Gold and Ding , 2013 ) .\",\n \"When a decision maker needs to select an action from a set of alternatives , the tendency to choose one alternative over the others is known as choice bias ( Gold and Ding , 2013 ) .\",\n \"Choice biases occur in perceptual tasks even in simple scenarios like the one described above , in which the stimuli carry similar levels of signal relative to a neutral point ( Newsome and Paré , 1988; Mareschal and Clifford , 2012; Jazayeri and Movshon , 2007; Milner et al . , 1992; Tadin et al . , 2003 ) .\",\n \"Choice biases are extensively studied in relation to how current decisions are influenced by previous decisions , what is known as choice history biases ( Abrahamyan et al . , 2016; Akaishi et al . , 2014; Fründ et al . , 2014; Urai et al . , 2017; Fischer and Whitney , 2014; Braun et al . , 2018; St John-Saaltink et al . , 2016; Fritsche et al . , 2017; Hermoso-Mendizabal et al . , 2019 ) .\",\n \"Depending on the perceptual scenario , people tend to repeat their previous choice ( Akaishi et al . , 2014; Braun et al . , 2018 ) , alternate between choices ( Fritsche et al . , 2017 ) or idiosyncratically repeat or alternate ( Urai et al . , 2017; Abrahamyan et al . , 2016 ) .\",\n \"Whether choice history biases reflect a bias in the sensory evidence or in the decision process is under debate ( Akaishi et al . , 2014; Fischer and Whitney , 2014; St John-Saaltink et al . , 2016; Fritsche et al . , 2017 ) .\",\n \"Unlike choice history biases , some choice biases are related to the overall idiosyncratic tendency to choose one alternative over the others , not conditioned by the previous choices .\",\n \"We will refer to these as global choice biases .\",\n \"The existence of these biases is acknowledged ( Gold and Ding , 2013; Kingdom and Prins , 2016; Morgan et al . , 2012; García-Pérez and Alcalá-Quintana , 2013; Peters et al . , 2016 ) and they are included in current models of perceptual decision making ( Abrahamyan et al . , 2016; Akaishi et al . , 2014; Urai et al . , 2017; Braun et al . , 2018; Hermoso-Mendizabal et al . , 2019 ) , but whether they reflect sensory or decisional processes has not been , to our knowledge , assessed ( we searched in Google Scholar several times , last one on February 2019 , the following keywords: choice biases , response biases , motor biases , perceptual biases and sensory biases; we identified the relevant articles and searched within the references cited; we also tracked the articles that cited the relevant articles ) .\",\n \"The problem to identify their origin is that , in standard perceptual paradigms , the tendency to choose one alternative more often is consistent with a biased sensory representation , but also with a bias in the decision process ( García-Pérez and Alcalá-Quintana , 2013; Gold and Ding , 2013 ) .\",\n \"Here , to disentangle the contribution of sensory and decisional processes to global choice bias , we measured choice behavior in a simple common perceptual discrimination task with two symmetric alternatives , but under two different task instructions .\",\n \"The instructions varied the response mapping between perception and the category of the alternatives .\"\n ],\n [\n \"On a given trial of the symmetric task ( the model for the asymmetric task is described in Supplementary Models ) a grating with orientation θi that could be clockwise or counterclockwise , is presented and the participant decides whether its orientation is consistent with clockwise acw or counterclockwise accw orientation .\",\n \"A simple standard signal detection theory ( SDT ) model assumes that: ( 1 ) the sensory evidence s associated to the stimulus is a random variable normally distributed with variance σ2 centered on μcw when θi is clockwise and centered on μccw when θi is counterclockwise; ( 2 ) to make choices , the participant sets a criterion β , computes the log likelihood ratio logΛ of s under the hypothesis that μ=μcw and the hypothesis that μ=μccw ( 1 ) logΛ ( s ) =log ( p ( s|μ=μcw ) p ( s|μ=μccw ) ) =log ( ( 2πσ2 ) −1/2e− ( s−μcw ) 22σ2 ( 2πσ2 ) −1/2e− ( s−μccw ) 22σ2 ) and chooses the action acw when logΛ ( s ) >β .\",\n \"In standard SDT models the origin and scale of the sensory evidence is arbitrary ( Wickens , 2001; Green and Swets , 1966 ) .\",\n \"A typical choice is ( Wickens , 2001; Green and Swets , 1966 ) ( 2 ) μcw=d′2μccw=−d′2σ=1which results in logΛ ( s ) =d′s .\",\n \"That is , one can use directly the evidence as a decision variable and choose acw when s>c=β/d′ ( Wickens , 2001; Green and Swets , 1966 ) .\",\n \"The probability of choosing clockwise when θi is clockwise is ( 3 ) P ( a=acw;μ=μcw ) =12π∫c∞e− ( s−μcw ) 22ds=Φ ( μcw−c ) where Φ is the standard cumulative normal function .\",\n \"To relate this probability to the magnitude of the stimulus , this standard SDT could be extended to include how the stimulus is transduced into sensory evidence ( García-Pérez and Alcalá-Quintana , 2013; Schneider and Bavelier , 2003; García-Pérez and Peli , 2014; Schneider and Komlos , 2008 ) .\",\n \"Adding the transduction provides a meaningful origin and scale to the sensory evidence .\",\n \"Assuming that the transduction is linear ( 4 ) μcw=aθi+bthe probability of choosing clockwise as a function of θi becomes ( 5 ) P ( a=acw;θi ) =Φ ( aθi+b−c ) =Φ ( θi− ( δD−δS ) σ ) Where ( 6 ) δS=ba=sensorybiasδD=ca=decisionalbiasσ=1awhich corresponds to a psychometric function centered on δD−δS with slope given by σ .\",\n \"Consider the trials in which the grating is presented in orientations around the horizontal orientation and the reference is on the right ( the reasoning that follows also holds for vertical orientations ) .\",\n \"If the orientation of the grating can take M different values , each value is presented n times and ki is the number of times that the participant pressed the down key ( clockwise ) , then the probability density function that defines statistical model for these trials is ( 7 ) f ( k1 , … , kM;δS , δD , σ ) =∏i=1M ( nki ) Φ ( θi− ( δD−δS ) σ ) ki ( 1−Φ ( θi− ( δD−δS ) σ ) ) n−ki Given that f depends on δD−δS , it is not possible to distinguish sensory from decisional biases ( Witt et al . , 2015; Schneider and Komlos , 2008; García-Pérez and Alcalá-Quintana , 2013 ) .\",\n \"Consider now the trials for the left reference .\",\n \"A criterion c associated to a bias to choose responses consistent with clockwise orientation for the right reference corresponds to a criterion −c for the left reference .\",\n \"The statistical model that incorporates responses for the two locations of the reference is ( 8 ) f ( k1 , … , k2M;δS , δD , σ ) =∏i=1M ( nki ) ( 1−ΦR ) n−ki∏i=M+12M ( nki ) ΦLki ( 1−ΦL ) n−kiwhere ( 9 ) ΦR=Φ ( θi− ( δD−δS ) σ ) ΦL=Φ ( θi− ( −δD−δS ) σ ) and the index i larger than M refers to the trials for the left reference .\",\n \"Now , by fitting two psychometric functions conjointly , δD and δs are not degenerated and could be estimated .\",\n \"This is the exact model that we fitted for 17 of the 34 groups ( 17 participants that discriminate orientations around the horizontal and the vertical orientation in Experiment 1 ) .\",\n \"For the rest of the groups , we also fitted this basic model , but adding some extra features that significantly improved the quality of the fit ( using likelihood ratio tests , see Materials and methods ) .\",\n \"The first feature is related to the variability of the sensory evidence σ .\",\n \"The basic model considers that this variability does not depend on where the participants imagined the reference , which results in the two psychometric functions having the same slope .\",\n \"We found this to be the case in 28 of the 34 fits , but in six groups we found that considering different variabilities ( σ1 and σ2 ) for the two locations of the reference was better .\",\n \"It might be that orienting attention to different parts of the visual field might change the uncertainty in the sensory evidence for some participants .\",\n \"The other feature that we added consisted in adding lapses , which are responses that are incorrect independently of the sensory evidence ( Kingdom and Prins , 2016; Gold and Ding , 2013 ) .\",\n \"This might occur , for example , when the participant misses the stimulus because of blinking or a loss of attention .\",\n \"To incorporate lapses into the basic model , ΦR need to be replaced by λ1+ ( 1−λ1−λ2 ) ΦR and ΦL by λ1′+ ( 1−λ1′−λ2′ ) ΦL ( Kingdom and Prins , 2016 ) .\",\n \"From the 34 groups , adding lapses improved the fit in 14 cases .\",\n \"From those , three fits required the four lapse parameters , 4 fits required two lapse parameters ( one for each psychometric function , λ1=λ2 ; λ1′=λ2′ ) and 7 fits required one lapse rate ( λ1=λ2=λ1′=λ2′ ) .\",\n \"Importantly , including a lapse rate parameter λ∗ placed on the right asymptote for the right reference ( 10 ) λ1+λ∗+ ( 1−λ1−λ1−λ∗ ) ΦRand placed on the left asymptote for the left reference ( 11 ) λ2+ ( 1−λ2−λ2−λ∗ ) ΦLdid not improve significantly the fit for any group , which indicates that decisional biases cannot be explained by a tendency of participants to select one alternative completely independently of the sensory evidence .\",\n \"Perceptual tasks with two symmetric alternatives have been also modeled using a high threshold model , called the indecision model ( García-Pérez and Alcalá-Quintana , 2013; García-Pérez and Peli , 2014 ) .\",\n \"This model is similar to the SDT model described , but divides the sensory axis in three regions delimited by c1 and c2 .\",\n \"When the sensory evidence is lower than c1 the participant chooses accw and when is larger than c2 , chooses acw .\",\n \"Critically , when the sensory evidence lies between c1 and c2—called interval of uncertainty—the model assumes that the observer is undecided and guesses the response ( choosing accw with probability ξ ) .\",\n \"The probability of choosing clockwise when θi is clockwise and the reference is on the right is ( 12 ) P ( a=acw;μ=μcw ) =12π∫c2∞e− ( s−μcw ) 22ds+ξ12π∫c1c2e− ( s−μcw ) 22ds=Φ ( μcw−c2 ) +ξ ( Φ ( c2−μcw ) −Φ ( c1−μcw ) ) For the left reference , assuming that c1 and c2 do not change , ξ should be replaced by 1−ξ .\",\n \"We fitted this model to the 19 groups with significant decisional biases and found that for all of them ( except for participant eight for the vertical condition in Experiment\",\n \"1 ) the SDT model was better using the Akaike information criterion ( Burnham and Anderson , 2004 ) .\",\n \"Given that the indecision model has two parameters more than the SDT model ( c is replaced by c1 and c2 and ξ is introduced ) , we also compared the SDT model to a simplified version of the indecision model with symmetric boundaries c1=−c2 ( García-Pérez and Alcalá-Quintana , 2013; García-Pérez and Peli , 2014 ) .\",\n \"In this case , the SDT model was better for all the groups .\"\n ],\n [\n \"We showed that , in a simple discrimination task with two symmetric alternatives , most people exhibit idiosyncratic global choice biases .\",\n \"Changing the stimulus-response mapping and testing a task with two asymmetric alternatives , we found that these biases reflect biases in the sensory evidence and in the decisional process .\",\n \"Specifically , about half of the participants showed biases consistent with only a sensory origin and about half of the participant biases with a contribution of sensory and decisional biases .\",\n \"We also found that a very few participants exhibited biases consistent with only a decisional origin .\",\n \"Across participants , the magnitude of the sensory bias was about three times as larger as the magnitude of the decisional bias .\",\n \"Our findings suggest that if a person has a bias to report that a perfectly horizontal canvas is tilted towards a given direction , it is likely that the largest contribution to the bias has a sensory origin .\",\n \"To our knowledge , these idiosyncratic sensory biases have not been linked to asymmetries in the neural representation of the stimulus .\",\n \"Possible asymmetries include a difference in the number of neurons tuned to clockwise and counterclockwise orientation if the decoding is of the population-vector type or a difference in the gain of neurons tuned to clockwise and counterclockwise orientation for more general decoding schemes ( Dayan and Abbott , 2001; Schwartz et al . , 2007 ) .\",\n \"More recently , it has been proposed that the asymmetries could also emerge from the dynamics of competing neural networks ( Lebovich et al . , 2018 ) .\",\n \"Importantly , for these asymmetries to bias perception , the downstream decoding mechanisms should have not been calibrated with the environmental property to compensate the asymmetries ( decoding ambiguity; Schwartz et al . , 2007 ) .\",\n \"An SDT model , in which the sensory biases were included as an intercept term in a linear transduction of the stimuli ( García-Pérez and Alcalá-Quintana , 2013; Schneider and Bavelier , 2003; García-Pérez and Peli , 2014; Schneider and Komlos , 2008 ) and decisional biases correspond to a shift in the criterion , fitted well the choice behavior of the participants .\",\n \"A model in which participants have a tendency to choose one alternative independently of the magnitude of the sensory evidence did not fit well the data; it rather predicts an asymmetry on the lapse rate across locations of the reference that was not observed .\",\n \"Finally , a model ( García-Pérez and Alcalá-Quintana , 2013; García-Pérez and Peli , 2014 ) in which participants guess when the sensory evidence lies within some uncertainty range was not parsimonious: first , to fit the data for the participants with pure sensory biases , the model needs that participants , when uncertain , guess the two alternatives equally often; second , for the participants with decisional biases , we found that the SDT model provided a better fit .\",\n \"Perceptual discrimination with two symmetric alternatives are often regarded as Type 1 tasks , tasks for which the responses could be designated as correct or incorrect ( Kingdom and Prins , 2016 ) .\",\n \"If a stimulus has positive signal ( for example , rightward motion ) relative to a neutral point ( no net motion ) , but the decision-maker chooses the alternative consistent with negative signal ( leftward motion ) , the response is considered an error .\",\n \"In this case , a psychometric function of the proportion of correct responses as a function of the unsigned signal is often fitted , and precision is estimated as the amount of signal that is required to reach a certain level of correct responses .\",\n \"Our results , however , suggest that in some cases a positive signal might be perceived consistently as a negative signal ( sensory bias ) .\",\n \"Consequently , it might be inappropriate to consider these responses as errors and , in case feedback is given , provide a negative reward .\",\n \"Our results suggest , thus , that perceptual discrimination tasks with two symmetric alternatives might be better regarded as Type 2 tasks , tasks for which there are not correct and incorrect responses ( Kingdom and Prins , 2016; Gold and Ding , 2013 ) .\",\n \"Accordingly , the psychometric function that should be fitted is the proportion of times that a category was selected ( for example , rightward motion ) as a function of the signed signal , and precision should be estimated using the slope of the psychometric function ( Gold and Ding , 2013 ) .\",\n \"Discrimination tasks with two symmetric alternatives are commonly used to assess how perception is affected by contextual cues ( Carrasco et al . , 2004; Schwartz et al . , 2007 ) , but in some scenarios it is unclear whether the context influences perception or biases decisions ( Carrasco , 2011; García-Pérez and Alcalá-Quintana , 2013; Morgan et al . , 2012; Störmer et al . , 2009; Schneider , 2011; Anton-Erxleben et al . , 2010; Jogan and Stocker , 2012; Mather and Sharman , 2015 ) .\",\n \"Our results indicate that , even when the symmetry of the task is not broken by the context , half of the participants exhibit decisional biases .\",\n \"This suggests that to reliably estimate sensory biases in the presence of contextual cues , it might be better to use manipulations of the stimulus-response mapping like the one that we used or tasks for which the potential contribution of decisional biases is reduced or eliminated ( Jogan and Stocker , 2012; Patten and Clifford , 2015; Morgan , 2014 ) .\",\n \"The perceptual discrimination task with two symmetric alternatives that we have used is also known as the method of single stimuli ( Morgan et al . , 2012 ) In this method , the decision-maker is asked to categorize the signal presented against an internal absolute reference that corresponds to a natural neutral point that the decision-maker possibly has learnt during her life ( Morgan et al . , 2012 ) —verticality or horizontality in our case .\",\n \"The sensory biases that we have measured are deviations from this internal reference .\",\n \"In the non-symmetric version of this task , the Yes-No task ( Green and Swets , 1966; Kingdom and Prins , 2016; Wickens , 2001 ) , the decision-maker decides , for example , whether a dim light is present or absent .\",\n \"For this task , decision-makers have also idiosyncratic tendencies to report that the signal is present or absent ( Green and Swets , 1966 ) .\",\n \"Given that , in this task , there is no natural neutral point , it is harder to assess whether these tendencies reflect differences in the strength that each decision-maker perceives or differences in how conservatively they report that the signal is present or absent ( Jin and Glickfeld , 2018 ) .\",\n \"Another popular task to assess perception is the two alternative forced choice task ( Green and Swets , 1966; Wickens , 2001; Kingdom and Prins , 2016 ) .\",\n \"In this task , two stimuli are presented in a random order—for example , a vertical grating and a grating slightly tilted clockwise—and the decision-maker decides in which interval the grating was more tilted clockwise .\",\n \"This task assesses the precision to discriminate similar orientations , but cannot measure sensory biases from verticality: a bias of the decision-maker to perceive orientations clockwise will affect the stimuli presented in the two intervals in the same direction without affecting the selection of the interval with the more tilted stimulus .\",\n \"To facilitate that participants internalize two different associations between perception and the responses , we asked them to compare the orientation of the stimulus to a reference imagined in two different locations .\",\n \"Given that the reference was not physically present , we think , however , that it is likely that participants were performing judgments of absolute orientation relative to an internal point of horizontality ( or verticality ) .\",\n \"As we found evidence that the sensory biases were consistent with a global perceptual rotation , if participants were performing judgments of relative orientation between the stimulus and the reference—instead of judgments of absolute orientation—then sensory biases should not be expected , as both the stimulus and the reference should be biased in the same direction .\",\n \"We found a good agreement between the sensory biases estimated from the symmetric and the asymmetric task .\",\n \"We also found a good agreement between tasks when the biases were estimated taking into account only one localization of the reference , and thus for which it was not possible to assess the contribution of sensory and decisional biases ( see legend for Figure 3A ) .\",\n \"This is expected given the large contribution of sensory biases to the total magnitude of the biases .\",\n \"This agreement across tasks contrasts with the disagreement found between the analogous tasks in the temporal domain .\",\n \"In the temporal domain , the analogous to the symmetric task is the temporal order judgment .\",\n \"Given , for example , an auditory and a visual event presented at some asynchrony , the task is to report whether the auditory or the visual event is perceived first .\",\n \"The location parameter of the psychometric fit for the proportion of ‘auditory perceived first’ responses as a function of the asynchrony provides a measure of the bias .\",\n \"The asymmetric task is the simultaneity judgment , in which is necessary to report whether the auditory and the visual event were perceived simultaneously .\",\n \"In this case , the bias is estimated as the asynchrony for which the proportion of simultaneous responses is maximum .\",\n \"It has been shown that the biases estimated from these two tasks do not agree ( Love et al . , 2013; Linares and Holcombe , 2014; van Eijk et al . , 2008 ) , which suggests that decision making for time perception might be more affected by decisional biases ( Linares and Holcombe , 2014; Shore et al . , 2001; Schneider and Bavelier , 2003 ) .\"\n ],\n [\n \"The study was approved by the ethical committee of the University of Barcelona and followed the requirements of the Helsinki convention .\",\n \"The participants , who did not know the hypothesis of the experiments , provided written consent to perform the experiments .\",\n \"Twenty-one participants were recruited for Experiment 1 and sixteen for Experiment 2 .\",\n \"Stimuli were generated using PsychoPy ( Peirce , 2007 ) , displayed on a Sony G520 CRT screen ( 40 cm width and 30 cm height; 60 Hz refresh rate ) and viewed from a distance of 57 cm in a dark room .\",\n \"They consisted in a Gabor patch ( standard deviation ( sd ) of the Gaussian envelope: 1 . 33 degrees of visual angle ( dva ) ; maximum luminance: 79 cd/m2 ) and a red Gaussian blob ( sd: 0 . 1 dva; maximum luminance: 19 cd/m2 ) on top of it that participants were asked to fixate during the experiment .\",\n \"Stimuli were presented against a uniform circular grey background ( diameter: 25 dva; luminance: 43 cd/m2 ) that was displayed in a black background ( luminance: 0 . 2 cd/m2 ) .\",\n \"The verticality of the Gabor was calibrated using a pendulum .\",\n \"Before the experiments , to facilitate the understanding of the instructions , a reference ( a green gaussian luminance profile; sd: 0 . 1 dva; maximum luminance: 29 cd/m2 ) was displayed for 5 to 10 trials at the corresponding cardinal location at a distance of 6 dva from the center of the fixation point .\",\n \"The data and the code to do the statistical analysis and create the figures is available at https://github . com/danilinares/2018LinaresAguilarLopezmoliner ( Linares , 2019; copy archived at https://github . com/elifesciences-publications/2018LinaresAguilarLopezmoliner ) .\",\n \"The model fitting , goodness of fit and model selection ( likelihood ratio test and Akaike information criterion ) was performed using the R package quickspy ( Linares and López-Moliner , 2016 ) , which under the development version allows fitting several psychometric functions conjointly .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["The contribution of sensory and decisional processes to perceptual decision making is still unclear , even in simple perceptual tasks .","When decision makers need to select an action from a set of balanced alternatives , any tendency to choose one alternative more often—choice bias—is consistent with a bias in the sensory evidence , but also with a preference to select that alternative independently of the sensory evidence .","To decouple sensory from decisional biases , here we asked humans to perform a simple perceptual discrimination task with two symmetric alternatives under two different task instructions .","The instructions varied the response mapping between perception and the category of the alternatives .","We found that from 32 participants , 30 exhibited sensory biases and 15 decisional biases .","The decisional biases were consistent with a criterion change in a simple signal detection theory model .","Perceptual decision making , thus , even in simple scenarios , is affected by sensory and decisional choice biases ."],"string":"[\n \"The contribution of sensory and decisional processes to perceptual decision making is still unclear , even in simple perceptual tasks .\",\n \"When decision makers need to select an action from a set of balanced alternatives , any tendency to choose one alternative more often—choice bias—is consistent with a bias in the sensory evidence , but also with a preference to select that alternative independently of the sensory evidence .\",\n \"To decouple sensory from decisional biases , here we asked humans to perform a simple perceptual discrimination task with two symmetric alternatives under two different task instructions .\",\n \"The instructions varied the response mapping between perception and the category of the alternatives .\",\n \"We found that from 32 participants , 30 exhibited sensory biases and 15 decisional biases .\",\n \"The decisional biases were consistent with a criterion change in a simple signal detection theory model .\",\n \"Perceptual decision making , thus , even in simple scenarios , is affected by sensory and decisional choice biases .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Imagine that every day , you split a chocolate bar into two and offer one half to your friend .","Even though you take care to divide the bar into equal pieces , your friend nearly always chooses the left half .","Why is that ?","One possibility is that sensory bias in her visual system makes her perceive the left half of the bar to be larger than the right .","But it is also possible that she does not see any difference between the two halves .","Instead she simply decides to pick the left half because she prefers doing so .","The above example illustrates a key problem in studying perception .","When asked to make a decision where there is no obviously correct answer such as deciding whether a painting is hanging perfectly straight people typically respond one way more often than the other .","But does this response bias reflect biased perception or biased decision making ?","Linares et al . have designed an experiment to tease apart these alternatives .","Healthy volunteers had to decide whether gratings were tilted slightly upward or slightly downward .","Almost all volunteers showed biases in their choice behavior in one of the two directions .","To decouple sensory biases from ‘decisional’ biases , the volunteers had to press a particular key to select ‘upward’ on some trials , but ‘downward’ on others .","This would not affect responding if the volunteers showed a decisional bias to press a key .","But it would affect responding if the volunteers showed a sensory bias .","The results revealed that both sensory and decisional biases influenced the volunteers’ choice behavior .","However , sensory biases were more common .","People diagnosed with psychiatric disorders like schizophrenia often respond differently on perceptual tasks compared to healthy volunteers .","Future studies should investigate whether this difference results from altered perception or altered decision making .","This information could help narrow down the neural circuits affected by these disorders ."],"string":"[\n \"Imagine that every day , you split a chocolate bar into two and offer one half to your friend .\",\n \"Even though you take care to divide the bar into equal pieces , your friend nearly always chooses the left half .\",\n \"Why is that ?\",\n \"One possibility is that sensory bias in her visual system makes her perceive the left half of the bar to be larger than the right .\",\n \"But it is also possible that she does not see any difference between the two halves .\",\n \"Instead she simply decides to pick the left half because she prefers doing so .\",\n \"The above example illustrates a key problem in studying perception .\",\n \"When asked to make a decision where there is no obviously correct answer such as deciding whether a painting is hanging perfectly straight people typically respond one way more often than the other .\",\n \"But does this response bias reflect biased perception or biased decision making ?\",\n \"Linares et al . have designed an experiment to tease apart these alternatives .\",\n \"Healthy volunteers had to decide whether gratings were tilted slightly upward or slightly downward .\",\n \"Almost all volunteers showed biases in their choice behavior in one of the two directions .\",\n \"To decouple sensory biases from ‘decisional’ biases , the volunteers had to press a particular key to select ‘upward’ on some trials , but ‘downward’ on others .\",\n \"This would not affect responding if the volunteers showed a decisional bias to press a key .\",\n \"But it would affect responding if the volunteers showed a sensory bias .\",\n \"The results revealed that both sensory and decisional biases influenced the volunteers’ choice behavior .\",\n \"However , sensory biases were more common .\",\n \"People diagnosed with psychiatric disorders like schizophrenia often respond differently on perceptual tasks compared to healthy volunteers .\",\n \"Future studies should investigate whether this difference results from altered perception or altered decision making .\",\n \"This information could help narrow down the neural circuits affected by these disorders .\"\n]","isConversational":false},"year":{"kind":"string","value":"2019"}}},{"rowIdx":86,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["cell biology"],"string":"[\n \"cell biology\"\n]","isConversational":false},"title":{"kind":"string","value":"Cardiac mitochondrial function depends on BUD23 mediated ribosome programming"},"id":{"kind":"string","value":"elife-50705-v2"},"sections":{"kind":"list like","value":[["Regulation of protein abundance within a cell is of fundamental importance to homeostasis , cell identity and responsiveness to changes in external environment .","Many layers of regulatory control have evolved to fine-tune gene expression to meet cellular needs , including transcriptional regulation , cis-acting genomic elements , epigenetic structures , and translational efficiency .","Selective control of mRNA translation at the level of the ribosome is emerging as important for protein dynamics within the cell; however , the regulatory mechanisms remain unclear ( Sonenberg and Hinnebusch , 2009 ) .","Protein synthesis is the most energetically-demanding process for a cell to perform and , therefore , mRNA translation is closely coupled to mitochondrial function ( Morita et al . , 2013 ) .","Mitochondria are fundamental cellular components of eukaryotes , generating the majority of cellular ATP .","The human mitochondrial genome contains only 37 genes , of which 13 are subunits of respiratory complexes , 22 encode mitochondrial tRNAs , and a further two encode rRNA .","Whilst the mitochondrion translates the 13 protein-coding genes of its own genome using a bespoke mitochondrial apparatus , the majority of the genes required for a functional mitochondrion are encoded in the cell nuclear genome and are , therefore , dependent on the cytosolic translational apparatus .","This relationship implies a fundamental role for the eukaryotic 80S ribosome in control of mitochondrial abundance and function .","Indeed , mitochondrial content and function varies between tissues and cell-types , as well as within a tissue , in response to external and internal stimuli including energy demand , oxidative stress and cellular signals ( Palmer et al . , 2011 ) .","For example , red blood cells contain no mitochondria whereas in cardiomyocytes mitochondria occupy 30% of cell volume to meet the exceptionally high ATP demand ( Piquereau et al . , 2013 ) .","Mitochondrial dysfunction has been linked to a wide range of cardiac disorders , often due to the aberrant production of ROS , and attendant cell death ( Kanaan and Harper , 2017; Ott et al . , 2007 ) .","BUD23 ( previously known as WBSCR22 ) was initially identified as a putative methyltransferase implicated in tumour metastases ( Nakazawa et al . , 2011 ) .","Its expression is responsive to diverse inflammatory and cancer pathologies ( Jangani et al . , 2014 ) , and reduction of BUD23 expression can affect cellular response to glucocorticoid and alter histone methylation ( Jangani et al . , 2014 ) .","However , BUD23 actions were diverse , and a unifying mechanism of action was elusive .","More recent studies identified ribosomal RNA as the preferred substrate of BUD23 ( Haag et al . , 2015; White et al . , 2008; Zorbas et al . , 2015 ) .","There is little understanding of the physiological role of BUD23 in a mammalian context although it is one of the genes deleted in Williams-Beuren syndrome ( WBS ) .","WBS patients have a complex phenotype with prominent neurological and morphological features .","In addition metabolic pathologies exist including diabetes and obesity ( Morris , 1993 ) .","The contribution of individual genes within the 22-gene interval to the human phenotype has not been determined , but the neurological and metabolic features suggest a bioenergetic contribution .","BUD23 is a ribosomal RNA methyltransferase which imparts a methyl mark on a key guanosine residue ( forming N7-methylguanosine ) located between the E and P site of the small ribosomal subunit that has been mapped to residue G1575 of yeast 18S rRNA and G1639 of human 18S rRNA ( Haag et al . , 2015; Õunap et al . , 2013; White et al . , 2008; Zorbas et al . , 2015 ) .","BUD23 protein is found in both the nucleus and the cytoplasm ( Õunap et al . , 2015 ) , and its depletion leads to a nuclear accumulation of 18SE-pre-RNA ( Haag et al . , 2015 ) .","In both yeast and human cells the methyltransferase catalytic activity of BUD23 is not required for the processing of 18S pre-RNA , or the synthesis of 40S subunits , indicating that BUD23 has an additional role , distinct from its methyltransferase activity ( Haag et al . , 2015; Zorbas et al . , 2015 ) .","In human cell lines it has been suggested that some 18S rRNA lacks the m7-G1639 mark , which may indicate a selective role in ribosome function ( Haag et al . , 2015 ) .","BUD23 stability requires interaction with TRMT112 ( TRM112 in yeast ) , an obligate binding partner , which is known to stabilise four client methyltransferase enzymes , all involved in the generation of the translational apparatus ( Bourgeois et al . , 2017; Létoquart et al . , 2014 ) .","The binding of TRMT112 with these four methyltransferases is competitive , which results in tight regulation of BUD23 protein concentration and enzymatic function at the level of protein stability .","Here we identify a novel mechanism linking the energy-demanding process of protein translation to mitochondrial dynamics through BUD23 .","We examine the role of BUD23 in ribosome function and discover that BUD23 preferentially promotes the selection of mRNA species with low GC-content 5’UTRs .","We also identify a role for BUD23 in mitochondrial transcript translation which impacts mitochondrial function in vitro .","We go on to examine the role of BUD23 in a murine in vivo system and discover that the production of mitochondrial proteins is dependent on BUD23 .","Finally , we examine the role of BUD23 in the mitochondrially-rich and energetically-demanding cardiac tissue and discover a cardiomyopathy phenotype leading to premature death .","These discoveries identify BUD23 as essential for mammalian mitochondrial function , with implications for human mitochondrial disease and cardiomyopathy ."],["We previously identified pleiotropic actions of BUD23 in airway epithelial cells , and attributed these to an epigenetic effect ( Jangani et al . , 2014 ) .","The discovery that BUD23 modifies ribosomal RNA rather than histone proteins prompted re-examination of the BUD23 role in physiology .","To identify candidate pathways affected by BUD23 we depleted expression in human airway epithelial cells ( Figure 1A , B ) .","This intervention caused a small but statistically significant decrease in cell proliferation ( Figure 1C ) , which is similar to the anti-proliferative effects caused by BUD23 deletion in yeast ( White et al . , 2008 ) .","To investigate comprehensively the role of BUD23 in ribosome function we performed polysome profiling .","The 48 hr knockdown of BUD23 resulted in a reduction of the 40S subunit peak , and a concomitant increase in the 60/80S peak ( Figure 1D , E ) .","There was , however , little change in the profile of the polysome fractions at this early time-point following transient BUD23 knockdown , allowing us to investigate the changes in mRNA substrate selection in the intact polysomes .","To test the global efficiency of the translational apparatus in cells lacking BUD23 , we used 35-S-methionine incorporation for 1 hr ( Figure 1—figure supplement 1 ) .","This demonstrated little overall impact on global protein translation rate , confirming that the identified polysomes were essentially competent at this time-point , despite the impact on 40S maturation .","In order to determine if BUD23-loss differentially impacted the translation of a specific subset of mRNAs , we profiled the translational efficiency ( TE ) of individual mRNA species within the polysome profiles .","To do this , the polysome fractions were divided into ‘heavy’ ( more than three ribosomes ) or ‘light’ ( less than three ribosomes ) and pooled , prior to RNA extraction and sequencing .","This defined a total of 14 , 527 transcripts , from which the relative proportion of each individual transcript in the heavy compared to the light fraction was calculated to give a translation efficiency ( TE ) score .","Average TE scores across the three replicates for each condition ( + /- BUD23 ) were then plotted against each other ( Figure 1F ) .","This revealed that the most efficiently translated transcripts in the control samples ( more heavily loaded with ribosomes ) , were most significantly reduced in the BUD23 knockdown condition .","To examine the identity of the most highly affected transcripts , a threshold of >1 or <-1 difference in the relative TE was set ( i . e . TE doubled or halved ) , and applied to Figure 1F , shown as trend lines .","There were 650 transcripts with a TE ratio <-1 and 95 transcripts where the TE ratio >1 after BUD23 depletion; indicating the marked loss of high efficiency mRNA translation .","The 650 transcripts for which TE was reduced ( TE ratio <-1 after BUD23 depletion ) constituted a highly connected network , with a significantly higher number of protein-protein interactions than would be expected from a random dataset of similar size , and an interaction analysis ( PPI ) enrichment p-value<1e−16 .","Gene set analysis of these transcripts revealed enrichment for genes involved in RNA processing , metabolic processes and organelle organisation among other high energy demand processes ( Figure 1—figure supplement 1 ) .","Surprisingly , the TE of mitochondrially-encoded transcripts was observed to be particularly strongly down-regulated ( Figure 1G ) .","In fact , all mitochondrial transcripts had a negative change in TE after BUD23 loss , with approximately 50% reduction in TE .","To investigate the mechanism underlying the BUD23 dependent selection of transcripts we examined specific mRNA features associated with translational efficiency .","Given the known influence of 5’UTR features on translation efficiency , we reasoned that specific features of the 5’UTR may be particularly important for BUD23 action .","A number of well-characterised 5’UTR features impact mRNA translation , including length , secondary structure , GC content , and specific motifs including the TOP sequence ( Gandin et al . , 2016; Meijer and Thomas , 2002 ) .","We saw no enrichment for TOP sequence content in the differentially translated mRNA species ( Figure 1—figure supplement 2 ) , and no correlation between change in TE score and 5’UTR length was observed ( Figure 1H ) .","We also noted that transcripts with shorter total lengths were not more likely to shift from the heavy to the light fraction ( Figure 1—figure supplement 3 ) .","There was , however , a highly significant correlation between change in TE score and GC content ( Pearson’s correlation value of −0 . 1337 , p-value<2 . 2e−16 ) ( Figure 1I ) .","This identifies an intersection between mRNA transcript 5’UTR GC content and BUD23 ribosome maturation .","There was no difference in the distribution of 5’UTR GC in mitochondrial-related transcripts relative to all transcripts detected in the polysome analysis ( Figure 1—figure supplement 3 ) .","Our studies reveal the consequences for the ribosome of BUD23 action and point to an impact on the cellular proteome .","Therefore , to investigate the downstream consequences of BUD23 loss we examined the cellular proteome by LC-MS/MS .","BUD23 knockdown downregulated 83 proteins and upregulated 64 , out of a total of 3255 identified ( Figure 2—source data 1 ) .","When tested using PANTHER GO cellular component ontology analysis , upregulated proteins were over-represented for components of the large ribosomal subunit ( 14/64 proteins ) , whilst downregulated proteins were over-represented for components of the small ribosomal subunit ( 22/83 proteins ) , ( Figure 2A , Figure 2—figure supplement 1 ) .","This imbalance in the relative abundance of ribosomal proteins , recapitulates the observation in the polysome profiles and confirms a major role for BUD23 in the maturation of the protein translation apparatus of the cell .","Interestingly , TRMT112 protein , an obligate partner for BUD23 , was also significantly down-regulated with BUD23 knock-down , which may indicate a reciprocal stabilising interaction .","Ontology analysis of the proteomics dataset was used to predict the functional consequences of BUD23 reduction ( PANTHER GO biological processes ) .","Over-representation analysis of the up-regulated proteins overlapped significantly with the ‘protein translation’ .","Analysis of the down-regulated proteins again showed a statistically significant over-representation for proteins associated with GO biological terms including ‘translation’ , as well as ‘rRNA metabolic process’ and ‘protein metabolic process’ which support a major effect on the translational apparatus ( Figure 2B ) .","Unexpectedly , there was also over-representation within the down-regulated proteins for the terms ‘Mitochondrion organization’ and ‘Cellular component biogenesis’ , suggesting a consequential impact on mitochondrial function , a major destination for new protein synthesis in the cell .","Because the TE of mitochondrial transcripts was down-regulated and mitochondrial organisation emerged from the proteome analysis , specific assays of mitochondrial activity were performed .","Depletion of BUD23 with any of three , independent siRNAs resulted in a > 50% reduction in oxidative phosphorylation ( OXPHOS ) and around a 25% reduction in Electron Transfer Capacity ( ETC ) ( Figure 2C , D , E ) but no observable difference in leak respiration ( LEAK ) , relative to control ( Figure 2—figure supplement 2 ) .","Moreover , no alteration in glycolytic capacity was observed indicating specificity for mitochondrial energetic pathways ( Figure 2F ) .","We also observed a reduction in mitochondrial mRNA abundance with BUD23 knockdown , but no change in mitochondrial genome copy number , implying a functional defect , rather than loss of mitochondrial mass ( Figure 2G , H , I ) , although these measurements were performed soon after siRNA knockdown , and so a later impact on mitochondrial mass resulting from prolonged BUD23 loss cannot be excluded .","To examine whether this mitochondrial transcription defect was specific to BUD23 deficiency , or part of a more general mechanism resulting from 40S/60S imbalance , we performed siRNA-mediated knockdowns of other known ribosome biogenesis factors , LTV1 and RIOK2 , as well as the ribosomal small subunit protein RPS27A ( also reported to result in ribosomal subunit imbalance ) ( Sloan et al . , 2019 ) .","Depletion of LTV1 , RIOK2 and RPS27A all resulted in significant decrease in 18S/28S ratio relative to control siRNA , indicating a similar 40S/60S imbalance as observed with BUD23 knockdown ( Figure 2G , Figure 2—figure supplement 2 ) .","LTV1 knockdown resulted in a similar decrease in mitochondrial transcript expression as observed in BUD23 knockdown ( Figure 2H ) .","However , RPS27A and RIOK2 knockdown-induced ribosome imbalance did not affect mitochondrial transcript expression ( Figure 2H , Figure 2—figure supplement 2 ) .","This indicates that disruption of 40S subunit biogenesis is not by itself sufficient to result in mitochondrial dysfunction .","To check that BUD23 was not regulating mitochondria directly , we blotted for the protein after sub-cellular fractionation .","BUD23 was found to be abundant in the cytosolic fraction but was not detected in the mitochondrial fraction after proteinase K digestion ( Figure 2J ) .","This suggests the mechanism of BUD23 action does not occur within the mitochondria .","Furthermore , TRMT112 , the obligate BUD23 binding partner , was also abundant in the cytosol , but barely detectable within mitochondria ( Figure 2J ) .","Together these data indicate that BUD23-dependent mitochondrial regulation is likely to be a downstream consequence of effects on ribosome composition , function , and cellular protein repertoire , rather than a direct effect of BUD23 protein within the mitochondria .","To investigate the physiological impact of BUD23 we generated Bud23 null mice .","We introduced a frameshift deletion within the Bud23 gene , which resulted in a null allele ( Figure 3A ) .","However , upon further breeding of heterozygous mice a clear deviation from the expected Mendelian ratio was observed in the resultant offspring .","Out of 74 pups born , 39 were found to be wild-type for the Bud23 gene , 35 were heterozygous , and zero were homozygous for the null allele ( Figure 3B ) , indicating embryonic lethality in Bud23-null mice .","This is in contrast to reports in yeast where BUD23 was shown to be non-essential for life ( White et al . , 2008 ) .","Furthermore , it is notable that approximately half the expected heterozygote birth rate was seen , implying an additional haplo-insufficiency developmental death rate .","Accordingly , we analysed embryos at day E10 . 5 and detected Mendelian ratios of wild-type , and heterozygous animals , but again no homozygous null embryos were seen ( Figure 3B ) , suggesting total Bud23 loss is incompatible with embryogenesis , but haploinsufficiency results in fetal death later in development .","Surprisingly , surviving adult heterozygous mice showed no difference compared to wildtype littermate controls in body weight , lean body weight or body fat between 10 and 30 weeks of age ( Figure 3C ) .","There was also no significant change in energy expenditure or respiratory exchange ratio in Bud23+/- mice relative to WT littermate controls ( Figure 3—figure supplement 1 ) , suggesting compensatory mechanisms to permit survival .","To circumvent embryonic lethality , we generated a floxed Bud23 allele mouse , allowing post-natal tissue-specific KO ( Figure 3A ) .","Mice homozygous for this floxed Bud23 ( Bud23fl/fl ) allele showed no observable differences from wild-type littermates .","To confirm the embryonic lethality phenotype using an independent genetic approach , we crossed these mice with a Cre-deleter mouse line ( Hprt-cre ) ( Nichol et al . , 2011 ) to globally delete Bud23 alleles .","Bud23+/- heterozygous crosses also failed to generate any homozygous null offspring ( 0 out of 28 cre positive offspring ) , confirming the observation in the global knockout mouse line ( Figure 3D ) .","We targeted Bud23 disruption to skeletal and cardiac muscle using Muscle Creatine Kinase ( Mck ) -cre as the driver ( Whitnall et al . , 2008 ) .","Cardiac muscle relies heavily on mitochondrial production of ATP in post-natal life , but less so during development , so offering us the chance to analyse BUD23 impact on a highly mitochondrially-dependent tissue .","Bud23fl/fl Mck-Cre+/- mice exhibited Cre-mediated deletion of exon 7 of the Bud23 gene in cardiac tissue , with a consequent reduction in BUD23 protein levels ( Figure 3E , F ) .","Loss of BUD23 in cardiac muscle resulted in sudden death between the age of 28–35 days ( Figure 4A ) .","Littermate control mice expressing Bud23fl/flMck-Cre-/- or Bud23fl/wt Mck-Cre+/- ( i . e . mice retaining at least one functional copy of the Bud23 allele ) were viable , fertile and did not die prematurely ( Figure 4A ) .","Mouse tissues from this line were therefore subsequently harvested at 26 days of age .","A significant reduction in the 18S/28S RNA ratio , consistent with the in vitro data using siRNA knockdown was observed at this time point ( Figure 4B ) .","Proteomic analysis of cardiac tissue recovered from 26 day old Bud23fl/flMck-Cre+/- mice and littermate controls ( Bud23fl/flMck-Cre-/- ) revealed a significant decrease in protein content per cell with loss of BUD23 ( Figure 4C ) , with clear separation by genotype ( Figure 4—figure supplement 1 ) .","When detected proteins were annotated by sub-cellular compartment , mitochondria were found to be subject to the largest loss of protein content , with small reciprocal increases in the other compartments ( Figure 4D ) , again identifying mitochondria as especially sensitive to BUD23 action .","Of the 2047 proteins identified by mass spectroscopy , 347 were found to be significantly up-regulated and 442 were found to be significantly down-regulated in the BUD23 deficient condition ( Figure 4E , Figure 4—source data 1 ) .","Whilst an approximately equivalent number of down and up regulated proteins was identified globally , when this list was refined to mitochondrial proteins only 20 were up-regulated compared to 220 down-regulated ( 50% of all down-regulated proteins ) ( Figure 4F ) .","By contrast , the up-regulated dataset was enriched for translational ( 60S proteins ) and proteasomal proteins ( Figure 4—figure supplement 1 ) .","Whilst correlation between change in protein abundance in BUD23-deficient heart tissue and change in mRNA TE of the corresponding transcripts in the BUD23 siRNA treated A549 cells was poor overall ( Figure 4—figure supplement 2 ) , it is striking that mitochondria are affected to a great extent in both models .","These observations show a strong selectivity for impaired expression of mitochondrial proteins as a result of BUD23 deficiency .","Within the set of mitochondrial proteins , we found that electron transport chain complexes I , IV and V ( ATP synthase ) were most strongly reduced in response to loss of BUD23 ( Figure 4—figure supplement 3 ) .","Furthermore , ontology analysis of all the significantly down-regulated proteins identified three main clusters of proteins: mitochondrial proteins , mitochondrial ribosomal proteins and proteins involved in energy metabolism , indicating a functional mechanism linking between the BUD23 dependent ribosomal defect , through reduced mitochondrial protein expression to result in disruption of energy metabolic pathways ( Figure 4G ) .","Therefore , BUD23 actions on the ribosome result in down-regulation of core mitochondrial proteins involved in ATP synthesis .","Given the observation of preferential dysregulation of mitochondrial proteins , we hypothesised that mitochondrial function was likely to be compromised in BUD23-deficient cardiac tissue .","Mitochondrial DNA copy number was reduced in BUD23 deficient cardiac tissue ( Figure 5A ) .","Functional analysis of mitochondrial function in cardiac homogenates using the Oroboros microrespirometer system revealed that mitochondrial OXPHOS and ETC were both significantly reduced in mice deficient for BUD23 , while LEAK respiration was unaffected ( Figure 5B ) .","The reduction in OXPHOS led to a lower respiratory control ratio ( RCR ) indicating that mitochondria from mice deficient for BUD23 were less efficient at producing ATP .","These effects were apparent regardless of which substrate was used for electron donation .","In addition to reduced mitochondrial capacity , mice deficient for BUD23 had lower citrate synthase activity , indicating a reduction in cardiac mitochondrial density ( Figure 5C ) , which accords with the measured loss of mitochondrial genome copy number ( Figure 5A ) .","To test whether haplo-insufficiency of BUD23 protein was enough to impair mitochondrial function we also tested heart homogenates from Bud23+/- mice .","These were found to have no significant difference in mitochondrial respiratory capacity across any mitochondrial state ( LEAK , OXPHOS , ETS ) , whether normalised to citrate synthase activity or protein content ( Figure 5—figure supplement 1 ) .","It was , however , noted that , compared to wild-types , Bud23+/- mice did exhibit significantly higher citrate synthase activities , indicating greater mitochondrial density .","This may result from increased mitochondrial biogenesis to compensate for defective oxidative phosphorylation function and may partly explain the reduced viability in Bud23+/- mice .","Interestingly , the impact of BUD23 loss on mitochondrial function appears to be sexually-dimorphic .","When mitochondrial respiration was normalised to citrate synthase , male heterozygous ( HZ ) mice exhibited a deficiency in OXPHOS and ETS relative to wildtype , whereas female mice did not ( Figure 5—figure supplement 1 ) .","This observation indicates a more severe phenotype in BUD23 deficient males , with further derangement of individual components of the electron transport chain .","Electron microscopy ( EM ) analysis revealed that despite decreased mitochondrial protein abundance and reduced mitochondrial function in the BUD23 cardiac tissue , the individual mitochondria appear to be formed normally with an equivalent number of cristae compared to the wildtype .","The intermyofibrillar mitochondria were typically arranged in a highly ordered pattern in the wildtype cardiac tissue , however , there was marked disorganisation in the BUD23-deficient cardiac tissue ( Figure 5D ) .","In addition , we identified the presence of numerous spherical electron dense inclusion bodies within the mitochondria of BUD23 deficient cardiac cells ( Figure 5D , yellow arrows ) .","Similar electron dense structures have been previously reported in myocardial mitochondria , which may accumulate in response to increased cellular bioenergetic demand ( Jacob et al . , 1994 ) .","The observation of these structures here may be consistent with attempted mitochondrial adaptation to energetic-demand challenge .","A frequent cellular adaptation to impaired oxidative phosphorylation is a switch to glycolytic ATP generation , and so we measured rate-limiting glycolytic gene expression ( Figure 5E ) .","We identified a significant induction of glucokinase ( GCK ) , and a trend to induction of phosphofructokinase ( PKM ) and hexokinase ( HK2 ) supporting such an adaptive switch to glycolysis .","A further adaptation predicted was an induction in mitochondrial biogenesis programmes .","We identified a significant induction in PGC1a and PGC1b , but no change in TFAM1 expression ( Figure 5E ) .","Taken together these results point to some of the expected cellular responses to loss of oxidative phosphorylation , but it was surprising that the changes were so few .","This implies that additional cellular injury may be proceeding , which acts to limit successful adaptation .","Impaired mitochondrial function may result in overproduction of reactive oxygen species , which can trigger the apoptotic cascade and lead to cell death .","Such a progression may explain the distorted and thinned ventricular walls observed in the cardiac Bud23-null mice .","We found induction of NRF2 ( Figure 5E ) , a key sentinel gene controlling cellular responses to reactive oxygen species stress , which led us to measure reactive oxygen species directly , and found that these were significantly higher in hearts lacking BUD23 expression ( Figure 5F ) , identifying a burden of increased reactive oxygen species in BUD23 deficient cardiomyocytes resulting from the action of deranged mitochondria .","Deficient mitochondrial generation of ATP in cardiac tissue is a key cause of cardiomyopathy .","Given the strong mitochondrial deficiencies observed in BUD23-deficient cardiac tissue , we predicted that the cardiomyopathy was the likely cause of sudden death .","Histological examination of hearts from Bud23fl/fl Mck-Cre+/- ( BUD23-deficient ) mice showed significant enlargement , ventricle dilatation and decreased ventricle wall thickness compared to littermate controls at 26 days old ( Figure 6A–D ) .","Littermate control mice expressing Bud23fl/flMck-Cre-/- or Bud23fl/wt Mck-Cre+/- ( i . e . mice retaining at least one functional copy of the Bud23 allele ) were viable , fertile and did not die prematurely ( Figure 3A ) .","Echocardiogram analysis of Bud23fl/fl Mck-Cre+/-mice at 26–28 days old showed that cardiac BUD23-deficiency resulted in systolic dysfunction indicated by significantly dilated left ventricle ( sD ) , as well as significantly decreased fractional shortening and relative wall thickness ( Figure 6E , Figure 6—source data 1 ) .",") .","There was , however , no observed pulmonary oedema or liver oedema ( Figure 6—figure supplement 1 , ) .","Further analysis of the heart proteomics data from Figure 4 showed an increase in a number of markers of fibrosis , including Col4a2 , CTGF , Galectin-3 and others ( Figure 6—figure supplement 1 ) .","Taken together , these observations indicate that Bud23fl/fl Mck-Cre+/- mice exhibit the early stages of heart failure around 26–28 days of age .","Electrocardiogram analysis of the same mice at the same age revealed significantly shorter QRS and corrected QT ( QTc ) intervals but no change in heart rate ( Figure 6F; Figure 6—source data 1 ) .",") ."],["BUD23 is tightly conserved through evolution and has recently emerged as a ribosomal RNA methyltransferase , although its physiological role in mammals is unexplored .","It is one of the genes deleted in Williams-Beuren Syndrome , a complex multi-gene deletion syndrome with neurological , and energy homeostatic features , although themediating role of individual genes to the phenotype has not been determined ( Morris , 1993 ) .","We now identify BUD23 as playing a major role in regulating the translational efficiency of specific transcripts through its role in ribosome maturation .","Indeed , translation of low 5’UTR GC-content mRNA species were particularly dependent on BUD23 expression , in contrast to the picture seen with simple ribosome deficiency in which transcripts with short/unstructured 5’UTRs were seen to be most affected ( Khajuria et al . , 2018 ) .","Furthermore , we identified a surprising role for BUD23 in maintaining mitochondrial oxidative phosphorylation capacity .","As protein translation is tightly coupled to ATP demand this raised the possibility that BUD23-dependent ribosomal maturation is critical to link the two processes in living cells and tissues ( Morita et al . , 2013 ) .","Indeed , BUD23 loss greatly impaired mitochondrial ATP generation , both in vitro and in vivo .","In mice , cardiomyocyte loss of BUD23 greatly impaired mitochondrial ATP generation leading to dilated cardiomyopathy and premature death , and global loss of BUD23 resulted in embryo-lethality .","BUD23 is thought to perform two functions in the generation of the translational apparatus: firstly , in the processing of pre-18S RNA into its mature form , and secondly imparting a m7-G methyl mark on a key residue .","As the catalytic activity of BUD23 is not required for efficient pre-18S RNA processing , these two functions appear to be independent of each other ( Haag et al . , 2015; Zorbas et al . , 2015 ) .","It is notable that both the BUD23 imparted m7-G methyl mark and structurally complex 5’UTRs are characteristic features of translation specific to Eukaryotes .","We speculate the co-evolution of this regulatory 18S methyl mark with the need for more complex regulation of differential translation .","Further work should focus on delineating the effect of BUD23-dependent 18S maturation , from the role of the methylation mark .","We used proteomics to examine the relative changes in protein abundance which result from targeted loss of BUD23 in cardiomyocytes .","This revealed a selective loss of mitochondrial proteins , explaining the observed reduction in mitochondrial capacity .","As BUD23 appears to have no direct role within mitochondria we propose that the mitochondrial phenotype results from impaired translation of nuclear-encoded genes with a role in mitochondrial homeostasis .","Indeed , we found that the translational efficiency of a number of mitochondria-targeted as well as mitochondrial regulating proteins were BUD23-dependent .","There are many potential candidates linking the change of translational efficiency to the mitochondrial phenotype .","Mammalian target of rapamycin ( mTOR ) , for example , is known to regulate mitochondrial activity and biogenesis by regulation of translation ( Morita et al . , 2013 ) and , interestingly , BUD23 loss resulted in impaired translation of mTOR itself .","We also observed a marked decrease in the translational efficiency of mitochondrially-encoded transcripts .","We initially generated and characterised a frame shift-generating null allele mouse .","The homozygote null animals were all lost prior to embryonic day e10 . 5 and we were unable to identify a cause but considered that a major defect in a mitochondrially–dependent organ such as heart may be a plausible mechanism .","We moved , therefore , to a conditional allele mouse , using which we confirmed the developmental lethality of the null allele by crossing with a global deleter-cre strain ( Hprt-cre ) .","Using the conditional allele mice with a muscle Cre driver we were able to profile the impact of BUD23 in a physiological role .","Loss of BUD23 in heart and skeletal muscle resulted in early death from cardiac failure .","This striking phenotype was accompanied by morphological changes including cardiac dilatation , ventricular wall thinning , and impaired ventricular contractility .","There was evidence for an attempted switch to glycolytic ATP generation in the hearts , with induction of rate-limiting glycolytic enzyme genes; and also adaptation to conditions of oxidative stress with induction of NRF2 .","Mitochondrial dysfunction frequently imposes a burden of oxidative stress on affected cells , which can result in apoptosis , thereby explaining the loss of cardiac muscle seen .","Detailed ultrastructure analysis of cardiac mitochondria revealed no major loss of mitochondrial volume , but there was a striking additional feature seen within the mitochondria: electron-dense , spheroid inclusions .","These were not observed in control hearts , and are not typically seen in mitochondria from individuals with mitochondrial myopathies , but have been reported before under conditions of extreme ATP demand , and to result from bilayer budding from the inner cristae , to generate additional ATP generating surface area ( Jacob et al . , 1994; Somlyo et al . , 1974 ) .","Therefore , we propose this may be a further adaptation to the impaired oxidative phosphorylation function within the mitochondria .","There is a growing recognition that transcriptional control mechanisms are insufficient to explain differentiated cell function .","Recent advances have identified differential translation of mRNA species dependent on 5’UTR features as an important control mechanism , potentially coupling global changes in protein translation with appropriate upregulation of mitochondrial ATP generation to meet demand 2 .","BUD23 , a highly conserved ribosomal RNA methyltransferase which lies within a multigene interval , deleted as a cause of Williams-Beuren syndrome , plays a specific role in highly ATP-hungry cells , promoting translation of mitochondrial proteins , and impacting on oxidative phosphorylation .","This function likely contributes to aspects of the human phenotype , in particular neurological dysfunction , glucose intolerance and obesity .","Complete global loss of Bud23 is embryonic lethal , and early post-natal death follows deletion of Bud23 in the heart .","The cardiac phenotype is severe , with marked mitochondrial dysfunction , and evidence of compensatory changes towards glycolytic ATP generation , and adaptation to oxidative stress .","Therefore , BUD23 plays a critical role in coupling protein translation to mitochondrial function , with implications for mitochondrial diseases , and cardiomyopathies ."],["All mice were routinely housed in 12:12 light/dark ( L:D ) cycles with ad libitum access to food and water .","All experiments were carried out in accordance with the Animals ( Scientific Procedures ) Act 1986 ( UK ) under Home Office protocol number 70/8768 and P3A97F3D1 .","In studies utilizing conditional targeted mice , littermate controls ( floxed/floxed ) were used as control; these mice carried no copies of the Cre or iCre .","Genotyping was performed on all experimental animals .","To assess body composition , whole body , fat mass and lean body mass was assessed using the EchoMRI system ( Echo Medical Systems ) .","To assess metabolic gas exchange , mice were individually housed in indirect calorimetry cages ( CLAMS , Columbus Instruments ) .","Mice were acclimatised to the cages for 48 hr prior to recording data .","Measurements of O2 consumption and CO2 production were recorded every 10 min for >72 hr .","RER was derived from these measures ( VCO2/VO2 ) , as was energy expenditure ( 3 . 815∗VO2 + 1 . 232∗VCO2 ) .","In order to conditionally KO the Bud23 gene by Cre recombinase expression we generated a transgenic mouse line with the critical exon seven flanked by LoxP sites .","Exon seven has an unequal splicing phase , meaning its Cre mediated excision would lead to a frame shift and KO of the Bud23 gene .","This exon is also flanked by large introns giving sufficient space to insert LoxP sites without perturbing normal gene splicing and regulation .","We used CRISPR ( clustered regularly interspaced short palindromic repeats ) -Cas9 in the mouse zygote to integrate these LoxP sites .","Briefly , sgRNA targeting this genomic region , and with minimal off target potential , were identified ( sgRNA-1 GGCATTGGGCCTACTTAAAG-GGG , sgRNA-2 AGTTGAAGGGTTCCATAATG-AGG , sgRNA-3 CTTTACAGCCCAAGACCACT-TGG ) and selected for use .","sgRNA was synthesised in vitro , using protocols developed by Bassett et al . ( 2013 ) .","Forward primers of the form GAAATTAATACGACTCACTATA GGN18-20GTTTTAGAGCTAGAAATAGC ( where GGN18-20 is the sgRNA sequence ) for each guide were combined with the universal reverse primer , AAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATAACGGACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAAC and used in a PCR reaction with high fidelity polymerase ( Phusion , NEB ) .","200 ng of the resulting amplicon were used overnight in an in vitro transcription reaction ( HiScribe , NEB , as per manufacturer’s instructions ) , before purification ( Megaclear , Ambion ) and quantification by nanodrop .","A double stranded DNA ( dsDNA ) repair template was designed to","( i ) incorporate loxP sites , with unique restriction sites for screening purposes , 300 bp upstream and 200 bp downstream of exon seven respectively and","( ii ) integrate silent shield mutations for all three sgRNA to prevent cutting of the repair template and","( iii ) contain 1000 bp 5’ and 3’ homology arms to the target region .","The above template was synthesised in a pUC57 vector ( Genscript , US ) , the linear 5’homology-loxP-exon7-loxP-3’homology fragment excised from the vector by restriction digest , gel extracted ( Bioline ) and further cleaned prior to microinjection ( Monarch PCR purification kit , NEB ) .","An injection mix of the three sgRNA ( 40 ng/ul each ) , Cas9 mRNA ( 100 ng/ul ) and the dsDNA repair template ( 10 ng/ul ) was prepared and directly microinjected into B6D2F1 ( Envigo ) zygote pronuclei using standard protocols .","Zygotes were cultured overnight and the resulting 2 cell embryos surgically implanted into the oviduct of day 0 . 5 post-coitum pseudopregnant mice .","Potential founder mice were screened by PCR and digest .","Animals positive for both 5’ and 3’ LoxP integration had the region fully verified by pCR-Blunt cloning followed by Sanger sequencing .","two founder mice were identified and then back-crossed to C57BL/6J wild-type mice to assess germline penetrance .","Additionally , the described CRISPR targeting strategy resulted in deletion of critical exon 7 , a global Bud23 KO line was generated from this founder animal .","DNA was extracted from ear snip or tail tip using REDExtract-N-Amp tissue PCR kit ( Sigma ) .","Primer sequences and PCR reaction conditions are listed in the key resources table and the primer sequences supplementary file .","PCR products were resolved on 1 . 5% agarose gels .","STR authenticated A549 cells were obtained from the European Collection of Cell Cultures ( ECACC Cat no: 86012804 ) and maintained in culture , incubated at 37°C in humidified air , 5% v/v CO2 .","Cultured cells were tested regularly for mycoplasma contamination .","A549 cells were cultured in DMEM growth medium ( Sigma , D6546 ) , supplemented with 2 mM L-Glutamine ( Sigma , G7513 ) , 10% foetal calf serum ( FCS ) ( Sigma , F96665 ) .","A549 cells were plated at 1 × 10^6 cells in a 10 cm cell culture dish and transfected with BUD23 specific or non-targeting control siRNA ( S41529 , S41530 , S41531 , negative control 1 , negative control 2 , Ambion Silencer Select ) using Dharmafect DF-1 according to the manufacturer’s guidelines .","A549 cell pellets were lysed in M-PER mammalian protein extraction reagent ( Thermo Scientific ) .","Whole mouse hearts were dissected , quartered and washed in ice cold PBS to remove blood before disruption of the tissue in 1x TBS supplemented with protease inhibitor cocktail using a TissueRuptor ( Qiagen ) .","SDS was added to a final concentration of 4% ( w/v ) prior to sonication .","Samples were then reduced with 100 mM DTT and heated for 3 min at 95°C .","Samples were prepared for label-free quantification in accordance with published protocols ( Hernandez-Valladares et al . , 2016 ) .","Digested samples were analysed by LC-MS/MS using an UltiMate 3000 Rapid Separation LC ( RSLC , Dionex Corporation , Sunnyvale , CA ) coupled to a Q Exactive HF ( Thermo Fisher Scientific , Waltham , MA ) mass spectrometer using a stepped normalized collision energy centered at 28 .","A549 peptide mixtures were separated using a multistep gradient from 95% A ( 0 . 1% FA in water ) and 5% B ( 0 . 1% FA in acetonitrile ) to 7% B at 1 min , 18% B at 58 min , 27% B in 72 min and 60% B at 74 min at 300 nL min−1 , using a 75 mm x 250 μm i . d . 1 . 7 μM CSH M-Class C18 , analytical column ( Waters ) , for a final run time of 90 min .","Cardiac samples were separated using a gradient from 5% B to 7% B at 1 min , 18% B at 81 . 5 min , 27% B at 102 min and 60% B at 104 min , with a final run time of 120 min .","The top 12 precursors were selected for fragmentation automatically by data dependant analysis during each cycle .","Ionization potential was set at 1900V with a survey scan window of 300–1750 m/z .","MS1 used a resolution of 120 , 000 with a target ion intensity of 3e6 .","Higher-energy collisional dissociation was used .","MS2 was set at a resolution of 60 , 000 with a target ion intensity of 2e5 .","Maximum injection times for MS1 and MS2 were 20 ms and 110 ms respectively .","Peptides were dynamically excluded for 15 s after one occurrence .","Mass spectra were analysed using MaxQuant version 1 . 6 . 0 . 16 ( Cox and Mann , 2008 ) , searching against either the Uniprot Mus musculus database ( accessed 21/12/2017 ) or Homo sapiens database ( accessed 12/03/2017 ) using the native Andromeda search engine .","Match between runs was selected with a retention time of 1 min , using default parameters for all other settings .","MS1 search tolerance was set at 20ppm , and MS2 at 4 . 5ppm .","Modifications were set as Carbamidomethyl ( C ) for fixed , and Oxidation ( M ) and Acetyl ( Protein N-term ) variable .","A minimum of 1 spectra was required for identification , while a minimum of 2 spectra were required for quantification .","Resulting data was processed using Perseus version 1 . 6 . 0 . 7 ( Tyanova et al . , 2016 ) and R statistical software ( Wickham , 2009; R Development Core Team , 2017 ) .","All proteins identified by site only , as a potential contaminant , or in the decoy reverse database were excluded .","Data was then filtered to exclude proteins with more than 3 ‘zero-values’ across all groups .","All proteins assigned to the GO term ‘blood microparticle’ were removed as well .","LFQ values were log ( 2 ) transformed and missing values imputated in Perseus using a normal distribution ( Lazar et al . , 2016 ) .","The width of the new distribution was 0 . 3 standard deviations of the data , and was shifted downwards by 1 . 8 standard deviations of the data .","A total of 6 . 68% of values were imputated .","Significance was determined using a T-test with a permutation based FDR ( FDR < 0 . 05 , s0 = 0 . 1 ) .","s0 represents the relative importance of the difference between the means , with non-zero s0 values taking fold change , and not only p-value , into account .","Estimations of copy numbers were performed using the proteomic ruler plug-in version 0 . 1 . 6 ( Wiśniewski et al . , 2014 ) .","Proteins were then mapped to subcellular localisations to determine percentage protein mass by organelle according to published protocols using the HeLa spatial proteome database ( Doll et al . , 2017; Itzhak et al . , 2016 ) .","Interaction networks were built using String 10 online software ( Szklarczyk et al . , 2017 ) using seven data sources ( textmining , experiments , databases , co-expression , neighborhood , gene fusion co-occurrence ) and imported into Cytoscape ( Shannon et al . , 2003 ) .","Network edges represent confidence .","A minimum interaction score of 0 . 4 ( medium confidence ) was required for network inclusion .","Enrichment analysis was performed using PANTHER ( Mi et al . , 2017 ) .","Lists of proteins were exported for the identified GO terms and then manually coloured using Cytoscape for visualization .","Transthoracic echocardiography ( TTE ) was performed using a Vevo 770 High-Resolution Imaging System ( Visualsonics Inc , Toronto , Canada ) and a 30MHz probe .","Mice were lightly anesthetized with 1–1 . 5% isoflurane via facemask , maintaining the heart rate at approximately 450 beats/min .","Views were taken in planes that approximated the parasternal short-axis view and the apical long-axis view .","M-mode tracings were used to determine left ventricular end-diastolic diameter ( LVDD ) and end-systolic diameter ( LVSD ) , posterior wall thickness in diastole ( LVPWD ) and systole ( LVPWS ) , and interventricular septum thickness in diastole ( IVSD ) and systole ( IVSS ) over three cardiac cycles .","The analysis was performed blinded to animal details .","LV fractional shortening ( FS ) was calculated using the formula FS = [ ( LVDD - LVSD ) / ( LVDD ) ] x 100 .","Relative wall thickness ( RWT ) was calculated using the equation RWT = ( IVDD + LVPWD ) /LVDD .","Mice were lightly anesthetized with 1–1 . 5% isoflurane via facemask .","The body temperature was maintained around 37°C using a heat pad .","The lead II ECG was recorded using Power Lab/4SP system ( Adinstruments ) from needle electrodes inserted subcutaneously into the left and right forelimbs and the right hindlimb .","The signal was acquired for about 5 min using Chart 7 software ( Adinstruments ) .","During offline analysis , the 5 min recording was examined for unusually shaped P , QRS , or ‘T’ waves and for time-varying phenomenon ( e . g . irregularities in interval durations ) .","Ectopic or abnormal beats are noted .","A representative 10–15 s segment of the recording was averaged to obtain the signal averaged ECG to calculate Heart rate , RR interval , P wave duration , PR interval , QRS , JT and QT durations .","QT duration was corrected ( QTc ) using the Bazett's formula ( Bazett , 1997 ) .","Polysome profiles were prepared according to the protocols previously published in: ( Gandin et al . , 2014 ) .","Two conditions were profiled , BUD23 siRNA and Control siRNA , with an n of 3 samples per condition .","Fractions were collected from each sample and RNA was purified using the Trizol extraction method .","RNA from fractions containing three or less ribosomes were pooled ( ‘Low translational efficiency ) and RNA from fractions containing more than three ribosomes ( ‘high efficiency’ ) were pooled .","Sub-polysomal fractions were similarly pooled .","RNA sequencing was used to quantify the expression of RNA species in each of the pooled samples .","Translational efficiency ( TE ) was defined as the relative abundance of a transcript in the high and low efficiency pooled fractions .","A TE score for each transcript was calculated by averaging the ratio derived from each of the three replicates in each condition .","TE was plotted for the two conditions using R software .","Translational efficiency ( TE ) was then compared to both the length and GC content of the 5' UTR region .","5' UTRs were defined using the TxDb . Hsapiens . UCSC . hg38 . knownGene ( Bioconductor Core Team and Bioconductor Package Maintainer , 2016 ) and org . Hs . eg . db ( Carlson , 2018 ) R packages .","Sequences of the regions were extracted using the twoBitToFa program from the BLAT suite ( Kent , 2002 ) and then GC content was calculated using the seqinr R package ( Charif and Lobry , 2007 ) .","The Pearson correlation coefficients , and their relative significances , were calculated using cor . test function from the stats R package ( R Development Core Team , 2017 ) .","Live cell metabolic assays were performed on A549 cells pre-treated with either BUD23 siRNA or control siRNA using a Seahorse XFe 96 analyser ( Agilent ) .","Mitochondrial stress tests was performed according to the manufacturer’s instructions ( Agilent ) .","2M Oligomycin ( OA ) , 1M FCCP and 0 . 5M rotenone/antimycin A ( AA+R ) were used for all conditions .","A549 cells were plated into Seahorse XF96 plates at 160 , 000 cells per well , utilising 16 wells per condition .","Cell density was normalised using SRB assay .","Experiments were performed in triplicate .","Of the three distinct experimental preparations available ( isolated mitochondria , permeabilized fibers , and tissue homogenates ) , we utilized cardiac homogenates for mitochondrial assessment ( see Goo et al . , 2013 for advantages of this method ) .","Ventricular tissue ( ~20 mg ) was weighed and transferred to 1 . 5 ml of ice-cold MiRO5 medium ( in mM: EGTA 0 . 5 , MgCl2 1 . 4 , taurine 20 , KH2P04 10 , HEPES 20 , BSA 1% , K-MES 60 mM , sucrose 110 mM , pH 7 . 1 , adjusted with 5 N KOH ) .","Tissue was homogenized for 3 s in 1 s bursts with a tissue homogenizer and loaded immediately ( 40 µg/ml ) into an Oroboros Oxygraph 2 k high resolution respirometry system ( Oroboros Instruments , Innsbruck , Austria ) for measurement of mitochondrial respiration combined with the Fluorescence-Sensor Green of the O2k-Fluo LED2-Module for H2O2 measurement ( see Makrecka-Kuka et al . , 2015 for full details ) .","Two identical respiration chambers ( chamber A and chamber B ) held at the same temperature were run in parallel for each experimental run .","All measurements of respiration rates were carried out at 37°C .","Oxygen electrodes were calibrated daily with air-saturated respiration solution Zero calibrations were achieved by injecting yeast into the experimental chambers .","Oxygen solubility in the assay medium was calculated as described previously ( Lienig and Forstner , 1984 ) .","H2O2 flux was measured simultaneously with respiration in the O2k-Fluorometer using the H2O2-sensitive probe Amplex UltraRed ( 10 μM ) with 1 U/mL horse radish peroxidase ( HRP ) and 5 U/mL superoxide dismutase ( SOD ) .","AmR in the presence of H2O2 is catalysed by HRP to the fluorescent product resorufin .","Calibrations were performed with two sequential injections of H2O2 at 0 . 1 μM steps .","Three parameters are commonly used to assess mitochondrial function ( for reviews , see Brand and Nicholls , 2011; Pesta and Gnaiger , 2012 ) .","Firstly , the capacity for oxidative phosphorylation ( OXPHOS ) is the respiratory capacity of mitochondria in the ADP-activated state of oxidative phosphorylation ( saturating concentrations of ADP , inorganic phosphate , oxygen , and defined substrates ) .","Secondly , LEAK respiration rate represents mitochondrial respiration that occurs in the absence of ATP generation , mainly to compensate for proton leak across the mitochondrial inner membrane .","Lastly , the respiratory electron transfer-pathway capacity ( ETC ) is mitochondrial respiration in the noncoupled state in the presence an uncoupler; this induces maximum oxygen flux through the electron transport chain .","The Respiratory Control Ratio ( RCR , calculated here as OXPHOS/LEAK ) provides a measure of the degree of coupling between oxidation and phosphorylation , or in other words , the efficiency of mitochondrial ATP production .","OXPHOS , LEAK and ETS were measured in the presence of Complex I substrates pyruvate and malate ( electron transfer through Complexes I-IV ) or Complex I+II substrates ( addition of succinate ) .","Additionally , respiratory flux with electron transfer through Complex IV alone was measured via the addition of the electron donor tetramethyl-phenylene-diamine ( TMPD ) .","The protocol used to measure these parameters was adapted from Pesta and Gnaiger ( 2012 ) .","Briefly , pyruvate ( 5 mmol l −1 ) , malate ( 0 . 25 mmol l−1 ) and glutamate ( 10 mmol l−1 ) are added as carbon substrates and to spark the citric acid cycle .","Under these conditions , mitochondria are in LEAK respiration with CI substrates in the absence of adenylates .","OXPHOS with CI substrates was achieved through addition of saturating levels of ADP ( 2 mM l−1 ) .","Following steady-state conditions , succinate ( 10 mmol l−1 ) was added to achieve OXPHOS with CI+CII substrates .","To uncouple respiration and achieve ETS with CI+CII substrates , carbonyl cyanide 4-trifluoromethoxyphenylhydrazone ( FCCP ) was carefully titrated to a maximum concentration of 0 . 25 μmol l−1 .","Rotenone was then added to achieve ETS with CII substrates and antimycin A ( 5 μmol l−1 ) was given to block Complex III and measure background non-mitochondrial residual oxygen consumption ( ROX ) .","OXPHOS through Complex IV alone was assessed by adding the electron donor TMPD ( 0 . 5 mmol l−1 ) .","To avoid oxidation of TMPD , ascorbate ( 2 mMol l−1 ) was added prior to TMPD injection .","The citrate synthase activity of cardiac homogenates were measured in frozen homogenates .","Briefly , maximal activity ( Vmax ) was determined with a spectrophotometer at 37°C with a Synergy HTX spectrophotometer ( BioTek , UK ) in assay buffer; 50 mM TRIS-HCl , pH 8 . 0 .","Citrate synthase activity was monitored in the presence or absence of oxaloacetate by the appearance of 5-thio-2-nitrobenzoic acid as a result of the reaction of free acetyl-CoA with 5 , 5’-dithiobis ( 2-nitrobenzoic acid ) at 412 nm over a 10 min incubation period ( assay buffer with 0 . 5 oxaloacetate , 0 . 3 mM acetyl-CoA , 0 . 15 mM 5 , 5-dithiobis-2-nitrobenzoic acid ) .","Extinction coefficients were empirically determined to quantify Vmax values .","Enzyme activities were normalised to total soluble protein , which was quantified according to Bradford ( Bradford , 1976 ) .","Total cell protein was isolated from cells using protein extraction buffer ( 50 mM Tris , 150 mM NaCl , 1% TritonX-100 , supplemented with protease inhibitors ) .","Total protein was isolated from snap-frozen tissue using RIPA buffer supplemented with protease inhibitor cocktail after disruption using Lysing Matrix D tubes ( MP Bio ) .","All experiments were performed according to current UK Home Office regulations and under approval of the relevant University of Manchester ( Manchester , UK ) local ethics committee .","Six mice were used for this study ( n = 3 control , n = 3 Bud 23 KO ) .","Immediately after euthanasia , hearts were harvested and small portions ( cubes with sides smaller than 0 . 5 mm ) of right and left ventricle were immersion fixed in 2 . 5% glutaraldehyde and 2% paraformaldehyde in 100 mM sodium cacodylate buffer ( pH 7 . 2 ) overnight .","Specimen preparation was performed , with small modifications , according to the Ellisman protocol ( Holcomb et al . , 2013 ) .","Briefly , after washings in sodium cacodylate , samples were sequentially stained in: 2% osmium tetroxide and 1 . 5% potassium ferrocyanide in 100 mM sodium cacodylate for 1 hr; 1% aqueous thiocarbohydrazide for 20 min; 2% aqueous osmium for 30 min; 1% aqueous uranyl acetate overnight and Walton’s lead aspartate for 30 min the following morning at 60°C .","Samples were washed 3 times for 10 min in double distilled water after each staining step .","Staining was followed by dehydration in ethanol ascending series ( 50% , 70% , 90% , 100% ) and infiltration with TAAB 812 hard resin mixed with propylene oxide .","After embedding , resin blocks were cured at 70° C for 40 hr .","Plastic blocks were sectioned at 80 nm thickness using a Leica UC6 ultramicrotome and sections were imaged using a FEI Tecnai12 Biotwin operated at 100kV .","Images were analysed using Fiji ( Schindelin et al . , 2012 ) .","Fractionation experiments were performed by differential centrifugation as previously described previously ( Rorbach et al . , 2014 ) .","Unless otherwise stated in the figure legend , parametric data was analysed by ANOVA with a Dunnett’s multiple comparisons test .","Non-parametric data was analysed using a Mann-Whitney test .","RNAseq data was analysed using edgeR ( Robinson et al . , 2010 ) .","For Mass Spectrometry data analysis please refer to the dedicated section ."]],"string":"[\n [\n \"Regulation of protein abundance within a cell is of fundamental importance to homeostasis , cell identity and responsiveness to changes in external environment .\",\n \"Many layers of regulatory control have evolved to fine-tune gene expression to meet cellular needs , including transcriptional regulation , cis-acting genomic elements , epigenetic structures , and translational efficiency .\",\n \"Selective control of mRNA translation at the level of the ribosome is emerging as important for protein dynamics within the cell; however , the regulatory mechanisms remain unclear ( Sonenberg and Hinnebusch , 2009 ) .\",\n \"Protein synthesis is the most energetically-demanding process for a cell to perform and , therefore , mRNA translation is closely coupled to mitochondrial function ( Morita et al . , 2013 ) .\",\n \"Mitochondria are fundamental cellular components of eukaryotes , generating the majority of cellular ATP .\",\n \"The human mitochondrial genome contains only 37 genes , of which 13 are subunits of respiratory complexes , 22 encode mitochondrial tRNAs , and a further two encode rRNA .\",\n \"Whilst the mitochondrion translates the 13 protein-coding genes of its own genome using a bespoke mitochondrial apparatus , the majority of the genes required for a functional mitochondrion are encoded in the cell nuclear genome and are , therefore , dependent on the cytosolic translational apparatus .\",\n \"This relationship implies a fundamental role for the eukaryotic 80S ribosome in control of mitochondrial abundance and function .\",\n \"Indeed , mitochondrial content and function varies between tissues and cell-types , as well as within a tissue , in response to external and internal stimuli including energy demand , oxidative stress and cellular signals ( Palmer et al . , 2011 ) .\",\n \"For example , red blood cells contain no mitochondria whereas in cardiomyocytes mitochondria occupy 30% of cell volume to meet the exceptionally high ATP demand ( Piquereau et al . , 2013 ) .\",\n \"Mitochondrial dysfunction has been linked to a wide range of cardiac disorders , often due to the aberrant production of ROS , and attendant cell death ( Kanaan and Harper , 2017; Ott et al . , 2007 ) .\",\n \"BUD23 ( previously known as WBSCR22 ) was initially identified as a putative methyltransferase implicated in tumour metastases ( Nakazawa et al . , 2011 ) .\",\n \"Its expression is responsive to diverse inflammatory and cancer pathologies ( Jangani et al . , 2014 ) , and reduction of BUD23 expression can affect cellular response to glucocorticoid and alter histone methylation ( Jangani et al . , 2014 ) .\",\n \"However , BUD23 actions were diverse , and a unifying mechanism of action was elusive .\",\n \"More recent studies identified ribosomal RNA as the preferred substrate of BUD23 ( Haag et al . , 2015; White et al . , 2008; Zorbas et al . , 2015 ) .\",\n \"There is little understanding of the physiological role of BUD23 in a mammalian context although it is one of the genes deleted in Williams-Beuren syndrome ( WBS ) .\",\n \"WBS patients have a complex phenotype with prominent neurological and morphological features .\",\n \"In addition metabolic pathologies exist including diabetes and obesity ( Morris , 1993 ) .\",\n \"The contribution of individual genes within the 22-gene interval to the human phenotype has not been determined , but the neurological and metabolic features suggest a bioenergetic contribution .\",\n \"BUD23 is a ribosomal RNA methyltransferase which imparts a methyl mark on a key guanosine residue ( forming N7-methylguanosine ) located between the E and P site of the small ribosomal subunit that has been mapped to residue G1575 of yeast 18S rRNA and G1639 of human 18S rRNA ( Haag et al . , 2015; Õunap et al . , 2013; White et al . , 2008; Zorbas et al . , 2015 ) .\",\n \"BUD23 protein is found in both the nucleus and the cytoplasm ( Õunap et al . , 2015 ) , and its depletion leads to a nuclear accumulation of 18SE-pre-RNA ( Haag et al . , 2015 ) .\",\n \"In both yeast and human cells the methyltransferase catalytic activity of BUD23 is not required for the processing of 18S pre-RNA , or the synthesis of 40S subunits , indicating that BUD23 has an additional role , distinct from its methyltransferase activity ( Haag et al . , 2015; Zorbas et al . , 2015 ) .\",\n \"In human cell lines it has been suggested that some 18S rRNA lacks the m7-G1639 mark , which may indicate a selective role in ribosome function ( Haag et al . , 2015 ) .\",\n \"BUD23 stability requires interaction with TRMT112 ( TRM112 in yeast ) , an obligate binding partner , which is known to stabilise four client methyltransferase enzymes , all involved in the generation of the translational apparatus ( Bourgeois et al . , 2017; Létoquart et al . , 2014 ) .\",\n \"The binding of TRMT112 with these four methyltransferases is competitive , which results in tight regulation of BUD23 protein concentration and enzymatic function at the level of protein stability .\",\n \"Here we identify a novel mechanism linking the energy-demanding process of protein translation to mitochondrial dynamics through BUD23 .\",\n \"We examine the role of BUD23 in ribosome function and discover that BUD23 preferentially promotes the selection of mRNA species with low GC-content 5’UTRs .\",\n \"We also identify a role for BUD23 in mitochondrial transcript translation which impacts mitochondrial function in vitro .\",\n \"We go on to examine the role of BUD23 in a murine in vivo system and discover that the production of mitochondrial proteins is dependent on BUD23 .\",\n \"Finally , we examine the role of BUD23 in the mitochondrially-rich and energetically-demanding cardiac tissue and discover a cardiomyopathy phenotype leading to premature death .\",\n \"These discoveries identify BUD23 as essential for mammalian mitochondrial function , with implications for human mitochondrial disease and cardiomyopathy .\"\n ],\n [\n \"We previously identified pleiotropic actions of BUD23 in airway epithelial cells , and attributed these to an epigenetic effect ( Jangani et al . , 2014 ) .\",\n \"The discovery that BUD23 modifies ribosomal RNA rather than histone proteins prompted re-examination of the BUD23 role in physiology .\",\n \"To identify candidate pathways affected by BUD23 we depleted expression in human airway epithelial cells ( Figure 1A , B ) .\",\n \"This intervention caused a small but statistically significant decrease in cell proliferation ( Figure 1C ) , which is similar to the anti-proliferative effects caused by BUD23 deletion in yeast ( White et al . , 2008 ) .\",\n \"To investigate comprehensively the role of BUD23 in ribosome function we performed polysome profiling .\",\n \"The 48 hr knockdown of BUD23 resulted in a reduction of the 40S subunit peak , and a concomitant increase in the 60/80S peak ( Figure 1D , E ) .\",\n \"There was , however , little change in the profile of the polysome fractions at this early time-point following transient BUD23 knockdown , allowing us to investigate the changes in mRNA substrate selection in the intact polysomes .\",\n \"To test the global efficiency of the translational apparatus in cells lacking BUD23 , we used 35-S-methionine incorporation for 1 hr ( Figure 1—figure supplement 1 ) .\",\n \"This demonstrated little overall impact on global protein translation rate , confirming that the identified polysomes were essentially competent at this time-point , despite the impact on 40S maturation .\",\n \"In order to determine if BUD23-loss differentially impacted the translation of a specific subset of mRNAs , we profiled the translational efficiency ( TE ) of individual mRNA species within the polysome profiles .\",\n \"To do this , the polysome fractions were divided into ‘heavy’ ( more than three ribosomes ) or ‘light’ ( less than three ribosomes ) and pooled , prior to RNA extraction and sequencing .\",\n \"This defined a total of 14 , 527 transcripts , from which the relative proportion of each individual transcript in the heavy compared to the light fraction was calculated to give a translation efficiency ( TE ) score .\",\n \"Average TE scores across the three replicates for each condition ( + /- BUD23 ) were then plotted against each other ( Figure 1F ) .\",\n \"This revealed that the most efficiently translated transcripts in the control samples ( more heavily loaded with ribosomes ) , were most significantly reduced in the BUD23 knockdown condition .\",\n \"To examine the identity of the most highly affected transcripts , a threshold of >1 or <-1 difference in the relative TE was set ( i . e . TE doubled or halved ) , and applied to Figure 1F , shown as trend lines .\",\n \"There were 650 transcripts with a TE ratio <-1 and 95 transcripts where the TE ratio >1 after BUD23 depletion; indicating the marked loss of high efficiency mRNA translation .\",\n \"The 650 transcripts for which TE was reduced ( TE ratio <-1 after BUD23 depletion ) constituted a highly connected network , with a significantly higher number of protein-protein interactions than would be expected from a random dataset of similar size , and an interaction analysis ( PPI ) enrichment p-value<1e−16 .\",\n \"Gene set analysis of these transcripts revealed enrichment for genes involved in RNA processing , metabolic processes and organelle organisation among other high energy demand processes ( Figure 1—figure supplement 1 ) .\",\n \"Surprisingly , the TE of mitochondrially-encoded transcripts was observed to be particularly strongly down-regulated ( Figure 1G ) .\",\n \"In fact , all mitochondrial transcripts had a negative change in TE after BUD23 loss , with approximately 50% reduction in TE .\",\n \"To investigate the mechanism underlying the BUD23 dependent selection of transcripts we examined specific mRNA features associated with translational efficiency .\",\n \"Given the known influence of 5’UTR features on translation efficiency , we reasoned that specific features of the 5’UTR may be particularly important for BUD23 action .\",\n \"A number of well-characterised 5’UTR features impact mRNA translation , including length , secondary structure , GC content , and specific motifs including the TOP sequence ( Gandin et al . , 2016; Meijer and Thomas , 2002 ) .\",\n \"We saw no enrichment for TOP sequence content in the differentially translated mRNA species ( Figure 1—figure supplement 2 ) , and no correlation between change in TE score and 5’UTR length was observed ( Figure 1H ) .\",\n \"We also noted that transcripts with shorter total lengths were not more likely to shift from the heavy to the light fraction ( Figure 1—figure supplement 3 ) .\",\n \"There was , however , a highly significant correlation between change in TE score and GC content ( Pearson’s correlation value of −0 . 1337 , p-value<2 . 2e−16 ) ( Figure 1I ) .\",\n \"This identifies an intersection between mRNA transcript 5’UTR GC content and BUD23 ribosome maturation .\",\n \"There was no difference in the distribution of 5’UTR GC in mitochondrial-related transcripts relative to all transcripts detected in the polysome analysis ( Figure 1—figure supplement 3 ) .\",\n \"Our studies reveal the consequences for the ribosome of BUD23 action and point to an impact on the cellular proteome .\",\n \"Therefore , to investigate the downstream consequences of BUD23 loss we examined the cellular proteome by LC-MS/MS .\",\n \"BUD23 knockdown downregulated 83 proteins and upregulated 64 , out of a total of 3255 identified ( Figure 2—source data 1 ) .\",\n \"When tested using PANTHER GO cellular component ontology analysis , upregulated proteins were over-represented for components of the large ribosomal subunit ( 14/64 proteins ) , whilst downregulated proteins were over-represented for components of the small ribosomal subunit ( 22/83 proteins ) , ( Figure 2A , Figure 2—figure supplement 1 ) .\",\n \"This imbalance in the relative abundance of ribosomal proteins , recapitulates the observation in the polysome profiles and confirms a major role for BUD23 in the maturation of the protein translation apparatus of the cell .\",\n \"Interestingly , TRMT112 protein , an obligate partner for BUD23 , was also significantly down-regulated with BUD23 knock-down , which may indicate a reciprocal stabilising interaction .\",\n \"Ontology analysis of the proteomics dataset was used to predict the functional consequences of BUD23 reduction ( PANTHER GO biological processes ) .\",\n \"Over-representation analysis of the up-regulated proteins overlapped significantly with the ‘protein translation’ .\",\n \"Analysis of the down-regulated proteins again showed a statistically significant over-representation for proteins associated with GO biological terms including ‘translation’ , as well as ‘rRNA metabolic process’ and ‘protein metabolic process’ which support a major effect on the translational apparatus ( Figure 2B ) .\",\n \"Unexpectedly , there was also over-representation within the down-regulated proteins for the terms ‘Mitochondrion organization’ and ‘Cellular component biogenesis’ , suggesting a consequential impact on mitochondrial function , a major destination for new protein synthesis in the cell .\",\n \"Because the TE of mitochondrial transcripts was down-regulated and mitochondrial organisation emerged from the proteome analysis , specific assays of mitochondrial activity were performed .\",\n \"Depletion of BUD23 with any of three , independent siRNAs resulted in a > 50% reduction in oxidative phosphorylation ( OXPHOS ) and around a 25% reduction in Electron Transfer Capacity ( ETC ) ( Figure 2C , D , E ) but no observable difference in leak respiration ( LEAK ) , relative to control ( Figure 2—figure supplement 2 ) .\",\n \"Moreover , no alteration in glycolytic capacity was observed indicating specificity for mitochondrial energetic pathways ( Figure 2F ) .\",\n \"We also observed a reduction in mitochondrial mRNA abundance with BUD23 knockdown , but no change in mitochondrial genome copy number , implying a functional defect , rather than loss of mitochondrial mass ( Figure 2G , H , I ) , although these measurements were performed soon after siRNA knockdown , and so a later impact on mitochondrial mass resulting from prolonged BUD23 loss cannot be excluded .\",\n \"To examine whether this mitochondrial transcription defect was specific to BUD23 deficiency , or part of a more general mechanism resulting from 40S/60S imbalance , we performed siRNA-mediated knockdowns of other known ribosome biogenesis factors , LTV1 and RIOK2 , as well as the ribosomal small subunit protein RPS27A ( also reported to result in ribosomal subunit imbalance ) ( Sloan et al . , 2019 ) .\",\n \"Depletion of LTV1 , RIOK2 and RPS27A all resulted in significant decrease in 18S/28S ratio relative to control siRNA , indicating a similar 40S/60S imbalance as observed with BUD23 knockdown ( Figure 2G , Figure 2—figure supplement 2 ) .\",\n \"LTV1 knockdown resulted in a similar decrease in mitochondrial transcript expression as observed in BUD23 knockdown ( Figure 2H ) .\",\n \"However , RPS27A and RIOK2 knockdown-induced ribosome imbalance did not affect mitochondrial transcript expression ( Figure 2H , Figure 2—figure supplement 2 ) .\",\n \"This indicates that disruption of 40S subunit biogenesis is not by itself sufficient to result in mitochondrial dysfunction .\",\n \"To check that BUD23 was not regulating mitochondria directly , we blotted for the protein after sub-cellular fractionation .\",\n \"BUD23 was found to be abundant in the cytosolic fraction but was not detected in the mitochondrial fraction after proteinase K digestion ( Figure 2J ) .\",\n \"This suggests the mechanism of BUD23 action does not occur within the mitochondria .\",\n \"Furthermore , TRMT112 , the obligate BUD23 binding partner , was also abundant in the cytosol , but barely detectable within mitochondria ( Figure 2J ) .\",\n \"Together these data indicate that BUD23-dependent mitochondrial regulation is likely to be a downstream consequence of effects on ribosome composition , function , and cellular protein repertoire , rather than a direct effect of BUD23 protein within the mitochondria .\",\n \"To investigate the physiological impact of BUD23 we generated Bud23 null mice .\",\n \"We introduced a frameshift deletion within the Bud23 gene , which resulted in a null allele ( Figure 3A ) .\",\n \"However , upon further breeding of heterozygous mice a clear deviation from the expected Mendelian ratio was observed in the resultant offspring .\",\n \"Out of 74 pups born , 39 were found to be wild-type for the Bud23 gene , 35 were heterozygous , and zero were homozygous for the null allele ( Figure 3B ) , indicating embryonic lethality in Bud23-null mice .\",\n \"This is in contrast to reports in yeast where BUD23 was shown to be non-essential for life ( White et al . , 2008 ) .\",\n \"Furthermore , it is notable that approximately half the expected heterozygote birth rate was seen , implying an additional haplo-insufficiency developmental death rate .\",\n \"Accordingly , we analysed embryos at day E10 . 5 and detected Mendelian ratios of wild-type , and heterozygous animals , but again no homozygous null embryos were seen ( Figure 3B ) , suggesting total Bud23 loss is incompatible with embryogenesis , but haploinsufficiency results in fetal death later in development .\",\n \"Surprisingly , surviving adult heterozygous mice showed no difference compared to wildtype littermate controls in body weight , lean body weight or body fat between 10 and 30 weeks of age ( Figure 3C ) .\",\n \"There was also no significant change in energy expenditure or respiratory exchange ratio in Bud23+/- mice relative to WT littermate controls ( Figure 3—figure supplement 1 ) , suggesting compensatory mechanisms to permit survival .\",\n \"To circumvent embryonic lethality , we generated a floxed Bud23 allele mouse , allowing post-natal tissue-specific KO ( Figure 3A ) .\",\n \"Mice homozygous for this floxed Bud23 ( Bud23fl/fl ) allele showed no observable differences from wild-type littermates .\",\n \"To confirm the embryonic lethality phenotype using an independent genetic approach , we crossed these mice with a Cre-deleter mouse line ( Hprt-cre ) ( Nichol et al . , 2011 ) to globally delete Bud23 alleles .\",\n \"Bud23+/- heterozygous crosses also failed to generate any homozygous null offspring ( 0 out of 28 cre positive offspring ) , confirming the observation in the global knockout mouse line ( Figure 3D ) .\",\n \"We targeted Bud23 disruption to skeletal and cardiac muscle using Muscle Creatine Kinase ( Mck ) -cre as the driver ( Whitnall et al . , 2008 ) .\",\n \"Cardiac muscle relies heavily on mitochondrial production of ATP in post-natal life , but less so during development , so offering us the chance to analyse BUD23 impact on a highly mitochondrially-dependent tissue .\",\n \"Bud23fl/fl Mck-Cre+/- mice exhibited Cre-mediated deletion of exon 7 of the Bud23 gene in cardiac tissue , with a consequent reduction in BUD23 protein levels ( Figure 3E , F ) .\",\n \"Loss of BUD23 in cardiac muscle resulted in sudden death between the age of 28–35 days ( Figure 4A ) .\",\n \"Littermate control mice expressing Bud23fl/flMck-Cre-/- or Bud23fl/wt Mck-Cre+/- ( i . e . mice retaining at least one functional copy of the Bud23 allele ) were viable , fertile and did not die prematurely ( Figure 4A ) .\",\n \"Mouse tissues from this line were therefore subsequently harvested at 26 days of age .\",\n \"A significant reduction in the 18S/28S RNA ratio , consistent with the in vitro data using siRNA knockdown was observed at this time point ( Figure 4B ) .\",\n \"Proteomic analysis of cardiac tissue recovered from 26 day old Bud23fl/flMck-Cre+/- mice and littermate controls ( Bud23fl/flMck-Cre-/- ) revealed a significant decrease in protein content per cell with loss of BUD23 ( Figure 4C ) , with clear separation by genotype ( Figure 4—figure supplement 1 ) .\",\n \"When detected proteins were annotated by sub-cellular compartment , mitochondria were found to be subject to the largest loss of protein content , with small reciprocal increases in the other compartments ( Figure 4D ) , again identifying mitochondria as especially sensitive to BUD23 action .\",\n \"Of the 2047 proteins identified by mass spectroscopy , 347 were found to be significantly up-regulated and 442 were found to be significantly down-regulated in the BUD23 deficient condition ( Figure 4E , Figure 4—source data 1 ) .\",\n \"Whilst an approximately equivalent number of down and up regulated proteins was identified globally , when this list was refined to mitochondrial proteins only 20 were up-regulated compared to 220 down-regulated ( 50% of all down-regulated proteins ) ( Figure 4F ) .\",\n \"By contrast , the up-regulated dataset was enriched for translational ( 60S proteins ) and proteasomal proteins ( Figure 4—figure supplement 1 ) .\",\n \"Whilst correlation between change in protein abundance in BUD23-deficient heart tissue and change in mRNA TE of the corresponding transcripts in the BUD23 siRNA treated A549 cells was poor overall ( Figure 4—figure supplement 2 ) , it is striking that mitochondria are affected to a great extent in both models .\",\n \"These observations show a strong selectivity for impaired expression of mitochondrial proteins as a result of BUD23 deficiency .\",\n \"Within the set of mitochondrial proteins , we found that electron transport chain complexes I , IV and V ( ATP synthase ) were most strongly reduced in response to loss of BUD23 ( Figure 4—figure supplement 3 ) .\",\n \"Furthermore , ontology analysis of all the significantly down-regulated proteins identified three main clusters of proteins: mitochondrial proteins , mitochondrial ribosomal proteins and proteins involved in energy metabolism , indicating a functional mechanism linking between the BUD23 dependent ribosomal defect , through reduced mitochondrial protein expression to result in disruption of energy metabolic pathways ( Figure 4G ) .\",\n \"Therefore , BUD23 actions on the ribosome result in down-regulation of core mitochondrial proteins involved in ATP synthesis .\",\n \"Given the observation of preferential dysregulation of mitochondrial proteins , we hypothesised that mitochondrial function was likely to be compromised in BUD23-deficient cardiac tissue .\",\n \"Mitochondrial DNA copy number was reduced in BUD23 deficient cardiac tissue ( Figure 5A ) .\",\n \"Functional analysis of mitochondrial function in cardiac homogenates using the Oroboros microrespirometer system revealed that mitochondrial OXPHOS and ETC were both significantly reduced in mice deficient for BUD23 , while LEAK respiration was unaffected ( Figure 5B ) .\",\n \"The reduction in OXPHOS led to a lower respiratory control ratio ( RCR ) indicating that mitochondria from mice deficient for BUD23 were less efficient at producing ATP .\",\n \"These effects were apparent regardless of which substrate was used for electron donation .\",\n \"In addition to reduced mitochondrial capacity , mice deficient for BUD23 had lower citrate synthase activity , indicating a reduction in cardiac mitochondrial density ( Figure 5C ) , which accords with the measured loss of mitochondrial genome copy number ( Figure 5A ) .\",\n \"To test whether haplo-insufficiency of BUD23 protein was enough to impair mitochondrial function we also tested heart homogenates from Bud23+/- mice .\",\n \"These were found to have no significant difference in mitochondrial respiratory capacity across any mitochondrial state ( LEAK , OXPHOS , ETS ) , whether normalised to citrate synthase activity or protein content ( Figure 5—figure supplement 1 ) .\",\n \"It was , however , noted that , compared to wild-types , Bud23+/- mice did exhibit significantly higher citrate synthase activities , indicating greater mitochondrial density .\",\n \"This may result from increased mitochondrial biogenesis to compensate for defective oxidative phosphorylation function and may partly explain the reduced viability in Bud23+/- mice .\",\n \"Interestingly , the impact of BUD23 loss on mitochondrial function appears to be sexually-dimorphic .\",\n \"When mitochondrial respiration was normalised to citrate synthase , male heterozygous ( HZ ) mice exhibited a deficiency in OXPHOS and ETS relative to wildtype , whereas female mice did not ( Figure 5—figure supplement 1 ) .\",\n \"This observation indicates a more severe phenotype in BUD23 deficient males , with further derangement of individual components of the electron transport chain .\",\n \"Electron microscopy ( EM ) analysis revealed that despite decreased mitochondrial protein abundance and reduced mitochondrial function in the BUD23 cardiac tissue , the individual mitochondria appear to be formed normally with an equivalent number of cristae compared to the wildtype .\",\n \"The intermyofibrillar mitochondria were typically arranged in a highly ordered pattern in the wildtype cardiac tissue , however , there was marked disorganisation in the BUD23-deficient cardiac tissue ( Figure 5D ) .\",\n \"In addition , we identified the presence of numerous spherical electron dense inclusion bodies within the mitochondria of BUD23 deficient cardiac cells ( Figure 5D , yellow arrows ) .\",\n \"Similar electron dense structures have been previously reported in myocardial mitochondria , which may accumulate in response to increased cellular bioenergetic demand ( Jacob et al . , 1994 ) .\",\n \"The observation of these structures here may be consistent with attempted mitochondrial adaptation to energetic-demand challenge .\",\n \"A frequent cellular adaptation to impaired oxidative phosphorylation is a switch to glycolytic ATP generation , and so we measured rate-limiting glycolytic gene expression ( Figure 5E ) .\",\n \"We identified a significant induction of glucokinase ( GCK ) , and a trend to induction of phosphofructokinase ( PKM ) and hexokinase ( HK2 ) supporting such an adaptive switch to glycolysis .\",\n \"A further adaptation predicted was an induction in mitochondrial biogenesis programmes .\",\n \"We identified a significant induction in PGC1a and PGC1b , but no change in TFAM1 expression ( Figure 5E ) .\",\n \"Taken together these results point to some of the expected cellular responses to loss of oxidative phosphorylation , but it was surprising that the changes were so few .\",\n \"This implies that additional cellular injury may be proceeding , which acts to limit successful adaptation .\",\n \"Impaired mitochondrial function may result in overproduction of reactive oxygen species , which can trigger the apoptotic cascade and lead to cell death .\",\n \"Such a progression may explain the distorted and thinned ventricular walls observed in the cardiac Bud23-null mice .\",\n \"We found induction of NRF2 ( Figure 5E ) , a key sentinel gene controlling cellular responses to reactive oxygen species stress , which led us to measure reactive oxygen species directly , and found that these were significantly higher in hearts lacking BUD23 expression ( Figure 5F ) , identifying a burden of increased reactive oxygen species in BUD23 deficient cardiomyocytes resulting from the action of deranged mitochondria .\",\n \"Deficient mitochondrial generation of ATP in cardiac tissue is a key cause of cardiomyopathy .\",\n \"Given the strong mitochondrial deficiencies observed in BUD23-deficient cardiac tissue , we predicted that the cardiomyopathy was the likely cause of sudden death .\",\n \"Histological examination of hearts from Bud23fl/fl Mck-Cre+/- ( BUD23-deficient ) mice showed significant enlargement , ventricle dilatation and decreased ventricle wall thickness compared to littermate controls at 26 days old ( Figure 6A–D ) .\",\n \"Littermate control mice expressing Bud23fl/flMck-Cre-/- or Bud23fl/wt Mck-Cre+/- ( i . e . mice retaining at least one functional copy of the Bud23 allele ) were viable , fertile and did not die prematurely ( Figure 3A ) .\",\n \"Echocardiogram analysis of Bud23fl/fl Mck-Cre+/-mice at 26–28 days old showed that cardiac BUD23-deficiency resulted in systolic dysfunction indicated by significantly dilated left ventricle ( sD ) , as well as significantly decreased fractional shortening and relative wall thickness ( Figure 6E , Figure 6—source data 1 ) .\",\n \") .\",\n \"There was , however , no observed pulmonary oedema or liver oedema ( Figure 6—figure supplement 1 , ) .\",\n \"Further analysis of the heart proteomics data from Figure 4 showed an increase in a number of markers of fibrosis , including Col4a2 , CTGF , Galectin-3 and others ( Figure 6—figure supplement 1 ) .\",\n \"Taken together , these observations indicate that Bud23fl/fl Mck-Cre+/- mice exhibit the early stages of heart failure around 26–28 days of age .\",\n \"Electrocardiogram analysis of the same mice at the same age revealed significantly shorter QRS and corrected QT ( QTc ) intervals but no change in heart rate ( Figure 6F; Figure 6—source data 1 ) .\",\n \") .\"\n ],\n [\n \"BUD23 is tightly conserved through evolution and has recently emerged as a ribosomal RNA methyltransferase , although its physiological role in mammals is unexplored .\",\n \"It is one of the genes deleted in Williams-Beuren Syndrome , a complex multi-gene deletion syndrome with neurological , and energy homeostatic features , although themediating role of individual genes to the phenotype has not been determined ( Morris , 1993 ) .\",\n \"We now identify BUD23 as playing a major role in regulating the translational efficiency of specific transcripts through its role in ribosome maturation .\",\n \"Indeed , translation of low 5’UTR GC-content mRNA species were particularly dependent on BUD23 expression , in contrast to the picture seen with simple ribosome deficiency in which transcripts with short/unstructured 5’UTRs were seen to be most affected ( Khajuria et al . , 2018 ) .\",\n \"Furthermore , we identified a surprising role for BUD23 in maintaining mitochondrial oxidative phosphorylation capacity .\",\n \"As protein translation is tightly coupled to ATP demand this raised the possibility that BUD23-dependent ribosomal maturation is critical to link the two processes in living cells and tissues ( Morita et al . , 2013 ) .\",\n \"Indeed , BUD23 loss greatly impaired mitochondrial ATP generation , both in vitro and in vivo .\",\n \"In mice , cardiomyocyte loss of BUD23 greatly impaired mitochondrial ATP generation leading to dilated cardiomyopathy and premature death , and global loss of BUD23 resulted in embryo-lethality .\",\n \"BUD23 is thought to perform two functions in the generation of the translational apparatus: firstly , in the processing of pre-18S RNA into its mature form , and secondly imparting a m7-G methyl mark on a key residue .\",\n \"As the catalytic activity of BUD23 is not required for efficient pre-18S RNA processing , these two functions appear to be independent of each other ( Haag et al . , 2015; Zorbas et al . , 2015 ) .\",\n \"It is notable that both the BUD23 imparted m7-G methyl mark and structurally complex 5’UTRs are characteristic features of translation specific to Eukaryotes .\",\n \"We speculate the co-evolution of this regulatory 18S methyl mark with the need for more complex regulation of differential translation .\",\n \"Further work should focus on delineating the effect of BUD23-dependent 18S maturation , from the role of the methylation mark .\",\n \"We used proteomics to examine the relative changes in protein abundance which result from targeted loss of BUD23 in cardiomyocytes .\",\n \"This revealed a selective loss of mitochondrial proteins , explaining the observed reduction in mitochondrial capacity .\",\n \"As BUD23 appears to have no direct role within mitochondria we propose that the mitochondrial phenotype results from impaired translation of nuclear-encoded genes with a role in mitochondrial homeostasis .\",\n \"Indeed , we found that the translational efficiency of a number of mitochondria-targeted as well as mitochondrial regulating proteins were BUD23-dependent .\",\n \"There are many potential candidates linking the change of translational efficiency to the mitochondrial phenotype .\",\n \"Mammalian target of rapamycin ( mTOR ) , for example , is known to regulate mitochondrial activity and biogenesis by regulation of translation ( Morita et al . , 2013 ) and , interestingly , BUD23 loss resulted in impaired translation of mTOR itself .\",\n \"We also observed a marked decrease in the translational efficiency of mitochondrially-encoded transcripts .\",\n \"We initially generated and characterised a frame shift-generating null allele mouse .\",\n \"The homozygote null animals were all lost prior to embryonic day e10 . 5 and we were unable to identify a cause but considered that a major defect in a mitochondrially–dependent organ such as heart may be a plausible mechanism .\",\n \"We moved , therefore , to a conditional allele mouse , using which we confirmed the developmental lethality of the null allele by crossing with a global deleter-cre strain ( Hprt-cre ) .\",\n \"Using the conditional allele mice with a muscle Cre driver we were able to profile the impact of BUD23 in a physiological role .\",\n \"Loss of BUD23 in heart and skeletal muscle resulted in early death from cardiac failure .\",\n \"This striking phenotype was accompanied by morphological changes including cardiac dilatation , ventricular wall thinning , and impaired ventricular contractility .\",\n \"There was evidence for an attempted switch to glycolytic ATP generation in the hearts , with induction of rate-limiting glycolytic enzyme genes; and also adaptation to conditions of oxidative stress with induction of NRF2 .\",\n \"Mitochondrial dysfunction frequently imposes a burden of oxidative stress on affected cells , which can result in apoptosis , thereby explaining the loss of cardiac muscle seen .\",\n \"Detailed ultrastructure analysis of cardiac mitochondria revealed no major loss of mitochondrial volume , but there was a striking additional feature seen within the mitochondria: electron-dense , spheroid inclusions .\",\n \"These were not observed in control hearts , and are not typically seen in mitochondria from individuals with mitochondrial myopathies , but have been reported before under conditions of extreme ATP demand , and to result from bilayer budding from the inner cristae , to generate additional ATP generating surface area ( Jacob et al . , 1994; Somlyo et al . , 1974 ) .\",\n \"Therefore , we propose this may be a further adaptation to the impaired oxidative phosphorylation function within the mitochondria .\",\n \"There is a growing recognition that transcriptional control mechanisms are insufficient to explain differentiated cell function .\",\n \"Recent advances have identified differential translation of mRNA species dependent on 5’UTR features as an important control mechanism , potentially coupling global changes in protein translation with appropriate upregulation of mitochondrial ATP generation to meet demand 2 .\",\n \"BUD23 , a highly conserved ribosomal RNA methyltransferase which lies within a multigene interval , deleted as a cause of Williams-Beuren syndrome , plays a specific role in highly ATP-hungry cells , promoting translation of mitochondrial proteins , and impacting on oxidative phosphorylation .\",\n \"This function likely contributes to aspects of the human phenotype , in particular neurological dysfunction , glucose intolerance and obesity .\",\n \"Complete global loss of Bud23 is embryonic lethal , and early post-natal death follows deletion of Bud23 in the heart .\",\n \"The cardiac phenotype is severe , with marked mitochondrial dysfunction , and evidence of compensatory changes towards glycolytic ATP generation , and adaptation to oxidative stress .\",\n \"Therefore , BUD23 plays a critical role in coupling protein translation to mitochondrial function , with implications for mitochondrial diseases , and cardiomyopathies .\"\n ],\n [\n \"All mice were routinely housed in 12:12 light/dark ( L:D ) cycles with ad libitum access to food and water .\",\n \"All experiments were carried out in accordance with the Animals ( Scientific Procedures ) Act 1986 ( UK ) under Home Office protocol number 70/8768 and P3A97F3D1 .\",\n \"In studies utilizing conditional targeted mice , littermate controls ( floxed/floxed ) were used as control; these mice carried no copies of the Cre or iCre .\",\n \"Genotyping was performed on all experimental animals .\",\n \"To assess body composition , whole body , fat mass and lean body mass was assessed using the EchoMRI system ( Echo Medical Systems ) .\",\n \"To assess metabolic gas exchange , mice were individually housed in indirect calorimetry cages ( CLAMS , Columbus Instruments ) .\",\n \"Mice were acclimatised to the cages for 48 hr prior to recording data .\",\n \"Measurements of O2 consumption and CO2 production were recorded every 10 min for >72 hr .\",\n \"RER was derived from these measures ( VCO2/VO2 ) , as was energy expenditure ( 3 . 815∗VO2 + 1 . 232∗VCO2 ) .\",\n \"In order to conditionally KO the Bud23 gene by Cre recombinase expression we generated a transgenic mouse line with the critical exon seven flanked by LoxP sites .\",\n \"Exon seven has an unequal splicing phase , meaning its Cre mediated excision would lead to a frame shift and KO of the Bud23 gene .\",\n \"This exon is also flanked by large introns giving sufficient space to insert LoxP sites without perturbing normal gene splicing and regulation .\",\n \"We used CRISPR ( clustered regularly interspaced short palindromic repeats ) -Cas9 in the mouse zygote to integrate these LoxP sites .\",\n \"Briefly , sgRNA targeting this genomic region , and with minimal off target potential , were identified ( sgRNA-1 GGCATTGGGCCTACTTAAAG-GGG , sgRNA-2 AGTTGAAGGGTTCCATAATG-AGG , sgRNA-3 CTTTACAGCCCAAGACCACT-TGG ) and selected for use .\",\n \"sgRNA was synthesised in vitro , using protocols developed by Bassett et al . ( 2013 ) .\",\n \"Forward primers of the form GAAATTAATACGACTCACTATA GGN18-20GTTTTAGAGCTAGAAATAGC ( where GGN18-20 is the sgRNA sequence ) for each guide were combined with the universal reverse primer , AAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATAACGGACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAAC and used in a PCR reaction with high fidelity polymerase ( Phusion , NEB ) .\",\n \"200 ng of the resulting amplicon were used overnight in an in vitro transcription reaction ( HiScribe , NEB , as per manufacturer’s instructions ) , before purification ( Megaclear , Ambion ) and quantification by nanodrop .\",\n \"A double stranded DNA ( dsDNA ) repair template was designed to\",\n \"( i ) incorporate loxP sites , with unique restriction sites for screening purposes , 300 bp upstream and 200 bp downstream of exon seven respectively and\",\n \"( ii ) integrate silent shield mutations for all three sgRNA to prevent cutting of the repair template and\",\n \"( iii ) contain 1000 bp 5’ and 3’ homology arms to the target region .\",\n \"The above template was synthesised in a pUC57 vector ( Genscript , US ) , the linear 5’homology-loxP-exon7-loxP-3’homology fragment excised from the vector by restriction digest , gel extracted ( Bioline ) and further cleaned prior to microinjection ( Monarch PCR purification kit , NEB ) .\",\n \"An injection mix of the three sgRNA ( 40 ng/ul each ) , Cas9 mRNA ( 100 ng/ul ) and the dsDNA repair template ( 10 ng/ul ) was prepared and directly microinjected into B6D2F1 ( Envigo ) zygote pronuclei using standard protocols .\",\n \"Zygotes were cultured overnight and the resulting 2 cell embryos surgically implanted into the oviduct of day 0 . 5 post-coitum pseudopregnant mice .\",\n \"Potential founder mice were screened by PCR and digest .\",\n \"Animals positive for both 5’ and 3’ LoxP integration had the region fully verified by pCR-Blunt cloning followed by Sanger sequencing .\",\n \"two founder mice were identified and then back-crossed to C57BL/6J wild-type mice to assess germline penetrance .\",\n \"Additionally , the described CRISPR targeting strategy resulted in deletion of critical exon 7 , a global Bud23 KO line was generated from this founder animal .\",\n \"DNA was extracted from ear snip or tail tip using REDExtract-N-Amp tissue PCR kit ( Sigma ) .\",\n \"Primer sequences and PCR reaction conditions are listed in the key resources table and the primer sequences supplementary file .\",\n \"PCR products were resolved on 1 . 5% agarose gels .\",\n \"STR authenticated A549 cells were obtained from the European Collection of Cell Cultures ( ECACC Cat no: 86012804 ) and maintained in culture , incubated at 37°C in humidified air , 5% v/v CO2 .\",\n \"Cultured cells were tested regularly for mycoplasma contamination .\",\n \"A549 cells were cultured in DMEM growth medium ( Sigma , D6546 ) , supplemented with 2 mM L-Glutamine ( Sigma , G7513 ) , 10% foetal calf serum ( FCS ) ( Sigma , F96665 ) .\",\n \"A549 cells were plated at 1 × 10^6 cells in a 10 cm cell culture dish and transfected with BUD23 specific or non-targeting control siRNA ( S41529 , S41530 , S41531 , negative control 1 , negative control 2 , Ambion Silencer Select ) using Dharmafect DF-1 according to the manufacturer’s guidelines .\",\n \"A549 cell pellets were lysed in M-PER mammalian protein extraction reagent ( Thermo Scientific ) .\",\n \"Whole mouse hearts were dissected , quartered and washed in ice cold PBS to remove blood before disruption of the tissue in 1x TBS supplemented with protease inhibitor cocktail using a TissueRuptor ( Qiagen ) .\",\n \"SDS was added to a final concentration of 4% ( w/v ) prior to sonication .\",\n \"Samples were then reduced with 100 mM DTT and heated for 3 min at 95°C .\",\n \"Samples were prepared for label-free quantification in accordance with published protocols ( Hernandez-Valladares et al . , 2016 ) .\",\n \"Digested samples were analysed by LC-MS/MS using an UltiMate 3000 Rapid Separation LC ( RSLC , Dionex Corporation , Sunnyvale , CA ) coupled to a Q Exactive HF ( Thermo Fisher Scientific , Waltham , MA ) mass spectrometer using a stepped normalized collision energy centered at 28 .\",\n \"A549 peptide mixtures were separated using a multistep gradient from 95% A ( 0 . 1% FA in water ) and 5% B ( 0 . 1% FA in acetonitrile ) to 7% B at 1 min , 18% B at 58 min , 27% B in 72 min and 60% B at 74 min at 300 nL min−1 , using a 75 mm x 250 μm i . d . 1 . 7 μM CSH M-Class C18 , analytical column ( Waters ) , for a final run time of 90 min .\",\n \"Cardiac samples were separated using a gradient from 5% B to 7% B at 1 min , 18% B at 81 . 5 min , 27% B at 102 min and 60% B at 104 min , with a final run time of 120 min .\",\n \"The top 12 precursors were selected for fragmentation automatically by data dependant analysis during each cycle .\",\n \"Ionization potential was set at 1900V with a survey scan window of 300–1750 m/z .\",\n \"MS1 used a resolution of 120 , 000 with a target ion intensity of 3e6 .\",\n \"Higher-energy collisional dissociation was used .\",\n \"MS2 was set at a resolution of 60 , 000 with a target ion intensity of 2e5 .\",\n \"Maximum injection times for MS1 and MS2 were 20 ms and 110 ms respectively .\",\n \"Peptides were dynamically excluded for 15 s after one occurrence .\",\n \"Mass spectra were analysed using MaxQuant version 1 . 6 . 0 . 16 ( Cox and Mann , 2008 ) , searching against either the Uniprot Mus musculus database ( accessed 21/12/2017 ) or Homo sapiens database ( accessed 12/03/2017 ) using the native Andromeda search engine .\",\n \"Match between runs was selected with a retention time of 1 min , using default parameters for all other settings .\",\n \"MS1 search tolerance was set at 20ppm , and MS2 at 4 . 5ppm .\",\n \"Modifications were set as Carbamidomethyl ( C ) for fixed , and Oxidation ( M ) and Acetyl ( Protein N-term ) variable .\",\n \"A minimum of 1 spectra was required for identification , while a minimum of 2 spectra were required for quantification .\",\n \"Resulting data was processed using Perseus version 1 . 6 . 0 . 7 ( Tyanova et al . , 2016 ) and R statistical software ( Wickham , 2009; R Development Core Team , 2017 ) .\",\n \"All proteins identified by site only , as a potential contaminant , or in the decoy reverse database were excluded .\",\n \"Data was then filtered to exclude proteins with more than 3 ‘zero-values’ across all groups .\",\n \"All proteins assigned to the GO term ‘blood microparticle’ were removed as well .\",\n \"LFQ values were log ( 2 ) transformed and missing values imputated in Perseus using a normal distribution ( Lazar et al . , 2016 ) .\",\n \"The width of the new distribution was 0 . 3 standard deviations of the data , and was shifted downwards by 1 . 8 standard deviations of the data .\",\n \"A total of 6 . 68% of values were imputated .\",\n \"Significance was determined using a T-test with a permutation based FDR ( FDR < 0 . 05 , s0 = 0 . 1 ) .\",\n \"s0 represents the relative importance of the difference between the means , with non-zero s0 values taking fold change , and not only p-value , into account .\",\n \"Estimations of copy numbers were performed using the proteomic ruler plug-in version 0 . 1 . 6 ( Wiśniewski et al . , 2014 ) .\",\n \"Proteins were then mapped to subcellular localisations to determine percentage protein mass by organelle according to published protocols using the HeLa spatial proteome database ( Doll et al . , 2017; Itzhak et al . , 2016 ) .\",\n \"Interaction networks were built using String 10 online software ( Szklarczyk et al . , 2017 ) using seven data sources ( textmining , experiments , databases , co-expression , neighborhood , gene fusion co-occurrence ) and imported into Cytoscape ( Shannon et al . , 2003 ) .\",\n \"Network edges represent confidence .\",\n \"A minimum interaction score of 0 . 4 ( medium confidence ) was required for network inclusion .\",\n \"Enrichment analysis was performed using PANTHER ( Mi et al . , 2017 ) .\",\n \"Lists of proteins were exported for the identified GO terms and then manually coloured using Cytoscape for visualization .\",\n \"Transthoracic echocardiography ( TTE ) was performed using a Vevo 770 High-Resolution Imaging System ( Visualsonics Inc , Toronto , Canada ) and a 30MHz probe .\",\n \"Mice were lightly anesthetized with 1–1 . 5% isoflurane via facemask , maintaining the heart rate at approximately 450 beats/min .\",\n \"Views were taken in planes that approximated the parasternal short-axis view and the apical long-axis view .\",\n \"M-mode tracings were used to determine left ventricular end-diastolic diameter ( LVDD ) and end-systolic diameter ( LVSD ) , posterior wall thickness in diastole ( LVPWD ) and systole ( LVPWS ) , and interventricular septum thickness in diastole ( IVSD ) and systole ( IVSS ) over three cardiac cycles .\",\n \"The analysis was performed blinded to animal details .\",\n \"LV fractional shortening ( FS ) was calculated using the formula FS = [ ( LVDD - LVSD ) / ( LVDD ) ] x 100 .\",\n \"Relative wall thickness ( RWT ) was calculated using the equation RWT = ( IVDD + LVPWD ) /LVDD .\",\n \"Mice were lightly anesthetized with 1–1 . 5% isoflurane via facemask .\",\n \"The body temperature was maintained around 37°C using a heat pad .\",\n \"The lead II ECG was recorded using Power Lab/4SP system ( Adinstruments ) from needle electrodes inserted subcutaneously into the left and right forelimbs and the right hindlimb .\",\n \"The signal was acquired for about 5 min using Chart 7 software ( Adinstruments ) .\",\n \"During offline analysis , the 5 min recording was examined for unusually shaped P , QRS , or ‘T’ waves and for time-varying phenomenon ( e . g . irregularities in interval durations ) .\",\n \"Ectopic or abnormal beats are noted .\",\n \"A representative 10–15 s segment of the recording was averaged to obtain the signal averaged ECG to calculate Heart rate , RR interval , P wave duration , PR interval , QRS , JT and QT durations .\",\n \"QT duration was corrected ( QTc ) using the Bazett's formula ( Bazett , 1997 ) .\",\n \"Polysome profiles were prepared according to the protocols previously published in: ( Gandin et al . , 2014 ) .\",\n \"Two conditions were profiled , BUD23 siRNA and Control siRNA , with an n of 3 samples per condition .\",\n \"Fractions were collected from each sample and RNA was purified using the Trizol extraction method .\",\n \"RNA from fractions containing three or less ribosomes were pooled ( ‘Low translational efficiency ) and RNA from fractions containing more than three ribosomes ( ‘high efficiency’ ) were pooled .\",\n \"Sub-polysomal fractions were similarly pooled .\",\n \"RNA sequencing was used to quantify the expression of RNA species in each of the pooled samples .\",\n \"Translational efficiency ( TE ) was defined as the relative abundance of a transcript in the high and low efficiency pooled fractions .\",\n \"A TE score for each transcript was calculated by averaging the ratio derived from each of the three replicates in each condition .\",\n \"TE was plotted for the two conditions using R software .\",\n \"Translational efficiency ( TE ) was then compared to both the length and GC content of the 5' UTR region .\",\n \"5' UTRs were defined using the TxDb . Hsapiens . UCSC . hg38 . knownGene ( Bioconductor Core Team and Bioconductor Package Maintainer , 2016 ) and org . Hs . eg . db ( Carlson , 2018 ) R packages .\",\n \"Sequences of the regions were extracted using the twoBitToFa program from the BLAT suite ( Kent , 2002 ) and then GC content was calculated using the seqinr R package ( Charif and Lobry , 2007 ) .\",\n \"The Pearson correlation coefficients , and their relative significances , were calculated using cor . test function from the stats R package ( R Development Core Team , 2017 ) .\",\n \"Live cell metabolic assays were performed on A549 cells pre-treated with either BUD23 siRNA or control siRNA using a Seahorse XFe 96 analyser ( Agilent ) .\",\n \"Mitochondrial stress tests was performed according to the manufacturer’s instructions ( Agilent ) .\",\n \"2M Oligomycin ( OA ) , 1M FCCP and 0 . 5M rotenone/antimycin A ( AA+R ) were used for all conditions .\",\n \"A549 cells were plated into Seahorse XF96 plates at 160 , 000 cells per well , utilising 16 wells per condition .\",\n \"Cell density was normalised using SRB assay .\",\n \"Experiments were performed in triplicate .\",\n \"Of the three distinct experimental preparations available ( isolated mitochondria , permeabilized fibers , and tissue homogenates ) , we utilized cardiac homogenates for mitochondrial assessment ( see Goo et al . , 2013 for advantages of this method ) .\",\n \"Ventricular tissue ( ~20 mg ) was weighed and transferred to 1 . 5 ml of ice-cold MiRO5 medium ( in mM: EGTA 0 . 5 , MgCl2 1 . 4 , taurine 20 , KH2P04 10 , HEPES 20 , BSA 1% , K-MES 60 mM , sucrose 110 mM , pH 7 . 1 , adjusted with 5 N KOH ) .\",\n \"Tissue was homogenized for 3 s in 1 s bursts with a tissue homogenizer and loaded immediately ( 40 µg/ml ) into an Oroboros Oxygraph 2 k high resolution respirometry system ( Oroboros Instruments , Innsbruck , Austria ) for measurement of mitochondrial respiration combined with the Fluorescence-Sensor Green of the O2k-Fluo LED2-Module for H2O2 measurement ( see Makrecka-Kuka et al . , 2015 for full details ) .\",\n \"Two identical respiration chambers ( chamber A and chamber B ) held at the same temperature were run in parallel for each experimental run .\",\n \"All measurements of respiration rates were carried out at 37°C .\",\n \"Oxygen electrodes were calibrated daily with air-saturated respiration solution Zero calibrations were achieved by injecting yeast into the experimental chambers .\",\n \"Oxygen solubility in the assay medium was calculated as described previously ( Lienig and Forstner , 1984 ) .\",\n \"H2O2 flux was measured simultaneously with respiration in the O2k-Fluorometer using the H2O2-sensitive probe Amplex UltraRed ( 10 μM ) with 1 U/mL horse radish peroxidase ( HRP ) and 5 U/mL superoxide dismutase ( SOD ) .\",\n \"AmR in the presence of H2O2 is catalysed by HRP to the fluorescent product resorufin .\",\n \"Calibrations were performed with two sequential injections of H2O2 at 0 . 1 μM steps .\",\n \"Three parameters are commonly used to assess mitochondrial function ( for reviews , see Brand and Nicholls , 2011; Pesta and Gnaiger , 2012 ) .\",\n \"Firstly , the capacity for oxidative phosphorylation ( OXPHOS ) is the respiratory capacity of mitochondria in the ADP-activated state of oxidative phosphorylation ( saturating concentrations of ADP , inorganic phosphate , oxygen , and defined substrates ) .\",\n \"Secondly , LEAK respiration rate represents mitochondrial respiration that occurs in the absence of ATP generation , mainly to compensate for proton leak across the mitochondrial inner membrane .\",\n \"Lastly , the respiratory electron transfer-pathway capacity ( ETC ) is mitochondrial respiration in the noncoupled state in the presence an uncoupler; this induces maximum oxygen flux through the electron transport chain .\",\n \"The Respiratory Control Ratio ( RCR , calculated here as OXPHOS/LEAK ) provides a measure of the degree of coupling between oxidation and phosphorylation , or in other words , the efficiency of mitochondrial ATP production .\",\n \"OXPHOS , LEAK and ETS were measured in the presence of Complex I substrates pyruvate and malate ( electron transfer through Complexes I-IV ) or Complex I+II substrates ( addition of succinate ) .\",\n \"Additionally , respiratory flux with electron transfer through Complex IV alone was measured via the addition of the electron donor tetramethyl-phenylene-diamine ( TMPD ) .\",\n \"The protocol used to measure these parameters was adapted from Pesta and Gnaiger ( 2012 ) .\",\n \"Briefly , pyruvate ( 5 mmol l −1 ) , malate ( 0 . 25 mmol l−1 ) and glutamate ( 10 mmol l−1 ) are added as carbon substrates and to spark the citric acid cycle .\",\n \"Under these conditions , mitochondria are in LEAK respiration with CI substrates in the absence of adenylates .\",\n \"OXPHOS with CI substrates was achieved through addition of saturating levels of ADP ( 2 mM l−1 ) .\",\n \"Following steady-state conditions , succinate ( 10 mmol l−1 ) was added to achieve OXPHOS with CI+CII substrates .\",\n \"To uncouple respiration and achieve ETS with CI+CII substrates , carbonyl cyanide 4-trifluoromethoxyphenylhydrazone ( FCCP ) was carefully titrated to a maximum concentration of 0 . 25 μmol l−1 .\",\n \"Rotenone was then added to achieve ETS with CII substrates and antimycin A ( 5 μmol l−1 ) was given to block Complex III and measure background non-mitochondrial residual oxygen consumption ( ROX ) .\",\n \"OXPHOS through Complex IV alone was assessed by adding the electron donor TMPD ( 0 . 5 mmol l−1 ) .\",\n \"To avoid oxidation of TMPD , ascorbate ( 2 mMol l−1 ) was added prior to TMPD injection .\",\n \"The citrate synthase activity of cardiac homogenates were measured in frozen homogenates .\",\n \"Briefly , maximal activity ( Vmax ) was determined with a spectrophotometer at 37°C with a Synergy HTX spectrophotometer ( BioTek , UK ) in assay buffer; 50 mM TRIS-HCl , pH 8 . 0 .\",\n \"Citrate synthase activity was monitored in the presence or absence of oxaloacetate by the appearance of 5-thio-2-nitrobenzoic acid as a result of the reaction of free acetyl-CoA with 5 , 5’-dithiobis ( 2-nitrobenzoic acid ) at 412 nm over a 10 min incubation period ( assay buffer with 0 . 5 oxaloacetate , 0 . 3 mM acetyl-CoA , 0 . 15 mM 5 , 5-dithiobis-2-nitrobenzoic acid ) .\",\n \"Extinction coefficients were empirically determined to quantify Vmax values .\",\n \"Enzyme activities were normalised to total soluble protein , which was quantified according to Bradford ( Bradford , 1976 ) .\",\n \"Total cell protein was isolated from cells using protein extraction buffer ( 50 mM Tris , 150 mM NaCl , 1% TritonX-100 , supplemented with protease inhibitors ) .\",\n \"Total protein was isolated from snap-frozen tissue using RIPA buffer supplemented with protease inhibitor cocktail after disruption using Lysing Matrix D tubes ( MP Bio ) .\",\n \"All experiments were performed according to current UK Home Office regulations and under approval of the relevant University of Manchester ( Manchester , UK ) local ethics committee .\",\n \"Six mice were used for this study ( n = 3 control , n = 3 Bud 23 KO ) .\",\n \"Immediately after euthanasia , hearts were harvested and small portions ( cubes with sides smaller than 0 . 5 mm ) of right and left ventricle were immersion fixed in 2 . 5% glutaraldehyde and 2% paraformaldehyde in 100 mM sodium cacodylate buffer ( pH 7 . 2 ) overnight .\",\n \"Specimen preparation was performed , with small modifications , according to the Ellisman protocol ( Holcomb et al . , 2013 ) .\",\n \"Briefly , after washings in sodium cacodylate , samples were sequentially stained in: 2% osmium tetroxide and 1 . 5% potassium ferrocyanide in 100 mM sodium cacodylate for 1 hr; 1% aqueous thiocarbohydrazide for 20 min; 2% aqueous osmium for 30 min; 1% aqueous uranyl acetate overnight and Walton’s lead aspartate for 30 min the following morning at 60°C .\",\n \"Samples were washed 3 times for 10 min in double distilled water after each staining step .\",\n \"Staining was followed by dehydration in ethanol ascending series ( 50% , 70% , 90% , 100% ) and infiltration with TAAB 812 hard resin mixed with propylene oxide .\",\n \"After embedding , resin blocks were cured at 70° C for 40 hr .\",\n \"Plastic blocks were sectioned at 80 nm thickness using a Leica UC6 ultramicrotome and sections were imaged using a FEI Tecnai12 Biotwin operated at 100kV .\",\n \"Images were analysed using Fiji ( Schindelin et al . , 2012 ) .\",\n \"Fractionation experiments were performed by differential centrifugation as previously described previously ( Rorbach et al . , 2014 ) .\",\n \"Unless otherwise stated in the figure legend , parametric data was analysed by ANOVA with a Dunnett’s multiple comparisons test .\",\n \"Non-parametric data was analysed using a Mann-Whitney test .\",\n \"RNAseq data was analysed using edgeR ( Robinson et al . , 2010 ) .\",\n \"For Mass Spectrometry data analysis please refer to the dedicated section .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Efficient mitochondrial function is required in tissues with high energy demand such as the heart , and mitochondrial dysfunction is associated with cardiovascular disease .","Expression of mitochondrial proteins is tightly regulated in response to internal and external stimuli .","Here we identify a novel mechanism regulating mitochondrial content and function , through BUD23-dependent ribosome generation .","BUD23 was required for ribosome maturation , normal 18S/28S stoichiometry and modulated the translation of mitochondrial transcripts in human A549 cells .","Deletion of Bud23 in murine cardiomyocytes reduced mitochondrial content and function , leading to severe cardiomyopathy and death .","We discovered that BUD23 selectively promotes ribosomal interaction with low GC-content 5’UTRs .","Taken together we identify a critical role for BUD23 in bioenergetics gene expression , by promoting efficient translation of mRNA transcripts with low 5’UTR GC content .","BUD23 emerges as essential to mouse development , and to postnatal cardiac function ."],"string":"[\n \"Efficient mitochondrial function is required in tissues with high energy demand such as the heart , and mitochondrial dysfunction is associated with cardiovascular disease .\",\n \"Expression of mitochondrial proteins is tightly regulated in response to internal and external stimuli .\",\n \"Here we identify a novel mechanism regulating mitochondrial content and function , through BUD23-dependent ribosome generation .\",\n \"BUD23 was required for ribosome maturation , normal 18S/28S stoichiometry and modulated the translation of mitochondrial transcripts in human A549 cells .\",\n \"Deletion of Bud23 in murine cardiomyocytes reduced mitochondrial content and function , leading to severe cardiomyopathy and death .\",\n \"We discovered that BUD23 selectively promotes ribosomal interaction with low GC-content 5’UTRs .\",\n \"Taken together we identify a critical role for BUD23 in bioenergetics gene expression , by promoting efficient translation of mRNA transcripts with low 5’UTR GC content .\",\n \"BUD23 emerges as essential to mouse development , and to postnatal cardiac function .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Cells need to make proteins to survive , so they have protein-making machines called ribosomes .","Ribosomes are themselves made out of proteins and RNA ( a molecule similar to DNA ) , and they are assembled by other proteins that bring ribosomal components together and modify them until the ribosomes are functional .","Mitochondria are compartments in the cell that are in charge of providing it with energy .","To do this they require several proteins produced by the ribosomes .","If not enough mitochondrial proteins are made , mitochondria cannot provide enough energy for the cell to survive .","One of the proteins involved in modifying ribosomes so they are functional is called BUD23 .","People with certain diseases , such as Williams-Beuren syndrome , do not make enough BUD23; but it was unknown what specific effects resulted from a loss of BUD23 .","To answer this question , Baxter et al . first genetically removed BUD23 from human cells , and then checked what happened to protein production .","They found that ribosomes in human cells with no BUD23 were different than in normal cells , and that cells without BUD23 produced different proteins , which did not always perform their roles correctly .","Proteins in the mitochondria are one of the main groups affected by the absence of BUD23 .","To determine what effects these modified mitochondrial proteins would have in an animal , Baxter et al . genetically modified mice so that they no longer produced BUD23 .","These mice developed heart problems caused by their mitochondria not working correctly and being unable to provide the energy the heart cells needed , eventually leading to heart failure .","Heart problems are common in people with Williams-Beuren syndrome .","Many diseases arise when a person’s mitochondria do not work properly , but it is often unclear why .","These experiments suggest that low levels of BUD23 or faulty ribosomes may be causing mitochondria to work poorly in some of these diseases , which could lead to the development of new therapies ."],"string":"[\n \"Cells need to make proteins to survive , so they have protein-making machines called ribosomes .\",\n \"Ribosomes are themselves made out of proteins and RNA ( a molecule similar to DNA ) , and they are assembled by other proteins that bring ribosomal components together and modify them until the ribosomes are functional .\",\n \"Mitochondria are compartments in the cell that are in charge of providing it with energy .\",\n \"To do this they require several proteins produced by the ribosomes .\",\n \"If not enough mitochondrial proteins are made , mitochondria cannot provide enough energy for the cell to survive .\",\n \"One of the proteins involved in modifying ribosomes so they are functional is called BUD23 .\",\n \"People with certain diseases , such as Williams-Beuren syndrome , do not make enough BUD23; but it was unknown what specific effects resulted from a loss of BUD23 .\",\n \"To answer this question , Baxter et al . first genetically removed BUD23 from human cells , and then checked what happened to protein production .\",\n \"They found that ribosomes in human cells with no BUD23 were different than in normal cells , and that cells without BUD23 produced different proteins , which did not always perform their roles correctly .\",\n \"Proteins in the mitochondria are one of the main groups affected by the absence of BUD23 .\",\n \"To determine what effects these modified mitochondrial proteins would have in an animal , Baxter et al . genetically modified mice so that they no longer produced BUD23 .\",\n \"These mice developed heart problems caused by their mitochondria not working correctly and being unable to provide the energy the heart cells needed , eventually leading to heart failure .\",\n \"Heart problems are common in people with Williams-Beuren syndrome .\",\n \"Many diseases arise when a person’s mitochondria do not work properly , but it is often unclear why .\",\n \"These experiments suggest that low levels of BUD23 or faulty ribosomes may be causing mitochondria to work poorly in some of these diseases , which could lead to the development of new therapies .\"\n]","isConversational":false},"year":{"kind":"string","value":"2020"}}},{"rowIdx":87,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["ecology","neuroscience"],"string":"[\n \"ecology\",\n \"neuroscience\"\n]","isConversational":false},"title":{"kind":"string","value":"Burst muscle performance predicts the speed, acceleration, and turning performance of Anna’s hummingbirds"},"id":{"kind":"string","value":"elife-11159-v2"},"sections":{"kind":"list like","value":[["The ability of an animal to change the speed and direction of movement , defined as maneuverability ( Dudley , 2002 ) , can determine its success at avoiding predators , obtaining food , and performing other behaviors that determine the margin between life and death ( Webb , 1976; Hedenström and Rosén , 2001; Walker et al . , 2005 ) .","Most biomechanical research on birds has focused on either brief ( e . g . , take off ) or steady state movements ( e . g . , forward flight ) that can be studied most readily in the laboratory .","Maneuverability is therefore one of the most important but least understood aspects of animal locomotion .","Warrick and coworkers ( Warrick et al . , 1988; Warrick and Dial , 1998 ) proposed that there are both intrinsic and facultative influences on maneuvering performance .","For animals that perform powered flight , intrinsic maneuverability is defined by the physical limitations imposed by morphology ( Norberg and Rayner , 1987 ) , but excess muscle capacity should allow them to facultatively overcome the costs of suboptimal morphology , achieving higher levels of performance by sacrificing efficiency .","Although compelling , this hypothesis has never been tested explicitly .","Wing morphology is defined using measures of size ( e . g . , area or length ) and non-dimensional measures of shape ( e . g . , aspect ratio ) .","Wing area and aspect ratio have significant and well known effects on the aerodynamics of flight in animals ( Pennycuick , 1975; Kruyt et al . , 2014; 2015 ) , and should affect maneuvering performance .","Wing morphology influences flight efficiency ( Feinsinger and Chaplin , 1975 ) , ecological roles ( Feinsinger , 1976; Feinsinger and Colwell , 1978; Warrick , 1998 ) and competitive ability ( Feinsinger and Chaplin , 1975; Feinsinger and Colwell , 1978; Feinsinger et al . , 1979; Altshuler , 2006 ) .","Because these previous studies focused on species and gender comparisons , less is known about how individual variation in wing morphology influences performance , especially with respect to maneuverability .","One complication is that different wing sizes and shapes can be favored depending on the specific maneuver performed , e . g . , yaw versus banked turns .","Given the diversity of flight behaviors , it is unclear if the requirements for maneuvering exert strong selection on wing morphology .","Muscle capacity affects the maximum aerodynamic force a flying animal can produce .","Aerodynamic force can be directed for performing maneuvers that require greater output than the minimum requirements for flight .","Excess muscle capacity can also be used to compensate for anatomical or spatial constraints on wing movement ( Warrick , 1998 ) .","Muscle output of hummingbirds has been quantified in several ways including oxygen consumption to determine metabolic input , wingbeat kinematics to estimate mechanical power output , and electromyography ( EMG ) to measure myoelectric input .","Considering hovering flight as the point of comparison , forward flight at the fastest speeds recorded in a wind tunnel requires about 20% more metabolic ( Clark and Dudley , 2010 ) and myoelectric input ( Tobalske et al . , 2010 ) .","Maximum sustained hovering performance has been studied by experimentally lowering air density to the lowest values in which birds are still able to hover .","These experiments revealed that hovering in hypodense air requires ~40% higher mechanical power output ( Chai and Dudley , 1995 ) and ~60% higher spatial recruitment of muscle fibers , as measured by the spike amplitude of the electromyogram recordings ( Altshuler et al . , 2010b ) , in comparison to hovering in normal air .","By far the most expensive flight behavior studied to date in hummingbirds is maximum load lifting , which requires 200–400% more mechanical power output ( Chai et al . , 1997; Chai and Millard , 1997; Altshuler et al . , 2010a ) , about 200% more spatial recruitment ( EMG spike amplitude ) , and 150% more temporal recruitment ( EMG spike frequency ) ( Altshuler et al . , 2010b ) compared to hovering .","Maximum load lifting is a transient behavior that uses the bird’s natural escape response to measure burst power output .","Thus , it is not surprising that this assay provides the maximum muscle capacity that has been measured in hummingbirds .","It is particularly useful for quantifying variation among and within species in burst muscle capacity .","Studies using the load lifting assay have revealed that maximum burst muscle capacity is related to hummingbird evolutionary ecology .","Altshuler and coworkers ( Altshuler et al . , 2004b; Altshuler , 2006 ) demonstrated that ecological role is more strongly related to load lifting ability than morphological parameters such as wing loading .","Load lifting ability is also associated with species- and gender-specific competitive ability at different elevations .","Altshuler ( Altshuler , 2006 ) suggested that the relationship between maximum muscle capacity and competitive ability may be mediated through maneuvering performance .","Unconstrained maneuvering performance of birds , including hummingbirds , has recently been quantified in the field without individual identification ( Shelton et al . , 2014; Sholtis et al . , 2015 ) .","Although field studies are valuable for quantifying average species performance , individual identification and large sample sizes are required to examine sources of within-species variation .","Here , we studied the free-flight maneuvering performance of Anna's hummingbirds ( Calypte anna ) in a large flight chamber ( Video 1 ) .","Flight maneuvers in a chamber are not expected to be the same as outdoors , and may have lower velocities and accelerations .","The benefit of this approach is that a large number of measurements from the same individuals can be combined with other data to examine how variation in the observed maneuvers is influenced by individual morphology and muscle capacity . 10 . 7554/eLife . 11159 . 003Video 1 . The multi-camera , automated tracking system filming two hummingbirds in the flight arena at 200 frames per second . Continuously tracked sequences are assigned an object number ( from 0 to 4 over this sequence ) .","Body position and orientation are calculated and reprojected onto the video of four cameras .","The videos are saved using a compression algorithm that only records the sections of the image that are moving ( Straw et al . 2011 ) .","Thus , birds disappear from the video when they land and stop moving .","The trajectory shown in Figure 1 is taken from the bird labeled #2 and begins at 5 . 1 seconds and ends at 8 . 05 seconds . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 003 We used a high-throughput computational approach to record the flight performance of 20 individuals alone and in the presence of a competitor .","Flight trajectories were parsed into a set of performance metrics based on body position and orientation .","The first goal of our study was to determine if voluntary maneuvering performance is repeatable within individuals .","Repeatability of maneuvering performance can arise either through a strong influence of fixed traits such as morphology and anatomy , or through other consistent influences , such as motivation .","We expect that repeatable measurements will be most useful for our second goal , determining how variation in maneuverability among individuals is influenced by natural variation in morphology and muscle capacity .","This also required measuring morphological traits and maximum burst performance for each individual .","Our third goal was to determine how motivation state induced by the presence of a competitor influenced maneuvering performance .","To address this question we compared flight trials with and without competitors ."],["The first stage of analysis was estimating instantaneous velocities , accelerations , and headings from the raw tracking data ( Figure 1—figure supplement 1 ) .","Translational velocity and acceleration were calculated by taking the first and second derivatives of an interpolation spline fit to the body position data ( splev and splrep functions , Scientific Python ) .","The velocities and accelerations were split into vertical and horizontal components .","The body orientation vector was represented in spherical coordinates as azimuth and pitch angles .","We took the first derivatives to obtain azimuth and pitch velocities .","Because the video tracking system did not allow a measurement of body roll , we decided to use a global coordinate system instead of a body axis-centered coordinate system .","In our frame of reference , pitch is a global measure defined relative to the horizontal plane .","Heading was calculated as the instantaneous direction of the horizontal translation velocity , and the heading velocity was calculated as the derivative of heading . 10 . 7554/eLife . 11159 . 004Figure 1 . A multi-camera , automated tracking system extracted hummingbird body position ( blue circle ) and orientation ( red line ) from solo and competitive flights . The trajectory shown for one bird","( a ) is also shown in Video 1 ( see Figure 1—figure supplement 1 for time series of position , velocity , and acceleration values ) .","Stereotyped maneuvers were classified in each trajectory","( b ) and between one and five performance metrics were calculated from each maneuver .","Maneuvers within a trajectory may be overlapping ( e . g . #4 , 5 , 6 ) .","The trajectory presented in b is a top down ( x-y projection ) view of the trajectory shown in a .","Body position and orientation were smoothed with an extended Kalman filter ( c , d ) .","The effects of four different sets of smoothing parameters are presented for an arcing turn ( maneuver #9 in","b ) and an upward acceleration ( maneuver #1 in","b ) .","Shown here are the unsmoothed position and orientation ( black trace and text ) , the chosen levels of smoothing ( blue ) , a lower level of smoothing ( green; 0 . 1 x Rpos; 0 . 1 x Rori ) , and a higher level of smoothing ( red; 10 x Rpos; 10 x Rori ) .","The chosen smoothing parameters for body position were determined by tracking multiple dropped objects and calibrating the Z-axis acceleration to gravity .","The chosen smoothing parameters for body orientation were determined by re-projecting the body axis vector onto the video .","The higher and lower levels of smoothing for body position presented in this figure were both deemed too extreme , when re-projected onto the video .","However , the level of smoothing for body orientation had minimal effect on the average yaw velocity . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 00410 . 7554/eLife . 11159 . 005Figure 1—figure supplement 1 . The representative trajectory from Figure 1 and Video 1 displayed through time . The upper three panels provide the position , velocity , and acceleration values .","The lower three panels provide the orientation vector , orientation angle and rotation velocity values .","The maneuvers extracted for this sequence are given by thick black lines below the traces . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 005 We then used the velocity , acceleration , and orientation data to search for a series of ten stereotyped maneuvers that were independent of time and distance scales ( Figure 1b ) .","Five of the maneuvers were sequences defined by changes in translational velocity: 1 ) 3D accelerations , 2 ) horizontal accelerations , 3 ) horizontal decelerations , 4 ) vertical upward accelerations , and 5 ) vertical downward accelerations .","Three maneuvers were sequences defined by changes in rotation: 6 ) pitch-up rotations , 7 ) pitch-down rotations , and 8 ) yaw turns .","Two of the maneuvers were defined as turns with translational components: 9 ) arcing turns and 10 ) pitch-roll turns .","These ten maneuvers are not meant to be mutually exclusive , exhaustive , or to divide the entire filming session into a set of discrete behaviors , but are instead intended to extract simple measurements that can be used as an assay for maneuvering performance .","The search criteria for the maneuvers are given in Table 1 . Because we assume that a new maneuver must involve a change in velocity , the first search parameter was to find sequences bounded by velocity maxima and minima , or vice versa .","We next describe the additional search parameters and the performance metrics used to quantify each maneuver . 10 . 7554/eLife . 11159 . 006Table 1 . Search parameters for the ten maneuvers analyzed in the study .","The definitions , units , and symbols for the 14 related performance metrics are also provided . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 006ManeuverSearch parametersPerformance metricUnitsSymbol3D accelerationStart: velocity xyz minimumEnd: velocity xyz maximumDistance xyz > 25 cmMaximum velocitym/sVelmaxHorizontal accelerationStart: velocity xy minimumEnd: velocity xy maximumDistance xy > 25 cmDistance z < 10 cmMaximum acceleration xym/s2AccHormaxHorizontal decelerationStart: velocity xy maximumEnd: velocity xy minimumDistance xy > 25 cmDistance z < 10 cmMaximum deceleration xym/s2AccDecmaxVertical upwards accelerationStart: velocity z minimumEnd: velocity z maximumDistance z > 25 cmMaximum acceleration zm/s2AccVUmaxVertical downwards accelerationStart: velocity z maximumEnd: velocity z minimumDistance z > 25 cmMaximum acceleration zm/s2AccVDmaxPitch-up rotationStart: pitch minimumEnd: pitch maximumDegrees rotated > 45 degDistance xyz < 10 cmAverage pitch velocityrev/sPitchUvel , avgPitch-down rotationStart: pitch maximumEnd: pitch minimumDegrees rotated > 45 degDistance xyz < 10 cmAverage pitch velocityrev/sPitchDvel , avgYaw turnStart: velocity yaw = 0 deg/sEnd: velocity yaw = 0 deg/sDegrees rotated > 90 degPitch maximum < 75 degDistance xyz < 10 cmAverage yaw velocityrev/sYawvel , avgArcing turnStart: Δ heading velocity > 0 . 25 rev/sEnd Δ heading velocity < 0 . 25 rev/sVelocity xy min > 50 cm/sDistance xy > 25 cmDistance z < 10 cmAverage xy velocity*radius*Centripetal acceleration*m/smm/s2Arcvel , avgArcradArccent , maxPitch roll turnStart: velocity maximumEnd: velocity maximumPitch maximum > 75 degDistance xy before velocityMin > 12 . 5 cmDistance xy after velocity Min < 12 . 5 cmDistance z < 10 cmtime†degrees turned†sdegPRTtimePRTdeg*for a 25 cm segment centered at the sharpest point of the turn†for a 25 cm segment centered at the minimum velocity xyz The five translational maneuvers were defined using velocity minima and maxima , and only sequences with at least 25 cm of travel were analyzed .","The 3D acceleration maneuvers started from a velocity minimum and ended with a velocity maximum .","The performance metric calculated for these maneuvers was the maximum translational velocity ( Velmax ) .","The horizontal acceleration maneuvers were bounded by horizontal velocity minima and maxima , and were constrained to no more than 10 cm of vertical distance traveled .","The performance metric calculated for these maneuvers was the maximum horizontal acceleration ( AccHormax ) .","The horizontal deceleration maneuvers and the corresponding performance metric , maximum horizontal deceleration ( DecHormax ) , were bounded by horizontal velocity maxima and minima .","The vertical upward acceleration and vertical downward acceleration maneuvers were bounded by vertical velocity minima and maxima .","The performance metrics calculated from these maneuvers were , respectively , maximum upward ( AccVUmax ) and maximum downward ( AccVDmax ) accelerations .","All translational accelerations and decelerations were expressed as positive values , so that higher values represent a higher level of performance .","We defined three rotational maneuvers: pitch-up rotations , pitch-down rotations , and yaw turns .","These sequences were bounded by the zero-crossings of the azimuthal and pitch velocities .","In contrast to translational maneuvers , which were defined by the maxima and the minima of the velocities , the rotational maneuvers begin and end with changes in rotational velocity direction .","Thus , the performance metrics calculated from these rotational maneuvers were the average rotational velocities over the whole maneuver instead of maximum accelerations or decelerations .","An additional constraint common to all three rotational maneuvers is that the linear distance traveled was less than 10 cm .","We chose 10 cm as a general cutoff here and elsewhere because this value is close to the body length of a bird and the wing span at mid-downstroke , thus providing a good threshold for distinguishing translational motion .","The pitch-up and pitch-down maneuvers were defined as having continuous pitch velocity in the upward or downward direction , respectively .","Only maneuvers with a total pitch rotation greater than 45° were analyzed .","From these maneuvers we calculated either the average pitch-up ( PitchUvel , avg ) or pitch-down ( PitchDvel , avg ) velocity as performance metrics .","Defining yaw turns is challenging because hummingbirds fly with an upright body posture .","When the body posture is near vertical , azimuthal rotation is implemented by rolling about the body axis , but when the body posture is near horizontal , azimuthal rotation is implemented by yawing the body axis .","We therefore define yaw turns as azimuthal changes in direction when the body pitch angle is below 75° .","An additional constraint specific to yaw turns was a requirement for at least 90° change in azimuth .","From these trajectories we measured the average yaw velocity ( Yawvel , avg ) as the performance metric .","In addition to five translational and three rotational maneuvers , we also considered two maneuvers that are complex turns with translational components .","Arcing turn maneuvers were defined as sequences with a heading velocity > 90°/sec , a minimum total translational velocity > 0 . 5 m/s , a total distance traveled > 25 cm , and a vertical distance traveled < 10 cm .","These search parameters reliably extract arcing turns that occur in the horizontal plane .","To compare arcing turns of different shapes and scales we clipped the trajectories to a length of 25 cm centered at the sharpest point of the turn .","From the clipped trajectory we analyzed three performance metrics , average velocity ( Arcvel , avg ) , radius ( Arcrad ) , and the maximum centripetal acceleration ( Arccent , max ) .","The latter two were calculated using the following equations: Arcrad=Arcdistance traveledΔHeadingrad Arccent , max=Arcvel , avg2Arcrad Pitch-roll turn maneuvers have been described in hummingbirds and are characterized by the following sequence:","a ) deceleration ,","b ) increase in pitch to near vertical ,","c ) azimuthal rotation by rolling the body , and","d ) acceleration in a new direction ( Clark , 2011 ) .","These maneuvers were identified by searching for sequences of deceleration followed by acceleration with a maximum pitch > 75° .","Just as we did for the yaw turns , we assume that above a pitch angle of 75° , the rotation is primarily dominated by a body axis roll , even if there may be a slight yawing component .","For this reason , we maintain the established 'pitch-roll' terminology to describe these types of turns .","These sequences were clipped to a linear distance of 25 cm centered on the point of the lowest translational velocity .","Only clipped sequences in which the total vertical displacement was less than 10 cm were analyzed .","The performance metrics for pitch-roll turns were the time taken ( PRTtime ) and the degrees turned ( PRTdeg ) .","Arcing turns and pitch-roll turns are two different mechanisms for generating a change in heading with no overlap in our data set by definition ( Table 1 ) .","We analyzed how morphology , burst capacity , and competitor presence influenced the relative use of these two turns .","The pitch-roll percent ( PRT% ) was defined as the number of pitch-roll turns divided by total the number of arcing and pitch-roll turns extracted from each trial .","Descriptive statistics for morphology and load lifting are provided in Table 2 . A large sample of values was obtained for each maneuvering performance metric ( Table 3 ) .","Figure 2 shows the distributions of trial means for all performance metrics . 10 . 7554/eLife . 11159 . 007Table 2 . Wing morphology and load lifting performance of male Anna’s hummingbirds ( n = 20 individuals ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 007TraitMeanRangeWing length50 . 97 mm[45 . 76 , 55 . 45]Wing area 1355 mm2[1051 , 1653]Wing aspect ratio7 . 73[7 . 13 , 8 . 46]Body mass4 . 64 g[4 . 09 , 5 . 61]Mass of weights lifted5 . 93 g[4 . 00 , 7 . 24]10 . 7554/eLife . 11159 . 008Table 3 . Descriptive statistics and sample sizes for maneuvering performance .","Grand mean values were calculated by first taking the mean of each bird’s trial averages ( i . e . , the bird means ) , and then taking the mean of the bird means ( n = 20 birds in 20 solo trials and 16 paired competition trials ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 008ManeuverabilityPerformance metric# TrajectoriesGrand mean[Range of means]Linear accelerationsVelmax71 , 0072 . 22 m/s[1 . 20 , 2 . 94]AccHormax47 , 2876 . 30 m/s2[2 . 96 , 8 . 83]DecHormax51 , 2456 . 67 m/s2[9 . 03 , 3 . 45]AccVUmax6 , 9353 . 78 m/s2[2 . 98 , 4 . 67]AccVDmax9 , 2843 . 58 m/s2[4 . 69 , 2 . 68]Rotational velocitiesPitchUvel , avg6 , 0851 . 13 rev/s[0 . 91 , 1 . 34]PitchDvel , avg14 , 8071 . 00 rev/s[1 . 19 , 0 . 78]Yawvel , avg12 , 6601 . 52 rev/s[1 . 32 , 1 . 75]Complex turnsPitch-rollPRTdeg17 , 133133 . 3 º[34 . 9 , 162 . 7]PRTtime17 , 1330 . 47 s[0 . 38 , 0 . 60]ArcingArcrad6 . 9450 . 48 m[0 . 14 , 0 . 70]Arcvel , avg6 , 9451 . 57 m/s[0 . 80 , 2 . 26]Arccent , max6 , 9456 . 59 m/s2[3 . 42 , 10 . 80]Use of turnsPRT% 24 , 0780 . 69[0 . 39 , 0 . 87]10 . 7554/eLife . 11159 . 009Figure 2 . Distributions of mean performance metric values for n = 52 bird-trial combinations . Only PRTdeg has statistically significant outliers .","See Figure 2—figure supplement 1 for distributions of residuals from the best-fit model in each case .","Note that two statistical outliers were omitted from the analysis of PRTdeg . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 00910 . 7554/eLife . 11159 . 010Figure 2—figure supplement 1 . Distributions of residuals from the best-fit model for each performance metric . Note that two statistical outliers were omitted from the analysis of PRTdeg . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 010 All performance metrics based on total and horizontal linear accelerations and complex turns were highly repeatable , with >80% of the variation in these metrics attributable to differences among individuals ( Figure 3 ) .","The rotational performance metrics and the percent of turns that were pitch-roll turns were moderately repeatable , with 40–70% of the variation in these metrics attributable to among-individual differences .","The vertical accelerations were not repeatable , as the 95% confidence intervals for repeatability of these metrics overlapped zero . 10 . 7554/eLife . 11159 . 011Figure 3 . Most maneuvering performance metrics are highly repeatable . Values > 70% are considered to have high repeatability , 40–70% moderate repeatability , and < 40% low repeatability .","A metric is considered not repeatable if its 95% confidence intervals overlap zero . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 011 The best-supported models for each maneuvering performance metric are given in Table 4 .","Burst muscle capacity was an important predictor for most of the maneuvering performance metrics .","Birds that lifted more weight ( accounting for their wing morphology ) tended to accelerate and decelerate faster , and they tended to perform maneuvers with higher velocity ( Figure 4 ) .","However , burst muscle capacity was not an important determinant of vertical acceleration and deceleration , as candidate models including burst performance as a predictor were not supported .","Birds that lifted more weight also executed pitch-up and pitch-down maneuvers with higher rotational velocities .","Burst capacity was not a strong determinant of yaw performance .","Although yaw velocity was somewhat positively related to burst capacity ( Figure 4 ) , candidate models of yaw velocity that included burst as a predictor were not well supported . 10 . 7554/eLife . 11159 . 012Table 4 . Maneuvering performance in relation to burst performance , wing morphology , and competitor presence ( n = 20 birds in 20 solo trials and 16 paired competition trials ) .","Standardized beta coefficients and R2GLMM ( m ) values are reported for either the best-fit model , or , if there was support for more than one model , the average of supported models .","The standardized beta coefficient is a measure of effect size that can be compared among predictors in the same model .","Relative importance is a measure of the weight of evidence in favor of a predictor on a scale from 0–1 , and is reported for burst capacity and wing morphology variables as these alone were subject to model selection .","Marginal R2GLMM ( m ) provides a measure of the combined explanatory power of fixed effects of interest ( competitor presence , burst muscle capacity , and wing morphology effects combined ) .","Details of all candidate models are provided in Supplementary file 1 . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 012ModelSupport forFixed effectsStd beta coef [95% CI]Relative importanceR2GLMM ( m ) Burst+ morphology+ competitorVelmaxburstcompetitor presencemassburstwing lengthwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) days post-capture–0 . 04 [–0 . 18 , 0 . 10]0 . 10 [–0 . 01 , 0 . 22]0 . 09 [0 . 00 , 0 . 18]–0 . 08 [–0 . 22 , 0 . 06]0 . 10 [–0 . 07 , 0 . 28]1 . 01 [0 . 59 , 1 . 42]1 . 06 [0 . 68 , 1 . 43]–0 . 07 [–0 . 24 , 0 . 11]----1 . 000 . 250 . 26------0 . 28AccHormaxburst + competitioncompetitor presencemassburstexperiment ( CA1 ) experiment ( CA2 ) days post-capture–0 . 46 [–0 . 82 , –0 . 11]0 . 20 [–0 . 28 , 0 . 69]0 . 39 [0 . 00 , 0 . 77]4 . 01 [2 . 46 , 5 . 56]3 . 68 [2 . 72 , 4 . 64]–0 . 39 [–1 . 09 , 0 . 32]----1 . 00------0 . 18DecHormaxburst + competitioncompetitor presencemassburstexperiment ( CA1 ) experiment ( CA2 ) days post-capture–0 . 47 [–0 . 78 , –0 . 16]0 . 31 [–0 . 13 , 0 . 74]0 . 41 [0 . 06 , 0 . 76]3 . 86 [2 . 47 , 5 . 25]3 . 64 [2 . 76 , 4 . 51]–0 . 24 [–0 . 88 , 0 . 39]----1 . 00------0 . 19AccVUmaxintercept-onlyNANANA0 ( NA ) AccVDmaxintercept-onlyNANANA0 ( NA ) PitchUvel , avgburstcompetitor presencemassburstexperiment ( CA1 ) experiment ( CA2 ) 0 . 02 [–0 . 02 , 0 . 06]0 . 00 [–0 . 04 , 0 . 04]0 . 03 [–0 . 01 , 0 . 07]0 . 14 [0 . 06 , 0 . 23]0 . 13 [0 . 03 , 0 . 23]----1 . 00----0 . 10PitchDvel , avgcompetition+ burstcompetitor presencemassburstwing lengthwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) 0 . 06 [0 . 01 , 0 . 10]0 . 01 [–0 . 04 , 0 . 05]0 . 03 [–0 . 01 , 0 . 08]0 . 04 [–0 . 03 , 0 . 12]–0 . 04 [–0 . 13 , 0 . 05]0 . 19 [0 . 03 , 0 . 34]0 . 22 [0 . 03 , 0 . 41]----0 . 660 . 370 . 28----0 . 18Yawvel , avgintercept-onlyNANANA0 ( NA ) PRTdegintercept-onlyNANANA0 ( NA ) PRTtimeburstcompetitor presencemassburstwing lengthwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) 0 . 00 [–0 . 01 , 0 . 01]–0 . 01 [–0 . 03 , 0 . 00]–0 . 02 [–0 . 03 , 0 . 00]–0 . 01 [–0 . 03 , 0 . 01]0 . 01 [–0 . 01 , 0 . 04]–0 . 08 [–0 . 12 , –0 . 03]–0 . 11 [–0 . 16 , –0 . 05]–--1 . 000 . 210 . 23–--0 . 29Arcradburstcompetitor presencemassburstwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) –0 . 02 [–0 . 07 , 0 . 03]0 . 01 [–0 . 03 , 0 . 06]0 . 06 [0 . 01 , 0 . 10]–0 . 06 [–0 . 15 , 0 . 03]0 . 25 [0 . 12 , 0 . 37]0 . 29 [0 . 06 , 0 . 52]----1 . 000 . 44–--0 . 22Arcvel , avgburstcompetitor presencemassburstexperiment ( CA1 ) experiment ( CA2 ) days post-capture–0 . 01 [–0 . 09 , 0 . 08]0 . 03 [–0 . 06 , 0 . 12]0 . 11 [0 . 04 , 0 . 19]0 . 89 [0 . 59 , 1 . 19]0 . 74 [0 . 56 , 0 . 92]–0 . 06 [–0 . 20 , 0 . 08]––1 . 00–----0 . 18Acccent , maxwing shapecompetitor presencemasswing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) days post-capture0 . 29 [–0 . 37 , 0 . 94]–0 . 20 [–0 . 74 , 0 . 34]1 . 09 [0 . 19 , 1 . 99]5 . 93 [4 . 02 , 7 . 84]0 . 85 [–1 . 59 , 3 . 28]–1 . 76 [–2 . 62 , –0 . 90]----1 . 00------0 . 36PRT% wing shape + competition+ burst + wing sizecompetitor presencemassburstwing lengthwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) –0 . 14 [–0 . 19 , –0 . 09]0 . 00 [–0 . 04 , 0 . 05]0 . 04 [0 . 00 , 0 . 09]–0 . 06 [–0 . 13 , 0 . 01]–0 . 16 [–0 . 24 , –0 . 07]0 . 17 [–0 . 03 , 0 . 36]0 . 44 [0 . 19 , 0 . 69]----1 . 000 . 611 . 00----0 . 2710 . 7554/eLife . 11159 . 013Figure 4 . Burst muscle capacity was associated with most maneuvering performance metrics . Each panel shows partial residuals for a performance metric ( y-axis ) in relation to burst muscle capacity ( x-axis ) for the most supported candidate model with burst capacity as a predictor .","Partial residual values ( y-axis ) account for the other fixed effects in that model .","Lines show model predictions assuming the median value of continuous predictors , and averaging across experiments and levels of competitor presence .","Prediction lines are dashed for metrics where burst performance was not present in any of the supported models .","Color is used to denote data points from the same bird ( online version only ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 013 Burst muscle capacity was also associated with some , but not all maneuvering performance metrics related to complex turns .","Birds that lifted more weight for their wing morphology tended to execute faster , larger radius arcing turns ( Figure 4 ) .","However , the centripetal acceleration of arcing turns was not associated with burst capacity .","Hummingbirds with higher load lifting capacity executed pitch-roll turns in less time .","Burst capacity was not a strong determinant of heading change during pitch-roll turns .","Lastly , birds with higher burst muscle capacity used pitch-roll turns for proportionately more of their heading changes .","Wing morphology , specifically the aspect ratio , was an important predictor for two performance metrics: centripetal acceleration and the percent of direction changes that were pitch-roll turns ( Figure 5 ) .","Hummingbirds with long , narrow wings tended to perform arcing turns with higher centripetal accelerations , relative to birds with short , wide wings .","Birds with higher aspect ratio wings also used proportionately more arcing turns than birds with low aspect ratio wings . 10 . 7554/eLife . 11159 . 014Figure 5 . Aspect ratio was associated with two maneuvering performance metrics . Each panel shows partial residual performance ( y-axis ) in relation to wing aspect ratio ( x-axis ) from a best-fit model that identified aspect ratio as an important predictor .","Note that the partial residuals for PRT% in","( b ) go above 1 because PRT% was modeled as a normally-distributed ( Gaussian ) variable .","All other features as in Figure 4 . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 014 Body mass was included in candidate models 1–7 because we had anticipated that body mass would have a strong influence on variation in maneuvering performance .","However , for every performance metric in Table 4 , the coefficient estimate for body mass had confidence intervals that broadly overlapped zero .","We did not detect a substantial effect of competitor presence on many of the performance metrics ( Table 4 ) .","Two metrics , horizontal acceleration and deceleration , were affected , but in the direction opposite to what we predicted .","Specifically , birds performed maneuvers with lower acceleration ( –0 . 46 m/s2 difference on average ) and lower deceleration ( –0 . 47 m/s2 ) in the presence of a competitor , relative to solo flight ( Figure 6a , b ) .","One metric , pitch-down velocity ( Figure 6c ) , did increase during competition as predicted ( 0 . 06 rev/s difference on average ) .","We had no prediction for how competition would influence the relative use of pitch-roll and arcing turns , but found that birds used proportionately more arcing turns in the presence of a competitor ( Figure 6d ) .","Specifically , 35% of direction changes were arcing turns on average ( and 65% pitch-roll ) when a competitor was present , whereas during solo flight , only 23% of direction changes were arcing turns ( and 77% pitch-roll ) on average . 10 . 7554/eLife . 11159 . 015Figure 6 . Competitor presence was associated with four maneuvering performance metrics . Each panel shows residual performance ( y-axis ) in relation to competitor presence from a best-fit model where competitor presence had a detected effect .","All other features as in Figure 4 . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 015"],["We collected a large number of free flight measurements for each of 20 individual hummingbirds to examine the biomechanical determinants of maneuverability .","Other studies have measured elements of maneuvering performance of hummingbirds in the field ( Clark , 2009; Sholtis et al . , 2015 ) and documented the maximum velocities , accelerations , and rotations obtained during specific maneuvers .","Our values for velocity and acceleration are considerably lower than either of the field studies , likely because of cage size .","However , the benefit of using a flight chamber is that it allowed us to evaluate the relative contributions of different factors to the performance we observed .","We found that hummingbirds maneuvered with highly repeatable performance ( Figure 3 ) .","Maximum weight lifted during load lifting trials predicted most of the performance metrics that we measured , independent of a bird’s wing size and shape , such that birds with higher burst muscle capacity flew faster , had higher horizontal accelerations , faster rotations , and higher performance during complex turns ( Figure 4 ) .","Aspect ratio predicted only two performance metrics , such that birds with higher aspect ratio wings performed turns with higher centripetal acceleration and a greater percentage of arcing turns ( Figure 5 ) .","When flying in the presence of a competitor , hummingbirds used faster pitch velocities , although they used slower horizontal accelerations and decelerations .","During competition trials birds also increased the proportion of arcing turns used ( Figure 6 ) .","Collectively , these results suggest that burst muscle capacity is a much more important predictor of flight maneuverability than within-species variation in body mass , wing morphology , and competition with conspecifics .","Why were body mass and wing size not associated with maneuvering performance ?","Wing morphology has well-known physical affects on flight performance: aspect ratio predicts aerodynamic efficiency , wing area is directly proportional to aerodynamic force , and wing length is a strong predictor of wingbeat frequency .","All of these morphological traits , along with body mass , could affect maneuverability in flight , either individually or in combination .","For example , wing loading ( the ratio of body mass to wing area or to area swept by the wings ) was initially thought to be a key predictor of hummingbird flight performance and behavioral ecology ( Feinsinger and Chaplin , 1975; Feinsinger , 1976; Feinsinger and Colwell , 1978; Feinsinger et al . , 1979 ) .","However , in our analysis the hypothesis that wing size and body mass together determine maneuvering performance was not supported for any performance metric ( see Supplementary file 1 ) .","We found it especially surprising that only wing shape ( and not wing size ) predicted maneuvering performance .","It is possible that other morphological traits may determine maneuvering performance , or that subtle relationships may have gone undetected , because our analysis was limited to 20 individuals of a single species .","It would be informative to expand this analysis to other species with potentially greater within-species variation in wing morphology , and to assess maneuverability across different hummingbird species with divergent morphologies .","Almost all of the performance metrics were highly repeatable , which indicates a potential role for intrinsic influences of wing morphology in determining maneuverability .","However , aspect ratio was the only morphological parameter that predicted performance , and only for a limited set of maneuvers .","Aspect ratio is a key determinant in wing efficiency for fixed wings , such as during gliding ( Pennycuick , 1983 ) , and it has recently been demonstrated that higher aspect ratio wings correspond to higher power factors in the revolving wings of hummingbirds ( Kruyt et al . , 2014 ) .","We found that aspect ratio had a strong effect on the few performance metrics that it predicted , but did not affect most features of maneuvering performance .","This suggests a limited role for aerodynamic efficiency in many features of maneuvering .","Burst muscle capacity predicted most of the performance metrics we considered , independently of any association with wing size or shape .","Load lifting is measured as a transient escape maneuver that is likely anaerobic and performed inefficiently .","All hummingbirds reach maximum load lifting performance at a geometric limit set by the amplitude of the wings: wing stroke amplitude cannot extend much past 180° without the two wings interfering with each other physically and aerodynamically ( Chai and Dudley , 1995; Chai et al . , 1997; Chai and Millard , 1997; Altshuler and Dudley , 2003 ) .","Maximum load lifting also elicits a substantial increase in wingbeat frequency as a constant fraction of baseline wingbeat frequency ( Altshuler and Dudley , 2003 ) .","Thus , maximum load lifting performance involves brief increases in muscle strain and muscle velocity to physically imposed limits .","The capacity to increase muscle strain and velocity has previously been shown to influence foraging behavior and competitive ability ( Altshuler , 2006 ) .","The results of the current study demonstrate that it also underlies multiple features of maneuvering performance .","The two performance metrics that were not repeatable are vertical accelerations and decelerations , which were expected to be important based on previous observations of hummingbird competitive interactions ( Altshuler , 2006 ) and mating displays ( Clark , 2009 ) .","Moreover , vertical performance was not well predicted by morphology , burst capacity , or competitor presence in this study .","The dimensions of our experimental chamber likely influenced our observations of vertical performance .","Hummingbirds in captivity tend to fly near the top of their cages , and the vertical dimension of the chamber ( 1 . 5 m ) may have limited vertical movement .","Male hummingbirds are extremely aggressive towards conspecifics ( Kodric-Brown and Brown , 1978; Carpenter et al . , 1983 ) and other species of hummingbirds ( Stiles and Wolf , 1970; Wolf et al . , 1976 ) .","The most territorial species will vigorously defend territories ( Carpenter et al . , 1983 ) and lekking sites ( Rico-Guevara and Araya-Salas , 2015 ) .","In staged competition studies , paired hummingbirds will also establish and defend territories ( Tiebout , 1993 ) .","We originally intended to use competition to elicit high levels of flight activity and maneuvering performance in territorial male Anna's hummingbirds ( Stiles , 1982 ) .","However , we found that competitor presence affected only a small number of the maneuvering performance metrics that we measured .","Pitch-down velocity increased with competition whereas horizontal acceleration and deceleration actually decreased .","We do not know why these three metrics ( in addition to PRT%; see below ) were strongly affected by competition or why they were affected in the directions observed .","However , there are several possible causes for why competitor presence did not affect the other metrics:","1 ) we were unable to elicit a high level of competition or territoriality;","2 ) the birds may have worked out dominance without the aggressive interactions normally seen outdoors; and/or","3 ) the interactions required to establish dominance may have been very brief ( Maynard Smith , 1974 ) such that they comprised only a minuscule sample of the maneuvers we analyzed .","This experiment was not designed to study the effects of maneuvering performance on competitive success , although this represents an important topic for future investigation .","Laboratory performance tests do not always reflect field behavior ( Irschick , 2003 ) and outdoor studies of maneuvering performance will be important for understanding the role of maneuverability in competitive interactions .","Recent advances in video tracking ( Theriault et al . , 2014; Shelton et al . , 2014 ) should make it possible to track individuals for multiple measurements .","The most substantial result of competitor presence was the increase in the use of arcing over pitch-roll turns .","These two types of turns represent different strategies for changing direction that differ in duration and amount of heading change .","Arcing turns require less time but are used for smaller heading changes , whereas pitch-roll turns are longer but can be used to change heading by 180° ( Figure 7 ) .","Given that hummingbird agonistic interactions can involve direct contact and stabbing with bills ( Tiebout , 1993; Clark and Russell , 2012; Rico-Guevara and Araya-Salas , 2015 ) , slow turns in place could make a bird more vulnerable during competition . 10 . 7554/eLife . 11159 . 016Figure 7 . Arcing and pitch-roll turns are two classes of complex maneuver that differ in turn magnitude and duration . Representative examples of arcing","( a ) and pitch-roll","( b ) turns are depicted from the above perspective .","Arcing turns ( Arc; orange ) and pitch-roll turns ( PRT; green ) differed in the degrees turned","( c ) and elapsed time","( d ) .","Circles represent bird-trial means ( n = 52 ) with grand means indicated with black lines .","Histograms for the pooled dataset of all maneuvers are given on the right .","The outliers for degrees turned in pitch-roll turns were included when calculating the grand means but not in the model analyses ( Table 4 ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 016 The relative use of arcing and pitch-roll turns was the only metric in our study that was influenced by all of morphology , burst muscle capacity , and competitor presence .","The minimum radius of an arcing turn is limited by the maximum centripetal acceleration that a bird can generate while maintaining lift .","The speed of a pitch-roll turn is limited by the ability to decelerate and then accelerate .","Birds with higher wing aspect ratio may have preferred arcing turns because they were able to generate higher centripetal accelerations .","Birds with higher burst muscle capacity may have favored pitch-roll turns because they had higher accelerating and decelerating performance .","These observations suggest the hypothesis that high aspect ratio and high burst capacity enhance maneuverability .","This hypothesis could be evaluated by comparing hummingbird species that differ in wing shape , foraging strategy , and burst capacity ( Altshuler et al . , 2004b; 2010a; Altshuler , 2006; Kruyt et al . , 2014 ) .","By constraining hummingbirds to fly in a large chamber we were able to track and measure a large sample of maneuvers attributed to individuals with known morphological traits and burst performance .","A major contribution of our study is the development of an assay of free flight maneuvering performance based on large numbers of stereotyped movements .","Using this method , we identify several performance metrics that were highly repeatable across trials for individual hummingbirds , strongly correlated with individual morphological and physiological characteristics , and largely uninfluenced by the added motivation of a conspecific competitor .","This approach to measuring maneuverability will be useful for future studies comparing maneuvering performance across different experimental manipulations , geographic ranges , or ecological , morphological , and phylogenetic groups ."],["We captured and filmed 20 adult male Anna's hummingbirds ( Calypte anna ) at the University of California , Riverside ( eight birds in July-October 2009; four birds in January-March 2010 ) and the University of British Columbia ( eight birds in December 2013-April 2014 ) .","The hummingbirds were housed in individual cages and fed ad libitum with a solution of artificial nectar ( Nektar-Plus , Nekton , Pforzheim , Germany ) and sucrose .","The flight arenas were large rectangular cages ( 3 x 1 . 5 x 1 . 5","m ) built with an aluminum frame and had either garden mesh ( California ) or clear acrylic ( British Columbia ) side panels .","The cages contained multiple perches and a single feeder hung from the roof of the cage .","Before the first trial , each bird was allowed to acclimate to the flight arena and learn where the perches and the feeder were located .","The trials began once the birds were actively exploring the cage and consistently visiting the feeder and both perches .","At this point , we recorded high-speed video of a two-hour solo trial for each bird .","Following solo trials ( between 0–23 days later ) , birds were paired and filmed for another two hours in competition trials .","One bird in each pair was marked with a small square of retro-reflective tape placed between the shoulder blades for identification .","The birds filmed in British Columbia had one competition trial and the birds in California had two competition trials .","In the latter case , the second trial consisted of previously unknown opponents that were chosen randomly from the remaining pool .","The competition trials involved chases , displacements , and aerial displays , but very little contact .","Regardless , we monitored the competition trials to ensure that no birds were harmed or excluded from the feeder .","Following each round of solo and competition trials , we performed asymptotic load lifting experiments using the techniques described in Chai et al . ( Chai et al . , 1997 ) , and subsequently used in other studies estimating maximum burst power output ( Chai and Millard , 1997; Altshuler et al . , 2004a; 2010b; Altshuler , 2006 ) .","Here , we use the mass of maximum number of beads lifted by each individual as a measure of burst performance .","Immediately following load lifting , we weighed the birds and photographed both wings in an outstretched position against white paper with a reference scale ( Chai and Dudley , 1995 ) .","We oriented the wing image and divided it into pixel wide strips representing the wing chords at each value of wing radius .","Values for aspect ratio , wing area , and wing length were then calculated based on equations in Ellington ( Ellington , 1984 ) .","We considered wing area and wing length as two potential measures of wing size , but these traits were highly correlated in our dataset ( R2 = 0 . 85 , p < 0 . 0001 , n = 20 ) .","Because these two traits did not vary independently in our relatively small sample of 20 hummingbirds , we could not consider them independently .","We therefore selected wing length as the more robust measure of wing size , because unlike area , wing length is less prone to measurement error as a result of variation in feather overlap when wings are positioned for measurements .","We verified that our results were consistent when using wing area instead of length , and thus these two traits should be considered interchangeable as measures of wing size in this study .","Because both wing morphology and muscle capacity may influence burst performance , we used burst performance controlled statistically for wing morphology as a measure of burst muscle capacity .","Further details are provided below in the Statistical analysis section .","All procedures were conducted under approval of the Institutional Animal Care and Use Committee at the University of California , Riverside and the Animal Care Committee at the University of British Columbia .","We used an automated tracking system to measure both body position and orientation of flying birds in three dimensions ( Video 1 ) .","A complete description of the tracking algorithm and hardware components is provided in ( Straw et al . , 2011 ) .","The core algorithms were written in Python ( Python Software Foundation , 2012 ) , and are available via github ( PyMVG: https://github . com/strawlab/pymvg; adskalman: https://github . com/astraw/adskalman; MultiCamSelfCal: https://github . com/strawlab/MultiCamSelfCal/ ) .","We adapted this system for recording hummingbird solo and competitive flight trajectories with four or five digital cameras ( GE680 , Allied Vision Technologies , Burnaby , Canada ) .","The cameras were mounted on the ceiling and recorded at 640 x 480 pixel resolution at 200 frames per second ( Figure 1a ) .","We calibrated the filming volume by moving a single light-emitting diode throughout the arena to acquire data for an automated self calibration algorithm ( Svoboda et al . , 2005 ) .","This algorithm provides a relative calibration ( non-linear warping distortion parameters and 3x4 camera calibration matrices ) across all cameras .","This calibration is brought into absolute terms ( the scale , rotation , and translation are found ) by matching a manually measured 3D model of the flight arena with reconstructed image coordinates using the ‘estsimt’ function of the MultiCamSelfCal toolbox ( Svoboda et al . , 2005 ) .","To minimize the effect of errors in the 3D tracking , we used a forward/reverse non-causal Kalman filter ( Rauch–Tung–Striebel smoother ) applied to the online state estimate of position and velocity from the realtime Kalman filter .","The smoothing parameters were chosen so that seven traces of a tracked , falling object yielded an average peak acceleration of 9 . 8 m/s2 .","The process covariance matrix we used is: Qpos=σ2×T3300T22000T3300T22000T3300T22T2200T000T2200T000T2200T where σ2 is 0 . 01 and T is the interval between frames ( 0 . 005 s ) .","The observation covariance matrix we used is: Rpos=0 . 0001440000 . 0001440000 . 000144 Figure 1c and d show examples of two trajectories with plots of the unsmoothed data , the data smoothed with Qpos and Rpos , and the effects of two different smoothing parameters ( Rpos x 10 , Rposx 0 . 1 ) .","Following establishment of the 3D trajectories , the tracking system assigned 3D body orientation vectors to each bird in each frame based on 2D estimates of the long axis of the body .","Body orientation was estimated using an algorithm that fit orientations to the body axis in each 2D image .","Each sequence of five consecutive images cropped around the bird was aligned at the optical center of intensity .","Averaging these images effectively eliminated the wings and emphasized the body .","Orientation was estimated by calculating the covariance matrix of the image luminance and then computing the eigensystem of this covariance matrix .","The eigenvector associated with the largest eigenvalue was taken as the orientation .","Orientation vector assignments were also smoothed with a Kalman filter using more restrictive smoothing parameters than were used to smooth body position .","To determine appropriate smoothing parameters we replotted the smoothed body orientation vectors onto a sample of videos , and visually chose the ones that provided the best fit .","The process covariance matrix used for body orientation ( Qori ) is the same as the process covariance matrix used for body position ( Qpos ) and the observation covariance matrix used is: Rori=0 . 000001440000 . 000001440000 . 00000144 Once the body orientations were calculated we used a dynamic programming algorithm to decide which end of the vector was the head and which end was the tail .","The direction of the head was chosen based on the direction of the previous orientation , the direction of travel , and the vertical up direction .","For each frame ( n ) , the 'cost' associated with the two possible orientations ( Ori→ , -Ori→ ) were calculated: CostOri=Speed ( Ori→n⋅Vel→modOri→nVel→mod ) + ( 1−Speed ) ( Ori→n⋅Up→Ori→nUp→ ) + ( 1−Speed ) ( Ori→n⋅Ori→n−1Ori→nOri→n−1 ) Cost−Ori=Speed ( −Ori→n⋅Vel→modOri→nVel→mod ) + ( 1−Speed ) ( −Ori→n⋅Up→Ori→nUp→ ) + ( 1−Speed ) ( −Ori→n⋅Ori→n−1Ori→nOri→n−1 ) where Ori→ is the body vector , Vel→mod is the modified velocity vector tipped up 15º towards the vertical direction .","Up→ is the vertical direction vector , Ori→n−1 is the orientation during the previous frame , and if the magnitude of the velocity is greater than 0 . 5m/s: Speed=Vel→ otherwise: Speed=0 . 5 This approach accounted for the tendency of hummingbirds to fly forwards and with an upright posture , but allowed for exceptions in the case of backwards flight , inversions , and dives , particularly if these occurred at low speeds .","The magnitudes of calculated accelerations and , to a lesser extent , velocities derived from position data were influenced by the specific smoothing parameters .","Examples of maneuvers with different smoothing parameters and their effects on the calculated performance metrics are given in Figure 1c and d .","This influence of smoothing parameters is a well known limitation of video tracking ( Walker , 1998 ) .","Thus , although acceleration values are comparable within a study , caution must be applied when comparing the magnitude of acceleration values among studies differing in camera frame rate , filming volume , calibrations , and smoothing parameters ( Walker , 1998 ) .","For our final performance metrics we used instantaneous body orientation and orientation velocity , but not orientation acceleration .","The automated tracking system extracted the 3D coordinates of multiple flying animals and saved each trajectory as a separate object ( Video 1 ) .","An object began when the tracking system detected new movement and ended when either the object stopped moving , the error in the 3D reprojection grew too large , or multiple objects came within 2 cm of each other .","In our experiments tracking hummingbird flight led to two problems in determining distinct objects .","The first is that very stable hovering can be misidentified as perching .","For example , as a bird went into an extended hovering bout , such as at a feeder , the tracking system detected the cessation of movement and ended the trajectory .","Conversely , when the bird perched at the end of a flight or in between two flights , especially if it continued to move its head or fluff its feathers , the tracking system treated the bird as moving and continued the trajectory .","Because our study focused on identifying and analyzing relatively long , moving trajectories , these types of errors did not cause problems .","The second challenge concerned identification of birds during close encounters in competition trials .","When two tracked objects became close to each other , even if they did not physically touch , the tracking system could not accurately distinguish them .","We used a conservative solution and terminated the trajectories whenever two birds came close enough that the tracked objects merged .","Birds were later identified manually by a team of digitizers who viewed the videos and assigned each object number to either the marked or unmarked bird .","The automated digitization produced a small number of extreme tracking errors , which we did not want to unduly influence statistical analyses .","We accordingly removed values >5 SDs more extreme than the mean for each performance metric .","The trimmed values comprised only 0–0 . 31% of the original pooled sample size for each metric .","We next calculated the mean of each performance metric for each bird-trial combination ( n= 52 means; 20 birds in 20 solo trials and 16 paired competition trials ) .","All statistical analyses were performed on these bird-trial means using R 3 . 1 . 1 ( R Development Core Team , 2014 ) , and the data used for the analysis are available online ( Segre et al . , 2015 ) .","Repeatability , or the intra-class correlation coefficient ( ICC ) , is defined as the proportion of variation that is attributable to differences among individuals ( Nakagawa and Schielzeth , 2010 ) .","We estimated repeatability for each performance metric from an intercept-only mixed effects model that included estimates of the population intercept ( i . e . , the grand mean ) as well as an individual intercept for each bird ( Nakagawa and Schielzeth , 2010 ) .","Such a model has two variance components , the variance of the random intercept values ( variance among individuals ) and a residual variance associated with the error term .","Repeatability is the variance among individuals divided by the total variance ( Nakagawa and Schielzeth , 2010 ) .","We used parametric bootstrapping with 5000 iterations to obtain confidence intervals for these repeatability estimates via the bootMer function in the lme4 ( v1 . 1 . 7 ) package .","Because our second question involved evaluating several possible scenarios for the influence of morphology and burst performance on maneuverability , we used an information-theoretic approach to multi-model inference ( Burnham and Anderson , 2010 ) .","Unlike dichotomous null hypothesis testing , this approach quantifies support for multiple hypotheses , and it avoids the problem of eliminating potentially important predictors when two or more alternative models are equally well supported .","The output for interpretation includes the effect size and relative importance of each predictor , and there are no null hypotheses or P values associated with this approach .","As a measure of effect size we report the standardized partial regression coefficient , std β , for each predictor , which can be used to compare their independent associations with a given response variable .","Unstandardized regression coefficients corresponding to units of the predictor variables are provided in Supplementary file 1 .","We also examined associations between burst performance , wing size , and wing shape because our load lifting assay may have incorporated effects of wing morphology as well as muscle capacity .","The mass of weights lifted during load lifting was not significantly associated with wing length in our sample of 20 individuals ( p = 0 . 87 ) , however , it was negatively associated with wing aspect ratio ( p = 0 . 04 ) controlling for site .","Thus in our model analyses we used residual burst performance controlling for wing aspect ratio and site as a measure of burst muscle capacity independent of a bird’s wing morphology .","We considered eight candidate mixed-effects models that could plausibly explain variation in each maneuvering performance metric ( Table 5 ) .","All candidate models included an individual intercept for each bird ( the random intercept term ) and were fit using the nlme ( v 3 . 1–117 ) package ( Zuur et al . , 2009 ) .","The intercept-only model included an estimate of the population intercept ( grand mean ) and random intercept terms , but no fixed effects .","Other candidate models are listed in Table 5 .","All models except the intercept-only model included the fixed effects of competitor presence , body mass , and experiment to account for potential effects of these factors .","Experiment had three levels , one for each round of trials ( California 2009 , 2010 , British Columbia 2014 ) to account for differences such as location , time of year , and filming conditions . 10 . 7554/eLife . 11159 . 017Table 5 . Candidate models of maneuvering performance .","All models include an intercept as well as a random effect of bird identity to account for repeated measures of individuals . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 017ModelFixed effectsDescription1 . Solo/comp + experiment + body mass + wing lengthWing size2 . Solo/comp + experiment + body mass + wing aspect ratioWing shape3 . Solo/comp + experiment + body mass + wing length + wing aspect ratioWing size & shape4 . Solo/comp + experiment + body mass + weight liftedBurst power5 . Solo/comp + experiment + body mass + weight lifted + wing lengthBurst power & wing size6 . Solo/comp + experiment + body mass + weight lifted + wing aspect ratioBurst power & wing shape7 . Solo/comp + experiment + body mass + weight lifted + wing length + wing aspect ratioBurst power , wing size & shape8 . Intercept-only*Candidate models 1-7 also include a fixed effect of days post-capture for the following metrics: Velmax , AccHormax , DecHormax , Arcvel , avg , and Arccent , max Two issues arose in the preliminary examination of data .","The first issue was that five of the performance metrics were significantly influenced by the number of days a bird had been in captivity .","We therefore included an additional fixed effect of the number of days since capture when analyzing these five metrics ( Table 5 ) .","The second issue was that one of the metrics , the heading change in pitch-roll turns ( PRTdeg ) , had three values that were significant outliers ( Grubb’s test , all G > 3 . 09 , all p<0 . 03; Figure 7 ) .","We determined that these three statistical outliers were not errors in the tracking system but were instead derived from one individual that used pitch-roll turns to make small heading changes , unlike the other birds .","We omitted these outliers from the analysis of heading change in pitch-roll turns to ensure that all fitted Gaussian models met the required assumptions , with no other outliers or problems of skew or heteroskedasticity .","The best-fit model for heading change in pitch-roll turns was the intercept-only model regardless of whether the outliers were included .","To quantify the variance explained by the fixed effects of interest in each model , we calculated the marginal R2GLMM","( m ) using the r . squaredGLMM function in the MuMIn ( v1 . 10 . 5 ) package ( Nakagawa and Schielzeth , 2013 ) .","This measure does not have all the properties of a traditional coefficient of determination , but like R2 it ranges from 0 to 1 , and it is an appropriate estimate of the variance explained by the fixed effects in a mixed model .","We removed the effect of experiment and the number of days post-capture when calculating R2GLMM","( m ) , because these were not effects of interest .","Thus , R2GLMM","( m ) provides a measure of the variance explained by the other supported fixed effect variables .","We evaluated the support for different models using the Akaike information criterion ( AICc ) adjusted for small sample sizes .","This was calculated using the MuMIn ( v 1 . 10 . 5 ) package with maximum likelihood estimation .","We defined the group of supported models as those with a difference in AICc < 2 from the best-fit model for each performance metric .","If no other models came within 2 AICc units of the best-fit model , we present effect size measures , their confidence intervals , and R2GLMM","( m ) for only that model .","Otherwise , we present averages of all supported models .","Details of all candidate models are provided in Supplementary file 1 .","Our third question concerned the influence of competitor presence on the performance metrics .","If the confidence interval for the coefficient estimate of competitor presence excluded zero , we examined the magnitude and direction of that effect .","Positive coefficient estimates indicate that performance was higher during competitive flights , whereas negative coefficients indicate that performance was lower in the presence of a competitor ."]],"string":"[\n [\n \"The ability of an animal to change the speed and direction of movement , defined as maneuverability ( Dudley , 2002 ) , can determine its success at avoiding predators , obtaining food , and performing other behaviors that determine the margin between life and death ( Webb , 1976; Hedenström and Rosén , 2001; Walker et al . , 2005 ) .\",\n \"Most biomechanical research on birds has focused on either brief ( e . g . , take off ) or steady state movements ( e . g . , forward flight ) that can be studied most readily in the laboratory .\",\n \"Maneuverability is therefore one of the most important but least understood aspects of animal locomotion .\",\n \"Warrick and coworkers ( Warrick et al . , 1988; Warrick and Dial , 1998 ) proposed that there are both intrinsic and facultative influences on maneuvering performance .\",\n \"For animals that perform powered flight , intrinsic maneuverability is defined by the physical limitations imposed by morphology ( Norberg and Rayner , 1987 ) , but excess muscle capacity should allow them to facultatively overcome the costs of suboptimal morphology , achieving higher levels of performance by sacrificing efficiency .\",\n \"Although compelling , this hypothesis has never been tested explicitly .\",\n \"Wing morphology is defined using measures of size ( e . g . , area or length ) and non-dimensional measures of shape ( e . g . , aspect ratio ) .\",\n \"Wing area and aspect ratio have significant and well known effects on the aerodynamics of flight in animals ( Pennycuick , 1975; Kruyt et al . , 2014; 2015 ) , and should affect maneuvering performance .\",\n \"Wing morphology influences flight efficiency ( Feinsinger and Chaplin , 1975 ) , ecological roles ( Feinsinger , 1976; Feinsinger and Colwell , 1978; Warrick , 1998 ) and competitive ability ( Feinsinger and Chaplin , 1975; Feinsinger and Colwell , 1978; Feinsinger et al . , 1979; Altshuler , 2006 ) .\",\n \"Because these previous studies focused on species and gender comparisons , less is known about how individual variation in wing morphology influences performance , especially with respect to maneuverability .\",\n \"One complication is that different wing sizes and shapes can be favored depending on the specific maneuver performed , e . g . , yaw versus banked turns .\",\n \"Given the diversity of flight behaviors , it is unclear if the requirements for maneuvering exert strong selection on wing morphology .\",\n \"Muscle capacity affects the maximum aerodynamic force a flying animal can produce .\",\n \"Aerodynamic force can be directed for performing maneuvers that require greater output than the minimum requirements for flight .\",\n \"Excess muscle capacity can also be used to compensate for anatomical or spatial constraints on wing movement ( Warrick , 1998 ) .\",\n \"Muscle output of hummingbirds has been quantified in several ways including oxygen consumption to determine metabolic input , wingbeat kinematics to estimate mechanical power output , and electromyography ( EMG ) to measure myoelectric input .\",\n \"Considering hovering flight as the point of comparison , forward flight at the fastest speeds recorded in a wind tunnel requires about 20% more metabolic ( Clark and Dudley , 2010 ) and myoelectric input ( Tobalske et al . , 2010 ) .\",\n \"Maximum sustained hovering performance has been studied by experimentally lowering air density to the lowest values in which birds are still able to hover .\",\n \"These experiments revealed that hovering in hypodense air requires ~40% higher mechanical power output ( Chai and Dudley , 1995 ) and ~60% higher spatial recruitment of muscle fibers , as measured by the spike amplitude of the electromyogram recordings ( Altshuler et al . , 2010b ) , in comparison to hovering in normal air .\",\n \"By far the most expensive flight behavior studied to date in hummingbirds is maximum load lifting , which requires 200–400% more mechanical power output ( Chai et al . , 1997; Chai and Millard , 1997; Altshuler et al . , 2010a ) , about 200% more spatial recruitment ( EMG spike amplitude ) , and 150% more temporal recruitment ( EMG spike frequency ) ( Altshuler et al . , 2010b ) compared to hovering .\",\n \"Maximum load lifting is a transient behavior that uses the bird’s natural escape response to measure burst power output .\",\n \"Thus , it is not surprising that this assay provides the maximum muscle capacity that has been measured in hummingbirds .\",\n \"It is particularly useful for quantifying variation among and within species in burst muscle capacity .\",\n \"Studies using the load lifting assay have revealed that maximum burst muscle capacity is related to hummingbird evolutionary ecology .\",\n \"Altshuler and coworkers ( Altshuler et al . , 2004b; Altshuler , 2006 ) demonstrated that ecological role is more strongly related to load lifting ability than morphological parameters such as wing loading .\",\n \"Load lifting ability is also associated with species- and gender-specific competitive ability at different elevations .\",\n \"Altshuler ( Altshuler , 2006 ) suggested that the relationship between maximum muscle capacity and competitive ability may be mediated through maneuvering performance .\",\n \"Unconstrained maneuvering performance of birds , including hummingbirds , has recently been quantified in the field without individual identification ( Shelton et al . , 2014; Sholtis et al . , 2015 ) .\",\n \"Although field studies are valuable for quantifying average species performance , individual identification and large sample sizes are required to examine sources of within-species variation .\",\n \"Here , we studied the free-flight maneuvering performance of Anna's hummingbirds ( Calypte anna ) in a large flight chamber ( Video 1 ) .\",\n \"Flight maneuvers in a chamber are not expected to be the same as outdoors , and may have lower velocities and accelerations .\",\n \"The benefit of this approach is that a large number of measurements from the same individuals can be combined with other data to examine how variation in the observed maneuvers is influenced by individual morphology and muscle capacity . 10 . 7554/eLife . 11159 . 003Video 1 . The multi-camera , automated tracking system filming two hummingbirds in the flight arena at 200 frames per second . Continuously tracked sequences are assigned an object number ( from 0 to 4 over this sequence ) .\",\n \"Body position and orientation are calculated and reprojected onto the video of four cameras .\",\n \"The videos are saved using a compression algorithm that only records the sections of the image that are moving ( Straw et al . 2011 ) .\",\n \"Thus , birds disappear from the video when they land and stop moving .\",\n \"The trajectory shown in Figure 1 is taken from the bird labeled #2 and begins at 5 . 1 seconds and ends at 8 . 05 seconds . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 003 We used a high-throughput computational approach to record the flight performance of 20 individuals alone and in the presence of a competitor .\",\n \"Flight trajectories were parsed into a set of performance metrics based on body position and orientation .\",\n \"The first goal of our study was to determine if voluntary maneuvering performance is repeatable within individuals .\",\n \"Repeatability of maneuvering performance can arise either through a strong influence of fixed traits such as morphology and anatomy , or through other consistent influences , such as motivation .\",\n \"We expect that repeatable measurements will be most useful for our second goal , determining how variation in maneuverability among individuals is influenced by natural variation in morphology and muscle capacity .\",\n \"This also required measuring morphological traits and maximum burst performance for each individual .\",\n \"Our third goal was to determine how motivation state induced by the presence of a competitor influenced maneuvering performance .\",\n \"To address this question we compared flight trials with and without competitors .\"\n ],\n [\n \"The first stage of analysis was estimating instantaneous velocities , accelerations , and headings from the raw tracking data ( Figure 1—figure supplement 1 ) .\",\n \"Translational velocity and acceleration were calculated by taking the first and second derivatives of an interpolation spline fit to the body position data ( splev and splrep functions , Scientific Python ) .\",\n \"The velocities and accelerations were split into vertical and horizontal components .\",\n \"The body orientation vector was represented in spherical coordinates as azimuth and pitch angles .\",\n \"We took the first derivatives to obtain azimuth and pitch velocities .\",\n \"Because the video tracking system did not allow a measurement of body roll , we decided to use a global coordinate system instead of a body axis-centered coordinate system .\",\n \"In our frame of reference , pitch is a global measure defined relative to the horizontal plane .\",\n \"Heading was calculated as the instantaneous direction of the horizontal translation velocity , and the heading velocity was calculated as the derivative of heading . 10 . 7554/eLife . 11159 . 004Figure 1 . A multi-camera , automated tracking system extracted hummingbird body position ( blue circle ) and orientation ( red line ) from solo and competitive flights . The trajectory shown for one bird\",\n \"( a ) is also shown in Video 1 ( see Figure 1—figure supplement 1 for time series of position , velocity , and acceleration values ) .\",\n \"Stereotyped maneuvers were classified in each trajectory\",\n \"( b ) and between one and five performance metrics were calculated from each maneuver .\",\n \"Maneuvers within a trajectory may be overlapping ( e . g . #4 , 5 , 6 ) .\",\n \"The trajectory presented in b is a top down ( x-y projection ) view of the trajectory shown in a .\",\n \"Body position and orientation were smoothed with an extended Kalman filter ( c , d ) .\",\n \"The effects of four different sets of smoothing parameters are presented for an arcing turn ( maneuver #9 in\",\n \"b ) and an upward acceleration ( maneuver #1 in\",\n \"b ) .\",\n \"Shown here are the unsmoothed position and orientation ( black trace and text ) , the chosen levels of smoothing ( blue ) , a lower level of smoothing ( green; 0 . 1 x Rpos; 0 . 1 x Rori ) , and a higher level of smoothing ( red; 10 x Rpos; 10 x Rori ) .\",\n \"The chosen smoothing parameters for body position were determined by tracking multiple dropped objects and calibrating the Z-axis acceleration to gravity .\",\n \"The chosen smoothing parameters for body orientation were determined by re-projecting the body axis vector onto the video .\",\n \"The higher and lower levels of smoothing for body position presented in this figure were both deemed too extreme , when re-projected onto the video .\",\n \"However , the level of smoothing for body orientation had minimal effect on the average yaw velocity . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 00410 . 7554/eLife . 11159 . 005Figure 1—figure supplement 1 . The representative trajectory from Figure 1 and Video 1 displayed through time . The upper three panels provide the position , velocity , and acceleration values .\",\n \"The lower three panels provide the orientation vector , orientation angle and rotation velocity values .\",\n \"The maneuvers extracted for this sequence are given by thick black lines below the traces . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 005 We then used the velocity , acceleration , and orientation data to search for a series of ten stereotyped maneuvers that were independent of time and distance scales ( Figure 1b ) .\",\n \"Five of the maneuvers were sequences defined by changes in translational velocity: 1 ) 3D accelerations , 2 ) horizontal accelerations , 3 ) horizontal decelerations , 4 ) vertical upward accelerations , and 5 ) vertical downward accelerations .\",\n \"Three maneuvers were sequences defined by changes in rotation: 6 ) pitch-up rotations , 7 ) pitch-down rotations , and 8 ) yaw turns .\",\n \"Two of the maneuvers were defined as turns with translational components: 9 ) arcing turns and 10 ) pitch-roll turns .\",\n \"These ten maneuvers are not meant to be mutually exclusive , exhaustive , or to divide the entire filming session into a set of discrete behaviors , but are instead intended to extract simple measurements that can be used as an assay for maneuvering performance .\",\n \"The search criteria for the maneuvers are given in Table 1 . Because we assume that a new maneuver must involve a change in velocity , the first search parameter was to find sequences bounded by velocity maxima and minima , or vice versa .\",\n \"We next describe the additional search parameters and the performance metrics used to quantify each maneuver . 10 . 7554/eLife . 11159 . 006Table 1 . Search parameters for the ten maneuvers analyzed in the study .\",\n \"The definitions , units , and symbols for the 14 related performance metrics are also provided . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 006ManeuverSearch parametersPerformance metricUnitsSymbol3D accelerationStart: velocity xyz minimumEnd: velocity xyz maximumDistance xyz > 25 cmMaximum velocitym/sVelmaxHorizontal accelerationStart: velocity xy minimumEnd: velocity xy maximumDistance xy > 25 cmDistance z < 10 cmMaximum acceleration xym/s2AccHormaxHorizontal decelerationStart: velocity xy maximumEnd: velocity xy minimumDistance xy > 25 cmDistance z < 10 cmMaximum deceleration xym/s2AccDecmaxVertical upwards accelerationStart: velocity z minimumEnd: velocity z maximumDistance z > 25 cmMaximum acceleration zm/s2AccVUmaxVertical downwards accelerationStart: velocity z maximumEnd: velocity z minimumDistance z > 25 cmMaximum acceleration zm/s2AccVDmaxPitch-up rotationStart: pitch minimumEnd: pitch maximumDegrees rotated > 45 degDistance xyz < 10 cmAverage pitch velocityrev/sPitchUvel , avgPitch-down rotationStart: pitch maximumEnd: pitch minimumDegrees rotated > 45 degDistance xyz < 10 cmAverage pitch velocityrev/sPitchDvel , avgYaw turnStart: velocity yaw = 0 deg/sEnd: velocity yaw = 0 deg/sDegrees rotated > 90 degPitch maximum < 75 degDistance xyz < 10 cmAverage yaw velocityrev/sYawvel , avgArcing turnStart: Δ heading velocity > 0 . 25 rev/sEnd Δ heading velocity < 0 . 25 rev/sVelocity xy min > 50 cm/sDistance xy > 25 cmDistance z < 10 cmAverage xy velocity*radius*Centripetal acceleration*m/smm/s2Arcvel , avgArcradArccent , maxPitch roll turnStart: velocity maximumEnd: velocity maximumPitch maximum > 75 degDistance xy before velocityMin > 12 . 5 cmDistance xy after velocity Min < 12 . 5 cmDistance z < 10 cmtime†degrees turned†sdegPRTtimePRTdeg*for a 25 cm segment centered at the sharpest point of the turn†for a 25 cm segment centered at the minimum velocity xyz The five translational maneuvers were defined using velocity minima and maxima , and only sequences with at least 25 cm of travel were analyzed .\",\n \"The 3D acceleration maneuvers started from a velocity minimum and ended with a velocity maximum .\",\n \"The performance metric calculated for these maneuvers was the maximum translational velocity ( Velmax ) .\",\n \"The horizontal acceleration maneuvers were bounded by horizontal velocity minima and maxima , and were constrained to no more than 10 cm of vertical distance traveled .\",\n \"The performance metric calculated for these maneuvers was the maximum horizontal acceleration ( AccHormax ) .\",\n \"The horizontal deceleration maneuvers and the corresponding performance metric , maximum horizontal deceleration ( DecHormax ) , were bounded by horizontal velocity maxima and minima .\",\n \"The vertical upward acceleration and vertical downward acceleration maneuvers were bounded by vertical velocity minima and maxima .\",\n \"The performance metrics calculated from these maneuvers were , respectively , maximum upward ( AccVUmax ) and maximum downward ( AccVDmax ) accelerations .\",\n \"All translational accelerations and decelerations were expressed as positive values , so that higher values represent a higher level of performance .\",\n \"We defined three rotational maneuvers: pitch-up rotations , pitch-down rotations , and yaw turns .\",\n \"These sequences were bounded by the zero-crossings of the azimuthal and pitch velocities .\",\n \"In contrast to translational maneuvers , which were defined by the maxima and the minima of the velocities , the rotational maneuvers begin and end with changes in rotational velocity direction .\",\n \"Thus , the performance metrics calculated from these rotational maneuvers were the average rotational velocities over the whole maneuver instead of maximum accelerations or decelerations .\",\n \"An additional constraint common to all three rotational maneuvers is that the linear distance traveled was less than 10 cm .\",\n \"We chose 10 cm as a general cutoff here and elsewhere because this value is close to the body length of a bird and the wing span at mid-downstroke , thus providing a good threshold for distinguishing translational motion .\",\n \"The pitch-up and pitch-down maneuvers were defined as having continuous pitch velocity in the upward or downward direction , respectively .\",\n \"Only maneuvers with a total pitch rotation greater than 45° were analyzed .\",\n \"From these maneuvers we calculated either the average pitch-up ( PitchUvel , avg ) or pitch-down ( PitchDvel , avg ) velocity as performance metrics .\",\n \"Defining yaw turns is challenging because hummingbirds fly with an upright body posture .\",\n \"When the body posture is near vertical , azimuthal rotation is implemented by rolling about the body axis , but when the body posture is near horizontal , azimuthal rotation is implemented by yawing the body axis .\",\n \"We therefore define yaw turns as azimuthal changes in direction when the body pitch angle is below 75° .\",\n \"An additional constraint specific to yaw turns was a requirement for at least 90° change in azimuth .\",\n \"From these trajectories we measured the average yaw velocity ( Yawvel , avg ) as the performance metric .\",\n \"In addition to five translational and three rotational maneuvers , we also considered two maneuvers that are complex turns with translational components .\",\n \"Arcing turn maneuvers were defined as sequences with a heading velocity > 90°/sec , a minimum total translational velocity > 0 . 5 m/s , a total distance traveled > 25 cm , and a vertical distance traveled < 10 cm .\",\n \"These search parameters reliably extract arcing turns that occur in the horizontal plane .\",\n \"To compare arcing turns of different shapes and scales we clipped the trajectories to a length of 25 cm centered at the sharpest point of the turn .\",\n \"From the clipped trajectory we analyzed three performance metrics , average velocity ( Arcvel , avg ) , radius ( Arcrad ) , and the maximum centripetal acceleration ( Arccent , max ) .\",\n \"The latter two were calculated using the following equations: Arcrad=Arcdistance traveledΔHeadingrad Arccent , max=Arcvel , avg2Arcrad Pitch-roll turn maneuvers have been described in hummingbirds and are characterized by the following sequence:\",\n \"a ) deceleration ,\",\n \"b ) increase in pitch to near vertical ,\",\n \"c ) azimuthal rotation by rolling the body , and\",\n \"d ) acceleration in a new direction ( Clark , 2011 ) .\",\n \"These maneuvers were identified by searching for sequences of deceleration followed by acceleration with a maximum pitch > 75° .\",\n \"Just as we did for the yaw turns , we assume that above a pitch angle of 75° , the rotation is primarily dominated by a body axis roll , even if there may be a slight yawing component .\",\n \"For this reason , we maintain the established 'pitch-roll' terminology to describe these types of turns .\",\n \"These sequences were clipped to a linear distance of 25 cm centered on the point of the lowest translational velocity .\",\n \"Only clipped sequences in which the total vertical displacement was less than 10 cm were analyzed .\",\n \"The performance metrics for pitch-roll turns were the time taken ( PRTtime ) and the degrees turned ( PRTdeg ) .\",\n \"Arcing turns and pitch-roll turns are two different mechanisms for generating a change in heading with no overlap in our data set by definition ( Table 1 ) .\",\n \"We analyzed how morphology , burst capacity , and competitor presence influenced the relative use of these two turns .\",\n \"The pitch-roll percent ( PRT% ) was defined as the number of pitch-roll turns divided by total the number of arcing and pitch-roll turns extracted from each trial .\",\n \"Descriptive statistics for morphology and load lifting are provided in Table 2 . A large sample of values was obtained for each maneuvering performance metric ( Table 3 ) .\",\n \"Figure 2 shows the distributions of trial means for all performance metrics . 10 . 7554/eLife . 11159 . 007Table 2 . Wing morphology and load lifting performance of male Anna’s hummingbirds ( n = 20 individuals ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 007TraitMeanRangeWing length50 . 97 mm[45 . 76 , 55 . 45]Wing area 1355 mm2[1051 , 1653]Wing aspect ratio7 . 73[7 . 13 , 8 . 46]Body mass4 . 64 g[4 . 09 , 5 . 61]Mass of weights lifted5 . 93 g[4 . 00 , 7 . 24]10 . 7554/eLife . 11159 . 008Table 3 . Descriptive statistics and sample sizes for maneuvering performance .\",\n \"Grand mean values were calculated by first taking the mean of each bird’s trial averages ( i . e . , the bird means ) , and then taking the mean of the bird means ( n = 20 birds in 20 solo trials and 16 paired competition trials ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 008ManeuverabilityPerformance metric# TrajectoriesGrand mean[Range of means]Linear accelerationsVelmax71 , 0072 . 22 m/s[1 . 20 , 2 . 94]AccHormax47 , 2876 . 30 m/s2[2 . 96 , 8 . 83]DecHormax51 , 2456 . 67 m/s2[9 . 03 , 3 . 45]AccVUmax6 , 9353 . 78 m/s2[2 . 98 , 4 . 67]AccVDmax9 , 2843 . 58 m/s2[4 . 69 , 2 . 68]Rotational velocitiesPitchUvel , avg6 , 0851 . 13 rev/s[0 . 91 , 1 . 34]PitchDvel , avg14 , 8071 . 00 rev/s[1 . 19 , 0 . 78]Yawvel , avg12 , 6601 . 52 rev/s[1 . 32 , 1 . 75]Complex turnsPitch-rollPRTdeg17 , 133133 . 3 º[34 . 9 , 162 . 7]PRTtime17 , 1330 . 47 s[0 . 38 , 0 . 60]ArcingArcrad6 . 9450 . 48 m[0 . 14 , 0 . 70]Arcvel , avg6 , 9451 . 57 m/s[0 . 80 , 2 . 26]Arccent , max6 , 9456 . 59 m/s2[3 . 42 , 10 . 80]Use of turnsPRT% 24 , 0780 . 69[0 . 39 , 0 . 87]10 . 7554/eLife . 11159 . 009Figure 2 . Distributions of mean performance metric values for n = 52 bird-trial combinations . Only PRTdeg has statistically significant outliers .\",\n \"See Figure 2—figure supplement 1 for distributions of residuals from the best-fit model in each case .\",\n \"Note that two statistical outliers were omitted from the analysis of PRTdeg . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 00910 . 7554/eLife . 11159 . 010Figure 2—figure supplement 1 . Distributions of residuals from the best-fit model for each performance metric . Note that two statistical outliers were omitted from the analysis of PRTdeg . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 010 All performance metrics based on total and horizontal linear accelerations and complex turns were highly repeatable , with >80% of the variation in these metrics attributable to differences among individuals ( Figure 3 ) .\",\n \"The rotational performance metrics and the percent of turns that were pitch-roll turns were moderately repeatable , with 40–70% of the variation in these metrics attributable to among-individual differences .\",\n \"The vertical accelerations were not repeatable , as the 95% confidence intervals for repeatability of these metrics overlapped zero . 10 . 7554/eLife . 11159 . 011Figure 3 . Most maneuvering performance metrics are highly repeatable . Values > 70% are considered to have high repeatability , 40–70% moderate repeatability , and < 40% low repeatability .\",\n \"A metric is considered not repeatable if its 95% confidence intervals overlap zero . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 011 The best-supported models for each maneuvering performance metric are given in Table 4 .\",\n \"Burst muscle capacity was an important predictor for most of the maneuvering performance metrics .\",\n \"Birds that lifted more weight ( accounting for their wing morphology ) tended to accelerate and decelerate faster , and they tended to perform maneuvers with higher velocity ( Figure 4 ) .\",\n \"However , burst muscle capacity was not an important determinant of vertical acceleration and deceleration , as candidate models including burst performance as a predictor were not supported .\",\n \"Birds that lifted more weight also executed pitch-up and pitch-down maneuvers with higher rotational velocities .\",\n \"Burst capacity was not a strong determinant of yaw performance .\",\n \"Although yaw velocity was somewhat positively related to burst capacity ( Figure 4 ) , candidate models of yaw velocity that included burst as a predictor were not well supported . 10 . 7554/eLife . 11159 . 012Table 4 . Maneuvering performance in relation to burst performance , wing morphology , and competitor presence ( n = 20 birds in 20 solo trials and 16 paired competition trials ) .\",\n \"Standardized beta coefficients and R2GLMM ( m ) values are reported for either the best-fit model , or , if there was support for more than one model , the average of supported models .\",\n \"The standardized beta coefficient is a measure of effect size that can be compared among predictors in the same model .\",\n \"Relative importance is a measure of the weight of evidence in favor of a predictor on a scale from 0–1 , and is reported for burst capacity and wing morphology variables as these alone were subject to model selection .\",\n \"Marginal R2GLMM ( m ) provides a measure of the combined explanatory power of fixed effects of interest ( competitor presence , burst muscle capacity , and wing morphology effects combined ) .\",\n \"Details of all candidate models are provided in Supplementary file 1 . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 012ModelSupport forFixed effectsStd beta coef [95% CI]Relative importanceR2GLMM ( m ) Burst+ morphology+ competitorVelmaxburstcompetitor presencemassburstwing lengthwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) days post-capture–0 . 04 [–0 . 18 , 0 . 10]0 . 10 [–0 . 01 , 0 . 22]0 . 09 [0 . 00 , 0 . 18]–0 . 08 [–0 . 22 , 0 . 06]0 . 10 [–0 . 07 , 0 . 28]1 . 01 [0 . 59 , 1 . 42]1 . 06 [0 . 68 , 1 . 43]–0 . 07 [–0 . 24 , 0 . 11]----1 . 000 . 250 . 26------0 . 28AccHormaxburst + competitioncompetitor presencemassburstexperiment ( CA1 ) experiment ( CA2 ) days post-capture–0 . 46 [–0 . 82 , –0 . 11]0 . 20 [–0 . 28 , 0 . 69]0 . 39 [0 . 00 , 0 . 77]4 . 01 [2 . 46 , 5 . 56]3 . 68 [2 . 72 , 4 . 64]–0 . 39 [–1 . 09 , 0 . 32]----1 . 00------0 . 18DecHormaxburst + competitioncompetitor presencemassburstexperiment ( CA1 ) experiment ( CA2 ) days post-capture–0 . 47 [–0 . 78 , –0 . 16]0 . 31 [–0 . 13 , 0 . 74]0 . 41 [0 . 06 , 0 . 76]3 . 86 [2 . 47 , 5 . 25]3 . 64 [2 . 76 , 4 . 51]–0 . 24 [–0 . 88 , 0 . 39]----1 . 00------0 . 19AccVUmaxintercept-onlyNANANA0 ( NA ) AccVDmaxintercept-onlyNANANA0 ( NA ) PitchUvel , avgburstcompetitor presencemassburstexperiment ( CA1 ) experiment ( CA2 ) 0 . 02 [–0 . 02 , 0 . 06]0 . 00 [–0 . 04 , 0 . 04]0 . 03 [–0 . 01 , 0 . 07]0 . 14 [0 . 06 , 0 . 23]0 . 13 [0 . 03 , 0 . 23]----1 . 00----0 . 10PitchDvel , avgcompetition+ burstcompetitor presencemassburstwing lengthwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) 0 . 06 [0 . 01 , 0 . 10]0 . 01 [–0 . 04 , 0 . 05]0 . 03 [–0 . 01 , 0 . 08]0 . 04 [–0 . 03 , 0 . 12]–0 . 04 [–0 . 13 , 0 . 05]0 . 19 [0 . 03 , 0 . 34]0 . 22 [0 . 03 , 0 . 41]----0 . 660 . 370 . 28----0 . 18Yawvel , avgintercept-onlyNANANA0 ( NA ) PRTdegintercept-onlyNANANA0 ( NA ) PRTtimeburstcompetitor presencemassburstwing lengthwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) 0 . 00 [–0 . 01 , 0 . 01]–0 . 01 [–0 . 03 , 0 . 00]–0 . 02 [–0 . 03 , 0 . 00]–0 . 01 [–0 . 03 , 0 . 01]0 . 01 [–0 . 01 , 0 . 04]–0 . 08 [–0 . 12 , –0 . 03]–0 . 11 [–0 . 16 , –0 . 05]–--1 . 000 . 210 . 23–--0 . 29Arcradburstcompetitor presencemassburstwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) –0 . 02 [–0 . 07 , 0 . 03]0 . 01 [–0 . 03 , 0 . 06]0 . 06 [0 . 01 , 0 . 10]–0 . 06 [–0 . 15 , 0 . 03]0 . 25 [0 . 12 , 0 . 37]0 . 29 [0 . 06 , 0 . 52]----1 . 000 . 44–--0 . 22Arcvel , avgburstcompetitor presencemassburstexperiment ( CA1 ) experiment ( CA2 ) days post-capture–0 . 01 [–0 . 09 , 0 . 08]0 . 03 [–0 . 06 , 0 . 12]0 . 11 [0 . 04 , 0 . 19]0 . 89 [0 . 59 , 1 . 19]0 . 74 [0 . 56 , 0 . 92]–0 . 06 [–0 . 20 , 0 . 08]––1 . 00–----0 . 18Acccent , maxwing shapecompetitor presencemasswing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) days post-capture0 . 29 [–0 . 37 , 0 . 94]–0 . 20 [–0 . 74 , 0 . 34]1 . 09 [0 . 19 , 1 . 99]5 . 93 [4 . 02 , 7 . 84]0 . 85 [–1 . 59 , 3 . 28]–1 . 76 [–2 . 62 , –0 . 90]----1 . 00------0 . 36PRT% wing shape + competition+ burst + wing sizecompetitor presencemassburstwing lengthwing aspect ratioexperiment ( CA1 ) experiment ( CA2 ) –0 . 14 [–0 . 19 , –0 . 09]0 . 00 [–0 . 04 , 0 . 05]0 . 04 [0 . 00 , 0 . 09]–0 . 06 [–0 . 13 , 0 . 01]–0 . 16 [–0 . 24 , –0 . 07]0 . 17 [–0 . 03 , 0 . 36]0 . 44 [0 . 19 , 0 . 69]----1 . 000 . 611 . 00----0 . 2710 . 7554/eLife . 11159 . 013Figure 4 . Burst muscle capacity was associated with most maneuvering performance metrics . Each panel shows partial residuals for a performance metric ( y-axis ) in relation to burst muscle capacity ( x-axis ) for the most supported candidate model with burst capacity as a predictor .\",\n \"Partial residual values ( y-axis ) account for the other fixed effects in that model .\",\n \"Lines show model predictions assuming the median value of continuous predictors , and averaging across experiments and levels of competitor presence .\",\n \"Prediction lines are dashed for metrics where burst performance was not present in any of the supported models .\",\n \"Color is used to denote data points from the same bird ( online version only ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 013 Burst muscle capacity was also associated with some , but not all maneuvering performance metrics related to complex turns .\",\n \"Birds that lifted more weight for their wing morphology tended to execute faster , larger radius arcing turns ( Figure 4 ) .\",\n \"However , the centripetal acceleration of arcing turns was not associated with burst capacity .\",\n \"Hummingbirds with higher load lifting capacity executed pitch-roll turns in less time .\",\n \"Burst capacity was not a strong determinant of heading change during pitch-roll turns .\",\n \"Lastly , birds with higher burst muscle capacity used pitch-roll turns for proportionately more of their heading changes .\",\n \"Wing morphology , specifically the aspect ratio , was an important predictor for two performance metrics: centripetal acceleration and the percent of direction changes that were pitch-roll turns ( Figure 5 ) .\",\n \"Hummingbirds with long , narrow wings tended to perform arcing turns with higher centripetal accelerations , relative to birds with short , wide wings .\",\n \"Birds with higher aspect ratio wings also used proportionately more arcing turns than birds with low aspect ratio wings . 10 . 7554/eLife . 11159 . 014Figure 5 . Aspect ratio was associated with two maneuvering performance metrics . Each panel shows partial residual performance ( y-axis ) in relation to wing aspect ratio ( x-axis ) from a best-fit model that identified aspect ratio as an important predictor .\",\n \"Note that the partial residuals for PRT% in\",\n \"( b ) go above 1 because PRT% was modeled as a normally-distributed ( Gaussian ) variable .\",\n \"All other features as in Figure 4 . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 014 Body mass was included in candidate models 1–7 because we had anticipated that body mass would have a strong influence on variation in maneuvering performance .\",\n \"However , for every performance metric in Table 4 , the coefficient estimate for body mass had confidence intervals that broadly overlapped zero .\",\n \"We did not detect a substantial effect of competitor presence on many of the performance metrics ( Table 4 ) .\",\n \"Two metrics , horizontal acceleration and deceleration , were affected , but in the direction opposite to what we predicted .\",\n \"Specifically , birds performed maneuvers with lower acceleration ( –0 . 46 m/s2 difference on average ) and lower deceleration ( –0 . 47 m/s2 ) in the presence of a competitor , relative to solo flight ( Figure 6a , b ) .\",\n \"One metric , pitch-down velocity ( Figure 6c ) , did increase during competition as predicted ( 0 . 06 rev/s difference on average ) .\",\n \"We had no prediction for how competition would influence the relative use of pitch-roll and arcing turns , but found that birds used proportionately more arcing turns in the presence of a competitor ( Figure 6d ) .\",\n \"Specifically , 35% of direction changes were arcing turns on average ( and 65% pitch-roll ) when a competitor was present , whereas during solo flight , only 23% of direction changes were arcing turns ( and 77% pitch-roll ) on average . 10 . 7554/eLife . 11159 . 015Figure 6 . Competitor presence was associated with four maneuvering performance metrics . Each panel shows residual performance ( y-axis ) in relation to competitor presence from a best-fit model where competitor presence had a detected effect .\",\n \"All other features as in Figure 4 . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 015\"\n ],\n [\n \"We collected a large number of free flight measurements for each of 20 individual hummingbirds to examine the biomechanical determinants of maneuverability .\",\n \"Other studies have measured elements of maneuvering performance of hummingbirds in the field ( Clark , 2009; Sholtis et al . , 2015 ) and documented the maximum velocities , accelerations , and rotations obtained during specific maneuvers .\",\n \"Our values for velocity and acceleration are considerably lower than either of the field studies , likely because of cage size .\",\n \"However , the benefit of using a flight chamber is that it allowed us to evaluate the relative contributions of different factors to the performance we observed .\",\n \"We found that hummingbirds maneuvered with highly repeatable performance ( Figure 3 ) .\",\n \"Maximum weight lifted during load lifting trials predicted most of the performance metrics that we measured , independent of a bird’s wing size and shape , such that birds with higher burst muscle capacity flew faster , had higher horizontal accelerations , faster rotations , and higher performance during complex turns ( Figure 4 ) .\",\n \"Aspect ratio predicted only two performance metrics , such that birds with higher aspect ratio wings performed turns with higher centripetal acceleration and a greater percentage of arcing turns ( Figure 5 ) .\",\n \"When flying in the presence of a competitor , hummingbirds used faster pitch velocities , although they used slower horizontal accelerations and decelerations .\",\n \"During competition trials birds also increased the proportion of arcing turns used ( Figure 6 ) .\",\n \"Collectively , these results suggest that burst muscle capacity is a much more important predictor of flight maneuverability than within-species variation in body mass , wing morphology , and competition with conspecifics .\",\n \"Why were body mass and wing size not associated with maneuvering performance ?\",\n \"Wing morphology has well-known physical affects on flight performance: aspect ratio predicts aerodynamic efficiency , wing area is directly proportional to aerodynamic force , and wing length is a strong predictor of wingbeat frequency .\",\n \"All of these morphological traits , along with body mass , could affect maneuverability in flight , either individually or in combination .\",\n \"For example , wing loading ( the ratio of body mass to wing area or to area swept by the wings ) was initially thought to be a key predictor of hummingbird flight performance and behavioral ecology ( Feinsinger and Chaplin , 1975; Feinsinger , 1976; Feinsinger and Colwell , 1978; Feinsinger et al . , 1979 ) .\",\n \"However , in our analysis the hypothesis that wing size and body mass together determine maneuvering performance was not supported for any performance metric ( see Supplementary file 1 ) .\",\n \"We found it especially surprising that only wing shape ( and not wing size ) predicted maneuvering performance .\",\n \"It is possible that other morphological traits may determine maneuvering performance , or that subtle relationships may have gone undetected , because our analysis was limited to 20 individuals of a single species .\",\n \"It would be informative to expand this analysis to other species with potentially greater within-species variation in wing morphology , and to assess maneuverability across different hummingbird species with divergent morphologies .\",\n \"Almost all of the performance metrics were highly repeatable , which indicates a potential role for intrinsic influences of wing morphology in determining maneuverability .\",\n \"However , aspect ratio was the only morphological parameter that predicted performance , and only for a limited set of maneuvers .\",\n \"Aspect ratio is a key determinant in wing efficiency for fixed wings , such as during gliding ( Pennycuick , 1983 ) , and it has recently been demonstrated that higher aspect ratio wings correspond to higher power factors in the revolving wings of hummingbirds ( Kruyt et al . , 2014 ) .\",\n \"We found that aspect ratio had a strong effect on the few performance metrics that it predicted , but did not affect most features of maneuvering performance .\",\n \"This suggests a limited role for aerodynamic efficiency in many features of maneuvering .\",\n \"Burst muscle capacity predicted most of the performance metrics we considered , independently of any association with wing size or shape .\",\n \"Load lifting is measured as a transient escape maneuver that is likely anaerobic and performed inefficiently .\",\n \"All hummingbirds reach maximum load lifting performance at a geometric limit set by the amplitude of the wings: wing stroke amplitude cannot extend much past 180° without the two wings interfering with each other physically and aerodynamically ( Chai and Dudley , 1995; Chai et al . , 1997; Chai and Millard , 1997; Altshuler and Dudley , 2003 ) .\",\n \"Maximum load lifting also elicits a substantial increase in wingbeat frequency as a constant fraction of baseline wingbeat frequency ( Altshuler and Dudley , 2003 ) .\",\n \"Thus , maximum load lifting performance involves brief increases in muscle strain and muscle velocity to physically imposed limits .\",\n \"The capacity to increase muscle strain and velocity has previously been shown to influence foraging behavior and competitive ability ( Altshuler , 2006 ) .\",\n \"The results of the current study demonstrate that it also underlies multiple features of maneuvering performance .\",\n \"The two performance metrics that were not repeatable are vertical accelerations and decelerations , which were expected to be important based on previous observations of hummingbird competitive interactions ( Altshuler , 2006 ) and mating displays ( Clark , 2009 ) .\",\n \"Moreover , vertical performance was not well predicted by morphology , burst capacity , or competitor presence in this study .\",\n \"The dimensions of our experimental chamber likely influenced our observations of vertical performance .\",\n \"Hummingbirds in captivity tend to fly near the top of their cages , and the vertical dimension of the chamber ( 1 . 5 m ) may have limited vertical movement .\",\n \"Male hummingbirds are extremely aggressive towards conspecifics ( Kodric-Brown and Brown , 1978; Carpenter et al . , 1983 ) and other species of hummingbirds ( Stiles and Wolf , 1970; Wolf et al . , 1976 ) .\",\n \"The most territorial species will vigorously defend territories ( Carpenter et al . , 1983 ) and lekking sites ( Rico-Guevara and Araya-Salas , 2015 ) .\",\n \"In staged competition studies , paired hummingbirds will also establish and defend territories ( Tiebout , 1993 ) .\",\n \"We originally intended to use competition to elicit high levels of flight activity and maneuvering performance in territorial male Anna's hummingbirds ( Stiles , 1982 ) .\",\n \"However , we found that competitor presence affected only a small number of the maneuvering performance metrics that we measured .\",\n \"Pitch-down velocity increased with competition whereas horizontal acceleration and deceleration actually decreased .\",\n \"We do not know why these three metrics ( in addition to PRT%; see below ) were strongly affected by competition or why they were affected in the directions observed .\",\n \"However , there are several possible causes for why competitor presence did not affect the other metrics:\",\n \"1 ) we were unable to elicit a high level of competition or territoriality;\",\n \"2 ) the birds may have worked out dominance without the aggressive interactions normally seen outdoors; and/or\",\n \"3 ) the interactions required to establish dominance may have been very brief ( Maynard Smith , 1974 ) such that they comprised only a minuscule sample of the maneuvers we analyzed .\",\n \"This experiment was not designed to study the effects of maneuvering performance on competitive success , although this represents an important topic for future investigation .\",\n \"Laboratory performance tests do not always reflect field behavior ( Irschick , 2003 ) and outdoor studies of maneuvering performance will be important for understanding the role of maneuverability in competitive interactions .\",\n \"Recent advances in video tracking ( Theriault et al . , 2014; Shelton et al . , 2014 ) should make it possible to track individuals for multiple measurements .\",\n \"The most substantial result of competitor presence was the increase in the use of arcing over pitch-roll turns .\",\n \"These two types of turns represent different strategies for changing direction that differ in duration and amount of heading change .\",\n \"Arcing turns require less time but are used for smaller heading changes , whereas pitch-roll turns are longer but can be used to change heading by 180° ( Figure 7 ) .\",\n \"Given that hummingbird agonistic interactions can involve direct contact and stabbing with bills ( Tiebout , 1993; Clark and Russell , 2012; Rico-Guevara and Araya-Salas , 2015 ) , slow turns in place could make a bird more vulnerable during competition . 10 . 7554/eLife . 11159 . 016Figure 7 . Arcing and pitch-roll turns are two classes of complex maneuver that differ in turn magnitude and duration . Representative examples of arcing\",\n \"( a ) and pitch-roll\",\n \"( b ) turns are depicted from the above perspective .\",\n \"Arcing turns ( Arc; orange ) and pitch-roll turns ( PRT; green ) differed in the degrees turned\",\n \"( c ) and elapsed time\",\n \"( d ) .\",\n \"Circles represent bird-trial means ( n = 52 ) with grand means indicated with black lines .\",\n \"Histograms for the pooled dataset of all maneuvers are given on the right .\",\n \"The outliers for degrees turned in pitch-roll turns were included when calculating the grand means but not in the model analyses ( Table 4 ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 016 The relative use of arcing and pitch-roll turns was the only metric in our study that was influenced by all of morphology , burst muscle capacity , and competitor presence .\",\n \"The minimum radius of an arcing turn is limited by the maximum centripetal acceleration that a bird can generate while maintaining lift .\",\n \"The speed of a pitch-roll turn is limited by the ability to decelerate and then accelerate .\",\n \"Birds with higher wing aspect ratio may have preferred arcing turns because they were able to generate higher centripetal accelerations .\",\n \"Birds with higher burst muscle capacity may have favored pitch-roll turns because they had higher accelerating and decelerating performance .\",\n \"These observations suggest the hypothesis that high aspect ratio and high burst capacity enhance maneuverability .\",\n \"This hypothesis could be evaluated by comparing hummingbird species that differ in wing shape , foraging strategy , and burst capacity ( Altshuler et al . , 2004b; 2010a; Altshuler , 2006; Kruyt et al . , 2014 ) .\",\n \"By constraining hummingbirds to fly in a large chamber we were able to track and measure a large sample of maneuvers attributed to individuals with known morphological traits and burst performance .\",\n \"A major contribution of our study is the development of an assay of free flight maneuvering performance based on large numbers of stereotyped movements .\",\n \"Using this method , we identify several performance metrics that were highly repeatable across trials for individual hummingbirds , strongly correlated with individual morphological and physiological characteristics , and largely uninfluenced by the added motivation of a conspecific competitor .\",\n \"This approach to measuring maneuverability will be useful for future studies comparing maneuvering performance across different experimental manipulations , geographic ranges , or ecological , morphological , and phylogenetic groups .\"\n ],\n [\n \"We captured and filmed 20 adult male Anna's hummingbirds ( Calypte anna ) at the University of California , Riverside ( eight birds in July-October 2009; four birds in January-March 2010 ) and the University of British Columbia ( eight birds in December 2013-April 2014 ) .\",\n \"The hummingbirds were housed in individual cages and fed ad libitum with a solution of artificial nectar ( Nektar-Plus , Nekton , Pforzheim , Germany ) and sucrose .\",\n \"The flight arenas were large rectangular cages ( 3 x 1 . 5 x 1 . 5\",\n \"m ) built with an aluminum frame and had either garden mesh ( California ) or clear acrylic ( British Columbia ) side panels .\",\n \"The cages contained multiple perches and a single feeder hung from the roof of the cage .\",\n \"Before the first trial , each bird was allowed to acclimate to the flight arena and learn where the perches and the feeder were located .\",\n \"The trials began once the birds were actively exploring the cage and consistently visiting the feeder and both perches .\",\n \"At this point , we recorded high-speed video of a two-hour solo trial for each bird .\",\n \"Following solo trials ( between 0–23 days later ) , birds were paired and filmed for another two hours in competition trials .\",\n \"One bird in each pair was marked with a small square of retro-reflective tape placed between the shoulder blades for identification .\",\n \"The birds filmed in British Columbia had one competition trial and the birds in California had two competition trials .\",\n \"In the latter case , the second trial consisted of previously unknown opponents that were chosen randomly from the remaining pool .\",\n \"The competition trials involved chases , displacements , and aerial displays , but very little contact .\",\n \"Regardless , we monitored the competition trials to ensure that no birds were harmed or excluded from the feeder .\",\n \"Following each round of solo and competition trials , we performed asymptotic load lifting experiments using the techniques described in Chai et al . ( Chai et al . , 1997 ) , and subsequently used in other studies estimating maximum burst power output ( Chai and Millard , 1997; Altshuler et al . , 2004a; 2010b; Altshuler , 2006 ) .\",\n \"Here , we use the mass of maximum number of beads lifted by each individual as a measure of burst performance .\",\n \"Immediately following load lifting , we weighed the birds and photographed both wings in an outstretched position against white paper with a reference scale ( Chai and Dudley , 1995 ) .\",\n \"We oriented the wing image and divided it into pixel wide strips representing the wing chords at each value of wing radius .\",\n \"Values for aspect ratio , wing area , and wing length were then calculated based on equations in Ellington ( Ellington , 1984 ) .\",\n \"We considered wing area and wing length as two potential measures of wing size , but these traits were highly correlated in our dataset ( R2 = 0 . 85 , p < 0 . 0001 , n = 20 ) .\",\n \"Because these two traits did not vary independently in our relatively small sample of 20 hummingbirds , we could not consider them independently .\",\n \"We therefore selected wing length as the more robust measure of wing size , because unlike area , wing length is less prone to measurement error as a result of variation in feather overlap when wings are positioned for measurements .\",\n \"We verified that our results were consistent when using wing area instead of length , and thus these two traits should be considered interchangeable as measures of wing size in this study .\",\n \"Because both wing morphology and muscle capacity may influence burst performance , we used burst performance controlled statistically for wing morphology as a measure of burst muscle capacity .\",\n \"Further details are provided below in the Statistical analysis section .\",\n \"All procedures were conducted under approval of the Institutional Animal Care and Use Committee at the University of California , Riverside and the Animal Care Committee at the University of British Columbia .\",\n \"We used an automated tracking system to measure both body position and orientation of flying birds in three dimensions ( Video 1 ) .\",\n \"A complete description of the tracking algorithm and hardware components is provided in ( Straw et al . , 2011 ) .\",\n \"The core algorithms were written in Python ( Python Software Foundation , 2012 ) , and are available via github ( PyMVG: https://github . com/strawlab/pymvg; adskalman: https://github . com/astraw/adskalman; MultiCamSelfCal: https://github . com/strawlab/MultiCamSelfCal/ ) .\",\n \"We adapted this system for recording hummingbird solo and competitive flight trajectories with four or five digital cameras ( GE680 , Allied Vision Technologies , Burnaby , Canada ) .\",\n \"The cameras were mounted on the ceiling and recorded at 640 x 480 pixel resolution at 200 frames per second ( Figure 1a ) .\",\n \"We calibrated the filming volume by moving a single light-emitting diode throughout the arena to acquire data for an automated self calibration algorithm ( Svoboda et al . , 2005 ) .\",\n \"This algorithm provides a relative calibration ( non-linear warping distortion parameters and 3x4 camera calibration matrices ) across all cameras .\",\n \"This calibration is brought into absolute terms ( the scale , rotation , and translation are found ) by matching a manually measured 3D model of the flight arena with reconstructed image coordinates using the ‘estsimt’ function of the MultiCamSelfCal toolbox ( Svoboda et al . , 2005 ) .\",\n \"To minimize the effect of errors in the 3D tracking , we used a forward/reverse non-causal Kalman filter ( Rauch–Tung–Striebel smoother ) applied to the online state estimate of position and velocity from the realtime Kalman filter .\",\n \"The smoothing parameters were chosen so that seven traces of a tracked , falling object yielded an average peak acceleration of 9 . 8 m/s2 .\",\n \"The process covariance matrix we used is: Qpos=σ2×T3300T22000T3300T22000T3300T22T2200T000T2200T000T2200T where σ2 is 0 . 01 and T is the interval between frames ( 0 . 005 s ) .\",\n \"The observation covariance matrix we used is: Rpos=0 . 0001440000 . 0001440000 . 000144 Figure 1c and d show examples of two trajectories with plots of the unsmoothed data , the data smoothed with Qpos and Rpos , and the effects of two different smoothing parameters ( Rpos x 10 , Rposx 0 . 1 ) .\",\n \"Following establishment of the 3D trajectories , the tracking system assigned 3D body orientation vectors to each bird in each frame based on 2D estimates of the long axis of the body .\",\n \"Body orientation was estimated using an algorithm that fit orientations to the body axis in each 2D image .\",\n \"Each sequence of five consecutive images cropped around the bird was aligned at the optical center of intensity .\",\n \"Averaging these images effectively eliminated the wings and emphasized the body .\",\n \"Orientation was estimated by calculating the covariance matrix of the image luminance and then computing the eigensystem of this covariance matrix .\",\n \"The eigenvector associated with the largest eigenvalue was taken as the orientation .\",\n \"Orientation vector assignments were also smoothed with a Kalman filter using more restrictive smoothing parameters than were used to smooth body position .\",\n \"To determine appropriate smoothing parameters we replotted the smoothed body orientation vectors onto a sample of videos , and visually chose the ones that provided the best fit .\",\n \"The process covariance matrix used for body orientation ( Qori ) is the same as the process covariance matrix used for body position ( Qpos ) and the observation covariance matrix used is: Rori=0 . 000001440000 . 000001440000 . 00000144 Once the body orientations were calculated we used a dynamic programming algorithm to decide which end of the vector was the head and which end was the tail .\",\n \"The direction of the head was chosen based on the direction of the previous orientation , the direction of travel , and the vertical up direction .\",\n \"For each frame ( n ) , the 'cost' associated with the two possible orientations ( Ori→ , -Ori→ ) were calculated: CostOri=Speed ( Ori→n⋅Vel→modOri→nVel→mod ) + ( 1−Speed ) ( Ori→n⋅Up→Ori→nUp→ ) + ( 1−Speed ) ( Ori→n⋅Ori→n−1Ori→nOri→n−1 ) Cost−Ori=Speed ( −Ori→n⋅Vel→modOri→nVel→mod ) + ( 1−Speed ) ( −Ori→n⋅Up→Ori→nUp→ ) + ( 1−Speed ) ( −Ori→n⋅Ori→n−1Ori→nOri→n−1 ) where Ori→ is the body vector , Vel→mod is the modified velocity vector tipped up 15º towards the vertical direction .\",\n \"Up→ is the vertical direction vector , Ori→n−1 is the orientation during the previous frame , and if the magnitude of the velocity is greater than 0 . 5m/s: Speed=Vel→ otherwise: Speed=0 . 5 This approach accounted for the tendency of hummingbirds to fly forwards and with an upright posture , but allowed for exceptions in the case of backwards flight , inversions , and dives , particularly if these occurred at low speeds .\",\n \"The magnitudes of calculated accelerations and , to a lesser extent , velocities derived from position data were influenced by the specific smoothing parameters .\",\n \"Examples of maneuvers with different smoothing parameters and their effects on the calculated performance metrics are given in Figure 1c and d .\",\n \"This influence of smoothing parameters is a well known limitation of video tracking ( Walker , 1998 ) .\",\n \"Thus , although acceleration values are comparable within a study , caution must be applied when comparing the magnitude of acceleration values among studies differing in camera frame rate , filming volume , calibrations , and smoothing parameters ( Walker , 1998 ) .\",\n \"For our final performance metrics we used instantaneous body orientation and orientation velocity , but not orientation acceleration .\",\n \"The automated tracking system extracted the 3D coordinates of multiple flying animals and saved each trajectory as a separate object ( Video 1 ) .\",\n \"An object began when the tracking system detected new movement and ended when either the object stopped moving , the error in the 3D reprojection grew too large , or multiple objects came within 2 cm of each other .\",\n \"In our experiments tracking hummingbird flight led to two problems in determining distinct objects .\",\n \"The first is that very stable hovering can be misidentified as perching .\",\n \"For example , as a bird went into an extended hovering bout , such as at a feeder , the tracking system detected the cessation of movement and ended the trajectory .\",\n \"Conversely , when the bird perched at the end of a flight or in between two flights , especially if it continued to move its head or fluff its feathers , the tracking system treated the bird as moving and continued the trajectory .\",\n \"Because our study focused on identifying and analyzing relatively long , moving trajectories , these types of errors did not cause problems .\",\n \"The second challenge concerned identification of birds during close encounters in competition trials .\",\n \"When two tracked objects became close to each other , even if they did not physically touch , the tracking system could not accurately distinguish them .\",\n \"We used a conservative solution and terminated the trajectories whenever two birds came close enough that the tracked objects merged .\",\n \"Birds were later identified manually by a team of digitizers who viewed the videos and assigned each object number to either the marked or unmarked bird .\",\n \"The automated digitization produced a small number of extreme tracking errors , which we did not want to unduly influence statistical analyses .\",\n \"We accordingly removed values >5 SDs more extreme than the mean for each performance metric .\",\n \"The trimmed values comprised only 0–0 . 31% of the original pooled sample size for each metric .\",\n \"We next calculated the mean of each performance metric for each bird-trial combination ( n= 52 means; 20 birds in 20 solo trials and 16 paired competition trials ) .\",\n \"All statistical analyses were performed on these bird-trial means using R 3 . 1 . 1 ( R Development Core Team , 2014 ) , and the data used for the analysis are available online ( Segre et al . , 2015 ) .\",\n \"Repeatability , or the intra-class correlation coefficient ( ICC ) , is defined as the proportion of variation that is attributable to differences among individuals ( Nakagawa and Schielzeth , 2010 ) .\",\n \"We estimated repeatability for each performance metric from an intercept-only mixed effects model that included estimates of the population intercept ( i . e . , the grand mean ) as well as an individual intercept for each bird ( Nakagawa and Schielzeth , 2010 ) .\",\n \"Such a model has two variance components , the variance of the random intercept values ( variance among individuals ) and a residual variance associated with the error term .\",\n \"Repeatability is the variance among individuals divided by the total variance ( Nakagawa and Schielzeth , 2010 ) .\",\n \"We used parametric bootstrapping with 5000 iterations to obtain confidence intervals for these repeatability estimates via the bootMer function in the lme4 ( v1 . 1 . 7 ) package .\",\n \"Because our second question involved evaluating several possible scenarios for the influence of morphology and burst performance on maneuverability , we used an information-theoretic approach to multi-model inference ( Burnham and Anderson , 2010 ) .\",\n \"Unlike dichotomous null hypothesis testing , this approach quantifies support for multiple hypotheses , and it avoids the problem of eliminating potentially important predictors when two or more alternative models are equally well supported .\",\n \"The output for interpretation includes the effect size and relative importance of each predictor , and there are no null hypotheses or P values associated with this approach .\",\n \"As a measure of effect size we report the standardized partial regression coefficient , std β , for each predictor , which can be used to compare their independent associations with a given response variable .\",\n \"Unstandardized regression coefficients corresponding to units of the predictor variables are provided in Supplementary file 1 .\",\n \"We also examined associations between burst performance , wing size , and wing shape because our load lifting assay may have incorporated effects of wing morphology as well as muscle capacity .\",\n \"The mass of weights lifted during load lifting was not significantly associated with wing length in our sample of 20 individuals ( p = 0 . 87 ) , however , it was negatively associated with wing aspect ratio ( p = 0 . 04 ) controlling for site .\",\n \"Thus in our model analyses we used residual burst performance controlling for wing aspect ratio and site as a measure of burst muscle capacity independent of a bird’s wing morphology .\",\n \"We considered eight candidate mixed-effects models that could plausibly explain variation in each maneuvering performance metric ( Table 5 ) .\",\n \"All candidate models included an individual intercept for each bird ( the random intercept term ) and were fit using the nlme ( v 3 . 1–117 ) package ( Zuur et al . , 2009 ) .\",\n \"The intercept-only model included an estimate of the population intercept ( grand mean ) and random intercept terms , but no fixed effects .\",\n \"Other candidate models are listed in Table 5 .\",\n \"All models except the intercept-only model included the fixed effects of competitor presence , body mass , and experiment to account for potential effects of these factors .\",\n \"Experiment had three levels , one for each round of trials ( California 2009 , 2010 , British Columbia 2014 ) to account for differences such as location , time of year , and filming conditions . 10 . 7554/eLife . 11159 . 017Table 5 . Candidate models of maneuvering performance .\",\n \"All models include an intercept as well as a random effect of bird identity to account for repeated measures of individuals . DOI: http://dx . doi . org/10 . 7554/eLife . 11159 . 017ModelFixed effectsDescription1 . Solo/comp + experiment + body mass + wing lengthWing size2 . Solo/comp + experiment + body mass + wing aspect ratioWing shape3 . Solo/comp + experiment + body mass + wing length + wing aspect ratioWing size & shape4 . Solo/comp + experiment + body mass + weight liftedBurst power5 . Solo/comp + experiment + body mass + weight lifted + wing lengthBurst power & wing size6 . Solo/comp + experiment + body mass + weight lifted + wing aspect ratioBurst power & wing shape7 . Solo/comp + experiment + body mass + weight lifted + wing length + wing aspect ratioBurst power , wing size & shape8 . Intercept-only*Candidate models 1-7 also include a fixed effect of days post-capture for the following metrics: Velmax , AccHormax , DecHormax , Arcvel , avg , and Arccent , max Two issues arose in the preliminary examination of data .\",\n \"The first issue was that five of the performance metrics were significantly influenced by the number of days a bird had been in captivity .\",\n \"We therefore included an additional fixed effect of the number of days since capture when analyzing these five metrics ( Table 5 ) .\",\n \"The second issue was that one of the metrics , the heading change in pitch-roll turns ( PRTdeg ) , had three values that were significant outliers ( Grubb’s test , all G > 3 . 09 , all p<0 . 03; Figure 7 ) .\",\n \"We determined that these three statistical outliers were not errors in the tracking system but were instead derived from one individual that used pitch-roll turns to make small heading changes , unlike the other birds .\",\n \"We omitted these outliers from the analysis of heading change in pitch-roll turns to ensure that all fitted Gaussian models met the required assumptions , with no other outliers or problems of skew or heteroskedasticity .\",\n \"The best-fit model for heading change in pitch-roll turns was the intercept-only model regardless of whether the outliers were included .\",\n \"To quantify the variance explained by the fixed effects of interest in each model , we calculated the marginal R2GLMM\",\n \"( m ) using the r . squaredGLMM function in the MuMIn ( v1 . 10 . 5 ) package ( Nakagawa and Schielzeth , 2013 ) .\",\n \"This measure does not have all the properties of a traditional coefficient of determination , but like R2 it ranges from 0 to 1 , and it is an appropriate estimate of the variance explained by the fixed effects in a mixed model .\",\n \"We removed the effect of experiment and the number of days post-capture when calculating R2GLMM\",\n \"( m ) , because these were not effects of interest .\",\n \"Thus , R2GLMM\",\n \"( m ) provides a measure of the variance explained by the other supported fixed effect variables .\",\n \"We evaluated the support for different models using the Akaike information criterion ( AICc ) adjusted for small sample sizes .\",\n \"This was calculated using the MuMIn ( v 1 . 10 . 5 ) package with maximum likelihood estimation .\",\n \"We defined the group of supported models as those with a difference in AICc < 2 from the best-fit model for each performance metric .\",\n \"If no other models came within 2 AICc units of the best-fit model , we present effect size measures , their confidence intervals , and R2GLMM\",\n \"( m ) for only that model .\",\n \"Otherwise , we present averages of all supported models .\",\n \"Details of all candidate models are provided in Supplementary file 1 .\",\n \"Our third question concerned the influence of competitor presence on the performance metrics .\",\n \"If the confidence interval for the coefficient estimate of competitor presence excluded zero , we examined the magnitude and direction of that effect .\",\n \"Positive coefficient estimates indicate that performance was higher during competitive flights , whereas negative coefficients indicate that performance was lower in the presence of a competitor .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Despite recent advances in the study of animal flight , the biomechanical determinants of maneuverability are poorly understood .","It is thought that maneuverability may be influenced by intrinsic body mass and wing morphology , and by physiological muscle capacity , but this hypothesis has not yet been evaluated because it requires tracking a large number of free flight maneuvers from known individuals .","We used an automated tracking system to record flight sequences from 20 Anna's hummingbirds flying solo and in competition in a large chamber .","We found that burst muscle capacity predicted most performance metrics .","Hummingbirds with higher burst capacity flew with faster velocities , accelerations , and rotations , and they used more demanding complex turns .","In contrast , body mass did not predict variation in maneuvering performance , and wing morphology predicted only the use of arcing turns and high centripetal accelerations .","Collectively , our results indicate that burst muscle capacity is a key predictor of maneuverability ."],"string":"[\n \"Despite recent advances in the study of animal flight , the biomechanical determinants of maneuverability are poorly understood .\",\n \"It is thought that maneuverability may be influenced by intrinsic body mass and wing morphology , and by physiological muscle capacity , but this hypothesis has not yet been evaluated because it requires tracking a large number of free flight maneuvers from known individuals .\",\n \"We used an automated tracking system to record flight sequences from 20 Anna's hummingbirds flying solo and in competition in a large chamber .\",\n \"We found that burst muscle capacity predicted most performance metrics .\",\n \"Hummingbirds with higher burst capacity flew with faster velocities , accelerations , and rotations , and they used more demanding complex turns .\",\n \"In contrast , body mass did not predict variation in maneuvering performance , and wing morphology predicted only the use of arcing turns and high centripetal accelerations .\",\n \"Collectively , our results indicate that burst muscle capacity is a key predictor of maneuverability .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["The ability of an animal to maneuver can determine its success at avoiding predators , catching prey , and outperforming its competitors .","However , little is known about the characteristics that determine maneuverability .","Why are some individuals more maneuverable than others ?","To investigate this question , Segre et al . used an automated video tracking system to track male Anna's hummingbirds as they flew around a large chamber .","These tracks were then compared with the physical characteristics of the birds to see which , if any , affect the birds’ maneuverability .","This revealed that body size did not affect how well the birds could maneuver .","Instead , the muscle capacity of the birds – their ability to generate force rapidly – determined how well the birds performed most types of movement .","Birds with higher muscle capacity flew faster , had faster accelerations and decelerations , could rotate their bodies more quickly , and performed more demanding and complex turns .","Segre et al . also found that wing shape is important for a type of maneuver called an arcing turn .","Hummingbirds with a more slender wing shape were able to execute more demanding arcing turns involving higher accelerations , and they used arcing turns more often than birds with wider wings .","Future research will aim to determine whether these relationships are also found in other species of birds ."],"string":"[\n \"The ability of an animal to maneuver can determine its success at avoiding predators , catching prey , and outperforming its competitors .\",\n \"However , little is known about the characteristics that determine maneuverability .\",\n \"Why are some individuals more maneuverable than others ?\",\n \"To investigate this question , Segre et al . used an automated video tracking system to track male Anna's hummingbirds as they flew around a large chamber .\",\n \"These tracks were then compared with the physical characteristics of the birds to see which , if any , affect the birds’ maneuverability .\",\n \"This revealed that body size did not affect how well the birds could maneuver .\",\n \"Instead , the muscle capacity of the birds – their ability to generate force rapidly – determined how well the birds performed most types of movement .\",\n \"Birds with higher muscle capacity flew faster , had faster accelerations and decelerations , could rotate their bodies more quickly , and performed more demanding and complex turns .\",\n \"Segre et al . also found that wing shape is important for a type of maneuver called an arcing turn .\",\n \"Hummingbirds with a more slender wing shape were able to execute more demanding arcing turns involving higher accelerations , and they used arcing turns more often than birds with wider wings .\",\n \"Future research will aim to determine whether these relationships are also found in other species of birds .\"\n]","isConversational":false},"year":{"kind":"string","value":"2015"}}},{"rowIdx":88,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["structural biology and molecular biophysics"],"string":"[\n \"structural biology and molecular biophysics\"\n]","isConversational":false},"title":{"kind":"string","value":"Cryo-EM structures of the autoinhibited E. coli ATP synthase in three rotational states"},"id":{"kind":"string","value":"elife-21598-v4"},"sections":{"kind":"list like","value":[["In most cells , the bulk of ATP , the principal source of cellular energy , is synthesized by ATP synthase .","This molecular generator couples ion flow across membranes with the addition of inorganic phosphate ( Pi ) to ADP thereby generating ATP ( Iino and Noji , 2013; Stewart et al . , 2014 ) .","Most bacteria , including Escherichia coli have only one type of rotary ATPase , referred to as F-type ATPase .","Like the analogous complexes in other kingdoms , it is based on two reversible motors , termed F1 and Fo ( Negrin et al . , 1980 ) , connected by central and peripheral stalks ( Wilkens and Capaldi , 1998a ) ( Figure 1 ) .","The Fo motor spans the membrane converting the potential energy of the proton motive force ( pmf ) into rotation of the central stalk that in turn drives conformational changes in the F1 catalytic sites . 10 . 7554/eLife . 21598 . 003Figure 1 . Schematic illustration showing the arrangement of subunits in E . coli F-ATPase . Subunits α in red , β in yellow , γ in blue , ε in green , c in grey , a in orange , b in magenta or pink , and δ in teal .","The proton path and ATP synthesis are labeled accordingly . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 003 The Fo motor is constructed from subunits a , b and c ( Figure 1 ) .","Subunit c assembles into a ring , thought , in E . coli , to have decameric stoichiometry ( Jiang et al . , 2001; Ballhausen et al . , 2009; Düser et al . , 2009; Ishmukhametov et al . , 2010 ) , whereas subunits a and b associate to form a helical bundle adjacent to this ring .","Recent sub nanometer electron cryo-microscopy ( cryo-EM ) reconstructions of F-type ( Allegretti et al . , 2015; Zhou et al . , 2015; Kühlbrandt and Davies , 2016; Hahn et al . , 2016 ) and the analogous V- and A-type ATPases ( Zhao et al . , 2015; Schep et al . , 2016 ) as well as a low-resolution crystal structure of Paracoccus denitrificans F-ATPase ( Morales-Rios et al . , 2015 ) are consistent with a two half-channel mechanism for the generation of rotation within the membrane ( Vik and Antonio , 1994; Junge et al . , 1997 ) .","All structures confirm that a four-helix bundle of subunit a , inclined by 20–30˚ to the membrane plane , forms a crucial structural component .","In this mechanism , protons from the bacterial periplasm access a conserved negatively charged carboxylate in subunit c ( Asp61 in E . coli [Hoppe et al . , 1982] ) through an aqueous half channel at the subunit a/c interface ( Steed and Fillingame , 2008 ) .","Neutralizing this carboxylate enables the c-ring to rotate within the hydrophobic membrane and to access a second aqueous half channel that opens to the cytoplasm into which the protons are released ( Pogoryelov et al . , 2010 ) .","A conserved arginine residue in helix-4 of subunit a ( Arg210 in E . coli ) prevents the c-ring rotating in the opposite direction and short-circuiting of the system ( Lightowlers et al . , 1987; Cain and Simoni , 1989; Mitome et al . , 2010 ) .","The sequential binding of protons in combination with thermal fluctuations generates rotation within the complex in a manner akin to a turbine ( Oster and Wang , 1999; Pogoryelov et al . , 2010; Aksimentiev et al . , 2004 ) .","The torque generated in the Fo motor is then transferred to the F1 motor by the central shaft consisting of subunits γ and ε ( Wilkens et al . , 1995 ) .","The N- and C-termini of subunit γ form a curved coiled-coil that extends into the central cavity of F1 .","The F1 motor is the chemical generator in which ATP is synthesized .","The motor comprises a ring of three heterodimers , each containing an active site at the interface of subunits α and β .","Within the F1 motor , each αβ dimer has a different conformation at any point in time and can be either empty , bound to ADP and Pi , or bound to ATP ( open , half-closed , closed ) ( Abrahams et al . , 1994; Yoshida et al . , 2001 ) .","These different catalytic states relate to the position of the curved coiled-coil of subunit γ in the central stalk , which drives the conformational changes associated with catalysis .","To enable the central stalk to rotate relative to the F1 αβ heterodimers , the Fo and F1 motors need to be coupled .","This coupling is mediated by the peripheral stalk that is constructed from subunits b and δ ( Figure 1 ) .","Subunit b forms an amphipathic homodimeric coiled-coil that spans the periphery of the complex linking subunit a with subunit α , whereas subunit δ provides additional coupling of the C-termini of the b and α subunits ( Carbajo et al . , 2005; Wilkens et al . , 2005 ) .","The bacterial F-ATPase can also function in reverse , employing ATP hydrolysis to generate a proton gradient across the membrane when needed ( von Ballmoos et al . , 2008 ) .","In vivo , subunit ε is believed to change conformation in an ATP-dependent manner to prevent rotation of the complex ( Capaldi et al . , 1992; Rodgers and Wilce , 2000; Yagi et al . , 2007; Imamura et al . , 2009 ) thereby conserving ATP when its concentration is low .","This regulatory function is mediated by the C-terminal domain of subunit ε ( εCTD ) that , when not bound to ATP , opens to an extended conformation and inserts into the αβ heterodimers .","The crystal structures of both the E . coli and Bacillus PS3 F1 motors in this autoinhibited state show the εCTD intercalating into the αβ heterodimers ( Cingolani and Duncan , 2011; Shirakihara et al . , 2015 ) .","However , in each structure , the F1 motor had been captured in a different conformation ( E . coli – half-closed , closed , open [Cingolani and Duncan , 2011] and Bacillus PS3 – open , closed , open [Shirakihara et al . , 2015] ) which could either relate to inter species differences or crystal contacts and crystallization conditions .","The crystal structure of the F-ATPase from P . denitrificans ( Morales-Rios et al . , 2015 ) , that is closely related to E . coli ( 38% sequence identity over all subunits ) , shows a similar overall architecture to bovine F1Fo ATP synthase as well as to the main features of A/V ATPases .","However , it is inhibited by the ζ-protein rather than by subunit ε , which is generally employed by bacterial F-ATPases for this purpose .","Moreover , this crystal structure shows only one conformation of the rotary catalytic cycle .","Here , we present three cryo-EM maps along with molecular models of E . coli F-ATPase in its autoinhibited state , determined to resolutions of 6 . 9 , 7 . 8 , and 8 . 5 Å .","In all three reconstructions , the εCTD is in an extended conformation , stabilizing an overall F1 motor conformation similar to that seen in the thermophilic Bacillus PS3 F1 ATPase structure .","Density for the peripheral stalk extends the entire length of the complex and its coiled-coil bifurcates towards the N-terminus to enter the membrane as two separate helices that clamp the a subunit to the c ring .","Moreover , our maps allowed us to interpret the complete F1-delta interface , showing the three α subunit N-termini in distinct orientations .","Each map also confirmed the c-ring stoichiometry to be decameric , which to date has been only characterized by crosslinking and single molecule analyses .","We used our maps in combination with published crosslinking and mutagenesis information to generate a molecular model of the complex in three states .","These models provide crucial structural information on a key complex that extends our understanding of the mechanism of rotary ATPases in general , together with information on the bacterial ATP synthase , which is seen as an important antimicrobial target in organisms related to E . coli such as Mycobacterium tuberculosis ( Ahmad et al . , 2013 ) ."],["Cysteine-free E . coli F-ATPase , as described in Ishmukhametov et al . ( 2005 ) where all 10 cysteines were replaced with alanines and a His-tag introduced on the β subunit , was solubilized in digitonin detergent and purified as described in the Materials and methods .","This procedure provided pure protein ( Figure 2—figure supplement 1 ) capable of ATP hydrolysis-driven proton pumping upon reconstitution into proteoliposomes ( Figure 2—figure supplement 1 ) .","N , N`-dicyclohexylcarbodiimide ( DCCD ) , a compound which selectively modifies Asp61 of subunit c at 50 µM ( Pogoryelov et al . , 2010 ) completely abolished proton pumping ( Figure 2—figure supplement 1B ) and inhibited 90% of ATPase activity of isolated protein ( Figure 2—figure supplement 1C ) .","Such inhibition indicates coupling between the F1 and Fo motors ( Cook et al . , 2003; Peskova and Nakamoto , 2000; Tsunoda et al . , 2000 ) .","Protein was further examined by cryo-EM without addition of nucleotides .","395 , 140 particles were picked , of which 216 , 711 were used in refinement .","Three different conformations of the complex were identified using 3D classification in RELION ( Scheres , 2012 ) .","The particles in each subset were then refined to generate sub-nanometre reconstructions , to a resolution of 6 . 9 , 7 . 8 and 8 . 5 Å ( Figure 2A–C , and Figure 2—figure supplements 2 and 3 ) .","In these three conformations , the central stalk was progressively rotated 120° relative to the peripheral stalk . 10 . 7554/eLife . 21598 . 004Figure 2 . The three states of the autoinhibited E . coli F-ATPase .","( A–B )","Cryo-EM maps shown as surface representation , states 1 , 2 and 3 , respectively , resulting from rotation of the central stalk by 120° .","( D–F )","Molecular models built into the cryo-EM maps shown as cartoon representation .","Subunits α in red , β in yellow , γ in blue , ε in green , c in grey , a in orange , b in magenta or pink and δ in teal . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00410 . 7554/eLife . 21598 . 005Figure 2—source data 1 . Data collection and image processing statistics . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00510 . 7554/eLife . 21598 . 006Figure 2—figure supplement 1 . Characterization of E . coli F1Fo ATP synthase used for cryo-EM .","( A ) SDS page gel showing protein purity .","Lane1: SeeBlue Plus2 marker .","Lane 2: E . coli F-ATPase .","( B ) Inhibition of ATP-hydrolysis-driven proton pumping by F1Fo in the presence of and without 50 µM DCCD .","The protein was reconstituted into proteoliposomes and assayed for ACMA quenching as described in the Materials and methods .","Red arrow marks addition of ATP , blue arrows indicate addition of the uncoupler FCCP .","( C ) Inhibition of ATP hydrolysis by isolated F1Fo in the presence of 50 µM DCCD .","The protein was inhibited with and without 50 µM DCCD and assayed with the ATP regenerating system as described in Materials and methods . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00610 . 7554/eLife . 21598 . 007Figure 2—figure supplement 2 . cryoEM analysis .","( A ) Representative micrograph with picked particles , scale bar = 500 Å .","( B ) 25 highest populated classes from round 1 2D classification , scale bar = 100 Å .","( C ) Gold standard FSC plots of States 1 , 2 and 3 . ( D ) Orientation distribution of particles in the State one reconstruction . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00710 . 7554/eLife . 21598 . 008Figure 2—figure supplement 3 . Flowchart describing cryoEM data analysis . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00810 . 7554/eLife . 21598 . 009Figure 2—figure supplement 4 . Examples of the electron density map of State 1 , to highlight strengths and weaknesses .","( A ) Extra density at the N-terminus of subunit β shows 6xHis-tag .","( B ) Helical register visible in the density for the peripheral stalk .","( C ) β barrel of the ε subunit .","( D ) Poor density for the c-ring . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00910 . 7554/eLife . 21598 . 010Figure 2—figure supplement 5 . Local resolution map of State 1 . ( A ) Overall complex .","( B ) Slice through F1 .","( C ) Slice through FO .","Dashed lines show area of cross section and scale on right shows color/resolution relationship .","Made with ResMap ( Kucukelbir et al . , 2014 ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01010 . 7554/eLife . 21598 . 011Figure 2—figure supplement 6 . Quality of the models built into the state one cryoEM map . The model presented is a mosaic refined model created by docking crystal , NMR , homology models and de novo built sections .","The colors represent the origin of these models prior to refinement .","Crystal structures in blue , NMR models in cyan , homology model using a crystal structure in green , modeled using other EM structure as guide in yellow and built into the map in red . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01110 . 7554/eLife . 21598 . 012Figure 2—figure supplement 7 . Position of the natural cysteines in E . coli F1Fo . CryoEM map of state one containing built model .","Yellow spheres depict the positions of Cys-Ala mutants in the cysteine-free sample used . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01210 . 7554/eLife . 21598 . 013Figure 2—figure supplement 8 . Transmembrane architecture of ( A ) E . coli , ( B ) P . denitrificans , ( C ) Y . lipolytica . Unassigned subunits of P . denitrificans shown in cyan and subunit 8 of Y . lipolytica shown in pink . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01310 . 7554/eLife . 21598 . 014Figure 2—figure supplement 9 . Comparison of peripheral stalk position between the three states; diagrams on left depict part of complex that each state is superposed to . State 1 , 2 and 3 in red , green and blue , respectively .","The F1 has been removed for clarity .","( A–C )","Structures superposed to the a subunit .","( D–F )","Structures superposed to the delta subunit .","Arrows and triangle show distance displaced between states .","Circles highlight possible hinge/swivel regions . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01410 . 7554/eLife . 21598 . 015Figure 2—figure supplement 10 . FSC curves showing the effects of masking on the refined map , with the gold-standard , corrected FSC curve ( black ) , FSC of the unmasked map ( green ) , FSC of the masked map ( blue ) , and FSC of the phase-randomized masked map ( red ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 015 Even though the resolution of the reconstructions varied throughout the complex , it was sufficient to resolve individual helices .","Additional density of the N-terminal His-tag of the β subunit , as well as helical and β sheet patterns observed in parts of the map in the F1 motor region illustrate the high quality of the maps , with the c-ring density being poorest ( Figure 2—figure supplement 4 ) .","Local resolution estimates showed the region corresponding to the F1 motor to be of highest quality , the Fo motor with moderate detail and , the detergent micelle being clearly the worst region of the map ( Figure 2—figure supplement 5 ) .","Docking of high-resolution crystal and NMR models of different components into the maps followed by manual building and refinement enabled virtually complete molecular models of the three different states to be built ( Figure 2D–F ) , with varying quality of the docked structures as indicated in Figure 2—figure supplement 6 . The positions of the Cys-Ala mutants are depicted in Figure 2—figure supplement 7 . The cryo-EM maps provided novel insights into the architecture and function of the E . coli F-ATPase .","Thus , although its overall architecture was similar to that of F-ATPase from P . denitrificans ( Morales-Rios et al . , 2015 ) and F1Fo ATP synthases from Bos Taurus ( Zhou et al . , 2015 ) , Yarrowia lipolytica ( Hahn et al . , 2016 ) and Polytomella ( Allegretti et al . , 2015 ) , with the catalytic F1 motor attached to a proton powered membrane Fo motor and single central and peripheral stalks , differences in the individual motors and peripheral stalk were apparent .","Comparison with the membrane-embedded motors from other sub nanometre cryo-EM maps indicated the E . coli F-ATPase had a simpler stator architecture , containing only seven helices in the a and b subunits rather than the eight seen in mitochondrial F-type ATP synthase ( Figure 2—figure supplement 8 ) , consistent with labeling approaches ( Wada et al . , 1999 ) .","This difference suggested that the extra helices present in other rotary ATPase subtypes could have additional functions such as the dimerization seen in mitochondria ( Hahn et al . , 2016 ) .","A movie generated by interpolation between the three states ( Video 1 ) indicated that the F1 motor rocks or wobbles during the catalytic cycle ( Kinosita et al . , 2000; Stewart et al . , 2012 ) as previously predicted , although of course the structures described here do not represent the complex in its uninhibited active synthesizing form .","Two pivot points , one near the peripheral stalk/Fo interface ( ~bArg36 ) ( Welch et al . , 2008 ) and one near the peripheral stalk/F1 interface ( ~bGln106 ) , enabled the stalk to accommodate this eccentric movement of F1 ( Figure 2—figure supplement 9 ) . 10 . 7554/eLife . 21598 . 016Video 1 . Interpolation between States 1 , 3 and 2 to simulate ATP synthesis by E . coli F-ATPase . A and B are rotated 90° about the y-axis . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 016 The maps showed a long right-handed coiled-coil dimer generated by the two b subunits of the peripheral stalk together with the globular δ subunit that anchors them to the catalytic head ( Figure 3A ) .","The quality of the map was sufficient to enable almost the entire of the δ subunit to be built as a polyalanine model ( Figure 2D–F ) , whereas previous structural information was limited to the N-terminal domain ( Wilkens et al . , 2005 ) .","Interestingly , the peripheral stalk contacted all three α subunits via their N-terminal helices , but did so asymmetrically employing three different interfaces with each α subunit ( Figure 4 ) .","Although the resolution of the map was insufficient to assign the precise interface , the binding of the peripheral stalk to three anchor points in different geometries would provide a molecular key that would result in the δ subunit binding in a single orientation across the top of the symmetrical αβ heterodimers . 10 . 7554/eLife . 21598 . 017Figure 3 . The peripheral and central stalks of E . coli F-ATPase .","( A ) The peripheral stalk is comprised of a globular head ( subunit δ in teal ) and a homodimeric coiled-coil ( subunits b in pink and magenta ) that bifurcates at the membrane interface to brace subunit a ( orange ) .","( B ) The εCTD is in an extended conformation , inhibiting the enzyme from rotating .","The arrow depicts the extended vs closed conformation of subunit ε .","DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01710 . 7554/eLife . 21598 . 018Figure 3—figure supplement 1 . Fitting of the of the autoinhibited E . coli F1-ATPase crystal structure ( pdb 3oaa ) into the State one cryoEM map of E . coli F-ATPase .","( A and B ) 3oaa rigid body fitted into the cryoEM map .","β1 needs substantial movement to fit the density well .","( C and D )","Final refined model of the E . coli F1-ATPase shows all subunits fitting well . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01810 . 7554/eLife . 21598 . 019Figure 3—figure supplement 2 . Stimulation of ATP hydrolase activity of isolated F1Fo by 0 . 4% LDAO . ATPase activity of the protein was registered with the ATP regenerating system as described in Materials and methods .","Red arrow marks addition of the protein .","Blue arrow marks addition of LDAO . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01910 . 7554/eLife . 21598 . 020Figure 4 . Subunit δ and peripheral stalk attachments to the α subunits . Top panel; left , the segmented cryoEM map viewed from the side and right , viewed from above with the orientation of views 1 , 2 and 3 depicted .","Bottom panel; detailed views of the three attachment points labeled 1 , 2 and 3 , with δ in teal , b in pink and magenta and α in red . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 020 The b subunits formed a homodimeric coiled-coil that spaned almost the entire complex ( 212 of 232 Å ) with their N-termini bifurcating just above the membrane to generate two separate helices within the membrane ( Figure 3A ) .","This was unexpected , albeit reminiscent of the yeast F-type ATP synthase dimer ( Figure 2—figure supplement 8C ) , where subunit eight is an evolutionary derivative of the bacterial b subunit ( Hahn et al . , 2016 ) .","Furthermore , NMR analysis of the transmembrane domain of E . coli F-ATPase b subunit ( Dmitriev et al . , 1999 ) showed a helical structure that was interrupted by a rigid 20° bend at residues 23–26 that result in a structure consistent with the b subunit bifurcation .","All three reconstructions showed the complex in its autoinhibited state , with clear density for the εCTD extending deep into the central cavity of the F1 enzyme ( Figure 3B ) .","Fitting of the E . coli α3β3γε crystal structure ( Cingolani and Duncan , 2011 ) into our cryo-EM maps showed that the β1 subunit had adopted a different more open conformation ( Figure 3—figure supplement 1 ) .","In the above crystal structure , the εCTD contacts more subunits in F1 ( α1 , α2 , β1 , β2 and γ ) compared to our cryo-EM reconstructions , where it contacted fewer subunits ( α1 , α2 , β2 and γ ) .","The conformation of our cryo-EM structure was more similar to that seen in the Bacillus PS3 structure ( Shirakihara et al . , 2015 ) , where it is proposed to be in an ‘open , closed , open’ conformation ( Figure 5 ) . 10 . 7554/eLife . 21598 . 021Figure 5 . Autoinhibted E . coli F1-ATPase conformation . The αβ hetrodimers of state 1 as viewed from the membrane with the peripheral stalk to the left of the figure .","The ‘open , closed , open’ conformation of the F1 motor is labeled and the positions of nucleotides are shown as blue surfaces . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 021 To ascertain the enzymatic state of the F1 motor , we generated a difference density map between the cryo-EM density and our built model .","Remarkably , clear peaks were seen at the nucleotide binding pockets , indicating that the non-catalytic binding sites in subunit α contained nucleotide , as well as the ‘closed’ αβ heterodimer ( Figure 5 ) .","These were modeled as ATP and ADP respectively based on known orientations of these nucleotides from high-resolution crystal structures .","Interestingly , this did not correspond to the nucleotide binding states of either the E . coli ( Cingolani and Duncan , 2011 ) or the PS3 crystal structures ( Shirakihara et al . , 2015 ) .","Single particle analysis revealed highly uniform homogeneity of the protein preparation , where 100% of F1Fo molecules observed demonstrated the extended conformation of εCTD .","Our structural data were supported by an enzymatic assay ( Figure 3—figure supplement 2 ) , where the extent of ATP hydrolysis inhibition of the protein by subunit ε was tested with N , N-dimethyldodecylamine N-oxide ( LDAO ) , a well-known activator of the subunit ε inhibited protein .","LDAO used at 0 . 4% ( weight/volume ) concentration has been shown to stimulate ATP hydrolysis by the E coli protein 3–4 times ( Dunn et al . , 1990; Peskova and Nakamoto , 2000 ) , including earlier studies ( Ishmukhametov et al . , 2008 , 2016 ) on the same cysteine-free F1Fo construct using the same batch of LDAO .","Prior to addition of LDAO , the hydrolysis rate was 0 . 75 µmol ATP/min/mg protein but surprisingly in presence of 0 . 4% LDAO it was stimulated ~13 times .","To our best knowledge , such a high LDAO stimulation of ATP hydrolysis by E . coli F1Fo was not described in the literature before and is consistent with our single particle data .","Density in Fo defined the overall architecture of the membrane-embedded motor together with two invaginations of the detergent micelle that have previously been proposed to facilitate proton translocation ( Allegretti et al . , 2015; Kühlbrandt and Davies , 2016 ) ( Figure 6—figure supplement 1 ) .","While the overall density of the c-ring was relatively weak , 10 peaks of density were clearly present when viewed from above ( Figure 6—figure supplement 2 ) , confirming the stoichiometry of the c-ring to be decameric in E . coli F-ATPase ( Jiang et al . , 2001; Ballhausen et al . , 2009; Düser et al . , 2009; Ishmukhametov et al . , 2010 ) .","Furthermore , density inside the c-ring corroborates data suggesting it to be filled with phospholipids ( Oberfeld et al . , 2006 ) .","By combining the helical density from the cryo-EM maps ( Figure 6A ) , with models previously suggested for the related bovine subunit , together with crosslinking data and transmembrane topography prediction for the E . coli F-ATPase ( Jiang and Fillingame , 1998; Valiyaveetil and Fillingame , 1998; Moore and Fillingame , 2008; Wada et al . , 1999 ) , it was possible to build a molecular model of the a subunit ( Figure 6B ) .","The crosslinks mapped to two clusters ( Figure 6—figure supplement 3 ) , allowing a likely sequence register for the model to be proposed .","This was consistent with the two half channel hypothesis , placing Arg210 of subunit a adjacent to Asp61 of the c-ring ( Figure 6B ) .","Interestingly , density for the c subunit is clearest adjacent to Arg210 of subunit a suggesting this area to be well ordered ( Figure 6—figure supplement 4 ) . 10 . 7554/eLife . 21598 . 022Figure 6 . The E . coli F-ATPase subunit a and the suggested path of proton translocation .","( A ) Density map of subunit a , shown as orange surface viewed from the c-ring .","Grey outline depicts invaginations of the detergent micelle , with arrows showing possible proton path .","( B ) Cartoon representation of subunit a with a horizontal stripe to depict the position of Asp61 on the c-ring ( red where Asp61 would be bound to a proton and blue when bound to Arg210 ) .","Functional mutants labeled as follows; essential arginine in blue , substitution with Arg210 resulting in functional complex in yellow , mutation to arginine resulting in a dysfunctional complex in teal and residues that are aqueous accessible in red .","Solid arrows show a possible proton path via two ‘half’ channels and dashed arrows show the path when bound to Asp61 of the c-ring and rotating . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02210 . 7554/eLife . 21598 . 023Figure 6—figure supplement 1 . Aqueous cavities of the E . coli FO motor . Left: overall map , with cross-section and viewpoints highlighted .","( a ) and","( b ) : cross-sections to show invaginations ( red arrows ) .","( c ) and","( d ) : views from cytoplasm and periplasm to show invaginations ( within red circles ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02310 . 7554/eLife . 21598 . 024Figure 6—figure supplement 2 . View of the State two map from F1 to show c-ring stoichiometry ( numbered ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02410 . 7554/eLife . 21598 . 025Figure 6—figure supplement 3 . Crosslinks of the E . coli FO motor . Yellow lines depict distances between Cα atoms known to crosslink when double cysteine mutants are introduced .","( A–C ) intramolecular crosslinks identified in subunit a .","The sequence was mapped to minimize the distance of these crosslinks , with them localizing to two groups .","( D–F ) intermolecular crosslinks added to model show that the docked sequence maps well to the c-ring , but not as well to the b subunit . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02510 . 7554/eLife . 21598 . 026Figure 6—figure supplement 4 . Strong density near Arg210 . Arg210 of subunit a shown as blue stick , subunit a shown as orange cartoon , c-ring shown as grey cartoon and State one density shown as white surface . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02610 . 7554/eLife . 21598 . 027Figure 6—figure supplement 5 . Functional mutants of E . coli F-ATPase subunit a . As in main text Figure 6B , but with residues labeled and referenced ( superscript number ) .","( 1 . Cain and Simoni , 1986; 2 . Lightowlers et al . , 1987; 3 . Lightowlers et al . , 1988; 4 . Cain and Simoni , 1989; 5 . Hatch et al . , 1995; 6 . Angevine and Fillingame , 2003; 7 . Angevine et al . , 2003 ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 027"],["The three models and maps were deposited in the pdb and emDB with codes 5T4O , EMD-8357 ( State 1 ) , 5 T4P , EMD-8358 ( State 2 ) , 5T4Q and EMD-8359 ( State 3 ) ."],["A cysteine-free version of E . coli F-ATPase cloned in plasmid pFV2 and expressed in E . coli DK8 strain was used ( Ishmukhametov et al . , 2005 ) .","Cells were grown at 37°C in LB medium supplemented with 100 μg/ml ampicillin , for 4–5 hr .","The harvested cells were resuspended in lysis buffer containing 50 mM Tris/Cl pH 8 . 0 , 100 mM NaCl , 5 mM MgCl2 , 0 . 1 mM EDTA , 2 . 5% glycerol and 1 μg/ml DNase I and processed by one pass in French press at 20 kPSI .","Cellular debris was removed by centrifuging at 7700 × g for 15 min , and the membranes were collected by ultracentrifugation at 100 , 000 × g for 1 hr .","The ATP synthase complex was extracted from membranes at 4°C for 1 hr by resuspending the pellet in extraction buffer consisting of 20 mM Tris/Cl , pH 8 . 0 , 300 mM NaCl , 2 mM MgCl2 , 100 mM sucrose , 20 mM imidazole , 10% glycerol , 4 mM digitonin and EDTA-free protease inhibitor tablets ( Roche ) .","The complex was then purified by binding on Talon resin ( Clontech ) and eluted in 150 mM imidazole .","The protein was further purified and sugars removed by size exclusion chromatography on a 16/60 Superose six column equilibrated in a buffer containing 20 mM Tris/Cl pH 8 . 0 , 100 mM NaCl , 4 mM digitonin and 2 mM MgCl2 .","The purified protein was then concentrated to 2 mg/ml for cryo-EM .","Seventy microgram of F1Fo was reconstituted into extrusion-preformed 100 nm soybean phosphatidylcholine liposomes exactly as descried ( Ishmukhametov et al . , 2016 ) .","Proton pumping by proteoliposomes was studied using quenching of a pH sensitive fluorescent probe 9-Amino-6-Chloro-2-Methoxyacridine ( ACMA ) exactly as described ( Ishmukhametov et al . , 2016 ) .","The assay was performed with 100 µl of proteoliposomes in 2-ml cuvettes .","The reaction was started with 0 . 25 mM ATP and stopped by 2 µM of the uncoupler FCCP .","ATP hydrolase activity and its stimulation by LDAO was measured with ATP regenerating system using 5 µg of the protein with 1 mM ATP exactly as described ( Ishmukhametov et al . , 2016 ) .","DCCD inhibition of proteoliposomes was done as described ( Ishmukhametov et al . , 2016 ) .","DCCD inhibition of pure F1Fo was done as described ( Ishmukhametov et al . , 2005 ) , with the following modification .","Ten microgram of the protein was incubated in 1 ml buffer A ( 50 mM MES , pH 6 . 4 , 100 mM KCl , 1 mM MgCl2 ) with 50 µM DCCD for 30 min at room temperature .","Control sample contained 1% ethanol instead of DCCD .","Reaction was started by mixing the inhibited protein with 1 ml of buffer A containing all the components of ATP regenerating system .","All the functional experiments presented here were repeated two to three times using the protein isolated by MS and shipped to RI at liquid N2 temperature .","Results of typical experiments are shown .","Aliquots of 4 μl of purified E . coli F-ATPase at a concentration of 3 . 58 μM were placed on glow-discharged holey carbon grids ( Quantifoils Copper R2/2 , 200 Mesh ) .","Grids were blotted for 2 s and flash-frozen in liquid ethane using an FEI Vitrobot Mark IV .","Grids were transferred to an FEI Titan Krios transmission electron microscope that was operating at 300 kV .","Images were recorded automatically using the FEI EPU software , yielding a pixel size of 1 . 4 Å .","A dose rate of 29 electrons ( spread over 20 frames ) per Å2 per second , and an exposure time of 2 s were used on the Falcon-II detector .","8640 movies were collected .","MotionCorr ( Li et al . , 2013 ) was used to correct local beam-induced motion and to align resulting frames .","Defocus and astigmatism values were estimated using CTFFIND4 ( Rohou and Grigorieff , 2015 ) , and 252 micrographs were excluded due to drift or excessive ice contamination .","1208 particles were manually picked and subjected to 2D classification to generate templates for autopicking in RELION ( Scheres , 2012 ) .","The automatically picked micrographs were manually inspected to remove false positives , finally yielding 395 , 140 particles .","These particles then underwent two rounds of 2D classification to generate 22 classes with 311 , 887 particles .","The final particles were classified into four 3D classes using a previously generated model from a low-resolution data set of the same sample ( unpublished ) , low-pass filtered to 60 Å .","The resolution was estimated using Fourier Shell Correlation ( FSC = 0 . 143 , gold-standard ) .","Three of the four classes containing 104 , 510 ( State 1 ) , 67 , 829 ( State","2 ) and 53 , 587 ( State","3 ) particles were movie-refined and post-processed in RELION producing maps at 7 . 4 , 7 . 8 , 8 . 5 Å , respectively ( Figure 2—figure supplement 10 ) .","State 1 was further processed using masked classification ( Bai et al . , 2015 ) with residual signal subtraction with a mask created by removing parts of the detergent micelle .","Three out of the four classes from this classification containing 95 , 345 particles were combined and refined to generate the final 6 . 9 Å map .","Figure 2—figure supplement 3 is a summary of these methods , shown as a flowchart .","Local resolution of different parts of the complex was estimated using RELION and ResMap ( Kucukelbir et al . , 2014 ) .","Crystal and NMR structures of subunits from E . coli ( αβγε - 3oaa [Cingolani and Duncan , 2011] , δ - 2a7u [Wilkens et al . , 2005] , b - 1b9u [Dmitriev et al . , 1999] , 1l2p [Del Rizzo et al . , 2002] and 2khk [Priya et al . , 2009] ) and related organisms ( c - 3u2f [Symersky et al . , 2012] and a - 5fik [Zhou et al . , 2015] ) were rigid body docked into the highest resolution cryo-EM map and the side chains ‘pruned’ to Cα .","The sequence was mapped to subunit a using crosslinks as restraints .","Subsequent manual model building and refinement was performed with Coot ( Emsley et al . , 2010 ) , Phenix ( Adams et al . , 2010 ) and Refmac ( Murshudov et al . , 2011 ) ( excluding subunit c due to weak density ) , with crosslinks again used as external restraints ( a summary of the types of models used to build the initial model can be found in Figure 2—figure supplement 6 ) .","Nucleotide occupancy was determined by first building the model without any nucleotide present , and then segmenting the map and selecting any density with 15% overlap with atoms and deleting this density .","Nucleotide was subsequently docked into this difference density using the known positions from previous structures .","Once a complete model was built of the highest resolution map , this was docked and refined to the other two maps to create three models .","The three models and maps were deposited in the pdb and emDataBank with codes 5T4O , EMD-8357 ( State 1 ) , 5 T4P , EMD-8358 ( State 2 ) , 5T4Q and EMD-8359 ( State 3 ) .","Data statistics shown in Figure 2—source data 1 ."]],"string":"[\n [\n \"In most cells , the bulk of ATP , the principal source of cellular energy , is synthesized by ATP synthase .\",\n \"This molecular generator couples ion flow across membranes with the addition of inorganic phosphate ( Pi ) to ADP thereby generating ATP ( Iino and Noji , 2013; Stewart et al . , 2014 ) .\",\n \"Most bacteria , including Escherichia coli have only one type of rotary ATPase , referred to as F-type ATPase .\",\n \"Like the analogous complexes in other kingdoms , it is based on two reversible motors , termed F1 and Fo ( Negrin et al . , 1980 ) , connected by central and peripheral stalks ( Wilkens and Capaldi , 1998a ) ( Figure 1 ) .\",\n \"The Fo motor spans the membrane converting the potential energy of the proton motive force ( pmf ) into rotation of the central stalk that in turn drives conformational changes in the F1 catalytic sites . 10 . 7554/eLife . 21598 . 003Figure 1 . Schematic illustration showing the arrangement of subunits in E . coli F-ATPase . Subunits α in red , β in yellow , γ in blue , ε in green , c in grey , a in orange , b in magenta or pink , and δ in teal .\",\n \"The proton path and ATP synthesis are labeled accordingly . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 003 The Fo motor is constructed from subunits a , b and c ( Figure 1 ) .\",\n \"Subunit c assembles into a ring , thought , in E . coli , to have decameric stoichiometry ( Jiang et al . , 2001; Ballhausen et al . , 2009; Düser et al . , 2009; Ishmukhametov et al . , 2010 ) , whereas subunits a and b associate to form a helical bundle adjacent to this ring .\",\n \"Recent sub nanometer electron cryo-microscopy ( cryo-EM ) reconstructions of F-type ( Allegretti et al . , 2015; Zhou et al . , 2015; Kühlbrandt and Davies , 2016; Hahn et al . , 2016 ) and the analogous V- and A-type ATPases ( Zhao et al . , 2015; Schep et al . , 2016 ) as well as a low-resolution crystal structure of Paracoccus denitrificans F-ATPase ( Morales-Rios et al . , 2015 ) are consistent with a two half-channel mechanism for the generation of rotation within the membrane ( Vik and Antonio , 1994; Junge et al . , 1997 ) .\",\n \"All structures confirm that a four-helix bundle of subunit a , inclined by 20–30˚ to the membrane plane , forms a crucial structural component .\",\n \"In this mechanism , protons from the bacterial periplasm access a conserved negatively charged carboxylate in subunit c ( Asp61 in E . coli [Hoppe et al . , 1982] ) through an aqueous half channel at the subunit a/c interface ( Steed and Fillingame , 2008 ) .\",\n \"Neutralizing this carboxylate enables the c-ring to rotate within the hydrophobic membrane and to access a second aqueous half channel that opens to the cytoplasm into which the protons are released ( Pogoryelov et al . , 2010 ) .\",\n \"A conserved arginine residue in helix-4 of subunit a ( Arg210 in E . coli ) prevents the c-ring rotating in the opposite direction and short-circuiting of the system ( Lightowlers et al . , 1987; Cain and Simoni , 1989; Mitome et al . , 2010 ) .\",\n \"The sequential binding of protons in combination with thermal fluctuations generates rotation within the complex in a manner akin to a turbine ( Oster and Wang , 1999; Pogoryelov et al . , 2010; Aksimentiev et al . , 2004 ) .\",\n \"The torque generated in the Fo motor is then transferred to the F1 motor by the central shaft consisting of subunits γ and ε ( Wilkens et al . , 1995 ) .\",\n \"The N- and C-termini of subunit γ form a curved coiled-coil that extends into the central cavity of F1 .\",\n \"The F1 motor is the chemical generator in which ATP is synthesized .\",\n \"The motor comprises a ring of three heterodimers , each containing an active site at the interface of subunits α and β .\",\n \"Within the F1 motor , each αβ dimer has a different conformation at any point in time and can be either empty , bound to ADP and Pi , or bound to ATP ( open , half-closed , closed ) ( Abrahams et al . , 1994; Yoshida et al . , 2001 ) .\",\n \"These different catalytic states relate to the position of the curved coiled-coil of subunit γ in the central stalk , which drives the conformational changes associated with catalysis .\",\n \"To enable the central stalk to rotate relative to the F1 αβ heterodimers , the Fo and F1 motors need to be coupled .\",\n \"This coupling is mediated by the peripheral stalk that is constructed from subunits b and δ ( Figure 1 ) .\",\n \"Subunit b forms an amphipathic homodimeric coiled-coil that spans the periphery of the complex linking subunit a with subunit α , whereas subunit δ provides additional coupling of the C-termini of the b and α subunits ( Carbajo et al . , 2005; Wilkens et al . , 2005 ) .\",\n \"The bacterial F-ATPase can also function in reverse , employing ATP hydrolysis to generate a proton gradient across the membrane when needed ( von Ballmoos et al . , 2008 ) .\",\n \"In vivo , subunit ε is believed to change conformation in an ATP-dependent manner to prevent rotation of the complex ( Capaldi et al . , 1992; Rodgers and Wilce , 2000; Yagi et al . , 2007; Imamura et al . , 2009 ) thereby conserving ATP when its concentration is low .\",\n \"This regulatory function is mediated by the C-terminal domain of subunit ε ( εCTD ) that , when not bound to ATP , opens to an extended conformation and inserts into the αβ heterodimers .\",\n \"The crystal structures of both the E . coli and Bacillus PS3 F1 motors in this autoinhibited state show the εCTD intercalating into the αβ heterodimers ( Cingolani and Duncan , 2011; Shirakihara et al . , 2015 ) .\",\n \"However , in each structure , the F1 motor had been captured in a different conformation ( E . coli – half-closed , closed , open [Cingolani and Duncan , 2011] and Bacillus PS3 – open , closed , open [Shirakihara et al . , 2015] ) which could either relate to inter species differences or crystal contacts and crystallization conditions .\",\n \"The crystal structure of the F-ATPase from P . denitrificans ( Morales-Rios et al . , 2015 ) , that is closely related to E . coli ( 38% sequence identity over all subunits ) , shows a similar overall architecture to bovine F1Fo ATP synthase as well as to the main features of A/V ATPases .\",\n \"However , it is inhibited by the ζ-protein rather than by subunit ε , which is generally employed by bacterial F-ATPases for this purpose .\",\n \"Moreover , this crystal structure shows only one conformation of the rotary catalytic cycle .\",\n \"Here , we present three cryo-EM maps along with molecular models of E . coli F-ATPase in its autoinhibited state , determined to resolutions of 6 . 9 , 7 . 8 , and 8 . 5 Å .\",\n \"In all three reconstructions , the εCTD is in an extended conformation , stabilizing an overall F1 motor conformation similar to that seen in the thermophilic Bacillus PS3 F1 ATPase structure .\",\n \"Density for the peripheral stalk extends the entire length of the complex and its coiled-coil bifurcates towards the N-terminus to enter the membrane as two separate helices that clamp the a subunit to the c ring .\",\n \"Moreover , our maps allowed us to interpret the complete F1-delta interface , showing the three α subunit N-termini in distinct orientations .\",\n \"Each map also confirmed the c-ring stoichiometry to be decameric , which to date has been only characterized by crosslinking and single molecule analyses .\",\n \"We used our maps in combination with published crosslinking and mutagenesis information to generate a molecular model of the complex in three states .\",\n \"These models provide crucial structural information on a key complex that extends our understanding of the mechanism of rotary ATPases in general , together with information on the bacterial ATP synthase , which is seen as an important antimicrobial target in organisms related to E . coli such as Mycobacterium tuberculosis ( Ahmad et al . , 2013 ) .\"\n ],\n [\n \"Cysteine-free E . coli F-ATPase , as described in Ishmukhametov et al . ( 2005 ) where all 10 cysteines were replaced with alanines and a His-tag introduced on the β subunit , was solubilized in digitonin detergent and purified as described in the Materials and methods .\",\n \"This procedure provided pure protein ( Figure 2—figure supplement 1 ) capable of ATP hydrolysis-driven proton pumping upon reconstitution into proteoliposomes ( Figure 2—figure supplement 1 ) .\",\n \"N , N`-dicyclohexylcarbodiimide ( DCCD ) , a compound which selectively modifies Asp61 of subunit c at 50 µM ( Pogoryelov et al . , 2010 ) completely abolished proton pumping ( Figure 2—figure supplement 1B ) and inhibited 90% of ATPase activity of isolated protein ( Figure 2—figure supplement 1C ) .\",\n \"Such inhibition indicates coupling between the F1 and Fo motors ( Cook et al . , 2003; Peskova and Nakamoto , 2000; Tsunoda et al . , 2000 ) .\",\n \"Protein was further examined by cryo-EM without addition of nucleotides .\",\n \"395 , 140 particles were picked , of which 216 , 711 were used in refinement .\",\n \"Three different conformations of the complex were identified using 3D classification in RELION ( Scheres , 2012 ) .\",\n \"The particles in each subset were then refined to generate sub-nanometre reconstructions , to a resolution of 6 . 9 , 7 . 8 and 8 . 5 Å ( Figure 2A–C , and Figure 2—figure supplements 2 and 3 ) .\",\n \"In these three conformations , the central stalk was progressively rotated 120° relative to the peripheral stalk . 10 . 7554/eLife . 21598 . 004Figure 2 . The three states of the autoinhibited E . coli F-ATPase .\",\n \"( A–B )\",\n \"Cryo-EM maps shown as surface representation , states 1 , 2 and 3 , respectively , resulting from rotation of the central stalk by 120° .\",\n \"( D–F )\",\n \"Molecular models built into the cryo-EM maps shown as cartoon representation .\",\n \"Subunits α in red , β in yellow , γ in blue , ε in green , c in grey , a in orange , b in magenta or pink and δ in teal . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00410 . 7554/eLife . 21598 . 005Figure 2—source data 1 . Data collection and image processing statistics . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00510 . 7554/eLife . 21598 . 006Figure 2—figure supplement 1 . Characterization of E . coli F1Fo ATP synthase used for cryo-EM .\",\n \"( A ) SDS page gel showing protein purity .\",\n \"Lane1: SeeBlue Plus2 marker .\",\n \"Lane 2: E . coli F-ATPase .\",\n \"( B ) Inhibition of ATP-hydrolysis-driven proton pumping by F1Fo in the presence of and without 50 µM DCCD .\",\n \"The protein was reconstituted into proteoliposomes and assayed for ACMA quenching as described in the Materials and methods .\",\n \"Red arrow marks addition of ATP , blue arrows indicate addition of the uncoupler FCCP .\",\n \"( C ) Inhibition of ATP hydrolysis by isolated F1Fo in the presence of 50 µM DCCD .\",\n \"The protein was inhibited with and without 50 µM DCCD and assayed with the ATP regenerating system as described in Materials and methods . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00610 . 7554/eLife . 21598 . 007Figure 2—figure supplement 2 . cryoEM analysis .\",\n \"( A ) Representative micrograph with picked particles , scale bar = 500 Å .\",\n \"( B ) 25 highest populated classes from round 1 2D classification , scale bar = 100 Å .\",\n \"( C ) Gold standard FSC plots of States 1 , 2 and 3 . ( D ) Orientation distribution of particles in the State one reconstruction . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00710 . 7554/eLife . 21598 . 008Figure 2—figure supplement 3 . Flowchart describing cryoEM data analysis . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00810 . 7554/eLife . 21598 . 009Figure 2—figure supplement 4 . Examples of the electron density map of State 1 , to highlight strengths and weaknesses .\",\n \"( A ) Extra density at the N-terminus of subunit β shows 6xHis-tag .\",\n \"( B ) Helical register visible in the density for the peripheral stalk .\",\n \"( C ) β barrel of the ε subunit .\",\n \"( D ) Poor density for the c-ring . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 00910 . 7554/eLife . 21598 . 010Figure 2—figure supplement 5 . Local resolution map of State 1 . ( A ) Overall complex .\",\n \"( B ) Slice through F1 .\",\n \"( C ) Slice through FO .\",\n \"Dashed lines show area of cross section and scale on right shows color/resolution relationship .\",\n \"Made with ResMap ( Kucukelbir et al . , 2014 ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01010 . 7554/eLife . 21598 . 011Figure 2—figure supplement 6 . Quality of the models built into the state one cryoEM map . The model presented is a mosaic refined model created by docking crystal , NMR , homology models and de novo built sections .\",\n \"The colors represent the origin of these models prior to refinement .\",\n \"Crystal structures in blue , NMR models in cyan , homology model using a crystal structure in green , modeled using other EM structure as guide in yellow and built into the map in red . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01110 . 7554/eLife . 21598 . 012Figure 2—figure supplement 7 . Position of the natural cysteines in E . coli F1Fo . CryoEM map of state one containing built model .\",\n \"Yellow spheres depict the positions of Cys-Ala mutants in the cysteine-free sample used . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01210 . 7554/eLife . 21598 . 013Figure 2—figure supplement 8 . Transmembrane architecture of ( A ) E . coli , ( B ) P . denitrificans , ( C ) Y . lipolytica . Unassigned subunits of P . denitrificans shown in cyan and subunit 8 of Y . lipolytica shown in pink . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01310 . 7554/eLife . 21598 . 014Figure 2—figure supplement 9 . Comparison of peripheral stalk position between the three states; diagrams on left depict part of complex that each state is superposed to . State 1 , 2 and 3 in red , green and blue , respectively .\",\n \"The F1 has been removed for clarity .\",\n \"( A–C )\",\n \"Structures superposed to the a subunit .\",\n \"( D–F )\",\n \"Structures superposed to the delta subunit .\",\n \"Arrows and triangle show distance displaced between states .\",\n \"Circles highlight possible hinge/swivel regions . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01410 . 7554/eLife . 21598 . 015Figure 2—figure supplement 10 . FSC curves showing the effects of masking on the refined map , with the gold-standard , corrected FSC curve ( black ) , FSC of the unmasked map ( green ) , FSC of the masked map ( blue ) , and FSC of the phase-randomized masked map ( red ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 015 Even though the resolution of the reconstructions varied throughout the complex , it was sufficient to resolve individual helices .\",\n \"Additional density of the N-terminal His-tag of the β subunit , as well as helical and β sheet patterns observed in parts of the map in the F1 motor region illustrate the high quality of the maps , with the c-ring density being poorest ( Figure 2—figure supplement 4 ) .\",\n \"Local resolution estimates showed the region corresponding to the F1 motor to be of highest quality , the Fo motor with moderate detail and , the detergent micelle being clearly the worst region of the map ( Figure 2—figure supplement 5 ) .\",\n \"Docking of high-resolution crystal and NMR models of different components into the maps followed by manual building and refinement enabled virtually complete molecular models of the three different states to be built ( Figure 2D–F ) , with varying quality of the docked structures as indicated in Figure 2—figure supplement 6 . The positions of the Cys-Ala mutants are depicted in Figure 2—figure supplement 7 . The cryo-EM maps provided novel insights into the architecture and function of the E . coli F-ATPase .\",\n \"Thus , although its overall architecture was similar to that of F-ATPase from P . denitrificans ( Morales-Rios et al . , 2015 ) and F1Fo ATP synthases from Bos Taurus ( Zhou et al . , 2015 ) , Yarrowia lipolytica ( Hahn et al . , 2016 ) and Polytomella ( Allegretti et al . , 2015 ) , with the catalytic F1 motor attached to a proton powered membrane Fo motor and single central and peripheral stalks , differences in the individual motors and peripheral stalk were apparent .\",\n \"Comparison with the membrane-embedded motors from other sub nanometre cryo-EM maps indicated the E . coli F-ATPase had a simpler stator architecture , containing only seven helices in the a and b subunits rather than the eight seen in mitochondrial F-type ATP synthase ( Figure 2—figure supplement 8 ) , consistent with labeling approaches ( Wada et al . , 1999 ) .\",\n \"This difference suggested that the extra helices present in other rotary ATPase subtypes could have additional functions such as the dimerization seen in mitochondria ( Hahn et al . , 2016 ) .\",\n \"A movie generated by interpolation between the three states ( Video 1 ) indicated that the F1 motor rocks or wobbles during the catalytic cycle ( Kinosita et al . , 2000; Stewart et al . , 2012 ) as previously predicted , although of course the structures described here do not represent the complex in its uninhibited active synthesizing form .\",\n \"Two pivot points , one near the peripheral stalk/Fo interface ( ~bArg36 ) ( Welch et al . , 2008 ) and one near the peripheral stalk/F1 interface ( ~bGln106 ) , enabled the stalk to accommodate this eccentric movement of F1 ( Figure 2—figure supplement 9 ) . 10 . 7554/eLife . 21598 . 016Video 1 . Interpolation between States 1 , 3 and 2 to simulate ATP synthesis by E . coli F-ATPase . A and B are rotated 90° about the y-axis . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 016 The maps showed a long right-handed coiled-coil dimer generated by the two b subunits of the peripheral stalk together with the globular δ subunit that anchors them to the catalytic head ( Figure 3A ) .\",\n \"The quality of the map was sufficient to enable almost the entire of the δ subunit to be built as a polyalanine model ( Figure 2D–F ) , whereas previous structural information was limited to the N-terminal domain ( Wilkens et al . , 2005 ) .\",\n \"Interestingly , the peripheral stalk contacted all three α subunits via their N-terminal helices , but did so asymmetrically employing three different interfaces with each α subunit ( Figure 4 ) .\",\n \"Although the resolution of the map was insufficient to assign the precise interface , the binding of the peripheral stalk to three anchor points in different geometries would provide a molecular key that would result in the δ subunit binding in a single orientation across the top of the symmetrical αβ heterodimers . 10 . 7554/eLife . 21598 . 017Figure 3 . The peripheral and central stalks of E . coli F-ATPase .\",\n \"( A ) The peripheral stalk is comprised of a globular head ( subunit δ in teal ) and a homodimeric coiled-coil ( subunits b in pink and magenta ) that bifurcates at the membrane interface to brace subunit a ( orange ) .\",\n \"( B ) The εCTD is in an extended conformation , inhibiting the enzyme from rotating .\",\n \"The arrow depicts the extended vs closed conformation of subunit ε .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01710 . 7554/eLife . 21598 . 018Figure 3—figure supplement 1 . Fitting of the of the autoinhibited E . coli F1-ATPase crystal structure ( pdb 3oaa ) into the State one cryoEM map of E . coli F-ATPase .\",\n \"( A and B ) 3oaa rigid body fitted into the cryoEM map .\",\n \"β1 needs substantial movement to fit the density well .\",\n \"( C and D )\",\n \"Final refined model of the E . coli F1-ATPase shows all subunits fitting well . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01810 . 7554/eLife . 21598 . 019Figure 3—figure supplement 2 . Stimulation of ATP hydrolase activity of isolated F1Fo by 0 . 4% LDAO . ATPase activity of the protein was registered with the ATP regenerating system as described in Materials and methods .\",\n \"Red arrow marks addition of the protein .\",\n \"Blue arrow marks addition of LDAO . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 01910 . 7554/eLife . 21598 . 020Figure 4 . Subunit δ and peripheral stalk attachments to the α subunits . Top panel; left , the segmented cryoEM map viewed from the side and right , viewed from above with the orientation of views 1 , 2 and 3 depicted .\",\n \"Bottom panel; detailed views of the three attachment points labeled 1 , 2 and 3 , with δ in teal , b in pink and magenta and α in red . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 020 The b subunits formed a homodimeric coiled-coil that spaned almost the entire complex ( 212 of 232 Å ) with their N-termini bifurcating just above the membrane to generate two separate helices within the membrane ( Figure 3A ) .\",\n \"This was unexpected , albeit reminiscent of the yeast F-type ATP synthase dimer ( Figure 2—figure supplement 8C ) , where subunit eight is an evolutionary derivative of the bacterial b subunit ( Hahn et al . , 2016 ) .\",\n \"Furthermore , NMR analysis of the transmembrane domain of E . coli F-ATPase b subunit ( Dmitriev et al . , 1999 ) showed a helical structure that was interrupted by a rigid 20° bend at residues 23–26 that result in a structure consistent with the b subunit bifurcation .\",\n \"All three reconstructions showed the complex in its autoinhibited state , with clear density for the εCTD extending deep into the central cavity of the F1 enzyme ( Figure 3B ) .\",\n \"Fitting of the E . coli α3β3γε crystal structure ( Cingolani and Duncan , 2011 ) into our cryo-EM maps showed that the β1 subunit had adopted a different more open conformation ( Figure 3—figure supplement 1 ) .\",\n \"In the above crystal structure , the εCTD contacts more subunits in F1 ( α1 , α2 , β1 , β2 and γ ) compared to our cryo-EM reconstructions , where it contacted fewer subunits ( α1 , α2 , β2 and γ ) .\",\n \"The conformation of our cryo-EM structure was more similar to that seen in the Bacillus PS3 structure ( Shirakihara et al . , 2015 ) , where it is proposed to be in an ‘open , closed , open’ conformation ( Figure 5 ) . 10 . 7554/eLife . 21598 . 021Figure 5 . Autoinhibted E . coli F1-ATPase conformation . The αβ hetrodimers of state 1 as viewed from the membrane with the peripheral stalk to the left of the figure .\",\n \"The ‘open , closed , open’ conformation of the F1 motor is labeled and the positions of nucleotides are shown as blue surfaces . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 021 To ascertain the enzymatic state of the F1 motor , we generated a difference density map between the cryo-EM density and our built model .\",\n \"Remarkably , clear peaks were seen at the nucleotide binding pockets , indicating that the non-catalytic binding sites in subunit α contained nucleotide , as well as the ‘closed’ αβ heterodimer ( Figure 5 ) .\",\n \"These were modeled as ATP and ADP respectively based on known orientations of these nucleotides from high-resolution crystal structures .\",\n \"Interestingly , this did not correspond to the nucleotide binding states of either the E . coli ( Cingolani and Duncan , 2011 ) or the PS3 crystal structures ( Shirakihara et al . , 2015 ) .\",\n \"Single particle analysis revealed highly uniform homogeneity of the protein preparation , where 100% of F1Fo molecules observed demonstrated the extended conformation of εCTD .\",\n \"Our structural data were supported by an enzymatic assay ( Figure 3—figure supplement 2 ) , where the extent of ATP hydrolysis inhibition of the protein by subunit ε was tested with N , N-dimethyldodecylamine N-oxide ( LDAO ) , a well-known activator of the subunit ε inhibited protein .\",\n \"LDAO used at 0 . 4% ( weight/volume ) concentration has been shown to stimulate ATP hydrolysis by the E coli protein 3–4 times ( Dunn et al . , 1990; Peskova and Nakamoto , 2000 ) , including earlier studies ( Ishmukhametov et al . , 2008 , 2016 ) on the same cysteine-free F1Fo construct using the same batch of LDAO .\",\n \"Prior to addition of LDAO , the hydrolysis rate was 0 . 75 µmol ATP/min/mg protein but surprisingly in presence of 0 . 4% LDAO it was stimulated ~13 times .\",\n \"To our best knowledge , such a high LDAO stimulation of ATP hydrolysis by E . coli F1Fo was not described in the literature before and is consistent with our single particle data .\",\n \"Density in Fo defined the overall architecture of the membrane-embedded motor together with two invaginations of the detergent micelle that have previously been proposed to facilitate proton translocation ( Allegretti et al . , 2015; Kühlbrandt and Davies , 2016 ) ( Figure 6—figure supplement 1 ) .\",\n \"While the overall density of the c-ring was relatively weak , 10 peaks of density were clearly present when viewed from above ( Figure 6—figure supplement 2 ) , confirming the stoichiometry of the c-ring to be decameric in E . coli F-ATPase ( Jiang et al . , 2001; Ballhausen et al . , 2009; Düser et al . , 2009; Ishmukhametov et al . , 2010 ) .\",\n \"Furthermore , density inside the c-ring corroborates data suggesting it to be filled with phospholipids ( Oberfeld et al . , 2006 ) .\",\n \"By combining the helical density from the cryo-EM maps ( Figure 6A ) , with models previously suggested for the related bovine subunit , together with crosslinking data and transmembrane topography prediction for the E . coli F-ATPase ( Jiang and Fillingame , 1998; Valiyaveetil and Fillingame , 1998; Moore and Fillingame , 2008; Wada et al . , 1999 ) , it was possible to build a molecular model of the a subunit ( Figure 6B ) .\",\n \"The crosslinks mapped to two clusters ( Figure 6—figure supplement 3 ) , allowing a likely sequence register for the model to be proposed .\",\n \"This was consistent with the two half channel hypothesis , placing Arg210 of subunit a adjacent to Asp61 of the c-ring ( Figure 6B ) .\",\n \"Interestingly , density for the c subunit is clearest adjacent to Arg210 of subunit a suggesting this area to be well ordered ( Figure 6—figure supplement 4 ) . 10 . 7554/eLife . 21598 . 022Figure 6 . The E . coli F-ATPase subunit a and the suggested path of proton translocation .\",\n \"( A ) Density map of subunit a , shown as orange surface viewed from the c-ring .\",\n \"Grey outline depicts invaginations of the detergent micelle , with arrows showing possible proton path .\",\n \"( B ) Cartoon representation of subunit a with a horizontal stripe to depict the position of Asp61 on the c-ring ( red where Asp61 would be bound to a proton and blue when bound to Arg210 ) .\",\n \"Functional mutants labeled as follows; essential arginine in blue , substitution with Arg210 resulting in functional complex in yellow , mutation to arginine resulting in a dysfunctional complex in teal and residues that are aqueous accessible in red .\",\n \"Solid arrows show a possible proton path via two ‘half’ channels and dashed arrows show the path when bound to Asp61 of the c-ring and rotating . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02210 . 7554/eLife . 21598 . 023Figure 6—figure supplement 1 . Aqueous cavities of the E . coli FO motor . Left: overall map , with cross-section and viewpoints highlighted .\",\n \"( a ) and\",\n \"( b ) : cross-sections to show invaginations ( red arrows ) .\",\n \"( c ) and\",\n \"( d ) : views from cytoplasm and periplasm to show invaginations ( within red circles ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02310 . 7554/eLife . 21598 . 024Figure 6—figure supplement 2 . View of the State two map from F1 to show c-ring stoichiometry ( numbered ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02410 . 7554/eLife . 21598 . 025Figure 6—figure supplement 3 . Crosslinks of the E . coli FO motor . Yellow lines depict distances between Cα atoms known to crosslink when double cysteine mutants are introduced .\",\n \"( A–C ) intramolecular crosslinks identified in subunit a .\",\n \"The sequence was mapped to minimize the distance of these crosslinks , with them localizing to two groups .\",\n \"( D–F ) intermolecular crosslinks added to model show that the docked sequence maps well to the c-ring , but not as well to the b subunit . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02510 . 7554/eLife . 21598 . 026Figure 6—figure supplement 4 . Strong density near Arg210 . Arg210 of subunit a shown as blue stick , subunit a shown as orange cartoon , c-ring shown as grey cartoon and State one density shown as white surface . DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 02610 . 7554/eLife . 21598 . 027Figure 6—figure supplement 5 . Functional mutants of E . coli F-ATPase subunit a . As in main text Figure 6B , but with residues labeled and referenced ( superscript number ) .\",\n \"( 1 . Cain and Simoni , 1986; 2 . Lightowlers et al . , 1987; 3 . Lightowlers et al . , 1988; 4 . Cain and Simoni , 1989; 5 . Hatch et al . , 1995; 6 . Angevine and Fillingame , 2003; 7 . Angevine et al . , 2003 ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 21598 . 027\"\n ],\n [\n \"The three models and maps were deposited in the pdb and emDB with codes 5T4O , EMD-8357 ( State 1 ) , 5 T4P , EMD-8358 ( State 2 ) , 5T4Q and EMD-8359 ( State 3 ) .\"\n ],\n [\n \"A cysteine-free version of E . coli F-ATPase cloned in plasmid pFV2 and expressed in E . coli DK8 strain was used ( Ishmukhametov et al . , 2005 ) .\",\n \"Cells were grown at 37°C in LB medium supplemented with 100 μg/ml ampicillin , for 4–5 hr .\",\n \"The harvested cells were resuspended in lysis buffer containing 50 mM Tris/Cl pH 8 . 0 , 100 mM NaCl , 5 mM MgCl2 , 0 . 1 mM EDTA , 2 . 5% glycerol and 1 μg/ml DNase I and processed by one pass in French press at 20 kPSI .\",\n \"Cellular debris was removed by centrifuging at 7700 × g for 15 min , and the membranes were collected by ultracentrifugation at 100 , 000 × g for 1 hr .\",\n \"The ATP synthase complex was extracted from membranes at 4°C for 1 hr by resuspending the pellet in extraction buffer consisting of 20 mM Tris/Cl , pH 8 . 0 , 300 mM NaCl , 2 mM MgCl2 , 100 mM sucrose , 20 mM imidazole , 10% glycerol , 4 mM digitonin and EDTA-free protease inhibitor tablets ( Roche ) .\",\n \"The complex was then purified by binding on Talon resin ( Clontech ) and eluted in 150 mM imidazole .\",\n \"The protein was further purified and sugars removed by size exclusion chromatography on a 16/60 Superose six column equilibrated in a buffer containing 20 mM Tris/Cl pH 8 . 0 , 100 mM NaCl , 4 mM digitonin and 2 mM MgCl2 .\",\n \"The purified protein was then concentrated to 2 mg/ml for cryo-EM .\",\n \"Seventy microgram of F1Fo was reconstituted into extrusion-preformed 100 nm soybean phosphatidylcholine liposomes exactly as descried ( Ishmukhametov et al . , 2016 ) .\",\n \"Proton pumping by proteoliposomes was studied using quenching of a pH sensitive fluorescent probe 9-Amino-6-Chloro-2-Methoxyacridine ( ACMA ) exactly as described ( Ishmukhametov et al . , 2016 ) .\",\n \"The assay was performed with 100 µl of proteoliposomes in 2-ml cuvettes .\",\n \"The reaction was started with 0 . 25 mM ATP and stopped by 2 µM of the uncoupler FCCP .\",\n \"ATP hydrolase activity and its stimulation by LDAO was measured with ATP regenerating system using 5 µg of the protein with 1 mM ATP exactly as described ( Ishmukhametov et al . , 2016 ) .\",\n \"DCCD inhibition of proteoliposomes was done as described ( Ishmukhametov et al . , 2016 ) .\",\n \"DCCD inhibition of pure F1Fo was done as described ( Ishmukhametov et al . , 2005 ) , with the following modification .\",\n \"Ten microgram of the protein was incubated in 1 ml buffer A ( 50 mM MES , pH 6 . 4 , 100 mM KCl , 1 mM MgCl2 ) with 50 µM DCCD for 30 min at room temperature .\",\n \"Control sample contained 1% ethanol instead of DCCD .\",\n \"Reaction was started by mixing the inhibited protein with 1 ml of buffer A containing all the components of ATP regenerating system .\",\n \"All the functional experiments presented here were repeated two to three times using the protein isolated by MS and shipped to RI at liquid N2 temperature .\",\n \"Results of typical experiments are shown .\",\n \"Aliquots of 4 μl of purified E . coli F-ATPase at a concentration of 3 . 58 μM were placed on glow-discharged holey carbon grids ( Quantifoils Copper R2/2 , 200 Mesh ) .\",\n \"Grids were blotted for 2 s and flash-frozen in liquid ethane using an FEI Vitrobot Mark IV .\",\n \"Grids were transferred to an FEI Titan Krios transmission electron microscope that was operating at 300 kV .\",\n \"Images were recorded automatically using the FEI EPU software , yielding a pixel size of 1 . 4 Å .\",\n \"A dose rate of 29 electrons ( spread over 20 frames ) per Å2 per second , and an exposure time of 2 s were used on the Falcon-II detector .\",\n \"8640 movies were collected .\",\n \"MotionCorr ( Li et al . , 2013 ) was used to correct local beam-induced motion and to align resulting frames .\",\n \"Defocus and astigmatism values were estimated using CTFFIND4 ( Rohou and Grigorieff , 2015 ) , and 252 micrographs were excluded due to drift or excessive ice contamination .\",\n \"1208 particles were manually picked and subjected to 2D classification to generate templates for autopicking in RELION ( Scheres , 2012 ) .\",\n \"The automatically picked micrographs were manually inspected to remove false positives , finally yielding 395 , 140 particles .\",\n \"These particles then underwent two rounds of 2D classification to generate 22 classes with 311 , 887 particles .\",\n \"The final particles were classified into four 3D classes using a previously generated model from a low-resolution data set of the same sample ( unpublished ) , low-pass filtered to 60 Å .\",\n \"The resolution was estimated using Fourier Shell Correlation ( FSC = 0 . 143 , gold-standard ) .\",\n \"Three of the four classes containing 104 , 510 ( State 1 ) , 67 , 829 ( State\",\n \"2 ) and 53 , 587 ( State\",\n \"3 ) particles were movie-refined and post-processed in RELION producing maps at 7 . 4 , 7 . 8 , 8 . 5 Å , respectively ( Figure 2—figure supplement 10 ) .\",\n \"State 1 was further processed using masked classification ( Bai et al . , 2015 ) with residual signal subtraction with a mask created by removing parts of the detergent micelle .\",\n \"Three out of the four classes from this classification containing 95 , 345 particles were combined and refined to generate the final 6 . 9 Å map .\",\n \"Figure 2—figure supplement 3 is a summary of these methods , shown as a flowchart .\",\n \"Local resolution of different parts of the complex was estimated using RELION and ResMap ( Kucukelbir et al . , 2014 ) .\",\n \"Crystal and NMR structures of subunits from E . coli ( αβγε - 3oaa [Cingolani and Duncan , 2011] , δ - 2a7u [Wilkens et al . , 2005] , b - 1b9u [Dmitriev et al . , 1999] , 1l2p [Del Rizzo et al . , 2002] and 2khk [Priya et al . , 2009] ) and related organisms ( c - 3u2f [Symersky et al . , 2012] and a - 5fik [Zhou et al . , 2015] ) were rigid body docked into the highest resolution cryo-EM map and the side chains ‘pruned’ to Cα .\",\n \"The sequence was mapped to subunit a using crosslinks as restraints .\",\n \"Subsequent manual model building and refinement was performed with Coot ( Emsley et al . , 2010 ) , Phenix ( Adams et al . , 2010 ) and Refmac ( Murshudov et al . , 2011 ) ( excluding subunit c due to weak density ) , with crosslinks again used as external restraints ( a summary of the types of models used to build the initial model can be found in Figure 2—figure supplement 6 ) .\",\n \"Nucleotide occupancy was determined by first building the model without any nucleotide present , and then segmenting the map and selecting any density with 15% overlap with atoms and deleting this density .\",\n \"Nucleotide was subsequently docked into this difference density using the known positions from previous structures .\",\n \"Once a complete model was built of the highest resolution map , this was docked and refined to the other two maps to create three models .\",\n \"The three models and maps were deposited in the pdb and emDataBank with codes 5T4O , EMD-8357 ( State 1 ) , 5 T4P , EMD-8358 ( State 2 ) , 5T4Q and EMD-8359 ( State 3 ) .\",\n \"Data statistics shown in Figure 2—source data 1 .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["A molecular model that provides a framework for interpreting the wealth of functional information obtained on the E . coli F-ATP synthase has been generated using cryo-electron microscopy .","Three different states that relate to rotation of the enzyme were observed , with the central stalk’s ε subunit in an extended autoinhibitory conformation in all three states .","The Fo motor comprises of seven transmembrane helices and a decameric c-ring and invaginations on either side of the membrane indicate the entry and exit channels for protons .","The proton translocating subunit contains near parallel helices inclined by ~30° to the membrane , a feature now synonymous with rotary ATPases .","For the first time in this rotary ATPase subtype , the peripheral stalk is resolved over its entire length of the complex , revealing the F1 attachment points and a coiled-coil that bifurcates toward the membrane with its helices separating to embrace subunit a from two sides ."],"string":"[\n \"A molecular model that provides a framework for interpreting the wealth of functional information obtained on the E . coli F-ATP synthase has been generated using cryo-electron microscopy .\",\n \"Three different states that relate to rotation of the enzyme were observed , with the central stalk’s ε subunit in an extended autoinhibitory conformation in all three states .\",\n \"The Fo motor comprises of seven transmembrane helices and a decameric c-ring and invaginations on either side of the membrane indicate the entry and exit channels for protons .\",\n \"The proton translocating subunit contains near parallel helices inclined by ~30° to the membrane , a feature now synonymous with rotary ATPases .\",\n \"For the first time in this rotary ATPase subtype , the peripheral stalk is resolved over its entire length of the complex , revealing the F1 attachment points and a coiled-coil that bifurcates toward the membrane with its helices separating to embrace subunit a from two sides .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["ATP synthase is a biological motor that produces a molecule called adenosine tri-phosphate ( ATP for short ) , which acts as the major store of chemical energy in cells .","A single molecule of ATP contains three phosphate groups: the cell can remove one of these phosphates to make a molecule called adenosine di-phosphate ( ADP ) and release energy to drive a variety of biological processes .","ATP synthase sits in the membranes that separate cell compartments or form barriers around cells .","When cells break down food they transport hydrogen ions across these membranes so that each side of the membrane has a different level ( or “concentration” ) of hydrogen ions .","Movement of hydrogen ions from an area with a high concentration to a low concentration causes ATP synthase to rotate like a turbine .","This rotation of the enzyme results in ATP synthase adding a phosphate group to ADP to make a new molecule of ATP .","In certain conditions cells need to switch off the ATP synthase and this is done by changing the shape of the central shaft in a process called autoinhibition , which blocks the rotation .","The ATP synthase from a bacterium known as E . coli – which is commonly found in the human gut –has been used as a model to study how this biological motor works .","However , since the precise details of the three-dimensional structure of ATP synthase have remained unclear it has been difficult to interpret the results of these studies .","Sobti et al . used a technique called Cryo-electron microscopy to investigate the structure of ATP synthase from E . coli .","This made it possible to develop a three-dimensional model of the ATP synthase in its autoinhibited form .","The structural data could also be split into three distinct shapes that relate to dwell points in the rotation of the motor where the rotation has been inhibited .","These models further our understanding of ATP synthases and provide a template to understand the findings of previous studies .","Further work will be needed to understand this essential biological process at the atomic level in both its inhibited and uninhibited form .","This will reveal the inner workings of a marvel of the natural world and may also lead to the discovery of new antibiotics against related bacteria that cause diseases in humans ."],"string":"[\n \"ATP synthase is a biological motor that produces a molecule called adenosine tri-phosphate ( ATP for short ) , which acts as the major store of chemical energy in cells .\",\n \"A single molecule of ATP contains three phosphate groups: the cell can remove one of these phosphates to make a molecule called adenosine di-phosphate ( ADP ) and release energy to drive a variety of biological processes .\",\n \"ATP synthase sits in the membranes that separate cell compartments or form barriers around cells .\",\n \"When cells break down food they transport hydrogen ions across these membranes so that each side of the membrane has a different level ( or “concentration” ) of hydrogen ions .\",\n \"Movement of hydrogen ions from an area with a high concentration to a low concentration causes ATP synthase to rotate like a turbine .\",\n \"This rotation of the enzyme results in ATP synthase adding a phosphate group to ADP to make a new molecule of ATP .\",\n \"In certain conditions cells need to switch off the ATP synthase and this is done by changing the shape of the central shaft in a process called autoinhibition , which blocks the rotation .\",\n \"The ATP synthase from a bacterium known as E . coli – which is commonly found in the human gut –has been used as a model to study how this biological motor works .\",\n \"However , since the precise details of the three-dimensional structure of ATP synthase have remained unclear it has been difficult to interpret the results of these studies .\",\n \"Sobti et al . used a technique called Cryo-electron microscopy to investigate the structure of ATP synthase from E . coli .\",\n \"This made it possible to develop a three-dimensional model of the ATP synthase in its autoinhibited form .\",\n \"The structural data could also be split into three distinct shapes that relate to dwell points in the rotation of the motor where the rotation has been inhibited .\",\n \"These models further our understanding of ATP synthases and provide a template to understand the findings of previous studies .\",\n \"Further work will be needed to understand this essential biological process at the atomic level in both its inhibited and uninhibited form .\",\n \"This will reveal the inner workings of a marvel of the natural world and may also lead to the discovery of new antibiotics against related bacteria that cause diseases in humans .\"\n]","isConversational":false},"year":{"kind":"string","value":"2016"}}},{"rowIdx":89,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["cell biology","neuroscience"],"string":"[\n \"cell biology\",\n \"neuroscience\"\n]","isConversational":false},"title":{"kind":"string","value":"Modeling of axonal endoplasmic reticulum network by spastic paraplegia proteins"},"id":{"kind":"string","value":"elife-23882-v2"},"sections":{"kind":"list like","value":[["Axons allow long-range bidirectional communication in neurons .","They carry action potentials along their plasma membrane from dendrites and cell body to presynaptic terminals; and they transport cell components and signaling complexes using motor proteins , anterogradely and retrogradely .","Another potential route for communication along axons is endoplasmic reticulum ( ER ) ; axons possess a network of ER tubules , which appears physically continuous both locally ( Tsukita and Ishikawa , 1976; Villegas et al . , 2014 ) and with ER throughout the neuron ( Lindsey and Ellisman , 1985; Terasaki et al . , 1994 ) .","The physical continuity of ER and its ensuing potential for long-distance communication has been likened to a ‘neuron within a neuron’ ( Berridge , 1998 ) .","Axonal ER appears mostly tubular and smooth , with some cisternae ( Tsukita and Ishikawa , 1976; Lindsey and Ellisman , 1985; Villegas et al . , 2014 ) .","Some rough ER is likely to be present too: numerous mRNAs are found in both growing and mature axons ( Zivraj et al . , 2010; Shigeoka et al . , 2016 ) , and local axonal translation can occur in response to injury ( Ben-Yaakov et al . , 2012; Perry et al . , 2012 ) .","However , rough ER sheets ( Tsukita and Ishikawa , 1976; Villegas et al . , 2014 ) , and markers of protein export and folding that are characteristic of rough ER ( Röper , 2007; O'Sullivan et al . , 2012 ) , are relatively sparse in mature axons .","Axonal ER therefore likely has major roles other than protein export; these could include lipid biosynthesis ( Tidhar and Futerman , 2013; Vance , 2015 ) , calcium homeostasis and signaling ( Ross , 2012 ) , and coordination of organelle physiology ( Phillips and Voeltz , 2016 ) .","The continuity of the ER network suggests that some of these roles might have long-range as well as local functions; indeed , an ER-dependent propagating calcium wave is seen after axotomy of Caenorhabditis elegans or mammalian dorsal root ganglion neurons ( Ghosh-Roy et al . , 2010; Cho et al . , 2013 ) .","A strong hint of the importance of ER in axons is found in Hereditary Spastic Paraplegia ( HSP ) , a group of axon degeneration disorders characterized by progressive spasticity and weakness of the lower limbs ( Blackstone et al . , 2011; Blackstone , 2012 ) .","Mutations affecting spastin , atlastin-1 , reticulon-2 , REEP1 and REEP2 account for most cases of autosomal dominant ‘pure’ HSP ( Hazan et al . , 1999; Zhao et al . , 2001; Züchner et al . , 2006; Montenegro et al . , 2012; Esteves et al . , 2014 ) .","These proteins share a common feature of one or two hydrophobic hairpin-loops inserted in the ER membrane , promoting ER membrane curvature in a process termed hydrophobic wedging ( Voeltz et al . , 2006 ) .","Proteins of the REEP and reticulon families localize preferentially to tubular or smooth ER , and their loss results in disruption of ER tubular organization ( Shibata et al . , 2006; Voeltz et al . , 2006; Park et al . , 2010; Shibata et al . , 2010 ) ; they may also contribute to modeling of rough ER sheets by stabilizing their curved edges ( Shibata et al . , 2009 ) .","What is the link between ER modeling and axon structure and function ?","HSP-causing mutations often appear to cause loss of protein expression or function ( Beetz et al . , 2013; Novarino et al . , 2014 ) , and the ability of hairpin-loop proteins to form homomeric and heteromeric complexes ( Shibata et al . , 2008 ) allows some point mutations to have dominant negative effects ( Züchner et al . , 2006; Beetz et al . , 2012 ) .","Therefore loss of normal ER modeling appears to compromise axon maintenance and function .","Given the roles of hairpin-loop proteins in ER modeling , we aimed to test the model that hairpin-loop-containing HSP proteins organize the axonal ER network .","Since reticulon and REEP family proteins are redundantly required for most peripheral ER tubules in yeast ( Voeltz et al . , 2006 ) , we focus on the requirement for these two families in axons .","We previously showed that knockdown of the Drosophila reticulon Rtnl1 causes expansion of epidermal ER sheets , and partial loss of smooth ER marker from distal but not proximal motor axons ( O'Sullivan et al . , 2012 ) .","Here we show that REEP proteins have similar roles .","We also show that simultaneous loss of reticulon and REEP family members leads to a range of axonal ER phenotypes , including a reduced network with fewer and larger tubules , and occasional gaps in the network .","Our work implicates hairpin-loop-containing HSP proteins as important players in the axonal ER network , and suggests further models for how the network is organized ."],["The reticulon and REEP families of double-hairpin-containing proteins are collectively responsible for formation or maintenance of most peripheral ER tubules in yeast ( Voeltz et al . , 2006 ) .","We previously showed that the Drosophila reticulon ortholog Rtnl1 was strongly localized in axons , and that its knockdown caused partial loss of a smooth ER marker in posterior larval segmental axons ( O'Sullivan et al . , 2012 ) .","To test the roles of Drosophila REEP proteins in axonal ER localization , we first dissected the ortholog relationships between the six Drosophila and six human REEP proteins .","Multiple sequence alignment of mammalian and Drosophila REEP protein sequences suggested that CG42678 was the single Drosophila ortholog of mammalian REEP1-REEP4 ( Figure 1A ) .","CG42678 has previously been designated Reep1 ( http://flybase . org/reports/FBgn0261564 . html ) , but we propose the name ReepA to reflect its orthology to the four mammalian genes REEP1-REEP4 .","Mammalian REEP5 and REEP6 appeared to share two Drosophila orthologs , CG8331 and CG4960 ( Figure 1 ) .","We designated CG8331 as ReepB because of its widespread expression ( www . flyatlas . org; Chintapalli et al . , 2007 ) and slower rate of evolutionary sequence divergence ( Supplementary file 1 ) .","We excluded CG4690 and three additional Drosophila REEP genes from further study , since their expression was restricted to testes and larval fat body ( www . flyatlas . org; Chintapalli et al . , 2007 ) , and their faster rate of evolutionary sequence divergence ( Supplementary file 1 , and reflected in longer branch lengths in Figure 1A ) , suggesting poorly conserved function . 10 . 7554/eLife . 23882 . 003Figure 1 . Drosophila Rtnl1 and REEP genes and products .","( A ) A dendrogram based on ClustalW sequence alignment of Drosophila and human REEP proteins shows two branches corresponding to human REEP1-4 ( CG42678 ) and human REEP5-6 .","Broken lines represent Drosophila REEP proteins that are evolving rapidly ( Supplementary file 1 ) , reflected by longer branch lengths .","Sequences used are NP_075063 . 1 , NP_057690 . 2 , NP_001001330 . 1 , NP_079508 . 2 , NP_005660 . 4 , NP_612402 . 1 , NP_726266 . 1 , NP_610936 . 2 , NP_651429 . 1 , NP_611831 . 2 , NP_572730 . 1 , NP_726366 . 1 .","( B ) Rtnl1 genomic and transcript map , showing the region deleted in Rtnl1− by excision of P-element NP7026 ( Wakefield and Tear , 2006; Figure 1—figure supplement 1 ) , the RHD domain ( Pfam 02453 ) and its coordinates in protein isoform G , the Rtnl1::YFP exon trap insertion CPTI001291 ( green triangle ) , and the fragment targeted by GD RNAi 7866 .","Map and coordinates are from the Drosophila Genome Browser ( www . flybase . org , version R6 . 04 ) , here and in subsequent panels; light regions in transcripts represent coding regions , dark shaded regions represent untranslated regions .","( C ) ReepA genomic and transcript map showing the region deleted in ReepA− by excision of P-element CB-0501–3 , the position of the DP1 domain ( Pfam 03134 ) and its coordinates in protein isoforms H and J . GFP insertion sites for ReepA1::GFP , ReepA2::GFP and ReepA3::GFP fusions are shown with green triangles .","( D ) ReepB genomic and transcript map showing the region deleted in ReepB− by excision of P-element EY05130 , the position of the DP1 domain ( Pfam 03134 ) and its coordinates in protein isoforms A and D ) .","( E–I )","Confocal sections showing localization of ReepA::GFP isoforms and ReepB::GFP .","( E ) Overlap of ReepA1::GFP and ReepB::GFP with anti-KDEL labeling in larval epidermal cells .","To facilitate display of weaker ReepA1::GFP , the GFP channel in wild-type control ( WT ) and ReepA::GFP images has been brightened four times as much as for ReepB::GFP .","( F ) Expression of ReepA3::GFP and ReepB::GFP in third instar ventral nerve cord .","The GFP channels for WT and ReepA::GFP have been brightened by twice as much as for ReepB::GFP .","VNCs are outlined with yellow dashed lines .","Arrowheads show ReepB::GFP extending into peripheral nerves .","( G ) A single confocal section of a peripheral nerve , showing ReepB::GFP localized continuously along its length .","( H ) Double labeling of an NMJ for ReepB::GFP and the mainly postsynaptic marker Dlg , showing ReepB::GFP in an axon emerging from nerve bundles ( arrowhead ) to extend to the NMJ , as well as in axons traversing the muscle surface ( arrow ) ; note the dispersed ReepB::GFP staining also in the underlying muscle .","( I ) Double labeling of ReepA::GFP and ReepB::GFP lines for GFP and Dlg ( mainly postsynaptic ) shows presynaptic expression of ReepB::GFP .","( J ) GFP expression in a wildtype negative control ( WT ) , or from Rtnl1::GFP expressed in two closely apposed motor neurons by m12-GAL4 .","Scale bars 10 µm , except F , 20 µm . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 00310 . 7554/eLife . 23882 . 004Figure 1—figure supplement 1 . Molecular lesions in Rtnl11 , ReepA541 and ReepB48 . BLASTN searches of the Drosophila genome , using the sequences of ( A ) Rtnl11 , ( B ) ReepA541 and ( C ) ReepB48 mutations as queries , show the coordinates of each molecular lesion in the genome , as gaps in alignments between mutant and genomic sequences .","The ReepA541 and ReepB48 sequences highlighted in yellow show P element footprints left at the excision sites . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 004 A Rtnl1::YFP exon trap , Rtnl1CPTI001291 ( Figure 1B ) was previously shown to localize to ER , including in axons ( O'Sullivan et al . , 2012 ) .","To study localization of ReepA and ReepB , we recombineered C-terminal GFP-tagged versions of these ( Figure 1C , D ) using P[acman] genomic clones ( Venken et al . , 2009 ) .","For ReepA , we generated EGFP fusions at three different C-termini ( Figure 1C ) , that we called ReepA1::GFP ( for protein isoforms D , E , G ) , ReepA2::GFP ( for protein isoforms H , I , J , K ) and ReepA3::GFP ( for protein isoforms L , R , S ) .","In epidermal cells , we detected weak expression of ReepA1::GFP , but not of the other ReepA::GFP fusions; ReepB::GFP showed stronger expression still , and both fusions overlapped with an ER marker ( KDEL; Figure 1E ) , similar to REEP proteins in other organisms ( Voeltz et al . , 2006; Shibata et al . , 2008; Park et al . , 2010 ) .","ReepB::GFP was also more strongly expressed in third instar larval CNS than ReepA3::GFP , which was the only ReepA::GFP fusion that we detected there ( Figure 1F ) .","ReepB::GFP , but no ReepA::GFP fusion , was also detected in segmental nerves ( Figure 1F , G ) , and in individual axons emerging from nerve bundles leading to the NMJ ( Figure 1H ) .","ReepB::GFP , but none of the ReepA::GFP fusions , localized as a mostly continual structure along the length of presynaptic terminals of neuromuscular junctions ( NMJs ) ( Figure 1H , I ) .","We did not have a UAS-ReepB construct available , but UAS-Rtnl1::GFP ( Rao et al . , 2016 ) expressed in two adjacent motor neurons using m12-GAL4 ( Xiong et al . , 2010 ) also showed strong axonal localization .","We therefore conclude that ReepB::GFP localizes in axonal and presynaptic ER , similar to Rtnl1::YFP ( O'Sullivan et al . , 2012 ) , but that none of the ReepA::GFP fusions is detectable in axonal or presynaptic ER .","An Rtnl1 loss-of-function mutant , Rtnl11 ( Wakefield and Tear , 2006 ) , hereafter referred to as Rtnl1− , lacks the hydrophobic hairpin loop domain ( RHD region ) that induces ER membrane curvature ( Figure 1B; Figure 1—figure supplement 1 ) .","We also used P-transposase-mediated imprecise excision to generate ReepA− and ReepB− mutants , both of which lack most of the curvature-mediating DP1 hairpin domains ( Figure 1B , C; Figure 1—figure supplement 1 ) .","First , we asked whether Drosophila reticulon and REEP proteins contribute to ER network organization in third instar larval epidermal cells; Rtnl1 knockdown leads to a more diffuse ER network organization and expansion of ER sheets in these cells , compared to wild-type controls ( O'Sullivan et al . , 2012 ) .","Rtnl1− mutant larvae also showed loss of ER network organization ( Figure 2A ) .","Intensity of KDEL labeling along a line from the nucleus to cell periphery displayed fluctuating intensity , reflecting the reticular distribution of KDEL in wild-type larvae , but less fluctuation in Rtnl1− larvae; overall levels of KDEL remained unchanged ( Figure 2A ) .","Similarly , loss of both ReepA and ReepB , but not loss of either gene alone , made KDEL levels in larval epidermal cells fluctuate less than in controls .","Mean intensity of KDEL staining was decreased in all ReepA and ReepB mutant genotypes , and fluctuation in intensity was increased in ReepB− larvae ( Figure 2B ) .","Since confocal analysis suggests altered organization of ER in epidermal cells , but does not have sufficient resolution to reveal the details of how it is altered , we performed electron microscopy of epidermal cells .","This revealed that ReepA− ReepB− double mutant epidermal cells showed longer ribosome studded sheet ER profiles , compared to controls , and to ReepA− and ReepB− single mutants ( Figure 2C ) .","This mutant phenotype is similar to , although less severe than loss of Rtnl1 ( O'Sullivan et al . , 2012 ) .","ReepA− ReepB− mutants also showed increased ER stress in epidermal cells ( Figure 2D ) but not in CNS ( Figure 2E ) .","Therefore , Rtnl1 and REEP proteins shape the ER network in Drosophila , and their loss disrupts ER organization , causing longer ER sheet profiles . 10 . 7554/eLife . 23882 . 005Figure 2 . Rtnl1− mutants and ReepA− ReepB− double mutants show disrupted ER organization in epidermal cells .","( A ) KDEL distribution in third instar larval epidermal cells appears diffuse in Rtnl1− larvae compared to a more reticular staining in wild-type ( WT ) larvae .","Micrographs show 1 . 5 µm z-projections of confocal sections .","KDEL intensity along yellow lines from nuclear envelope to the cell periphery shows less fluctuation in Rtnl1− ( blue dotted line on line graph ) than in WT larvae ( black line on line graph ) .","Fluctuation of intensity is quantified using the local normalized variance of intensity ( variance/mean for rolling 10-pixel windows ) over a 12 µm line for each cell .","n = 61 epidermal cells from 12 larvae , 3–5 cells from each larva , from 3 independent experiments .","( B ) KDEL distribution in third instar larval epidermal cells shows less spatial fluctuation in ReepA− ReepB− double mutant larvae compared to ReepA+ ( WT ) , and ReepA− and ReepB− single mutant larvae .","This is confirmed by quantification as in A ) ; overall KDEL intensity is reduced in all ReepA and ReepB mutant genotypes .","n = 45–61 epidermal cells in total from 12 to 16 different larvae , 3–5 cells from each larva , from 6 independent experiments .","Intensity datapoints are experimentwise averages of multiple larvae , compared by paired t-tests , since larva-wise data were too significantly different to pool across experiments in this case .","( C ) Electron micrographs show increased ER sheet profile length in ReepA− ReepB− double mutant third instar larval epidermal cells compared to single ReepA− and ReepB− mutants and controls .","Arrowheads , ER sheets; M , mitochondria .","( n = 56–58 cells in total from three independent larvae , 18–19 cells from each . ( D , E )","ReepA− ReepB− double mutant larvae have an increased ER stress response compared to a ReepA+ ( WT ) control , measured by Xbp1::GFP expression , in larval epidermis ( D ) but not in neuronal cell bodies ( E ) .","Graphs show quantification of Xbp1::GFP staining intensity relative to controls .","( n = 12–15 larvae ) .","All graphs show individual datapoints and mean ±SEM .","Occasional outlier datapoints off the top of the scale are omitted from graphs but included in statistical analyses .","ns , p>0 . 05; *p<0 . 02; **p<0 . 005; ***p<0 . 0003; ****p<0 . 0001 , two-tailed Student’s t-test .","Scale bars: A , B , D , E , 10 µm; C , 0 . 5 µm ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 005 To understand the roles of reticulon and REEPs in axonal ER organization , we labeled axonal ER by expressing Acsl::myc ( O'Sullivan et al . , 2012 ) in two adjacent motor neurons using m12-GAL4 ( Xiong et al . , 2010 ) .","Loss of Rtnl1 caused partial loss of Acsl::myc from posterior ( segment A6 ) axons , but not from anterior axons ( segment A2 ) ; it also caused Acsl::myc staining to appear more irregular in posterior but not anterior axons , reflected in a higher coefficient of variation ( SD/mean ) of Acsl::myc staining intensity along the length of posterior axons lacking Rtnl1 , compared to wild-type ( Figure 3A ) .","These Rtnl1 loss-of-function phenotypes were found either on targeted knockdown of Rtnl1 in these motor axons or in Rtnl1− mutant larvae , and the Rtnl1− mutant phenotypes could be partially rescued by one copy of an Rtnl1Pacman genomic clone ( Figure 3A ) . 10 . 7554/eLife . 23882 . 006Figure 3 . Loss of either Rtnl1 or ReepB leads to partial loss of smooth ER marker from distal motor axons .","( A ) Effects of Rtnl1 loss on ER , visualized by Acsl::myc expressed in two adjacent motor axons using m12-GAL4 .","Images show effects of Rtnl1 knockdown , Rtnl1− mutation , rescue of Rtnl1− by Rtnl1Pacman and respective ReepA+ control ( WT ) axons .","Graphs show mean staining intensity ( top graphs ) or coefficient of variation of intensity ( bottom graphs ) as a measurement of staining variability along the length of each axon .","Rtnl1 loss leads to partial loss of Acsl::myc in posterior but not anterior axons ( top graphs ) , and to some disorganization of posterior axonal ER seen by increased coefficient of variation of Acsl::myc staining intensity .","There is partial rescue of Rtnl1 phenotypes by one copy of a Rtnl1Pacman genomic clone .","n = 15–20 larvae per genotype pooled from 5 independent experiments for RNAi; n = 24–36 larvae per genotype from 10 to 12 independent experiments for Rtnl1− mutant .","Graphs show individual datapoints with mean ±SEM; ns , p>0 . 05; *p<0 . 03; **p<0 . 005; two-tailed unpaired Student’s t-test; two-tailed paired Student’s t-test was used when larva-wise data were too significantly different across experiments ( p<0 . 05 , ANOVA ) to pool .","( B ) Effects of ReepA or ReepB loss on axonal ER .","ReepB knockdown or ReepB− mutation , but not ReepA mutation , causes partial loss of smooth ER marker Acsl::myc in posterior but not in anterior motor axons .","The ReepB− phenotype can be mostly rescued by one copy of a genomic ReepB::GFP clone .","A ReepA− ReepB− double mutant shows a similar phenotype to a ReepB− single mutant .","ReepA− ReepB− double mutant , but not single ReepA− or ReepB− mutants , show increased coefficient of variation of Acsl::myc staining levels in posterior axons; n = 7–47 larvae from 4 to 11 independent experiments .","All graphs show individual datapoints with mean ±SEM; occasional outlier datapoints off the top of the scale are omitted from graphs but included in statistical analyses .","ns , p>0 . 05; *p<0 . 04 **p<0 . 01 .","ReepB RNAi was analyzed using two-tailed unpaired Student’s t-tests; multiple comparisons of ReepA and ReepB mutant genotypes were analyzed by ANOVA followed by post-hoc Tukey HSD tests .","Scale bars , 10 µm . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 006 ReepA− mutants showed no loss of Acsl::myc from axons .","Similar to Rtnl1 loss of function , ReepB− mutants showed partial loss of Acsl::myc from posterior but not anterior motor axons ( Figure 3B ) ; this phenotype was partially rescued with one copy of a genomic ReepB::GFP clone , and was also observed on ReepB knockdown in m12-GAL4-expressing neurons ( Figure 3B ) .","A ReepA− ReepB− double mutant showed loss of Acsl::myc from posterior axons , that was similar to that seen in ReepB− mutants ( Figure 3B ) .","ReepA− ReepB− double mutant larvae , but not ReepA− single mutants , showed an increased coefficient of variation of Acsl::myc staining intensity in posterior but not in anterior axons; ReepB mutant and knockdown axons both showed a slight increase in coefficient of variation in posterior axons , although this was not significant for ReepB mutants ( Figure 3B ) .","In summary , loss of either Rtnl1 or at least ReepB alters axonal ER distribution in posterior motor axons , therefore supporting roles for these two protein families in axonal ER organization; in contrast loss of ReepA has either subtle or no effects .","Since loss of both reticulon and REEP families in yeast removes most peripheral ER tubules ( Voeltz et al . , 2006 ) , we tested whether loss of both protein families in Drosophila might have similarly severe effects on axonal ER .","Rtnl1− ReepA− ReepB− triple mutants were viable and fertile as adults , but survived poorly beyond two weeks of adulthood , compared to 4–5 weeks for wild-type adults .","Triple mutant larvae showed increased fluctuation of Acsl::myc staining intensity along motor axons compared to wild-type , mainly in middle parts ( segment A4 and A5 ) of longer axons .","At its most extreme , this manifested as fragmentation of Acsl::myc labeling , which was never seen in wild-type axons ( Figure 4A , B ) .","When gaps in labeling were found in the central regions of axons , labeling in the anterior and posterior parts of the same axons was usually continuous ( Figure 4A ) .","Double labeling of plasma membrane ( CD4::tdGFP ) and axonal ER ( Acsl::myc ) showed no effect on axonal plasma membrane in Rtnl1− ReepA− ReepB− triple mutant compared to control axons ( Figure 4—figure supplement 1 ) , implying that the phenotype was limited to ER and did not affect axon integrity .","To compare genotypes and labels , we quantified irregular labeling along a 45 µm stretch of axon traversing the larval A4/A5 region in two ways: first we measured gaps in labeling , defined as intensity below a threshold that could consistently distinguish labeled axons above background labeling in the nerve; and second , we quantified the coefficient of variation of labeling intensity along the axon ( Figure 4C , D ) .","Rtnl1 loss-of-function axons , but not ReepA− ReepB− mutant axons , also showed a mild ER fragmentation phenotype ( Figure 4A , B ) .","Rtnl1 ( RNAi ) ReepA− ReepB− larvae also showed a fragmentation phenotype similar to Rtnl1− ReepA− ReepB− , suggesting that loss of Rtnl1 is essential for it ( Figure 4B ) .","Rtnl1− ReepB− double mutant axons also showed a similar phenotype to Rtnl1− ReepA− ReepB− triple mutants .","Therefore , loss of Rtnl1 causes a mild irregular organization of axonal ER , and this phenotype is exacerbated by loss of ReepB , with no detectable contribution of ReepA loss; and loss of both Reep proteins has no apparent effect . 10 . 7554/eLife . 23882 . 007Figure 4 . Loss of hairpin proteins leads to discontinuity of axonal ER staining .","( A ) Rtnl1− larvae and Rtnl1− ReepA− ReepB− triple mutant larvae sometimes show fragmented axonal ER labeling in the middle parts ( segment A5 ) of long motor axons that express Acsl::myc under control of m12-GAL4 .","Anterior ( segment A2 ) and posterior ( segment A6 ) portions of the same motor axons show continuous ER labeling .","Arrowheads show gaps in Acsl::myc staining; brighter versions of the same images show gaps in staining in mutants but not wild-type ( WT ) , even when brightness of remaining staining is saturating .","( B ) Panels show three examples of the range of Acsl::myc distributions found in the middle parts of long motor axons of each genotype of ReepA+ ( WT ) , Rtnl1− , Rtnl1− rescued with a Rtnl1Pacman construct , ReepA− ReepB− and Rtnl1− ReepB− double mutants , Rtnl1− ReepA− ReepB− triple mutant , and Rtnl1 ( RNAi ) ReepA− ReepB− larvae .","A variety of phenotypes , from continuous to fragmented Acsl::myc labeling , are found in genotypes that lack Rtnl1 , and tend to be more severe in genotypes that also lack ReepB .","Insets show examples of gaps in ER continuity ( arrowheads ) at higher magnification .","Scale bars 10 µm , and 5 µm in higher zoom images .","( C ) Percentages of a 45 µm length in the middle ( A4/A5 segment ) of each axon that lacks Acsl::myc staining , using an intensity threshold of 20 on a scale of 0–255 .","Individual axons are plotted , together with median and interquartile range; comparisons use Mann-Whitney U-tests .","All genotypes that lack Rtnl1 show more gaps than WT; Rtnl1− ReepA− ReepB− triple mutants do not differ from Rtnl1 ( RNAi ) ReepA− ReepB− or Rtnl1− ReepB− double mutants , but are significantly more severe than Rtnl1− single mutants .","( D ) The coefficient of variation of Acsl::myc labeling in middle axon portions is increased in genotypes lacking Rtnl1 , relative to controls .","Graphs show individual datapoints with mean ±SEM; occasional outlier datapoints off the top of the scale are omitted from graphs but included in statistical analyses .","Comparison between Rtnl1− and Rtnl1− ReepA− ReepB− mutant genotypes was analyzed by two-tailed Student’s t-test , other multiple comparisons by ANOVA followed by Dunnett’s T3 test .","ns , p>0 . 05; *p<0 . 04; **p<0 . 01; ***p<0 . 0005; ****p<0 . 0001; individual axons are plotted from 3 to 20 independent experiments , each with 2–3 different larvae for each genotype . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 00710 . 7554/eLife . 23882 . 008Figure 4—figure supplement 1 . Plasma membrane integrity is not affected in Rtnl1− ReepA− ReepB− triple mutant larvae ) .","Coefficient of variation for plasma membrane labeling along axon length is unchanged in Rtnl1− ReepA− ReepB− triple mutant compared to control ( WT ) axons , whereas the ER to plasma membrane ratio of the coefficient of variation is increased .","Higher zoom images of the boxed areas show gaps ( yellow arrowheads ) and brightly labeled axonal ER accumulations .","Graphs show mean ±SEM; *p<0 . 02 , two-tailed Student’s t-test; n = 5 different larvae from 3 independent experiments . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 008 To confirm the physical continuity of ER in normal axons , and that gaps in Acsl::myc staining in Rtnl1− ReepA− ReepB− triple mutant axons reflect physical gaps in the ER network , we also visualized ER using fluorescent protein markers .","The lipase CG9186::GFP ( Thiel et al . , 2013 ) , Rtnl1::GFP ( Rao et al . , 2016 ) and tdTomato::Sec61β ( Summerville et al . , 2016 ) all localize to ER , and all showed continuous labeling in wild-type ( ReepA+ ) axons ( Figure 5A; Figure 5—figure supplement 1A ) .","However , CG9186::GFP showed occasional gaps in triple mutant axons ( Figure 5A ) , with at least one gap per larva visible by confocal microscopy in 6/17 larvae .","Since confocal imaging might not reveal all physical discontinuities , and since gaps in marker distribution might not mean gaps in ER distribution , we probed the physical continuity of CG9186::GFP labeling using fluorescence recovery after photobleaching ( FRAP ) .","After bleaching a 12 µm length of axon in either wild-type axons or in Rtnl1− ReepA− ReepB− triple mutant axons lacking gaps , we observed rapid recovery of fluorescence from both ends of the bleached region ( Figure 5; Figure 5—figure supplement 1B; Video 1 ) , suggesting no physical barrier to CG9186::GFP diffusion .","The kinetics of recovery were similar between wildtype and triple mutant axons ( Figure 5B–D ) .","We also performed FRAP on regions next to gaps in CG9186::GFP labeling in triple mutant axons ( Figure 5A ) ; here we observed good recovery of fluorescence from the end of the bleached region opposite the gap , but no recovery across the gap ( Figure 5; Figure 5—figure supplement 1C; Videos 2–4 ) , suggesting that gaps in CG9186::GFP labelling were also physical barriers to diffusion of CG9186::GFP , and hence gaps in the ER network . 10 . 7554/eLife . 23882 . 009Figure 5 . Live imaging of ER in wild type and hairpin mutant axons .","( A ) Representative FRAP assay images from ReepA+ ( WT ) or Rtnl1− ReepA− ReepB− triple mutant axons in which ER was visualized using CG9186::GFP expression driven in two motor axons by m12-GAL4 .","One triple mutant axon shows a gap in labeling ( arrowheads ) .","Regions of interest ( 12 µm , between dashed lines ) were photobleached .","Fluorescence was visualized before photobleaching ( ‘pre’ ) , and during recovery over 200 s .","A kymograph was generated for each axon indicated by an arrow in the top panel .","Most areas of intense and less intense ER labeling remain stable over time ( e . g . black arrows below kymographs ) ; occasional movements of ER features are indicated by white arrows in kymographs .","( B ) Relative fluorescence intensities within the photobleached region were plotted ( mean ±SEM ) during recovery for wild-type axons or for triple mutant axons lacking ER gaps .","( C–D )","Quantification of half recovery time ( C ) and rate constant ( D ) for wild-type and triple mutant axons lacking gaps , with median and interquartile range .","Data in B–D are from seven recordings from 4 wild-type larvae , and 20 recordings from 14 mutant larvae .","ns , p>0 . 05; Mann-Whitney U test .","Scale bars , 5 µm .","( E ) Representative kymographs from time-lapse recording of CG9186::GFP in unbleached single ReepA+ ( WT ) or Rtnl1− ReepA− ReepB− triple mutant axons lacking gaps .","The left panel shows retrograde movement of a ~ 3 µm length of ER labeling ( arrows ) ; the right panel shows anterograde movement of a ~ 5 µm stretch of ER labeling ( arrows ) .","Retrograde movements of labeled puncta are seen in both panels ( arrowheads ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 00910 . 7554/eLife . 23882 . 010Figure 5—figure supplement 1 . Continuity , stability and movement of fluorescent ER labeling in wild-type ( WT ) and Rtnl1− ReepA− ReepB− triple mutant axons .","( A ) Time series and kymographs of two closely apposed motor axons expressing either Rtnl1::GFP or Sec61bβ::tdTomato ER markers under control of m12-GAL4 .","Images show continuity of ER staining in axons , and kymographs show stability of most ER features over the acquisition time .","( B ) Kymographs showing fluorescence recovery after photobleaching in two ReepA+ ( WT ) axons , or in two Rtnl1− ReepA− ReepB− triple mutant axons lacking ER gaps .","Recovery proceeds from both ends of the bleached region ( dotted lines ) .","Most ER features seen before bleaching ( top line of kymographs ) can also be seen after recovery , implying that recovery proceeds by protein diffusion , not by movement of ER structures .","( C ) Kymographs showing fluorescence recovery after photobleaching immediately distal to gaps in ER staining , in four Rtnl1− ReepA− ReepB− triple mutant axons .","Recovery proceeds only from the distal end of the bleached region ( dotted lines ) opposite the gap , but not across the proximal gap .","While most ER gaps and many other features are stable , one axon shows a retrogradely moving particle ( white arrowhead ) , and one shows anterograde movement of ER that closes one gap and opens another more proximally ( white asterisk ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01010 . 7554/eLife . 23882 . 011Video 1 . Representative time-lapse microscopy of FRAP analysis in wild-type axons . Wild-type ( WT ) axons expressing CG9186::GFP under control of m12-GAL4 were photobleached as described in Figure 5A .","The video shows 2 s of pre-bleach images and 200 s of post-bleach images at one frame every 5 s .","The bleached area is shown with a rectangle in the bleaching frame . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01110 . 7554/eLife . 23882 . 012Video 2 . Representative time-lapse microscopy of FRAP analysis in Rtnl1− ReepA− ReepB− triple mutant axons . Rtnl1− ReepA− ReepB− triple mutant axons expressing CG9186::GFP were photobleached and imaged as in Video 1 , in a region immediately distal to a gap ( between white arrows ) in ER labeling in one axon . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01210 . 7554/eLife . 23882 . 013Video 3 . FRAP analysis in Rtnl1− ReepA− ReepB− triple mutant axons showing a CG9186::GFP labeled particle passing retrogradely across the gap region . Rtnl1− ReepA− ReepB− triple mutant axons expressing CG9186::GFP were photobleached and imaged immediately distal to an ER gap as in Video","2 . A labeled particle ( white arrow ) can be seen moving retrogradely across the ER gap ( white arrowhead ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01310 . 7554/eLife . 23882 . 014Video 4 . FRAP analysis in Rtnl1− ReepA− ReepB− triple mutant axons showing dynamic ER gap generation . Rtnl1− ReepA− ReepB− triple mutant axons expressing CG9186::GFP were photobleached and imaged immediately distal to an ER gap as in Video","2 . Anterograde movement of a 5 µm length of ER simultaneously closes up a gap distal to it ( first white arrow ) and opens up a new gap ( second white arrow ) proximal to it .","A kymograph from this preparation is shown in Figure 5—figure supplement 1C . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 014 During live imaging , we also followed the stability of the ER network in both wild-type and mutant axons .","Features such as higher or lower levels of CG9186::GFP labelling ( which could reflect variation in the numbers of tubules , or the presence of structures like cisternae ) were mostly stable throughout a 200 s acquisition period ( Figure 5A , E; Figure 5—figure supplement 1 ) .","However , we also observed some dynamic features , including anterograde or retrograde movement of more brightly labeled regions ( Figure 5A , E ) perhaps representing movement of ER tubules detached from the ER network , and retrograde movement of labeled puncta ( Figure 5E; Figure 5—figure supplement 1C; Video 3 ) , at around 0 . 3 µm/s .","While most gaps in ER labeling in triple mutants were stable over 200 s ( 7 out of 8 gaps imaged , including Videos 2–3 ) , in one case a gap was closed up by anterograde movement of a 5 µm length of ER proximal to the gap , which simultaneously opened up a new gap proximal to the moving section ( Figure 5—figure supplement 1C; Video 4 ) .","In one case a labeled punctum moved retrogradely across an ER gap without pausing ( Figure 5—figure supplement 1C ) , implying that microtubule-based transport was intact .","Therefore , live imaging and photobleaching of CG9186::GFP both suggest that ER is normally continuous in axons , that loss of reticulon and REEP proteins leads to occasional gaps in ER labeling that represent physical breaks of ER continuity , and that ER network organization in axons shows both static and dynamic features .","We also tested for possible defects in axon transport by staining for abnormal accumulation of the synaptic vesicle protein CSP in axons .","Rtnl1− larvae and Rtnl1− ReepA− ReepB− triple mutant larvae showed large accumulations of CSP in many peripheral nerves .","The large accumulations of CSP in Rtnl1− larvae could be rescued by two copies of a Rtnl1Pacman genomic clone ( Figure 6 ) . 10 . 7554/eLife . 23882 . 015Figure 6 . Loss of Rtnl1 causes mild accumulation of synaptic vesicles in axons .","( A ) Peripheral nerves of Rtnl1− and Rtnl1− ReepA− ReepB− triple mutant larvae show larger accumulations of synaptic vesicle protein CSP ( e . g . yellow arrows ) , and smaller elongated CSP puncta ( e . g . yellow arrowheads ) .","In contrast , control larvae ( WT ) and ReepA− ReepB− double mutant larvae show an even distribution of small round CSP puncta .","CSP accumulations are not significantly bigger in Rtnl1− ReepA− ReepB− triple mutants than in Rtnl1− mutants .","The CSP accumulations in Rtnl1− larvae can be rescued by two copies of a Rtnl1Pacman genomic clone .","All axons shown are crossing abdominal segment A2 .","( B ) Graph shows mean ±SEM; n = 16–92 axons from 8 to 46 larvae , from 3 different experiments .","ns , p>0 . 05; **p<0 . 006; ****p<0 . 0001 , two-tailed Student’s t-test .","Scale bar , 10 µm ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 015 To better understand wild-type axonal ER organization , and the mutant phenotypes seen in confocal microscopy , we performed electron microscopy ( EM ) on 60-nm-thick serial sections of third instar peripheral nerves .","Peripheral nerves contain both motor and sensory axons , arranged in fascicles , and wrapped in three main classes of glial cell ( Stork et al . , 2008; Matzat et al . , 2015 ) .","We used ROTO staining ( Tapia et al . , 2012; Terasaki et al . , 2013 ) to preferentially highlight cellular membranes including ER .","Wild-type larvae showed a network of ER tubules , in every axon that could be observed ( Figure 7A left; serial sections in Video 5 ) .","Tubule outer diameter averaged around 40 nm ( Figure 7B , C ) , and axons contained an average of around 1 . 6 ER tubules in each cross-section ( Figure 7D , E ) .","Reconstruction ( Figure 7F left; Supplementary file 2 ) showed a tubular network with multiple branches , some dead ends , and continuity along nearly every axon sectioned .","ER tubules often showed proximity to mitochondria or plasma membrane ( Figure 7G left ) .","We also found occasional structures resembling small patches of ER sheets with an adjoining cisterna ( Figure 7H ) , continuous with the tubular ER network ( Video 6 ) , at a frequency averaging around one per 20 µm per axon . 10 . 7554/eLife . 23882 . 016Figure 7 . Loss of hairpin proteins leads to fewer but enlarged ER tubules in larval peripheral nerve axons .","( A ) EMs of peripheral nerve axons from wild-type ReepA+ ( WT , left ) or Rtnl1− ReepA− ReepB− triple mutant ( right ) larvae .","Arrowheads indicate ER tubules ( seen as continuous structures in serial sections in Videos 5 and 7 ) .","Triple mutant axons show enlarged ER tubules , sometimes with a clear lumen ( arrows ) , seldom observed in wild-type .","Quantification of ER tubule diameter ( B , C ) and tubules per axon cross-section ( D , E ) in control and mutant larvae .","Data from individual ER tubules or axons are shown in B and D , averaged larval values , mean ±SEM in C and E . ( F ) 3D reconstruction of a 4 . 5 µm axon segment from wild-type ( left ) or mutant ( right ) peripheral nerves , generated from 75 serial 60 nm sections , showing ER ( green ) , mitochondria ( magenta ) and plasma membrane ( gray ) .","Interactive versions of the reconstructions are in Supplementary files 2 and","3 . ( G ) Electron micrographs of peripheral nerve axons from wild-type ( ReepA+ , left ) or Rtnl1− ReepA− ReepB− triple mutant ( right ) larvae , showing proximity of ER to mitochondria ( arrowheads ) or plasma membrane ( arrows ) .","( H ) Representative EM of short ER sheet ( arrowhead ) and cisterna ( asterisk ) from a wild-type larva; further sections in Video 6 .","( I ) section of an axonal swelling from a wild-type larva showing mitochondria ( m ) , vesicles ( v ) , and a large clear cisterna ( c ) ; further sections in Video 8 .","( J ) Serial EM sections show ER discontinuity in two mutant axons: axon 1 lacks ER tubules in sections z6-z14 and axon 3 in sections z1-z11; neighboring axons ( e . g . axon 2 ) show a continuous ER network .","( K ) 3D reconstruction of a 4 . 5 µm axon segment from mutant peripheral nerves , generated from 75 serial 60 nm sections , showing multiple gaps ( indicated by arrows ) .","ER is in green and plasma membrane in gray .","Raw EM data and an interactive version of the reconstruction are in Video 9 and Supplementary file 4 , respectively .","( L ) Frequency of axons with gaps in the 4 . 5 µm lengths analyzed .","ER gap length ( M ) , numbers of ER gaps per µm ( N ) , and proportion of axon length with gaps ( O ) in affected axons .","In M–O , top graphs show data from individual ER gaps ( M ) or individual axons ( N–O ) , with second and third quartiles and 5th and 95th percentiles; bottom graphs show averaged larval values , mean ±SEM .","In all graphs , ns p>0 . 05; *p<0 . 04; **p<0 . 003; ****p<0 . 0001; Mann-Whitney U test for B , D , top graphs in M–O; two-tailed Student’s T test for C , E , L , bottom graphs in M–O .","Scale bars , 500 nm . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01610 . 7554/eLife . 23882 . 017Video 5 . Serial EM sections of a wild-type peripheral nerve , showing continuity of tubular membrane structures through multiple sections . The sixth section in the series is shown in Figure 7A ( left ) .","ER tubules were identified as darkly stained structures present for multiple sections .","See Figure 7A for annotations . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01710 . 7554/eLife . 23882 . 018Video 6 . Serial EM sections of a sheet-like ER structure in a wild-type axon , shown in Figure 7H . Arrows indicate short ER sheets and associated cisternae , present across multiple sections .","Note that these structures usually appear continuous with tubules in adjacent sections . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 018 If Rtnl1− ReepA− ReepB− triple mutant larvae have less ER membrane curvature , we would expect them to have larger ER tubules , fewer tubules per section , loss of tubules , or a combination of these .","EM revealed all these mutant phenotypes to varying degrees ( Figure 7 ) .","ER tubule diameter was increased to around 60 nm ( Figure 7B , C ) , allowing a lumen to be seen in some tubules ( Figure 7A right; Video 7 ) that was rarely seen in wild-type ( Figure 7A left; Video 5 ) , and most triple mutant axons exhibited only a single ER tubule ( Figure 7D , E ) .","Reconstructions ( Figure 7F right; Supplementary file 3 ) showed a less extensive ER network in mutant axons .","Frequent contacts of ER with mitochondria and plasma membrane were found in both wild-type and Rtnl1− ReepA− ReepB− mutant axons ( Figure 7G ) .","Both wild-type and mutant axons showed swellings containing mitochondria and clusters of vesicles resembling synaptic vesicles ( Figure 7I; Video 8 ) .","We did not detect large swellings with synaptic vesicle materials similar to those seen with confocal microscopy ( Figure 6 ) , but this may reflect the much shorter lengths of nerve examined by EM , around 5 µm compared to around 50 µm in confocal . 10 . 7554/eLife . 23882 . 019Video 7 . Serial EM sections of a Rtnl1− ReepA− ReepB− peripheral nerve showing continuity of tubular membrane structures through multiple sections . The sixth section in the series is shown in Figure 7A ( right ) .","ER tubules were identified as darkly stained structures present for multiple sections .","See Figure 7A for annotations . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01910 . 7554/eLife . 23882 . 020Video 8 . Serial EM sections of a wild-type axonal swelling shown in Figure 7I . An axon with a swelling is highlighted in the first frame .","Sections show accumulated mitochondria , vesicles , and a large clear cisterna , as labeled in Figure 7I . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 020 Serial EM sections also revealed variable fragmentation of ER in Rtnl1− ReepA− ReepB− mutant axons ( Figure 7J , K ) , consistent with that seen using confocal microscopy ( Figure 4 ) .","Some discontinuity of the ER network was observed in about 10% of wild-type or Rtnl1− ReepA− ReepB− mutant axons ( Figure 7L ) .","However , gaps in Rtnl1− ReepA− ReepB− mutant axons ( Figure 7J , K; Video 9; Supplementary file 4 ) were longer ( Figure 7M ) , and slightly more numerous ( but not significantly when averaged across larvae; Figure 7N ) than in wild-type axons , resulting in nearly a four-fold increase in the length of affected axons that lacked ER tubules ( Figure 7O ) . 10 . 7554/eLife . 23882 . 021Video 9 . Serial EM sections of a 4 . 5 µm segment of a Rtnl1− ReepA− ReepB− mutant axon with disrupted continuity of ER , used for 3D reconstruction in Figure 7K . An axon with gaps in its ER network is highlighted in the first frame .","Continuous ER tubules were identified as the presence of signals at the same position for three or more sections .","Given the varying brightness and contrast of EM sections , faint staining that coincided with a tubule signal in adjacent sections was also considered as an ER tubule .","Complete loss of ER tubules from three or more sections was defined as a gap . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 021 Rtnl1− ReepA− ReepB− mutant peripheral nerves also showed glial cell phenotypes .","Wild-type peripheral nerves are surrounded by an outer perineurial glial cell , and just beneath this a subperineurial glial cell; axons or axon fascicles are wrapped imperfectly by a wrapping glia cell ( Stork et al . , 2008; Matzat et al . , 2015 ) .","All glial classes , but particularly subperineurial glia , showed a trend towards increased ER sheet profile length compared to control cells ( Figure 8A–I ) , similar to Rtnl1 knockdown ( O'Sullivan et al . , 2012 ) or ReepA− ReepB− double mutant ( Figure 2 ) epidermal cells .","Triple mutant wrapping glia also displayed more extensive wrapping , sometimes completely ensheathing axons , which was rarely observed in control nerve sections ( Figure 8J–M; Figure 8—figure supplement 1 ) . 10 . 7554/eLife . 23882 . 022Figure 8 . Loss of reticulon and REEP proteins leads to ER disorganization in glial cells and hyper-wrapping of peripheral axons .","( A–B )","EMs of peripheral nerve sections from ReepA+ ( WT ) ( A ) or Rtnl1− ReepA− ReepB− triple mutant ( B ) larvae .","Perineurial , subperineurial , and wrapping glial cells are shaded green , blue , and magenta , respectively .","( C–H )","Higher magnification images of perineurial ( C , F ) , subperineurial ( D , G ) , and wrapping ( E , H ) glia from wild-type ( C–E ) or mutant ( F–H ) nerve sections , showing ER tubules ( white arrows; confirmed as tubules by presence in adjacent sections ) and sheets ( white arrowheads ) .","Note the longer ER sheet profiles and fewer ER tubules in the subperineurial ( G ) and wrapping glial cells ( H ) of mutant nerves .","Asterisks show mitochondria; yellow arrowheads show glial plasma membrane , identified by its continuity .","( I ) ER sheet profile length in control and mutant glial cells from 3 wild-type and 4 mutant larvae .","Top graphs represent all individual sheet profiles , with second and third quartiles and 5th and 95th percentiles; bottom graphs show averaged larval values , showing mean ±SEM .","Two-way ANOVA showed a significant effect of genotype ( p<0 . 002 ) but not glial class ( p>0 . 3 ) on ER sheet length , with no interaction between factors ( p>0 . 3 ) .","( J , K )","Sketches of wild-type ( J ) and mutant ( K ) wrapping glial cells , showing excess processes in the triple mutant .","Asterisks indicate completely wrapped one-axon or two-axon fascicles , rarely seen in wild-type nerve sections .","More examples of each phenotype are in Figure 8—figure supplement 1 .","( L–M )","Quantification of wrapping glial membrane profile length per nerve cross-section ( L ) and percentage of axons that are wrapped individually or as two-axon fascicles ( M ) in wild-type and mutant nerve sections ( mean ±SEM ) .","ns , p>0 . 05; *p<0 . 05; ***p<0 . 001; ****p<0 . 0001 .","Mann-Whitney U test ( I ) , top graphs ) ; two-tailed Student’s t-test ( I ) , bottom graphs; L , M , ) .","Scale bars , 1 µm . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 02210 . 7554/eLife . 23882 . 023Figure 8—figure supplement 1 . Sketches of wrapping glia in wild-type ( ReepA+ ) and Rtnl1− ReepA− ReepB− mutant peripheral nerves , similar to those in Figure 8J , K . Sketches of wild-type ( ReepA+ , upper ) and mutant ( lower ) wrapping glial cells were abstracted from cross-sections of peripheral nerves .","Excess formation of processes were observed in triple mutants . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 023"],["The existence of a tubular axonal ER network has been known for decades .","Nevertheless , the cellular mechanisms that organize a compartment that is usually distributed throughout cells , along the great lengths of axons , are until now largely unknown .","The finding that several causative genes for the axon degenerative disease HSP encode ER modeling proteins , suggests a link between ER modeling and axon function or maintenance , and provides candidate proteins that may be instrumental in structure and function of the axon ER network .","These candidates include several hairpin-loop-containing HSP proteins , of the spastin , atlastin , reticulon , REEP , and Arl6IP1 families , that influence ER structure in situations including yeast , mammalian cultured cells , and neuronal cell bodies in vivo ( Shibata et al . , 2006; Voeltz et al . , 2006; Hu et al . , 2008; Shibata et al . , 2009; Park et al . , 2010; Shibata et al . , 2010 ) .","The HSP-related protein families that model ER , and some other proteins that interact with them , share a common feature of one or two intramembrane hairpin loops that can insert into the cytosolic face of the ER membrane , thereby recognizing or inducing curvature .","This property makes the reticulon and REEP ( DP1 ) families together responsible for most peripheral ER tubules in yeast , and contribute to the curved edges of ER sheets ( Voeltz et al . , 2006; Hu et al . , 2008 ) .","The latter property may explain the expansion of ER sheets in Drosophila lacking the reticulon Rtnl1 ( O'Sullivan et al . , 2012 ) and in REEP1 homozygous mutant mice ( Beetz et al . , 2013 ) .","Given this background , we set out to test how far the reticulon and REEP families contribute to axonal ER organization .","REEP1 homozygous mutant mice were not previously tested for effects on axonal ER , although knockdown of Drosophila Rtnl1 led to partial loss of smooth ER marker in distal motor axons ( O'Sullivan et al . , 2012 ) .","Here we build on this work by analyzing mutants of all the widely expressed and highest conserved members of the reticulon and REEP families in Drosophila: Rtnl1 , an ortholog of all four human reticulons ( O'Sullivan et al . , 2012 ) ; ReepA , an ortholog of human REEP1-REEP4; and ReepB , an ortholog of human REEP5-REEP6 .","We monitored phenotypes of these mutants , singly and in combination , by confocal microscopy of axonal ER markers in small numbers of motor neurons , live imaging and photobleaching of ER , and EM using membrane-specific staining .","Confocal microscopy revealed a partial loss of ER marker in distal but not in more proximal motor axons in Rtnl1 and in ReepB mutants .","Although the only REEP genes identified as causative for HSP are REEP1 and REEP2 , loss of their ortholog ReepA had at most only mild effects on axonal ER ( Figures 3 and 4 ) .","The stronger axonal ER phenotypes of ReepB or Rtnl1 loss of function , compared to ReepA− mutants , is consistent with the higher levels of ReepB and Rtnl1 expression , judged by transcriptomics ( www . flyatlas . com ) , and the detection of ReepB and Rtnl1 but not ReepA fusions in peripheral nerves ( Figure 1F , G; O'Sullivan et al . , 2012 ) or individual axons ( Figure 1H , J ) .","The partial loss of distal axonal ER marker in both mutants and motor-neuron knockdowns , of either Rtnl1 or ReepB ( Figure 3; O'Sullivan et al . , 2012 ) , and qualitatively similar ER fragmentation phenotypes in Rtnl1 ReepA ReepB triple loss-of-function genotypes , obtained using either Rtnl1 mutants or motor-neuron knockdown ( Figure 4 ) , suggest that the axonal ER phenotypes are cell autonomous in motor neurons .","Given the joint and partly redundant requirement of the reticulon and REEP families for ER tubule formation in yeast ( Voeltz et al . , 2006 ) , we tested whether this was also true for axonal ER .","Flies lacking Rtnl1 , ReepA and ReepB – probably equivalent to mammals lacking all four reticulons and all six REEPs , and homozygous viable – indeed showed more extreme ER phenotypes than axons lacking either Rtnl1 or ReepA and ReepB alone ( Figures 3 and 4 ) .","However , only a fraction of mutant axons showed severe fragmentation , and even affected axons still had continuous labeling of axons with ER marker through much of their length .","ER fragmentation in the middle parts of long axons might be a consequence of axon expansion during larval growth , in which the somatic and presynaptic ends of the axon are gradually pulled apart , with insufficient ER-modeling proteins to maintain the expanding tubular network throughout the axoplasm .","Therefore reticulon and REEP proteins are present in axons and have roles in ER organization there – but since triple mutant axons still mostly possess ER , there must be additional proteins required too .","These might be found among the increasing number of other HSP genes that encode ER proteins with possible hairpins , such as Arl6IP1/SPG61 , which affects ER organization ( Novarino et al . , 2014; Yamamoto et al . , 2014; Fowler and O'Sullivan , 2016 ) , or C19orf12/SPG43 ( Landouré et al . , 2013 ) .","The variable nature of the triple mutant fragmentation phenotype might reflect stochastic variation in the amounts of such proteins , the amount of ER present , external factors like physical stresses during larval movement , or dynamic fluctuations in the local levels and connections of ER tubules .","The tubular ER network is highly dynamic in non-neuronal cells ( Nixon-Abell et al . , 2016; Valm et al . , 2017 ) and in axons our live imaging suggests dynamic features superimposed on a structure that is largely stable over the 1–2 min of imaging ( Figure 5 ) .","EM examination of wild-type axonal ultrastructure revealed that ER tubules were effectively ubiquitous in Drosophila peripheral nerve sections , as seen previously in mammalian neurons ( Tsukita and Ishikawa , 1976; Villegas et al . , 2014 ) , albeit with fewer tubules , presumably reflecting the smaller diameters of the axons examined here .","Reconstruction over several µm showed a continuous network of ER tubules in most axons examined , in agreement with the ER continuity found in neurons by lipid dye labeling ( Terasaki et al . , 1994 ) , and recently in reconstructions of serial sections from focused ion beam SEM ( Wu et al . , 2017 ) .","However , in a few axons , we found short lengths of axon with no detectable ER ( Figure 7J–O ) .","There could be several reasons for this: Terasaki et al . ( 1994 ) only assessed continuity in dendrites , cell body and proximal axon; some of the gaps we observe in EM could be short transient gaps in a dynamic network; some of the apparent gaps could be ‘thin ER’ observed using higher-resolution focused-ion-beam SEM ( Wu et al . , 2017 ) , but that might be missed using our approach; larval axons with low diameters might be intrinsically more susceptible to occasional gaps in the ER network , than wider axons with more tubules; and we might occasionally miss an ER tubule due to weaker staining , or close proximity to other structures like plasma membrane .","We also observed occasional small ER sheet-like structures in wild-type axons ( Figure 7H; Video 6 ) .","Although we do not see ribosomes on these , this could be due to lack of staining by the ROTO protocol .","As discussed above , rough ER and translation are relatively sparse in axons , but low levels of rough ER are possible , and consistent with the occasional sheet structures observed here .","EM also showed phenotypes consistent with loss of ER membrane curvature in Rtnl1− ReepA− ReepB− triple mutant axons ( Figure 7 ) .","Mutant axons had ER tubules of larger diameter , fewer tubules per axon cross-section , and consistent with our confocal data ( Figure 4 ) , longer gaps in the ER network than wild-type , although most parts of most mutant axons examined still had a continuous ER network ( Figure 7 ) .","These mutant phenotypes could potentially have physiological consequences .","Larger tubules could potentially store and release more calcium than thinner ones , while the reduced network could make the role of the ER in calcium buffering or release more localized .","The less extensive ER network in mutants might also reduce the amount of contact between ER and other organelles , with consequences for calcium and lipid homeostasis that require these contacts , or for regulation of mitochondrial fission ( Friedman et al . , 2011 ) – although the continuing proximity of ER to mitochondria and plasma membrane in mutants means that any effects are presumably quantitative rather than qualitative .","The reduced curvature of ER membrane in mutants might also influence their protein composition , since many membrane proteins have mechanisms for recognizing differential membrane curvature ( Antonny , 2011 ) .","The occasional lack of continuity could prevent propagation of ER-dependent Ca2+ signals like those seen in injured mammalian sensory neurons ( Cho et al . , 2013 ) ; it could also cause local impairments in Ca2+ or lipid homeostasis that could lead to local transport inhibition , as is the case for mitochondrial transport ( Wang and Schwarz , 2009 ) , although lack of ER continuity does not appear to directly prevent axon transport ( Figure 5—figure supplement 1C; Video 3 ) .","Sporadic lack of ER continuity might explain the preferential sensitivity of distal longer axons to HSPs , since these would be more likely to suffer from a gap in ER continuity to the cell body , compared to proximal or shorter axons .","In this model , disease-causing alleles in single hairpin-encoding genes could promote degeneration in distal motor axons by increasing the probability of such gaps , dependent on factors such as age , axon length or diameter , and ER tubule density and dynamics .","The apparent ubiquity of ER in axons , the extent of its continuity over long distances , and the preferential susceptibility of distal longer axons to mutations that affect ER-modeling proteins , all point to important physiological roles of this compartment and of its continuity .","In this work we have begun to reveal the mechanisms that determine its organization .","We have shown roles for two protein families that contain HSP disease gene products , in influencing the shape of individual tubules and the axonal ER network , with potential physiological consequences that would also be affected by mutations in these genes .","Understanding the consequences of axonal ER structural defects for ER dynamics and axonal physiology , both in the genotypes we have described here , and in other genotypes that might also affect axonal ER organization , will provide models for the potential physiological defects in HSP and other axon degeneration diseases ."],["ReepA541 ( referred to as ReepA− ) , and ReepB48 ( referred as ReepB− ) mutants were generated by imprecise excision of P elements CB-0501–3 ( RRID:DGGR_123207 ) and EY05130 ( RRID:BDSC_16636 ) shown in Figure 1 .","A precise excision generated in these experiments , ReepA+C591 ( referred to as ReepA+ ) was used as a genetic background control where feasible .","Rtnl11 , referred as Rtnl1− , was a gift from G . Tear ( Wakefield and Tear , 2006; FlyBase ID FBal0246222 ) .","Rtnl1− , ReepA− and ReepB− recombinants were generated by meiotic recombination on the second chromosome , and recombinants were screened using PCR primers ( Supplementary file","5 ) to diagnose wild-type or mutant alleles of all three genes .","Mutant and wild-type stocks were frequently genotyped to ensure that experimental flies were not contaminated .","For knockdown experiments , either UAS-Rtnl1-RNAi line 7866 ( construct GD900 , which has no predicted off-targets; FlyBase ID FBti0098310 ) , or the w1118 control stock , 60000 ( both obtained from the Vienna Drosophila RNAi Center , www . vdrc . at ) , or the UAS-ReepB-RNAi line 8331 R-3 ( National Institute of Genetics Fly Stock Center , Japan; FlyBase ID FBal0275953 ) was crossed with UAS-Dcr2; CyO/If; m12-GAL4 , UAS-Acsl::myc .","UAS-Dcr2 ( RRID:BDSC_24648; Dietzl et al . , 2007 ) was also present for knockdown in Rtnl1 ( RNAi ) ReepA− ReepB− larvae .","Other fly stocks used were P{UAS-Acsl . 715 . MycC}3 ( RRID:BDSC_32330; Zhang et al . , 2009 ) , PBac{681 . P . FSVS-1}Rtnl1CPTI001291 ( RRID:DGGR_115146; Wakefield and Tear , 2006 ) , m12-GAL4 ( Xiong et al . , 2010; RRID:BDSC_2702 ) , UAS-Rtnl1::GFP ( Rao et al . , 2016 ) , {UAS-Xbp1 . EGFP . LG}4 ( RRID:BDSC_39719; Ryoo et al . , 2007 ) and UAS-tdTomato::Sec61β ( RRID:BDSC_64746; Summerville et al . , 2016 ) .","A second-chromosome insertion of P{UAS-CG9186::GFP} ( Thiel et al . , 2013 ) was mobilized onto the third chromosome using the P transposase source P{∆2–3}99B ( Robertson et al . , 1988; RRID:BDSC_3612 ) , and expressed using m12-GAL4 in either a ReepA+ or an Rtnl1− ReepA− ReepB− mutant second chromosome background .","For rescue of Rtnl11 we generated transgenic flies carrying P[acman] clone CH322-124P15 inserted at attP2 on chromosome 3 ( Bloomington stock 25710 ) , referred to as Rtnl1Pacman ( Venken et al . , 2009 ) .","For C-terminal EGFP-LAP-tagging of ReepA and ReepB we used recombineering with the P[acman] system with minor modifications ( Venken et al . , 2009 ) , utilizing the CH322-97D15 ( ReepA ) and CH322-16N11 ( ReepB ) BAC clones ( BPRC; http://bacpac . chori . org ) .","An EGFP-LAP-tagging cassette was amplified from R6Kamp-LAP ( GFP ) ( Poser et al . , 2008 ) using primers ( Supplementary file","5 ) with 50 bp homology to the corresponding Reep clone and around 20 bp of homology to the tagging cassette , and used to transform the recombineering E . coli SW102 strain .","Correct clones were verified at every step by PCR and/or restriction digestion .","DNA for Drosophila transformation was extracted using a PureLink HiPure Maxiprep kit ( Invitrogen ) , and injected into y w M ( eGFP , vasa-integrase , dmRFP ) ZH-2A; M ( attP ) ZH-51D or y w M ( eGFP , vasa-integrase , dmRFP ) ZH-2A; M ( attP ) ZH-86Fb ( RRID:BDSC_24483 or RRID:BDSC_24749 , respectively ) at the Department of Genetics embryo injection facility , University of Cambridge , UK .","Transformant lines were screened for the presence of the insert by PCR and GFP fluorescence .","The primers used are described in Supplementary file 5 .","BLAST sequence searches were used to define genome coordinates of P-element excisions , and Pfam domain coordinates in coding regions , and compare protein divergence rates .","They were performed at the National Center for Biotechnology Information ( www . ncbi . nlm . nih . gov ) .","REEP dendrograms were drawn from a ClustalW alignment ( Larkin et al . , 2007 ) using the neighbor-joining algorithm in MEGA 5 . 05 ( Tamura et al . , 2011 ) .","Third instar larvae were dissected in chilled Ca2+-free HL3 solution ( Stewart et al . , 1994 ) , and fixed for 30 min in PBS with 4% formaldehyde .","Dissected Drosophila preparations were permeabilized in PBS containing 0 . 3% Triton X-100 ( PBT ) at room temperature , and blocked in PBT with 4% bovine serum albumin for 30 min at room temperature .","Primary antibodies were: Csp ( 6D6 , RRID:AB_10013286 , 1:50; Zinsmaier et al . , 1994 ) , Dlg ( 4F3 , RRID:AB_2314321 , 1:100; Parnas et al . , 2001 ) , ( both from the Developmental Studies Hybridoma Bank , Iowa , USA ) , GFP ( Ab6556 , RRID:AB_305564 , 1:600 Abcam , UK ) , HRP ( P97899 , RRID:AB_2314650 , 1:300 , Sigma ) , KDEL ( Ab50601 , RRID:AB_880636 , 1:25 , Abcam , UK ) , myc ( 2272 , RRID:AB_331667 , 1:25 , Cell Signaling , USA ) .","Fixed preparations were mounted in Vectashield ( Vector Laboratories , USA , RRID:AB_2336789 ) , and images were collected using EZ-C1 acquisition software ( Nikon ) on a Nikon Eclipse C1si confocal microscope ( Nikon Instruments , UK ) .","Images were captured using 10x/0 . 30NA , or a 60x/1 . 4NA oil objective .","Confocal images were analyzed blind to genotype using ImageJ ( Schneider et al . , 2012 ) .","Images of entire epidermal cells were obtained as z-projections of three consecutive sections .","Using the line tool of ImageJ a 12 µm line was drawn from the nuclear envelope towards the periphery of each cell analyzed .","Pixel intensity along the line was recorded in an Excel file .","Local variance of intensity was calculated by dividing the rolling variance of the intensity ( in 10-pixel windows ) , by the rolling mean intensity , all along the line .","Proximal ( anterior ) axons were imaged from segment A2 , middle images were from the end of segment A4 and A5 , distal ( posterior ) axons were imaged from segment A6 of third instar larvae .","Mean gray intensity for single-axon images was measured by drawing a 45 µm line , either along both M12-GAL4-expressing axons ( where they could not be separated ) , or along the most strongly labeled axon ( where they appeared as separate axons ) , and quantifying gray intensity ( 0–255 ) by ImageJ; occasional images with saturated pixels were excluded from analysis after blinding .","Coefficient of variation was calculated by dividing the standard deviation of staining intensity by the mean; occasional images with faint staining throughout the axon were excluded from analysis after blinding .","Gaps were defined as regions where staining intensity was less than 20 ( out of 255 ) , after background subtraction .","FRAP experiments were performed on a Nikon Eclipse C1si confocal microscope ( Nikon Instruments , UK ) , using a 20 mW Argon laser and a 40× 0 . 8 N . A . water dipping objective .","Third instar larvae were dissected and incubated in chilled Ca2+-free HL3 solution ( Stewart et al . , 1994 ) .","Time-lapse images were acquired on a single focal plane every 0 . 5 or 1 s for 40 loops .","For FRAP , a defined region of interest ( 12 × 4 µm ) was photobleached at full laser power ( 488 nm ) for two iterations at a scan speed of 0 . 5 frame/s .","Postbleaching images were acquired on a single focal plane at 15% laser power once every 5 s for 200 s .","Experiments were completed within 20 min after dissection .","Kymographs were generated for a hand-drawn line selection along the axon using the Multiple Kymograph plugin in Fiji ( Schindelin et al . , 2012 ) .","Average fluorescence intensity of each axon in each frame was measured by creating a line selection along the axon .","After subtracting background ( average intensity in a nearby non-GFP-expressing region within the segmental nerve ) , the intensity of the bleached region was normalized to the average intensity cross the axon length in the same frame .","Then the data were further normalized by taking the prebleaching intensity as 100% and bleach intensity as 0 .","Normalized postbleaching data were plotted and fitted to a single exponential function to calculate rate constant ( k ) and half time ( t1/2 ) : I ( t ) =A ( 1-e−kt ) , in which I ( t ) represents the fluorescence intensity at time point t , A the highest postbleaching intensity .","For epidermal cell EM , larvae were prepared and fixed as described by O'Sullivan et al . ( 2012 ) .","Transverse sections were cut on a Leica Ultracut UCT ultra-microtome at 70 nm , using a diamond knife , and contrasted with uranyl acetate and lead citrate ( for epidermal cells ) .","Sections were viewed using a Tecnai G2 electron microscope operated at 120 kV , and an AMT XR60B camera running the Deben software in the Cambridge Advanced Imaging Centre , School of Biology , University of Cambridge .","For EM of peripheral nerves , we used a ROTO protocol ( Tapia et al . , 2012; Terasaki et al . , 2013 ) to highlight membranes .","Third instar larvae were dissected in HL3 solution and fixed in 0 . 05 M sodium cacodylate ( pH 7 . 4 ) containing 4% formaldehyde , 2% vacuum distilled glutaraldehyde , and 0 . 2% CaCl2 ) , at 4°C for 6 hr .","Larvae were dissected as for confocal analysis , but leaving overlying organs such as gut and fat body attached , to reduce loss of peripheral nerves during processing .","Preparations were then washed 3 times for 10 min each at 4°C using cold cacodylate buffer with 2 mM CaCl2 .","A solution of 3% potassium ferricyanide in 0 . 3 M cacodylate buffer with 4 mM CaCl2 was mixed with an equal volume of 4% aqueous osmium tetroxide; larval preparations were incubated in this solution at 4°C for 1–12 hr , then rinsed with deionized water at room temperature 5 times for 3 min each .","Thiocarbohydrazide solution was prepared by adding 0 . 1 g thiocarbohydrazide to 10 ml deionized water , kept in a 60°C oven in a secondary embedding pot for 1 hr , swirled every 10 min to facilitate dissolution , and filtered through two 9 cm filter papers just before use .","Larval preparations were incubated in thiocarbohydrazide solution for 20–30 min at room temperature and covered with foil to protect from light .","Then they were rinsed with deionized water at room temperature 5 times for 3 min each , incubated in 2% osmium tetroxide for 30–60 min at room temperature , and rinsed with deionized water at room temperature 5 times for 3 min each .","Preparations were incubated in 1% uranyl acetate ( maleate-buffered to pH 5 . 5 ) at 4°C overnight and rinsed with deionized water at room temperature 5 times for 3 min each .","Then they were incubated in lead aspartate solution ( 0 . 66 g lead nitrate dissolved in 100 ml 0 . 03 M aspartic acid , pH adjusted to 5 . 5 with 1 M KOH ) at 60°C for 30 min and rinsed with deionized water at room temperature 5 times for 3 min .","Then they were dehydrated twice with 50% , 70% , 90% and 100% ethanol , twice with dried ethanol , twice with dried acetone and twice with dry acetonitrile .","Preparations were incubated in 50/50 acetonitrile/Quetol 651 overnight at room temperature , three times for 24 hr each in Quetol epoxy resin 651 ( Agar Scientific , Stansted , UK ) and three times for 24 hr each in Quetolepoxy resin 651 with BDMA ( dimethylbenzylamine ) .","They were then incubated at 60°C for at least 48 hr .","Serial 60-nm-thick transverse sections were cut in the larval abdominal region , visualized using scanning EM , and images were aligned for analysis of serial sections and reconstruction , as described by Terasaki et al . ( 2013 ) .","To quantify axonal ER tubule diameter , non-axonal staining was removed manually , and ER tubule profiles were identified based on the local threshold in a single cross-section , and the presence of signals at the same position in adjacent sections .","The minimum Feret diameter of each tubule was measured using ImageJ Fiji ( https://fiji . sc ) via the Analyze Particles command .","ER numbers per axon were counted manually for all axons detected in the nerve .","3D reconstruction was carried out using the Fiji TrakEM2 plug-in .","To quantify gaps in the tubule network , each axon was analyzed throughout the entire stack of sections .","To allow for occasional lightly stained or blurred sections , only complete loss of ER tubules from three or more consecutive sections in an axon was defined as a gap .","Continuous ER tubules were identified as the presence of signals at the same position for three or more consecutive sections .","Given the varying brightness and contrast of EM sections , faint staining that coincided with a tubule signal in adjacent sections was also considered as an ER tubule .","Color shading and sketches drawn in Figure 8 were processed in Adobe Photoshop CS6 .","For quantification of glial ER sheet length , individual ER sheet profiles were measured using Fiji via the line tool and Measure command .","Wrapped axons were defined as one-axon or two-axon fascicles which were completely wrapped by glial cells and isolated from other neighboring axons .","Statistical analyses were performed in GraphPad Prism 6 or IBM SPSS 22 .","Data were analyzed by two-tailed Student’s t-tests or ANOVA followed by post-hoc tests ( for comparison of datapoints that were means of raw measurements and hence expected to be normally distributed ) , or by Mann-Whitney U tests ( for data that were not normally distributed ) .","Multiple comparisons after ANOVA were performed by a Tukey HSD test when equal variances were found , or otherwise by Dunnett’s T3 test .","Bar graphs and scatter plots show mean ±SEM; box plots show median with interquartile range , and the 5% and 95% percentiles as whiskers .","Sample sizes are reported in figures .","No outliers were excluded from analysis after quantification; data were only excluded from analysis if they were unanalyzable , i . e . when confocal images were too faint or saturated , and EM images lacked clearly identifiable plasma or ER membranes .","Most data were analyzed and presented as individual axon or larval datapoints , unpaired between genotypes .","For two comparisons where ANOVA showed significant ( p<0 . 05 ) experiment-to-experiment differences within a genotype ( Figures 2B and 3A ) , data were averaged within each independent experiment to produce experiment-wise datapoints , which were compared using two-tailed paired t-tests .","P levels are indicated as ns p>0 . 05 , *p<0 . 05 , **p<0 . 01 , ***p<0 . 001 , or ****p<0 . 0001 , except where indicated otherwise , when criteria could be defined more stringently ."]],"string":"[\n [\n \"Axons allow long-range bidirectional communication in neurons .\",\n \"They carry action potentials along their plasma membrane from dendrites and cell body to presynaptic terminals; and they transport cell components and signaling complexes using motor proteins , anterogradely and retrogradely .\",\n \"Another potential route for communication along axons is endoplasmic reticulum ( ER ) ; axons possess a network of ER tubules , which appears physically continuous both locally ( Tsukita and Ishikawa , 1976; Villegas et al . , 2014 ) and with ER throughout the neuron ( Lindsey and Ellisman , 1985; Terasaki et al . , 1994 ) .\",\n \"The physical continuity of ER and its ensuing potential for long-distance communication has been likened to a ‘neuron within a neuron’ ( Berridge , 1998 ) .\",\n \"Axonal ER appears mostly tubular and smooth , with some cisternae ( Tsukita and Ishikawa , 1976; Lindsey and Ellisman , 1985; Villegas et al . , 2014 ) .\",\n \"Some rough ER is likely to be present too: numerous mRNAs are found in both growing and mature axons ( Zivraj et al . , 2010; Shigeoka et al . , 2016 ) , and local axonal translation can occur in response to injury ( Ben-Yaakov et al . , 2012; Perry et al . , 2012 ) .\",\n \"However , rough ER sheets ( Tsukita and Ishikawa , 1976; Villegas et al . , 2014 ) , and markers of protein export and folding that are characteristic of rough ER ( Röper , 2007; O'Sullivan et al . , 2012 ) , are relatively sparse in mature axons .\",\n \"Axonal ER therefore likely has major roles other than protein export; these could include lipid biosynthesis ( Tidhar and Futerman , 2013; Vance , 2015 ) , calcium homeostasis and signaling ( Ross , 2012 ) , and coordination of organelle physiology ( Phillips and Voeltz , 2016 ) .\",\n \"The continuity of the ER network suggests that some of these roles might have long-range as well as local functions; indeed , an ER-dependent propagating calcium wave is seen after axotomy of Caenorhabditis elegans or mammalian dorsal root ganglion neurons ( Ghosh-Roy et al . , 2010; Cho et al . , 2013 ) .\",\n \"A strong hint of the importance of ER in axons is found in Hereditary Spastic Paraplegia ( HSP ) , a group of axon degeneration disorders characterized by progressive spasticity and weakness of the lower limbs ( Blackstone et al . , 2011; Blackstone , 2012 ) .\",\n \"Mutations affecting spastin , atlastin-1 , reticulon-2 , REEP1 and REEP2 account for most cases of autosomal dominant ‘pure’ HSP ( Hazan et al . , 1999; Zhao et al . , 2001; Züchner et al . , 2006; Montenegro et al . , 2012; Esteves et al . , 2014 ) .\",\n \"These proteins share a common feature of one or two hydrophobic hairpin-loops inserted in the ER membrane , promoting ER membrane curvature in a process termed hydrophobic wedging ( Voeltz et al . , 2006 ) .\",\n \"Proteins of the REEP and reticulon families localize preferentially to tubular or smooth ER , and their loss results in disruption of ER tubular organization ( Shibata et al . , 2006; Voeltz et al . , 2006; Park et al . , 2010; Shibata et al . , 2010 ) ; they may also contribute to modeling of rough ER sheets by stabilizing their curved edges ( Shibata et al . , 2009 ) .\",\n \"What is the link between ER modeling and axon structure and function ?\",\n \"HSP-causing mutations often appear to cause loss of protein expression or function ( Beetz et al . , 2013; Novarino et al . , 2014 ) , and the ability of hairpin-loop proteins to form homomeric and heteromeric complexes ( Shibata et al . , 2008 ) allows some point mutations to have dominant negative effects ( Züchner et al . , 2006; Beetz et al . , 2012 ) .\",\n \"Therefore loss of normal ER modeling appears to compromise axon maintenance and function .\",\n \"Given the roles of hairpin-loop proteins in ER modeling , we aimed to test the model that hairpin-loop-containing HSP proteins organize the axonal ER network .\",\n \"Since reticulon and REEP family proteins are redundantly required for most peripheral ER tubules in yeast ( Voeltz et al . , 2006 ) , we focus on the requirement for these two families in axons .\",\n \"We previously showed that knockdown of the Drosophila reticulon Rtnl1 causes expansion of epidermal ER sheets , and partial loss of smooth ER marker from distal but not proximal motor axons ( O'Sullivan et al . , 2012 ) .\",\n \"Here we show that REEP proteins have similar roles .\",\n \"We also show that simultaneous loss of reticulon and REEP family members leads to a range of axonal ER phenotypes , including a reduced network with fewer and larger tubules , and occasional gaps in the network .\",\n \"Our work implicates hairpin-loop-containing HSP proteins as important players in the axonal ER network , and suggests further models for how the network is organized .\"\n ],\n [\n \"The reticulon and REEP families of double-hairpin-containing proteins are collectively responsible for formation or maintenance of most peripheral ER tubules in yeast ( Voeltz et al . , 2006 ) .\",\n \"We previously showed that the Drosophila reticulon ortholog Rtnl1 was strongly localized in axons , and that its knockdown caused partial loss of a smooth ER marker in posterior larval segmental axons ( O'Sullivan et al . , 2012 ) .\",\n \"To test the roles of Drosophila REEP proteins in axonal ER localization , we first dissected the ortholog relationships between the six Drosophila and six human REEP proteins .\",\n \"Multiple sequence alignment of mammalian and Drosophila REEP protein sequences suggested that CG42678 was the single Drosophila ortholog of mammalian REEP1-REEP4 ( Figure 1A ) .\",\n \"CG42678 has previously been designated Reep1 ( http://flybase . org/reports/FBgn0261564 . html ) , but we propose the name ReepA to reflect its orthology to the four mammalian genes REEP1-REEP4 .\",\n \"Mammalian REEP5 and REEP6 appeared to share two Drosophila orthologs , CG8331 and CG4960 ( Figure 1 ) .\",\n \"We designated CG8331 as ReepB because of its widespread expression ( www . flyatlas . org; Chintapalli et al . , 2007 ) and slower rate of evolutionary sequence divergence ( Supplementary file 1 ) .\",\n \"We excluded CG4690 and three additional Drosophila REEP genes from further study , since their expression was restricted to testes and larval fat body ( www . flyatlas . org; Chintapalli et al . , 2007 ) , and their faster rate of evolutionary sequence divergence ( Supplementary file 1 , and reflected in longer branch lengths in Figure 1A ) , suggesting poorly conserved function . 10 . 7554/eLife . 23882 . 003Figure 1 . Drosophila Rtnl1 and REEP genes and products .\",\n \"( A ) A dendrogram based on ClustalW sequence alignment of Drosophila and human REEP proteins shows two branches corresponding to human REEP1-4 ( CG42678 ) and human REEP5-6 .\",\n \"Broken lines represent Drosophila REEP proteins that are evolving rapidly ( Supplementary file 1 ) , reflected by longer branch lengths .\",\n \"Sequences used are NP_075063 . 1 , NP_057690 . 2 , NP_001001330 . 1 , NP_079508 . 2 , NP_005660 . 4 , NP_612402 . 1 , NP_726266 . 1 , NP_610936 . 2 , NP_651429 . 1 , NP_611831 . 2 , NP_572730 . 1 , NP_726366 . 1 .\",\n \"( B ) Rtnl1 genomic and transcript map , showing the region deleted in Rtnl1− by excision of P-element NP7026 ( Wakefield and Tear , 2006; Figure 1—figure supplement 1 ) , the RHD domain ( Pfam 02453 ) and its coordinates in protein isoform G , the Rtnl1::YFP exon trap insertion CPTI001291 ( green triangle ) , and the fragment targeted by GD RNAi 7866 .\",\n \"Map and coordinates are from the Drosophila Genome Browser ( www . flybase . org , version R6 . 04 ) , here and in subsequent panels; light regions in transcripts represent coding regions , dark shaded regions represent untranslated regions .\",\n \"( C ) ReepA genomic and transcript map showing the region deleted in ReepA− by excision of P-element CB-0501–3 , the position of the DP1 domain ( Pfam 03134 ) and its coordinates in protein isoforms H and J . GFP insertion sites for ReepA1::GFP , ReepA2::GFP and ReepA3::GFP fusions are shown with green triangles .\",\n \"( D ) ReepB genomic and transcript map showing the region deleted in ReepB− by excision of P-element EY05130 , the position of the DP1 domain ( Pfam 03134 ) and its coordinates in protein isoforms A and D ) .\",\n \"( E–I )\",\n \"Confocal sections showing localization of ReepA::GFP isoforms and ReepB::GFP .\",\n \"( E ) Overlap of ReepA1::GFP and ReepB::GFP with anti-KDEL labeling in larval epidermal cells .\",\n \"To facilitate display of weaker ReepA1::GFP , the GFP channel in wild-type control ( WT ) and ReepA::GFP images has been brightened four times as much as for ReepB::GFP .\",\n \"( F ) Expression of ReepA3::GFP and ReepB::GFP in third instar ventral nerve cord .\",\n \"The GFP channels for WT and ReepA::GFP have been brightened by twice as much as for ReepB::GFP .\",\n \"VNCs are outlined with yellow dashed lines .\",\n \"Arrowheads show ReepB::GFP extending into peripheral nerves .\",\n \"( G ) A single confocal section of a peripheral nerve , showing ReepB::GFP localized continuously along its length .\",\n \"( H ) Double labeling of an NMJ for ReepB::GFP and the mainly postsynaptic marker Dlg , showing ReepB::GFP in an axon emerging from nerve bundles ( arrowhead ) to extend to the NMJ , as well as in axons traversing the muscle surface ( arrow ) ; note the dispersed ReepB::GFP staining also in the underlying muscle .\",\n \"( I ) Double labeling of ReepA::GFP and ReepB::GFP lines for GFP and Dlg ( mainly postsynaptic ) shows presynaptic expression of ReepB::GFP .\",\n \"( J ) GFP expression in a wildtype negative control ( WT ) , or from Rtnl1::GFP expressed in two closely apposed motor neurons by m12-GAL4 .\",\n \"Scale bars 10 µm , except F , 20 µm . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 00310 . 7554/eLife . 23882 . 004Figure 1—figure supplement 1 . Molecular lesions in Rtnl11 , ReepA541 and ReepB48 . BLASTN searches of the Drosophila genome , using the sequences of ( A ) Rtnl11 , ( B ) ReepA541 and ( C ) ReepB48 mutations as queries , show the coordinates of each molecular lesion in the genome , as gaps in alignments between mutant and genomic sequences .\",\n \"The ReepA541 and ReepB48 sequences highlighted in yellow show P element footprints left at the excision sites . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 004 A Rtnl1::YFP exon trap , Rtnl1CPTI001291 ( Figure 1B ) was previously shown to localize to ER , including in axons ( O'Sullivan et al . , 2012 ) .\",\n \"To study localization of ReepA and ReepB , we recombineered C-terminal GFP-tagged versions of these ( Figure 1C , D ) using P[acman] genomic clones ( Venken et al . , 2009 ) .\",\n \"For ReepA , we generated EGFP fusions at three different C-termini ( Figure 1C ) , that we called ReepA1::GFP ( for protein isoforms D , E , G ) , ReepA2::GFP ( for protein isoforms H , I , J , K ) and ReepA3::GFP ( for protein isoforms L , R , S ) .\",\n \"In epidermal cells , we detected weak expression of ReepA1::GFP , but not of the other ReepA::GFP fusions; ReepB::GFP showed stronger expression still , and both fusions overlapped with an ER marker ( KDEL; Figure 1E ) , similar to REEP proteins in other organisms ( Voeltz et al . , 2006; Shibata et al . , 2008; Park et al . , 2010 ) .\",\n \"ReepB::GFP was also more strongly expressed in third instar larval CNS than ReepA3::GFP , which was the only ReepA::GFP fusion that we detected there ( Figure 1F ) .\",\n \"ReepB::GFP , but no ReepA::GFP fusion , was also detected in segmental nerves ( Figure 1F , G ) , and in individual axons emerging from nerve bundles leading to the NMJ ( Figure 1H ) .\",\n \"ReepB::GFP , but none of the ReepA::GFP fusions , localized as a mostly continual structure along the length of presynaptic terminals of neuromuscular junctions ( NMJs ) ( Figure 1H , I ) .\",\n \"We did not have a UAS-ReepB construct available , but UAS-Rtnl1::GFP ( Rao et al . , 2016 ) expressed in two adjacent motor neurons using m12-GAL4 ( Xiong et al . , 2010 ) also showed strong axonal localization .\",\n \"We therefore conclude that ReepB::GFP localizes in axonal and presynaptic ER , similar to Rtnl1::YFP ( O'Sullivan et al . , 2012 ) , but that none of the ReepA::GFP fusions is detectable in axonal or presynaptic ER .\",\n \"An Rtnl1 loss-of-function mutant , Rtnl11 ( Wakefield and Tear , 2006 ) , hereafter referred to as Rtnl1− , lacks the hydrophobic hairpin loop domain ( RHD region ) that induces ER membrane curvature ( Figure 1B; Figure 1—figure supplement 1 ) .\",\n \"We also used P-transposase-mediated imprecise excision to generate ReepA− and ReepB− mutants , both of which lack most of the curvature-mediating DP1 hairpin domains ( Figure 1B , C; Figure 1—figure supplement 1 ) .\",\n \"First , we asked whether Drosophila reticulon and REEP proteins contribute to ER network organization in third instar larval epidermal cells; Rtnl1 knockdown leads to a more diffuse ER network organization and expansion of ER sheets in these cells , compared to wild-type controls ( O'Sullivan et al . , 2012 ) .\",\n \"Rtnl1− mutant larvae also showed loss of ER network organization ( Figure 2A ) .\",\n \"Intensity of KDEL labeling along a line from the nucleus to cell periphery displayed fluctuating intensity , reflecting the reticular distribution of KDEL in wild-type larvae , but less fluctuation in Rtnl1− larvae; overall levels of KDEL remained unchanged ( Figure 2A ) .\",\n \"Similarly , loss of both ReepA and ReepB , but not loss of either gene alone , made KDEL levels in larval epidermal cells fluctuate less than in controls .\",\n \"Mean intensity of KDEL staining was decreased in all ReepA and ReepB mutant genotypes , and fluctuation in intensity was increased in ReepB− larvae ( Figure 2B ) .\",\n \"Since confocal analysis suggests altered organization of ER in epidermal cells , but does not have sufficient resolution to reveal the details of how it is altered , we performed electron microscopy of epidermal cells .\",\n \"This revealed that ReepA− ReepB− double mutant epidermal cells showed longer ribosome studded sheet ER profiles , compared to controls , and to ReepA− and ReepB− single mutants ( Figure 2C ) .\",\n \"This mutant phenotype is similar to , although less severe than loss of Rtnl1 ( O'Sullivan et al . , 2012 ) .\",\n \"ReepA− ReepB− mutants also showed increased ER stress in epidermal cells ( Figure 2D ) but not in CNS ( Figure 2E ) .\",\n \"Therefore , Rtnl1 and REEP proteins shape the ER network in Drosophila , and their loss disrupts ER organization , causing longer ER sheet profiles . 10 . 7554/eLife . 23882 . 005Figure 2 . Rtnl1− mutants and ReepA− ReepB− double mutants show disrupted ER organization in epidermal cells .\",\n \"( A ) KDEL distribution in third instar larval epidermal cells appears diffuse in Rtnl1− larvae compared to a more reticular staining in wild-type ( WT ) larvae .\",\n \"Micrographs show 1 . 5 µm z-projections of confocal sections .\",\n \"KDEL intensity along yellow lines from nuclear envelope to the cell periphery shows less fluctuation in Rtnl1− ( blue dotted line on line graph ) than in WT larvae ( black line on line graph ) .\",\n \"Fluctuation of intensity is quantified using the local normalized variance of intensity ( variance/mean for rolling 10-pixel windows ) over a 12 µm line for each cell .\",\n \"n = 61 epidermal cells from 12 larvae , 3–5 cells from each larva , from 3 independent experiments .\",\n \"( B ) KDEL distribution in third instar larval epidermal cells shows less spatial fluctuation in ReepA− ReepB− double mutant larvae compared to ReepA+ ( WT ) , and ReepA− and ReepB− single mutant larvae .\",\n \"This is confirmed by quantification as in A ) ; overall KDEL intensity is reduced in all ReepA and ReepB mutant genotypes .\",\n \"n = 45–61 epidermal cells in total from 12 to 16 different larvae , 3–5 cells from each larva , from 6 independent experiments .\",\n \"Intensity datapoints are experimentwise averages of multiple larvae , compared by paired t-tests , since larva-wise data were too significantly different to pool across experiments in this case .\",\n \"( C ) Electron micrographs show increased ER sheet profile length in ReepA− ReepB− double mutant third instar larval epidermal cells compared to single ReepA− and ReepB− mutants and controls .\",\n \"Arrowheads , ER sheets; M , mitochondria .\",\n \"( n = 56–58 cells in total from three independent larvae , 18–19 cells from each . ( D , E )\",\n \"ReepA− ReepB− double mutant larvae have an increased ER stress response compared to a ReepA+ ( WT ) control , measured by Xbp1::GFP expression , in larval epidermis ( D ) but not in neuronal cell bodies ( E ) .\",\n \"Graphs show quantification of Xbp1::GFP staining intensity relative to controls .\",\n \"( n = 12–15 larvae ) .\",\n \"All graphs show individual datapoints and mean ±SEM .\",\n \"Occasional outlier datapoints off the top of the scale are omitted from graphs but included in statistical analyses .\",\n \"ns , p>0 . 05; *p<0 . 02; **p<0 . 005; ***p<0 . 0003; ****p<0 . 0001 , two-tailed Student’s t-test .\",\n \"Scale bars: A , B , D , E , 10 µm; C , 0 . 5 µm ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 005 To understand the roles of reticulon and REEPs in axonal ER organization , we labeled axonal ER by expressing Acsl::myc ( O'Sullivan et al . , 2012 ) in two adjacent motor neurons using m12-GAL4 ( Xiong et al . , 2010 ) .\",\n \"Loss of Rtnl1 caused partial loss of Acsl::myc from posterior ( segment A6 ) axons , but not from anterior axons ( segment A2 ) ; it also caused Acsl::myc staining to appear more irregular in posterior but not anterior axons , reflected in a higher coefficient of variation ( SD/mean ) of Acsl::myc staining intensity along the length of posterior axons lacking Rtnl1 , compared to wild-type ( Figure 3A ) .\",\n \"These Rtnl1 loss-of-function phenotypes were found either on targeted knockdown of Rtnl1 in these motor axons or in Rtnl1− mutant larvae , and the Rtnl1− mutant phenotypes could be partially rescued by one copy of an Rtnl1Pacman genomic clone ( Figure 3A ) . 10 . 7554/eLife . 23882 . 006Figure 3 . Loss of either Rtnl1 or ReepB leads to partial loss of smooth ER marker from distal motor axons .\",\n \"( A ) Effects of Rtnl1 loss on ER , visualized by Acsl::myc expressed in two adjacent motor axons using m12-GAL4 .\",\n \"Images show effects of Rtnl1 knockdown , Rtnl1− mutation , rescue of Rtnl1− by Rtnl1Pacman and respective ReepA+ control ( WT ) axons .\",\n \"Graphs show mean staining intensity ( top graphs ) or coefficient of variation of intensity ( bottom graphs ) as a measurement of staining variability along the length of each axon .\",\n \"Rtnl1 loss leads to partial loss of Acsl::myc in posterior but not anterior axons ( top graphs ) , and to some disorganization of posterior axonal ER seen by increased coefficient of variation of Acsl::myc staining intensity .\",\n \"There is partial rescue of Rtnl1 phenotypes by one copy of a Rtnl1Pacman genomic clone .\",\n \"n = 15–20 larvae per genotype pooled from 5 independent experiments for RNAi; n = 24–36 larvae per genotype from 10 to 12 independent experiments for Rtnl1− mutant .\",\n \"Graphs show individual datapoints with mean ±SEM; ns , p>0 . 05; *p<0 . 03; **p<0 . 005; two-tailed unpaired Student’s t-test; two-tailed paired Student’s t-test was used when larva-wise data were too significantly different across experiments ( p<0 . 05 , ANOVA ) to pool .\",\n \"( B ) Effects of ReepA or ReepB loss on axonal ER .\",\n \"ReepB knockdown or ReepB− mutation , but not ReepA mutation , causes partial loss of smooth ER marker Acsl::myc in posterior but not in anterior motor axons .\",\n \"The ReepB− phenotype can be mostly rescued by one copy of a genomic ReepB::GFP clone .\",\n \"A ReepA− ReepB− double mutant shows a similar phenotype to a ReepB− single mutant .\",\n \"ReepA− ReepB− double mutant , but not single ReepA− or ReepB− mutants , show increased coefficient of variation of Acsl::myc staining levels in posterior axons; n = 7–47 larvae from 4 to 11 independent experiments .\",\n \"All graphs show individual datapoints with mean ±SEM; occasional outlier datapoints off the top of the scale are omitted from graphs but included in statistical analyses .\",\n \"ns , p>0 . 05; *p<0 . 04 **p<0 . 01 .\",\n \"ReepB RNAi was analyzed using two-tailed unpaired Student’s t-tests; multiple comparisons of ReepA and ReepB mutant genotypes were analyzed by ANOVA followed by post-hoc Tukey HSD tests .\",\n \"Scale bars , 10 µm . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 006 ReepA− mutants showed no loss of Acsl::myc from axons .\",\n \"Similar to Rtnl1 loss of function , ReepB− mutants showed partial loss of Acsl::myc from posterior but not anterior motor axons ( Figure 3B ) ; this phenotype was partially rescued with one copy of a genomic ReepB::GFP clone , and was also observed on ReepB knockdown in m12-GAL4-expressing neurons ( Figure 3B ) .\",\n \"A ReepA− ReepB− double mutant showed loss of Acsl::myc from posterior axons , that was similar to that seen in ReepB− mutants ( Figure 3B ) .\",\n \"ReepA− ReepB− double mutant larvae , but not ReepA− single mutants , showed an increased coefficient of variation of Acsl::myc staining intensity in posterior but not in anterior axons; ReepB mutant and knockdown axons both showed a slight increase in coefficient of variation in posterior axons , although this was not significant for ReepB mutants ( Figure 3B ) .\",\n \"In summary , loss of either Rtnl1 or at least ReepB alters axonal ER distribution in posterior motor axons , therefore supporting roles for these two protein families in axonal ER organization; in contrast loss of ReepA has either subtle or no effects .\",\n \"Since loss of both reticulon and REEP families in yeast removes most peripheral ER tubules ( Voeltz et al . , 2006 ) , we tested whether loss of both protein families in Drosophila might have similarly severe effects on axonal ER .\",\n \"Rtnl1− ReepA− ReepB− triple mutants were viable and fertile as adults , but survived poorly beyond two weeks of adulthood , compared to 4–5 weeks for wild-type adults .\",\n \"Triple mutant larvae showed increased fluctuation of Acsl::myc staining intensity along motor axons compared to wild-type , mainly in middle parts ( segment A4 and A5 ) of longer axons .\",\n \"At its most extreme , this manifested as fragmentation of Acsl::myc labeling , which was never seen in wild-type axons ( Figure 4A , B ) .\",\n \"When gaps in labeling were found in the central regions of axons , labeling in the anterior and posterior parts of the same axons was usually continuous ( Figure 4A ) .\",\n \"Double labeling of plasma membrane ( CD4::tdGFP ) and axonal ER ( Acsl::myc ) showed no effect on axonal plasma membrane in Rtnl1− ReepA− ReepB− triple mutant compared to control axons ( Figure 4—figure supplement 1 ) , implying that the phenotype was limited to ER and did not affect axon integrity .\",\n \"To compare genotypes and labels , we quantified irregular labeling along a 45 µm stretch of axon traversing the larval A4/A5 region in two ways: first we measured gaps in labeling , defined as intensity below a threshold that could consistently distinguish labeled axons above background labeling in the nerve; and second , we quantified the coefficient of variation of labeling intensity along the axon ( Figure 4C , D ) .\",\n \"Rtnl1 loss-of-function axons , but not ReepA− ReepB− mutant axons , also showed a mild ER fragmentation phenotype ( Figure 4A , B ) .\",\n \"Rtnl1 ( RNAi ) ReepA− ReepB− larvae also showed a fragmentation phenotype similar to Rtnl1− ReepA− ReepB− , suggesting that loss of Rtnl1 is essential for it ( Figure 4B ) .\",\n \"Rtnl1− ReepB− double mutant axons also showed a similar phenotype to Rtnl1− ReepA− ReepB− triple mutants .\",\n \"Therefore , loss of Rtnl1 causes a mild irregular organization of axonal ER , and this phenotype is exacerbated by loss of ReepB , with no detectable contribution of ReepA loss; and loss of both Reep proteins has no apparent effect . 10 . 7554/eLife . 23882 . 007Figure 4 . Loss of hairpin proteins leads to discontinuity of axonal ER staining .\",\n \"( A ) Rtnl1− larvae and Rtnl1− ReepA− ReepB− triple mutant larvae sometimes show fragmented axonal ER labeling in the middle parts ( segment A5 ) of long motor axons that express Acsl::myc under control of m12-GAL4 .\",\n \"Anterior ( segment A2 ) and posterior ( segment A6 ) portions of the same motor axons show continuous ER labeling .\",\n \"Arrowheads show gaps in Acsl::myc staining; brighter versions of the same images show gaps in staining in mutants but not wild-type ( WT ) , even when brightness of remaining staining is saturating .\",\n \"( B ) Panels show three examples of the range of Acsl::myc distributions found in the middle parts of long motor axons of each genotype of ReepA+ ( WT ) , Rtnl1− , Rtnl1− rescued with a Rtnl1Pacman construct , ReepA− ReepB− and Rtnl1− ReepB− double mutants , Rtnl1− ReepA− ReepB− triple mutant , and Rtnl1 ( RNAi ) ReepA− ReepB− larvae .\",\n \"A variety of phenotypes , from continuous to fragmented Acsl::myc labeling , are found in genotypes that lack Rtnl1 , and tend to be more severe in genotypes that also lack ReepB .\",\n \"Insets show examples of gaps in ER continuity ( arrowheads ) at higher magnification .\",\n \"Scale bars 10 µm , and 5 µm in higher zoom images .\",\n \"( C ) Percentages of a 45 µm length in the middle ( A4/A5 segment ) of each axon that lacks Acsl::myc staining , using an intensity threshold of 20 on a scale of 0–255 .\",\n \"Individual axons are plotted , together with median and interquartile range; comparisons use Mann-Whitney U-tests .\",\n \"All genotypes that lack Rtnl1 show more gaps than WT; Rtnl1− ReepA− ReepB− triple mutants do not differ from Rtnl1 ( RNAi ) ReepA− ReepB− or Rtnl1− ReepB− double mutants , but are significantly more severe than Rtnl1− single mutants .\",\n \"( D ) The coefficient of variation of Acsl::myc labeling in middle axon portions is increased in genotypes lacking Rtnl1 , relative to controls .\",\n \"Graphs show individual datapoints with mean ±SEM; occasional outlier datapoints off the top of the scale are omitted from graphs but included in statistical analyses .\",\n \"Comparison between Rtnl1− and Rtnl1− ReepA− ReepB− mutant genotypes was analyzed by two-tailed Student’s t-test , other multiple comparisons by ANOVA followed by Dunnett’s T3 test .\",\n \"ns , p>0 . 05; *p<0 . 04; **p<0 . 01; ***p<0 . 0005; ****p<0 . 0001; individual axons are plotted from 3 to 20 independent experiments , each with 2–3 different larvae for each genotype . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 00710 . 7554/eLife . 23882 . 008Figure 4—figure supplement 1 . Plasma membrane integrity is not affected in Rtnl1− ReepA− ReepB− triple mutant larvae ) .\",\n \"Coefficient of variation for plasma membrane labeling along axon length is unchanged in Rtnl1− ReepA− ReepB− triple mutant compared to control ( WT ) axons , whereas the ER to plasma membrane ratio of the coefficient of variation is increased .\",\n \"Higher zoom images of the boxed areas show gaps ( yellow arrowheads ) and brightly labeled axonal ER accumulations .\",\n \"Graphs show mean ±SEM; *p<0 . 02 , two-tailed Student’s t-test; n = 5 different larvae from 3 independent experiments . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 008 To confirm the physical continuity of ER in normal axons , and that gaps in Acsl::myc staining in Rtnl1− ReepA− ReepB− triple mutant axons reflect physical gaps in the ER network , we also visualized ER using fluorescent protein markers .\",\n \"The lipase CG9186::GFP ( Thiel et al . , 2013 ) , Rtnl1::GFP ( Rao et al . , 2016 ) and tdTomato::Sec61β ( Summerville et al . , 2016 ) all localize to ER , and all showed continuous labeling in wild-type ( ReepA+ ) axons ( Figure 5A; Figure 5—figure supplement 1A ) .\",\n \"However , CG9186::GFP showed occasional gaps in triple mutant axons ( Figure 5A ) , with at least one gap per larva visible by confocal microscopy in 6/17 larvae .\",\n \"Since confocal imaging might not reveal all physical discontinuities , and since gaps in marker distribution might not mean gaps in ER distribution , we probed the physical continuity of CG9186::GFP labeling using fluorescence recovery after photobleaching ( FRAP ) .\",\n \"After bleaching a 12 µm length of axon in either wild-type axons or in Rtnl1− ReepA− ReepB− triple mutant axons lacking gaps , we observed rapid recovery of fluorescence from both ends of the bleached region ( Figure 5; Figure 5—figure supplement 1B; Video 1 ) , suggesting no physical barrier to CG9186::GFP diffusion .\",\n \"The kinetics of recovery were similar between wildtype and triple mutant axons ( Figure 5B–D ) .\",\n \"We also performed FRAP on regions next to gaps in CG9186::GFP labeling in triple mutant axons ( Figure 5A ) ; here we observed good recovery of fluorescence from the end of the bleached region opposite the gap , but no recovery across the gap ( Figure 5; Figure 5—figure supplement 1C; Videos 2–4 ) , suggesting that gaps in CG9186::GFP labelling were also physical barriers to diffusion of CG9186::GFP , and hence gaps in the ER network . 10 . 7554/eLife . 23882 . 009Figure 5 . Live imaging of ER in wild type and hairpin mutant axons .\",\n \"( A ) Representative FRAP assay images from ReepA+ ( WT ) or Rtnl1− ReepA− ReepB− triple mutant axons in which ER was visualized using CG9186::GFP expression driven in two motor axons by m12-GAL4 .\",\n \"One triple mutant axon shows a gap in labeling ( arrowheads ) .\",\n \"Regions of interest ( 12 µm , between dashed lines ) were photobleached .\",\n \"Fluorescence was visualized before photobleaching ( ‘pre’ ) , and during recovery over 200 s .\",\n \"A kymograph was generated for each axon indicated by an arrow in the top panel .\",\n \"Most areas of intense and less intense ER labeling remain stable over time ( e . g . black arrows below kymographs ) ; occasional movements of ER features are indicated by white arrows in kymographs .\",\n \"( B ) Relative fluorescence intensities within the photobleached region were plotted ( mean ±SEM ) during recovery for wild-type axons or for triple mutant axons lacking ER gaps .\",\n \"( C–D )\",\n \"Quantification of half recovery time ( C ) and rate constant ( D ) for wild-type and triple mutant axons lacking gaps , with median and interquartile range .\",\n \"Data in B–D are from seven recordings from 4 wild-type larvae , and 20 recordings from 14 mutant larvae .\",\n \"ns , p>0 . 05; Mann-Whitney U test .\",\n \"Scale bars , 5 µm .\",\n \"( E ) Representative kymographs from time-lapse recording of CG9186::GFP in unbleached single ReepA+ ( WT ) or Rtnl1− ReepA− ReepB− triple mutant axons lacking gaps .\",\n \"The left panel shows retrograde movement of a ~ 3 µm length of ER labeling ( arrows ) ; the right panel shows anterograde movement of a ~ 5 µm stretch of ER labeling ( arrows ) .\",\n \"Retrograde movements of labeled puncta are seen in both panels ( arrowheads ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 00910 . 7554/eLife . 23882 . 010Figure 5—figure supplement 1 . Continuity , stability and movement of fluorescent ER labeling in wild-type ( WT ) and Rtnl1− ReepA− ReepB− triple mutant axons .\",\n \"( A ) Time series and kymographs of two closely apposed motor axons expressing either Rtnl1::GFP or Sec61bβ::tdTomato ER markers under control of m12-GAL4 .\",\n \"Images show continuity of ER staining in axons , and kymographs show stability of most ER features over the acquisition time .\",\n \"( B ) Kymographs showing fluorescence recovery after photobleaching in two ReepA+ ( WT ) axons , or in two Rtnl1− ReepA− ReepB− triple mutant axons lacking ER gaps .\",\n \"Recovery proceeds from both ends of the bleached region ( dotted lines ) .\",\n \"Most ER features seen before bleaching ( top line of kymographs ) can also be seen after recovery , implying that recovery proceeds by protein diffusion , not by movement of ER structures .\",\n \"( C ) Kymographs showing fluorescence recovery after photobleaching immediately distal to gaps in ER staining , in four Rtnl1− ReepA− ReepB− triple mutant axons .\",\n \"Recovery proceeds only from the distal end of the bleached region ( dotted lines ) opposite the gap , but not across the proximal gap .\",\n \"While most ER gaps and many other features are stable , one axon shows a retrogradely moving particle ( white arrowhead ) , and one shows anterograde movement of ER that closes one gap and opens another more proximally ( white asterisk ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01010 . 7554/eLife . 23882 . 011Video 1 . Representative time-lapse microscopy of FRAP analysis in wild-type axons . Wild-type ( WT ) axons expressing CG9186::GFP under control of m12-GAL4 were photobleached as described in Figure 5A .\",\n \"The video shows 2 s of pre-bleach images and 200 s of post-bleach images at one frame every 5 s .\",\n \"The bleached area is shown with a rectangle in the bleaching frame . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01110 . 7554/eLife . 23882 . 012Video 2 . Representative time-lapse microscopy of FRAP analysis in Rtnl1− ReepA− ReepB− triple mutant axons . Rtnl1− ReepA− ReepB− triple mutant axons expressing CG9186::GFP were photobleached and imaged as in Video 1 , in a region immediately distal to a gap ( between white arrows ) in ER labeling in one axon . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01210 . 7554/eLife . 23882 . 013Video 3 . FRAP analysis in Rtnl1− ReepA− ReepB− triple mutant axons showing a CG9186::GFP labeled particle passing retrogradely across the gap region . Rtnl1− ReepA− ReepB− triple mutant axons expressing CG9186::GFP were photobleached and imaged immediately distal to an ER gap as in Video\",\n \"2 . A labeled particle ( white arrow ) can be seen moving retrogradely across the ER gap ( white arrowhead ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01310 . 7554/eLife . 23882 . 014Video 4 . FRAP analysis in Rtnl1− ReepA− ReepB− triple mutant axons showing dynamic ER gap generation . Rtnl1− ReepA− ReepB− triple mutant axons expressing CG9186::GFP were photobleached and imaged immediately distal to an ER gap as in Video\",\n \"2 . Anterograde movement of a 5 µm length of ER simultaneously closes up a gap distal to it ( first white arrow ) and opens up a new gap ( second white arrow ) proximal to it .\",\n \"A kymograph from this preparation is shown in Figure 5—figure supplement 1C . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 014 During live imaging , we also followed the stability of the ER network in both wild-type and mutant axons .\",\n \"Features such as higher or lower levels of CG9186::GFP labelling ( which could reflect variation in the numbers of tubules , or the presence of structures like cisternae ) were mostly stable throughout a 200 s acquisition period ( Figure 5A , E; Figure 5—figure supplement 1 ) .\",\n \"However , we also observed some dynamic features , including anterograde or retrograde movement of more brightly labeled regions ( Figure 5A , E ) perhaps representing movement of ER tubules detached from the ER network , and retrograde movement of labeled puncta ( Figure 5E; Figure 5—figure supplement 1C; Video 3 ) , at around 0 . 3 µm/s .\",\n \"While most gaps in ER labeling in triple mutants were stable over 200 s ( 7 out of 8 gaps imaged , including Videos 2–3 ) , in one case a gap was closed up by anterograde movement of a 5 µm length of ER proximal to the gap , which simultaneously opened up a new gap proximal to the moving section ( Figure 5—figure supplement 1C; Video 4 ) .\",\n \"In one case a labeled punctum moved retrogradely across an ER gap without pausing ( Figure 5—figure supplement 1C ) , implying that microtubule-based transport was intact .\",\n \"Therefore , live imaging and photobleaching of CG9186::GFP both suggest that ER is normally continuous in axons , that loss of reticulon and REEP proteins leads to occasional gaps in ER labeling that represent physical breaks of ER continuity , and that ER network organization in axons shows both static and dynamic features .\",\n \"We also tested for possible defects in axon transport by staining for abnormal accumulation of the synaptic vesicle protein CSP in axons .\",\n \"Rtnl1− larvae and Rtnl1− ReepA− ReepB− triple mutant larvae showed large accumulations of CSP in many peripheral nerves .\",\n \"The large accumulations of CSP in Rtnl1− larvae could be rescued by two copies of a Rtnl1Pacman genomic clone ( Figure 6 ) . 10 . 7554/eLife . 23882 . 015Figure 6 . Loss of Rtnl1 causes mild accumulation of synaptic vesicles in axons .\",\n \"( A ) Peripheral nerves of Rtnl1− and Rtnl1− ReepA− ReepB− triple mutant larvae show larger accumulations of synaptic vesicle protein CSP ( e . g . yellow arrows ) , and smaller elongated CSP puncta ( e . g . yellow arrowheads ) .\",\n \"In contrast , control larvae ( WT ) and ReepA− ReepB− double mutant larvae show an even distribution of small round CSP puncta .\",\n \"CSP accumulations are not significantly bigger in Rtnl1− ReepA− ReepB− triple mutants than in Rtnl1− mutants .\",\n \"The CSP accumulations in Rtnl1− larvae can be rescued by two copies of a Rtnl1Pacman genomic clone .\",\n \"All axons shown are crossing abdominal segment A2 .\",\n \"( B ) Graph shows mean ±SEM; n = 16–92 axons from 8 to 46 larvae , from 3 different experiments .\",\n \"ns , p>0 . 05; **p<0 . 006; ****p<0 . 0001 , two-tailed Student’s t-test .\",\n \"Scale bar , 10 µm ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 015 To better understand wild-type axonal ER organization , and the mutant phenotypes seen in confocal microscopy , we performed electron microscopy ( EM ) on 60-nm-thick serial sections of third instar peripheral nerves .\",\n \"Peripheral nerves contain both motor and sensory axons , arranged in fascicles , and wrapped in three main classes of glial cell ( Stork et al . , 2008; Matzat et al . , 2015 ) .\",\n \"We used ROTO staining ( Tapia et al . , 2012; Terasaki et al . , 2013 ) to preferentially highlight cellular membranes including ER .\",\n \"Wild-type larvae showed a network of ER tubules , in every axon that could be observed ( Figure 7A left; serial sections in Video 5 ) .\",\n \"Tubule outer diameter averaged around 40 nm ( Figure 7B , C ) , and axons contained an average of around 1 . 6 ER tubules in each cross-section ( Figure 7D , E ) .\",\n \"Reconstruction ( Figure 7F left; Supplementary file 2 ) showed a tubular network with multiple branches , some dead ends , and continuity along nearly every axon sectioned .\",\n \"ER tubules often showed proximity to mitochondria or plasma membrane ( Figure 7G left ) .\",\n \"We also found occasional structures resembling small patches of ER sheets with an adjoining cisterna ( Figure 7H ) , continuous with the tubular ER network ( Video 6 ) , at a frequency averaging around one per 20 µm per axon . 10 . 7554/eLife . 23882 . 016Figure 7 . Loss of hairpin proteins leads to fewer but enlarged ER tubules in larval peripheral nerve axons .\",\n \"( A ) EMs of peripheral nerve axons from wild-type ReepA+ ( WT , left ) or Rtnl1− ReepA− ReepB− triple mutant ( right ) larvae .\",\n \"Arrowheads indicate ER tubules ( seen as continuous structures in serial sections in Videos 5 and 7 ) .\",\n \"Triple mutant axons show enlarged ER tubules , sometimes with a clear lumen ( arrows ) , seldom observed in wild-type .\",\n \"Quantification of ER tubule diameter ( B , C ) and tubules per axon cross-section ( D , E ) in control and mutant larvae .\",\n \"Data from individual ER tubules or axons are shown in B and D , averaged larval values , mean ±SEM in C and E . ( F ) 3D reconstruction of a 4 . 5 µm axon segment from wild-type ( left ) or mutant ( right ) peripheral nerves , generated from 75 serial 60 nm sections , showing ER ( green ) , mitochondria ( magenta ) and plasma membrane ( gray ) .\",\n \"Interactive versions of the reconstructions are in Supplementary files 2 and\",\n \"3 . ( G ) Electron micrographs of peripheral nerve axons from wild-type ( ReepA+ , left ) or Rtnl1− ReepA− ReepB− triple mutant ( right ) larvae , showing proximity of ER to mitochondria ( arrowheads ) or plasma membrane ( arrows ) .\",\n \"( H ) Representative EM of short ER sheet ( arrowhead ) and cisterna ( asterisk ) from a wild-type larva; further sections in Video 6 .\",\n \"( I ) section of an axonal swelling from a wild-type larva showing mitochondria ( m ) , vesicles ( v ) , and a large clear cisterna ( c ) ; further sections in Video 8 .\",\n \"( J ) Serial EM sections show ER discontinuity in two mutant axons: axon 1 lacks ER tubules in sections z6-z14 and axon 3 in sections z1-z11; neighboring axons ( e . g . axon 2 ) show a continuous ER network .\",\n \"( K ) 3D reconstruction of a 4 . 5 µm axon segment from mutant peripheral nerves , generated from 75 serial 60 nm sections , showing multiple gaps ( indicated by arrows ) .\",\n \"ER is in green and plasma membrane in gray .\",\n \"Raw EM data and an interactive version of the reconstruction are in Video 9 and Supplementary file 4 , respectively .\",\n \"( L ) Frequency of axons with gaps in the 4 . 5 µm lengths analyzed .\",\n \"ER gap length ( M ) , numbers of ER gaps per µm ( N ) , and proportion of axon length with gaps ( O ) in affected axons .\",\n \"In M–O , top graphs show data from individual ER gaps ( M ) or individual axons ( N–O ) , with second and third quartiles and 5th and 95th percentiles; bottom graphs show averaged larval values , mean ±SEM .\",\n \"In all graphs , ns p>0 . 05; *p<0 . 04; **p<0 . 003; ****p<0 . 0001; Mann-Whitney U test for B , D , top graphs in M–O; two-tailed Student’s T test for C , E , L , bottom graphs in M–O .\",\n \"Scale bars , 500 nm . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01610 . 7554/eLife . 23882 . 017Video 5 . Serial EM sections of a wild-type peripheral nerve , showing continuity of tubular membrane structures through multiple sections . The sixth section in the series is shown in Figure 7A ( left ) .\",\n \"ER tubules were identified as darkly stained structures present for multiple sections .\",\n \"See Figure 7A for annotations . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01710 . 7554/eLife . 23882 . 018Video 6 . Serial EM sections of a sheet-like ER structure in a wild-type axon , shown in Figure 7H . Arrows indicate short ER sheets and associated cisternae , present across multiple sections .\",\n \"Note that these structures usually appear continuous with tubules in adjacent sections . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 018 If Rtnl1− ReepA− ReepB− triple mutant larvae have less ER membrane curvature , we would expect them to have larger ER tubules , fewer tubules per section , loss of tubules , or a combination of these .\",\n \"EM revealed all these mutant phenotypes to varying degrees ( Figure 7 ) .\",\n \"ER tubule diameter was increased to around 60 nm ( Figure 7B , C ) , allowing a lumen to be seen in some tubules ( Figure 7A right; Video 7 ) that was rarely seen in wild-type ( Figure 7A left; Video 5 ) , and most triple mutant axons exhibited only a single ER tubule ( Figure 7D , E ) .\",\n \"Reconstructions ( Figure 7F right; Supplementary file 3 ) showed a less extensive ER network in mutant axons .\",\n \"Frequent contacts of ER with mitochondria and plasma membrane were found in both wild-type and Rtnl1− ReepA− ReepB− mutant axons ( Figure 7G ) .\",\n \"Both wild-type and mutant axons showed swellings containing mitochondria and clusters of vesicles resembling synaptic vesicles ( Figure 7I; Video 8 ) .\",\n \"We did not detect large swellings with synaptic vesicle materials similar to those seen with confocal microscopy ( Figure 6 ) , but this may reflect the much shorter lengths of nerve examined by EM , around 5 µm compared to around 50 µm in confocal . 10 . 7554/eLife . 23882 . 019Video 7 . Serial EM sections of a Rtnl1− ReepA− ReepB− peripheral nerve showing continuity of tubular membrane structures through multiple sections . The sixth section in the series is shown in Figure 7A ( right ) .\",\n \"ER tubules were identified as darkly stained structures present for multiple sections .\",\n \"See Figure 7A for annotations . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 01910 . 7554/eLife . 23882 . 020Video 8 . Serial EM sections of a wild-type axonal swelling shown in Figure 7I . An axon with a swelling is highlighted in the first frame .\",\n \"Sections show accumulated mitochondria , vesicles , and a large clear cisterna , as labeled in Figure 7I . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 020 Serial EM sections also revealed variable fragmentation of ER in Rtnl1− ReepA− ReepB− mutant axons ( Figure 7J , K ) , consistent with that seen using confocal microscopy ( Figure 4 ) .\",\n \"Some discontinuity of the ER network was observed in about 10% of wild-type or Rtnl1− ReepA− ReepB− mutant axons ( Figure 7L ) .\",\n \"However , gaps in Rtnl1− ReepA− ReepB− mutant axons ( Figure 7J , K; Video 9; Supplementary file 4 ) were longer ( Figure 7M ) , and slightly more numerous ( but not significantly when averaged across larvae; Figure 7N ) than in wild-type axons , resulting in nearly a four-fold increase in the length of affected axons that lacked ER tubules ( Figure 7O ) . 10 . 7554/eLife . 23882 . 021Video 9 . Serial EM sections of a 4 . 5 µm segment of a Rtnl1− ReepA− ReepB− mutant axon with disrupted continuity of ER , used for 3D reconstruction in Figure 7K . An axon with gaps in its ER network is highlighted in the first frame .\",\n \"Continuous ER tubules were identified as the presence of signals at the same position for three or more sections .\",\n \"Given the varying brightness and contrast of EM sections , faint staining that coincided with a tubule signal in adjacent sections was also considered as an ER tubule .\",\n \"Complete loss of ER tubules from three or more sections was defined as a gap . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 021 Rtnl1− ReepA− ReepB− mutant peripheral nerves also showed glial cell phenotypes .\",\n \"Wild-type peripheral nerves are surrounded by an outer perineurial glial cell , and just beneath this a subperineurial glial cell; axons or axon fascicles are wrapped imperfectly by a wrapping glia cell ( Stork et al . , 2008; Matzat et al . , 2015 ) .\",\n \"All glial classes , but particularly subperineurial glia , showed a trend towards increased ER sheet profile length compared to control cells ( Figure 8A–I ) , similar to Rtnl1 knockdown ( O'Sullivan et al . , 2012 ) or ReepA− ReepB− double mutant ( Figure 2 ) epidermal cells .\",\n \"Triple mutant wrapping glia also displayed more extensive wrapping , sometimes completely ensheathing axons , which was rarely observed in control nerve sections ( Figure 8J–M; Figure 8—figure supplement 1 ) . 10 . 7554/eLife . 23882 . 022Figure 8 . Loss of reticulon and REEP proteins leads to ER disorganization in glial cells and hyper-wrapping of peripheral axons .\",\n \"( A–B )\",\n \"EMs of peripheral nerve sections from ReepA+ ( WT ) ( A ) or Rtnl1− ReepA− ReepB− triple mutant ( B ) larvae .\",\n \"Perineurial , subperineurial , and wrapping glial cells are shaded green , blue , and magenta , respectively .\",\n \"( C–H )\",\n \"Higher magnification images of perineurial ( C , F ) , subperineurial ( D , G ) , and wrapping ( E , H ) glia from wild-type ( C–E ) or mutant ( F–H ) nerve sections , showing ER tubules ( white arrows; confirmed as tubules by presence in adjacent sections ) and sheets ( white arrowheads ) .\",\n \"Note the longer ER sheet profiles and fewer ER tubules in the subperineurial ( G ) and wrapping glial cells ( H ) of mutant nerves .\",\n \"Asterisks show mitochondria; yellow arrowheads show glial plasma membrane , identified by its continuity .\",\n \"( I ) ER sheet profile length in control and mutant glial cells from 3 wild-type and 4 mutant larvae .\",\n \"Top graphs represent all individual sheet profiles , with second and third quartiles and 5th and 95th percentiles; bottom graphs show averaged larval values , showing mean ±SEM .\",\n \"Two-way ANOVA showed a significant effect of genotype ( p<0 . 002 ) but not glial class ( p>0 . 3 ) on ER sheet length , with no interaction between factors ( p>0 . 3 ) .\",\n \"( J , K )\",\n \"Sketches of wild-type ( J ) and mutant ( K ) wrapping glial cells , showing excess processes in the triple mutant .\",\n \"Asterisks indicate completely wrapped one-axon or two-axon fascicles , rarely seen in wild-type nerve sections .\",\n \"More examples of each phenotype are in Figure 8—figure supplement 1 .\",\n \"( L–M )\",\n \"Quantification of wrapping glial membrane profile length per nerve cross-section ( L ) and percentage of axons that are wrapped individually or as two-axon fascicles ( M ) in wild-type and mutant nerve sections ( mean ±SEM ) .\",\n \"ns , p>0 . 05; *p<0 . 05; ***p<0 . 001; ****p<0 . 0001 .\",\n \"Mann-Whitney U test ( I ) , top graphs ) ; two-tailed Student’s t-test ( I ) , bottom graphs; L , M , ) .\",\n \"Scale bars , 1 µm . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 02210 . 7554/eLife . 23882 . 023Figure 8—figure supplement 1 . Sketches of wrapping glia in wild-type ( ReepA+ ) and Rtnl1− ReepA− ReepB− mutant peripheral nerves , similar to those in Figure 8J , K . Sketches of wild-type ( ReepA+ , upper ) and mutant ( lower ) wrapping glial cells were abstracted from cross-sections of peripheral nerves .\",\n \"Excess formation of processes were observed in triple mutants . DOI: http://dx . doi . org/10 . 7554/eLife . 23882 . 023\"\n ],\n [\n \"The existence of a tubular axonal ER network has been known for decades .\",\n \"Nevertheless , the cellular mechanisms that organize a compartment that is usually distributed throughout cells , along the great lengths of axons , are until now largely unknown .\",\n \"The finding that several causative genes for the axon degenerative disease HSP encode ER modeling proteins , suggests a link between ER modeling and axon function or maintenance , and provides candidate proteins that may be instrumental in structure and function of the axon ER network .\",\n \"These candidates include several hairpin-loop-containing HSP proteins , of the spastin , atlastin , reticulon , REEP , and Arl6IP1 families , that influence ER structure in situations including yeast , mammalian cultured cells , and neuronal cell bodies in vivo ( Shibata et al . , 2006; Voeltz et al . , 2006; Hu et al . , 2008; Shibata et al . , 2009; Park et al . , 2010; Shibata et al . , 2010 ) .\",\n \"The HSP-related protein families that model ER , and some other proteins that interact with them , share a common feature of one or two intramembrane hairpin loops that can insert into the cytosolic face of the ER membrane , thereby recognizing or inducing curvature .\",\n \"This property makes the reticulon and REEP ( DP1 ) families together responsible for most peripheral ER tubules in yeast , and contribute to the curved edges of ER sheets ( Voeltz et al . , 2006; Hu et al . , 2008 ) .\",\n \"The latter property may explain the expansion of ER sheets in Drosophila lacking the reticulon Rtnl1 ( O'Sullivan et al . , 2012 ) and in REEP1 homozygous mutant mice ( Beetz et al . , 2013 ) .\",\n \"Given this background , we set out to test how far the reticulon and REEP families contribute to axonal ER organization .\",\n \"REEP1 homozygous mutant mice were not previously tested for effects on axonal ER , although knockdown of Drosophila Rtnl1 led to partial loss of smooth ER marker in distal motor axons ( O'Sullivan et al . , 2012 ) .\",\n \"Here we build on this work by analyzing mutants of all the widely expressed and highest conserved members of the reticulon and REEP families in Drosophila: Rtnl1 , an ortholog of all four human reticulons ( O'Sullivan et al . , 2012 ) ; ReepA , an ortholog of human REEP1-REEP4; and ReepB , an ortholog of human REEP5-REEP6 .\",\n \"We monitored phenotypes of these mutants , singly and in combination , by confocal microscopy of axonal ER markers in small numbers of motor neurons , live imaging and photobleaching of ER , and EM using membrane-specific staining .\",\n \"Confocal microscopy revealed a partial loss of ER marker in distal but not in more proximal motor axons in Rtnl1 and in ReepB mutants .\",\n \"Although the only REEP genes identified as causative for HSP are REEP1 and REEP2 , loss of their ortholog ReepA had at most only mild effects on axonal ER ( Figures 3 and 4 ) .\",\n \"The stronger axonal ER phenotypes of ReepB or Rtnl1 loss of function , compared to ReepA− mutants , is consistent with the higher levels of ReepB and Rtnl1 expression , judged by transcriptomics ( www . flyatlas . com ) , and the detection of ReepB and Rtnl1 but not ReepA fusions in peripheral nerves ( Figure 1F , G; O'Sullivan et al . , 2012 ) or individual axons ( Figure 1H , J ) .\",\n \"The partial loss of distal axonal ER marker in both mutants and motor-neuron knockdowns , of either Rtnl1 or ReepB ( Figure 3; O'Sullivan et al . , 2012 ) , and qualitatively similar ER fragmentation phenotypes in Rtnl1 ReepA ReepB triple loss-of-function genotypes , obtained using either Rtnl1 mutants or motor-neuron knockdown ( Figure 4 ) , suggest that the axonal ER phenotypes are cell autonomous in motor neurons .\",\n \"Given the joint and partly redundant requirement of the reticulon and REEP families for ER tubule formation in yeast ( Voeltz et al . , 2006 ) , we tested whether this was also true for axonal ER .\",\n \"Flies lacking Rtnl1 , ReepA and ReepB – probably equivalent to mammals lacking all four reticulons and all six REEPs , and homozygous viable – indeed showed more extreme ER phenotypes than axons lacking either Rtnl1 or ReepA and ReepB alone ( Figures 3 and 4 ) .\",\n \"However , only a fraction of mutant axons showed severe fragmentation , and even affected axons still had continuous labeling of axons with ER marker through much of their length .\",\n \"ER fragmentation in the middle parts of long axons might be a consequence of axon expansion during larval growth , in which the somatic and presynaptic ends of the axon are gradually pulled apart , with insufficient ER-modeling proteins to maintain the expanding tubular network throughout the axoplasm .\",\n \"Therefore reticulon and REEP proteins are present in axons and have roles in ER organization there – but since triple mutant axons still mostly possess ER , there must be additional proteins required too .\",\n \"These might be found among the increasing number of other HSP genes that encode ER proteins with possible hairpins , such as Arl6IP1/SPG61 , which affects ER organization ( Novarino et al . , 2014; Yamamoto et al . , 2014; Fowler and O'Sullivan , 2016 ) , or C19orf12/SPG43 ( Landouré et al . , 2013 ) .\",\n \"The variable nature of the triple mutant fragmentation phenotype might reflect stochastic variation in the amounts of such proteins , the amount of ER present , external factors like physical stresses during larval movement , or dynamic fluctuations in the local levels and connections of ER tubules .\",\n \"The tubular ER network is highly dynamic in non-neuronal cells ( Nixon-Abell et al . , 2016; Valm et al . , 2017 ) and in axons our live imaging suggests dynamic features superimposed on a structure that is largely stable over the 1–2 min of imaging ( Figure 5 ) .\",\n \"EM examination of wild-type axonal ultrastructure revealed that ER tubules were effectively ubiquitous in Drosophila peripheral nerve sections , as seen previously in mammalian neurons ( Tsukita and Ishikawa , 1976; Villegas et al . , 2014 ) , albeit with fewer tubules , presumably reflecting the smaller diameters of the axons examined here .\",\n \"Reconstruction over several µm showed a continuous network of ER tubules in most axons examined , in agreement with the ER continuity found in neurons by lipid dye labeling ( Terasaki et al . , 1994 ) , and recently in reconstructions of serial sections from focused ion beam SEM ( Wu et al . , 2017 ) .\",\n \"However , in a few axons , we found short lengths of axon with no detectable ER ( Figure 7J–O ) .\",\n \"There could be several reasons for this: Terasaki et al . ( 1994 ) only assessed continuity in dendrites , cell body and proximal axon; some of the gaps we observe in EM could be short transient gaps in a dynamic network; some of the apparent gaps could be ‘thin ER’ observed using higher-resolution focused-ion-beam SEM ( Wu et al . , 2017 ) , but that might be missed using our approach; larval axons with low diameters might be intrinsically more susceptible to occasional gaps in the ER network , than wider axons with more tubules; and we might occasionally miss an ER tubule due to weaker staining , or close proximity to other structures like plasma membrane .\",\n \"We also observed occasional small ER sheet-like structures in wild-type axons ( Figure 7H; Video 6 ) .\",\n \"Although we do not see ribosomes on these , this could be due to lack of staining by the ROTO protocol .\",\n \"As discussed above , rough ER and translation are relatively sparse in axons , but low levels of rough ER are possible , and consistent with the occasional sheet structures observed here .\",\n \"EM also showed phenotypes consistent with loss of ER membrane curvature in Rtnl1− ReepA− ReepB− triple mutant axons ( Figure 7 ) .\",\n \"Mutant axons had ER tubules of larger diameter , fewer tubules per axon cross-section , and consistent with our confocal data ( Figure 4 ) , longer gaps in the ER network than wild-type , although most parts of most mutant axons examined still had a continuous ER network ( Figure 7 ) .\",\n \"These mutant phenotypes could potentially have physiological consequences .\",\n \"Larger tubules could potentially store and release more calcium than thinner ones , while the reduced network could make the role of the ER in calcium buffering or release more localized .\",\n \"The less extensive ER network in mutants might also reduce the amount of contact between ER and other organelles , with consequences for calcium and lipid homeostasis that require these contacts , or for regulation of mitochondrial fission ( Friedman et al . , 2011 ) – although the continuing proximity of ER to mitochondria and plasma membrane in mutants means that any effects are presumably quantitative rather than qualitative .\",\n \"The reduced curvature of ER membrane in mutants might also influence their protein composition , since many membrane proteins have mechanisms for recognizing differential membrane curvature ( Antonny , 2011 ) .\",\n \"The occasional lack of continuity could prevent propagation of ER-dependent Ca2+ signals like those seen in injured mammalian sensory neurons ( Cho et al . , 2013 ) ; it could also cause local impairments in Ca2+ or lipid homeostasis that could lead to local transport inhibition , as is the case for mitochondrial transport ( Wang and Schwarz , 2009 ) , although lack of ER continuity does not appear to directly prevent axon transport ( Figure 5—figure supplement 1C; Video 3 ) .\",\n \"Sporadic lack of ER continuity might explain the preferential sensitivity of distal longer axons to HSPs , since these would be more likely to suffer from a gap in ER continuity to the cell body , compared to proximal or shorter axons .\",\n \"In this model , disease-causing alleles in single hairpin-encoding genes could promote degeneration in distal motor axons by increasing the probability of such gaps , dependent on factors such as age , axon length or diameter , and ER tubule density and dynamics .\",\n \"The apparent ubiquity of ER in axons , the extent of its continuity over long distances , and the preferential susceptibility of distal longer axons to mutations that affect ER-modeling proteins , all point to important physiological roles of this compartment and of its continuity .\",\n \"In this work we have begun to reveal the mechanisms that determine its organization .\",\n \"We have shown roles for two protein families that contain HSP disease gene products , in influencing the shape of individual tubules and the axonal ER network , with potential physiological consequences that would also be affected by mutations in these genes .\",\n \"Understanding the consequences of axonal ER structural defects for ER dynamics and axonal physiology , both in the genotypes we have described here , and in other genotypes that might also affect axonal ER organization , will provide models for the potential physiological defects in HSP and other axon degeneration diseases .\"\n ],\n [\n \"ReepA541 ( referred to as ReepA− ) , and ReepB48 ( referred as ReepB− ) mutants were generated by imprecise excision of P elements CB-0501–3 ( RRID:DGGR_123207 ) and EY05130 ( RRID:BDSC_16636 ) shown in Figure 1 .\",\n \"A precise excision generated in these experiments , ReepA+C591 ( referred to as ReepA+ ) was used as a genetic background control where feasible .\",\n \"Rtnl11 , referred as Rtnl1− , was a gift from G . Tear ( Wakefield and Tear , 2006; FlyBase ID FBal0246222 ) .\",\n \"Rtnl1− , ReepA− and ReepB− recombinants were generated by meiotic recombination on the second chromosome , and recombinants were screened using PCR primers ( Supplementary file\",\n \"5 ) to diagnose wild-type or mutant alleles of all three genes .\",\n \"Mutant and wild-type stocks were frequently genotyped to ensure that experimental flies were not contaminated .\",\n \"For knockdown experiments , either UAS-Rtnl1-RNAi line 7866 ( construct GD900 , which has no predicted off-targets; FlyBase ID FBti0098310 ) , or the w1118 control stock , 60000 ( both obtained from the Vienna Drosophila RNAi Center , www . vdrc . at ) , or the UAS-ReepB-RNAi line 8331 R-3 ( National Institute of Genetics Fly Stock Center , Japan; FlyBase ID FBal0275953 ) was crossed with UAS-Dcr2; CyO/If; m12-GAL4 , UAS-Acsl::myc .\",\n \"UAS-Dcr2 ( RRID:BDSC_24648; Dietzl et al . , 2007 ) was also present for knockdown in Rtnl1 ( RNAi ) ReepA− ReepB− larvae .\",\n \"Other fly stocks used were P{UAS-Acsl . 715 . MycC}3 ( RRID:BDSC_32330; Zhang et al . , 2009 ) , PBac{681 . P . FSVS-1}Rtnl1CPTI001291 ( RRID:DGGR_115146; Wakefield and Tear , 2006 ) , m12-GAL4 ( Xiong et al . , 2010; RRID:BDSC_2702 ) , UAS-Rtnl1::GFP ( Rao et al . , 2016 ) , {UAS-Xbp1 . EGFP . LG}4 ( RRID:BDSC_39719; Ryoo et al . , 2007 ) and UAS-tdTomato::Sec61β ( RRID:BDSC_64746; Summerville et al . , 2016 ) .\",\n \"A second-chromosome insertion of P{UAS-CG9186::GFP} ( Thiel et al . , 2013 ) was mobilized onto the third chromosome using the P transposase source P{∆2–3}99B ( Robertson et al . , 1988; RRID:BDSC_3612 ) , and expressed using m12-GAL4 in either a ReepA+ or an Rtnl1− ReepA− ReepB− mutant second chromosome background .\",\n \"For rescue of Rtnl11 we generated transgenic flies carrying P[acman] clone CH322-124P15 inserted at attP2 on chromosome 3 ( Bloomington stock 25710 ) , referred to as Rtnl1Pacman ( Venken et al . , 2009 ) .\",\n \"For C-terminal EGFP-LAP-tagging of ReepA and ReepB we used recombineering with the P[acman] system with minor modifications ( Venken et al . , 2009 ) , utilizing the CH322-97D15 ( ReepA ) and CH322-16N11 ( ReepB ) BAC clones ( BPRC; http://bacpac . chori . org ) .\",\n \"An EGFP-LAP-tagging cassette was amplified from R6Kamp-LAP ( GFP ) ( Poser et al . , 2008 ) using primers ( Supplementary file\",\n \"5 ) with 50 bp homology to the corresponding Reep clone and around 20 bp of homology to the tagging cassette , and used to transform the recombineering E . coli SW102 strain .\",\n \"Correct clones were verified at every step by PCR and/or restriction digestion .\",\n \"DNA for Drosophila transformation was extracted using a PureLink HiPure Maxiprep kit ( Invitrogen ) , and injected into y w M ( eGFP , vasa-integrase , dmRFP ) ZH-2A; M ( attP ) ZH-51D or y w M ( eGFP , vasa-integrase , dmRFP ) ZH-2A; M ( attP ) ZH-86Fb ( RRID:BDSC_24483 or RRID:BDSC_24749 , respectively ) at the Department of Genetics embryo injection facility , University of Cambridge , UK .\",\n \"Transformant lines were screened for the presence of the insert by PCR and GFP fluorescence .\",\n \"The primers used are described in Supplementary file 5 .\",\n \"BLAST sequence searches were used to define genome coordinates of P-element excisions , and Pfam domain coordinates in coding regions , and compare protein divergence rates .\",\n \"They were performed at the National Center for Biotechnology Information ( www . ncbi . nlm . nih . gov ) .\",\n \"REEP dendrograms were drawn from a ClustalW alignment ( Larkin et al . , 2007 ) using the neighbor-joining algorithm in MEGA 5 . 05 ( Tamura et al . , 2011 ) .\",\n \"Third instar larvae were dissected in chilled Ca2+-free HL3 solution ( Stewart et al . , 1994 ) , and fixed for 30 min in PBS with 4% formaldehyde .\",\n \"Dissected Drosophila preparations were permeabilized in PBS containing 0 . 3% Triton X-100 ( PBT ) at room temperature , and blocked in PBT with 4% bovine serum albumin for 30 min at room temperature .\",\n \"Primary antibodies were: Csp ( 6D6 , RRID:AB_10013286 , 1:50; Zinsmaier et al . , 1994 ) , Dlg ( 4F3 , RRID:AB_2314321 , 1:100; Parnas et al . , 2001 ) , ( both from the Developmental Studies Hybridoma Bank , Iowa , USA ) , GFP ( Ab6556 , RRID:AB_305564 , 1:600 Abcam , UK ) , HRP ( P97899 , RRID:AB_2314650 , 1:300 , Sigma ) , KDEL ( Ab50601 , RRID:AB_880636 , 1:25 , Abcam , UK ) , myc ( 2272 , RRID:AB_331667 , 1:25 , Cell Signaling , USA ) .\",\n \"Fixed preparations were mounted in Vectashield ( Vector Laboratories , USA , RRID:AB_2336789 ) , and images were collected using EZ-C1 acquisition software ( Nikon ) on a Nikon Eclipse C1si confocal microscope ( Nikon Instruments , UK ) .\",\n \"Images were captured using 10x/0 . 30NA , or a 60x/1 . 4NA oil objective .\",\n \"Confocal images were analyzed blind to genotype using ImageJ ( Schneider et al . , 2012 ) .\",\n \"Images of entire epidermal cells were obtained as z-projections of three consecutive sections .\",\n \"Using the line tool of ImageJ a 12 µm line was drawn from the nuclear envelope towards the periphery of each cell analyzed .\",\n \"Pixel intensity along the line was recorded in an Excel file .\",\n \"Local variance of intensity was calculated by dividing the rolling variance of the intensity ( in 10-pixel windows ) , by the rolling mean intensity , all along the line .\",\n \"Proximal ( anterior ) axons were imaged from segment A2 , middle images were from the end of segment A4 and A5 , distal ( posterior ) axons were imaged from segment A6 of third instar larvae .\",\n \"Mean gray intensity for single-axon images was measured by drawing a 45 µm line , either along both M12-GAL4-expressing axons ( where they could not be separated ) , or along the most strongly labeled axon ( where they appeared as separate axons ) , and quantifying gray intensity ( 0–255 ) by ImageJ; occasional images with saturated pixels were excluded from analysis after blinding .\",\n \"Coefficient of variation was calculated by dividing the standard deviation of staining intensity by the mean; occasional images with faint staining throughout the axon were excluded from analysis after blinding .\",\n \"Gaps were defined as regions where staining intensity was less than 20 ( out of 255 ) , after background subtraction .\",\n \"FRAP experiments were performed on a Nikon Eclipse C1si confocal microscope ( Nikon Instruments , UK ) , using a 20 mW Argon laser and a 40× 0 . 8 N . A . water dipping objective .\",\n \"Third instar larvae were dissected and incubated in chilled Ca2+-free HL3 solution ( Stewart et al . , 1994 ) .\",\n \"Time-lapse images were acquired on a single focal plane every 0 . 5 or 1 s for 40 loops .\",\n \"For FRAP , a defined region of interest ( 12 × 4 µm ) was photobleached at full laser power ( 488 nm ) for two iterations at a scan speed of 0 . 5 frame/s .\",\n \"Postbleaching images were acquired on a single focal plane at 15% laser power once every 5 s for 200 s .\",\n \"Experiments were completed within 20 min after dissection .\",\n \"Kymographs were generated for a hand-drawn line selection along the axon using the Multiple Kymograph plugin in Fiji ( Schindelin et al . , 2012 ) .\",\n \"Average fluorescence intensity of each axon in each frame was measured by creating a line selection along the axon .\",\n \"After subtracting background ( average intensity in a nearby non-GFP-expressing region within the segmental nerve ) , the intensity of the bleached region was normalized to the average intensity cross the axon length in the same frame .\",\n \"Then the data were further normalized by taking the prebleaching intensity as 100% and bleach intensity as 0 .\",\n \"Normalized postbleaching data were plotted and fitted to a single exponential function to calculate rate constant ( k ) and half time ( t1/2 ) : I ( t ) =A ( 1-e−kt ) , in which I ( t ) represents the fluorescence intensity at time point t , A the highest postbleaching intensity .\",\n \"For epidermal cell EM , larvae were prepared and fixed as described by O'Sullivan et al . ( 2012 ) .\",\n \"Transverse sections were cut on a Leica Ultracut UCT ultra-microtome at 70 nm , using a diamond knife , and contrasted with uranyl acetate and lead citrate ( for epidermal cells ) .\",\n \"Sections were viewed using a Tecnai G2 electron microscope operated at 120 kV , and an AMT XR60B camera running the Deben software in the Cambridge Advanced Imaging Centre , School of Biology , University of Cambridge .\",\n \"For EM of peripheral nerves , we used a ROTO protocol ( Tapia et al . , 2012; Terasaki et al . , 2013 ) to highlight membranes .\",\n \"Third instar larvae were dissected in HL3 solution and fixed in 0 . 05 M sodium cacodylate ( pH 7 . 4 ) containing 4% formaldehyde , 2% vacuum distilled glutaraldehyde , and 0 . 2% CaCl2 ) , at 4°C for 6 hr .\",\n \"Larvae were dissected as for confocal analysis , but leaving overlying organs such as gut and fat body attached , to reduce loss of peripheral nerves during processing .\",\n \"Preparations were then washed 3 times for 10 min each at 4°C using cold cacodylate buffer with 2 mM CaCl2 .\",\n \"A solution of 3% potassium ferricyanide in 0 . 3 M cacodylate buffer with 4 mM CaCl2 was mixed with an equal volume of 4% aqueous osmium tetroxide; larval preparations were incubated in this solution at 4°C for 1–12 hr , then rinsed with deionized water at room temperature 5 times for 3 min each .\",\n \"Thiocarbohydrazide solution was prepared by adding 0 . 1 g thiocarbohydrazide to 10 ml deionized water , kept in a 60°C oven in a secondary embedding pot for 1 hr , swirled every 10 min to facilitate dissolution , and filtered through two 9 cm filter papers just before use .\",\n \"Larval preparations were incubated in thiocarbohydrazide solution for 20–30 min at room temperature and covered with foil to protect from light .\",\n \"Then they were rinsed with deionized water at room temperature 5 times for 3 min each , incubated in 2% osmium tetroxide for 30–60 min at room temperature , and rinsed with deionized water at room temperature 5 times for 3 min each .\",\n \"Preparations were incubated in 1% uranyl acetate ( maleate-buffered to pH 5 . 5 ) at 4°C overnight and rinsed with deionized water at room temperature 5 times for 3 min each .\",\n \"Then they were incubated in lead aspartate solution ( 0 . 66 g lead nitrate dissolved in 100 ml 0 . 03 M aspartic acid , pH adjusted to 5 . 5 with 1 M KOH ) at 60°C for 30 min and rinsed with deionized water at room temperature 5 times for 3 min .\",\n \"Then they were dehydrated twice with 50% , 70% , 90% and 100% ethanol , twice with dried ethanol , twice with dried acetone and twice with dry acetonitrile .\",\n \"Preparations were incubated in 50/50 acetonitrile/Quetol 651 overnight at room temperature , three times for 24 hr each in Quetol epoxy resin 651 ( Agar Scientific , Stansted , UK ) and three times for 24 hr each in Quetolepoxy resin 651 with BDMA ( dimethylbenzylamine ) .\",\n \"They were then incubated at 60°C for at least 48 hr .\",\n \"Serial 60-nm-thick transverse sections were cut in the larval abdominal region , visualized using scanning EM , and images were aligned for analysis of serial sections and reconstruction , as described by Terasaki et al . ( 2013 ) .\",\n \"To quantify axonal ER tubule diameter , non-axonal staining was removed manually , and ER tubule profiles were identified based on the local threshold in a single cross-section , and the presence of signals at the same position in adjacent sections .\",\n \"The minimum Feret diameter of each tubule was measured using ImageJ Fiji ( https://fiji . sc ) via the Analyze Particles command .\",\n \"ER numbers per axon were counted manually for all axons detected in the nerve .\",\n \"3D reconstruction was carried out using the Fiji TrakEM2 plug-in .\",\n \"To quantify gaps in the tubule network , each axon was analyzed throughout the entire stack of sections .\",\n \"To allow for occasional lightly stained or blurred sections , only complete loss of ER tubules from three or more consecutive sections in an axon was defined as a gap .\",\n \"Continuous ER tubules were identified as the presence of signals at the same position for three or more consecutive sections .\",\n \"Given the varying brightness and contrast of EM sections , faint staining that coincided with a tubule signal in adjacent sections was also considered as an ER tubule .\",\n \"Color shading and sketches drawn in Figure 8 were processed in Adobe Photoshop CS6 .\",\n \"For quantification of glial ER sheet length , individual ER sheet profiles were measured using Fiji via the line tool and Measure command .\",\n \"Wrapped axons were defined as one-axon or two-axon fascicles which were completely wrapped by glial cells and isolated from other neighboring axons .\",\n \"Statistical analyses were performed in GraphPad Prism 6 or IBM SPSS 22 .\",\n \"Data were analyzed by two-tailed Student’s t-tests or ANOVA followed by post-hoc tests ( for comparison of datapoints that were means of raw measurements and hence expected to be normally distributed ) , or by Mann-Whitney U tests ( for data that were not normally distributed ) .\",\n \"Multiple comparisons after ANOVA were performed by a Tukey HSD test when equal variances were found , or otherwise by Dunnett’s T3 test .\",\n \"Bar graphs and scatter plots show mean ±SEM; box plots show median with interquartile range , and the 5% and 95% percentiles as whiskers .\",\n \"Sample sizes are reported in figures .\",\n \"No outliers were excluded from analysis after quantification; data were only excluded from analysis if they were unanalyzable , i . e . when confocal images were too faint or saturated , and EM images lacked clearly identifiable plasma or ER membranes .\",\n \"Most data were analyzed and presented as individual axon or larval datapoints , unpaired between genotypes .\",\n \"For two comparisons where ANOVA showed significant ( p<0 . 05 ) experiment-to-experiment differences within a genotype ( Figures 2B and 3A ) , data were averaged within each independent experiment to produce experiment-wise datapoints , which were compared using two-tailed paired t-tests .\",\n \"P levels are indicated as ns p>0 . 05 , *p<0 . 05 , **p<0 . 01 , ***p<0 . 001 , or ****p<0 . 0001 , except where indicated otherwise , when criteria could be defined more stringently .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Axons contain a smooth tubular endoplasmic reticulum ( ER ) network that is thought to be continuous with ER throughout the neuron; the mechanisms that form this axonal network are unknown .","Mutations affecting reticulon or REEP proteins , with intramembrane hairpin domains that model ER membranes , cause an axon degenerative disease , hereditary spastic paraplegia ( HSP ) .","We show that Drosophila axons have a dynamic axonal ER network , which these proteins help to model .","Loss of HSP hairpin proteins causes ER sheet expansion , partial loss of ER from distal motor axons , and occasional discontinuities in axonal ER .","Ultrastructural analysis reveals an extensive ER network in axons , which shows larger and fewer tubules in larvae that lack reticulon and REEP proteins , consistent with loss of membrane curvature .","Therefore HSP hairpin-containing proteins are required for shaping and continuity of axonal ER , thus suggesting roles for ER modeling in axon maintenance and function ."],"string":"[\n \"Axons contain a smooth tubular endoplasmic reticulum ( ER ) network that is thought to be continuous with ER throughout the neuron; the mechanisms that form this axonal network are unknown .\",\n \"Mutations affecting reticulon or REEP proteins , with intramembrane hairpin domains that model ER membranes , cause an axon degenerative disease , hereditary spastic paraplegia ( HSP ) .\",\n \"We show that Drosophila axons have a dynamic axonal ER network , which these proteins help to model .\",\n \"Loss of HSP hairpin proteins causes ER sheet expansion , partial loss of ER from distal motor axons , and occasional discontinuities in axonal ER .\",\n \"Ultrastructural analysis reveals an extensive ER network in axons , which shows larger and fewer tubules in larvae that lack reticulon and REEP proteins , consistent with loss of membrane curvature .\",\n \"Therefore HSP hairpin-containing proteins are required for shaping and continuity of axonal ER , thus suggesting roles for ER modeling in axon maintenance and function .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["The way we move – from simple motions like reaching out to grab something , to playing the piano or dancing – is coordinated in our brain .","These processes involve many regions and steps , in which nerve cells transport signals along projections known as axons .","Axons rely on sophisticated ‘engineering’ to work properly over long distances and are vulnerable to diseases that disrupt their engineering .","For example , in genetic diseases called ‘hereditary spastic paraplegias’ , damages to the ‘distal’ end of axons – the end furthest from the nerve cell body – cause paralysis of the lower body .","Axons have several internal structures that make sure everything works properly .","One of these structures is the endoplasmic reticulum , which is a network of tubular membranes that runs lengthwise along the axon .","It is known that spastic paraplegias are sometimes caused by mutations affecting proteins that help to build and shape the endoplasmic reticulum , for example , the proteins of the reticulon and REEP families .","However , until now it was not known how the ER forms its network in the axons and if this is influenced by these proteins .","To see whether reticulons and REEPs affect the shape of the endoplasmic reticulum , Yalçιn et al . used healthy fruit fly larvae , and genetically modified ones that lacked the proteins .","The results show that in healthy flies , the tubular network runs continuously along the axons .","When either reticulon or REEP proteins were removed , the distal axons contained less endoplasmic reticulum .","In mutant fly larvae that lacked both protein families , the endoplasmic reticulum was more interrupted and contained more gaps than in normal larvae .","Using high-magnification electron microscopy confirmed these findings , and showed that the tubules of the endoplasmic reticulum in mutant axons were larger , but fewer .","A next step will be to test whether these mutations also affect how the axons work and communicate over long distances .","A better knowledge of the role of the endoplasmic reticulum in axons will help us to understand how damages to it could affect hereditary spastic paraplegias and other degenerative conditions ."],"string":"[\n \"The way we move – from simple motions like reaching out to grab something , to playing the piano or dancing – is coordinated in our brain .\",\n \"These processes involve many regions and steps , in which nerve cells transport signals along projections known as axons .\",\n \"Axons rely on sophisticated ‘engineering’ to work properly over long distances and are vulnerable to diseases that disrupt their engineering .\",\n \"For example , in genetic diseases called ‘hereditary spastic paraplegias’ , damages to the ‘distal’ end of axons – the end furthest from the nerve cell body – cause paralysis of the lower body .\",\n \"Axons have several internal structures that make sure everything works properly .\",\n \"One of these structures is the endoplasmic reticulum , which is a network of tubular membranes that runs lengthwise along the axon .\",\n \"It is known that spastic paraplegias are sometimes caused by mutations affecting proteins that help to build and shape the endoplasmic reticulum , for example , the proteins of the reticulon and REEP families .\",\n \"However , until now it was not known how the ER forms its network in the axons and if this is influenced by these proteins .\",\n \"To see whether reticulons and REEPs affect the shape of the endoplasmic reticulum , Yalçιn et al . used healthy fruit fly larvae , and genetically modified ones that lacked the proteins .\",\n \"The results show that in healthy flies , the tubular network runs continuously along the axons .\",\n \"When either reticulon or REEP proteins were removed , the distal axons contained less endoplasmic reticulum .\",\n \"In mutant fly larvae that lacked both protein families , the endoplasmic reticulum was more interrupted and contained more gaps than in normal larvae .\",\n \"Using high-magnification electron microscopy confirmed these findings , and showed that the tubules of the endoplasmic reticulum in mutant axons were larger , but fewer .\",\n \"A next step will be to test whether these mutations also affect how the axons work and communicate over long distances .\",\n \"A better knowledge of the role of the endoplasmic reticulum in axons will help us to understand how damages to it could affect hereditary spastic paraplegias and other degenerative conditions .\"\n]","isConversational":false},"year":{"kind":"string","value":"2017"}}},{"rowIdx":90,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["developmental biology","neuroscience"],"string":"[\n \"developmental biology\",\n \"neuroscience\"\n]","isConversational":false},"title":{"kind":"string","value":"Large-scale phenotypic drug screen identifies neuroprotectants in zebrafish and mouse models of retinitis pigmentosa"},"id":{"kind":"string","value":"elife-57245-v4"},"sections":{"kind":"list like","value":[["Inherited retinal diseases ( IRDs ) encompass a group of genetically linked retinopathies characterized by progressive photoreceptor death ( Duncan et al . , 2018 ) .","IRDs lead to irreversible vision loss , for which treatment strategies are limited .","Retinitis pigmentosa ( RP ) , the most common IRD with approximately 1 . 5–2 . 5 million RP patients a worldwide ( Dias et al . , 2018; Hartong et al . , 2006; Verbakel et al . , 2018 ) , is characterized by early onset night blindness , gradual loss of visual field , and eventual loss of central vision ( Ferrari et al . , 2011; Hamel , 2006 ) .","The initial pathological feature of RP is selective rod photoreceptor cell death causing night blindness , which is then followed by loss of cone photoreceptors and eventual full blindness ( Léveillard et al . , 2014 ) .","Mutations in more than 70 genes have been linked to RP ( Dias et al . , 2018; https://sph . uth . edu/retnet/ ) .","How these mutations affect gene function or initiate aberrant photoreceptor cell loss is largely unknown .","Given the genetic diversity , pan-disease therapeutics are highly desirable .","Accordingly , the purpose of our study was to identify compounds capable of promoting rod cell survival in multiple RP models and across species .","As RP progression is relatively protracted , and even small numbers of surviving rod photoreceptors can preserve cone photoreceptor function ( Guadagni et al . , 2015; Hartong et al . , 2006; Punzo et al . , 2012 ) , pharmacological interventions aimed at slowing rod photoreceptor death are sought ( Duncan et al . , 2018; Wubben et al . , 2019 ) .","However , currently there are no effective therapies for promoting rod photoreceptor survival .","As a means of discovering new pharmacological treatments , target-directed high-throughput screening ( HTS ) approaches have been highly successful in identifying compounds that bind to and/or modulate disease-implicated molecules .","However , many promising leads fail during late-stage animal model testing or clinical trials ( Munos , 2009; Sams-Dodd , 2013; Scannell et al . , 2012 ) .","This trend has renewed interest in phenotypic drug discovery ( PDD ) , a complementary approach where drug effects are evaluated in cells or living disease models ( Bickle , 2010; Lee et al . , 2012; Swinney , 2013 ) .","A number of first-in-class drugs were recently discovered using PDD ( Eder et al . , 2014; Swinney and Anthony , 2011; Swinney , 2013 ) .","To expand opportunity on this front , we developed a PDD platform enabling quantitative HTS ( qHTS; Inglese et al . , 2006 ) in zebrafish ( Walker et al . , 2012; Wang et al . , 2015a; White et al . , 2016 ) .","Zebrafish offer several distinct advantages as a retinal disease modeling system ( Angueyra and Kindt , 2018; Richardson et al . , 2017 ) .","First , the structure of the zebrafish retina is similar to the other vertebrates ( Angueyra and Kindt , 2018; Richardson et al . , 2017; Schmitt and Dowling , 1999 ) .","In particular , the zebrafish retina is ‘cone rich’ like the human retina .","Second , about 70% of human genes have at least one ortholog in zebrafish ( Howe et al . , 2013 ) .","Moreover , all RP-associated genes listed in RetNet ( https://sph . uth . edu/retnet/ ) have conserved zebrafish orthologs .","Third , the zebrafish retinal system develops quickly being fairly mature by day five of development ( Brockerhoff et al . , 1995; Moyano et al . , 2013; Schmitt and Dowling , 1999 ) .","Fourth , zebrafish are amendable to large-scale chemical screening due to their high fecundity rate , small size , and ease of visualizing and quantifying a variety of phenotypes ( Mathias et al . , 2012; Zon and Peterson , 2005 ) .","To streamline such screens , we developed a high-throughput plate reader-based method for quantifying reporter gene expression in vivo ( ‘ARQiv’; Walker et al . , 2012 ) .","Recently , we adapted the ARQiv system to human stem-cell-derived retinal organoids ( Vergara et al . , 2017 ) to enable cross-species PDD .","To realize full throughput potential , ARQiv was combined with robotics-based automation to create ‘ARQiv-HTS’ ( Wang et al . , 2015b; White et al . , 2016 ) .","Here , to identify neuroprotective compounds promoting rod photoreceptor survival , ARQiv-HTS was used to perform a large-scale chemical screen in a transgenic zebrafish model enabling inducible rod photoreceptor cell death ( Walker et al . , 2012; White et al . , 2017 ) .","In these fish , YFP-NTR ( a yellow fluorescent protein-nitroreductase bacterial enzyme fusion protein ) is selectively expressed in rod photoreceptors .","NTR expression enables inducible ablation of rod cells upon exposure to prodrug substrates , such as metronidazole ( Mtz ) , which are thought to cause apoptosis through activated metabolites that cause DNA damage ( Curado et al . , 2007; White and Mumm , 2013 ) .","Close to 3000 largely human-approved drugs were tested across six concentrations ( i . e . using qHTS principles , Inglese et al . , 2006 ) in more than 350 , 000 zebrafish larvae .","Statistically , 113 hits were identified as hits and 42 of the top performing compounds advanced through confirmation and orthogonal assays .","Eleven compounds passed all secondary tests and moved forward as lead drug candidates .","Validation tests in an orthogonal series of mouse RP models followed , reasoning that drugs effective across species and/or independent of disease mutation may translate better clinically .","All leads were screened in primary mouse photoreceptor cell cultures and a subset of leads was evaluated in retinal explants from the retinal degeneration 1 ( Pde6brd1 , hereafter rd1 ) mutant mouse model of RP .","Finally , one of three top-performing leads was tested in retinal degeneration 10 mutant retinas in vivo ( Pde6brd10 , hereafter rd10 ) .","An analysis of potential mechanisms of action ( MOA ) suggested both shared and complementary targets/pathways .","The role of Poly ( ADP-ribose ) polymerase ( PARP ) , the most common shared target , was evaluated using chemical inhibition , genetic targeting , and western blot analysis .","The results implicated PARP as a key mediator of NTR/Mtz-mediated rod cell ablation , and PARP1 inhibition as a shared MOA for a subset of leads .","As PARP-dependent cell death pathways are implicated in the etiology of Parkinson’s disease ( Kam et al . , 2018 ) , RP ( Power et al . , 2020 ) , and other neurodegenerative disorders ( Fan et al . , 2017; Fatokun et al . , 2014 ) , NTR/Mtz-mediated cell ablation may serve as an inducible and titratable methodology for modeling neurodegenerative disease .","To interrogate complementary MOA , lead compounds were tested in pairs in mouse photoreceptor cell culture assays and in zebrafish .","Intriguingly , enhanced survival effects were common in mouse photoreceptor cell cultures , while additive and even synergistic effects were evident in zebrafish for the majority of pairs tested .","Taken together , our results suggest drugs providing neuroprotective effects across diverse RP models , between species , and/or through complementary MOA , will provide promising new therapeutic opportunities for IRD/RP patients ."],["We initiated our study using a robotics-automated large-scale in vivo screening system to identify compounds that promoted rod photoreceptor survival in a transgenic zebrafish line enabling targeted ablation of rod photoreceptors , Tg ( rho:YFP-Eco . NfsB ) gmc500 , hereafter , rho:YFP-NTR ( Walker et al . , 2012; White et al . , 2017 ) .","In this line , a 3 . 7 kb rhodopsin ( rho ) promoter fragment ( Hamaoka et al . , 2002 ) drives transgene expression exclusively in rod photoreceptor cells ( Figure 1A ) .","The transgene is a fusion protein linking a yellow fluorescent protein ( YFP ) reporter to a nitroreductase prodrug converting enzyme ( NTR , encoded by the nfsB gene from Escherichia coli ) .","NTR expression enables pro-drug inducible targeted cell ablation ( Curado et al . , 2007; Pisharath et al . , 2007 ) .","Exposing rho:YFP-NTR fish to the prodrug metronidazole ( Mtz ) leads to the selective death of rod photoreceptors and concomitant loss of YFP ( Figure 1A–C ) , physiologically mimicking the onset of RP ( Hamel , 2006 ) .","An immunohistological analysis of rod and cone photoreceptor markers was performed on 7 days post-fertilization ( dpf ) zebrafish retinal sections to test if Mtz-induced ablation was specific to rod cells .","In non-ablated controls , rod outer segment labeling was well correlated with YFP expression ( Figure 1—figure supplement 1A , B; -Mtz , arrows ) .","In Mtz-treated retinas , rod outer segment labeling and YFP expression were markedly diminished ( Figure 1—figure supplement 1A , B; +Mtz ) .","Cone photoreceptor labeling showed no overlap with YFP expression , and cone cell labeling appeared similar in non-ablated and Mtz-treated retinas ( Figure 1—figure supplement 1C ) , suggesting cone cells were not affected by Mtz exposure or by acute rod cell loss .","Prior studies suggest NTR/Mtz-mediated ablation occurs by apoptosis ( Chang et al . , 1993; Doonan et al . , 2003; Portera-Cailliau et al . , 1994; Zeiss et al . , 2004 ) and loss of rods in RP has also been linked to apoptosis ( Chang et al . , 1993; Doonan et al . , 2003; Portera-Cailliau et al . , 1994; Zeiss et al . , 2004 ) .","Thus , we reasoned rho:YFP-NTR fish could be used to identify compounds that protected rods from apoptosis .","To identify neuroprotective compounds with this model ( i . e . drugs that sustain YFP expression after Mtz exposure ) , we used our established plate reader-based ARQiv-HTS assay ( Figure 1D; Walker et al . , 2012; White et al . , 2016 ) .","We first determined optimal conditions for inducing rod cell loss while maintaining larval health in a 96-well plate format .","Major aspects of retinal cytogenesis are largely complete by five dpf in zebrafish ( Schmitt and Dowling , 1999; Stenkamp , 2011 ) .","Reporter expression in rho:YFP-NTR larvae has also stabilized by this time point ( Unal Eroglu et al . , 2018 ) , consistent with rho expression ( Raymond et al . , 1995 ) .","We therefore chose 5 dpf to initiate Mtz-induced rod cell ablation .","We previously determined that rod cell loss reached a nadir at 7 dpf following a 24 hr pulse of 10 mM Mtz at 5 dpf ( Walker et al . , 2012 ) .","We reasoned that a 48 hr Mtz exposure initiating at 5 dpf would maximize the signal window to test for neuroprotective effects .","Concluding the experiment by 7 dpf also avoids challenges associated with feeding , as zebrafish can subsist on their yolk sac up to that time point ( Hernandez et al . , 2018; Jardine and Litvak , 2003 ) .","However , 10 mM Mtz treatments extending beyond 24 hr become increasingly toxic ( Mathias et al . , 2014 ) and removing Mtz from microtiter plates after a 24 hr pulse could not be easily automated .","We therefore sought a 48 hr Mtz treatment regimen sufficient for inducing maximal rod cell loss by 7 dpf that showed no evidence of general toxicity .","Five concentrations of Mtz were tested across a twofold dilution series from 10 mM to 625 μM .","YFP reporter signals were quantified daily from 5 to 8 dpf using ARQiv .","Changes in YFP levels were calculated as percentages normalized to non-ablated YFP controls .","The data showed concentration-dependent reductions in YFP with maximal loss observed at 7 dpf ( Figure 1B ) .","Mtz exposures at 0 . 625 , 1 . 25 , 2 . 5 , 5 , and 10 mM led to a 55 , 79 , 87 , 87 , and 83% decrease in YFP detection , respectively ( Supplementary file 1a ) .","Although no lethality was observed for any condition , signs of distress were evident for 10 mM Mtz exposures ( i . e . reduced motility ) .","At ≤5 mM Mtz , however , no signs of stress were observed .","As 2 . 5 mM Mtz was the lowest concentration producing maximal loss of YFP detection , and confocal imaging verified YFP loss following a 2-day ( 5–7 dpf ) 2 . 5 mM Mtz treatment ( Figure 1C ) , this condition was selected as the treatment regimen for the large-scale screen .","Previously established power analysis methods using ablated and non-ablated controls ( White et al . , 2016 ) determined that a sample size of nine larvae was sufficient to detect a 50% neuroprotective effect .","For ease of dispensing , microtiter plate formatting , and to account for larval dispensing errors , we increased the sample size to 16 larvae per condition for the primary screen .","We also reasoned that this would also allow us to detect subtler neuroprotective effects .","The strictly standardized mean difference quality control ( SSMD QC ) score was 1 . 67 , indicating the assay was of sufficient quality to justify a large-scale screening effort ( Zhang , 2011 ) .","To establish a positive control , we tested 17 compounds and one compound ‘cocktail’ previously implicated as neuroprotectants in mammalian RP models ( Supplementary file 2a ) .","Unfortunately , none sustained YFP expression at the concentrations tested ( 4 μM to 125 nM ) .","However , none have proven effective in RP patients either , suggesting a lack of conservation of neuroprotective effects across species .","Our screen was designed to address this issue by testing for neuroprotective effects of these compounds across species and between different RP models .","Fortunately , a compound identified as a retinal cell neuroprotectant by the Zack lab ( manuscript in preparation ) did show-dose-dependent effects on YFP levels .","This compound was therefore used as a positive control ( POS ) for assay performance throughout the large-scale screen .","The Johns Hopkins Drug Library ( JHDL ) was chosen for the large-scale screen .","The JHDL is comprised of nearly 3000 compounds , most being human-approved drugs ( Shim and Liu , 2014 ) .","To minimize false discovery rates , all compounds were tested using qHTS principles ( Inglese et al . , 2006 ) – that is , across six concentrations ( 4 μM - 125 nM ) using a twofold dilution series .","The screen largely followed published ARQiv-HTS methodologies ( Wang et al . , 2015b; White et al . , 2016 ) , with additional assay-specific details ( Figure 1D , steps 1–8 ) .","In all , 2934 compounds were screened and more than 350 , 000 transgenic zebrafish larvae evaluated .","Real-time data analysis was performed as previously detailed ( White et al . , 2016 ) to generate: ( 1 ) a plot of YFP signal levels , ( 2 ) a plot of SSMD scores across all tested concentrations , ( 3 ) a signal intensity heat map of each plate , and ( 4 ) an SSMD score table ( Figure 1D , step 7 ) .","Compounds producing SSMD scores of ≥one were considered potential hits and flagged for visual inspection to assess fluorescence and general morphology using a stereo fluorescence microscope .","This step facilitated elimination of false-positive compounds producing aberrant fluorescence due to larval toxicity ( six drugs ) or compound autofluorescence ( 27 drugs; Figure 1C , step 8; Supplementary file 2b ) .","Additionally , this allowed visual confirmation of sustained YFP expression within the retina .","At the conclusion of the primary screen , 113 compounds were identified as hits ( Supplementary file 2c ) .","Hit compounds were classified according to the highest SSMD score achieved across all concentrations tested , and whether concentration-dependent effects were observed .","SSMD scores suggested one drug produced a strong effect ( SSMD of 2–3 ) ; three had semi-strong effects ( 1 . 645–2 ) , 21 showed moderate effects ( 1 . 28–1 . 645 ) , and 88 had fairly moderate effects ( 1–1 . 28 ) ( Supplementary file 2c ) .","Forty-two drugs showed concentration-dependent effects , while 71 exhibited discontinuous or singular concentration effects .","‘On label’ mechanism of action ( MOA ) information available for the 113 hits implicated more than 50 targets/pathways in rod cell neuroprotection ( Supplementary file 2d ) .","We next performed a series of confirmatory and orthogonal assays to evaluate a subset of 42 hit compounds prioritized by SSMD score , dose-response profile , and/or implicated MOA ( Supplementary file 2c , highlighted compounds ) .","Having extensive MOA data is a key advantage of testing human-approved compounds which we leveraged in a previous large-scale zebrafish PDD screen ( Wang et al . , 2015b ) .","Similarly here , as studies have suggested inflammation plays a key role in retinal degeneration and regeneration ( Hollyfield et al . , 2008; Mitchell et al . , 2018; White et al . , 2017; Yoshida et al . , 2013 ) , hits implicated as modulators of inflammatory signaling were included as part of the prioritization scheme .","In addition , several compounds that did not produce concentration-dependent effects were selected to test whether this criterion was useful in predicting reproducibility .","All compounds were obtained from new sources to ensure reagent authenticity .","To confirm survival promoting effects , three repeats of the primary screening assay ( Figure 2A ) were conducted , but using a wider concentration range to account for differences in reagent quality ( from 100 μM to 1 . 28 nM , fivefold dilution series ) .","If toxicity was observed at higher concentrations , dilution series were initiated at 10 µM or 1 µM .","Using this strategy , 11 of the 42 prioritized hit compounds were confirmed as lead candidates ( 26% ) .","Estimated effects on rod cell survival – that is , YFP signal levels relative to +Mtz/0 . 1% DMSO controls – ranged from 38 to 9% ( Figure 2B; lead drug candidate names , abbreviations , PubChem CID , and chemical structures are provided in Table 1 ) .","We next asked whether there was a correlation between SSMD scores and/or concentration-dependent effects and confirmation rates .","Among 19 selected hit compounds with higher SSMD scores ( ≥1 . 3 ) , seven ( 37% ) were confirmed; among 23 with lower SSMD scores ( 1–1 . 28 ) , four ( 17% ) were confirmed .","Of 27 selected hit compounds with a concentration-dependent trend , eight ( 30% ) were confirmed .","Conversely , of 15 compounds that did not show a concentration-dependent trend , 3 ( 20% ) were confirmed .","Among the 11 confirmed leads , 8 ( 73% ) showed dose-dependent effects and 7 ( 64% ) had higher SSMD scores ( >1 . 28 ) .","These results suggest that prioritizing hit compounds by both relative SSMD scores and dose-dependent trends provides predictive value for confirming activity , consistent with qHTS principles ( Inglese et al . , 2006 ) .","As rod cell death is induced by NTR-mediated reduction of the prodrug Mtz in our model , it is possible that some lead compounds simply suppressed NTR enzymatic activity .","To test this , NTR activity was evaluated in the presence of each lead compound by assaying the reduction kinetics of the prodrug CB1954 in vitro ( Prosser et al . , 2010 ) .","To ensure any potential for NTR inhibition was accounted for , all compounds were tested at 300 μM ( except MIC and CHL which precipitated at higher concentrations and were tested at 50 µM ) that is , ~100-fold greater than neuroprotective concentrations .","Compounds were deemed potential inhibitors if NTR activity was less than 75% of the NTR alone control .","Seven compounds showed no evidence of NTR inhibition by this criterion , but four did: warfarin ( WAR ) , ciclopirox olamine ( CPO ) , calcimycin ( CAL ) , and sulindac ( SUL; Figure 2—figure supplement 1A , B ) .","However , IC50 measures ranged from 150 µM ( for CPO ) to 350 µM ( for SUL , Figure 2—figure supplement 1B ) , approximately one to two orders of magnitude higher than observed neuroprotective concentrations ( i . e . 0 . 4–20 µM ) .","When NTR reduction of Mtz was tested for WAR , CPO , and CAL at 300 μM , similarly weak inhibitory effects were observed ( Figure 2—figure supplement 1B ) .","The differences in concentrations between neuroprotective and NTR inhibitory activities diminish the likelihood that these leads act directly on NTR .","To test this further , lead compounds were assayed for neuroprotective effects in NTR-independent mouse RP models ( see below ) .","To control for the possibility that lead candidates promoted rod photoreceptor development , rather than neuroprotection , YFP levels were quantified in rho:YFP-NTR larvae exposed solely to lead drugs from 5 to 7 dpf ( Figure 2—figure supplement 2A ) .","Retinoic acid ( RA , 1 . 25 μM ) was used as a positive control as it promotes rod fates during development in zebrafish ( Hyatt et al . , 1996 ) .","RA-treated fish displayed statistically significant increases in YFP levels compared to untreated controls ( Figure 2—figure supplement 2B ) .","In contrast , none of the retinas treated with lead compounds exhibited increased YFP expression , suggesting they do not promote rod photoreceptor cell fate .","In fact , three lead compounds cloxyquin ( CLO ) , cortexolone ( COR ) and CPO reduced YFP signals relative to the untreated control ( Figure 2—figure supplement 2B , asterisks ) , suggesting negative effects on rod cell development .","It is well known that the zebrafish retina regenerates ( Gorsuch and Hyde , 2014; Lenkowski and Raymond , 2014; Wan and Goldman , 2016 ) .","Therefore , to determine whether lead compounds acted by stimulating regeneration , we used a previously described ARQiv assay designed to detect changes in rod cell replacement kinetics ( Walker et al . , 2012; White et al . , 2017 ) .","Briefly , rho:YFP-NTR larvae were first treated with 10 mM Mtz at 5 dpf for 24 hr to induce rod cell loss .","At 6 dpf , Mtz was washed out and larvae were treated with lead compounds at concentrations corresponding to maximal neuroprotective effects and YFP levels quantified at 9 dpf ( Figure 2—figure supplement 3A ) .","Dexamethasone , which accelerates rod cell regeneration kinetics ( White et al . , 2017 ) , was used as a positive control .","The results showed that none of the compounds increased rod cell regeneration rates ( Figure 2—figure supplement 3B ) , while four compounds , dihydroartemisinin ( DHA ) , CLO , CPO and MIC inhibited regeneration .","These data suggest that lead compounds do not increase YFP levels by promoting rod cell regeneration .","To test if lead compounds simply increased YFP signal intensity , rather than promoted rod cell survival , intravital microscopy was used to image Mtz-treated retinas ±lead compounds and non-ablated ( -Mtz ) controls .","Mtz-treated retinas had reduced numbers of YFP-NTR-expressing rods ( Figure 3 , +Mtz ) .","Conversely , rods in control retinas displayed robust YFP signal throughout the retina and elongated morphologies suggestive of healthy outer segments ( Figure 3 , -Mtz ) .","In retinas exposed to Mtz and lead compounds , YFP signal loss was attenuated and rods typically displayed elongated morphologies ( Figure 3 ) .","However , for some compounds , cells appeared rounded , suggesting degeneration was not fully inhibited ( e . g . miconazole , MIC; Figure 3 ) .","To confirm lead compounds increased rod cell survival , confocal stacks of YFP-expressing rods were 3D-rendered and fluorescence volume quantified using Imaris software-based automated image analysis ( White et al . , 2017 ) .","The data showed a statistically significant increase in YFP volumes for all drug-treated groups ( Figure 3—figure supplement 1 ) , confirming that lead compounds promoted increased rod cell numbers and/or preserved outer segment morphology .","To further investigate if lead compounds preserved rod outer segments , we used a newly developed zebrafish transgenic line , Tg ( rho:GAP-YFP-2A-nfsB_Vv F70A/F108Y ) jh405 ( hereafter rho:YFP-NTR2 . 0 ) .","In this line , rods co-express membrane-tagged YFP and an improved NTR ( ‘NTR 2 . 0’ ) .","The membrane-tagged YFP facilitates improved imaging of rod outer segments , while NTR 2 . 0 enables cell ablation at a reduced concentration of Mtz ( e . g . 10 µM versus 2 . 5 mM; Sharrock et al . , 2020 ) .","To assess if lead compounds preserved outer segment morphologies , intravital confocal imaging was performed as above but using the rho:YFP-NTR2 . 0 line and 10 µM Mtz treatments .","Qualitative image analysis suggested the majority , six of seven tested lead compounds , preserved rod outer segments ( Figure 3—figure supplement 2A ) .","Reasoning that preservation of outer segments would equate to increased rod cell size , relative to rounded morphologies of dead or stressed cells , average rod cell sizes were calculated as the total YFP volume divided by the number of rod cells per each 3D confocal stack ( i . e . average YFP volume per cell ) .","Compared to rod ablated controls , five of seven lead compounds tested promoted statistically significant increases in rod cell volumes ( Figure 3—figure supplement 2B ) .","These data suggest some lead compounds were able to maintain rod outer segment morphology , thus potentially maintaining rod cell function .","To identify new therapeutics for RP patients , we reasoned compounds producing neuroprotective effects across fish and mammalian RP models would likely target conserved mechanisms , and thus be more likely to translate successfully in the clinic .","We therefore tested the efficacy of lead compounds in a series of mouse models of retinal photoreceptor degeneration .","First , we tested compounds for the capacity to protect mouse primary photoreceptor cells from stressor-induced cell death in culture .","Photoreceptors were isolated from postnatal day four ( P4 ) QRX mice , a transgenic line in which GFP expression is restricted to photoreceptors ( Wang et al . , 2004 ) , and grown as previously described ( Fuller et al . , 2014 ) .","To induce photoreceptor cell death , thapsigargin ( 0 . 25 µM ) or tunicamycin ( 0 . 6 µg/mL ) were added .","These ‘stressor’ compounds deplete endoplasmic reticulum ( ER ) calcium levels ( Thastrup et al . , 1990 ) and inhibit protein glycosylation ( Fliesler et al . , 1984 ) , respectively ( Lai et al . , 2007b ) .","In turn , they elicit an unfolded protein response ( UPR ) ( Wang et al . , 2015b ) and related ER stress ( Oslowski and Urano , 2011; Zhang et al . , 2014 ) , both implicated in the etiology of RP ( Griciuc et al . , 2011; Rana et al . , 2014 ) .","To test for survival effects , all eleven lead compounds were screened at seven concentrations across a threefold dilution series ( from 30 µM to 40 nM ) .","After 48 hr in stressor ± lead compounds , cells were imaged using an automated high-content screening system .","Photoreceptor survival was assessed by automated quantification of GFP-expressing cells ( Figure 4A ) .","Nine of 11 lead compounds , proved protective in the thapsigargin-induced cell death assay ( Figure 4B ) .","MIC , CLO and SUL also protected photoreceptors from tunicamycin-induced cell death ( Figure 4—figure supplement 1 ) .","Thus , nine of 11 lead compounds promoted the survival of mouse primary photoreceptor cells in at least one stressor-induced cell death assay , and three were neuroprotective in both assays .","We next tested lead compounds for survival effects in rd1 mouse retinal explant cultures .","The rd1 mouse model of RP exhibits early onset rod cell degeneration caused by a mutation in the Pde6b gene ( Danciger et al . , 1990; Pittler and Baehr , 1991; Sidman and Green , 1965 ) , which is an ortholog of the human RP-associated PDE6B ( Khramtsov et al . , 1993; McLaughlin et al . , 1993 ) .","In rd1 mice , photoreceptor degeneration begins around P10 and progresses rapidly .","By P21 , only a single row of photoreceptor cells remain in the outer nuclear layer ( ONL; Tansley , 1951 ) making it an excellent system for screening potential neuroprotectants ( Beeson et al . , 2016 ) .","Here , retinal explants from P10 rd1 mice were isolated and cultured ex vivo ( Bandyopadhyay and Rohrer , 2010 ) .","Eight lead compounds , all but CAL , SUL , and Zinc pyrithione ( ZPT ) , were tested for survival effects at three concentrations across a fivefold dilution series centered on the concentration most effective in fish RP models .","After 11 days in culture , explants were fixed , stained and the number of photoreceptor rows counted ( Figure 5A ) .","Neuroprotective effects were defined as a concentration-dependent increase in the number of photoreceptor rows remaining in the ONL relative to untreated controls ( p≤0 . 05 ) .","An average of 1 . 2 ±0 . 19 rows of photoreceptors remained in the ONL of control explants .","Three lead drugs , CPO , DHA , and ART , increased the number of surviving photoreceptor layers ( Figure 5B ) .","Representative images of control ( DMSO and POS ) and lead-treated explants demonstrate photoreceptor layer protection ( Figure 5C ) .","CPO , DHA , and ART thus promoted rod cell survival in the fish RP model and two mouse cell culture RP models .","However , high concentration CPO treatments ( 15 µM ) led to disruption of retinal histology due to induction of proliferation in the inner nuclear layer ( INL ) and ONL .","Therefore , as DHA is the active metabolite of ART-related compounds , we proceeded with testing DHA in an in vivo mouse RP model .","To test the potential of DHA as a ‘pan-disease’ therapeutic , that is , whether it was effective across multiple genetic RP models , it was tested for photoreceptor survival effects in the rd10 mouse model of RP ( Pde6brd10 ) .","The rd10 line was selected for in vivo experiments due to its slower rate of photoreceptor degeneration relative to other rd mutants , thus allowing a prolonged window for pharmacological intervention .","DHA was selected based on its superior performance in rd1 retinal explant assays ( Figure 5 ) and because it is the active metabolite of a second lead drug candidate , ART , recently shown to protect photoreceptors from light damage ( Lu and Xie , 2019 ) .","To create a long-release formulation , DHA was encapsulated in PLGA polymers ( Poly ( D , L-lactide-co-glycolide ) ) .","Release kinetics assays in PBS/0 . 1% DMSO in vitro suggested DHA would reach maximal concentrations after 30 days , and remain stable for at least 20 days thereafter ( Figure 6A ) .","We noted that release was minimal in the absence of DMSO , but were reluctant to include DMSO for injections due to the potential for deleterious effects on the retina ( Tsai et al . , 2009 ) .","PLGA-DHA was injected into the vitreous of one eye of rd10 mice at P14 , with the contralateral eye serving as a vehicle injection control .","At P32 , 18 days after injection , and when DHA levels were predicted to reach 80% of maximal concentration based on in vitro releases kinetics , rd10 mouse eyes were processed for immunohistochemistry ( Figure 6B ) and ONL thickness quantified ( Figure 6C ) .","Despite initial promising results in pilot assays , no reproducible neuroprotective effects were observed .","It is possible that in the absence of DMSO release kinetics may have been limiting in vivo .","We plan to address this possibility in future studies .","Previously , we used ‘on label’ information to explore MOA of hit compounds identified during an in vivo phenotypic screen of the JHDL ( Wang et al . , 2015b ) .","Recently , we have become interested in additional advantages afforded by HTS-based MOA data and whole-organism phenotypic screening , such as polypharmacology ( Dar et al . , 2012; Rennekamp and Peterson , 2015; Rihel et al . , 2010 ) .","Accordingly , we applied a target-agnostic MOA analysis process by evaluating lead compound performance in HTS and ultra-HTS studies archived on PubChem ( https://pubchem . ncbi . nlm . nih . gov/ ) .","Many lead compounds exhibited shared target/pathway activities , suggesting in-common MOA ( Table 2 ) .","Among shared targets , Tyrosyl-DNA Phosphodiesterase 1 ( TDP1 ) , a DNA repair enzyme , was the most popular ( eight leads ) .","To test whether TDP1 inhibition promoted rod photoreceptor survival , we tested three TDP1 inhibitors in rod:YFP-NTR fish: paromomycin ( PM ) , thiostrepton ( ThS ) , and methyl-3 , 4-dephostatin ( MD ) ( Huang et al . , 2011; Liao et al . , 2006 ) .","Both PM and ThS showed neuroprotective effects ( Table 3 , Figure 7 ) .","This result was surprising given that NTR reduction of Mtz is thought to cause DNA damage-induced cell death ( Curado et al . , 2007 ) .","Thus , inhibition of a DNA repair enzyme would be expected to enhance , not inhibit , NTR/Mtz-mediated cell death .","However , an alternative means of disrupting TDP1 activity is by inhibiting Poly ( ADP-ribose ) Polymerases ( PARPs ) .","Indeed , in a comprehensive 400 , 000 compound qHTS assay designed to identify TDP1 inhibitors , all five hits turned out to be PARP inhibitors ( Murai et al . , 2014 ) .","PARPs also mediate DNA repair but , interestingly , hyperactivation of PARP1 leads to a specific form of DNA damage-induced cell death , termed parthanatos ( Fatokun et al . , 2014; Wang et al . , 2016 ) .","PARP is also involved in a cGMP-dependent cell death pathway implicated in RP ( Power et al . , 2020 ) .","We therefore assayed PARP inhibitors for the capacity to promote rod cell survival in rod:YFP-NTR fish .","All eight PARP inhibitors tested had neuroprotective activity , ranging from 9% to 20% ( Figure 7 ) .","To account for other cell death mechanisms implicated in neurodegeneration , we tested an inhibitor of necroptosis ( necrostatin-1; NEC ) , and three inhibitors of apoptosis ( Ac-DEVD-CHO , CASI , and CASVII , Table 3 ) .","Surprisingly , NEC promoted rod cell survival , while the apoptosis inhibitors did not ( Figure 7 , Table 3 ) .","Interestingly , paired testing of PARP and necroptosis inhibitors resulted in additive effects ( Figure 7—figure supplement 1 , Supplementary file 1b ) , suggesting either differential responses to NTR/Mtz-induced DNA damage among rod cells , or that inhibiting both pathways keeps cells from activating the other mechanism of cell death when only one pathway is blocked .","To further dissect cell death pathways mediating NTR/Mtz-induced rod cell ablation , we used a ‘redundant’ CRISPR/Cas9 gene targeting approach ( i . e . four guide RNA targeting ) , which enables phenotyping in injected zebrafish embryos/larvae ( Wu et al . , 2018 ) .","Key factors for apoptosis ( casp3a and casp3b ) , PARP-dependent cell death pathways parthanatos and/or cGMP-dependent cell death ( parp1 ) , and necroptosis ( ripk1l ) were targeted .","In addition , tdp1 was targeted as a DNA repair control .","For each targeted gene , four gRNAs and Cas9 protein were co-injected into one-cell stage embryos .","Injected and uninjected ( control ) larvae were treated ±2 . 5 mM Mtz at five dpf and rod-YFP levels quantified at six dpf by ARQiv .","For Mtz-ablated larvae , knockdown of parp1 markedly increased rod cell survival ( 39% compared to 20% in the uninjected control , p=4E-12 ) , and ripk1l knockdown modestly protected rods ( 26% compared to 22% in the uninjected control , p=0 . 05 ) .","However , no improvement in survival was observed for either casp3a/b or tdp1 knockdown ( Figure 8A ) .","In no Mtz controls , knockdown of parp1 , ripk1l or tdp1 had no effect on rod cell numbers .","Conversely , casp3a/3b double knockdown increased rod cell numbers ( Figure 8B ) suggesting inhibition of developmental apoptosis and confirming efficacy of casp3a/3b knockdown .","Confocal imaging confirmed parp1 knockdown effects on rod cell survival and casp3a/3b knockdown effects on rod cell development ( Figure 8—figure supplement 1A ) .","Reduced gene expression was verified by qPCR ( Figure 8—figure supplement 1B , Supplementary file 3 ) .","These results are consistent with the NTR/Mtz system eliciting DNA damage-associated cell death mediated by parp1 , that is , parthanatos , in rod cells .","To further test for activation of PARP signaling during NTR/Mtz-induced rod cell death , accumulation of polymerized poly ( ADP-ribose ) ( PAR ) , a downstream effector of PARP1 activation ( Kam et al . , 2018 ) , was analyzed by western blot .","five dpf rho:YFP-NTR larvae were treated ±2 . 5 mM Mtz and protein samples collected from ~30 fish at 3 , 6 , 12 , 24 , and 48 hr post-Mtz treatment initiation .","Western blots showed clear accumulation of PAR in Mtz-treated fish at 24 and 48 hr post-Mtz ( Figure 8—figure supplement 1C ) which was verified quantitatively ( Figure 8—figure supplement 1D ) .","Collectively , results of cell death pathway analyses are concordant across chemical inhibition , genetic knockdown , and biochemical assays , strongly suggesting NTR/Mtz-induced cell ablation is mediated by PARP-dependent cell death ( e . g . parthanatos , cGMP-dependent cell death ) , a target/pathway implicated across multiple mammalian models of RP ( Arango-Gonzalez et al . , 2014 ) .","To determine if lead compound effects were mediated by inhibition of PARP signaling , parp1 expression was knocked down prior to testing for effects on rod cell survival in rho:YFP-NTR larvae .","Four leads predicted to be PARP inhibitors ( MIC , CLO , CPO , DHA ) , one predicted to be a non-PARP inhibitor ( WAR ) , and a known PARP inhibitor control ( BMN ) were tested .","Two lead compounds , MIC and DHA and BMN ( the PARP inhibitor control ) produced no statistically significant enhanced neuroprotective effects over parp1 knockdown .","Conversely , WAR , CLO and CPO increased rod cell survival in parp1 knockdown fish ( Figure 9A ) .","qPCR confirmed mRNA knockdown ( Figure 9B ) and controls showed no effects of parp1 knockdown on rod cell development ( Figure 9C , -Mtz ) and enhanced rod cell survival upon Mtz treatment ( Figure 9C , +Mtz ) .","These data confirm that PARP inhibition mediates the neuroprotective effects of some lead compounds ( e . g . MIC and DHA ) while others are PARP-independent .","The latter result suggests lead compounds may act via complementary neuroprotective mechanisms .","Combined , these findings are consistent with rod cell loss in our fish RP model occurring , at least partially , by PARP-dependent cell death .","PubChem data also suggested potential complementary MOA , that is , multiple independent targets across lead compounds ( Table 2 ) .","We therefore hypothesized that combining lead compounds may produce enhanced , additive , or even synergistic effects .","To investigate this , combinatorial lead drug treatments were tested in primary mouse photoreceptor cell cultures .","As above , photoreceptors were isolated from QRX mouse retinas and used in stressor-treated primary cell culture assays .","All 11 lead compounds were tested at a selected concentration individually and in pairs and the survival of GFP-labeled photoreceptors analyzed by high-content imaging and automated cell counting .","The results showed no wholly additive effects , that is , survival percentages equaling or excelling the summed effect of individual compound assays .","However , enhanced effects – i . e . , better median survival than for either compound alone – were observed for 30 of 55 pairs in thapsigargin-treated cultures and for 20 of 55 pairs in tunicamycin-treated cultures ( Figure 4—figure supplement 2 ) .","Finally , seven lead compounds were tested alone and in pairs in rho:YFP-NTR zebrafish using optimal effective concentrations .","Pairs producing survival effects equal to or greater than the sum of their individual values ( ±10% ) were considered additive .","By this criterion , 10 of 19 viable pairs ( two pairs proved lethal ) exhibited additive effects ( Figure 10 , bolded; Supplementary file 1c ) and three pairs produced supra-additive effects ( i . e . ≥25% greater than summed effects ) suggesting possible synergy ( Figure 10 , asterisks ) .","The maximal average rod cell survival effect was 58% ( WAR + CPO ) .","Several compounds showed broadly additive effects , for example , WAR was additive with all six drugs and COR with four of five pairs ( one pair proving lethal ) .","Additive effects suggest multiple signaling pathways are involved in NTR/Mtz-induced photoreceptor degeneration , thus that combinatorial drug regimens may be required to achieve maximal therapeutic benefits for RP patients ."],["Identifying effective neuroprotective therapies for RP and other IRDs stands as a critical unmet need for the field ( Duncan et al . , 2018; Wubben et al . , 2019 ) .","Although , neurotrophic factors , anti-apoptotic agents , nutritional supplements , and antioxidants have shown neuroprotective effects in animal models of RP ( Dias et al . , 2018 ) .","Unfortunately , these reagents have produced , at best , only limited benefits for patients to date and , for some , mild improvements are offset by adverse side effects associated with long-term use ( Dias et al . , 2018 ) .","For example , ciliary neurotrophic factor ( CNTF ) was shown to be effective in protecting photoreceptors in mouse ( Cayouette et al . , 1998 ) , dog ( Tao et al . , 2002 ) , and chicken ( Fuhrmann et al . , 2003 ) models of retinal degeneration .","However , CNTF failed to improve either visual acuity or field sensitivity in short- and long-term RP clinical trials ( Birch et al . , 2016; Ciliary Neurotrophic Factor Retinitis Pigmentosa Study Groups et al . , 2013; Argus II Study Group et al . , 2015 ) .","Clinical trials of Vitamin A in combination with Vitamin E ( Berson et al . , 1993 ) , docosahexaenoic acid ( Berson et al . , 2004 ) , lutein ( Berson et al . , 2010 ) , or valproic acid ( Birch et al . , 2018 ) were reported to produce some benefits for RP patients , but only in subpopulations , and some of these studies have been controversial ( Massof and Finkelstein , 1993 ) .","Our strategy for addressing this challenge was to: ( 1 ) scale up the number of compounds tested directly in complex living disease models and ( 2 ) perform a cross-species screening cascade that starts with small animal models amenable to HTS and proceeds to mammalian models .","We hypothesize that compounds producing beneficial outcomes across evolutionarily diverse species will target conserved MOA and thus stand a higher chance of successful translation .","As a generalized strategy , large-scale drug discovery screens using small animal models are showing increasing promise across multiple disease paradigms ( Cagan et al . , 2019; Cully , 2019; Kitcher et al . , 2019; MacRae and Peterson , 2015 ) .","Here , using a large-scale in vivo drug screening platform ( Figure 1 ) , we tested 2934 largely human-approved compounds for neuroprotective effects across six concentrations in >350 , 000 larval zebrafish models of RP .","The primary screen implicated 113 compounds as neuroprotectants ( Supplementary file 2c ) .","Confirmatory repeats and a series of four orthogonal assays validated 11 of 42 prioritized hit compounds in protecting zebrafish rod photoreceptors from cell death ( Figures 2 and 3 , Figure 2—figure supplements 1–3 , Figure 3—figure supplement 1 and Table 1 ) .","Importantly , investigations of lead compound MOA , led to the discovery that NTR/Mtz-mediated rod cell death appears to proceed through alternative cell death pathways ( Figure 8 ) recently linked to photoreceptor degeneration in mammalian models of RP ( Arango-Gonzalez et al . , 2014; Paquet-Durand et al . , 2007; Power et al . , 2020; Sancho-Pelluz et al . , 2008; Tolone et al . , 2019 ) .","A summary schematic of the entire screening cascade is provided in Figure 11 .","Confirmatory tests showed survival effects of lead compounds varied from 9 to 38% ( Figure 2 ) , which might suggest limited therapeutic potential .","However , even small numbers of surviving rod cells can have a significant impact on cone photoreceptor survival ( Guadagni et al . , 2015; Hartong et al . , 2006; Punzo et al . , 2012 ) .","This effect is mediated by the rod-derived cone viability factor ( RdCVF ) ( Léveillard et al . , 2004 ) which stimulates glucose metabolism to promote cone survival ( Aït-Ali et al . , 2015 ) .","Therefore , as cone cells provide high-acuity daytime vision , protecting small numbers of rod cell has significant therapeutic potential for RP patients .","To test this further , lead compounds will need to be tested for the ability to prolong cone survival and function in rod-cone degeneration models .","Visualization by in vivo confocal imaging showed variations in surviving cell morphologies across lead compounds ( Figure 3 and Figure 3—figure supplement 2 ) , suggesting .","variations in the extent of rod cell protection that could impact function .","In particular , maintenance of the outer segment , a specialized primary cilium in which phototransduction occurs , is necessary for photoreceptor function .","Volumetric quantification of surviving rod cells further suggested lead compounds differ in their capacity to maintain outer segments .","Visual function tests will need to be performed to test this possibility and thus to prioritize lead compounds for in vivo testing in mammalian RP models .","To test conservation of neuroprotective effects across species , lead compounds were evaluated in three mouse IRD/RP models .","Nine of 11 leads were confirmed as neuroprotectants in primary photoreceptor cell cultures ( Figure 4 ) .","Three of eight leads assayed using rd1 retinal explant cultures were also validated ( Figure 5 ) ; two being active in both paradigms , DHA and ART .","We chose the rd10 RP model for in vivo testing because it undergoes a slower rate of rod cell loss than rd1; spanning from approximately P16 to P35 ( Chang et al . , 2007; Gargini et al . , 2007 ) .","DHA was the most promising compound for these tests as it had shown strong effects across assays and was amenable to a long-term release formulation designed to sustain drug action over weeks to months ( PLGA-DHA , Figure 6A ) .","In addition , DHA is the active metabolite of artemisinin , another of our cross-species confirmed leads that was shown to have neuroprotective activity in rat models of stress-induced neuronal damage and light-induced photoreceptor degeneration ( Yan et al . , 2017 ) .","Unfortunately , we did not observe increased photoreceptor survival in rd10 retinas injected with PLGA encapsulated DHA .","A potential issue that needs to be resolved is the release kinetics of encapsulated DHA in vivo .","DHA has poor water solubility , therefore in vitro release kinetics were obtained in the presence of 0 . 1% DMSO .","However , because intravitreal injection of DMSO can have deleterious effects ( Tsai et al . , 2009 ) we avoided this potential complication .","To assess DHA release in vivo , pharmacokinetic assays such as high-performance liquid chromatography ( Ayalasomayajula and Kompella , 2004; Shen et al . , 2014 ) will be used to quantify lead compound concentrations in the retina for future tests .","Another possible explanation is that rd1 and rd10 models have differential responses to DHA .","This has been reported for the histone-deacetylase inhibitor valproic acid , which shows opposing effects in rd1 ( neuroprotective ) and rd10 ( deleterious ) mice ( Mitton et al . , 2014 ) and across four different frog models of RP ( Vent-Schmidt et al . , 2017 ) .","In addition valproic acid has produced inconsistent results in clinical trials with RP patients ( Chen et al . , 2019; Todd and Zelinka , 2017; Totan et al . , 2017 ) .","These results emphasize the need for a more thorough understanding of IRD and RP disease mechanisms and downstream cell pathways to support the development of both personalized and pan-disease therapeutics .","Numerous IRD/RP-linked mutations have been identified ( Dias et al . , 2018; https://sph . uth . edu/retnet/ ) implicating an array of disease mechanisms ( Dharmat et al . , 2020 ) .","However , cell death pathways common across different IRD/RP patient subpopulations may provide pan-disease targets for neuroprotective therapies .","Apoptosis has long been thought to be the mechanism by which rod photoreceptors die in IRD/RP ( Chang et al . , 1993; Doonan et al . , 2003; Portera-Cailliau et al . , 1994; Zeiss et al . , 2004 ) .","However , these reports relied on terminal deoxynucleotidyl transferase dUTP nick end labeling ( TUNEL ) , which does not distinguish apoptosis from other types of cell death ( Ansari et al . , 1993; Charriaut-Marlangue and Ben-Ari , 1995; Grasl-Kraupp et al . , 1995; Dmitrieva and Burg , 2007; Kanoh et al . , 1999; Nishiyama et al . , 1996 ) .","More recent evaluations of multiple apoptosis-related markers ( e . g . BAX , cytochrome c , caspase-9 , cleaved caspase-3 ) suggest apoptosis occurs in only a minority of RP models ( Arango-Gonzalez et al . , 2014; Sancho-Pelluz et al . , 2008 ) .","Moreover , inhibition of apoptosis does not block cell death in many mouse photoreceptor degeneration models ( Hamann et al . , 2009; Yoshizawa et al . , 2002 ) , suggesting other pathways may mediate rod and/or cone cell death in retinal degenerative disease .","Recently , non-apoptotic cell death mechanisms have been implicated in IRD/RP .","In a comprehensive biochemical analysis of 10 mammalian RP models—involving mutations in cnga3 , cngb1 , pde6a , pde6b , pde6c , prph2 , rho , and rpe65—non-apoptotic cell death signatures were found to be common across all models tested ( Arango-Gonzalez et al . , 2014 ) .","Conversely , definitive apoptotic markers were found only for the S334ter ( rho ) rat model .","Shared features included activation of poly ( ADP-ribose ) polymerase ( PARP ) , histone deacetylase ( HDAC ) , and calpain , as well as accumulation of cyclic guanosine monophosphate ( cGMP ) and poly ( ADP-ribose ) ( PAR ) .","For PARP , chemical inhibitors and a knock out line provided further confirmation ( Jiao et al . , 2016; Paquet-Durand et al . , 2007; Sahaboglu et al . , 2017; Sahaboglu et al . , 2016; Sahaboglu et al . , 2010 ) .","Prior reports had suggested the NTR/Mtz system elicits caspase-3 activation and apoptotic cell death ( Chen et al . , 2011 ) .","Initially , we had used this as a rationale for pursuing a neuroprotective screen with the rho:YFP-NTR line- however , the recent reports outlined above suggested apoptosis may have limited relevance to IRD/RP .","Interestingly , when we tested cell death processes implicated in photoreceptor degeneration directly , an inhibitor of apoptosis ( BEL ) did not promote rod cell survival , whereas inhibition of necroptosis ( NEC ) and PARP were neuroprotective in our fish RP model ( Figure 7 , Table 3 ) .","Serendipitously , an exploration of shared lead compound MOA helped to clarify cell death mechanism ( s ) mediating NTR/Mtz-induced rod cell ablation .","To explore molecular MOA of our lead compounds , we searched bioactivity data from prior HTS and ultra HTS assays ( PubChem ) .","The results suggested both shared and independent MOA .","The most common shared target was TDP1 ( eight of 11 lead compounds; Table 2 ) .","An initial test confirmed two of three TDP1 inhibitors tested , though neuroprotective activity was relatively weak ( ~10% survival; Figure 7 ) .","TDP1 is a DNA repair enzyme that repairs topoisomerase I-induced DNA damage ( Dexheimer et al . , 2008; El-Khamisy , 2011 ) .","Interestingly , a qHTS cell-based screen of 400 , 000 compounds for inhibitors of human TDP1 found that all five confirmed compounds actually inhibited PARP activity not TDP1 ( Murai et al . , 2014 ) .","This is consistent with findings showing that TDP1 acts in conjunction with PARP1 ( Das et al . , 2014; Lebedeva et al . , 2015 ) , thus PARP inhibition can indirectly affect TDP1 activity .","Combined with the results discussed above , we were motivated to test whether PARP inhibition was protective against NTR/Mtz-induced rod cell death .","All eight PARP inhibitors tested promoted rod cell survival in our fish RP model ( Figure 7 and Table 3 ) .","CRISPR/Cas9-based knockdown of parp1 resulted in a statistically significant reduction of Mtz-induced rod cell death ( Figure 8A ) .","Moreover , PAR polymer , an indicator of PARP activation , accumulated in Mtz-treated fish ( Figure 8—figure supplement 1C–D ) .","Collectively , these results suggest PARP activation plays an important role in NTR/Mtz-induced rod cell death .","PARP1 overactivation initiates a caspase-independent form of DNA damage-induced cell death , termed parthanatos ( Fan et al . , 2017 ) .","Parthanatos has been strongly implicated in the etiology of Parkinson’s disease ( Kam et al . , 2018 ) and in a variety of neurodegenerative conditions as well ( Fan et al . , 2017; Fatokun et al . , 2014 ) , including RP/IRDs ( Power et al . , 2020 ) .","Importantly , PARP is also a key component of the cGMP-dependent cell death pathway which has been linked to photoreceptor degeneration ( Iribarne and Masai , 2017; Tolone et al . , 2019; Power et al . , 2020 ) .","Thus , the NTR/Mtz-mediated cell ablation system may provide a novel approach for modeling neurodegenerative diseases associated with PARP-dependent cell death that is both inducible and titratable .","To investigate cell death more broadly , we also explored the roles of necroptosis , apoptosis and tdp1 signaling in the zebrafish-inducible RP model .","Interestingly , inhibiting necroptosis but not apoptosis promoted rod cell survival in rho:YFP-NTR fish ( Figure 7 and Table 3 ) .","Consistent with this , CRISPR/Cas9-based knockdown of ripk1l protected rods , but casp3a/3b knockdown did not ( Figure 8 ) .","Necroptosis is primarily associated with secondary cone cell death in RP models ( Murakami et al . , 2015; Murakami et al . , 2012; Yang et al . , 2017 ) but has also been implicated in rod cell death in IRBP mutant RP models ( Sato et al . , 2013 ) and/or may damage photoreceptors indirectly via necroptotic microglia signaling ( Huang et al . , 2018 ) .","Combined , these results suggest that necroptosis , PARP1-dependent parthanatos , and/or c-GMP-dependent cell death mediate NTR/Mtz-induced rod cell ablation .","The potential relevance of these alternative cell death pathways to heritable photoreceptor degeneration may explain the relatively high rate of validation we observed in cross-species tests of lead compounds .","To evaluate whether lead compound effects were mediated by inhibition of PARP signaling , a subset of lead compounds and a PARP inhibitor control ( BMN ) were analyzed for survival effects in parp1 knockdown fish .","Among four leads tested that were implicated as parp1 inhibitors , two ( MIC and DHA ) were no longer protective when parp1 expression was suppressed ( Figure 9 ) , suggesting parp1 inhibition mediated their neuroprotective effects .","PARP1 has also been shown to have a role in stem cell reprogramming ( Chiou et al . , 2013; Doege et al . , 2012; Weber et al . , 2013 ) .","PARP inhibition might therefore block Müller glia dedifferentiation , diverting injury-induced activity from a regenerative to neuroprotective program ( Bringmann et al . , 2009 ) .","Finally , MOA analyses also implicated additional , potentially complementary , lead targets ( Table 2 ) .","Compounds acting through independent targets/pathways have the potential to produce additive or even synergistic effects .","This possibility was confirmed for 10 of 19 paired lead compounds tested in rho:YFP-NTR fish ( Figure 9 , Supplementary file 1c ) .","Analogous tests in mouse photoreceptor cell cultures showed a trend toward enhanced survival effects of paired leads but no fully additive effects ( Figure 4—figure supplement 2 ) .","This discrepancy could be due to cell culture assays lacking in vivo complexity , such as signals requiring intact tissue or interactions between cell types or with other tissues .","This result highlights a potential advantage of whole organism phenotypic drug screening: the ability to identify compounds/conditions that would be missed in more simplified assay systems .","That combinatorial assays could be performed efficiently and rapidly also exemplifies key advantages the zebrafish system affords phenotypic drug discovery , for example , versatility and low cost .","In summary , we identified 11 lead compounds promoting rod cell survival in an inducible zebrafish RP model using a large-scale in vivo phenotypic drug discovery platform .","Nine lead compounds were also effective as neuroprotectants in primary mouse photoreceptor cell cultures and three promoted photoreceptor survival in rd1 mouse retinal explants .","MOA studies indicated a subset of lead compounds may protect rod cells by inhibiting PARP-dependent cell death pathways and/or necroptosis .","Combinatorial assays in fish showed additive and even synergistic effects , suggesting some lead compounds target independent neuroprotective pathways ( summarized in Figure 11 ) .","We hypothesize that compounds producing beneficial outcomes across diverse animal disease models likely target highly conserved MOA and may therefore stand an increased likelihood of successfully translating to the clinic .","Further , our data suggest polypharmacological targeting of complementary neuroprotective mechanisms has the potential to maximize therapeutic benefits for IRD/RP patients ."],["All animal studies described herein were performed in accordance with both the Association for Research in Vision and Ophthalmology ( ARVO ) statement on the ‘Use of Animals in Ophthalmic and Vision Research’ and the National Institutes of Health ( NIH ) Office of Laboratory Animal Welfare ( OLAW ) policies regarding studies conducted in vertebrate species .","Animal protocols were approved by the Animal Care and Use Committees of the Johns Hopkins University School of Medicine ( protocol # FI19M489 and #MO20M253 ) and Medical University of South Carolina ( protocol #2018–00399 ) .","Zebrafish were maintained using established temperature and light cycle conditions ( 28 . 5°C , 14 hr of light/10 hr of dark ) .","A previously generated zebrafish transgenic line , Tg ( rho:YFP-Eco . NfsB ) gmc500 ( hereafter , rho:YFP-NTR ) expresses a fusion protein of enhanced yellow fluorescent protein ( YFP ) and a bacterial nitroreductase ( NTR ) enzyme ( encoded by the E . coli nfsB gene ) selectively in rod photoreceptors ( Walker et al . , 2012 ) .","When rho:YFP-NTR fish are exposed to the prodrug metronidazole ( Mtz ) , NTR reduces Mtz to a DNA damage-inducing cytotoxic derivative , resulting in the death of NTR-expressing cells ( Chen et al . , 2011; Curado et al . , 2007 ) .","A newly generated transgenic line , Tg ( rho:GAP-YFP-2A-nfsB_Vv F70A/F108Y ) jh405 ( hereafter rho:YFP-NTR2 . 0 ) expresses a membrane tagged YFP and an improved version of NTR which enables induction of cell death at much lower concentrations of Mtz ( Sharrock et al . , 2020 ) .","These lines were propagated in a pigmentation mutant with reduced iridophore numbers , roya9 ( roy ) , to facilitate YFP reporter signal detection in vivo .","Non-transgenic roy fish were used to define reporter signal cutoff values for fluorescent microplate reader assays .","For each large-scale drug screening assays , 6000–12 , 000 eggs were collected by group breeding ~300 adult rho:YFP-NTR fish ( White et al . , 2016 ) .","rho:YFP-NTR larvae were treated with 2 . 5 mM Mtz or 0 . 1% DMSO at 5 dpf for 2 days ( 10 fish per condition ) , and collected at 7 dpf for immunostaining using previously reported methods ( Unal Eroglu et al . , 2018 ) .","Briefly , larvae were fixed in 4% PFA at 4°C overnight , infiltrated with 30% sucrose , and embedded in OCT .","Cryosections of 10 µm thickness were made for immunofluorescence staining .","Each section was blocked with 1x PBS containing 0 . 5% Triton X-100% and 5% goat serum for one hour followed by primary antibody staining overnight .","The next day , each section was washed and stained with the secondary antibody for 2 hr .","All slides were mounted and underwent confocal imaging .","The primary antibodies used in this study were zpr-1 ( anti-Arrestin3a; 1:200 dilution; from ZIRC ) , 1d1 ( anti-Rhodopsin; 1:50; gift from Dr . James M . Fadool ) and 4C12 ( uncharacterized rod cell antigen , 1:100 dilution; gift from Dr . James M . Fadool ) .","The secondary antibody used was goat anti-mouse Alexa 647 ( 1:1000; Invitrogen ) .","rho:YFP-NTR larvae were separated into six groups of ~23 larvae per group .","Each group was treated with varying Mtz concentrations ( 10 , 5 , 2 . 5 , 1 . 25 , 0 . 625 , 0 mM ) , from 5 to 8 days post-fertilization ( dpf ) .","YFP signals were quantified daily by fluorescence microplate reader ( TECAN Infinite M1000 PRO; excitation 514 nm , bandwidth 5 nm; emission 538 nm , bandwidth 10 nm ) to track changes in rod photoreceptor cell numbers relative to Mtz concentration ( Walker et al . , 2012; White et al . , 2016 ) .","Two technical repeats were performed and results aggregated across experiments by normalizing to non-ablated controls .","To calculate sample size ( N ) and evaluate assay quality , two 96-well plates of larvae were treated with 2 . 5 mM Mtz/0 . 1% DMSO ( ablated ‘+Mtz’ control ) or 0 . 1% DMSO only ( non-ablated ‘-Mtz’ control ) from 5 to 7 dpf ( 196 fish per condition ) .","YFP signals were then quantified at seven dpf ( due to the large sample size , a single technical replicate was performed ) .","Power calculations were used to determine sample sizes across a range of error rates and effect sizes ( White et al . , 2016 ) .","This analysis suggested a sample size of nine per condition tested would sufficiently minimize false-discoveries ( i . e . type I and type II error rates of 0 . 05 and 0 . 05 , respectively ) and account for drug effects reaching 50% of signal window of the positive ( non-ablated ) control .","However , to account for plating errors and other confounding variables , we chose to increase the sample size to 16 .","A schematic of the primary screening process is presented in Figure 1C .","For the primary screen , 2934 compounds from John Hopkins Drug Library ( JHDL ) were screened across six concentrations ( 4 μM-125 nM using a twofold dilution series ) .","The JHDL consists of ~2200 drugs approved for use in humans ( e . g . FDA approved ) with the remainder approved for clinical trials ( Chong et al . , 2006 ) .","The ARQiv screening process has been detailed previously ( White et al . , 2016 ) and was adapted here for large-scale quantification of YFP-expressing rod photoreceptors ( Walker et al . , 2012; White et al . , 2016 ) .","rho:YFP-NTR embryos were collected and raised in zebrafish E3 embryo media ( 5 mM NaCl; 0 . 17 mM KCl; 0 . 33 mM CaCl; 0 . 33 mM MgSO4 ) .","At 16 hr post fertilization ( hpf ) , N-phenylthiourea ( PTU ) was added to E3 media ( E3/PTU ) at a final concentration of 0 . 2 mM to promote ocular transparency by inhibiting melanosome maturation in the retinal pigment epithelium .","At 4 dpf , visual screens were performed to remove larvae with abnormal morphology or low YFP expression levels .","Stock drug and DMSO ( negative control ) solutions were automatically dispensed and diluted across a 96-well plate containing E3/PTU using a robotic liquid handling system ( Hudson Robotics ) .","At 5 dpf , a COPAS-XL ( Complex Object Parametric Analyzer and Sorter , Union Biometrica ) was used to dispense single larvae into individual wells containing either drug or DMSO; the final DMSO concentration was 0 . 1% across all conditions .","After a 4 hr pre-exposure to test drugs or DMSO alone , larvae were treated with 2 . 5 mM Mtz to induce rod photoreceptor death .","Larvae were maintained under these conditions for 2 days until 7n dpf and then anesthetized with clove oil ( 50 ppm final concentration ) ; YFP signals were measured as described above .","Larvae exposed to 2 . 5 mM Mtz/0 . 1% DMSO without any tested drug served as controls ( ‘+Mtz’ ) for maximal rod cell ablation .","Larvae treated solely with 0 . 1% DMSO served as non-ablated controls ( ‘-Mtz’ ) to calculate maximal YFP signal levels .","Non-transgenic larvae were used to establish a signal cutoff value , as previously described ( White et al . , 2016 ) .","A customized R code program was applied for real-time data analysis of compound performance relative to controls , including dose-response curves and strictly standardized mean difference ( SSMD ) scores ( https://github . com/mummlab/ARQiv2; Ding and Zhang , 2021 ) .","Compound concentrations producing a SSMD score ≥1 were considered potential ‘hits’ and evaluated visually using fluorescence microscopy to eliminate non-specific fluorescence; e . g . , dead larvae or autofluorescent compounds .","After the initial screen , 42 top-performing ‘hit’ compounds were selected for confirmatory and orthogonal assays .","Hit drugs were obtained from new sources and tested across a wider range of concentrations ( fivefold dilution series , eight total concentrations , n=30 fish/condition ) .","Based on the toxicity profile of each drug , the starting concentration was either 100 , 10 , or 1 μM .","For validation assays , YFP signals were measured by fluorescence microplate reader with at least three experimental repeats conducted and SSMD scores calculated .","All experimental results were normalized and pooled to calculate effect sizes , confidence intervals , sample sizes ( N ) and p-values .","p Values were calculated using Student’s t-test followed by Bonferroni correction for multiple comparisons resulting in an adjusted significance level of 0 . 005 ( α=0 . 005 ) .","For in vivo confocal imaging assays , five dpf rho:YFP-NTR larvae were treated with drug plus 2 . 5 mM Mtz/0 . 1% DMSO ( ablated +Mtz controls ) , or 0 . 1% DMSO ( non-ablated -Mtz controls ) for 48 hr and processed for intravital imaging at 7 dpf .","Similar treatment strategy was applied to five dpf rho:YFP-NTR2 . 0 larvae except the Mtz treatment was reduced to 10 μM .","All compounds were tested at the maximal effective concentration in 0 . 1% DMSO .","Larvae from each group were anesthetized in 0 . 016% tricaine and embedded on their sides in 1% low melt agarose gel .","An Olympus Fluoview FV1000 confocal microscope with a 20x water immersion objective ( 0 . 95 NA ) was used to collect 30–40 z-stack images of YFP-expressing rod cells across the whole retina at 4 μm intervals .","These image slices were stacked into a single maximal intensity projection image using Image J . To provide greater detail of rod photoreceptor morphology , a region in the dorsal-nasal quadrant was imaged using a 60× water immersion objective ( 1 . 10 NA ) at 4 . 18 μm intervals .","Three contiguous image slices were then stacked into a single maximal intensity projection image using Image J . As a means of assessing rod photoreceptor survival , intravital confocal retinal images were collected of YFP-expressing rod cells per condition and processed for volumetric quantification of YFP signal volume using automated image analysis software ( Imaris 3 . 9 . 1; see White et al . , 2017 ) .","Briefly , YFP-expressing rod cells volumes were automatically rendered using the same parameters across all treatment groups with the sum of all volumes used to estimate rod photoreceptor numbers per each retina .","For images from rho:YFP-NTR2 . 0 larvae , the sum of rod cell YFP volumes in each retina was normalized to the number of YFP-expressing cells to estimate the average rod photoreceptor cell size per each condition .","The average number of cells quantified per retina was 18 . 52 , with sample sizes ranging from 9 to 19 fish per condition .","The eleven lead compounds were tested to eliminate false-positives that directly inhibited E . coli NTR enzymatic activity ( from the nfsB_Ec gene ) using an established anticancer prodrug ( CB1954 ) reduction assay ( Prosser et al . , 2010 ) .","The highest soluble concentration of the test drug ( up to 300 μM maximum ) or 0 . 1% DMSO ( rate control ) was added to a master mix of 250 μM CB1954/1 μM NTR/250 μM NADPH/10 mM Tris ( pH 7 . 0 ) to determine any inhibitory effect of the compounds on NTR-mediated reduction of CB1954 .","Reactions were conducted in 100 μl volume in a 96-well plate format .","CB1954 reduction kinetics were assayed at 420 nm unless the test compound exhibited confounding absorbance at 420 nm , in which case NADPH depletion was monitored at 340 nm .","The reaction rates for CB1954 reduction in the presence of tested compounds were compared to the DMSO control .","Compounds with NTR activity <75% of controls were deemed potentially inhibitory and IC50 values ( the concentration required to restrict NTR activity to 50% of the DMSO control ) determined .","Compounds were also evaluated for the ability to inhibit NTR-mediated reduction of Mtz .","For this assay , the highest soluble concentration of the test drug ( up to 300 μM maximum ) or DMSO ( rate control ) was added to a master mix of 500 μM Mtz/17 μM NTR/100 μM NADPH/0 . 55 μM glucose dehydrogenase/5 mM glucose/50 mM sodium phosphate buffer ( pH 7 . 0; the glucose dehydrogenase used for cofactor regeneration , maintaining a steady-state of NADPH to allow Mtz reduction to be monitored directly at 340 nm without interference from NADPH oxidation ) .","Reactions were conducted in 100 μl volume and assayed at 340 nm using 1 mm quartz cuvettes ( Rich et al . , 2018 ) .","IC50 values for potentially inhibitory compounds were measured as with CB1954 .","All kinetic assays were conducted in triplicate with a single purified batch of NfsB_Ec protein .","To evaluate whether lead compounds promoted rod photoreceptor development , rho:YFP-NTR larvae were handled as described for the primary screen with the following exceptions: at five dpf , larvae were exposed solely to test compounds; larvae were not treated with Mtz .","YFP reporter signals were quantified by fluorescence microplate reader at 7 dpf and compared to non-ablated ‘-Mtz’ controls treated only with 0 . 1% DMSO .","Sample sizes ranged from 58 to 88 across at least two experimental repeats .","To determine whether lead compounds stimulated retinal regeneration , rho:YFP-NTR larvae were handled as described for the primary screen with the following exceptions: at 5 dpf , larvae were incubated with either 10 mM Mtz or 0 . 1% DMSO for 24 hr .","At 6 dpf , larvae were then placed in new 0 . 1% DMSO/E3/PTU media containing test compounds ( or DMSO alone ) for 3 days .","YFP signal intensity was measured at 9 dpf .","Sample sizes ranged from 31 to 83 across at least two experimental repeats .","A transgenic mouse line with an IRES-GFP cassette integrated at the QRX locus ( QRX-IRES-EGFP ) has GFP reporter expression restricted to retinal photoreceptors , and was used here to isolate primary photoreceptor cells .","Mice were housed with 12 hr light/12 hr dark cycles , at 22°C , 30–70% relative humidity and food/water ad libitum .","Primary mouse retinal photoreceptor cells were isolated and prepared for culture as previously described ( Fuller et al . , 2014 ) .","Briefly , murine retinas were isolated at postnatal day four ( P4 ) .","Retinal tissue was dissociated into a single-cell suspension by incubating whole tissue in activated papain in Hibernate-E without Calcium ( BrainBits ) for 15 min at 37°C .","Cells were resuspended in culture media ( Neurobasal , 2% B-27 , 0 . 5 mM L-Glutamine and 1x final Penicillin/streptomycin; all Life Technologies ) and seeded onto poly-D-lysine coated 384 well tissue culture plates .","To test lead compounds for survival effects , thapsigargin and tunicamycin ( Sigma ) and were used as stressor compounds to induce photoreceptor cell death .","Stressors and test compounds were added at the time of primary cell seeding .","Eleven lead compounds were tested at seven concentrations across a threefold dilution series ( from 30 µM to 40 nM ) .","For paired tests , each compound was applied at a single concentration , WAR ( 3 . 33 µM ) , CPO ( 0 . 04 µM ) , DHA ( 1 . 11 µM ) , ART ( 10 µM ) , ZPT ( 0 . 12 µM ) , CAL ( 0 . 04 µM ) , COR ( 3 . 33 µM ) , CHL ( 3 . 33 µM ) , SUL ( 3 . 33 µM ) , MIC ( 10 µM ) , and CLO ( 1 . 11 µM ) .","After a 48 hr treatment , cells were stained with Hoescht and ethidium homodimer; nine field images were acquired via an automated imager ( Cellomics Vti; ThermoFisher ) using a 20x objective and images analyzed .","Photoreceptor number and the percent of photoreceptor to retinal cell population per well was determined by automated quantification of the number of live Hoechst 33342-stained , GFP-expressing cells using a custom algorithm ( Neuronal Profiling software package; ThermoFisher ) .","A minimum of two experimental repeats was performed for each test with individual assays consisting of six biological replicates per condition tested .","Mice were generated from retinal degeneration 1 ( Pde6brd1 , hereafter rd1 ) breeding pairs ( gift from Dr . Debora Farber; UCLA ) and housed under a 12:12 light:dark cycle with access to food and water ad libitum .","All chemicals used for organ cultures were tissue culture grade and purchased from Invitrogen unless otherwise noted .","Cultures of retina with attached retinal pigment epithelium ( RPE ) were grown according to published protocols ( Ogilvie et al . , 1999; Pinzón-Duarte et al . , 2000; Rohrer and Ogilvie , 2003 ) with modifications ( Bandyopadhyay and Rohrer , 2010 ) .","P10 pups were decapitated and heads were rinsed in 70% ethanol; whole eyes were dissected and placed in ice-cold Hanks balanced salt solution plus glucose ( 6 . 5 g/L ) .","Eyes were first incubated in 1 mL of high-glucose Hanks balanced salt solution/0 . 5 mg/ml proteinase K ( 37°C , 7 min ) and then in Neurobasal medium ( Life Technologies ) plus 10% fetal calf serum to stop enzymatic activity .","The retina with attached RPE was dissected free from the choroid and sclera after removing the anterior chamber , lens and vitreous .","Relaxing cuts were made to flatten the tissue prior to transfer to the upper compartment of a Costar Transwell chamber ( RPE layer faced-down ) using a drop of Neurobasal medium .","Neurobasal media with B-27 supplement ( Life Technologies ) was placed in the lower compartment .","For each of the cohorts , three to fourindividual P10 retina/RPE explants were placed in culture .","The cultures were kept in an incubator ( 5% CO2 , balanced air , 100% humidity , 37°C ) and the lower compartment media changed every 2 days ( neither antimitotics nor antibiotics were required ) .","Test compounds were added to the culture media and refreshed every 2 days for 11 days .","At completion , explants were fixed in 4% PFA , sectioned 14 µm thick , and stained with toluidine blue ( Bandyopadhyay and Rohrer , 2010 ) .","For each culture , 10 counts of rows of photoreceptors were performed in the center of the explant; the average of this cell counts provides the mean for each retina , with the average of all retinas providing the mean ± SEM per culture condition .","Cell counts were compared by repeated measure ANOVA to assess dose-response treatment effects followed by a Bonferroni posthoc analysis to determine if differences between control and lead compounds compared across the entire dose range were statistically significant .","An ANOVA across groups but within a particular dose , using a Bonferroni/Dunn test to reduce type I errors , was used to determine if lead compounds promoted a statistically significantly increase in cell numbers compared to control media at low , middle , and/or high concentrations .","PLGA-DHA microparticles were prepared using a single emulsion solvent evaporation method .","Briefly , 200 mg PLGA ( Resomer RG 502 hr , Poly ( D ) , L-lactide-co-glycolide; Evonik Corporation ) was dissolved in 1 mL of dichloromethane ( DCM , Sigma-Aldrich ) , and mixing with 40 mg DHA ( TCI , Tokyo , Japan ) dissolved in 0 . 125 ml dimethyl sulfoxide ( DMSO , Sigma-Aldrich ) .","The mixture was homogenized ( L4RT , Silverson Machines ) at 7000 RPM for 1 min .","The homogenized mixture was then poured into a solution containing 1% polyvinyl alcohol ( 25 kDa , 88% hydrolys , Polysciences , Warrington , PA ) in water under continuous stirring .","Particles were hardened by allowing solvent to evaporate while stirring at room temperature for 2 hr .","Particles were collected via centrifugation ( International Equipment Co ) at 1000 x g for 5 min , and washed three times with HyPure cell culture grade water ( endotoxin-free , HyClone , Logan , UT ) and re-collected by centrifugation three times .","The washed particles were then lyophilized and stored frozen until use .","Microparticles were resuspended at the desired concentration prior to injection in a sodium hyaluronate solution ( Healon ) diluted fivefold with endotoxin-free HyPure water at the desired concentration prior to injection .","Particle size distribution was determined using a Coulter Multisizer 4 ( Beckman Coulter , Inc , Miami , FL ) .","Particles were resuspended in double distilled water and added dropwise to 100 ml of ISOTON II solution until the coincidence of the particles was between 8% and 10% .","At least 100 , 000 particles were sized for each batch of particles to determine the mean particle size and size distribution .","To determine the drug loading , microparticles were dissolved in DMSO and the total drug content was calculated by measuring the UV absorbance at 289 nm after react with NaOH in triplicate ( C . -S . Lai et al . , 2007a ) .","The average microparticle size was 14 . 2 ± 1 . 9 µm and the DHA loading was 20 . 3% ± 0 . 7 ( w/w ) across three experimental replicates .","Release kinetics were obtained by resuspending microparticles in 1 ml phosphate buffered saline ( PBS containing 0 . 1% DMSO , pH 7 . 4 ) and incubating at 37°C on a platform shaker ( 140 RPM ) .","Supernatant was collected at predetermined intervals by centrifugation at 2000 x g for 5 min .","Drug-containing supernatant was collected and particles were resuspended in 1 ml of fresh 1xPBS containing 0 . 1% DMSO .","DHA concentration in the collected supernatant was assayed via absorbance at 289 nm in triplicate for each sample ( Figure 4—figure supplement 1 ) .","B6 . CXB1-Pde6brd10/J animals ( Jackson Laboratories stock #004297 ) were maintained as homozygotes .","On P14 , animals were anesthetized with ketamine/xylazine ( 115 mg/kg and 5 mg/kg , respectively ) , and placed on a heating pad to maintain body temperature throughout injection and recovery .","Glass needles were pulled and a bore size of approximately 75 µm was made to allow for the movement of the microspheres into and out of the needle .","A PLI-100 picospritzer ( Harvard Apparatus ) was used for intravitreal injections with settings of 350 ms injection time at 30 psi , which yielded the injection of approximately 500 nL .","One eye was randomly selected for injection with vehicle and the other was used for microsphere injections .","Following injection , GenTeal gel ( Novartis ) was applied to the eyes to prevent corneal drying .","Animals were monitored until they were awake and moving normally .","A total of 17 mice was used across three technical repeats .","Eighteen days post-injection , animals were anesthetized with isoflurane and decapitated; the eyes were removed and placed in 4% PFA for 30 min .","After removing the cornea and lens , the eye cups were placed back into fixative for 2 hr on ice .","After thorough washing with 1× PBS , eyecups were soaked in 15% sucrose ( w/v ) in 1× PBS for 3 hr , then incubated in 30% sucrose ( w/v ) /1× PBS overnight .","Samples were embedded in OCT compound ( TissueTek ) and frozen on dry ice .","Cryosections of 12–14 µm thickness were cut on a Leica CM3050S cryostat and were mounted on Superfrost Plus slides ( Fisher Scientific ) .","Slides were rehydrated with 1× PBS and then incubated in blocking buffer ( 10% Normal donkey serum/ 0 . 3% Triton X-100/1× PBS ) for 2 hr , and were then incubated for 30 hr at 4°C with primary antibodies diluted in blocking buffer .","Sections were washed in 1× PBS and incubated for 2 hr at room temperature with secondary antibodies diluted in blocking buffer .","Following additional washes , sections were incubated with DAPI and mounted with FluoroGel ( Electron Microscopy Sciences ) .","Sections were examined using a Zeiss Axioplan two epifluorescence microscope fitted with an Axioscope camera .","Following acquisition with a 40× air objective , images were analyzed in ImageJ using the measure tool to examine Outer Nuclear Layer ( ONL ) thickness .","Four to 5 measurements were taken at 75 µm intervals for 200–500 µm both superior and inferior from the optic nerve head .","The measured values from multiple sections were averaged for each eye .","Data were analyzed and statistical tests were performed in GraphPad Prism 7 .","An Olympus FV1000 confocal microscope was also used to image sections stained with multiple dyes .","To evaluate tdp1 and parp1 as a potential shared target across multiple leads , as well as cell death pathways implicated in photoreceptor degeneration , eight PARP inhibitors , one necroptosis inhibitor , three apoptosis inhibitors , and three Tdp1 inhibitors were selected for testing using the confirmation screening protocol ( Figure 2A ) .","Based on the toxicity profile of each inhibitor , upper concentrations were adjusted to 64 , 32 , or 8 µM and a twofold dilution series was used to test a total of 10 concentrations for rod cell survival effects in rho:YFP-NTR larvae .","Sample sizes varied from 27 to 173 across two to six experimental replicates for each tested compound .","YFP signal levels were measured by fluorescence microplate reader ( i . e . ARQiv assay ) .","Results across conditions were normalized to non-ablated controls and pooled across experimental repeats to calculate effect sizes , confidence intervals , and p-values .","P values were calculated by performing Student’s t-tests .","Bonferroni correction for multiple comparisons resulted in an adjusted significance level of 0 . 003 ( α=0 . 003 ) .","Note: assessments of TDP1 inhibitors followed the fivefold dilution series protocol used for secondary confirmation assays described above .","For paired inhibitor tests , three concentrations of BMN ( parp inhibitor ) and three concentrations of NEC ( necroptosis inhibitor ) were combined .","YFP quantification and statistical analysis of rod cell survival in Mtz-treated rho:YFP-NTR fish were performed as above .","Five genes , parp1 , ripk1l , casp3a , casp3b , tdp1 , were selected for CRISPR/Cas9-based knockdown in rho:YFP-NTR zebrafish using an established ‘redundant targeting’ methodology ( Wu et al . , 2018 ) .","Four sgRNAs were designed to target each gene .","The oligonucleotide for generating each sgRNA was either those suggested by Wu et al . , 2018 , Table S2 or designed using CRISPRscan ( https://www . crisprscan . org , Supplementary file 3; Moreno-Mateos et al . , 2015 ) .","Preparation of sgRNAs followed the published method ( Wu et al . , 2018 ) .","Briefly , four sgRNAs of each gene were in vitro transcribed using the HiScribe T7 High Yield RNA Synthesis kit ( New England BioLabs ) and purified using the NEB Monarch RNA Cleanup kit .","A mixture of four sgRNAs ( 1 ng in total ) and Cas9 protein ( 2 . 5 µM , IDT ) was injected into rho:YFP-NTR embryos at the one-cell stage for targeting parp1 , ripk1l and tdp1 .","Four sgRNAs each for casp3a and casp3b were pooled to co-target both genes .","Injected and uninjected embryos were treated ±Mtz ( 2 . 5 mM ) at 5 dpf .","Rod cell survival was evaluated by fluorescence microplate reader-based quantification of YFP at 6 dpf .","Student’s t-test was used to compare rod survival between injected and uninjected ( control ) fish .","Each gene knockdown experiment was conducted in triplicate or quadruplicate .","A subset of lead compounds ( and a PARP inhibitor control , BMN ) was also tested in parp1 knockdown fish to determine if Parp1 inhibition was required for survival effects .","Leads were applied to larvae one hour before Mtz administration and YFP quantification performed as above .","Each drug test was conducted in duplicate or triplicate .","To calculate p-values relative to +Mtz controls , Student’s t-tests were performed .","Bonferroni correction for multiple comparisons resulted in an adjusted significance level of 0 . 006 ( α=0 . 006 ) .","To quantify gene expression levels in rho:YFP-NTR larvae ±CRISPR/Cas9-based gene knockdown , 16–24 larvae from each condition were pooled for mRNA extraction at six dpf .","RNA was purified using NEB Monarch RNA Cleanup kit and reverse transcribed to cDNA using qScript cDNA synthesis kit ( QuantaBio ) according to the instruction from the manufacturer .","Quantitative PCR was conducted using designed primers ( Supplementary file","3 ) and PowerUp SYBR Green Master Mix ( ABI ) in QuantaStudio ( ABI ) .","Delta delta CT analysis was performed to calculate relative fold change in gene expression levels between knockdown larvae and controls .","Each experiment was performed in triplicate or quadruplicate .","To further evaluate Parp-dependent signaling downstream of Mtz treatments in rho:YFP-NTR larvae , PAR accumulation , a vetted measure of Parp activity , was quantified by western blot following the published method ( Kam et al . , 2018 ) .","Briefly , ~30 rho:YFP-NTR fish were treated ±2 . 5 mM Mtz at five dpf were collected at 3 , 6 , 12 , 24 , 48 hr post treatment ( hpt ) .","Larvae were homogenized and lysed using RIPA buffer ( Sigma ) supplemented with protease inhibitors ( Roche ) .","Protein concentrations were quantified using Pierce BCA kit ( ThermoScientific ) following the manufacturer’s instruction .","Protein samples were separated on Tris-Glycine gel and then transferred onto nitrocellulose membranes .","The membrane was blocked with 5% non-fat milk in TBST ( Tris-buffered saline with 0 . 1% Tween 20 ) at room temperature for 1 hr and then blotted with the primary antibody , anti-human-PAR ( 1:2000 , Dowson lab reagent ) , at 4°C overnight .","After rinsing in TBST for three times , the secondary antibody , anti-Human IgG-HRP ( 1:10 , 000 , Abcam ) was applied for 1 hr at room temperature .","Anti-beta-actin-HRP antibody ( 1:20 , 000 ) was used as a loading control .","Immunolabeled bands were visualized by ECL substrate and quantified in Fiji .","24 hr post-Mtz PAR accumulation assays were performed in triplicate .","Mann-Whitney U test was performed and a p-value of ≤0 . 05 used to indicate statistical significance .","To test for additive neuroprotective effects in zebrafish , seven lead compounds were tested at optimal concentrations either individually or in pairs using the primary screen protocol .","All experimental results were normalized to controls and pooled per compound tested to calculate effect sizes , confidence intervals , and p-values using Student’s t-test .","Bonferroni correction for multiple comparisons resulted in an adjusted significance level of 0 . 002 ( α=0 . 002 ) .","Sample sizes for each paired condition varied from 86 to 160 across at least four experimental replicates .","For the primary screening , under the R environment , a customized ARQiv data analysis package was used to calculate sample size ( n ) , quality control strictly standardized mean difference ( SSMD ) and SSMD scores as previously described ( White et al . , 2016 ) .","The following results of each drug were derived: ( 1 ) a plot of signal to background ratio at all tested concentrations , ( 2 ) a plot and a table of SSMD scores , and ( 3 ) a signal intensity heat map of each drug plate ( 96-well plate view ) .","To combine and analyze the data from different experiments , data normalization was conducted by ( Si -X-neg ) / ( X- Pos -X-neg ) .","Si is the signal of ith reading , X-Pos is mean of positive controls , and X-Neg is mean of negative controls .","To compare between experimental conditions and controls , Student’s t-tests were performed followed by Bonferroni correction for multiple comparisons ( α/n ) .","Relative effects , 95% confidence intervals and p-values were calculated .","For mouse photoreceptor cell culture assays and volumetric confocal image analyses , adjusted p-values were calculated by performing Mann-Whitney U tests followed by false discovery rate correction for multiple comparisons ."]],"string":"[\n [\n \"Inherited retinal diseases ( IRDs ) encompass a group of genetically linked retinopathies characterized by progressive photoreceptor death ( Duncan et al . , 2018 ) .\",\n \"IRDs lead to irreversible vision loss , for which treatment strategies are limited .\",\n \"Retinitis pigmentosa ( RP ) , the most common IRD with approximately 1 . 5–2 . 5 million RP patients a worldwide ( Dias et al . , 2018; Hartong et al . , 2006; Verbakel et al . , 2018 ) , is characterized by early onset night blindness , gradual loss of visual field , and eventual loss of central vision ( Ferrari et al . , 2011; Hamel , 2006 ) .\",\n \"The initial pathological feature of RP is selective rod photoreceptor cell death causing night blindness , which is then followed by loss of cone photoreceptors and eventual full blindness ( Léveillard et al . , 2014 ) .\",\n \"Mutations in more than 70 genes have been linked to RP ( Dias et al . , 2018; https://sph . uth . edu/retnet/ ) .\",\n \"How these mutations affect gene function or initiate aberrant photoreceptor cell loss is largely unknown .\",\n \"Given the genetic diversity , pan-disease therapeutics are highly desirable .\",\n \"Accordingly , the purpose of our study was to identify compounds capable of promoting rod cell survival in multiple RP models and across species .\",\n \"As RP progression is relatively protracted , and even small numbers of surviving rod photoreceptors can preserve cone photoreceptor function ( Guadagni et al . , 2015; Hartong et al . , 2006; Punzo et al . , 2012 ) , pharmacological interventions aimed at slowing rod photoreceptor death are sought ( Duncan et al . , 2018; Wubben et al . , 2019 ) .\",\n \"However , currently there are no effective therapies for promoting rod photoreceptor survival .\",\n \"As a means of discovering new pharmacological treatments , target-directed high-throughput screening ( HTS ) approaches have been highly successful in identifying compounds that bind to and/or modulate disease-implicated molecules .\",\n \"However , many promising leads fail during late-stage animal model testing or clinical trials ( Munos , 2009; Sams-Dodd , 2013; Scannell et al . , 2012 ) .\",\n \"This trend has renewed interest in phenotypic drug discovery ( PDD ) , a complementary approach where drug effects are evaluated in cells or living disease models ( Bickle , 2010; Lee et al . , 2012; Swinney , 2013 ) .\",\n \"A number of first-in-class drugs were recently discovered using PDD ( Eder et al . , 2014; Swinney and Anthony , 2011; Swinney , 2013 ) .\",\n \"To expand opportunity on this front , we developed a PDD platform enabling quantitative HTS ( qHTS; Inglese et al . , 2006 ) in zebrafish ( Walker et al . , 2012; Wang et al . , 2015a; White et al . , 2016 ) .\",\n \"Zebrafish offer several distinct advantages as a retinal disease modeling system ( Angueyra and Kindt , 2018; Richardson et al . , 2017 ) .\",\n \"First , the structure of the zebrafish retina is similar to the other vertebrates ( Angueyra and Kindt , 2018; Richardson et al . , 2017; Schmitt and Dowling , 1999 ) .\",\n \"In particular , the zebrafish retina is ‘cone rich’ like the human retina .\",\n \"Second , about 70% of human genes have at least one ortholog in zebrafish ( Howe et al . , 2013 ) .\",\n \"Moreover , all RP-associated genes listed in RetNet ( https://sph . uth . edu/retnet/ ) have conserved zebrafish orthologs .\",\n \"Third , the zebrafish retinal system develops quickly being fairly mature by day five of development ( Brockerhoff et al . , 1995; Moyano et al . , 2013; Schmitt and Dowling , 1999 ) .\",\n \"Fourth , zebrafish are amendable to large-scale chemical screening due to their high fecundity rate , small size , and ease of visualizing and quantifying a variety of phenotypes ( Mathias et al . , 2012; Zon and Peterson , 2005 ) .\",\n \"To streamline such screens , we developed a high-throughput plate reader-based method for quantifying reporter gene expression in vivo ( ‘ARQiv’; Walker et al . , 2012 ) .\",\n \"Recently , we adapted the ARQiv system to human stem-cell-derived retinal organoids ( Vergara et al . , 2017 ) to enable cross-species PDD .\",\n \"To realize full throughput potential , ARQiv was combined with robotics-based automation to create ‘ARQiv-HTS’ ( Wang et al . , 2015b; White et al . , 2016 ) .\",\n \"Here , to identify neuroprotective compounds promoting rod photoreceptor survival , ARQiv-HTS was used to perform a large-scale chemical screen in a transgenic zebrafish model enabling inducible rod photoreceptor cell death ( Walker et al . , 2012; White et al . , 2017 ) .\",\n \"In these fish , YFP-NTR ( a yellow fluorescent protein-nitroreductase bacterial enzyme fusion protein ) is selectively expressed in rod photoreceptors .\",\n \"NTR expression enables inducible ablation of rod cells upon exposure to prodrug substrates , such as metronidazole ( Mtz ) , which are thought to cause apoptosis through activated metabolites that cause DNA damage ( Curado et al . , 2007; White and Mumm , 2013 ) .\",\n \"Close to 3000 largely human-approved drugs were tested across six concentrations ( i . e . using qHTS principles , Inglese et al . , 2006 ) in more than 350 , 000 zebrafish larvae .\",\n \"Statistically , 113 hits were identified as hits and 42 of the top performing compounds advanced through confirmation and orthogonal assays .\",\n \"Eleven compounds passed all secondary tests and moved forward as lead drug candidates .\",\n \"Validation tests in an orthogonal series of mouse RP models followed , reasoning that drugs effective across species and/or independent of disease mutation may translate better clinically .\",\n \"All leads were screened in primary mouse photoreceptor cell cultures and a subset of leads was evaluated in retinal explants from the retinal degeneration 1 ( Pde6brd1 , hereafter rd1 ) mutant mouse model of RP .\",\n \"Finally , one of three top-performing leads was tested in retinal degeneration 10 mutant retinas in vivo ( Pde6brd10 , hereafter rd10 ) .\",\n \"An analysis of potential mechanisms of action ( MOA ) suggested both shared and complementary targets/pathways .\",\n \"The role of Poly ( ADP-ribose ) polymerase ( PARP ) , the most common shared target , was evaluated using chemical inhibition , genetic targeting , and western blot analysis .\",\n \"The results implicated PARP as a key mediator of NTR/Mtz-mediated rod cell ablation , and PARP1 inhibition as a shared MOA for a subset of leads .\",\n \"As PARP-dependent cell death pathways are implicated in the etiology of Parkinson’s disease ( Kam et al . , 2018 ) , RP ( Power et al . , 2020 ) , and other neurodegenerative disorders ( Fan et al . , 2017; Fatokun et al . , 2014 ) , NTR/Mtz-mediated cell ablation may serve as an inducible and titratable methodology for modeling neurodegenerative disease .\",\n \"To interrogate complementary MOA , lead compounds were tested in pairs in mouse photoreceptor cell culture assays and in zebrafish .\",\n \"Intriguingly , enhanced survival effects were common in mouse photoreceptor cell cultures , while additive and even synergistic effects were evident in zebrafish for the majority of pairs tested .\",\n \"Taken together , our results suggest drugs providing neuroprotective effects across diverse RP models , between species , and/or through complementary MOA , will provide promising new therapeutic opportunities for IRD/RP patients .\"\n ],\n [\n \"We initiated our study using a robotics-automated large-scale in vivo screening system to identify compounds that promoted rod photoreceptor survival in a transgenic zebrafish line enabling targeted ablation of rod photoreceptors , Tg ( rho:YFP-Eco . NfsB ) gmc500 , hereafter , rho:YFP-NTR ( Walker et al . , 2012; White et al . , 2017 ) .\",\n \"In this line , a 3 . 7 kb rhodopsin ( rho ) promoter fragment ( Hamaoka et al . , 2002 ) drives transgene expression exclusively in rod photoreceptor cells ( Figure 1A ) .\",\n \"The transgene is a fusion protein linking a yellow fluorescent protein ( YFP ) reporter to a nitroreductase prodrug converting enzyme ( NTR , encoded by the nfsB gene from Escherichia coli ) .\",\n \"NTR expression enables pro-drug inducible targeted cell ablation ( Curado et al . , 2007; Pisharath et al . , 2007 ) .\",\n \"Exposing rho:YFP-NTR fish to the prodrug metronidazole ( Mtz ) leads to the selective death of rod photoreceptors and concomitant loss of YFP ( Figure 1A–C ) , physiologically mimicking the onset of RP ( Hamel , 2006 ) .\",\n \"An immunohistological analysis of rod and cone photoreceptor markers was performed on 7 days post-fertilization ( dpf ) zebrafish retinal sections to test if Mtz-induced ablation was specific to rod cells .\",\n \"In non-ablated controls , rod outer segment labeling was well correlated with YFP expression ( Figure 1—figure supplement 1A , B; -Mtz , arrows ) .\",\n \"In Mtz-treated retinas , rod outer segment labeling and YFP expression were markedly diminished ( Figure 1—figure supplement 1A , B; +Mtz ) .\",\n \"Cone photoreceptor labeling showed no overlap with YFP expression , and cone cell labeling appeared similar in non-ablated and Mtz-treated retinas ( Figure 1—figure supplement 1C ) , suggesting cone cells were not affected by Mtz exposure or by acute rod cell loss .\",\n \"Prior studies suggest NTR/Mtz-mediated ablation occurs by apoptosis ( Chang et al . , 1993; Doonan et al . , 2003; Portera-Cailliau et al . , 1994; Zeiss et al . , 2004 ) and loss of rods in RP has also been linked to apoptosis ( Chang et al . , 1993; Doonan et al . , 2003; Portera-Cailliau et al . , 1994; Zeiss et al . , 2004 ) .\",\n \"Thus , we reasoned rho:YFP-NTR fish could be used to identify compounds that protected rods from apoptosis .\",\n \"To identify neuroprotective compounds with this model ( i . e . drugs that sustain YFP expression after Mtz exposure ) , we used our established plate reader-based ARQiv-HTS assay ( Figure 1D; Walker et al . , 2012; White et al . , 2016 ) .\",\n \"We first determined optimal conditions for inducing rod cell loss while maintaining larval health in a 96-well plate format .\",\n \"Major aspects of retinal cytogenesis are largely complete by five dpf in zebrafish ( Schmitt and Dowling , 1999; Stenkamp , 2011 ) .\",\n \"Reporter expression in rho:YFP-NTR larvae has also stabilized by this time point ( Unal Eroglu et al . , 2018 ) , consistent with rho expression ( Raymond et al . , 1995 ) .\",\n \"We therefore chose 5 dpf to initiate Mtz-induced rod cell ablation .\",\n \"We previously determined that rod cell loss reached a nadir at 7 dpf following a 24 hr pulse of 10 mM Mtz at 5 dpf ( Walker et al . , 2012 ) .\",\n \"We reasoned that a 48 hr Mtz exposure initiating at 5 dpf would maximize the signal window to test for neuroprotective effects .\",\n \"Concluding the experiment by 7 dpf also avoids challenges associated with feeding , as zebrafish can subsist on their yolk sac up to that time point ( Hernandez et al . , 2018; Jardine and Litvak , 2003 ) .\",\n \"However , 10 mM Mtz treatments extending beyond 24 hr become increasingly toxic ( Mathias et al . , 2014 ) and removing Mtz from microtiter plates after a 24 hr pulse could not be easily automated .\",\n \"We therefore sought a 48 hr Mtz treatment regimen sufficient for inducing maximal rod cell loss by 7 dpf that showed no evidence of general toxicity .\",\n \"Five concentrations of Mtz were tested across a twofold dilution series from 10 mM to 625 μM .\",\n \"YFP reporter signals were quantified daily from 5 to 8 dpf using ARQiv .\",\n \"Changes in YFP levels were calculated as percentages normalized to non-ablated YFP controls .\",\n \"The data showed concentration-dependent reductions in YFP with maximal loss observed at 7 dpf ( Figure 1B ) .\",\n \"Mtz exposures at 0 . 625 , 1 . 25 , 2 . 5 , 5 , and 10 mM led to a 55 , 79 , 87 , 87 , and 83% decrease in YFP detection , respectively ( Supplementary file 1a ) .\",\n \"Although no lethality was observed for any condition , signs of distress were evident for 10 mM Mtz exposures ( i . e . reduced motility ) .\",\n \"At ≤5 mM Mtz , however , no signs of stress were observed .\",\n \"As 2 . 5 mM Mtz was the lowest concentration producing maximal loss of YFP detection , and confocal imaging verified YFP loss following a 2-day ( 5–7 dpf ) 2 . 5 mM Mtz treatment ( Figure 1C ) , this condition was selected as the treatment regimen for the large-scale screen .\",\n \"Previously established power analysis methods using ablated and non-ablated controls ( White et al . , 2016 ) determined that a sample size of nine larvae was sufficient to detect a 50% neuroprotective effect .\",\n \"For ease of dispensing , microtiter plate formatting , and to account for larval dispensing errors , we increased the sample size to 16 larvae per condition for the primary screen .\",\n \"We also reasoned that this would also allow us to detect subtler neuroprotective effects .\",\n \"The strictly standardized mean difference quality control ( SSMD QC ) score was 1 . 67 , indicating the assay was of sufficient quality to justify a large-scale screening effort ( Zhang , 2011 ) .\",\n \"To establish a positive control , we tested 17 compounds and one compound ‘cocktail’ previously implicated as neuroprotectants in mammalian RP models ( Supplementary file 2a ) .\",\n \"Unfortunately , none sustained YFP expression at the concentrations tested ( 4 μM to 125 nM ) .\",\n \"However , none have proven effective in RP patients either , suggesting a lack of conservation of neuroprotective effects across species .\",\n \"Our screen was designed to address this issue by testing for neuroprotective effects of these compounds across species and between different RP models .\",\n \"Fortunately , a compound identified as a retinal cell neuroprotectant by the Zack lab ( manuscript in preparation ) did show-dose-dependent effects on YFP levels .\",\n \"This compound was therefore used as a positive control ( POS ) for assay performance throughout the large-scale screen .\",\n \"The Johns Hopkins Drug Library ( JHDL ) was chosen for the large-scale screen .\",\n \"The JHDL is comprised of nearly 3000 compounds , most being human-approved drugs ( Shim and Liu , 2014 ) .\",\n \"To minimize false discovery rates , all compounds were tested using qHTS principles ( Inglese et al . , 2006 ) – that is , across six concentrations ( 4 μM - 125 nM ) using a twofold dilution series .\",\n \"The screen largely followed published ARQiv-HTS methodologies ( Wang et al . , 2015b; White et al . , 2016 ) , with additional assay-specific details ( Figure 1D , steps 1–8 ) .\",\n \"In all , 2934 compounds were screened and more than 350 , 000 transgenic zebrafish larvae evaluated .\",\n \"Real-time data analysis was performed as previously detailed ( White et al . , 2016 ) to generate: ( 1 ) a plot of YFP signal levels , ( 2 ) a plot of SSMD scores across all tested concentrations , ( 3 ) a signal intensity heat map of each plate , and ( 4 ) an SSMD score table ( Figure 1D , step 7 ) .\",\n \"Compounds producing SSMD scores of ≥one were considered potential hits and flagged for visual inspection to assess fluorescence and general morphology using a stereo fluorescence microscope .\",\n \"This step facilitated elimination of false-positive compounds producing aberrant fluorescence due to larval toxicity ( six drugs ) or compound autofluorescence ( 27 drugs; Figure 1C , step 8; Supplementary file 2b ) .\",\n \"Additionally , this allowed visual confirmation of sustained YFP expression within the retina .\",\n \"At the conclusion of the primary screen , 113 compounds were identified as hits ( Supplementary file 2c ) .\",\n \"Hit compounds were classified according to the highest SSMD score achieved across all concentrations tested , and whether concentration-dependent effects were observed .\",\n \"SSMD scores suggested one drug produced a strong effect ( SSMD of 2–3 ) ; three had semi-strong effects ( 1 . 645–2 ) , 21 showed moderate effects ( 1 . 28–1 . 645 ) , and 88 had fairly moderate effects ( 1–1 . 28 ) ( Supplementary file 2c ) .\",\n \"Forty-two drugs showed concentration-dependent effects , while 71 exhibited discontinuous or singular concentration effects .\",\n \"‘On label’ mechanism of action ( MOA ) information available for the 113 hits implicated more than 50 targets/pathways in rod cell neuroprotection ( Supplementary file 2d ) .\",\n \"We next performed a series of confirmatory and orthogonal assays to evaluate a subset of 42 hit compounds prioritized by SSMD score , dose-response profile , and/or implicated MOA ( Supplementary file 2c , highlighted compounds ) .\",\n \"Having extensive MOA data is a key advantage of testing human-approved compounds which we leveraged in a previous large-scale zebrafish PDD screen ( Wang et al . , 2015b ) .\",\n \"Similarly here , as studies have suggested inflammation plays a key role in retinal degeneration and regeneration ( Hollyfield et al . , 2008; Mitchell et al . , 2018; White et al . , 2017; Yoshida et al . , 2013 ) , hits implicated as modulators of inflammatory signaling were included as part of the prioritization scheme .\",\n \"In addition , several compounds that did not produce concentration-dependent effects were selected to test whether this criterion was useful in predicting reproducibility .\",\n \"All compounds were obtained from new sources to ensure reagent authenticity .\",\n \"To confirm survival promoting effects , three repeats of the primary screening assay ( Figure 2A ) were conducted , but using a wider concentration range to account for differences in reagent quality ( from 100 μM to 1 . 28 nM , fivefold dilution series ) .\",\n \"If toxicity was observed at higher concentrations , dilution series were initiated at 10 µM or 1 µM .\",\n \"Using this strategy , 11 of the 42 prioritized hit compounds were confirmed as lead candidates ( 26% ) .\",\n \"Estimated effects on rod cell survival – that is , YFP signal levels relative to +Mtz/0 . 1% DMSO controls – ranged from 38 to 9% ( Figure 2B; lead drug candidate names , abbreviations , PubChem CID , and chemical structures are provided in Table 1 ) .\",\n \"We next asked whether there was a correlation between SSMD scores and/or concentration-dependent effects and confirmation rates .\",\n \"Among 19 selected hit compounds with higher SSMD scores ( ≥1 . 3 ) , seven ( 37% ) were confirmed; among 23 with lower SSMD scores ( 1–1 . 28 ) , four ( 17% ) were confirmed .\",\n \"Of 27 selected hit compounds with a concentration-dependent trend , eight ( 30% ) were confirmed .\",\n \"Conversely , of 15 compounds that did not show a concentration-dependent trend , 3 ( 20% ) were confirmed .\",\n \"Among the 11 confirmed leads , 8 ( 73% ) showed dose-dependent effects and 7 ( 64% ) had higher SSMD scores ( >1 . 28 ) .\",\n \"These results suggest that prioritizing hit compounds by both relative SSMD scores and dose-dependent trends provides predictive value for confirming activity , consistent with qHTS principles ( Inglese et al . , 2006 ) .\",\n \"As rod cell death is induced by NTR-mediated reduction of the prodrug Mtz in our model , it is possible that some lead compounds simply suppressed NTR enzymatic activity .\",\n \"To test this , NTR activity was evaluated in the presence of each lead compound by assaying the reduction kinetics of the prodrug CB1954 in vitro ( Prosser et al . , 2010 ) .\",\n \"To ensure any potential for NTR inhibition was accounted for , all compounds were tested at 300 μM ( except MIC and CHL which precipitated at higher concentrations and were tested at 50 µM ) that is , ~100-fold greater than neuroprotective concentrations .\",\n \"Compounds were deemed potential inhibitors if NTR activity was less than 75% of the NTR alone control .\",\n \"Seven compounds showed no evidence of NTR inhibition by this criterion , but four did: warfarin ( WAR ) , ciclopirox olamine ( CPO ) , calcimycin ( CAL ) , and sulindac ( SUL; Figure 2—figure supplement 1A , B ) .\",\n \"However , IC50 measures ranged from 150 µM ( for CPO ) to 350 µM ( for SUL , Figure 2—figure supplement 1B ) , approximately one to two orders of magnitude higher than observed neuroprotective concentrations ( i . e . 0 . 4–20 µM ) .\",\n \"When NTR reduction of Mtz was tested for WAR , CPO , and CAL at 300 μM , similarly weak inhibitory effects were observed ( Figure 2—figure supplement 1B ) .\",\n \"The differences in concentrations between neuroprotective and NTR inhibitory activities diminish the likelihood that these leads act directly on NTR .\",\n \"To test this further , lead compounds were assayed for neuroprotective effects in NTR-independent mouse RP models ( see below ) .\",\n \"To control for the possibility that lead candidates promoted rod photoreceptor development , rather than neuroprotection , YFP levels were quantified in rho:YFP-NTR larvae exposed solely to lead drugs from 5 to 7 dpf ( Figure 2—figure supplement 2A ) .\",\n \"Retinoic acid ( RA , 1 . 25 μM ) was used as a positive control as it promotes rod fates during development in zebrafish ( Hyatt et al . , 1996 ) .\",\n \"RA-treated fish displayed statistically significant increases in YFP levels compared to untreated controls ( Figure 2—figure supplement 2B ) .\",\n \"In contrast , none of the retinas treated with lead compounds exhibited increased YFP expression , suggesting they do not promote rod photoreceptor cell fate .\",\n \"In fact , three lead compounds cloxyquin ( CLO ) , cortexolone ( COR ) and CPO reduced YFP signals relative to the untreated control ( Figure 2—figure supplement 2B , asterisks ) , suggesting negative effects on rod cell development .\",\n \"It is well known that the zebrafish retina regenerates ( Gorsuch and Hyde , 2014; Lenkowski and Raymond , 2014; Wan and Goldman , 2016 ) .\",\n \"Therefore , to determine whether lead compounds acted by stimulating regeneration , we used a previously described ARQiv assay designed to detect changes in rod cell replacement kinetics ( Walker et al . , 2012; White et al . , 2017 ) .\",\n \"Briefly , rho:YFP-NTR larvae were first treated with 10 mM Mtz at 5 dpf for 24 hr to induce rod cell loss .\",\n \"At 6 dpf , Mtz was washed out and larvae were treated with lead compounds at concentrations corresponding to maximal neuroprotective effects and YFP levels quantified at 9 dpf ( Figure 2—figure supplement 3A ) .\",\n \"Dexamethasone , which accelerates rod cell regeneration kinetics ( White et al . , 2017 ) , was used as a positive control .\",\n \"The results showed that none of the compounds increased rod cell regeneration rates ( Figure 2—figure supplement 3B ) , while four compounds , dihydroartemisinin ( DHA ) , CLO , CPO and MIC inhibited regeneration .\",\n \"These data suggest that lead compounds do not increase YFP levels by promoting rod cell regeneration .\",\n \"To test if lead compounds simply increased YFP signal intensity , rather than promoted rod cell survival , intravital microscopy was used to image Mtz-treated retinas ±lead compounds and non-ablated ( -Mtz ) controls .\",\n \"Mtz-treated retinas had reduced numbers of YFP-NTR-expressing rods ( Figure 3 , +Mtz ) .\",\n \"Conversely , rods in control retinas displayed robust YFP signal throughout the retina and elongated morphologies suggestive of healthy outer segments ( Figure 3 , -Mtz ) .\",\n \"In retinas exposed to Mtz and lead compounds , YFP signal loss was attenuated and rods typically displayed elongated morphologies ( Figure 3 ) .\",\n \"However , for some compounds , cells appeared rounded , suggesting degeneration was not fully inhibited ( e . g . miconazole , MIC; Figure 3 ) .\",\n \"To confirm lead compounds increased rod cell survival , confocal stacks of YFP-expressing rods were 3D-rendered and fluorescence volume quantified using Imaris software-based automated image analysis ( White et al . , 2017 ) .\",\n \"The data showed a statistically significant increase in YFP volumes for all drug-treated groups ( Figure 3—figure supplement 1 ) , confirming that lead compounds promoted increased rod cell numbers and/or preserved outer segment morphology .\",\n \"To further investigate if lead compounds preserved rod outer segments , we used a newly developed zebrafish transgenic line , Tg ( rho:GAP-YFP-2A-nfsB_Vv F70A/F108Y ) jh405 ( hereafter rho:YFP-NTR2 . 0 ) .\",\n \"In this line , rods co-express membrane-tagged YFP and an improved NTR ( ‘NTR 2 . 0’ ) .\",\n \"The membrane-tagged YFP facilitates improved imaging of rod outer segments , while NTR 2 . 0 enables cell ablation at a reduced concentration of Mtz ( e . g . 10 µM versus 2 . 5 mM; Sharrock et al . , 2020 ) .\",\n \"To assess if lead compounds preserved outer segment morphologies , intravital confocal imaging was performed as above but using the rho:YFP-NTR2 . 0 line and 10 µM Mtz treatments .\",\n \"Qualitative image analysis suggested the majority , six of seven tested lead compounds , preserved rod outer segments ( Figure 3—figure supplement 2A ) .\",\n \"Reasoning that preservation of outer segments would equate to increased rod cell size , relative to rounded morphologies of dead or stressed cells , average rod cell sizes were calculated as the total YFP volume divided by the number of rod cells per each 3D confocal stack ( i . e . average YFP volume per cell ) .\",\n \"Compared to rod ablated controls , five of seven lead compounds tested promoted statistically significant increases in rod cell volumes ( Figure 3—figure supplement 2B ) .\",\n \"These data suggest some lead compounds were able to maintain rod outer segment morphology , thus potentially maintaining rod cell function .\",\n \"To identify new therapeutics for RP patients , we reasoned compounds producing neuroprotective effects across fish and mammalian RP models would likely target conserved mechanisms , and thus be more likely to translate successfully in the clinic .\",\n \"We therefore tested the efficacy of lead compounds in a series of mouse models of retinal photoreceptor degeneration .\",\n \"First , we tested compounds for the capacity to protect mouse primary photoreceptor cells from stressor-induced cell death in culture .\",\n \"Photoreceptors were isolated from postnatal day four ( P4 ) QRX mice , a transgenic line in which GFP expression is restricted to photoreceptors ( Wang et al . , 2004 ) , and grown as previously described ( Fuller et al . , 2014 ) .\",\n \"To induce photoreceptor cell death , thapsigargin ( 0 . 25 µM ) or tunicamycin ( 0 . 6 µg/mL ) were added .\",\n \"These ‘stressor’ compounds deplete endoplasmic reticulum ( ER ) calcium levels ( Thastrup et al . , 1990 ) and inhibit protein glycosylation ( Fliesler et al . , 1984 ) , respectively ( Lai et al . , 2007b ) .\",\n \"In turn , they elicit an unfolded protein response ( UPR ) ( Wang et al . , 2015b ) and related ER stress ( Oslowski and Urano , 2011; Zhang et al . , 2014 ) , both implicated in the etiology of RP ( Griciuc et al . , 2011; Rana et al . , 2014 ) .\",\n \"To test for survival effects , all eleven lead compounds were screened at seven concentrations across a threefold dilution series ( from 30 µM to 40 nM ) .\",\n \"After 48 hr in stressor ± lead compounds , cells were imaged using an automated high-content screening system .\",\n \"Photoreceptor survival was assessed by automated quantification of GFP-expressing cells ( Figure 4A ) .\",\n \"Nine of 11 lead compounds , proved protective in the thapsigargin-induced cell death assay ( Figure 4B ) .\",\n \"MIC , CLO and SUL also protected photoreceptors from tunicamycin-induced cell death ( Figure 4—figure supplement 1 ) .\",\n \"Thus , nine of 11 lead compounds promoted the survival of mouse primary photoreceptor cells in at least one stressor-induced cell death assay , and three were neuroprotective in both assays .\",\n \"We next tested lead compounds for survival effects in rd1 mouse retinal explant cultures .\",\n \"The rd1 mouse model of RP exhibits early onset rod cell degeneration caused by a mutation in the Pde6b gene ( Danciger et al . , 1990; Pittler and Baehr , 1991; Sidman and Green , 1965 ) , which is an ortholog of the human RP-associated PDE6B ( Khramtsov et al . , 1993; McLaughlin et al . , 1993 ) .\",\n \"In rd1 mice , photoreceptor degeneration begins around P10 and progresses rapidly .\",\n \"By P21 , only a single row of photoreceptor cells remain in the outer nuclear layer ( ONL; Tansley , 1951 ) making it an excellent system for screening potential neuroprotectants ( Beeson et al . , 2016 ) .\",\n \"Here , retinal explants from P10 rd1 mice were isolated and cultured ex vivo ( Bandyopadhyay and Rohrer , 2010 ) .\",\n \"Eight lead compounds , all but CAL , SUL , and Zinc pyrithione ( ZPT ) , were tested for survival effects at three concentrations across a fivefold dilution series centered on the concentration most effective in fish RP models .\",\n \"After 11 days in culture , explants were fixed , stained and the number of photoreceptor rows counted ( Figure 5A ) .\",\n \"Neuroprotective effects were defined as a concentration-dependent increase in the number of photoreceptor rows remaining in the ONL relative to untreated controls ( p≤0 . 05 ) .\",\n \"An average of 1 . 2 ±0 . 19 rows of photoreceptors remained in the ONL of control explants .\",\n \"Three lead drugs , CPO , DHA , and ART , increased the number of surviving photoreceptor layers ( Figure 5B ) .\",\n \"Representative images of control ( DMSO and POS ) and lead-treated explants demonstrate photoreceptor layer protection ( Figure 5C ) .\",\n \"CPO , DHA , and ART thus promoted rod cell survival in the fish RP model and two mouse cell culture RP models .\",\n \"However , high concentration CPO treatments ( 15 µM ) led to disruption of retinal histology due to induction of proliferation in the inner nuclear layer ( INL ) and ONL .\",\n \"Therefore , as DHA is the active metabolite of ART-related compounds , we proceeded with testing DHA in an in vivo mouse RP model .\",\n \"To test the potential of DHA as a ‘pan-disease’ therapeutic , that is , whether it was effective across multiple genetic RP models , it was tested for photoreceptor survival effects in the rd10 mouse model of RP ( Pde6brd10 ) .\",\n \"The rd10 line was selected for in vivo experiments due to its slower rate of photoreceptor degeneration relative to other rd mutants , thus allowing a prolonged window for pharmacological intervention .\",\n \"DHA was selected based on its superior performance in rd1 retinal explant assays ( Figure 5 ) and because it is the active metabolite of a second lead drug candidate , ART , recently shown to protect photoreceptors from light damage ( Lu and Xie , 2019 ) .\",\n \"To create a long-release formulation , DHA was encapsulated in PLGA polymers ( Poly ( D , L-lactide-co-glycolide ) ) .\",\n \"Release kinetics assays in PBS/0 . 1% DMSO in vitro suggested DHA would reach maximal concentrations after 30 days , and remain stable for at least 20 days thereafter ( Figure 6A ) .\",\n \"We noted that release was minimal in the absence of DMSO , but were reluctant to include DMSO for injections due to the potential for deleterious effects on the retina ( Tsai et al . , 2009 ) .\",\n \"PLGA-DHA was injected into the vitreous of one eye of rd10 mice at P14 , with the contralateral eye serving as a vehicle injection control .\",\n \"At P32 , 18 days after injection , and when DHA levels were predicted to reach 80% of maximal concentration based on in vitro releases kinetics , rd10 mouse eyes were processed for immunohistochemistry ( Figure 6B ) and ONL thickness quantified ( Figure 6C ) .\",\n \"Despite initial promising results in pilot assays , no reproducible neuroprotective effects were observed .\",\n \"It is possible that in the absence of DMSO release kinetics may have been limiting in vivo .\",\n \"We plan to address this possibility in future studies .\",\n \"Previously , we used ‘on label’ information to explore MOA of hit compounds identified during an in vivo phenotypic screen of the JHDL ( Wang et al . , 2015b ) .\",\n \"Recently , we have become interested in additional advantages afforded by HTS-based MOA data and whole-organism phenotypic screening , such as polypharmacology ( Dar et al . , 2012; Rennekamp and Peterson , 2015; Rihel et al . , 2010 ) .\",\n \"Accordingly , we applied a target-agnostic MOA analysis process by evaluating lead compound performance in HTS and ultra-HTS studies archived on PubChem ( https://pubchem . ncbi . nlm . nih . gov/ ) .\",\n \"Many lead compounds exhibited shared target/pathway activities , suggesting in-common MOA ( Table 2 ) .\",\n \"Among shared targets , Tyrosyl-DNA Phosphodiesterase 1 ( TDP1 ) , a DNA repair enzyme , was the most popular ( eight leads ) .\",\n \"To test whether TDP1 inhibition promoted rod photoreceptor survival , we tested three TDP1 inhibitors in rod:YFP-NTR fish: paromomycin ( PM ) , thiostrepton ( ThS ) , and methyl-3 , 4-dephostatin ( MD ) ( Huang et al . , 2011; Liao et al . , 2006 ) .\",\n \"Both PM and ThS showed neuroprotective effects ( Table 3 , Figure 7 ) .\",\n \"This result was surprising given that NTR reduction of Mtz is thought to cause DNA damage-induced cell death ( Curado et al . , 2007 ) .\",\n \"Thus , inhibition of a DNA repair enzyme would be expected to enhance , not inhibit , NTR/Mtz-mediated cell death .\",\n \"However , an alternative means of disrupting TDP1 activity is by inhibiting Poly ( ADP-ribose ) Polymerases ( PARPs ) .\",\n \"Indeed , in a comprehensive 400 , 000 compound qHTS assay designed to identify TDP1 inhibitors , all five hits turned out to be PARP inhibitors ( Murai et al . , 2014 ) .\",\n \"PARPs also mediate DNA repair but , interestingly , hyperactivation of PARP1 leads to a specific form of DNA damage-induced cell death , termed parthanatos ( Fatokun et al . , 2014; Wang et al . , 2016 ) .\",\n \"PARP is also involved in a cGMP-dependent cell death pathway implicated in RP ( Power et al . , 2020 ) .\",\n \"We therefore assayed PARP inhibitors for the capacity to promote rod cell survival in rod:YFP-NTR fish .\",\n \"All eight PARP inhibitors tested had neuroprotective activity , ranging from 9% to 20% ( Figure 7 ) .\",\n \"To account for other cell death mechanisms implicated in neurodegeneration , we tested an inhibitor of necroptosis ( necrostatin-1; NEC ) , and three inhibitors of apoptosis ( Ac-DEVD-CHO , CASI , and CASVII , Table 3 ) .\",\n \"Surprisingly , NEC promoted rod cell survival , while the apoptosis inhibitors did not ( Figure 7 , Table 3 ) .\",\n \"Interestingly , paired testing of PARP and necroptosis inhibitors resulted in additive effects ( Figure 7—figure supplement 1 , Supplementary file 1b ) , suggesting either differential responses to NTR/Mtz-induced DNA damage among rod cells , or that inhibiting both pathways keeps cells from activating the other mechanism of cell death when only one pathway is blocked .\",\n \"To further dissect cell death pathways mediating NTR/Mtz-induced rod cell ablation , we used a ‘redundant’ CRISPR/Cas9 gene targeting approach ( i . e . four guide RNA targeting ) , which enables phenotyping in injected zebrafish embryos/larvae ( Wu et al . , 2018 ) .\",\n \"Key factors for apoptosis ( casp3a and casp3b ) , PARP-dependent cell death pathways parthanatos and/or cGMP-dependent cell death ( parp1 ) , and necroptosis ( ripk1l ) were targeted .\",\n \"In addition , tdp1 was targeted as a DNA repair control .\",\n \"For each targeted gene , four gRNAs and Cas9 protein were co-injected into one-cell stage embryos .\",\n \"Injected and uninjected ( control ) larvae were treated ±2 . 5 mM Mtz at five dpf and rod-YFP levels quantified at six dpf by ARQiv .\",\n \"For Mtz-ablated larvae , knockdown of parp1 markedly increased rod cell survival ( 39% compared to 20% in the uninjected control , p=4E-12 ) , and ripk1l knockdown modestly protected rods ( 26% compared to 22% in the uninjected control , p=0 . 05 ) .\",\n \"However , no improvement in survival was observed for either casp3a/b or tdp1 knockdown ( Figure 8A ) .\",\n \"In no Mtz controls , knockdown of parp1 , ripk1l or tdp1 had no effect on rod cell numbers .\",\n \"Conversely , casp3a/3b double knockdown increased rod cell numbers ( Figure 8B ) suggesting inhibition of developmental apoptosis and confirming efficacy of casp3a/3b knockdown .\",\n \"Confocal imaging confirmed parp1 knockdown effects on rod cell survival and casp3a/3b knockdown effects on rod cell development ( Figure 8—figure supplement 1A ) .\",\n \"Reduced gene expression was verified by qPCR ( Figure 8—figure supplement 1B , Supplementary file 3 ) .\",\n \"These results are consistent with the NTR/Mtz system eliciting DNA damage-associated cell death mediated by parp1 , that is , parthanatos , in rod cells .\",\n \"To further test for activation of PARP signaling during NTR/Mtz-induced rod cell death , accumulation of polymerized poly ( ADP-ribose ) ( PAR ) , a downstream effector of PARP1 activation ( Kam et al . , 2018 ) , was analyzed by western blot .\",\n \"five dpf rho:YFP-NTR larvae were treated ±2 . 5 mM Mtz and protein samples collected from ~30 fish at 3 , 6 , 12 , 24 , and 48 hr post-Mtz treatment initiation .\",\n \"Western blots showed clear accumulation of PAR in Mtz-treated fish at 24 and 48 hr post-Mtz ( Figure 8—figure supplement 1C ) which was verified quantitatively ( Figure 8—figure supplement 1D ) .\",\n \"Collectively , results of cell death pathway analyses are concordant across chemical inhibition , genetic knockdown , and biochemical assays , strongly suggesting NTR/Mtz-induced cell ablation is mediated by PARP-dependent cell death ( e . g . parthanatos , cGMP-dependent cell death ) , a target/pathway implicated across multiple mammalian models of RP ( Arango-Gonzalez et al . , 2014 ) .\",\n \"To determine if lead compound effects were mediated by inhibition of PARP signaling , parp1 expression was knocked down prior to testing for effects on rod cell survival in rho:YFP-NTR larvae .\",\n \"Four leads predicted to be PARP inhibitors ( MIC , CLO , CPO , DHA ) , one predicted to be a non-PARP inhibitor ( WAR ) , and a known PARP inhibitor control ( BMN ) were tested .\",\n \"Two lead compounds , MIC and DHA and BMN ( the PARP inhibitor control ) produced no statistically significant enhanced neuroprotective effects over parp1 knockdown .\",\n \"Conversely , WAR , CLO and CPO increased rod cell survival in parp1 knockdown fish ( Figure 9A ) .\",\n \"qPCR confirmed mRNA knockdown ( Figure 9B ) and controls showed no effects of parp1 knockdown on rod cell development ( Figure 9C , -Mtz ) and enhanced rod cell survival upon Mtz treatment ( Figure 9C , +Mtz ) .\",\n \"These data confirm that PARP inhibition mediates the neuroprotective effects of some lead compounds ( e . g . MIC and DHA ) while others are PARP-independent .\",\n \"The latter result suggests lead compounds may act via complementary neuroprotective mechanisms .\",\n \"Combined , these findings are consistent with rod cell loss in our fish RP model occurring , at least partially , by PARP-dependent cell death .\",\n \"PubChem data also suggested potential complementary MOA , that is , multiple independent targets across lead compounds ( Table 2 ) .\",\n \"We therefore hypothesized that combining lead compounds may produce enhanced , additive , or even synergistic effects .\",\n \"To investigate this , combinatorial lead drug treatments were tested in primary mouse photoreceptor cell cultures .\",\n \"As above , photoreceptors were isolated from QRX mouse retinas and used in stressor-treated primary cell culture assays .\",\n \"All 11 lead compounds were tested at a selected concentration individually and in pairs and the survival of GFP-labeled photoreceptors analyzed by high-content imaging and automated cell counting .\",\n \"The results showed no wholly additive effects , that is , survival percentages equaling or excelling the summed effect of individual compound assays .\",\n \"However , enhanced effects – i . e . , better median survival than for either compound alone – were observed for 30 of 55 pairs in thapsigargin-treated cultures and for 20 of 55 pairs in tunicamycin-treated cultures ( Figure 4—figure supplement 2 ) .\",\n \"Finally , seven lead compounds were tested alone and in pairs in rho:YFP-NTR zebrafish using optimal effective concentrations .\",\n \"Pairs producing survival effects equal to or greater than the sum of their individual values ( ±10% ) were considered additive .\",\n \"By this criterion , 10 of 19 viable pairs ( two pairs proved lethal ) exhibited additive effects ( Figure 10 , bolded; Supplementary file 1c ) and three pairs produced supra-additive effects ( i . e . ≥25% greater than summed effects ) suggesting possible synergy ( Figure 10 , asterisks ) .\",\n \"The maximal average rod cell survival effect was 58% ( WAR + CPO ) .\",\n \"Several compounds showed broadly additive effects , for example , WAR was additive with all six drugs and COR with four of five pairs ( one pair proving lethal ) .\",\n \"Additive effects suggest multiple signaling pathways are involved in NTR/Mtz-induced photoreceptor degeneration , thus that combinatorial drug regimens may be required to achieve maximal therapeutic benefits for RP patients .\"\n ],\n [\n \"Identifying effective neuroprotective therapies for RP and other IRDs stands as a critical unmet need for the field ( Duncan et al . , 2018; Wubben et al . , 2019 ) .\",\n \"Although , neurotrophic factors , anti-apoptotic agents , nutritional supplements , and antioxidants have shown neuroprotective effects in animal models of RP ( Dias et al . , 2018 ) .\",\n \"Unfortunately , these reagents have produced , at best , only limited benefits for patients to date and , for some , mild improvements are offset by adverse side effects associated with long-term use ( Dias et al . , 2018 ) .\",\n \"For example , ciliary neurotrophic factor ( CNTF ) was shown to be effective in protecting photoreceptors in mouse ( Cayouette et al . , 1998 ) , dog ( Tao et al . , 2002 ) , and chicken ( Fuhrmann et al . , 2003 ) models of retinal degeneration .\",\n \"However , CNTF failed to improve either visual acuity or field sensitivity in short- and long-term RP clinical trials ( Birch et al . , 2016; Ciliary Neurotrophic Factor Retinitis Pigmentosa Study Groups et al . , 2013; Argus II Study Group et al . , 2015 ) .\",\n \"Clinical trials of Vitamin A in combination with Vitamin E ( Berson et al . , 1993 ) , docosahexaenoic acid ( Berson et al . , 2004 ) , lutein ( Berson et al . , 2010 ) , or valproic acid ( Birch et al . , 2018 ) were reported to produce some benefits for RP patients , but only in subpopulations , and some of these studies have been controversial ( Massof and Finkelstein , 1993 ) .\",\n \"Our strategy for addressing this challenge was to: ( 1 ) scale up the number of compounds tested directly in complex living disease models and ( 2 ) perform a cross-species screening cascade that starts with small animal models amenable to HTS and proceeds to mammalian models .\",\n \"We hypothesize that compounds producing beneficial outcomes across evolutionarily diverse species will target conserved MOA and thus stand a higher chance of successful translation .\",\n \"As a generalized strategy , large-scale drug discovery screens using small animal models are showing increasing promise across multiple disease paradigms ( Cagan et al . , 2019; Cully , 2019; Kitcher et al . , 2019; MacRae and Peterson , 2015 ) .\",\n \"Here , using a large-scale in vivo drug screening platform ( Figure 1 ) , we tested 2934 largely human-approved compounds for neuroprotective effects across six concentrations in >350 , 000 larval zebrafish models of RP .\",\n \"The primary screen implicated 113 compounds as neuroprotectants ( Supplementary file 2c ) .\",\n \"Confirmatory repeats and a series of four orthogonal assays validated 11 of 42 prioritized hit compounds in protecting zebrafish rod photoreceptors from cell death ( Figures 2 and 3 , Figure 2—figure supplements 1–3 , Figure 3—figure supplement 1 and Table 1 ) .\",\n \"Importantly , investigations of lead compound MOA , led to the discovery that NTR/Mtz-mediated rod cell death appears to proceed through alternative cell death pathways ( Figure 8 ) recently linked to photoreceptor degeneration in mammalian models of RP ( Arango-Gonzalez et al . , 2014; Paquet-Durand et al . , 2007; Power et al . , 2020; Sancho-Pelluz et al . , 2008; Tolone et al . , 2019 ) .\",\n \"A summary schematic of the entire screening cascade is provided in Figure 11 .\",\n \"Confirmatory tests showed survival effects of lead compounds varied from 9 to 38% ( Figure 2 ) , which might suggest limited therapeutic potential .\",\n \"However , even small numbers of surviving rod cells can have a significant impact on cone photoreceptor survival ( Guadagni et al . , 2015; Hartong et al . , 2006; Punzo et al . , 2012 ) .\",\n \"This effect is mediated by the rod-derived cone viability factor ( RdCVF ) ( Léveillard et al . , 2004 ) which stimulates glucose metabolism to promote cone survival ( Aït-Ali et al . , 2015 ) .\",\n \"Therefore , as cone cells provide high-acuity daytime vision , protecting small numbers of rod cell has significant therapeutic potential for RP patients .\",\n \"To test this further , lead compounds will need to be tested for the ability to prolong cone survival and function in rod-cone degeneration models .\",\n \"Visualization by in vivo confocal imaging showed variations in surviving cell morphologies across lead compounds ( Figure 3 and Figure 3—figure supplement 2 ) , suggesting .\",\n \"variations in the extent of rod cell protection that could impact function .\",\n \"In particular , maintenance of the outer segment , a specialized primary cilium in which phototransduction occurs , is necessary for photoreceptor function .\",\n \"Volumetric quantification of surviving rod cells further suggested lead compounds differ in their capacity to maintain outer segments .\",\n \"Visual function tests will need to be performed to test this possibility and thus to prioritize lead compounds for in vivo testing in mammalian RP models .\",\n \"To test conservation of neuroprotective effects across species , lead compounds were evaluated in three mouse IRD/RP models .\",\n \"Nine of 11 leads were confirmed as neuroprotectants in primary photoreceptor cell cultures ( Figure 4 ) .\",\n \"Three of eight leads assayed using rd1 retinal explant cultures were also validated ( Figure 5 ) ; two being active in both paradigms , DHA and ART .\",\n \"We chose the rd10 RP model for in vivo testing because it undergoes a slower rate of rod cell loss than rd1; spanning from approximately P16 to P35 ( Chang et al . , 2007; Gargini et al . , 2007 ) .\",\n \"DHA was the most promising compound for these tests as it had shown strong effects across assays and was amenable to a long-term release formulation designed to sustain drug action over weeks to months ( PLGA-DHA , Figure 6A ) .\",\n \"In addition , DHA is the active metabolite of artemisinin , another of our cross-species confirmed leads that was shown to have neuroprotective activity in rat models of stress-induced neuronal damage and light-induced photoreceptor degeneration ( Yan et al . , 2017 ) .\",\n \"Unfortunately , we did not observe increased photoreceptor survival in rd10 retinas injected with PLGA encapsulated DHA .\",\n \"A potential issue that needs to be resolved is the release kinetics of encapsulated DHA in vivo .\",\n \"DHA has poor water solubility , therefore in vitro release kinetics were obtained in the presence of 0 . 1% DMSO .\",\n \"However , because intravitreal injection of DMSO can have deleterious effects ( Tsai et al . , 2009 ) we avoided this potential complication .\",\n \"To assess DHA release in vivo , pharmacokinetic assays such as high-performance liquid chromatography ( Ayalasomayajula and Kompella , 2004; Shen et al . , 2014 ) will be used to quantify lead compound concentrations in the retina for future tests .\",\n \"Another possible explanation is that rd1 and rd10 models have differential responses to DHA .\",\n \"This has been reported for the histone-deacetylase inhibitor valproic acid , which shows opposing effects in rd1 ( neuroprotective ) and rd10 ( deleterious ) mice ( Mitton et al . , 2014 ) and across four different frog models of RP ( Vent-Schmidt et al . , 2017 ) .\",\n \"In addition valproic acid has produced inconsistent results in clinical trials with RP patients ( Chen et al . , 2019; Todd and Zelinka , 2017; Totan et al . , 2017 ) .\",\n \"These results emphasize the need for a more thorough understanding of IRD and RP disease mechanisms and downstream cell pathways to support the development of both personalized and pan-disease therapeutics .\",\n \"Numerous IRD/RP-linked mutations have been identified ( Dias et al . , 2018; https://sph . uth . edu/retnet/ ) implicating an array of disease mechanisms ( Dharmat et al . , 2020 ) .\",\n \"However , cell death pathways common across different IRD/RP patient subpopulations may provide pan-disease targets for neuroprotective therapies .\",\n \"Apoptosis has long been thought to be the mechanism by which rod photoreceptors die in IRD/RP ( Chang et al . , 1993; Doonan et al . , 2003; Portera-Cailliau et al . , 1994; Zeiss et al . , 2004 ) .\",\n \"However , these reports relied on terminal deoxynucleotidyl transferase dUTP nick end labeling ( TUNEL ) , which does not distinguish apoptosis from other types of cell death ( Ansari et al . , 1993; Charriaut-Marlangue and Ben-Ari , 1995; Grasl-Kraupp et al . , 1995; Dmitrieva and Burg , 2007; Kanoh et al . , 1999; Nishiyama et al . , 1996 ) .\",\n \"More recent evaluations of multiple apoptosis-related markers ( e . g . BAX , cytochrome c , caspase-9 , cleaved caspase-3 ) suggest apoptosis occurs in only a minority of RP models ( Arango-Gonzalez et al . , 2014; Sancho-Pelluz et al . , 2008 ) .\",\n \"Moreover , inhibition of apoptosis does not block cell death in many mouse photoreceptor degeneration models ( Hamann et al . , 2009; Yoshizawa et al . , 2002 ) , suggesting other pathways may mediate rod and/or cone cell death in retinal degenerative disease .\",\n \"Recently , non-apoptotic cell death mechanisms have been implicated in IRD/RP .\",\n \"In a comprehensive biochemical analysis of 10 mammalian RP models—involving mutations in cnga3 , cngb1 , pde6a , pde6b , pde6c , prph2 , rho , and rpe65—non-apoptotic cell death signatures were found to be common across all models tested ( Arango-Gonzalez et al . , 2014 ) .\",\n \"Conversely , definitive apoptotic markers were found only for the S334ter ( rho ) rat model .\",\n \"Shared features included activation of poly ( ADP-ribose ) polymerase ( PARP ) , histone deacetylase ( HDAC ) , and calpain , as well as accumulation of cyclic guanosine monophosphate ( cGMP ) and poly ( ADP-ribose ) ( PAR ) .\",\n \"For PARP , chemical inhibitors and a knock out line provided further confirmation ( Jiao et al . , 2016; Paquet-Durand et al . , 2007; Sahaboglu et al . , 2017; Sahaboglu et al . , 2016; Sahaboglu et al . , 2010 ) .\",\n \"Prior reports had suggested the NTR/Mtz system elicits caspase-3 activation and apoptotic cell death ( Chen et al . , 2011 ) .\",\n \"Initially , we had used this as a rationale for pursuing a neuroprotective screen with the rho:YFP-NTR line- however , the recent reports outlined above suggested apoptosis may have limited relevance to IRD/RP .\",\n \"Interestingly , when we tested cell death processes implicated in photoreceptor degeneration directly , an inhibitor of apoptosis ( BEL ) did not promote rod cell survival , whereas inhibition of necroptosis ( NEC ) and PARP were neuroprotective in our fish RP model ( Figure 7 , Table 3 ) .\",\n \"Serendipitously , an exploration of shared lead compound MOA helped to clarify cell death mechanism ( s ) mediating NTR/Mtz-induced rod cell ablation .\",\n \"To explore molecular MOA of our lead compounds , we searched bioactivity data from prior HTS and ultra HTS assays ( PubChem ) .\",\n \"The results suggested both shared and independent MOA .\",\n \"The most common shared target was TDP1 ( eight of 11 lead compounds; Table 2 ) .\",\n \"An initial test confirmed two of three TDP1 inhibitors tested , though neuroprotective activity was relatively weak ( ~10% survival; Figure 7 ) .\",\n \"TDP1 is a DNA repair enzyme that repairs topoisomerase I-induced DNA damage ( Dexheimer et al . , 2008; El-Khamisy , 2011 ) .\",\n \"Interestingly , a qHTS cell-based screen of 400 , 000 compounds for inhibitors of human TDP1 found that all five confirmed compounds actually inhibited PARP activity not TDP1 ( Murai et al . , 2014 ) .\",\n \"This is consistent with findings showing that TDP1 acts in conjunction with PARP1 ( Das et al . , 2014; Lebedeva et al . , 2015 ) , thus PARP inhibition can indirectly affect TDP1 activity .\",\n \"Combined with the results discussed above , we were motivated to test whether PARP inhibition was protective against NTR/Mtz-induced rod cell death .\",\n \"All eight PARP inhibitors tested promoted rod cell survival in our fish RP model ( Figure 7 and Table 3 ) .\",\n \"CRISPR/Cas9-based knockdown of parp1 resulted in a statistically significant reduction of Mtz-induced rod cell death ( Figure 8A ) .\",\n \"Moreover , PAR polymer , an indicator of PARP activation , accumulated in Mtz-treated fish ( Figure 8—figure supplement 1C–D ) .\",\n \"Collectively , these results suggest PARP activation plays an important role in NTR/Mtz-induced rod cell death .\",\n \"PARP1 overactivation initiates a caspase-independent form of DNA damage-induced cell death , termed parthanatos ( Fan et al . , 2017 ) .\",\n \"Parthanatos has been strongly implicated in the etiology of Parkinson’s disease ( Kam et al . , 2018 ) and in a variety of neurodegenerative conditions as well ( Fan et al . , 2017; Fatokun et al . , 2014 ) , including RP/IRDs ( Power et al . , 2020 ) .\",\n \"Importantly , PARP is also a key component of the cGMP-dependent cell death pathway which has been linked to photoreceptor degeneration ( Iribarne and Masai , 2017; Tolone et al . , 2019; Power et al . , 2020 ) .\",\n \"Thus , the NTR/Mtz-mediated cell ablation system may provide a novel approach for modeling neurodegenerative diseases associated with PARP-dependent cell death that is both inducible and titratable .\",\n \"To investigate cell death more broadly , we also explored the roles of necroptosis , apoptosis and tdp1 signaling in the zebrafish-inducible RP model .\",\n \"Interestingly , inhibiting necroptosis but not apoptosis promoted rod cell survival in rho:YFP-NTR fish ( Figure 7 and Table 3 ) .\",\n \"Consistent with this , CRISPR/Cas9-based knockdown of ripk1l protected rods , but casp3a/3b knockdown did not ( Figure 8 ) .\",\n \"Necroptosis is primarily associated with secondary cone cell death in RP models ( Murakami et al . , 2015; Murakami et al . , 2012; Yang et al . , 2017 ) but has also been implicated in rod cell death in IRBP mutant RP models ( Sato et al . , 2013 ) and/or may damage photoreceptors indirectly via necroptotic microglia signaling ( Huang et al . , 2018 ) .\",\n \"Combined , these results suggest that necroptosis , PARP1-dependent parthanatos , and/or c-GMP-dependent cell death mediate NTR/Mtz-induced rod cell ablation .\",\n \"The potential relevance of these alternative cell death pathways to heritable photoreceptor degeneration may explain the relatively high rate of validation we observed in cross-species tests of lead compounds .\",\n \"To evaluate whether lead compound effects were mediated by inhibition of PARP signaling , a subset of lead compounds and a PARP inhibitor control ( BMN ) were analyzed for survival effects in parp1 knockdown fish .\",\n \"Among four leads tested that were implicated as parp1 inhibitors , two ( MIC and DHA ) were no longer protective when parp1 expression was suppressed ( Figure 9 ) , suggesting parp1 inhibition mediated their neuroprotective effects .\",\n \"PARP1 has also been shown to have a role in stem cell reprogramming ( Chiou et al . , 2013; Doege et al . , 2012; Weber et al . , 2013 ) .\",\n \"PARP inhibition might therefore block Müller glia dedifferentiation , diverting injury-induced activity from a regenerative to neuroprotective program ( Bringmann et al . , 2009 ) .\",\n \"Finally , MOA analyses also implicated additional , potentially complementary , lead targets ( Table 2 ) .\",\n \"Compounds acting through independent targets/pathways have the potential to produce additive or even synergistic effects .\",\n \"This possibility was confirmed for 10 of 19 paired lead compounds tested in rho:YFP-NTR fish ( Figure 9 , Supplementary file 1c ) .\",\n \"Analogous tests in mouse photoreceptor cell cultures showed a trend toward enhanced survival effects of paired leads but no fully additive effects ( Figure 4—figure supplement 2 ) .\",\n \"This discrepancy could be due to cell culture assays lacking in vivo complexity , such as signals requiring intact tissue or interactions between cell types or with other tissues .\",\n \"This result highlights a potential advantage of whole organism phenotypic drug screening: the ability to identify compounds/conditions that would be missed in more simplified assay systems .\",\n \"That combinatorial assays could be performed efficiently and rapidly also exemplifies key advantages the zebrafish system affords phenotypic drug discovery , for example , versatility and low cost .\",\n \"In summary , we identified 11 lead compounds promoting rod cell survival in an inducible zebrafish RP model using a large-scale in vivo phenotypic drug discovery platform .\",\n \"Nine lead compounds were also effective as neuroprotectants in primary mouse photoreceptor cell cultures and three promoted photoreceptor survival in rd1 mouse retinal explants .\",\n \"MOA studies indicated a subset of lead compounds may protect rod cells by inhibiting PARP-dependent cell death pathways and/or necroptosis .\",\n \"Combinatorial assays in fish showed additive and even synergistic effects , suggesting some lead compounds target independent neuroprotective pathways ( summarized in Figure 11 ) .\",\n \"We hypothesize that compounds producing beneficial outcomes across diverse animal disease models likely target highly conserved MOA and may therefore stand an increased likelihood of successfully translating to the clinic .\",\n \"Further , our data suggest polypharmacological targeting of complementary neuroprotective mechanisms has the potential to maximize therapeutic benefits for IRD/RP patients .\"\n ],\n [\n \"All animal studies described herein were performed in accordance with both the Association for Research in Vision and Ophthalmology ( ARVO ) statement on the ‘Use of Animals in Ophthalmic and Vision Research’ and the National Institutes of Health ( NIH ) Office of Laboratory Animal Welfare ( OLAW ) policies regarding studies conducted in vertebrate species .\",\n \"Animal protocols were approved by the Animal Care and Use Committees of the Johns Hopkins University School of Medicine ( protocol # FI19M489 and #MO20M253 ) and Medical University of South Carolina ( protocol #2018–00399 ) .\",\n \"Zebrafish were maintained using established temperature and light cycle conditions ( 28 . 5°C , 14 hr of light/10 hr of dark ) .\",\n \"A previously generated zebrafish transgenic line , Tg ( rho:YFP-Eco . NfsB ) gmc500 ( hereafter , rho:YFP-NTR ) expresses a fusion protein of enhanced yellow fluorescent protein ( YFP ) and a bacterial nitroreductase ( NTR ) enzyme ( encoded by the E . coli nfsB gene ) selectively in rod photoreceptors ( Walker et al . , 2012 ) .\",\n \"When rho:YFP-NTR fish are exposed to the prodrug metronidazole ( Mtz ) , NTR reduces Mtz to a DNA damage-inducing cytotoxic derivative , resulting in the death of NTR-expressing cells ( Chen et al . , 2011; Curado et al . , 2007 ) .\",\n \"A newly generated transgenic line , Tg ( rho:GAP-YFP-2A-nfsB_Vv F70A/F108Y ) jh405 ( hereafter rho:YFP-NTR2 . 0 ) expresses a membrane tagged YFP and an improved version of NTR which enables induction of cell death at much lower concentrations of Mtz ( Sharrock et al . , 2020 ) .\",\n \"These lines were propagated in a pigmentation mutant with reduced iridophore numbers , roya9 ( roy ) , to facilitate YFP reporter signal detection in vivo .\",\n \"Non-transgenic roy fish were used to define reporter signal cutoff values for fluorescent microplate reader assays .\",\n \"For each large-scale drug screening assays , 6000–12 , 000 eggs were collected by group breeding ~300 adult rho:YFP-NTR fish ( White et al . , 2016 ) .\",\n \"rho:YFP-NTR larvae were treated with 2 . 5 mM Mtz or 0 . 1% DMSO at 5 dpf for 2 days ( 10 fish per condition ) , and collected at 7 dpf for immunostaining using previously reported methods ( Unal Eroglu et al . , 2018 ) .\",\n \"Briefly , larvae were fixed in 4% PFA at 4°C overnight , infiltrated with 30% sucrose , and embedded in OCT .\",\n \"Cryosections of 10 µm thickness were made for immunofluorescence staining .\",\n \"Each section was blocked with 1x PBS containing 0 . 5% Triton X-100% and 5% goat serum for one hour followed by primary antibody staining overnight .\",\n \"The next day , each section was washed and stained with the secondary antibody for 2 hr .\",\n \"All slides were mounted and underwent confocal imaging .\",\n \"The primary antibodies used in this study were zpr-1 ( anti-Arrestin3a; 1:200 dilution; from ZIRC ) , 1d1 ( anti-Rhodopsin; 1:50; gift from Dr . James M . Fadool ) and 4C12 ( uncharacterized rod cell antigen , 1:100 dilution; gift from Dr . James M . Fadool ) .\",\n \"The secondary antibody used was goat anti-mouse Alexa 647 ( 1:1000; Invitrogen ) .\",\n \"rho:YFP-NTR larvae were separated into six groups of ~23 larvae per group .\",\n \"Each group was treated with varying Mtz concentrations ( 10 , 5 , 2 . 5 , 1 . 25 , 0 . 625 , 0 mM ) , from 5 to 8 days post-fertilization ( dpf ) .\",\n \"YFP signals were quantified daily by fluorescence microplate reader ( TECAN Infinite M1000 PRO; excitation 514 nm , bandwidth 5 nm; emission 538 nm , bandwidth 10 nm ) to track changes in rod photoreceptor cell numbers relative to Mtz concentration ( Walker et al . , 2012; White et al . , 2016 ) .\",\n \"Two technical repeats were performed and results aggregated across experiments by normalizing to non-ablated controls .\",\n \"To calculate sample size ( N ) and evaluate assay quality , two 96-well plates of larvae were treated with 2 . 5 mM Mtz/0 . 1% DMSO ( ablated ‘+Mtz’ control ) or 0 . 1% DMSO only ( non-ablated ‘-Mtz’ control ) from 5 to 7 dpf ( 196 fish per condition ) .\",\n \"YFP signals were then quantified at seven dpf ( due to the large sample size , a single technical replicate was performed ) .\",\n \"Power calculations were used to determine sample sizes across a range of error rates and effect sizes ( White et al . , 2016 ) .\",\n \"This analysis suggested a sample size of nine per condition tested would sufficiently minimize false-discoveries ( i . e . type I and type II error rates of 0 . 05 and 0 . 05 , respectively ) and account for drug effects reaching 50% of signal window of the positive ( non-ablated ) control .\",\n \"However , to account for plating errors and other confounding variables , we chose to increase the sample size to 16 .\",\n \"A schematic of the primary screening process is presented in Figure 1C .\",\n \"For the primary screen , 2934 compounds from John Hopkins Drug Library ( JHDL ) were screened across six concentrations ( 4 μM-125 nM using a twofold dilution series ) .\",\n \"The JHDL consists of ~2200 drugs approved for use in humans ( e . g . FDA approved ) with the remainder approved for clinical trials ( Chong et al . , 2006 ) .\",\n \"The ARQiv screening process has been detailed previously ( White et al . , 2016 ) and was adapted here for large-scale quantification of YFP-expressing rod photoreceptors ( Walker et al . , 2012; White et al . , 2016 ) .\",\n \"rho:YFP-NTR embryos were collected and raised in zebrafish E3 embryo media ( 5 mM NaCl; 0 . 17 mM KCl; 0 . 33 mM CaCl; 0 . 33 mM MgSO4 ) .\",\n \"At 16 hr post fertilization ( hpf ) , N-phenylthiourea ( PTU ) was added to E3 media ( E3/PTU ) at a final concentration of 0 . 2 mM to promote ocular transparency by inhibiting melanosome maturation in the retinal pigment epithelium .\",\n \"At 4 dpf , visual screens were performed to remove larvae with abnormal morphology or low YFP expression levels .\",\n \"Stock drug and DMSO ( negative control ) solutions were automatically dispensed and diluted across a 96-well plate containing E3/PTU using a robotic liquid handling system ( Hudson Robotics ) .\",\n \"At 5 dpf , a COPAS-XL ( Complex Object Parametric Analyzer and Sorter , Union Biometrica ) was used to dispense single larvae into individual wells containing either drug or DMSO; the final DMSO concentration was 0 . 1% across all conditions .\",\n \"After a 4 hr pre-exposure to test drugs or DMSO alone , larvae were treated with 2 . 5 mM Mtz to induce rod photoreceptor death .\",\n \"Larvae were maintained under these conditions for 2 days until 7n dpf and then anesthetized with clove oil ( 50 ppm final concentration ) ; YFP signals were measured as described above .\",\n \"Larvae exposed to 2 . 5 mM Mtz/0 . 1% DMSO without any tested drug served as controls ( ‘+Mtz’ ) for maximal rod cell ablation .\",\n \"Larvae treated solely with 0 . 1% DMSO served as non-ablated controls ( ‘-Mtz’ ) to calculate maximal YFP signal levels .\",\n \"Non-transgenic larvae were used to establish a signal cutoff value , as previously described ( White et al . , 2016 ) .\",\n \"A customized R code program was applied for real-time data analysis of compound performance relative to controls , including dose-response curves and strictly standardized mean difference ( SSMD ) scores ( https://github . com/mummlab/ARQiv2; Ding and Zhang , 2021 ) .\",\n \"Compound concentrations producing a SSMD score ≥1 were considered potential ‘hits’ and evaluated visually using fluorescence microscopy to eliminate non-specific fluorescence; e . g . , dead larvae or autofluorescent compounds .\",\n \"After the initial screen , 42 top-performing ‘hit’ compounds were selected for confirmatory and orthogonal assays .\",\n \"Hit drugs were obtained from new sources and tested across a wider range of concentrations ( fivefold dilution series , eight total concentrations , n=30 fish/condition ) .\",\n \"Based on the toxicity profile of each drug , the starting concentration was either 100 , 10 , or 1 μM .\",\n \"For validation assays , YFP signals were measured by fluorescence microplate reader with at least three experimental repeats conducted and SSMD scores calculated .\",\n \"All experimental results were normalized and pooled to calculate effect sizes , confidence intervals , sample sizes ( N ) and p-values .\",\n \"p Values were calculated using Student’s t-test followed by Bonferroni correction for multiple comparisons resulting in an adjusted significance level of 0 . 005 ( α=0 . 005 ) .\",\n \"For in vivo confocal imaging assays , five dpf rho:YFP-NTR larvae were treated with drug plus 2 . 5 mM Mtz/0 . 1% DMSO ( ablated +Mtz controls ) , or 0 . 1% DMSO ( non-ablated -Mtz controls ) for 48 hr and processed for intravital imaging at 7 dpf .\",\n \"Similar treatment strategy was applied to five dpf rho:YFP-NTR2 . 0 larvae except the Mtz treatment was reduced to 10 μM .\",\n \"All compounds were tested at the maximal effective concentration in 0 . 1% DMSO .\",\n \"Larvae from each group were anesthetized in 0 . 016% tricaine and embedded on their sides in 1% low melt agarose gel .\",\n \"An Olympus Fluoview FV1000 confocal microscope with a 20x water immersion objective ( 0 . 95 NA ) was used to collect 30–40 z-stack images of YFP-expressing rod cells across the whole retina at 4 μm intervals .\",\n \"These image slices were stacked into a single maximal intensity projection image using Image J . To provide greater detail of rod photoreceptor morphology , a region in the dorsal-nasal quadrant was imaged using a 60× water immersion objective ( 1 . 10 NA ) at 4 . 18 μm intervals .\",\n \"Three contiguous image slices were then stacked into a single maximal intensity projection image using Image J . As a means of assessing rod photoreceptor survival , intravital confocal retinal images were collected of YFP-expressing rod cells per condition and processed for volumetric quantification of YFP signal volume using automated image analysis software ( Imaris 3 . 9 . 1; see White et al . , 2017 ) .\",\n \"Briefly , YFP-expressing rod cells volumes were automatically rendered using the same parameters across all treatment groups with the sum of all volumes used to estimate rod photoreceptor numbers per each retina .\",\n \"For images from rho:YFP-NTR2 . 0 larvae , the sum of rod cell YFP volumes in each retina was normalized to the number of YFP-expressing cells to estimate the average rod photoreceptor cell size per each condition .\",\n \"The average number of cells quantified per retina was 18 . 52 , with sample sizes ranging from 9 to 19 fish per condition .\",\n \"The eleven lead compounds were tested to eliminate false-positives that directly inhibited E . coli NTR enzymatic activity ( from the nfsB_Ec gene ) using an established anticancer prodrug ( CB1954 ) reduction assay ( Prosser et al . , 2010 ) .\",\n \"The highest soluble concentration of the test drug ( up to 300 μM maximum ) or 0 . 1% DMSO ( rate control ) was added to a master mix of 250 μM CB1954/1 μM NTR/250 μM NADPH/10 mM Tris ( pH 7 . 0 ) to determine any inhibitory effect of the compounds on NTR-mediated reduction of CB1954 .\",\n \"Reactions were conducted in 100 μl volume in a 96-well plate format .\",\n \"CB1954 reduction kinetics were assayed at 420 nm unless the test compound exhibited confounding absorbance at 420 nm , in which case NADPH depletion was monitored at 340 nm .\",\n \"The reaction rates for CB1954 reduction in the presence of tested compounds were compared to the DMSO control .\",\n \"Compounds with NTR activity <75% of controls were deemed potentially inhibitory and IC50 values ( the concentration required to restrict NTR activity to 50% of the DMSO control ) determined .\",\n \"Compounds were also evaluated for the ability to inhibit NTR-mediated reduction of Mtz .\",\n \"For this assay , the highest soluble concentration of the test drug ( up to 300 μM maximum ) or DMSO ( rate control ) was added to a master mix of 500 μM Mtz/17 μM NTR/100 μM NADPH/0 . 55 μM glucose dehydrogenase/5 mM glucose/50 mM sodium phosphate buffer ( pH 7 . 0; the glucose dehydrogenase used for cofactor regeneration , maintaining a steady-state of NADPH to allow Mtz reduction to be monitored directly at 340 nm without interference from NADPH oxidation ) .\",\n \"Reactions were conducted in 100 μl volume and assayed at 340 nm using 1 mm quartz cuvettes ( Rich et al . , 2018 ) .\",\n \"IC50 values for potentially inhibitory compounds were measured as with CB1954 .\",\n \"All kinetic assays were conducted in triplicate with a single purified batch of NfsB_Ec protein .\",\n \"To evaluate whether lead compounds promoted rod photoreceptor development , rho:YFP-NTR larvae were handled as described for the primary screen with the following exceptions: at five dpf , larvae were exposed solely to test compounds; larvae were not treated with Mtz .\",\n \"YFP reporter signals were quantified by fluorescence microplate reader at 7 dpf and compared to non-ablated ‘-Mtz’ controls treated only with 0 . 1% DMSO .\",\n \"Sample sizes ranged from 58 to 88 across at least two experimental repeats .\",\n \"To determine whether lead compounds stimulated retinal regeneration , rho:YFP-NTR larvae were handled as described for the primary screen with the following exceptions: at 5 dpf , larvae were incubated with either 10 mM Mtz or 0 . 1% DMSO for 24 hr .\",\n \"At 6 dpf , larvae were then placed in new 0 . 1% DMSO/E3/PTU media containing test compounds ( or DMSO alone ) for 3 days .\",\n \"YFP signal intensity was measured at 9 dpf .\",\n \"Sample sizes ranged from 31 to 83 across at least two experimental repeats .\",\n \"A transgenic mouse line with an IRES-GFP cassette integrated at the QRX locus ( QRX-IRES-EGFP ) has GFP reporter expression restricted to retinal photoreceptors , and was used here to isolate primary photoreceptor cells .\",\n \"Mice were housed with 12 hr light/12 hr dark cycles , at 22°C , 30–70% relative humidity and food/water ad libitum .\",\n \"Primary mouse retinal photoreceptor cells were isolated and prepared for culture as previously described ( Fuller et al . , 2014 ) .\",\n \"Briefly , murine retinas were isolated at postnatal day four ( P4 ) .\",\n \"Retinal tissue was dissociated into a single-cell suspension by incubating whole tissue in activated papain in Hibernate-E without Calcium ( BrainBits ) for 15 min at 37°C .\",\n \"Cells were resuspended in culture media ( Neurobasal , 2% B-27 , 0 . 5 mM L-Glutamine and 1x final Penicillin/streptomycin; all Life Technologies ) and seeded onto poly-D-lysine coated 384 well tissue culture plates .\",\n \"To test lead compounds for survival effects , thapsigargin and tunicamycin ( Sigma ) and were used as stressor compounds to induce photoreceptor cell death .\",\n \"Stressors and test compounds were added at the time of primary cell seeding .\",\n \"Eleven lead compounds were tested at seven concentrations across a threefold dilution series ( from 30 µM to 40 nM ) .\",\n \"For paired tests , each compound was applied at a single concentration , WAR ( 3 . 33 µM ) , CPO ( 0 . 04 µM ) , DHA ( 1 . 11 µM ) , ART ( 10 µM ) , ZPT ( 0 . 12 µM ) , CAL ( 0 . 04 µM ) , COR ( 3 . 33 µM ) , CHL ( 3 . 33 µM ) , SUL ( 3 . 33 µM ) , MIC ( 10 µM ) , and CLO ( 1 . 11 µM ) .\",\n \"After a 48 hr treatment , cells were stained with Hoescht and ethidium homodimer; nine field images were acquired via an automated imager ( Cellomics Vti; ThermoFisher ) using a 20x objective and images analyzed .\",\n \"Photoreceptor number and the percent of photoreceptor to retinal cell population per well was determined by automated quantification of the number of live Hoechst 33342-stained , GFP-expressing cells using a custom algorithm ( Neuronal Profiling software package; ThermoFisher ) .\",\n \"A minimum of two experimental repeats was performed for each test with individual assays consisting of six biological replicates per condition tested .\",\n \"Mice were generated from retinal degeneration 1 ( Pde6brd1 , hereafter rd1 ) breeding pairs ( gift from Dr . Debora Farber; UCLA ) and housed under a 12:12 light:dark cycle with access to food and water ad libitum .\",\n \"All chemicals used for organ cultures were tissue culture grade and purchased from Invitrogen unless otherwise noted .\",\n \"Cultures of retina with attached retinal pigment epithelium ( RPE ) were grown according to published protocols ( Ogilvie et al . , 1999; Pinzón-Duarte et al . , 2000; Rohrer and Ogilvie , 2003 ) with modifications ( Bandyopadhyay and Rohrer , 2010 ) .\",\n \"P10 pups were decapitated and heads were rinsed in 70% ethanol; whole eyes were dissected and placed in ice-cold Hanks balanced salt solution plus glucose ( 6 . 5 g/L ) .\",\n \"Eyes were first incubated in 1 mL of high-glucose Hanks balanced salt solution/0 . 5 mg/ml proteinase K ( 37°C , 7 min ) and then in Neurobasal medium ( Life Technologies ) plus 10% fetal calf serum to stop enzymatic activity .\",\n \"The retina with attached RPE was dissected free from the choroid and sclera after removing the anterior chamber , lens and vitreous .\",\n \"Relaxing cuts were made to flatten the tissue prior to transfer to the upper compartment of a Costar Transwell chamber ( RPE layer faced-down ) using a drop of Neurobasal medium .\",\n \"Neurobasal media with B-27 supplement ( Life Technologies ) was placed in the lower compartment .\",\n \"For each of the cohorts , three to fourindividual P10 retina/RPE explants were placed in culture .\",\n \"The cultures were kept in an incubator ( 5% CO2 , balanced air , 100% humidity , 37°C ) and the lower compartment media changed every 2 days ( neither antimitotics nor antibiotics were required ) .\",\n \"Test compounds were added to the culture media and refreshed every 2 days for 11 days .\",\n \"At completion , explants were fixed in 4% PFA , sectioned 14 µm thick , and stained with toluidine blue ( Bandyopadhyay and Rohrer , 2010 ) .\",\n \"For each culture , 10 counts of rows of photoreceptors were performed in the center of the explant; the average of this cell counts provides the mean for each retina , with the average of all retinas providing the mean ± SEM per culture condition .\",\n \"Cell counts were compared by repeated measure ANOVA to assess dose-response treatment effects followed by a Bonferroni posthoc analysis to determine if differences between control and lead compounds compared across the entire dose range were statistically significant .\",\n \"An ANOVA across groups but within a particular dose , using a Bonferroni/Dunn test to reduce type I errors , was used to determine if lead compounds promoted a statistically significantly increase in cell numbers compared to control media at low , middle , and/or high concentrations .\",\n \"PLGA-DHA microparticles were prepared using a single emulsion solvent evaporation method .\",\n \"Briefly , 200 mg PLGA ( Resomer RG 502 hr , Poly ( D ) , L-lactide-co-glycolide; Evonik Corporation ) was dissolved in 1 mL of dichloromethane ( DCM , Sigma-Aldrich ) , and mixing with 40 mg DHA ( TCI , Tokyo , Japan ) dissolved in 0 . 125 ml dimethyl sulfoxide ( DMSO , Sigma-Aldrich ) .\",\n \"The mixture was homogenized ( L4RT , Silverson Machines ) at 7000 RPM for 1 min .\",\n \"The homogenized mixture was then poured into a solution containing 1% polyvinyl alcohol ( 25 kDa , 88% hydrolys , Polysciences , Warrington , PA ) in water under continuous stirring .\",\n \"Particles were hardened by allowing solvent to evaporate while stirring at room temperature for 2 hr .\",\n \"Particles were collected via centrifugation ( International Equipment Co ) at 1000 x g for 5 min , and washed three times with HyPure cell culture grade water ( endotoxin-free , HyClone , Logan , UT ) and re-collected by centrifugation three times .\",\n \"The washed particles were then lyophilized and stored frozen until use .\",\n \"Microparticles were resuspended at the desired concentration prior to injection in a sodium hyaluronate solution ( Healon ) diluted fivefold with endotoxin-free HyPure water at the desired concentration prior to injection .\",\n \"Particle size distribution was determined using a Coulter Multisizer 4 ( Beckman Coulter , Inc , Miami , FL ) .\",\n \"Particles were resuspended in double distilled water and added dropwise to 100 ml of ISOTON II solution until the coincidence of the particles was between 8% and 10% .\",\n \"At least 100 , 000 particles were sized for each batch of particles to determine the mean particle size and size distribution .\",\n \"To determine the drug loading , microparticles were dissolved in DMSO and the total drug content was calculated by measuring the UV absorbance at 289 nm after react with NaOH in triplicate ( C . -S . Lai et al . , 2007a ) .\",\n \"The average microparticle size was 14 . 2 ± 1 . 9 µm and the DHA loading was 20 . 3% ± 0 . 7 ( w/w ) across three experimental replicates .\",\n \"Release kinetics were obtained by resuspending microparticles in 1 ml phosphate buffered saline ( PBS containing 0 . 1% DMSO , pH 7 . 4 ) and incubating at 37°C on a platform shaker ( 140 RPM ) .\",\n \"Supernatant was collected at predetermined intervals by centrifugation at 2000 x g for 5 min .\",\n \"Drug-containing supernatant was collected and particles were resuspended in 1 ml of fresh 1xPBS containing 0 . 1% DMSO .\",\n \"DHA concentration in the collected supernatant was assayed via absorbance at 289 nm in triplicate for each sample ( Figure 4—figure supplement 1 ) .\",\n \"B6 . CXB1-Pde6brd10/J animals ( Jackson Laboratories stock #004297 ) were maintained as homozygotes .\",\n \"On P14 , animals were anesthetized with ketamine/xylazine ( 115 mg/kg and 5 mg/kg , respectively ) , and placed on a heating pad to maintain body temperature throughout injection and recovery .\",\n \"Glass needles were pulled and a bore size of approximately 75 µm was made to allow for the movement of the microspheres into and out of the needle .\",\n \"A PLI-100 picospritzer ( Harvard Apparatus ) was used for intravitreal injections with settings of 350 ms injection time at 30 psi , which yielded the injection of approximately 500 nL .\",\n \"One eye was randomly selected for injection with vehicle and the other was used for microsphere injections .\",\n \"Following injection , GenTeal gel ( Novartis ) was applied to the eyes to prevent corneal drying .\",\n \"Animals were monitored until they were awake and moving normally .\",\n \"A total of 17 mice was used across three technical repeats .\",\n \"Eighteen days post-injection , animals were anesthetized with isoflurane and decapitated; the eyes were removed and placed in 4% PFA for 30 min .\",\n \"After removing the cornea and lens , the eye cups were placed back into fixative for 2 hr on ice .\",\n \"After thorough washing with 1× PBS , eyecups were soaked in 15% sucrose ( w/v ) in 1× PBS for 3 hr , then incubated in 30% sucrose ( w/v ) /1× PBS overnight .\",\n \"Samples were embedded in OCT compound ( TissueTek ) and frozen on dry ice .\",\n \"Cryosections of 12–14 µm thickness were cut on a Leica CM3050S cryostat and were mounted on Superfrost Plus slides ( Fisher Scientific ) .\",\n \"Slides were rehydrated with 1× PBS and then incubated in blocking buffer ( 10% Normal donkey serum/ 0 . 3% Triton X-100/1× PBS ) for 2 hr , and were then incubated for 30 hr at 4°C with primary antibodies diluted in blocking buffer .\",\n \"Sections were washed in 1× PBS and incubated for 2 hr at room temperature with secondary antibodies diluted in blocking buffer .\",\n \"Following additional washes , sections were incubated with DAPI and mounted with FluoroGel ( Electron Microscopy Sciences ) .\",\n \"Sections were examined using a Zeiss Axioplan two epifluorescence microscope fitted with an Axioscope camera .\",\n \"Following acquisition with a 40× air objective , images were analyzed in ImageJ using the measure tool to examine Outer Nuclear Layer ( ONL ) thickness .\",\n \"Four to 5 measurements were taken at 75 µm intervals for 200–500 µm both superior and inferior from the optic nerve head .\",\n \"The measured values from multiple sections were averaged for each eye .\",\n \"Data were analyzed and statistical tests were performed in GraphPad Prism 7 .\",\n \"An Olympus FV1000 confocal microscope was also used to image sections stained with multiple dyes .\",\n \"To evaluate tdp1 and parp1 as a potential shared target across multiple leads , as well as cell death pathways implicated in photoreceptor degeneration , eight PARP inhibitors , one necroptosis inhibitor , three apoptosis inhibitors , and three Tdp1 inhibitors were selected for testing using the confirmation screening protocol ( Figure 2A ) .\",\n \"Based on the toxicity profile of each inhibitor , upper concentrations were adjusted to 64 , 32 , or 8 µM and a twofold dilution series was used to test a total of 10 concentrations for rod cell survival effects in rho:YFP-NTR larvae .\",\n \"Sample sizes varied from 27 to 173 across two to six experimental replicates for each tested compound .\",\n \"YFP signal levels were measured by fluorescence microplate reader ( i . e . ARQiv assay ) .\",\n \"Results across conditions were normalized to non-ablated controls and pooled across experimental repeats to calculate effect sizes , confidence intervals , and p-values .\",\n \"P values were calculated by performing Student’s t-tests .\",\n \"Bonferroni correction for multiple comparisons resulted in an adjusted significance level of 0 . 003 ( α=0 . 003 ) .\",\n \"Note: assessments of TDP1 inhibitors followed the fivefold dilution series protocol used for secondary confirmation assays described above .\",\n \"For paired inhibitor tests , three concentrations of BMN ( parp inhibitor ) and three concentrations of NEC ( necroptosis inhibitor ) were combined .\",\n \"YFP quantification and statistical analysis of rod cell survival in Mtz-treated rho:YFP-NTR fish were performed as above .\",\n \"Five genes , parp1 , ripk1l , casp3a , casp3b , tdp1 , were selected for CRISPR/Cas9-based knockdown in rho:YFP-NTR zebrafish using an established ‘redundant targeting’ methodology ( Wu et al . , 2018 ) .\",\n \"Four sgRNAs were designed to target each gene .\",\n \"The oligonucleotide for generating each sgRNA was either those suggested by Wu et al . , 2018 , Table S2 or designed using CRISPRscan ( https://www . crisprscan . org , Supplementary file 3; Moreno-Mateos et al . , 2015 ) .\",\n \"Preparation of sgRNAs followed the published method ( Wu et al . , 2018 ) .\",\n \"Briefly , four sgRNAs of each gene were in vitro transcribed using the HiScribe T7 High Yield RNA Synthesis kit ( New England BioLabs ) and purified using the NEB Monarch RNA Cleanup kit .\",\n \"A mixture of four sgRNAs ( 1 ng in total ) and Cas9 protein ( 2 . 5 µM , IDT ) was injected into rho:YFP-NTR embryos at the one-cell stage for targeting parp1 , ripk1l and tdp1 .\",\n \"Four sgRNAs each for casp3a and casp3b were pooled to co-target both genes .\",\n \"Injected and uninjected embryos were treated ±Mtz ( 2 . 5 mM ) at 5 dpf .\",\n \"Rod cell survival was evaluated by fluorescence microplate reader-based quantification of YFP at 6 dpf .\",\n \"Student’s t-test was used to compare rod survival between injected and uninjected ( control ) fish .\",\n \"Each gene knockdown experiment was conducted in triplicate or quadruplicate .\",\n \"A subset of lead compounds ( and a PARP inhibitor control , BMN ) was also tested in parp1 knockdown fish to determine if Parp1 inhibition was required for survival effects .\",\n \"Leads were applied to larvae one hour before Mtz administration and YFP quantification performed as above .\",\n \"Each drug test was conducted in duplicate or triplicate .\",\n \"To calculate p-values relative to +Mtz controls , Student’s t-tests were performed .\",\n \"Bonferroni correction for multiple comparisons resulted in an adjusted significance level of 0 . 006 ( α=0 . 006 ) .\",\n \"To quantify gene expression levels in rho:YFP-NTR larvae ±CRISPR/Cas9-based gene knockdown , 16–24 larvae from each condition were pooled for mRNA extraction at six dpf .\",\n \"RNA was purified using NEB Monarch RNA Cleanup kit and reverse transcribed to cDNA using qScript cDNA synthesis kit ( QuantaBio ) according to the instruction from the manufacturer .\",\n \"Quantitative PCR was conducted using designed primers ( Supplementary file\",\n \"3 ) and PowerUp SYBR Green Master Mix ( ABI ) in QuantaStudio ( ABI ) .\",\n \"Delta delta CT analysis was performed to calculate relative fold change in gene expression levels between knockdown larvae and controls .\",\n \"Each experiment was performed in triplicate or quadruplicate .\",\n \"To further evaluate Parp-dependent signaling downstream of Mtz treatments in rho:YFP-NTR larvae , PAR accumulation , a vetted measure of Parp activity , was quantified by western blot following the published method ( Kam et al . , 2018 ) .\",\n \"Briefly , ~30 rho:YFP-NTR fish were treated ±2 . 5 mM Mtz at five dpf were collected at 3 , 6 , 12 , 24 , 48 hr post treatment ( hpt ) .\",\n \"Larvae were homogenized and lysed using RIPA buffer ( Sigma ) supplemented with protease inhibitors ( Roche ) .\",\n \"Protein concentrations were quantified using Pierce BCA kit ( ThermoScientific ) following the manufacturer’s instruction .\",\n \"Protein samples were separated on Tris-Glycine gel and then transferred onto nitrocellulose membranes .\",\n \"The membrane was blocked with 5% non-fat milk in TBST ( Tris-buffered saline with 0 . 1% Tween 20 ) at room temperature for 1 hr and then blotted with the primary antibody , anti-human-PAR ( 1:2000 , Dowson lab reagent ) , at 4°C overnight .\",\n \"After rinsing in TBST for three times , the secondary antibody , anti-Human IgG-HRP ( 1:10 , 000 , Abcam ) was applied for 1 hr at room temperature .\",\n \"Anti-beta-actin-HRP antibody ( 1:20 , 000 ) was used as a loading control .\",\n \"Immunolabeled bands were visualized by ECL substrate and quantified in Fiji .\",\n \"24 hr post-Mtz PAR accumulation assays were performed in triplicate .\",\n \"Mann-Whitney U test was performed and a p-value of ≤0 . 05 used to indicate statistical significance .\",\n \"To test for additive neuroprotective effects in zebrafish , seven lead compounds were tested at optimal concentrations either individually or in pairs using the primary screen protocol .\",\n \"All experimental results were normalized to controls and pooled per compound tested to calculate effect sizes , confidence intervals , and p-values using Student’s t-test .\",\n \"Bonferroni correction for multiple comparisons resulted in an adjusted significance level of 0 . 002 ( α=0 . 002 ) .\",\n \"Sample sizes for each paired condition varied from 86 to 160 across at least four experimental replicates .\",\n \"For the primary screening , under the R environment , a customized ARQiv data analysis package was used to calculate sample size ( n ) , quality control strictly standardized mean difference ( SSMD ) and SSMD scores as previously described ( White et al . , 2016 ) .\",\n \"The following results of each drug were derived: ( 1 ) a plot of signal to background ratio at all tested concentrations , ( 2 ) a plot and a table of SSMD scores , and ( 3 ) a signal intensity heat map of each drug plate ( 96-well plate view ) .\",\n \"To combine and analyze the data from different experiments , data normalization was conducted by ( Si -X-neg ) / ( X- Pos -X-neg ) .\",\n \"Si is the signal of ith reading , X-Pos is mean of positive controls , and X-Neg is mean of negative controls .\",\n \"To compare between experimental conditions and controls , Student’s t-tests were performed followed by Bonferroni correction for multiple comparisons ( α/n ) .\",\n \"Relative effects , 95% confidence intervals and p-values were calculated .\",\n \"For mouse photoreceptor cell culture assays and volumetric confocal image analyses , adjusted p-values were calculated by performing Mann-Whitney U tests followed by false discovery rate correction for multiple comparisons .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Retinitis pigmentosa ( RP ) and associated inherited retinal diseases ( IRDs ) are caused by rod photoreceptor degeneration , necessitating therapeutics promoting rod photoreceptor survival .","To address this , we tested compounds for neuroprotective effects in multiple zebrafish and mouse RP models , reasoning drugs effective across species and/or independent of disease mutation may translate better clinically .","We first performed a large-scale phenotypic drug screen for compounds promoting rod cell survival in a larval zebrafish model of inducible RP .","We tested 2934 compounds , mostly human-approved drugs , across six concentrations , resulting in 113 compounds being identified as hits .","Secondary tests of 42 high-priority hits confirmed eleven lead candidates .","Leads were then evaluated in a series of mouse RP models in an effort to identify compounds effective across species and RP models , that is , potential pan-disease therapeutics .","Nine of 11 leads exhibited neuroprotective effects in mouse primary photoreceptor cultures , and three promoted photoreceptor survival in mouse rd1 retinal explants .","Both shared and complementary mechanisms of action were implicated across leads .","Shared target tests implicated parp1-dependent cell death in our zebrafish RP model .","Complementation tests revealed enhanced and additive/synergistic neuroprotective effects of paired drug combinations in mouse photoreceptor cultures and zebrafish , respectively .","These results highlight the value of cross-species/multi-model phenotypic drug discovery and suggest combinatorial drug therapies may provide enhanced therapeutic benefits for RP patients ."],"string":"[\n \"Retinitis pigmentosa ( RP ) and associated inherited retinal diseases ( IRDs ) are caused by rod photoreceptor degeneration , necessitating therapeutics promoting rod photoreceptor survival .\",\n \"To address this , we tested compounds for neuroprotective effects in multiple zebrafish and mouse RP models , reasoning drugs effective across species and/or independent of disease mutation may translate better clinically .\",\n \"We first performed a large-scale phenotypic drug screen for compounds promoting rod cell survival in a larval zebrafish model of inducible RP .\",\n \"We tested 2934 compounds , mostly human-approved drugs , across six concentrations , resulting in 113 compounds being identified as hits .\",\n \"Secondary tests of 42 high-priority hits confirmed eleven lead candidates .\",\n \"Leads were then evaluated in a series of mouse RP models in an effort to identify compounds effective across species and RP models , that is , potential pan-disease therapeutics .\",\n \"Nine of 11 leads exhibited neuroprotective effects in mouse primary photoreceptor cultures , and three promoted photoreceptor survival in mouse rd1 retinal explants .\",\n \"Both shared and complementary mechanisms of action were implicated across leads .\",\n \"Shared target tests implicated parp1-dependent cell death in our zebrafish RP model .\",\n \"Complementation tests revealed enhanced and additive/synergistic neuroprotective effects of paired drug combinations in mouse photoreceptor cultures and zebrafish , respectively .\",\n \"These results highlight the value of cross-species/multi-model phenotypic drug discovery and suggest combinatorial drug therapies may provide enhanced therapeutic benefits for RP patients .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Photoreceptors are the cells responsible for vision .","They are part of the retina: the light-sensing tissue at the back of the eye .","They come in two types: rods and cones .","Rods specialise in night vision , while cones specialise in daytime colour vision .","The death of these cells can cause a disease , called retinitis pigmentosa , that leads to vision loss .","Symptoms often start in childhood with a gradual loss of night vision .","Later on , loss of cone photoreceptors can lead to total blindness .","Unfortunately , there are no treatments available that protect photoreceptor cells from dying .","Research has identified drugs that can protect photoreceptors in animal models , but these drugs have failed in humans .","The classic way to look for new treatments is to find drugs that target molecules implicated in a disease , and then test them to see if they are effective .","Unfortunately , many drugs identified in this way fail in later stages of testing , either because they are ineffective , or because they have unacceptable side effects .","One way to reverse this trend is to first test whether a drug is effective at curing a disease in animals , and later determining what it does at a molecular level .","This could reveal whether drugs can protect photoreceptors before research to discover their molecular targets begins .","Tests like this across different species could maximise the chances of finding a drug that works in humans , because if a drug works in several species , it is more likely to have shared target molecules across species .","Applying this reasoning , Zhang et al . tested around 3 , 000 drug candidates for treating retinitis pigmentosa in a strain of zebrafish that undergoes photoreceptor degeneration similar to the human disease .","Most of these drug candidates already have approval for use in humans , meaning that if they were found to be effective for treating retinitis pigmentosa , they could be fast-tracked for use in people .","Zhang et al . found three compounds that helped photoreceptors survive both in zebrafish and in retinas grown in the laboratory derived from a mouse strain with degeneration similar to retinitis pigmentosa .","Tests to find out how these three compounds worked at the molecular level revealed that they interfered with a protein that can trigger cell death .","The tests also found other promising compounds , many of which offered increased protection when combined in pairs .","Worldwide there are between 1 . 5 and 2 . 5 million people with retinitis pigmentosa .","With this disease , loss of vision happens slowly , so identifying drugs that could slow or stop the process could help many people .","These results suggest that placing animal testing earlier in the drug discovery process could complement traditional target-based methods .","The compounds identified here , and the information about how they work , could expand potential treatment research .","The next step in this research is to test whether the drugs identified by Zhang et al . protect mammals other than mice from the degeneration seen in retinitis pigmentosa ."],"string":"[\n \"Photoreceptors are the cells responsible for vision .\",\n \"They are part of the retina: the light-sensing tissue at the back of the eye .\",\n \"They come in two types: rods and cones .\",\n \"Rods specialise in night vision , while cones specialise in daytime colour vision .\",\n \"The death of these cells can cause a disease , called retinitis pigmentosa , that leads to vision loss .\",\n \"Symptoms often start in childhood with a gradual loss of night vision .\",\n \"Later on , loss of cone photoreceptors can lead to total blindness .\",\n \"Unfortunately , there are no treatments available that protect photoreceptor cells from dying .\",\n \"Research has identified drugs that can protect photoreceptors in animal models , but these drugs have failed in humans .\",\n \"The classic way to look for new treatments is to find drugs that target molecules implicated in a disease , and then test them to see if they are effective .\",\n \"Unfortunately , many drugs identified in this way fail in later stages of testing , either because they are ineffective , or because they have unacceptable side effects .\",\n \"One way to reverse this trend is to first test whether a drug is effective at curing a disease in animals , and later determining what it does at a molecular level .\",\n \"This could reveal whether drugs can protect photoreceptors before research to discover their molecular targets begins .\",\n \"Tests like this across different species could maximise the chances of finding a drug that works in humans , because if a drug works in several species , it is more likely to have shared target molecules across species .\",\n \"Applying this reasoning , Zhang et al . tested around 3 , 000 drug candidates for treating retinitis pigmentosa in a strain of zebrafish that undergoes photoreceptor degeneration similar to the human disease .\",\n \"Most of these drug candidates already have approval for use in humans , meaning that if they were found to be effective for treating retinitis pigmentosa , they could be fast-tracked for use in people .\",\n \"Zhang et al . found three compounds that helped photoreceptors survive both in zebrafish and in retinas grown in the laboratory derived from a mouse strain with degeneration similar to retinitis pigmentosa .\",\n \"Tests to find out how these three compounds worked at the molecular level revealed that they interfered with a protein that can trigger cell death .\",\n \"The tests also found other promising compounds , many of which offered increased protection when combined in pairs .\",\n \"Worldwide there are between 1 . 5 and 2 . 5 million people with retinitis pigmentosa .\",\n \"With this disease , loss of vision happens slowly , so identifying drugs that could slow or stop the process could help many people .\",\n \"These results suggest that placing animal testing earlier in the drug discovery process could complement traditional target-based methods .\",\n \"The compounds identified here , and the information about how they work , could expand potential treatment research .\",\n \"The next step in this research is to test whether the drugs identified by Zhang et al . protect mammals other than mice from the degeneration seen in retinitis pigmentosa .\"\n]","isConversational":false},"year":{"kind":"string","value":"2021"}}},{"rowIdx":91,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["developmental biology","genetics and genomics"],"string":"[\n \"developmental biology\",\n \"genetics and genomics\"\n]","isConversational":false},"title":{"kind":"string","value":"Deployment of a retinal determination gene network drives directed cell migration in the sea urchin embryo"},"id":{"kind":"string","value":"elife-08827-v2"},"sections":{"kind":"list like","value":[["Much research has been done to understand the cell biology underlying the events of directed cell migration; however , how the events are encoded in the genome and how gene regulatory networks ( GRNs ) control this process are works in progress .","Cell migration is the directed movement of a single cell or a collective group of cells through the early embryo or the adult body .","Many different cell types undergo chemotaxis-dependent directed migration during development , including primordial germ cells ( PGCs ) , Drosophila border cells , the zebrafish posterior lateral line , Drosophila tracheal cells , vertebrate neural crest cells , and vertebrate anterior mesoderm ( Reig et al . , 2014 ) .","The ability for cells to undergo directed migration towards a target location requires the use of different signal transduction mechanisms to remodel their actin cytoskeleton in a directed fashion such that they extend filopodia , lamellipodia , and blebs to create a polarized leading edge .","In the sea urchin , a near complete developmental GRN describes the specification of endomesoderm ( McClay , 2011; Peter and Davidson , 2011 ) .","Studies of this specification network have made the sea urchin a viable model for extending the study of how GRNs can explain control of complex cell behaviors ( Saunders and McClay , 2014 ) .","The migration of the sea urchin small micromeres serves as a powerful experimental model for connecting the genomic regulatory control of morphogenesis to an upstream GRN .","Small micromere cells arise from an asymmetric cleavage of the micromeres at the embryonic fifth cleavage ( Figure 1A ) .","These cells divide once within the vegetal plate to produce eight cells ( Pehrson and Cohen , 1986 ) .","The eight small micromeres migrate along with the growing archenteron during gastrulation until they reach the animal pole ( Yajima and Wessel , 2012; Campanale et al . , 2014 ) .","Post-migration , the small micromeres incorporate into the coelomic pouches , which are found on either side of the developing esophagus ( Hyman , 1955; Pehrson and Cohen , 1986; Luo et al . , 2012 ) . 10 . 7554/eLife . 08827 . 003Figure 1 . Small micromere movements during gastrulation .","( A ) The small micromeres arise at the vegetal pole at the asymmetric fifth cleavage from the micromere lineage ( red ) .","During gastrulation , they remain at the tip of the gut .","They migrate through the top of the blastocoel and enter the posterior half of the coelomic pouches by prism stage .","( B ) To study small micromere movements and migration at a higher resolution , membrane-GFP-labeled micromeres were transplanted to a H2b-RFP-labeled host .","( C ) Small micromeres actively changed shape throughout gastrulation ( extend filopodia and lamellipodia ) until they reach the tip of the archenteron .","Skeletogenic lineages are labeled as ‘Skel’ , and small micromeres are labeled as ‘SM’ .","See also , Video 1 . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 003 The coelomic pouches , a mesodermal sub-type , appear at the tip of the forming archenteron .","Their specification is initiated early in development by Delta/Notch signaling ( Sherwood and McClay , 1999; Sweet et al . , 2002 ) .","During gastrulation , several other mesodermal cell types undergo epithelial-to-mesenchymal transitions ( EMTs ) into the blastocoel where they take on different roles in the embryo .","The mesodermal cell sheet remaining at the tip of the archenteron at the end of gastrulation forms the two coelomic pouches on either side of the foregut ( Figure 1A ) .","Only those small micromeres that reach the left coelomic pouch , which will become the future adult rudiment , will survive until adulthood .","During metamorphosis of the indirect developing sea urchin , the embryonic small micromeres incorporate into the adult rudiment's left somatocoel , which will later give rise to the gonads of the adult animal and possibly other tissues ( Hyman , 1955 ) .","Previous research has suggested that the small micromeres contribute to the adult PGCs; however , it has not been shown directly whether the small micromeres contribute to additional adult tissues or whether the only source of the adult PGCs are the small micromeres ( Pehrson and Cohen , 1986; Voronina et al . , 2008; Juliano et al . , 2010a; Juliano et al . , 2010b; Yajima and Wessel , 2010; Yajima and Wessel , 2012; Wessel et al . , 2014 ) .","The possibility remains that the small micromeres go on to produce multiple cell types including , but not limited to PGCs ( Yajima and Wessel , 2015 ) .","Recent publications tracked small micromeres as they moved from the vegetal pole to the coelomic pouches ( Yajima and Wessel , 2012; Campanale et al . , 2014 ) .","In Yajima et al . , it was observed that during gut invagination , the small micromeres did not change position relative to the adjacent mesoderm cells of the advancing archenteron .","It was concluded that once they reach the tip of the archenteron , the small micromeres must actively migrate to the left and right coelomic pouches ( Yajima and Wessel , 2012 ) .","Seemingly contradictory evidence from Campanale et al . described an active migration throughout gastrulation and post-gastrulation as they make their way to the coelomic pouch ( Campanale et al . , 2014 ) .","While both studies conclude that there is an active migration post-gastrulation , we clarified whether the small micromeres acquired their active movement before or after gastrulation .","When removed from their endogenous location and placed ectopically , we find that the small micromeres are autonomous and active while migrating throughout gastrulation to make it ‘home’ to the coelomic pouch .","Their active motility during gastrulation is essential for the ectopic small micromeres to undergo directed ‘homing’ migration to the coelomic pouch .","In order to understand this homing ability at a systems level , we took advantage of the robust homing nature of the small micromeres to determine the GRN at work during their migration .","Using a homing assay to quantitatively score the ability of the coelomic pouch cells to direct movement of the small micromeres , we constructed a GRN that governs effector genes directly responsible for the homing mechanism .","We find that a specific subcircuit composed of Pax6 , Six3 , Six1/2 , Eya , and Dach1 that is canonically found in retinal development has been deployed for this distinct developmental process .","Here , we describe a GRN for homing behavior and uncover a striking similarity of our morphogenesis GRN to the retinal determination gene network ( RDGN ) ."],["The mechanisms by which the small micromeres make their way to the coelomic pouch is a topic of debate: one paper cites an active migratory event while another a passive translocation during gastrulation ( Yajima and Wessel , 2012; Campanale et al . , 2014 ) .","We followed the migratory process at a higher cellular resolution using time-lapse to resolve the seemingly contradictory data on the movement of small micromeres during gastrulation .","Small micromeres were fluorescently labeled , so they could be uniquely followed throughout gastrulation allowing a finer analysis of their behavior than if those cells were not specifically identified ( Figure 1B and Video 1 ) .","Host embryos were labeled by injection of histone2B-RFP mRNA and micromere-donor embryos with a membrane-GFP mRNA .","At the 16-cell stage , one donor fluorescent micromere was transplanted to the vegetal pole where it replaced a surgically removed micromere ( Figure 1B ) .","Each fourth cleavage micromere gives rise to two lineages: the small micromere ( the proposed primordial germ cell ) and the large micromere ( the skeletogenic mesoderm ) .","The transplant surgery was performed at the 16-cell stage to ensure the survival of the small micromere daughters for two reasons .","First , it assured that the small micromeres ( which arise at fifth cleavage by an asymmetric division of the micromere ) were not damaged by separation from the large micromeres during micromanipulation since they had not yet been born .","Second , when the small micromeres appeared , their only initial connection to the embryo was via the spindle remnant to the large micromere—this attachment to the large micromere was crucial for the incorporation of the transplant into the host embryo .","Both reasons for surgically transplanting the micromere ensured the re-establishment and survival of a high percentage of small micromeres in the host embryo ( Table 1 shows scoring of transplant efficacy ) .","At the beginning of gastrulation , this recombinant approach resulted in 8 fluorescent large micromere progeny and 2 fluorescent small micromeres ( Figure 1C and Video 1 ) .","It was easy to distinguish the donor small micromere progeny from the large micromere progeny when the large micromeres fused with non-fluorescent host mesoderm cells to form the skeletal syncytium thereby greatly diluting their fluorescence ( Figure 1C and Video 1 ) . 10 . 7554/eLife . 08827 . 004Video 1 . Movement of small micromeres throughout gastrulation . Membrane-GFP labeled micromeres were transplanted to a H2b-RFP-labeled host .","An image was acquired every 3 min for 6 hr starting at 12 hr post fertilization .","Videos were projected from multiple z-stacks ( every 3 μm , spanning the blastocoel of the embryo ) using SoftWorx for DeltaVision .","AVIs were then rotated and translated ( due to swimming embryos ) in FIJI .","Frame rate = 5 frames per second . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 00410 . 7554/eLife . 08827 . 005Table 1 . Transplant efficacy for microsurgeriesDOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 005MicrosurgerySuccessful transplants ( % ) Failed transplants ( % ) Ctrl host/Ctrl micromeres ( Figure 5 ) 51 ± 949 ± 9Ctrl host/MO micromeres ( Figure 5 ) 40 ± 2060 ± 20MO host/Ctrl micromeres ( Figure 5 ) 48 ± 1752 ± 17 Chimeric embryos made in this manner were examined over six hours of development to follow cell behavior .","As seen in Figure 1C and Video 1 , throughout invagination of the archenteron , the small micromeres were highly active .","They underwent many cell shape changes , blebbing , and pseudopodia extensions as was also seen by Campanale et al . They extended filopodia and lamellipodia through the basal lamina surrounding the archenteron .","Nevertheless , they retained an adhesion and were not displaced relative to the adjacent invaginating epithelial cells ( Figure 1C and Video 1 , 12 hr 00 min–18 hr 00 min ) .","When gastrulation was completed at 20 hpf ( early prism stage ) , small micromere behavior changed in tandem with a change in the basal lamina .","At a time corresponding to the end of gastrulation , laminin immunostaining indicated that the basal lamina was undergoing remodeling , as there were large gaps in the laminin component of the basement membrane at the tip of the archenteron ( Figure 2A ) .","Through those gaps , the small micromeres delaminated ( Figure 2A ) , moved to the blastocoelar side of the remodeling basal lamina , and then actively migrated to the posterior end of the left or right coelomic pouch where they were then re-enveloped with a basal lamina layer by 2 dpf ( Figure 2B ) .","The migratory behavior of the small micromeres suggested that they underwent an EMT at the beginning of their migratory phase over the archenteron and to their coelomic pouch .","These cells de-adhered , breached a basal lamina , changed shape , actively migrated within the blastocoel , and then rejoined an epithelium , demonstrating a number of EMT , and perhaps mesenchymal–epithelial transition ( MET ) , properties . 10 . 7554/eLife . 08827 . 006Figure 2 . Laminin remodeling at the tip of the archenteron facilitates migration of the small micromeres to the posterior end of the coelomic pouch .","( A ) Once the micromeres reach the tip of the gut , they undergo an epithelial–mesenchymal transition ( EMT ) .","By immunostaining , we see that laminin ( red ) , at the time , is reduced as the small micromeres ( SM , green ) breach the top of the gut .","( B ) Once they reach the posterior coelomic pouch , laminin ( red ) surrounds the small micromeres ( green ) and NSM to encapsulate the coelomic pouch ( yellow dashed circle ) .","( C , D )","Ectopically placed small micromeres underwent an EMT coincident with the endogenous EMT event of the skeletogenic cells .","Laminin ( red ) is absent both at the site of skeletogenic cell ingression ( indicated by white arrows at the site of ingression ) and at the site of ectopic small micromere ( SM , green ) ingression ( indicated by a yellow arrow ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 006 In many embryonic systems , cells home , or migrate in a directed fashion , from their place of origin to a distant target site .","Since that movement is directed , in each case there must be a mechanism to provide guidance ( Richardson and Lehmann , 2010; Yajima and Wessel , 2012; Campanale et al . , 2014; Reig et al . , 2014 ) .","Knowing that the small micromeres were actively changing cell shape then moving to a new location at the end of gastrulation , we tested to see if small micromeres would undergo a directed migration to the coelomic pouch when placed in an ectopic location .","Normally , small micromeres start their trajectory at the vegetal pole , which is the most posterior location in the embryo and terminate in the coelomic pouches near the anterior end of the embryo .","Accordingly , small micromeres were placed ectopically at varying locations along the anterior/posterior axis of 16- to 60-cell staged embryos , and the frequency of their ability to find their way to the coelomic pouch at 2 days post fertilization ( dpf ) was scored ( Figure 3 ) .","Micromeres placed at the vegetal pole served as a control since that is their endogenous starting location .","Donors placed in the endogenous location homed 86 . 6% ( n = 213 ) of the time ( Figure 3A ) .","13 . 4% of observed cases that were considered ‘no homing’ were instances where we could not find fluorescent small micromeres in the host embryo's coelomic pouch .","Those that did not make it to the coelomic pouch became lost in the blastocoelar space , which we scored as ‘no homing’ .","Micromeres were then placed ectopically between the Veg1 and Veg2 tier and in the animal half between the An1 and An2 tier of presumptive ectodermal cells ( scored together as equator ) , at the animal pole , or in the blastocoel .","No matter where they were placed , the vast majority of small micromeres transplanted to ectopic positions anywhere in the embryo made it home ( Equator = 93 . 7% ( n = 74 ) , Animal Pole = 78 . 6% ( n = 11 ) , Blastocoel = 78 . 6% ( n = 11 ) ) ( Figure 3B–D ) .","Interestingly , the majority of transplants were scored to be in the left coelomic pouch after homing suggesting a survival mechanism to ensure the proposed primordial germ cells make it onto adulthood to contribute to the gamete pool . 10 . 7554/eLife . 08827 . 007Figure 3 . Ectopically placed micromeres are able to find their way to the coelomic pouches via a directed homing mechanism from any ectopic location .","( A ) The endogenous , vegetal pole , ( B ) the equator ( p < 0 . 004 ) , ( C ) the animal pole ( p < 0 . 004 ) , or ( D ) inside of the blastocoel ( p < 0 . 004 ) .","Illustrations on the left designate their ectopic placement ( ectopic micromere = green , host embryo = red ) .","Yellow dashed lines indicate the location of the whole coelomic pouch with the ectopically placed micromeres labeled in green .","Graphs depict the percentage of embryos scored with the given ability to home ( Yes vs No ) , ‘n’ equals the number of embryos scored with the phenotype , and p-values were calculated using a χ2 test .","‘**’ denotes a statistically significant ( p < 0 . 05 ) p-value . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 00710 . 7554/eLife . 08827 . 008Figure 3—figure supplement 1 . Ectopic small micromeres arrive at the coelomic pouches independently . Small micromeres were ectopically placed as fourth division micromeres .","The ectopic skeletogenic cells joined the endogenous skeletogenic cells while the ectopic small micromeres reached the coelomic pouch independently of the endogenous small micromeres ( vegetally placed ) no matter their starting location , either the equator ( A , B ) or the animal pole ( C , D ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 008 Next , we wanted to see how the small micromeres homed .","Did they move from the ectopic position within the plane of the epithelium , or did they ingress from the ectopic location into the blastocoel and find the coelomic pouch from there ?","If they did ingress into the blastocoel , were they first attracted to the endogenous small micromeres before migrating to the coelomic pouch mesoderm , or did the small micromeres autonomously move to the coelomic pouches ?","Could it be that the ectopically placed small micromeres precociously ingressed through the basement membrane and thus headed directly to their final destination ?","To see when and where the ectopic small micromeres found their way home , we transplanted differentially labeled micromeres to two different locations ( one at the animal pole [Figure 3—figure supplement 1A] and one at the equator [Figure 3—figure supplement 1B] ) to host embryos .","We then assayed at intervals up to 24 hpf and saw that the ectopically placed micromeres arrived independently of the endogenous micromeres ( Figure 3—figure supplement 1A , B ) .","The transplanted small micromeres remained at the ectopic site until the primary mesenchyme cells ( PMCs ) ingressed .","At that time ( 9–10 hpf ) , the ectopic small micromeres also ingressed into the blastocoel ( Figure 2C , D ) .","Because the ectopic small micromeres ingressed at the time of skeletogenic cell ingression , we wondered whether isolated small micromeres could ingress on their own .","By ectopically transplanting just 32-cell stage small micromeres ( as opposed to the usual 16-cell stage micromere , which would result in large and small micromere cells ) ( Figure 2C ) , we observed the small micromeres autonomously invade through the laminin layer independent of large micromeres .","The ectopic small micromeres breached the laminin layer ( Figure 2D , yellow arrow ) at the same time as the PMC EMT ( Figure 2D , white arrows ) suggesting that the small micromeres create the hole through which they breach and invade , even though the endogenous small micromeres actually breach the basement membrane about 6–8 hr later when they leave the tip of the archenteron .","Once we learned that the ectopically placed small micromeres home to the coelomic pouch from any location in the embryo , we next investigated whether the signal to which they were responding was coming from the coelomic pouch mesoderm .","The question was whether the small micromeres home directly to the coelomic pouch region or indirectly .","First , to ask if the mesoderm is the lineage to which the small micromeres home , we knocked out Delta , a signal necessary for mesodermal specification ( Sherwood and McClay , 1999; Sweet et al . , 2002 ) .","The lack of mesoderm , indeed , had an effect on homing .","In the experiment , Delta was knocked down in host embryos and control ( unperturbed ) micromeres were placed ectopically at the equator of those embryos ( Figure 4A ) .","The small micromeres did not home to any significant location in the embryo ( coelomic pouches were non-existent in the KD ) and became lost in the blastocoelar space as is seen in cases we deemed ‘no homing’ 65% of the time ( n = 23 , p < 0 . 0001 ) ( Figure 4C ) .","35% of the time , there was an incomplete knockdown of the Delta signal resulting in a coelomic pouch and the small micromeres homed to that location .","Further investigation of the cell types and sub-territories within the coelomic pouch allowed us to elaborate on the genomic control underpinning the homing mechanism . 10 . 7554/eLife . 08827 . 009Figure 4 . Specification of the non-skeletogenic mesoderm ( NSM ) by Delta signaling is required for homing . Control micromeres labeled with membrane-GFP were ectopically transplanted to either a control TMR ( red ) host or a Delta–MO TMR ( red ) host ( A ) .","When control micromeres were transplanted to a control host , homing of the small micromeres occurs 100% ( n = 9 ) of the time ( B ) .","When NSM specification was blocked in a host embryo using a delta morpholino , homing of transplanted control micromeres was significantly reduced and only homed 53% of the time ( n = 23 , p < 0 . 0001 ) ( C , white arrows ) .","( D ) The graph depicts the percentage of embryos ( Control Host vs Delta Host ) seen with the given ability to home ( Black = Homing vs Gray = No Homing ) , ‘n’ equals the number of embryos scored in each case , and p-values were calculated using a χ2 test .","‘**’ denotes a statistically significant ( p < 0 . 05 ) p-value . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 009 Next , we wanted to know which transcription factors within each territory of the final coelomic pouch ( aboral or oral mesoderm ) were upstream of the homing mechanism .","We surveyed a list of known coelomic pouch and small micromere transcription factors , and genes found to maintain expression levels during homing were included in our list of potential upstream regulators .","The genes further investigated were ( 1 ) oral coelomic pouch transcription factors FoxC , FoxF , PitX2; ( 2 ) aboral coelomic pouch transcription factors Dach1 , Pax6 , SoxE , Six3 , Six1/2 , and transcriptional co-activator Eya; and ( 3 ) small micromere transcription factors FoxC and Pitx2 .","Accession numbers are found in Table 2 . 10 . 7554/eLife . 08827 . 010Table 2 . Plasmids used in this studyDOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 010PlasmidFragmentAccession numberDach1Full CDSKR181947DeltaFull CDSKR181946EyaFull CDSKR181945FoxCFull CDSKR181944FoxFFull CDSKR181943Pax6Full CDSKR181942PitX2Full CDSKR181941Six3Full CDSKR181940SoxEFull CDSKR181939 The experiment was to transplant a labeled micromere to a differentially labeled host embryo and assay the small micromeres' ability to home in either of the two perturbed scenarios: either with a morpholino-perturbed host embryo or morpholino-perturbed micromere transplant .","For each transcription factor , we transplanted control micromeres to control hosts ( a positive control for homing ) , perturbed micromeres transplanted to control hosts , and control micromeres transplanted to perturbed hosts .","Transplants were done at the 16-cell stage and scored at two days post-fertilization .","Transplant efficacy was also scored and can be found in Table 1 .","Only those transplants containing small micromere were scored for homing or no homing .","Failed transplants were cases where no small micromeres were observed in the embryo .","One transcription factor , Forkhead family member FoxC , was found to affect small micromere homing when only knocked down in the micromere .","FoxC is expressed in both the oral coelomic pouch mesoderm and the small micromeres; however , only when it was perturbed in the small micromere did it affect homing , and only 26% ( n = 19 , p < 0 . 001 ) of the time did the small micromeres successfully make it to the coelomic pouch ( Figure 5A ) . 10 . 7554/eLife . 08827 . 011Figure 5 . Homing is controlled by upstream transcription factors . By selectively knocking down specific transcription factors in either an ectopically placed micromere or the host embryo , transcription factors for the homing were identified .","( A ) FoxC , a transcription factor expressed in the small micromeres and oral mesoderm was necessary for homing of the small micromere ( n = 19 , p < 0 . 001 ) , but had no inhibitory effect on homing when knocked down in just the mesoderm ( n = 13 ) .","( B ) Dach1 , an aboral mesoderm transcription factor , affected homing when knocked down in the mesoderm ( n = 10 , p < 0 . 0001 ) but did not affect homing when perturbed in the small micromere ( n = 13 ) .","( C ) Pax6 , affected homing when knocked down in the mesoderm ( n = 10 , p < 0 . 004 ) but had no effect on homing when perturbed in the small micromere ( n = 14 ) .","( D ) Aboral mesoderm transcription factor , Six3 affected homing when knocked down in the mesoderm ( n = 9 , p < 0 . 00009 ) but had no effect homing when perturbed only in the small micromere ( n = 9 ) .","( E ) Aboral transcriptional co-activator , Eya , affected homing when knocked down in the mesoderm ( n = 7 , p < 0 . 002 ) but did not affect homing when perturbed in the small micromeres ( n = 8 ) .","( F ) Aboral transcription factor , Six1/2 , affected homing when knocked down in the mesoderm , as well ( n = 17 , p < 0 . 02 ) and did not affect homing when knocked down in the micromere ( n = 11 ) .","Yellow circles indicate the location of the coelomic pouch .","White arrows indicate ‘lost’ micromeres .","The graph depicts the percentage of embryos ( Control Host vs MO micromere vs MO Host ) seen with the given ability to home ( Black-Homing vs Gray-No Homing ) , ‘n’ equals the number of embryos scored in each case p-values were calculated using a χ2 test .","‘**’ denotes a statistically significant ( p < 0 . 05 ) p-value . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01110 . 7554/eLife . 08827 . 012Figure 5—figure supplement 1 . Coelomic pouch transcription factors that do not affect homing . Transcription factors FoxF ( A ) , SoxE ( B ) , and PitX2 ( C ) are not seen to affect homing when they are perturbed in either the small micromeres or the NSM .","The graph depicts the percentage of embryos ( Control Host vs Delta Host ) seen with the given ability to home ( Black = Homing vs Gray = No Homing ) , ‘n’ equals the number of embryos scored in each case , and p-values were calculated using a χ2 test .","No p-values were deemed significant . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01210 . 7554/eLife . 08827 . 013Figure 5—figure supplement 2 . Whole embryo perturbation of coelomic pouch transcription factors assayed for vasa expression . Embryos injected with a Control-MO and stained for vasa mRNA expression was seen to appropriately segregate small micromeres in the left and right coelomic pouches ( A ) .","Embryos perturbed for Delta ( B ) , FoxC ( C ) , Dach1 ( D ) , Six3 ( E ) , Six1/2 ( F ) , Eya ( G ) , and Pax6 ( H ) were assayed for vasa expression in the coelomic pouches to test whether migration defects were seen in endogenous knockdown situations .","Small micromeres were seen to properly segregate to coelomic pouches , or in the case of Delta to the side of the archenteron ( since there are no coelomic pouches in a Delta–MO ) in each perturbation . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 013 The first coelomic pouch transcription factor we found to be involved in homing was a Ski/Sno family member , Dachshund1 .","When control micromeres were transplanted to a Dachshund1 morpholino perturbed host , the control small micromeres were only able to home in 10% ( n = 10 , p < 0 . 0001 ) of scored cases ( Figure 5B ) .","In contrast , when Dach1 was knocked down in the small micromeres only , and the host embryo was a control , homing ability was scored as successful in 77% of cases .","We concluded that Dach1 was necessary in the targeted tissue for homing to occur .","Next , we found the paired box family transcription factor Pax6 was upstream of the small micromeres' ability to home .","When control micromeres were transplanted to a Pax6 knockdown host , the small micromeres were only able to home 40% ( n = 10 , p < 0 . 004 ) of the time ( Figure 5C ) .","We then found the six family transcription factor , Six3 , to regulate homing .","When control micromeres were transplanted to a Six3 knockdown host , the small micromeres were able to home 11% ( n = 9 , p < 0 . 00009 ) of scored cases ( Figure 5D ) .","Next , we saw transcriptional co-activator and tyrosine phosphatase , Eya , to exhibit an effect on homing .","When control micromeres were transplanted to an Eya knockdown host , small micromeres were able to home 28 . 5% of the time ( n = 7 , p < 0 . 002 ) ( Figure 5E ) .","Finally , we found another six family transcription factor , Six1/2 , to affect homing .","Control micromeres transplanted to a Six1/2 knockdown host were able to home 65% ( n = 17 , p < 0 . 02 ) of scored cases ( Figure 5F ) .","This effect is modest , but statistically significant relative to controls within that experiment .","Transcription factors FoxF , SoxE , FoxC , and Pitx2 did not appear necessary for homing when perturbed in the coelomic pouch mesoderm ( Figure 5—figure supplement 1 ) .","It is important to note that when the same knockdowns are stained for endogenous vasa expression ( labeling the small micromeres ) , small micromeres are still able to reach the coelomic pouches ( Figure 5—figure supplement 2 ) .","We observe that the control endogenous small micromeres are within the coelomic pouches in tight clusters ( Figure 5—figure supplement 2A ) while the knockdowns exhibit various levels of disorganization ( Figure 5—figure supplement 2B–H ) .","Since the endogenous small micromeres normally home over a very short distance , the knockdown small micromeres , as might be expected , remain close to the coelomic pouches—they do not disperse from that vicinity , but they do not tightly cluster in the correct location .","Also , at this time , there is prevalent cell rearrangement at the tip of the archenteron that leads to the fusion of the oral ectoderm and foregut endoderm ( unpublished observations ) .","Thus of the eleven genes screened for homing behavior , only six were necessary for homing .","We next wanted to better understand how the five coelomic pouch mesodermal transcription factors scored as necessary for homing might relate in the network circuit that runs in that tissue .","Each of the five transcription factors involved in controlling homing is expressed in the aboral coelomic pouch .","Luo and Su showed that transcription factors Dach1 , Eya , Six1/2 , and Pax6 are co-localized in the aboral coelomic pouch ( Luo et al . , 2012 ) .","Six3 was found to be expressed in the aboral mesoderm early in development , and by double fluorescent in situ hybridization , we see six3 to also overlap with aboral marker , eya ( Figure 6B ) ( Wei et al . , 2011 ) .","Six3 was seen to be tightly apposed with vasa expression but is exclusive from small micromeres .","Also , six3 expression is distinct from the oral , myosin , domain of expression ( Figure 6A , C ) .","Expression domains within the coelomic pouch are illustrated in Figure 6D .","Luo and Su also noted that transcription factors Dach1 and FoxC are expressed in the oral coelomic pouch ( Luo et al . , 2012 ) .","Transcription factor FoxC , however , is expressed in both the oral coelomic pouch and the small micromeres at the time of small micromere homing ( Luo et al . , 2012 ) . 10 . 7554/eLife . 08827 . 014Figure 6 . Expression domains within the coelomic pouch . Double fluorescent in situ hybridization of six3 with vasa shows a tight apposition of their expression domains but a lack of co-localization ( A–A′′′′ ) .","Z projection of six3 and vasa is seen in ( A ) , and individual Z sections are seen from aboral to oral most locations in the embryo in ( A′–A′′′′ ) .","Insets on panels A–A′′′′ show a zoomed perspective of just the left coelomic pouch apposed expression of six3 and vasa .","six3 and eya were seen to overlap in the aboral coelomic pouch in both the Z projection ( B ) and Z sections ( B′′ and B′′′ ) but not in Z sections ( B′ or B′′′′ ) .","myosin expression was seen to be in the most oral expression domain of the coelomic pouch , distinct of six3 ( C–C′′′′ ) .","An illustration demonstrating the expression domains observed in the coelomic pouch from our data and the data of Luo and Su , 2012 is displayed in ( D ) from both the lateral and oral views . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01410 . 7554/eLife . 08827 . 015Figure 6—figure supplement 1 . Spatial and temporal expression of homing genes by whole mount in situ hybridization . Dach1 , dachshund , is first detected early in development .","It becomes spatially restricted around mesenchyme blastula to the coelomic pouch mesoderm ( A3 ) .","During gastrulation , it is found throughout and at the tip of the archenteron in the presumptive coelomic pouch cells ( A5-8 ) and is maintained in that tissue and gut throughout coelomic pouch coalescence with the small micromeres at the pluteus stage ( A7-8 ) .","six3's earliest detectable expression is at early blastula ( B1 ) .","It is found in the animal pole region , and expression in the coelomic pouch mesoderm begins around mesenchyme blastula ( B3 ) .","Throughout gastrulation , six3 is found in the apical plate domain and the future coelomic pouch cells ( B3-4 ) .","As was seen in Wei 2009 , the six3 apical domain of expression is responsible for giving rise to neural fates .","Post-gastrulation , six3 is expressed in both coelomic pouches ( B8 ) .","Six1/2 is seen to be expressed in the NSM beginning at mesenchyme blastula ( C3 ) .","It maintains expression in the NSM throughout early development and is in the coelomic pouches by early pluteus ( C7 ) .","Genes activated at mid-gastrula stage at the tip of the archenteron and endure throughout gastrulation include foxc and eya .","mRNA expression is detected in the coelomic pouch mesoderm at the tip of the archenteron and end up in the coelomic pouches .","eya begins expression at mid-gastrula in the aboral coelomic pouch at the tip of the archenteron ( D4 ) .","eya expression by 32 hpf is found in the left coelomic pouch ( D8 ) .","pax6 expression begins at mid to late gastrula stage in the aboral coelomic pouch at the tip of the archenteron , roughly at the time of ectopic small micromere homing ( E5 ) , and by 32 hpf is found in the left coelomic pouch and two lateral patches of ectoderm presumed to be neural in fate ( E8 ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 015 To determine the order of expression of the transcription factors that could potentially act in a homing subcircuit , we first performed whole mount in situ hybridization on all five homing genes ( dach1 , pax6 , eya , six1/2 , and six3 ) at eight time points ranging from pre-hatched blastula ( before small micromere migration along the archenteron ) to early pluteus ( after the small micromeres and coelomic pouches have coalesced ) with a 2- to 4-hr step-time to establish relative timing of expression ( Figure 6—figure supplement 1 ) .","By surveying this gene set , we conclude that their co-localization in the coelomic pouches and temporal expression patterns suggest that these transcription factors may be interacting ( diagram , Figure 6D ) .","To ask how or if these genes interact in a gene network to drive the morphogenetic cassette , we undertook a perturbation analysis to build a homing GRN upstream of the morphogenetic movement .","Each transcription factor was perturbed and the effect of that perturbation on the others was examined .","When Dach1 was perturbed using a morpholino specific to the start of translation , we saw by in situ hybridization an increased expression of dach1 ( indicating self repression ) .","Knowing that dach1 is expressed in the endoderm earlier in development from our time course data , we concluded that the upregulation seen in the gut was a result of dach1 self-repression not only in the gut but also later in the coelomic pouch mesoderm expression .","A down-regulation of six3 and a down-regulation of eya transcripts was also seen in the Dach1 perturbation .","These data indicate that Dach1 positively regulates six3 and eya and inhibits the overexpression of itself by repression ( Figure 7 ) . 10 . 7554/eLife . 08827 . 016Figure 7 . Perturbation analysis unveils regulatory linkages in a homing GRN . Perturbations using a Dach1 morpholino ( MO ) showed Dach1 to be upstream of dach1 , six3 , and eya .","A Six3-MO caused a downregulation of eya and pax6 .","Pertubations with an Eya-MO showed Eya to be upstream of six3 , six1/2 , and eya .","Six1/2 is seen to be upstream of six3 , eya , and six1/2 .","Finally , a Pax6-MO caused a downregulation of six3 , eya , and pax6 .","Arrows represent up or downregulation seen for each panel .","‘n’ equals the total number of embryos scored , and the adjacent percentage designates the percent of embryos scored with the shown effect .","Morpholino morphology phenotypes at 24 hpf and 48 hpf are presented in Figure 7—figure supplement 1 .","Unpublished morpholino data has been validated using a second , distinct morpholino ( see Figure 7—figure supplement 2 ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01610 . 7554/eLife . 08827 . 017Figure 7—figure supplement 1 . Morpholino phenotypes . Morpholino morphologies and general health for all morpholinos used were assayed at 24 and 48 hpf .","Each morpholino was imaged at two time points and multiple focal planes .","GFP images are shown as proof of injection .","Since many of the morphants display a short arm phenotype at 2 dpf , specific focus on their forming skeleton was focused at 24 hpf . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01710 . 7554/eLife . 08827 . 018Figure 7—figure supplement 2 . Morpholino effects seen in this paper that have not previously been validated were confirmed using a second morpholino designed in a location distinct to the site of the first morpholino .","( A ) Using the same Ctrl-MO as was used in Figure 5 , 73% of the time , eya mRNA expression is present in the pattern shown .","( B ) The second morpholino designed to the translation start site of Eya , matched results of the first morpholino with eya mRNA downregulated in 96% of cases .","( C ) The second Pax6 morpholino , also designed to block translation , also matched the perturbation effect of the first morpholino in that 68% of the embryos scored were downregulated for eya mRNA .","N is equal to the number of embryos scored . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 018 When Six3 was perturbed using a start site morpholino , we saw a self up-regulation of six3 but only in the apical expression domain , a down-regulation of eya , and a down-regulation of pax6 transcripts .","We conclude from these data that Six3 positively regulates eya and pax6 in the coelomic pouch mesoderm .","Perturbing Eya in the same way showed a down-regulation of six3 , six1/2 , and a self-down-regulation of eya .","Knowing that Eya is a transcriptional co-activator of Six1/2 , Eya must positively regulate itself , six3 , and six1/2 by co-acting with Six1/2 .","Next , Six1/2 was perturbed , and , as expected , regulatory interactions found reflected those found in the Eya knockdowns indicating that Eya and Six1/2 do , in fact , act synergistically as has been seen before ( Heanue et al . , 1999; Materna et al . , 2013 ) .","Finally , Pax6 perturbation caused down-regulation of six3 , eya , and itself .","Pax6 , therefore , positively regulates six3 , eya , and itself ( Figure 7 ) .","When the network was laid out in a Biotapestry model , the subcircuit controlling the homing mechanism was readily discerned as a coherent feed-forward loop composed of multiple feedback loops that sustain its robust nature ( Figure 8A ) ( Longabaugh et al . , 2005 ) .","It also displayed a striking resemblance to another circuit well known in development: the RDGN .","The Pax-Six-Eya-Dach RDGN is necessary for specification of retinal cell types in many systems ( Purcell et al . , 2005; Kozmik et al . , 2007; Kumar , 2009; Salzer et al . , 2010; Weasner and Kumar , 2012 ) .","The same five factors contribute to a coherent feed-forward circuit for Drosophila eye specification ( Kumar , 2009 ) ( Figure 8B ) .","Few linkages between the eye and small micromere homing circuits are different ( Figure 8 ) . 10 . 7554/eLife . 08827 . 019Figure 8 . Homing GRN subcircuit shows striking resemblance to Drosophila RDGN .","( A ) The perturbation analysis was mapped as a Biotapestry network model .","( B ) A retinal gene network subcircuit extracted from Drosophila ( Kumar , 2009 ) shows a very similar circuit with few regulatory linkage changes in comparison . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01910 . 7554/eLife . 08827 . 020Figure 8—figure supplement 1 . The RDGN subcircuit is conserved throughout evolution . Similar putative GRN models were deduced based on publications in systems also known to utilize RDGN components and assembled in Biotapestry: ( A ) Demosponge canal systems , ( B ) vertebrate skeletal myogenesis , and ( C ) the mouse otic placode . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 020"],["The data from the chimera experiments show that normally small micromeres retain an adhesion to adjacent cells during archenteron extension .","When they arrive at the tip of the archenteron , they then leave the epithelium using an EMT through a laminin-containing basement membrane , migrate to the posterior end of the coelomic pouch , and transition back to an epithelium as they coalesce with the coelomic pouch .","Ectopically placed small micromeres behave differently from the endogenous small micromeres in that they fail to retain epithelial adhesions through gastrulation and instead go through an EMT at the time PMCs ingress .","The fact that the ectopically placed small micromeres lose adhesion and go through EMT independent of the timing of archenteron tip formation implies that there must be a non-autonomous cue to keep the endogenous small micromeres in place during gastrulation .","Without that local cue , the small micromeres actively home from the beginning of gastrulation as seen with the behavior of ectopically placed small micromeres .","Further , there must be a non-autonomous signal that is recognized by small micromeres when they reach the tip of the forming archenteron that cues them to initiate the EMT and migration .","The behavior of the ectopic cells suggest that the migratory capacity can be activated at the beginning of gastrulation , but since they normally do not move from the epithelium until at the tip of the archenteron , another event may stimulate them to complete their journey to the coelomic pouches .","Observations of ectopically placed small micromeres also reveal that the cells undergo directed cell migration to get to the coelomic pouch mesoderm .","The cells find their way to the coelomic pouch no matter their starting location .","In other systems cells become polarized in response to a chemoattractant signal at the leading edge of the cell while the remainder of the cell is cued to receive inhibitory or repulsion signals responsible for increasing its sensitivity to its target location ( as reviewed by Richardson and Lehmann , 2010; Reig et al . , 2014 ) .","Small micromeres appear to take a fairly direct route to their eventual target no matter where they are initially placed .","Since the normal route involves movement through part of the blastocoel , and since ectopic small micromeres also use the blastocoelar route , the putative homing signal likely is produced earlier than the end of gastrulation since the ectopic small micromeres move through the blastocoel to the correct target well before gastrulation is completed .","At present , it is only speculation that there is a chemoattraction mechanism at play .","If it were chemoattraction , the cue must work at the individual cell level since single small micromeres respond to the cue when ectopically placed .","The transcriptional regulation upstream of that putative signal offers approaches for its identification .","Potentially , as an alternative hypothesis , the coelomic pouch NSM could be drawing the small micromeres to their final site by filopodia , since it is known that these cells as well as the small micromeres extend filopodia ( Hardin , 1988; Campanale et al . , 2014 ) .","These hypotheses will need to be explored further .","The goal of determining upstream transcriptional regulation of the homing process requires connection of the aforementioned network ( Figure 8A ) to chemoattraction mechanisms ( both attractant and repulsion ) directly responsible for the directing migration .","In addition , other subcircuits are necessary to drive the EMT , for reception of the putative signal , and for directed movement in response .","Transcriptional control of directed cell migration is not well studied at a systems level .","There are a few examples , however , that suggest the way these systems work .","Findings in the tunicate , Ciona intestinalis , describe the distinct transcriptional regulation of a cellular module that drives migration of heart precursors .","The heart GRN drives membrane protrusions and RhoGTPase signaling , tail retraction , adhesion , and polarity ( Christiaen et al . , 2008 ) .","One important feature of that study was that the cells constitutively expressed many of the factors required for these cell behaviors .","It only took one or two factors to be under direct control of the developmental GRN to regulate the timing , directionality , onset , and termination of the heart cell precursors' journey .","In the sea urchin , five different GRN subcircuits were shown to drive five different cellular components of an EMT ( Saunders and McClay , 2014 ) .","It is highly likely , like Ciona heart cells , that the sea urchin developmental GRN will control directly expression of only a fraction of the proteins used for de-adhesion , motility , invasion of the basement membrane , altered cell polarity , and cell shape changes , all components of an EMT , while many other proteins in those processes will be constitutively expressed .","Using small micromere homing as an assay , we were able to begin to unravel the intricacy of GRN subcircuits that control cell migration to a target site .","This work thus demonstrates how morphogenesis can be controlled by conserved specification GRN subcircuits .","The small micromere homing example reveals something else in morphogenetic movement control .","We tend to think of the result of the movement: the small micromeres become associated with the aboral coelomic pouch through homing .","Yet , to accomplish that feat , both the cells that home and the target cells must enact a series of coordinated , but independent processes .","Here , we found a set of genes that control the production of the putative homing signal .","Being that the subcircuit is of coherent feed-forward topology , we suspect it will be found at the most distal part of a larger GRN; therefore , it will act downstream of the specification circuitry of the signal-producing cell type .","Connecting that upstream subcircuit to the downstream signal ( s ) will provide a tool for explaining one component of the homing mechanism that we know to be transcriptionally controlled .","Recognizable RDGN components are deployed in a variety of developmental processes in different organisms .","After pooling published data from three different fields of study ( demosponge canal systems , vertebrate skeletal muscle , and mouse otic placodes ) , we were able to construct putative GRN models ( Figure 8—figure supplement 1 ) .","Although each respective subcircuit is embedded in a larger specification GRN , when extracted as a subcircuit , it appears that the RDGN feed-forward topology has been deeply conserved ( Torres et al . , 1996; Heanue et al . , 1999; Relaix and Buckingham , 1999; Xu et al . , 1999; Ridgeway and Skerjanc , 2001; Kardon et al . , 2002; Laclef et al . , 2003; Riley and Phillips , 2003; Zheng , 2003; Ozaki , 2004; Silver , 2004; Relaix et al . , 2013; Rivera et al . , 2013 ) .","Only now that we have been able to make a complete RDGN subcircuit in the sea urchin , can we hypothesize about what evolutionary mechanism may be at play .","With the same transcription factors involved in a similar feed-forward process in such a variety of tissues , we hypothesize that the RDGN subcircuit is ancient and has been repeatedly deployed as a unit to provide feed-forward control .","If this subcircuit is redeployed throughout evolution as a unit , as seems to be the case given the frequency of its presence , the mechanism that keeps it intact is unknown .","Evolving as a modular unit must mean that the subcircuit serves as an indispensible or important component in a network full of modular components .","The assembly of the coherent feed-forward circuit , once established , appears to remain intact in separate lineages for the same reason: that circuit must be critical for the GRNs in which it is embedded .","Its presence raises the question of whether there are other ancient or more recent subcircuits that have evolved as modules .","Since analysis of developmental GRNs is a relatively young field , it is possible that a number of ancient modular subcircuits evolve as units to assist in the diversification of cell types and morphogenesis .","One obvious outcome of this mode of inheritance would be a contribution to robustness in developmental mechanisms .","And perhaps once a subcircuit is particularly well integrated and suited for a given type of regulatory function , each of its members ( in our case , transcription factors ) might act as a constraint on the others because any reduction in integration would be less adaptive .","What is intriguing is how ancient the RDGN must be .","It is a circuit that is used because it efficiently provides an excellent feed-forward circuit that stabilizes the process to which it is attached .","It is best studied in retinal determination but its ancestral role is unknown and could equally well be directed cell migration ."],["Adult Lytechinus variegatus were obtained from Reeftopia ( Key West , FL , United States ) or the Duke University Marine Lab ( Beaufort , NC , United States ) .","Gametes were obtained by injection of 0 . 5 M KCl into the adult coelom .","Embryos were cultured at 23°C in artificial seawater ( ASW ) .","The full-length coding sequences for NSM transcription factors , Delta , and RDGN genes were obtained by designing primers against a transcriptome data set .","Nucleotide sequences were annotated and deposited on GenBank ( Accession Numbers: Table 2 ) .","Morpholino antisense oligonucleotides were designed against the start site of translation by GeneTools .","Morpholino sequences and concentrations are as listed in Table 3 .","Morpholinos were diluted in molecular-grade H20 and either FITC or TMR injectable dyes .","Morpholino experiments were carried out on three biological triplicates , cultured at 23°C , and assayed by in situ hybridization .","Phenotypes of morpholinos at 24 hpf and 48 hpf were imaged to show the overall health and morphology of knockdowns ( Figure 7—figure supplement 1 ) .","Published morpholinos listed in Table 3 were verified using appropriate controls explained in the citation referenced .","Two morpholinos were assayed to test the specificity of Six3 , FoxF , and FoxC .","Appropriate efficacy controls were also carried out previously for Six3 , FoxF , FoxC , Delta , Six1/2 , and Dach1 as is cited in Table 3 .","Previously unpublished morpholinos were validated by ordering a second , distinct morpholino to the start of translation or 5′ UTR and were verified using in situ hybridization ( Figure 7—figure supplement 2 ) .","mRNA for injection was transcribed in vivo using Ambion mMessage mMachine .","Concentrations for mRNA injections: 300 ng/μl Histone2B-GFP , 500 ng/μl Histone2B-RFP , 500 ng/μl membrane-RFP , and 300 ng/μl membrane-GFP .","For injection , mRNAs were diluted in 20% glycerol in diethylpyrocarbonate-treated H2O . 10 . 7554/eLife . 08827 . 021Table 3 . Morpholino anti-sense oligonucleotide sequences used in this studyDOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 021MASOMASO sequenceMorpholino typeWorking concentrationLvPax6-MO1GTTGACCTGGCATAGCAGCATTTACTranslation blocking1 . 2 mMLvPax6-MO2CATGTCCCCGTGACCCATAGTTTTCTranslation blocking0 . 3 mMLvEya-MO1GCGCTGAAGCTATTTGACATGCTGTTranslation blocking0 . 75 mMLvEya-MO2GTTGAAACCTGTTTGACTGTAGGCCTranslation blocking1 mMLvSix3-MOATGTTTCCGACTCCGTCCAAACCATTranslation blocking ( Wei et al . , 2009 ) 0 . 75 mMLvSix1/2-MOCCCAAGTCCGTGGCAAGGATAAGATTranslation blocking ( Ransick and Davidson , 2012 ) 0 . 5 mMLvDach1-MOAGTAGGCGGTGGACTTCCCATTTTCTranslation blocking ( Peter and Davidson , 2011 ) 0 . 5 mMLvFoxC-MOTGAAGCGTACATTGGCATGGATGTTTranslation blocking ( Andrikou et al . , 2015 ) 0 . 75 mMLvDelta-MOGTGCAGCCGATTCGTTATTCCTTTTranslation blocking ( Sweet et al . , 2002 ) 0 . 4 mMLvFoxF-MOTCTAATTGAGTCATCTGGAGAGTGTTranslation blocking ( Andrikou et al . , 2015 ) 0 . 75 mMLvSoxE-MOGCTCTAAACTCTCAGGGCTACTCATTranslation blocking0 . 75 mMLvPitX2-MOACTGGTTCATCGCTGCTGATTAATTTranslation blocking0 . 6 mMControl-MOCCTCTTACCTCAGTTACAATTTATATranslation blockingExptDependent At 2 . 5 hpf , one micromere from an injected donor embryo was transplanted onto a differentially injected 16-cell host embryo in the place of one discarded host micromere or at an ectopic location .","Microsurgery was performed with fine glass needles and Narishige micromanipulators .","Detailed methods of injections and transplants were followed as previously described ( Logan et al . , 1999 ) .","Transplant efficacy was scored in Table 1 .","At 11 hpf , embryos were de-ciliated in 2× hypertonic artificial seawater , mounted in 1× artificial seawater containing 10 μM p-methoxy-phenyl isoxazoline ( Semenova et al . , 2011 ) on a slide coated in 1% protamine sulfate and sealed with a combination of Vasoline , lanolin , and paraffinwax ( also known as V . A . L . A . P . ) .","Images were acquired using Coolsnap high-resolution CCD camera on the DeltaVision Elite with 40×/0 . 65–1 . 35 Oil UAPO40X0I3/340 DIC objective .","Images were collected at 3-min intervals beginning at 12 hpf and ending at 18 hpf or later and projected as videos using SoftWorx for DeltaVision .","Embryos were fixed in 4% paraformaldehyde for 10 min , washed in PBST , blocked in 4% normal goat serum ( in PBST ) for 45 min , and incubated in polyclonal anti-Laminin antibody ( 1:250 ) ( ab11575; Abcam [Cambridge , MA , United States] ) and anti-GFP ( 1:2000 ) ( ab13970; Abcam ) overnight at 4°C .","Embryos were then washed in PBST , incubated in a 1:200 dilution of either Cy2- or Cy3-conjugated secondary antibody , and finally , stained with Hoechst's ( 1:1000 ) ( Molecular Probes [Eugene , OR , United States] ) to label all nuclei .","Images were acquired using Zeiss LSM 510 upright confocal with a 40×/1 . 4 Oil Plan-Apochromat objective .","Z-slices were spaced 1 . 0 μm apart spanning the diameter of the embryo .","RNA in situ hybridization was performed using Digoxigenin-11-UTP-labeled probes as previously described ( McIntyre et al . , 2013 ) .","Briefly , embryos were fixed in 4% paraformaldehyde overnight at 4°C , washed with ASW , and stored in methanol at −20°C .","RNA probes were synthesized in vitro and used at 1 ng/μl .","Hybridization took place at 60–65°C .","Probes were visualized using alkaline phosphatase-conjugated anti-DIG antibody ( 1:1500 , Roche [Indianapolis , IN , United States] ) .","Finally , color was developed using NBT/BCIP ( Roche ) .","For double fluorescent in situ hybridization , a second probe ( labeled with Fluorescein-12-UTP ) was hybridized .","Expression was visualized using the Tyramide Signal Amplification system ( TSA-plus kit , Perkin Elmer [Waltham , MA , United States] ) .","Embryos were visualized with a Zeiss Axioplan2 upright microscope .","GraphPad software was used to analyze a 2 × 2 contingency table with rows corresponding to the condition ( controls and knockdown micromeres or controls and knockdown hosts ) and columns corresponding to outcomes ( ‘homing’ or ‘no homing’ ) .","Chi-squared was then calculated , and a p-value was declared significant at a value of less than 0 . 05 ."]],"string":"[\n [\n \"Much research has been done to understand the cell biology underlying the events of directed cell migration; however , how the events are encoded in the genome and how gene regulatory networks ( GRNs ) control this process are works in progress .\",\n \"Cell migration is the directed movement of a single cell or a collective group of cells through the early embryo or the adult body .\",\n \"Many different cell types undergo chemotaxis-dependent directed migration during development , including primordial germ cells ( PGCs ) , Drosophila border cells , the zebrafish posterior lateral line , Drosophila tracheal cells , vertebrate neural crest cells , and vertebrate anterior mesoderm ( Reig et al . , 2014 ) .\",\n \"The ability for cells to undergo directed migration towards a target location requires the use of different signal transduction mechanisms to remodel their actin cytoskeleton in a directed fashion such that they extend filopodia , lamellipodia , and blebs to create a polarized leading edge .\",\n \"In the sea urchin , a near complete developmental GRN describes the specification of endomesoderm ( McClay , 2011; Peter and Davidson , 2011 ) .\",\n \"Studies of this specification network have made the sea urchin a viable model for extending the study of how GRNs can explain control of complex cell behaviors ( Saunders and McClay , 2014 ) .\",\n \"The migration of the sea urchin small micromeres serves as a powerful experimental model for connecting the genomic regulatory control of morphogenesis to an upstream GRN .\",\n \"Small micromere cells arise from an asymmetric cleavage of the micromeres at the embryonic fifth cleavage ( Figure 1A ) .\",\n \"These cells divide once within the vegetal plate to produce eight cells ( Pehrson and Cohen , 1986 ) .\",\n \"The eight small micromeres migrate along with the growing archenteron during gastrulation until they reach the animal pole ( Yajima and Wessel , 2012; Campanale et al . , 2014 ) .\",\n \"Post-migration , the small micromeres incorporate into the coelomic pouches , which are found on either side of the developing esophagus ( Hyman , 1955; Pehrson and Cohen , 1986; Luo et al . , 2012 ) . 10 . 7554/eLife . 08827 . 003Figure 1 . Small micromere movements during gastrulation .\",\n \"( A ) The small micromeres arise at the vegetal pole at the asymmetric fifth cleavage from the micromere lineage ( red ) .\",\n \"During gastrulation , they remain at the tip of the gut .\",\n \"They migrate through the top of the blastocoel and enter the posterior half of the coelomic pouches by prism stage .\",\n \"( B ) To study small micromere movements and migration at a higher resolution , membrane-GFP-labeled micromeres were transplanted to a H2b-RFP-labeled host .\",\n \"( C ) Small micromeres actively changed shape throughout gastrulation ( extend filopodia and lamellipodia ) until they reach the tip of the archenteron .\",\n \"Skeletogenic lineages are labeled as ‘Skel’ , and small micromeres are labeled as ‘SM’ .\",\n \"See also , Video 1 . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 003 The coelomic pouches , a mesodermal sub-type , appear at the tip of the forming archenteron .\",\n \"Their specification is initiated early in development by Delta/Notch signaling ( Sherwood and McClay , 1999; Sweet et al . , 2002 ) .\",\n \"During gastrulation , several other mesodermal cell types undergo epithelial-to-mesenchymal transitions ( EMTs ) into the blastocoel where they take on different roles in the embryo .\",\n \"The mesodermal cell sheet remaining at the tip of the archenteron at the end of gastrulation forms the two coelomic pouches on either side of the foregut ( Figure 1A ) .\",\n \"Only those small micromeres that reach the left coelomic pouch , which will become the future adult rudiment , will survive until adulthood .\",\n \"During metamorphosis of the indirect developing sea urchin , the embryonic small micromeres incorporate into the adult rudiment's left somatocoel , which will later give rise to the gonads of the adult animal and possibly other tissues ( Hyman , 1955 ) .\",\n \"Previous research has suggested that the small micromeres contribute to the adult PGCs; however , it has not been shown directly whether the small micromeres contribute to additional adult tissues or whether the only source of the adult PGCs are the small micromeres ( Pehrson and Cohen , 1986; Voronina et al . , 2008; Juliano et al . , 2010a; Juliano et al . , 2010b; Yajima and Wessel , 2010; Yajima and Wessel , 2012; Wessel et al . , 2014 ) .\",\n \"The possibility remains that the small micromeres go on to produce multiple cell types including , but not limited to PGCs ( Yajima and Wessel , 2015 ) .\",\n \"Recent publications tracked small micromeres as they moved from the vegetal pole to the coelomic pouches ( Yajima and Wessel , 2012; Campanale et al . , 2014 ) .\",\n \"In Yajima et al . , it was observed that during gut invagination , the small micromeres did not change position relative to the adjacent mesoderm cells of the advancing archenteron .\",\n \"It was concluded that once they reach the tip of the archenteron , the small micromeres must actively migrate to the left and right coelomic pouches ( Yajima and Wessel , 2012 ) .\",\n \"Seemingly contradictory evidence from Campanale et al . described an active migration throughout gastrulation and post-gastrulation as they make their way to the coelomic pouch ( Campanale et al . , 2014 ) .\",\n \"While both studies conclude that there is an active migration post-gastrulation , we clarified whether the small micromeres acquired their active movement before or after gastrulation .\",\n \"When removed from their endogenous location and placed ectopically , we find that the small micromeres are autonomous and active while migrating throughout gastrulation to make it ‘home’ to the coelomic pouch .\",\n \"Their active motility during gastrulation is essential for the ectopic small micromeres to undergo directed ‘homing’ migration to the coelomic pouch .\",\n \"In order to understand this homing ability at a systems level , we took advantage of the robust homing nature of the small micromeres to determine the GRN at work during their migration .\",\n \"Using a homing assay to quantitatively score the ability of the coelomic pouch cells to direct movement of the small micromeres , we constructed a GRN that governs effector genes directly responsible for the homing mechanism .\",\n \"We find that a specific subcircuit composed of Pax6 , Six3 , Six1/2 , Eya , and Dach1 that is canonically found in retinal development has been deployed for this distinct developmental process .\",\n \"Here , we describe a GRN for homing behavior and uncover a striking similarity of our morphogenesis GRN to the retinal determination gene network ( RDGN ) .\"\n ],\n [\n \"The mechanisms by which the small micromeres make their way to the coelomic pouch is a topic of debate: one paper cites an active migratory event while another a passive translocation during gastrulation ( Yajima and Wessel , 2012; Campanale et al . , 2014 ) .\",\n \"We followed the migratory process at a higher cellular resolution using time-lapse to resolve the seemingly contradictory data on the movement of small micromeres during gastrulation .\",\n \"Small micromeres were fluorescently labeled , so they could be uniquely followed throughout gastrulation allowing a finer analysis of their behavior than if those cells were not specifically identified ( Figure 1B and Video 1 ) .\",\n \"Host embryos were labeled by injection of histone2B-RFP mRNA and micromere-donor embryos with a membrane-GFP mRNA .\",\n \"At the 16-cell stage , one donor fluorescent micromere was transplanted to the vegetal pole where it replaced a surgically removed micromere ( Figure 1B ) .\",\n \"Each fourth cleavage micromere gives rise to two lineages: the small micromere ( the proposed primordial germ cell ) and the large micromere ( the skeletogenic mesoderm ) .\",\n \"The transplant surgery was performed at the 16-cell stage to ensure the survival of the small micromere daughters for two reasons .\",\n \"First , it assured that the small micromeres ( which arise at fifth cleavage by an asymmetric division of the micromere ) were not damaged by separation from the large micromeres during micromanipulation since they had not yet been born .\",\n \"Second , when the small micromeres appeared , their only initial connection to the embryo was via the spindle remnant to the large micromere—this attachment to the large micromere was crucial for the incorporation of the transplant into the host embryo .\",\n \"Both reasons for surgically transplanting the micromere ensured the re-establishment and survival of a high percentage of small micromeres in the host embryo ( Table 1 shows scoring of transplant efficacy ) .\",\n \"At the beginning of gastrulation , this recombinant approach resulted in 8 fluorescent large micromere progeny and 2 fluorescent small micromeres ( Figure 1C and Video 1 ) .\",\n \"It was easy to distinguish the donor small micromere progeny from the large micromere progeny when the large micromeres fused with non-fluorescent host mesoderm cells to form the skeletal syncytium thereby greatly diluting their fluorescence ( Figure 1C and Video 1 ) . 10 . 7554/eLife . 08827 . 004Video 1 . Movement of small micromeres throughout gastrulation . Membrane-GFP labeled micromeres were transplanted to a H2b-RFP-labeled host .\",\n \"An image was acquired every 3 min for 6 hr starting at 12 hr post fertilization .\",\n \"Videos were projected from multiple z-stacks ( every 3 μm , spanning the blastocoel of the embryo ) using SoftWorx for DeltaVision .\",\n \"AVIs were then rotated and translated ( due to swimming embryos ) in FIJI .\",\n \"Frame rate = 5 frames per second . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 00410 . 7554/eLife . 08827 . 005Table 1 . Transplant efficacy for microsurgeriesDOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 005MicrosurgerySuccessful transplants ( % ) Failed transplants ( % ) Ctrl host/Ctrl micromeres ( Figure 5 ) 51 ± 949 ± 9Ctrl host/MO micromeres ( Figure 5 ) 40 ± 2060 ± 20MO host/Ctrl micromeres ( Figure 5 ) 48 ± 1752 ± 17 Chimeric embryos made in this manner were examined over six hours of development to follow cell behavior .\",\n \"As seen in Figure 1C and Video 1 , throughout invagination of the archenteron , the small micromeres were highly active .\",\n \"They underwent many cell shape changes , blebbing , and pseudopodia extensions as was also seen by Campanale et al . They extended filopodia and lamellipodia through the basal lamina surrounding the archenteron .\",\n \"Nevertheless , they retained an adhesion and were not displaced relative to the adjacent invaginating epithelial cells ( Figure 1C and Video 1 , 12 hr 00 min–18 hr 00 min ) .\",\n \"When gastrulation was completed at 20 hpf ( early prism stage ) , small micromere behavior changed in tandem with a change in the basal lamina .\",\n \"At a time corresponding to the end of gastrulation , laminin immunostaining indicated that the basal lamina was undergoing remodeling , as there were large gaps in the laminin component of the basement membrane at the tip of the archenteron ( Figure 2A ) .\",\n \"Through those gaps , the small micromeres delaminated ( Figure 2A ) , moved to the blastocoelar side of the remodeling basal lamina , and then actively migrated to the posterior end of the left or right coelomic pouch where they were then re-enveloped with a basal lamina layer by 2 dpf ( Figure 2B ) .\",\n \"The migratory behavior of the small micromeres suggested that they underwent an EMT at the beginning of their migratory phase over the archenteron and to their coelomic pouch .\",\n \"These cells de-adhered , breached a basal lamina , changed shape , actively migrated within the blastocoel , and then rejoined an epithelium , demonstrating a number of EMT , and perhaps mesenchymal–epithelial transition ( MET ) , properties . 10 . 7554/eLife . 08827 . 006Figure 2 . Laminin remodeling at the tip of the archenteron facilitates migration of the small micromeres to the posterior end of the coelomic pouch .\",\n \"( A ) Once the micromeres reach the tip of the gut , they undergo an epithelial–mesenchymal transition ( EMT ) .\",\n \"By immunostaining , we see that laminin ( red ) , at the time , is reduced as the small micromeres ( SM , green ) breach the top of the gut .\",\n \"( B ) Once they reach the posterior coelomic pouch , laminin ( red ) surrounds the small micromeres ( green ) and NSM to encapsulate the coelomic pouch ( yellow dashed circle ) .\",\n \"( C , D )\",\n \"Ectopically placed small micromeres underwent an EMT coincident with the endogenous EMT event of the skeletogenic cells .\",\n \"Laminin ( red ) is absent both at the site of skeletogenic cell ingression ( indicated by white arrows at the site of ingression ) and at the site of ectopic small micromere ( SM , green ) ingression ( indicated by a yellow arrow ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 006 In many embryonic systems , cells home , or migrate in a directed fashion , from their place of origin to a distant target site .\",\n \"Since that movement is directed , in each case there must be a mechanism to provide guidance ( Richardson and Lehmann , 2010; Yajima and Wessel , 2012; Campanale et al . , 2014; Reig et al . , 2014 ) .\",\n \"Knowing that the small micromeres were actively changing cell shape then moving to a new location at the end of gastrulation , we tested to see if small micromeres would undergo a directed migration to the coelomic pouch when placed in an ectopic location .\",\n \"Normally , small micromeres start their trajectory at the vegetal pole , which is the most posterior location in the embryo and terminate in the coelomic pouches near the anterior end of the embryo .\",\n \"Accordingly , small micromeres were placed ectopically at varying locations along the anterior/posterior axis of 16- to 60-cell staged embryos , and the frequency of their ability to find their way to the coelomic pouch at 2 days post fertilization ( dpf ) was scored ( Figure 3 ) .\",\n \"Micromeres placed at the vegetal pole served as a control since that is their endogenous starting location .\",\n \"Donors placed in the endogenous location homed 86 . 6% ( n = 213 ) of the time ( Figure 3A ) .\",\n \"13 . 4% of observed cases that were considered ‘no homing’ were instances where we could not find fluorescent small micromeres in the host embryo's coelomic pouch .\",\n \"Those that did not make it to the coelomic pouch became lost in the blastocoelar space , which we scored as ‘no homing’ .\",\n \"Micromeres were then placed ectopically between the Veg1 and Veg2 tier and in the animal half between the An1 and An2 tier of presumptive ectodermal cells ( scored together as equator ) , at the animal pole , or in the blastocoel .\",\n \"No matter where they were placed , the vast majority of small micromeres transplanted to ectopic positions anywhere in the embryo made it home ( Equator = 93 . 7% ( n = 74 ) , Animal Pole = 78 . 6% ( n = 11 ) , Blastocoel = 78 . 6% ( n = 11 ) ) ( Figure 3B–D ) .\",\n \"Interestingly , the majority of transplants were scored to be in the left coelomic pouch after homing suggesting a survival mechanism to ensure the proposed primordial germ cells make it onto adulthood to contribute to the gamete pool . 10 . 7554/eLife . 08827 . 007Figure 3 . Ectopically placed micromeres are able to find their way to the coelomic pouches via a directed homing mechanism from any ectopic location .\",\n \"( A ) The endogenous , vegetal pole , ( B ) the equator ( p < 0 . 004 ) , ( C ) the animal pole ( p < 0 . 004 ) , or ( D ) inside of the blastocoel ( p < 0 . 004 ) .\",\n \"Illustrations on the left designate their ectopic placement ( ectopic micromere = green , host embryo = red ) .\",\n \"Yellow dashed lines indicate the location of the whole coelomic pouch with the ectopically placed micromeres labeled in green .\",\n \"Graphs depict the percentage of embryos scored with the given ability to home ( Yes vs No ) , ‘n’ equals the number of embryos scored with the phenotype , and p-values were calculated using a χ2 test .\",\n \"‘**’ denotes a statistically significant ( p < 0 . 05 ) p-value . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 00710 . 7554/eLife . 08827 . 008Figure 3—figure supplement 1 . Ectopic small micromeres arrive at the coelomic pouches independently . Small micromeres were ectopically placed as fourth division micromeres .\",\n \"The ectopic skeletogenic cells joined the endogenous skeletogenic cells while the ectopic small micromeres reached the coelomic pouch independently of the endogenous small micromeres ( vegetally placed ) no matter their starting location , either the equator ( A , B ) or the animal pole ( C , D ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 008 Next , we wanted to see how the small micromeres homed .\",\n \"Did they move from the ectopic position within the plane of the epithelium , or did they ingress from the ectopic location into the blastocoel and find the coelomic pouch from there ?\",\n \"If they did ingress into the blastocoel , were they first attracted to the endogenous small micromeres before migrating to the coelomic pouch mesoderm , or did the small micromeres autonomously move to the coelomic pouches ?\",\n \"Could it be that the ectopically placed small micromeres precociously ingressed through the basement membrane and thus headed directly to their final destination ?\",\n \"To see when and where the ectopic small micromeres found their way home , we transplanted differentially labeled micromeres to two different locations ( one at the animal pole [Figure 3—figure supplement 1A] and one at the equator [Figure 3—figure supplement 1B] ) to host embryos .\",\n \"We then assayed at intervals up to 24 hpf and saw that the ectopically placed micromeres arrived independently of the endogenous micromeres ( Figure 3—figure supplement 1A , B ) .\",\n \"The transplanted small micromeres remained at the ectopic site until the primary mesenchyme cells ( PMCs ) ingressed .\",\n \"At that time ( 9–10 hpf ) , the ectopic small micromeres also ingressed into the blastocoel ( Figure 2C , D ) .\",\n \"Because the ectopic small micromeres ingressed at the time of skeletogenic cell ingression , we wondered whether isolated small micromeres could ingress on their own .\",\n \"By ectopically transplanting just 32-cell stage small micromeres ( as opposed to the usual 16-cell stage micromere , which would result in large and small micromere cells ) ( Figure 2C ) , we observed the small micromeres autonomously invade through the laminin layer independent of large micromeres .\",\n \"The ectopic small micromeres breached the laminin layer ( Figure 2D , yellow arrow ) at the same time as the PMC EMT ( Figure 2D , white arrows ) suggesting that the small micromeres create the hole through which they breach and invade , even though the endogenous small micromeres actually breach the basement membrane about 6–8 hr later when they leave the tip of the archenteron .\",\n \"Once we learned that the ectopically placed small micromeres home to the coelomic pouch from any location in the embryo , we next investigated whether the signal to which they were responding was coming from the coelomic pouch mesoderm .\",\n \"The question was whether the small micromeres home directly to the coelomic pouch region or indirectly .\",\n \"First , to ask if the mesoderm is the lineage to which the small micromeres home , we knocked out Delta , a signal necessary for mesodermal specification ( Sherwood and McClay , 1999; Sweet et al . , 2002 ) .\",\n \"The lack of mesoderm , indeed , had an effect on homing .\",\n \"In the experiment , Delta was knocked down in host embryos and control ( unperturbed ) micromeres were placed ectopically at the equator of those embryos ( Figure 4A ) .\",\n \"The small micromeres did not home to any significant location in the embryo ( coelomic pouches were non-existent in the KD ) and became lost in the blastocoelar space as is seen in cases we deemed ‘no homing’ 65% of the time ( n = 23 , p < 0 . 0001 ) ( Figure 4C ) .\",\n \"35% of the time , there was an incomplete knockdown of the Delta signal resulting in a coelomic pouch and the small micromeres homed to that location .\",\n \"Further investigation of the cell types and sub-territories within the coelomic pouch allowed us to elaborate on the genomic control underpinning the homing mechanism . 10 . 7554/eLife . 08827 . 009Figure 4 . Specification of the non-skeletogenic mesoderm ( NSM ) by Delta signaling is required for homing . Control micromeres labeled with membrane-GFP were ectopically transplanted to either a control TMR ( red ) host or a Delta–MO TMR ( red ) host ( A ) .\",\n \"When control micromeres were transplanted to a control host , homing of the small micromeres occurs 100% ( n = 9 ) of the time ( B ) .\",\n \"When NSM specification was blocked in a host embryo using a delta morpholino , homing of transplanted control micromeres was significantly reduced and only homed 53% of the time ( n = 23 , p < 0 . 0001 ) ( C , white arrows ) .\",\n \"( D ) The graph depicts the percentage of embryos ( Control Host vs Delta Host ) seen with the given ability to home ( Black = Homing vs Gray = No Homing ) , ‘n’ equals the number of embryos scored in each case , and p-values were calculated using a χ2 test .\",\n \"‘**’ denotes a statistically significant ( p < 0 . 05 ) p-value . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 009 Next , we wanted to know which transcription factors within each territory of the final coelomic pouch ( aboral or oral mesoderm ) were upstream of the homing mechanism .\",\n \"We surveyed a list of known coelomic pouch and small micromere transcription factors , and genes found to maintain expression levels during homing were included in our list of potential upstream regulators .\",\n \"The genes further investigated were ( 1 ) oral coelomic pouch transcription factors FoxC , FoxF , PitX2; ( 2 ) aboral coelomic pouch transcription factors Dach1 , Pax6 , SoxE , Six3 , Six1/2 , and transcriptional co-activator Eya; and ( 3 ) small micromere transcription factors FoxC and Pitx2 .\",\n \"Accession numbers are found in Table 2 . 10 . 7554/eLife . 08827 . 010Table 2 . Plasmids used in this studyDOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 010PlasmidFragmentAccession numberDach1Full CDSKR181947DeltaFull CDSKR181946EyaFull CDSKR181945FoxCFull CDSKR181944FoxFFull CDSKR181943Pax6Full CDSKR181942PitX2Full CDSKR181941Six3Full CDSKR181940SoxEFull CDSKR181939 The experiment was to transplant a labeled micromere to a differentially labeled host embryo and assay the small micromeres' ability to home in either of the two perturbed scenarios: either with a morpholino-perturbed host embryo or morpholino-perturbed micromere transplant .\",\n \"For each transcription factor , we transplanted control micromeres to control hosts ( a positive control for homing ) , perturbed micromeres transplanted to control hosts , and control micromeres transplanted to perturbed hosts .\",\n \"Transplants were done at the 16-cell stage and scored at two days post-fertilization .\",\n \"Transplant efficacy was also scored and can be found in Table 1 .\",\n \"Only those transplants containing small micromere were scored for homing or no homing .\",\n \"Failed transplants were cases where no small micromeres were observed in the embryo .\",\n \"One transcription factor , Forkhead family member FoxC , was found to affect small micromere homing when only knocked down in the micromere .\",\n \"FoxC is expressed in both the oral coelomic pouch mesoderm and the small micromeres; however , only when it was perturbed in the small micromere did it affect homing , and only 26% ( n = 19 , p < 0 . 001 ) of the time did the small micromeres successfully make it to the coelomic pouch ( Figure 5A ) . 10 . 7554/eLife . 08827 . 011Figure 5 . Homing is controlled by upstream transcription factors . By selectively knocking down specific transcription factors in either an ectopically placed micromere or the host embryo , transcription factors for the homing were identified .\",\n \"( A ) FoxC , a transcription factor expressed in the small micromeres and oral mesoderm was necessary for homing of the small micromere ( n = 19 , p < 0 . 001 ) , but had no inhibitory effect on homing when knocked down in just the mesoderm ( n = 13 ) .\",\n \"( B ) Dach1 , an aboral mesoderm transcription factor , affected homing when knocked down in the mesoderm ( n = 10 , p < 0 . 0001 ) but did not affect homing when perturbed in the small micromere ( n = 13 ) .\",\n \"( C ) Pax6 , affected homing when knocked down in the mesoderm ( n = 10 , p < 0 . 004 ) but had no effect on homing when perturbed in the small micromere ( n = 14 ) .\",\n \"( D ) Aboral mesoderm transcription factor , Six3 affected homing when knocked down in the mesoderm ( n = 9 , p < 0 . 00009 ) but had no effect homing when perturbed only in the small micromere ( n = 9 ) .\",\n \"( E ) Aboral transcriptional co-activator , Eya , affected homing when knocked down in the mesoderm ( n = 7 , p < 0 . 002 ) but did not affect homing when perturbed in the small micromeres ( n = 8 ) .\",\n \"( F ) Aboral transcription factor , Six1/2 , affected homing when knocked down in the mesoderm , as well ( n = 17 , p < 0 . 02 ) and did not affect homing when knocked down in the micromere ( n = 11 ) .\",\n \"Yellow circles indicate the location of the coelomic pouch .\",\n \"White arrows indicate ‘lost’ micromeres .\",\n \"The graph depicts the percentage of embryos ( Control Host vs MO micromere vs MO Host ) seen with the given ability to home ( Black-Homing vs Gray-No Homing ) , ‘n’ equals the number of embryos scored in each case p-values were calculated using a χ2 test .\",\n \"‘**’ denotes a statistically significant ( p < 0 . 05 ) p-value . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01110 . 7554/eLife . 08827 . 012Figure 5—figure supplement 1 . Coelomic pouch transcription factors that do not affect homing . Transcription factors FoxF ( A ) , SoxE ( B ) , and PitX2 ( C ) are not seen to affect homing when they are perturbed in either the small micromeres or the NSM .\",\n \"The graph depicts the percentage of embryos ( Control Host vs Delta Host ) seen with the given ability to home ( Black = Homing vs Gray = No Homing ) , ‘n’ equals the number of embryos scored in each case , and p-values were calculated using a χ2 test .\",\n \"No p-values were deemed significant . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01210 . 7554/eLife . 08827 . 013Figure 5—figure supplement 2 . Whole embryo perturbation of coelomic pouch transcription factors assayed for vasa expression . Embryos injected with a Control-MO and stained for vasa mRNA expression was seen to appropriately segregate small micromeres in the left and right coelomic pouches ( A ) .\",\n \"Embryos perturbed for Delta ( B ) , FoxC ( C ) , Dach1 ( D ) , Six3 ( E ) , Six1/2 ( F ) , Eya ( G ) , and Pax6 ( H ) were assayed for vasa expression in the coelomic pouches to test whether migration defects were seen in endogenous knockdown situations .\",\n \"Small micromeres were seen to properly segregate to coelomic pouches , or in the case of Delta to the side of the archenteron ( since there are no coelomic pouches in a Delta–MO ) in each perturbation . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 013 The first coelomic pouch transcription factor we found to be involved in homing was a Ski/Sno family member , Dachshund1 .\",\n \"When control micromeres were transplanted to a Dachshund1 morpholino perturbed host , the control small micromeres were only able to home in 10% ( n = 10 , p < 0 . 0001 ) of scored cases ( Figure 5B ) .\",\n \"In contrast , when Dach1 was knocked down in the small micromeres only , and the host embryo was a control , homing ability was scored as successful in 77% of cases .\",\n \"We concluded that Dach1 was necessary in the targeted tissue for homing to occur .\",\n \"Next , we found the paired box family transcription factor Pax6 was upstream of the small micromeres' ability to home .\",\n \"When control micromeres were transplanted to a Pax6 knockdown host , the small micromeres were only able to home 40% ( n = 10 , p < 0 . 004 ) of the time ( Figure 5C ) .\",\n \"We then found the six family transcription factor , Six3 , to regulate homing .\",\n \"When control micromeres were transplanted to a Six3 knockdown host , the small micromeres were able to home 11% ( n = 9 , p < 0 . 00009 ) of scored cases ( Figure 5D ) .\",\n \"Next , we saw transcriptional co-activator and tyrosine phosphatase , Eya , to exhibit an effect on homing .\",\n \"When control micromeres were transplanted to an Eya knockdown host , small micromeres were able to home 28 . 5% of the time ( n = 7 , p < 0 . 002 ) ( Figure 5E ) .\",\n \"Finally , we found another six family transcription factor , Six1/2 , to affect homing .\",\n \"Control micromeres transplanted to a Six1/2 knockdown host were able to home 65% ( n = 17 , p < 0 . 02 ) of scored cases ( Figure 5F ) .\",\n \"This effect is modest , but statistically significant relative to controls within that experiment .\",\n \"Transcription factors FoxF , SoxE , FoxC , and Pitx2 did not appear necessary for homing when perturbed in the coelomic pouch mesoderm ( Figure 5—figure supplement 1 ) .\",\n \"It is important to note that when the same knockdowns are stained for endogenous vasa expression ( labeling the small micromeres ) , small micromeres are still able to reach the coelomic pouches ( Figure 5—figure supplement 2 ) .\",\n \"We observe that the control endogenous small micromeres are within the coelomic pouches in tight clusters ( Figure 5—figure supplement 2A ) while the knockdowns exhibit various levels of disorganization ( Figure 5—figure supplement 2B–H ) .\",\n \"Since the endogenous small micromeres normally home over a very short distance , the knockdown small micromeres , as might be expected , remain close to the coelomic pouches—they do not disperse from that vicinity , but they do not tightly cluster in the correct location .\",\n \"Also , at this time , there is prevalent cell rearrangement at the tip of the archenteron that leads to the fusion of the oral ectoderm and foregut endoderm ( unpublished observations ) .\",\n \"Thus of the eleven genes screened for homing behavior , only six were necessary for homing .\",\n \"We next wanted to better understand how the five coelomic pouch mesodermal transcription factors scored as necessary for homing might relate in the network circuit that runs in that tissue .\",\n \"Each of the five transcription factors involved in controlling homing is expressed in the aboral coelomic pouch .\",\n \"Luo and Su showed that transcription factors Dach1 , Eya , Six1/2 , and Pax6 are co-localized in the aboral coelomic pouch ( Luo et al . , 2012 ) .\",\n \"Six3 was found to be expressed in the aboral mesoderm early in development , and by double fluorescent in situ hybridization , we see six3 to also overlap with aboral marker , eya ( Figure 6B ) ( Wei et al . , 2011 ) .\",\n \"Six3 was seen to be tightly apposed with vasa expression but is exclusive from small micromeres .\",\n \"Also , six3 expression is distinct from the oral , myosin , domain of expression ( Figure 6A , C ) .\",\n \"Expression domains within the coelomic pouch are illustrated in Figure 6D .\",\n \"Luo and Su also noted that transcription factors Dach1 and FoxC are expressed in the oral coelomic pouch ( Luo et al . , 2012 ) .\",\n \"Transcription factor FoxC , however , is expressed in both the oral coelomic pouch and the small micromeres at the time of small micromere homing ( Luo et al . , 2012 ) . 10 . 7554/eLife . 08827 . 014Figure 6 . Expression domains within the coelomic pouch . Double fluorescent in situ hybridization of six3 with vasa shows a tight apposition of their expression domains but a lack of co-localization ( A–A′′′′ ) .\",\n \"Z projection of six3 and vasa is seen in ( A ) , and individual Z sections are seen from aboral to oral most locations in the embryo in ( A′–A′′′′ ) .\",\n \"Insets on panels A–A′′′′ show a zoomed perspective of just the left coelomic pouch apposed expression of six3 and vasa .\",\n \"six3 and eya were seen to overlap in the aboral coelomic pouch in both the Z projection ( B ) and Z sections ( B′′ and B′′′ ) but not in Z sections ( B′ or B′′′′ ) .\",\n \"myosin expression was seen to be in the most oral expression domain of the coelomic pouch , distinct of six3 ( C–C′′′′ ) .\",\n \"An illustration demonstrating the expression domains observed in the coelomic pouch from our data and the data of Luo and Su , 2012 is displayed in ( D ) from both the lateral and oral views . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01410 . 7554/eLife . 08827 . 015Figure 6—figure supplement 1 . Spatial and temporal expression of homing genes by whole mount in situ hybridization . Dach1 , dachshund , is first detected early in development .\",\n \"It becomes spatially restricted around mesenchyme blastula to the coelomic pouch mesoderm ( A3 ) .\",\n \"During gastrulation , it is found throughout and at the tip of the archenteron in the presumptive coelomic pouch cells ( A5-8 ) and is maintained in that tissue and gut throughout coelomic pouch coalescence with the small micromeres at the pluteus stage ( A7-8 ) .\",\n \"six3's earliest detectable expression is at early blastula ( B1 ) .\",\n \"It is found in the animal pole region , and expression in the coelomic pouch mesoderm begins around mesenchyme blastula ( B3 ) .\",\n \"Throughout gastrulation , six3 is found in the apical plate domain and the future coelomic pouch cells ( B3-4 ) .\",\n \"As was seen in Wei 2009 , the six3 apical domain of expression is responsible for giving rise to neural fates .\",\n \"Post-gastrulation , six3 is expressed in both coelomic pouches ( B8 ) .\",\n \"Six1/2 is seen to be expressed in the NSM beginning at mesenchyme blastula ( C3 ) .\",\n \"It maintains expression in the NSM throughout early development and is in the coelomic pouches by early pluteus ( C7 ) .\",\n \"Genes activated at mid-gastrula stage at the tip of the archenteron and endure throughout gastrulation include foxc and eya .\",\n \"mRNA expression is detected in the coelomic pouch mesoderm at the tip of the archenteron and end up in the coelomic pouches .\",\n \"eya begins expression at mid-gastrula in the aboral coelomic pouch at the tip of the archenteron ( D4 ) .\",\n \"eya expression by 32 hpf is found in the left coelomic pouch ( D8 ) .\",\n \"pax6 expression begins at mid to late gastrula stage in the aboral coelomic pouch at the tip of the archenteron , roughly at the time of ectopic small micromere homing ( E5 ) , and by 32 hpf is found in the left coelomic pouch and two lateral patches of ectoderm presumed to be neural in fate ( E8 ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 015 To determine the order of expression of the transcription factors that could potentially act in a homing subcircuit , we first performed whole mount in situ hybridization on all five homing genes ( dach1 , pax6 , eya , six1/2 , and six3 ) at eight time points ranging from pre-hatched blastula ( before small micromere migration along the archenteron ) to early pluteus ( after the small micromeres and coelomic pouches have coalesced ) with a 2- to 4-hr step-time to establish relative timing of expression ( Figure 6—figure supplement 1 ) .\",\n \"By surveying this gene set , we conclude that their co-localization in the coelomic pouches and temporal expression patterns suggest that these transcription factors may be interacting ( diagram , Figure 6D ) .\",\n \"To ask how or if these genes interact in a gene network to drive the morphogenetic cassette , we undertook a perturbation analysis to build a homing GRN upstream of the morphogenetic movement .\",\n \"Each transcription factor was perturbed and the effect of that perturbation on the others was examined .\",\n \"When Dach1 was perturbed using a morpholino specific to the start of translation , we saw by in situ hybridization an increased expression of dach1 ( indicating self repression ) .\",\n \"Knowing that dach1 is expressed in the endoderm earlier in development from our time course data , we concluded that the upregulation seen in the gut was a result of dach1 self-repression not only in the gut but also later in the coelomic pouch mesoderm expression .\",\n \"A down-regulation of six3 and a down-regulation of eya transcripts was also seen in the Dach1 perturbation .\",\n \"These data indicate that Dach1 positively regulates six3 and eya and inhibits the overexpression of itself by repression ( Figure 7 ) . 10 . 7554/eLife . 08827 . 016Figure 7 . Perturbation analysis unveils regulatory linkages in a homing GRN . Perturbations using a Dach1 morpholino ( MO ) showed Dach1 to be upstream of dach1 , six3 , and eya .\",\n \"A Six3-MO caused a downregulation of eya and pax6 .\",\n \"Pertubations with an Eya-MO showed Eya to be upstream of six3 , six1/2 , and eya .\",\n \"Six1/2 is seen to be upstream of six3 , eya , and six1/2 .\",\n \"Finally , a Pax6-MO caused a downregulation of six3 , eya , and pax6 .\",\n \"Arrows represent up or downregulation seen for each panel .\",\n \"‘n’ equals the total number of embryos scored , and the adjacent percentage designates the percent of embryos scored with the shown effect .\",\n \"Morpholino morphology phenotypes at 24 hpf and 48 hpf are presented in Figure 7—figure supplement 1 .\",\n \"Unpublished morpholino data has been validated using a second , distinct morpholino ( see Figure 7—figure supplement 2 ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01610 . 7554/eLife . 08827 . 017Figure 7—figure supplement 1 . Morpholino phenotypes . Morpholino morphologies and general health for all morpholinos used were assayed at 24 and 48 hpf .\",\n \"Each morpholino was imaged at two time points and multiple focal planes .\",\n \"GFP images are shown as proof of injection .\",\n \"Since many of the morphants display a short arm phenotype at 2 dpf , specific focus on their forming skeleton was focused at 24 hpf . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01710 . 7554/eLife . 08827 . 018Figure 7—figure supplement 2 . Morpholino effects seen in this paper that have not previously been validated were confirmed using a second morpholino designed in a location distinct to the site of the first morpholino .\",\n \"( A ) Using the same Ctrl-MO as was used in Figure 5 , 73% of the time , eya mRNA expression is present in the pattern shown .\",\n \"( B ) The second morpholino designed to the translation start site of Eya , matched results of the first morpholino with eya mRNA downregulated in 96% of cases .\",\n \"( C ) The second Pax6 morpholino , also designed to block translation , also matched the perturbation effect of the first morpholino in that 68% of the embryos scored were downregulated for eya mRNA .\",\n \"N is equal to the number of embryos scored . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 018 When Six3 was perturbed using a start site morpholino , we saw a self up-regulation of six3 but only in the apical expression domain , a down-regulation of eya , and a down-regulation of pax6 transcripts .\",\n \"We conclude from these data that Six3 positively regulates eya and pax6 in the coelomic pouch mesoderm .\",\n \"Perturbing Eya in the same way showed a down-regulation of six3 , six1/2 , and a self-down-regulation of eya .\",\n \"Knowing that Eya is a transcriptional co-activator of Six1/2 , Eya must positively regulate itself , six3 , and six1/2 by co-acting with Six1/2 .\",\n \"Next , Six1/2 was perturbed , and , as expected , regulatory interactions found reflected those found in the Eya knockdowns indicating that Eya and Six1/2 do , in fact , act synergistically as has been seen before ( Heanue et al . , 1999; Materna et al . , 2013 ) .\",\n \"Finally , Pax6 perturbation caused down-regulation of six3 , eya , and itself .\",\n \"Pax6 , therefore , positively regulates six3 , eya , and itself ( Figure 7 ) .\",\n \"When the network was laid out in a Biotapestry model , the subcircuit controlling the homing mechanism was readily discerned as a coherent feed-forward loop composed of multiple feedback loops that sustain its robust nature ( Figure 8A ) ( Longabaugh et al . , 2005 ) .\",\n \"It also displayed a striking resemblance to another circuit well known in development: the RDGN .\",\n \"The Pax-Six-Eya-Dach RDGN is necessary for specification of retinal cell types in many systems ( Purcell et al . , 2005; Kozmik et al . , 2007; Kumar , 2009; Salzer et al . , 2010; Weasner and Kumar , 2012 ) .\",\n \"The same five factors contribute to a coherent feed-forward circuit for Drosophila eye specification ( Kumar , 2009 ) ( Figure 8B ) .\",\n \"Few linkages between the eye and small micromere homing circuits are different ( Figure 8 ) . 10 . 7554/eLife . 08827 . 019Figure 8 . Homing GRN subcircuit shows striking resemblance to Drosophila RDGN .\",\n \"( A ) The perturbation analysis was mapped as a Biotapestry network model .\",\n \"( B ) A retinal gene network subcircuit extracted from Drosophila ( Kumar , 2009 ) shows a very similar circuit with few regulatory linkage changes in comparison . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 01910 . 7554/eLife . 08827 . 020Figure 8—figure supplement 1 . The RDGN subcircuit is conserved throughout evolution . Similar putative GRN models were deduced based on publications in systems also known to utilize RDGN components and assembled in Biotapestry: ( A ) Demosponge canal systems , ( B ) vertebrate skeletal myogenesis , and ( C ) the mouse otic placode . DOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 020\"\n ],\n [\n \"The data from the chimera experiments show that normally small micromeres retain an adhesion to adjacent cells during archenteron extension .\",\n \"When they arrive at the tip of the archenteron , they then leave the epithelium using an EMT through a laminin-containing basement membrane , migrate to the posterior end of the coelomic pouch , and transition back to an epithelium as they coalesce with the coelomic pouch .\",\n \"Ectopically placed small micromeres behave differently from the endogenous small micromeres in that they fail to retain epithelial adhesions through gastrulation and instead go through an EMT at the time PMCs ingress .\",\n \"The fact that the ectopically placed small micromeres lose adhesion and go through EMT independent of the timing of archenteron tip formation implies that there must be a non-autonomous cue to keep the endogenous small micromeres in place during gastrulation .\",\n \"Without that local cue , the small micromeres actively home from the beginning of gastrulation as seen with the behavior of ectopically placed small micromeres .\",\n \"Further , there must be a non-autonomous signal that is recognized by small micromeres when they reach the tip of the forming archenteron that cues them to initiate the EMT and migration .\",\n \"The behavior of the ectopic cells suggest that the migratory capacity can be activated at the beginning of gastrulation , but since they normally do not move from the epithelium until at the tip of the archenteron , another event may stimulate them to complete their journey to the coelomic pouches .\",\n \"Observations of ectopically placed small micromeres also reveal that the cells undergo directed cell migration to get to the coelomic pouch mesoderm .\",\n \"The cells find their way to the coelomic pouch no matter their starting location .\",\n \"In other systems cells become polarized in response to a chemoattractant signal at the leading edge of the cell while the remainder of the cell is cued to receive inhibitory or repulsion signals responsible for increasing its sensitivity to its target location ( as reviewed by Richardson and Lehmann , 2010; Reig et al . , 2014 ) .\",\n \"Small micromeres appear to take a fairly direct route to their eventual target no matter where they are initially placed .\",\n \"Since the normal route involves movement through part of the blastocoel , and since ectopic small micromeres also use the blastocoelar route , the putative homing signal likely is produced earlier than the end of gastrulation since the ectopic small micromeres move through the blastocoel to the correct target well before gastrulation is completed .\",\n \"At present , it is only speculation that there is a chemoattraction mechanism at play .\",\n \"If it were chemoattraction , the cue must work at the individual cell level since single small micromeres respond to the cue when ectopically placed .\",\n \"The transcriptional regulation upstream of that putative signal offers approaches for its identification .\",\n \"Potentially , as an alternative hypothesis , the coelomic pouch NSM could be drawing the small micromeres to their final site by filopodia , since it is known that these cells as well as the small micromeres extend filopodia ( Hardin , 1988; Campanale et al . , 2014 ) .\",\n \"These hypotheses will need to be explored further .\",\n \"The goal of determining upstream transcriptional regulation of the homing process requires connection of the aforementioned network ( Figure 8A ) to chemoattraction mechanisms ( both attractant and repulsion ) directly responsible for the directing migration .\",\n \"In addition , other subcircuits are necessary to drive the EMT , for reception of the putative signal , and for directed movement in response .\",\n \"Transcriptional control of directed cell migration is not well studied at a systems level .\",\n \"There are a few examples , however , that suggest the way these systems work .\",\n \"Findings in the tunicate , Ciona intestinalis , describe the distinct transcriptional regulation of a cellular module that drives migration of heart precursors .\",\n \"The heart GRN drives membrane protrusions and RhoGTPase signaling , tail retraction , adhesion , and polarity ( Christiaen et al . , 2008 ) .\",\n \"One important feature of that study was that the cells constitutively expressed many of the factors required for these cell behaviors .\",\n \"It only took one or two factors to be under direct control of the developmental GRN to regulate the timing , directionality , onset , and termination of the heart cell precursors' journey .\",\n \"In the sea urchin , five different GRN subcircuits were shown to drive five different cellular components of an EMT ( Saunders and McClay , 2014 ) .\",\n \"It is highly likely , like Ciona heart cells , that the sea urchin developmental GRN will control directly expression of only a fraction of the proteins used for de-adhesion , motility , invasion of the basement membrane , altered cell polarity , and cell shape changes , all components of an EMT , while many other proteins in those processes will be constitutively expressed .\",\n \"Using small micromere homing as an assay , we were able to begin to unravel the intricacy of GRN subcircuits that control cell migration to a target site .\",\n \"This work thus demonstrates how morphogenesis can be controlled by conserved specification GRN subcircuits .\",\n \"The small micromere homing example reveals something else in morphogenetic movement control .\",\n \"We tend to think of the result of the movement: the small micromeres become associated with the aboral coelomic pouch through homing .\",\n \"Yet , to accomplish that feat , both the cells that home and the target cells must enact a series of coordinated , but independent processes .\",\n \"Here , we found a set of genes that control the production of the putative homing signal .\",\n \"Being that the subcircuit is of coherent feed-forward topology , we suspect it will be found at the most distal part of a larger GRN; therefore , it will act downstream of the specification circuitry of the signal-producing cell type .\",\n \"Connecting that upstream subcircuit to the downstream signal ( s ) will provide a tool for explaining one component of the homing mechanism that we know to be transcriptionally controlled .\",\n \"Recognizable RDGN components are deployed in a variety of developmental processes in different organisms .\",\n \"After pooling published data from three different fields of study ( demosponge canal systems , vertebrate skeletal muscle , and mouse otic placodes ) , we were able to construct putative GRN models ( Figure 8—figure supplement 1 ) .\",\n \"Although each respective subcircuit is embedded in a larger specification GRN , when extracted as a subcircuit , it appears that the RDGN feed-forward topology has been deeply conserved ( Torres et al . , 1996; Heanue et al . , 1999; Relaix and Buckingham , 1999; Xu et al . , 1999; Ridgeway and Skerjanc , 2001; Kardon et al . , 2002; Laclef et al . , 2003; Riley and Phillips , 2003; Zheng , 2003; Ozaki , 2004; Silver , 2004; Relaix et al . , 2013; Rivera et al . , 2013 ) .\",\n \"Only now that we have been able to make a complete RDGN subcircuit in the sea urchin , can we hypothesize about what evolutionary mechanism may be at play .\",\n \"With the same transcription factors involved in a similar feed-forward process in such a variety of tissues , we hypothesize that the RDGN subcircuit is ancient and has been repeatedly deployed as a unit to provide feed-forward control .\",\n \"If this subcircuit is redeployed throughout evolution as a unit , as seems to be the case given the frequency of its presence , the mechanism that keeps it intact is unknown .\",\n \"Evolving as a modular unit must mean that the subcircuit serves as an indispensible or important component in a network full of modular components .\",\n \"The assembly of the coherent feed-forward circuit , once established , appears to remain intact in separate lineages for the same reason: that circuit must be critical for the GRNs in which it is embedded .\",\n \"Its presence raises the question of whether there are other ancient or more recent subcircuits that have evolved as modules .\",\n \"Since analysis of developmental GRNs is a relatively young field , it is possible that a number of ancient modular subcircuits evolve as units to assist in the diversification of cell types and morphogenesis .\",\n \"One obvious outcome of this mode of inheritance would be a contribution to robustness in developmental mechanisms .\",\n \"And perhaps once a subcircuit is particularly well integrated and suited for a given type of regulatory function , each of its members ( in our case , transcription factors ) might act as a constraint on the others because any reduction in integration would be less adaptive .\",\n \"What is intriguing is how ancient the RDGN must be .\",\n \"It is a circuit that is used because it efficiently provides an excellent feed-forward circuit that stabilizes the process to which it is attached .\",\n \"It is best studied in retinal determination but its ancestral role is unknown and could equally well be directed cell migration .\"\n ],\n [\n \"Adult Lytechinus variegatus were obtained from Reeftopia ( Key West , FL , United States ) or the Duke University Marine Lab ( Beaufort , NC , United States ) .\",\n \"Gametes were obtained by injection of 0 . 5 M KCl into the adult coelom .\",\n \"Embryos were cultured at 23°C in artificial seawater ( ASW ) .\",\n \"The full-length coding sequences for NSM transcription factors , Delta , and RDGN genes were obtained by designing primers against a transcriptome data set .\",\n \"Nucleotide sequences were annotated and deposited on GenBank ( Accession Numbers: Table 2 ) .\",\n \"Morpholino antisense oligonucleotides were designed against the start site of translation by GeneTools .\",\n \"Morpholino sequences and concentrations are as listed in Table 3 .\",\n \"Morpholinos were diluted in molecular-grade H20 and either FITC or TMR injectable dyes .\",\n \"Morpholino experiments were carried out on three biological triplicates , cultured at 23°C , and assayed by in situ hybridization .\",\n \"Phenotypes of morpholinos at 24 hpf and 48 hpf were imaged to show the overall health and morphology of knockdowns ( Figure 7—figure supplement 1 ) .\",\n \"Published morpholinos listed in Table 3 were verified using appropriate controls explained in the citation referenced .\",\n \"Two morpholinos were assayed to test the specificity of Six3 , FoxF , and FoxC .\",\n \"Appropriate efficacy controls were also carried out previously for Six3 , FoxF , FoxC , Delta , Six1/2 , and Dach1 as is cited in Table 3 .\",\n \"Previously unpublished morpholinos were validated by ordering a second , distinct morpholino to the start of translation or 5′ UTR and were verified using in situ hybridization ( Figure 7—figure supplement 2 ) .\",\n \"mRNA for injection was transcribed in vivo using Ambion mMessage mMachine .\",\n \"Concentrations for mRNA injections: 300 ng/μl Histone2B-GFP , 500 ng/μl Histone2B-RFP , 500 ng/μl membrane-RFP , and 300 ng/μl membrane-GFP .\",\n \"For injection , mRNAs were diluted in 20% glycerol in diethylpyrocarbonate-treated H2O . 10 . 7554/eLife . 08827 . 021Table 3 . Morpholino anti-sense oligonucleotide sequences used in this studyDOI: http://dx . doi . org/10 . 7554/eLife . 08827 . 021MASOMASO sequenceMorpholino typeWorking concentrationLvPax6-MO1GTTGACCTGGCATAGCAGCATTTACTranslation blocking1 . 2 mMLvPax6-MO2CATGTCCCCGTGACCCATAGTTTTCTranslation blocking0 . 3 mMLvEya-MO1GCGCTGAAGCTATTTGACATGCTGTTranslation blocking0 . 75 mMLvEya-MO2GTTGAAACCTGTTTGACTGTAGGCCTranslation blocking1 mMLvSix3-MOATGTTTCCGACTCCGTCCAAACCATTranslation blocking ( Wei et al . , 2009 ) 0 . 75 mMLvSix1/2-MOCCCAAGTCCGTGGCAAGGATAAGATTranslation blocking ( Ransick and Davidson , 2012 ) 0 . 5 mMLvDach1-MOAGTAGGCGGTGGACTTCCCATTTTCTranslation blocking ( Peter and Davidson , 2011 ) 0 . 5 mMLvFoxC-MOTGAAGCGTACATTGGCATGGATGTTTranslation blocking ( Andrikou et al . , 2015 ) 0 . 75 mMLvDelta-MOGTGCAGCCGATTCGTTATTCCTTTTranslation blocking ( Sweet et al . , 2002 ) 0 . 4 mMLvFoxF-MOTCTAATTGAGTCATCTGGAGAGTGTTranslation blocking ( Andrikou et al . , 2015 ) 0 . 75 mMLvSoxE-MOGCTCTAAACTCTCAGGGCTACTCATTranslation blocking0 . 75 mMLvPitX2-MOACTGGTTCATCGCTGCTGATTAATTTranslation blocking0 . 6 mMControl-MOCCTCTTACCTCAGTTACAATTTATATranslation blockingExptDependent At 2 . 5 hpf , one micromere from an injected donor embryo was transplanted onto a differentially injected 16-cell host embryo in the place of one discarded host micromere or at an ectopic location .\",\n \"Microsurgery was performed with fine glass needles and Narishige micromanipulators .\",\n \"Detailed methods of injections and transplants were followed as previously described ( Logan et al . , 1999 ) .\",\n \"Transplant efficacy was scored in Table 1 .\",\n \"At 11 hpf , embryos were de-ciliated in 2× hypertonic artificial seawater , mounted in 1× artificial seawater containing 10 μM p-methoxy-phenyl isoxazoline ( Semenova et al . , 2011 ) on a slide coated in 1% protamine sulfate and sealed with a combination of Vasoline , lanolin , and paraffinwax ( also known as V . A . L . A . P . ) .\",\n \"Images were acquired using Coolsnap high-resolution CCD camera on the DeltaVision Elite with 40×/0 . 65–1 . 35 Oil UAPO40X0I3/340 DIC objective .\",\n \"Images were collected at 3-min intervals beginning at 12 hpf and ending at 18 hpf or later and projected as videos using SoftWorx for DeltaVision .\",\n \"Embryos were fixed in 4% paraformaldehyde for 10 min , washed in PBST , blocked in 4% normal goat serum ( in PBST ) for 45 min , and incubated in polyclonal anti-Laminin antibody ( 1:250 ) ( ab11575; Abcam [Cambridge , MA , United States] ) and anti-GFP ( 1:2000 ) ( ab13970; Abcam ) overnight at 4°C .\",\n \"Embryos were then washed in PBST , incubated in a 1:200 dilution of either Cy2- or Cy3-conjugated secondary antibody , and finally , stained with Hoechst's ( 1:1000 ) ( Molecular Probes [Eugene , OR , United States] ) to label all nuclei .\",\n \"Images were acquired using Zeiss LSM 510 upright confocal with a 40×/1 . 4 Oil Plan-Apochromat objective .\",\n \"Z-slices were spaced 1 . 0 μm apart spanning the diameter of the embryo .\",\n \"RNA in situ hybridization was performed using Digoxigenin-11-UTP-labeled probes as previously described ( McIntyre et al . , 2013 ) .\",\n \"Briefly , embryos were fixed in 4% paraformaldehyde overnight at 4°C , washed with ASW , and stored in methanol at −20°C .\",\n \"RNA probes were synthesized in vitro and used at 1 ng/μl .\",\n \"Hybridization took place at 60–65°C .\",\n \"Probes were visualized using alkaline phosphatase-conjugated anti-DIG antibody ( 1:1500 , Roche [Indianapolis , IN , United States] ) .\",\n \"Finally , color was developed using NBT/BCIP ( Roche ) .\",\n \"For double fluorescent in situ hybridization , a second probe ( labeled with Fluorescein-12-UTP ) was hybridized .\",\n \"Expression was visualized using the Tyramide Signal Amplification system ( TSA-plus kit , Perkin Elmer [Waltham , MA , United States] ) .\",\n \"Embryos were visualized with a Zeiss Axioplan2 upright microscope .\",\n \"GraphPad software was used to analyze a 2 × 2 contingency table with rows corresponding to the condition ( controls and knockdown micromeres or controls and knockdown hosts ) and columns corresponding to outcomes ( ‘homing’ or ‘no homing’ ) .\",\n \"Chi-squared was then calculated , and a p-value was declared significant at a value of less than 0 . 05 .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Gene regulatory networks ( GRNs ) provide a systems-level orchestration of an organism's genome encoded anatomy .","As biological networks are revealed , they continue to answer many questions including knowledge of how GRNs control morphogenetic movements and how GRNs evolve .","The migration of the small micromeres to the coelomic pouches in the sea urchin embryo provides an exceptional model for understanding the genomic regulatory control of morphogenesis .","An assay using the robust homing potential of these cells reveals a ‘coherent feed-forward’ transcriptional subcircuit composed of Pax6 , Six3 , Six1/2 , Eya , and Dach1 that is responsible for the directed homing mechanism of these multipotent progenitors .","The linkages of that circuit are strikingly similar to a circuit involved in retinal specification in Drosophila suggesting that systems-level tasks can be highly conserved even though the tasks drive unrelated processes in different animals ."],"string":"[\n \"Gene regulatory networks ( GRNs ) provide a systems-level orchestration of an organism's genome encoded anatomy .\",\n \"As biological networks are revealed , they continue to answer many questions including knowledge of how GRNs control morphogenetic movements and how GRNs evolve .\",\n \"The migration of the small micromeres to the coelomic pouches in the sea urchin embryo provides an exceptional model for understanding the genomic regulatory control of morphogenesis .\",\n \"An assay using the robust homing potential of these cells reveals a ‘coherent feed-forward’ transcriptional subcircuit composed of Pax6 , Six3 , Six1/2 , Eya , and Dach1 that is responsible for the directed homing mechanism of these multipotent progenitors .\",\n \"The linkages of that circuit are strikingly similar to a circuit involved in retinal specification in Drosophila suggesting that systems-level tasks can be highly conserved even though the tasks drive unrelated processes in different animals .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Within an animal embryo , groups of cells tend to move , or migrate , between different areas before they form into tissues and organs .","These cell migrations are regulated by hundreds of genes , which must be expressed at the right time and in the right place .","Cells use proteins called transcription factors to regulate the expression of genes .","These proteins work together in circuit board-like networks called gene regulatory networks in order to drive different aspects of development , including cell migration .","The sea urchin is a useful model organism to study how animals develop .","This is because these marine animals express many of the same genes as humans , but they can be easily manipulated and studied in the laboratory .","In a developing sea urchin embryo , cells called the small micromeres move towards one end of animal and get incorporated into a pocket-like structure known as the coelomic pouch .","From this pouch , these cells mature and eventually contribute to the adult germ cells ( the precursors to the sperm and eggs ) .","Martik and McClay have now analyzed how small micromeres make their way to their final location in the coelomic pouch .","Micromeres were labeled with a dye that fluoresces green so that they could be tracked under a microscope .","This revealed that , like other moving cells , micromeres actively change their shape as they migrate .","Furthermore , when micromeres were experimentally moved to abnormal locations in the sea urchin embryo , they were still able to actively home in on the coelomic pouch no matter their starting location .","Martik and McClay then identified five transcription factors expressed in the coelomic pouch in the sea urchin that are involved in this homing activity .","Reducing the expression of any of these transcription factors was enough to hinder the ability of the micromeres to find their way to the coelomic pouch .","Further experiments and analysis then revealed that these five transcription factors work together in a sub-circuit , which is in turn embedded in a larger gene regulatory network .","This sub-circuit that drives cell migration is unexpectedly similar to another circuit in the fruit fly Drosophila .","Intriguingly , the sub-circuit in the fly controls eye development , which is unrelated to cell homing and migration .","These observations raise the possibility that this circuit has been conserved as a unit over millions of years of evolution and redeployed in new networks under completely different circumstances .","The data also suggest the possibility that additional conserved sub-circuits will be identified as more systems are analyzed in detail ."],"string":"[\n \"Within an animal embryo , groups of cells tend to move , or migrate , between different areas before they form into tissues and organs .\",\n \"These cell migrations are regulated by hundreds of genes , which must be expressed at the right time and in the right place .\",\n \"Cells use proteins called transcription factors to regulate the expression of genes .\",\n \"These proteins work together in circuit board-like networks called gene regulatory networks in order to drive different aspects of development , including cell migration .\",\n \"The sea urchin is a useful model organism to study how animals develop .\",\n \"This is because these marine animals express many of the same genes as humans , but they can be easily manipulated and studied in the laboratory .\",\n \"In a developing sea urchin embryo , cells called the small micromeres move towards one end of animal and get incorporated into a pocket-like structure known as the coelomic pouch .\",\n \"From this pouch , these cells mature and eventually contribute to the adult germ cells ( the precursors to the sperm and eggs ) .\",\n \"Martik and McClay have now analyzed how small micromeres make their way to their final location in the coelomic pouch .\",\n \"Micromeres were labeled with a dye that fluoresces green so that they could be tracked under a microscope .\",\n \"This revealed that , like other moving cells , micromeres actively change their shape as they migrate .\",\n \"Furthermore , when micromeres were experimentally moved to abnormal locations in the sea urchin embryo , they were still able to actively home in on the coelomic pouch no matter their starting location .\",\n \"Martik and McClay then identified five transcription factors expressed in the coelomic pouch in the sea urchin that are involved in this homing activity .\",\n \"Reducing the expression of any of these transcription factors was enough to hinder the ability of the micromeres to find their way to the coelomic pouch .\",\n \"Further experiments and analysis then revealed that these five transcription factors work together in a sub-circuit , which is in turn embedded in a larger gene regulatory network .\",\n \"This sub-circuit that drives cell migration is unexpectedly similar to another circuit in the fruit fly Drosophila .\",\n \"Intriguingly , the sub-circuit in the fly controls eye development , which is unrelated to cell homing and migration .\",\n \"These observations raise the possibility that this circuit has been conserved as a unit over millions of years of evolution and redeployed in new networks under completely different circumstances .\",\n \"The data also suggest the possibility that additional conserved sub-circuits will be identified as more systems are analyzed in detail .\"\n]","isConversational":false},"year":{"kind":"string","value":"2015"}}},{"rowIdx":92,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["biochemistry and chemical biology"],"string":"[\n \"biochemistry and chemical biology\"\n]","isConversational":false},"title":{"kind":"string","value":"A chemical probe of CARM1 alters epigenetic plasticity against breast cancer cell invasion"},"id":{"kind":"string","value":"elife-47110-v3"},"sections":{"kind":"list like","value":[["Numerous biological events are orchestrated epigenetically upon defining cellular fates ( Atlasi and Stunnenberg , 2017; Berdasco and Esteller , 2019 ) .","Among the key epigenetic regulators are protein methyltransferases ( PMTs ) , which can render downstream signals by modifying specific Arg or Lys residues of their substrates with S-adenosyl-L-methionine ( SAM ) as a methyl donor cofactor ( Luo , 2018 ) .","Significant efforts have been made to identify the PMT-dependent epigenetic cues that are dysregulated or addicted under specific disease settings such as cancer ( Berdasco and Esteller , 2019 ) .","Many PMTs are implicated as vulnerable targets against cancer malignancy ( Kaniskan et al . , 2018; Luo , 2018 ) .","The pro-cancerous mechanism of these PMTs can be attributed to their methyltransferase activities , which act individually or in combination to upregulate oncogenes , downregulate tumor suppressors , and maintain cancer-cell-addicted homeostasis ( Berdasco and Esteller , 2019; Blanc and Richard , 2017 ) .","Pharmacological inhibition of these epigenetic events thus presents promising anti-cancer strategies ( Berdasco and Esteller , 2019 ) , as exemplified by the development of the clinical inhibitors of DOT1L ( Bernt et al . , 2011; Daigle et al . , 2011 ) , EZH2 ( Kim et al . , 2013; Konze et al . , 2013; McCabe et al . , 2012; Qi et al . , 2012; Qi et al . , 2017 ) , and PRMT5 ( Bonday et al . , 2018; Chan-Penebre et al . , 2015 ) .","Protein arginine methyltransferases ( PRMTs ) act on their substrates to yield three different forms of methylated arginine: asymmetric dimethylarginine ( ADMA ) , symmetric dimethylarginine ( SDMA ) , and monomethylarginine ( MMA ) , which are the terminal products of Type I , II and III PRMTs , respectively ( Blanc and Richard , 2017; Yang and Bedford , 2013 ) .","Among the important Type I PRMTs is CARM1 ( PRMT4 ) , which regulates multiple aspects of transcription by methylating diverse targets including RNAPII , SRC3 , C/EBPβ , PAX3/7 , SOX2/9 , RUNX1 , Notch1 , p300 , CBP , p/CIP , Med12 , and BAF155 ( Blanc and Richard , 2017; Hein et al . , 2015; Vu et al . , 2013; Wang et al . , 2015; Wang et al . , 2014a; Yang and Bedford , 2013 ) .","The physiological function of CARM1 has been linked to the differentiation and maturation of embryonic stem cells to form immune cells , adipocytes , chondrocytes , myocytes , and lung tissues ( Blanc and Richard , 2017; Yang and Bedford , 2013 ) .","The requirement of CARM1 is implicated in multiple cancers , with its methyltransferase activity particularly addicted by hematopoietic malignancies and metastatic breast cancer ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018; Wang et al . , 2014a ) .","Our prior efforts using in vivo mouse and in vitro cell models uncovered the role of CARM1 in promoting breast cancer metastasis ( Wang et al . , 2014a ) .","Mechanistically , CARM1 methylates Arg1064 of BAF155 and thus facilitates the recruitment of the BAF155-containing SWI/SNF complex to a specific subset of gene loci that are essential for breast cancer metastasis .","CARM1 thus emerges as a novel anti-cancer target ( Wang et al . , 2014a ) .","Although this cancer relevance inspired the development of CARM1 inhibitors ( Kaniskan et al . , 2018; Scheer et al . , 2019 ) , many small-molecule CARM1 inhibitors lack target selectivity or cellular activity ( Kaniskan et al . , 2018 ) , two essential criteria of chemical probes ( Frye , 2010 ) .","To the best of our knowledge , EZM2302 ( Drew et al . , 2017; Greenblatt et al . , 2018 ) , TP-064 ( Nakayama et al . , 2018 ) and SKI-73 ( 6a in this work , www . thesgc . org/chemical-probes/SKI-73 ) , which were developed by Epizyme , Takeda/SGC ( Structural Genomic Consortium ) , and our team , respectively , are the only selective and cell-active CARM1 chemical probes .","EZM2302 and TP-064 were developed from conventional small-molecule scaffolds occupying the substrate-binding pocket of CARM1 ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018 ) .","The potential utility of EZM2302 and TP-064 is implicated by their selective anti-proliferative effects on hematopoietic cancer cells , in particular multiple myeloma cells ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018 ) .","However , definitive molecular mechanisms of the CARM1 addiction in these contexts remain elusive ( Greenblatt et al . , 2018 ) .","Here , we report the characterization and novel utility of SKI-73 , a chemical probe of CARM1 with pro-drug properties .","SKI-73 ( 6a in this work ) can readily penetrate cell membranes and then be processed into two active CARM1 inhibitors that contain 6′−homosinefungin ( HSF ) as their core scaffold ( Scheer et al . , 2019; Wu et al . , 2016 ) .","Notably , the two inhibitors can accumulate inside cells to remarkably high concentrations and for a prolonged period .","The potency , selectivity , modes of action , on-target engagement , and off-target effects of these compounds were characterized with multiple orthogonal assays in vitro and under cellular settings .","The pharmacological inhibition of CARM1 by SKI-73 ( 6a ) recapitulates the anti-invasion effect of the genetic perturbation of CARM1 .","In the context of cellular heterogeneity , we developed a cell-cycle-aware algorithm for single-cell RNA-seq ( scRNA-seq ) analysis and dissected the invasion-prone subset of breast cancer cells that is sensitive to SKI-73 ( 6a ) treatment .","Our scRNA-seq analysis provides the unprecedented insight that pharmacological inhibition of CARM1 alters epigenetic plasticity and suppresses invasion by suppressing the most invasive subpopulation of breast cancer cells ."],["Upon developing cofactor-competitive PMT inhibitors ( Wu et al . , 2016; Zheng et al . , 2012 ) , we tailored the SAM analog sinefungin ( Figure 1a ) around its 6′-amino moiety to potentially engage CARM1’s substrate-binding pocket .","6′-homosinefungin ( HSF , i . e . , 1 ) , a sinefungin analog with the insertion of 6′−methylene moiety , was discovered for its general high affinity to Type I PRMTs ( Figure 1a , b , Figure 1—figure supplement 1 , Supplementary file 1-Table A ) .","As a SAM mimic , 1 binds to the Type I PRMTs ( namely PRMT1 , CARM1 , PRMT6 and PRMT8 ) with IC50 of 13‒300 nM ( Figure 1a , c , Supplementary file 1-Table A ) .","Its relative affinity to Type I PRMTs aligns with that of the SAM mimics SAH and SNF ( around 20-fold lower IC50 of 1 versus SAH and SNF , Figure 1a , c , Supplementary file 1-Table A ) .","This observation argues that 1 retains the structural features of SAH and SNF to engage PRMTs and meanwhile leverages its 6′-methyleneamine group for additional interaction .","To further explore the 6′-region of HSF , we synthesized HSF derivatives from the same precursor 3 ( Figure 1b , Figure 1—figure supplement 2 ) , by further expanding the 6′-methylene amine moiety with different substituents .","The HSF derivative 2a ( Figure 1b ) was identified for its preferential binding to CARM1 with IC50 = 30 ± 3 nM and >10 fold selectivity over other seven human PRMTs and 26 methyltransferases of other classes ( Figure 1c , Supplementary file 1-Table A ) .","The structural difference between 2a and 1 ( Figure 1b ) suggests that the N-benzyl substituent enables 2a to engage CARM1 through a distinct mechanism ( see results below ) .","With 2a as a lead , we then explored its α-amino carboxylate moiety with different amides from the common precursor three and then the intermediate 4 ( Figure 1—figure supplement 2 ) , which led to the discovery of 5a .","This engagement of CARM1 with 2a is expected to be largely maintained by 5a .","Here , 5a shows an IC50 of 43 ± 7 nM against CARM1 and a >10-fold selectivity over the panel of 33 diverse methyltransferases ( Figure 1c , Supplementary file 1-Table A ) .","In comparison , the negative control compounds 2b ( Bn-SNF ) ( Zheng et al . , 2012 ) and 5b ( Figure 1b , Figure 1—figure supplement 3 ) , which differ from 2a and 5a only by the 6′-methylene group , poorly inhibit CARM1 ( IC50 = 22 ± 1 µM and 1 . 91 ± 0 . 03 µM ) ( Figure 1c , Supplementary file 1-Table A ) .","The dramatic increase of the potency of 2a and 5a in contrast to 2b and 5b supports an essential role of the 6′-methylene moiety upon binding CARM1 .","Distinguished from the SAM mimics SAH , SNF and 1 as nonspecific PMT inhibitors , 2a and 5a were developed as potent and selective SAM analogs .","With 2a and 5a characterized as CARM1 inhibitors , we leveraged orthogonal in vitro assays to explore their modes of interaction ( Figure 2a ) .","To examine whether 2a and 5a are SAM- or substrate-competitive , CARM1 inhibition by 2a and 5a was assessed in the presence of various concentrations of SAM cofactor and H3 peptide substrate ( Figure 2b , c ) .","IC50 values of 2a and 5a showed a linear positive correlation with SAM concentrations , as expected for SAM-competitive inhibitors ( Daigle et al . , 2011; Luo , 2018; Zheng et al . , 2012 ) .","The Kd values of 2a and 5a ( Kd , 2a = 17 ± 8 nM; Kd , 5a = 9 ± 5 nM ) were extrapolated from the y-axis intercepts upon fitting the equation IC50 = [SAM]×Kd/Km , SAM+Kd ( Figure 2b ) ( Segel , 1993 ) .","Km , SAM of 0 . 21 ± 0 . 09 µM and 0 . 28 ± 0 . 14 µM ( an averaged Km , SAM = 0 . 25 µM ) for competition with 2a and 5a , respectively , can also be derived through the ratio of the y-axis intercepts to the slopes ( Figure 2b and Materials and methods ) ( Segel , 1993 ) .","By contrast , the presence of the H3 peptide substrate had negligible effect on the binding of 2a and 5a , indicating their substrate-noncompetitive character ( Figure 2c ) .","The SAM analogs 2a and 5a were thus characterized as SAM-competitive , substrate-noncompetitive inhibitors of CARM1 .","For the direct binding of 2a and 5a to CARM1 , the CARM1-binding kinetics of 2a and 5a were examined using surface plasmon resonance ( SPR ) ( Figure 2d ) .","The SPR signal progression of 2a and 5a fits with a biphasic rather than a mono-phasic binding mode , with the lower Ki1 , 2a = 0 . 06 ± 0 . 02 µM , Ki1 , 5a = 0 . 10 ± 0 . 01 µM , and the higher Ki2 , 5b = 0 . 54 ± 0 . 07 µM , Ki2 , 2a = 0 . 4 ± 0 . 1 µM , probably because of the multi-phase binding kinetics of 2a and 5a ( Figure 2d ) .","To cross validate the binding of 2a and 5a to CARM1 , we conducted an in vitro thermal shift assay , for which ligand binding is expected to increase CARM1’s thermal stability ( Blum et al . , 2014 ) .","The binding of 2a and 5a ( at 5 μM concentration ) increased the melting temperature ( Tm ) of CARM1 by 4 . 4°C and 6 . 5°C , respectively ( Figure 2e , Tm , 2a = 44 . 2 ± 0 . 4 °C and Tm , 5a = 46 . 3 ± 0 . 3 °C versus Tm , DMSO = 39 . 8 ± 0 . 3 °C as control ) .","By contrast , the binding of SAM and 1 show much reduced effects on Tm of CARM1 ( Figure 2e , Tm , SAM = 40 . 1 ± 0 . 3 °C and Tm , 1 = 42 . 8 ± 0 . 4 °C versus Tm , DMSO = 39 . 8 ± 0 . 3 °C ) .","Therefore , although the affinities of 1 , 2a and 5a to CARM1 are comparable ( IC50 = 13‒43 nM , Figure 1c ) , their well-separated effects on Tm suggest that these inhibitors engage CARM1 differentially ( see results below ) .","The two orthogonal biochemical assays thus verified the tight binding of 2a and 5a with CARM1 .","To further seek a structural rationale for 5a and 2a for CARM1 inhibition , we solved the X-ray structure of CARM1 in complex with 5a with resolution of 2 . 00 Å and modeled the CARM1 binding of 2a ( Figure 3 , Materials and methods ) .","The overall topology of the CARM1–5a complex is indistinguishable with the V-shaped subunit of the CARM1 dimer in complex with SNF and 1 ( details in the next section ) , which is typical of the Rossmann fold of Class I methyltransferases ( Figure 3a , b , Table","1 ) ( Luo , 2018 ) .","However , 5a adopts a noncanonical pose with its 6′-N-benzyl moiety in a binding pocket that used to be occupied by the α-amino carboxylate moiety of canonical ligands such as SAH , SNF and 1 ( Figures 3c and 4 ) , while the α-amino methoxyphenethyl amide moiety of 5a protrudes into the substrate-binding pocket ( Boriack-Sjodin et al . , 2016; Sack et al . , 2011 ) .","This noncanonical mode is consistent with the SAM-competitive character of 5a ( Figure 2b ) .","In contrast to the noncanonical mode , Arg168 in the CARM1–5a complex adopts an alternative orientation ( two possible configurations ) , accompanied by an altered conformation of Glu257 , to accommodate the 6′-N-benzyl moiety of 5a ( Figure 3d ) .","The α-amino amide moiety of 5a also engages CARM1 through the combined outcomes of a hydrogen-bond network and hydrophobic interactions with nearby resides ( Figure 3e ) .","Interestingly , the overlaid structures of CARM1 in complex with 5a and a substrate peptide implicate a steric clash and thus a potential for binding competition between 5a and a CARM1 substrate ( Figure 3f ) .","However , the apparent substrate-noncompetitive character of 5a ( Figure 2c ) suggests that this steric clash might be avoided if there is no significant energy penalty when the substrate Arg adopts alternative conformation ( s ) .","The binding mode of the CARM1–2a complex was modeled via molecular docking followed by molecular dynamics ( MD ) simulation ( Materials and methods ) .","Here , we uncovered two distinct poses of 2a ( Binding Pose 1/2 or BP1/2 , Figure 3g , Figure 3—figure supplement 1 ) .","BP1 was characterized by the direct interaction between the α-amino carboxylate moiety of 2a and the guanidinium of Arg168 , whereas BP2 features a titled orientation of Arg168 to accommodate the 6′-N-benzyl moiety of 2a ( Figure 3g , Figure 3—figure supplement 1 ) .","The BP1 and BP2 of 2a closely resemble those of 1 and 5a , respectively , in terms of the orientations of Arg168 and the α-amino carboxylate moiety of the ligands .","When the same modeling protocol was applied to the CARM1–SNF complex , only the canonical pose was identified ( Materials and methods ) .","Energy calculation indicated that both BP1 and BP2 are stable with comparable binding free energies .","Interestingly , the side chain configurations of His414 in both BP1 and BP2 are different from those in the CARM1–5a complex and the CARM1–SNF complex ( Figure 3g ) .","Collectively , 5a and 2a , though structurally related to the SAM analogs 1 and SNF , engage CARM1 via distinct modes of interaction .","Upon comparing the CARM1 structure in complex with 5a and 2a , we observed the additional hydrogen-bond and hydrophobic interactions of 5a that involve its α-amino amide moiety ( Figure 3e ) .","Interestingly , these interactions do not increase but rather decrease the affinity of 5a to CARM1 by two-fold ( Kd , 2a = 17 ± 8 nM versus Kd , 5a = 9 ± 5 nM , Figure 2b ) .","By contrast , there is a significant 10-fold increase of affinity between 2b and 5b ( Figure 1c , Supplementary file 1-Table A ) .","These observations suggest that , although 5b facilitates CARM1’s engagement better than 2b via the former’s α-amino amide moiety , such an effect is dispensed with in the presence of the 6′-methylene ( N-benzyl ) amine moiety of 5a and 2a .","Given the tight CARM1 binding by 1 , we solved the X-ray structure of CARM1 in complex with 1 ( HSF ) with a resolution of 2 . 00 Å ( Figure 4 ) .","The overall folding of the CARM1–1 complex ( PDB: 4IKP ) is similar to those of 1 in complex with SNF and SAH ( PDB: 2Y1X , 2Y1W ) with a V-shape subunit in a dimer of dimers ( Figure 4a , Table","2 ) ( Sack et al . , 2011 ) .","However , the CARM1–1 complex is distinct for multiple configurations of its ligand and interactions via its 6′-methyleneamine moiety ( Figure 4b–d , Figure 4—figure supplement 1 ) .","In the CARM1–1 complex , 1 can adopt four alternative configurations ( Configuration I in Chains B , D; Configuration II in Chain C; Configuration III and IV in Chain A ) accompanied with the structural accommodation of the adjacent residues and water hydrogen bonds ( Figure 4b–d , Supplementary file 1-Table B–D ) .","By contrast , only a single configuration of the ligands was observed in SAH- or SNF-bound CARM1 ( Figure 4c , d , PDB: 2Y1X , 2Y1W ) ( Sack et al . , 2011 ) .","Detailed structural comparison of CARM1 in complex with SNF , SAH and 1 further revealed that 1 maintains common interactions observed in the CARM1–SNF and CARM1–SAH complexes with several exceptions ( Figure 4c , d , Supplementary file 1-Table B–D ) .","Most noticeably , the CARM1–1 complex gains the strong hydrogen bonds via the 6′-methyleneamine moiety of the ligand with","( i ) the backbone carbonyl of CARM1’s Glu258 ( Configuration I , II in Chains B , C , D and Configuration III in Chain A ) or","( ii ) the side chain of Tyr154 ( Configuration IV in Chain A ) , together with several less conserved water hydrogen bonds ( Figure 4c , d , Supplementary file 1-Table B–D ) .","By contrast , the 6′-amine of SNF in the CARM1–SNF complex forms weaker hydrogen bonds with Glu258 and may fewer water hydrogen bonds ( Figure 4c , Supplementary file 1-Table B , C , PDB: 2Y1W ) .","Comparable interactions are completely absent from the CARM1–SAH complex ( Figure 4d , Supplementary file 1-Table B–D , PDB: 2Y1X ) .","The desired hydrogen-bond networks of the 6′-methyleneamine moiety of 1 with CARM1 , which are present in the CARM1-1 complex but absent from the CARM1–SNF and CARM1–SAH complexes , can rationalize the significant decrease of IC50 from SNF and SAH to 1 .","Another key difference among CARM1–1 , CARM1–SNF and CARM1–SAH complexes lies in the region around the carboxylate moiety of these ligands .","In Chains A and C of the CARM1–1 complex , the carboxylate moiety of the ligand forms an ionic bond with Arg169 and a hydrogen bond with Gln160 ( Figure 4b , c , d , Supplementary file 1-Table B , C ) .","Such interactions are absent from CARM1–SNF and CARM1–SAH complexes ( Figure 4d ) .","By contrast , in Chains B and D of the CARM1–1 complex , the same carboxylate moiety forms the ionic bonds with Arg169 and a water hydrogen bond ( Figure 4b , c , d , Supplementary file 1-Table B , C ) .","To accommodate the latter conformation , the Gln160 residue flips toward the 3′-ribosyl hydroxyl moiety of 1 to form a new hydrogen bond ( Figure 4b , c , Supplementary file 1-Table B , C ) .","Similar interaction patterns can also be found in the CARM1–SNF and CARM1–SAH complexes ( Figure 4 , Supplementary file 1-Table B–D ) .","With regards to the rest of the CARM1–ligand interactions , the CARM1 complexes with 1 , SAH and SNF are nearly identical except for slightly altered water hydrogen bonds ( Figure 4 , Supplementary file 1-Table S2–S4 ) .","Here , the α-amino moiety of these ligands forms hydrogen bonds with the carbonyl backbone of Gly193 , as well as two water hydrogen bonds; their 2′ , 3′-ribosyl hydroxyl groups form two hydrogen bonds with the side chain of CARM1’s Glu215; adenine’s N1 and N6 form the hydrogen bonds with Asn243 and Glu244/Ser272 , respectively; and the adenine ring of these ligands is buried within a hydrophobic pocket .","By contrast , there are fewer conserved water hydrogen bonds , such as those involved with the carboxylate and 3′-ribosyl hydroxyl moieties of 1 in Chain A of the CARM1–1 complex ( Figure 4 , Supplementary file 1-Table C ) .","By contrast , adenine-N7 in the CARM1–SNF and CARM1–SAH complexes forms water hydrogen bonds bridged to Ser272 , which are absent from the CARM1–1 complex ( Figure 4 , Supplementary file 1-Table C , D ) .","Collectively , the general high affinity of 1 , SNF and SAH ( Figure 4 , Supplementary file 1-Table B–D ) arises from the combined hydrophilic and hydrophobic interactions of these ligands with CARM1 .","However , in comparison with SNF and SAH , 1 gains the extra interactions via its 6′-methyleneamine moiety ( Figure 4c , d , Supplementary file 1-Table B–D ) .","In addition , 1 adopts the canonical pose with its α-amino carboxylate moiety interacting with Arg168 , which is similar to that of SNF and SAH but different from the noncanonical pose of 2a and 5a , upon binding CARM1 ( Figure 4d ) .","Although the in vitro characterization demonstrated the potency and selectivity of 2a and 5a against CARM1 , we anticipated their poor membrane permeability as observed for structurally related analogs such as SAH and SNF ( Figure 1a ) ( Boriack-Sjodin et al . , 2016; Sack et al . , 2011 ) .","The lack of membrane penetration is probably due to their primary amine moiety , which has pKa of ~10 and is fully protonated at a physiological pH of 7 . 4 .","Given the essential roles of the 9′−amine moiety of 2a and 5a in CARM1 binding ( Figure 3e ) , we envisioned overcoming the membrane permeability issue through a pro-drug strategy by cloaking this amine moiety with a redox-triggered trimethyl-locked quinone butanoate moiety ( TML , Figure 5a ) ( Levine and Raines , 2012 ) .","We thus prepared 6a and its control compound 6b by derivatizing 5a and 5b with the TML moiety ( Figure 1b , Figure 1—figure supplements 1 and 2 ) .","To assess the cellular activity of 6a , we relied on our prior knowledge that CARM1 methylates the Arg1064 of BAF155 , a core component of the SWI/SNF chromatin-remodeling complex , and CARM1 knockout abolishes this posttranslational modification in MCF-7 cells ( Wang et al . , 2014a ) .","Treatment of MCF-7 cells with 10 µM of 6a fully suppressed this methylation mark , whereas treatment with 2a and 5a did not affect this mark ( Figure 5b ) .","We thus demonstrated the prodrug-like cellular activity of 6a .","To further evaluate 6a as a chemical probe against CARM1 , we quantified the efficiency by which 6a engages CARM1 in a cellular context and thus suppresses the CARM1-dependent invasion by breast cancer cells .","Because of the pro-drug character of 6a and its control compound 6b , we first developed quantitative LC-MS/MS methods to examine their cellular fates for CARM1 engagement ( see Materials and methods ) .","Upon treatment of MDA-MB-231 cells with 6a , we observed its time- and dose-dependent intracellular accumulation ( Figure 5c , Figure 5—figure supplements 1 and 2 ) .","We anticipated the conversion of the pro-drug 6a into 5a , but a striking finding is that 6a can also be readily processed into 2a inside cells ( Figure 5c , Figure 5—figure supplements 1 and 2 ) .","Remarkably , >100 μM of 2a can be accumulated inside cells for 2 days after 6 hr treatment with a single dose of 5‒10 µM 6a .","This observation probably reflects a slow efflux and thus effective intracellular retention of 2a due to its polar α-amino acid zwitterion moiety .","Given that cellular CARM1 inhibition is involved with multiple species ( 2a , 5a and 6a ) in competition with SAM , we modeled the ligand occupancy of cellular CARM1 on the basis of their Kd values ( Kd , 2a=17 nM , Kd , 5a=9 nM , Kd , 6a=0 . 28 µM and Kd , SAM≈Km , SAM=0 . 25 µM ) and MS-quantified intracellular concentrations ( Figure 5d , Figure 5—figure supplement 3 , Equations 5–7 , see Materials and methods ) .","The SAM cofactor , whose intracellular concentration was determined to be 89 ± 16 µM ( Figure 5c , d ) , is expected to occupy >99 . 5% CARM1 with residual <0 . 5% as the apo-enzyme under a native setting .","With single doses of 6a of 2 . 5‒10 µM , the combined CARM1 occupancy by 2a , 5a and their pro-drug precursor 6a rapidly reached the plateau of >95% within 6 hr , and was maintained at this level for at least 48 hr ( Figure 5e , Figure 5—figure supplements 1 and 2 ) .","Notably , treatment with 6a concentrations as low as 0 . 5 µM is sufficient to reach 60% target engagement within 10 hr and to maintain this occupancy for 48 hr ( Figure 5e , Figure 5—figure supplements 1 and 2 ) .","The time- and dose-dependent progression of the CARM1 occupancy by these ligands thus provides quantitative guidance upon the treatment of MDA-MB-231 cells with 6a .","Given that 2a is the predominant metabolic product of 6a within cells ( Figure 5c , e ) and also shows certain affinity to SMYD2 ( ~10 fold higher IC50 in comparison with CARM1 , Figure 1c , Supplementary file 1-Table A ) , we evaluated SMYD2 engagement of 2a for its potential off-target effect .","In a similar manner to that described for the ligand occupancy of cellular CARM1 , we modeled the occupancy of cellular SMYD2 by 2a on the basis of Kd , 2a=150 nM and Kd , SAM=60 nM for SMYD2 ( Figure 5e , Figure 5—figure supplement 4 , Materials and methods ) .","Largely because of the high affinity of 2a to SAM and thus a 37-fold larger Kd , 2a/Kd , SAM ratio of SMYD2 relative to CARM1 ( Kd , 2a/Kd , SAM of 2 . 5 and 0 . 068 for SMYD2 and CARM1 , respectively ) , the occupancy of SMYD2 by 2a is below 20% ( Figure 5e , Figure 5—figure supplement 4 ) under the efficacy doses of 6a ( see cellular data for 6a below ) .","We were thus less concerned about the SMYD2-associated off-target effect under our assay conditions .","The metabolic stability of 6a was also evaluated with a microsomal stability assay ( Figure 5c , e , Figure 5—figure supplement 5 , see Materials and methods ) .","In the presence of rat liver microsomes , 6a showed decent stability with 24% residual 6a after one-hour incubation .","Here , the conversion of 6a into 5a accounted for 40% of the microsome-processed 6a; no production of 2a was detected .","Such observation suggests that NQO1 , the putative enzyme candidate to reduce the TML moiety in 6a or 6b , is present in microsomes as well as in tumor cells ( Dias et al . , 2018; Huang et al . , 2016 ) .","By contrast , peptidase enzymes that are expected to process 5a into 2a are absent from microsomes but rich in tumor cells .","We then conducted a cellular thermal shift assay ( CETSA ) , in which ligand binding is expected to increase CARM1’s thermal stability in a cellular context ( Jafari et al . , 2014 ) .","Our data showed that the treatment of MDA-MB-231 cells with 6a but with not the control compound 6b increased cellular Tm and thus the thermal stability of CARM1 by 4 . 3 ± 0 . 6°C ( Figure 5f ) .","The distinct effect of 6a in contrast to 6b on the cellular Tm of CARM1 aligns well with the 4 . 1‒6 . 2°C difference in the in vitro Tm of CARM1 upon binding 2a and 5a versus SAM ( Figure 2e ) .","Here , 6b can penetrate cell membranes and be processed into 5b and 2b in a similar manner as 6a ( Figure 5—figure supplement 6 ) .","These observations thus present the cellular evidence to show that CARM1 engages 2a and 5a .","To further characterize 6a as a CARM1 chemical probe , we examined the Arg1064 methylation of BAF155 and the Arg455/Arg460 methylation of PABP1 , two well-characterized cellular methylation marks of CARM1 , upon treating MDA-MB-231 cells with 6a ( Lee and Bedford , 2002; Wang et al . , 2014a ) .","These methylation marks can be fully suppressed by 6a in a dose-dependent manner ( Figure 6a ) .","The resultant EC50 values of 0 . 45‒0 . 75 µM ( Figure 6b ) are well correlated with the modeled 60% cellular occupancy of CARM1 upon treatment with 0 . 5 µM 6a for 48 hr ( Figure 5e ) .","By contrast , the treatment of the negative control compound 6b showed no effect on these methylation marks ( Figure 6a ) .","We therefore demonstrated the robust use of 6a ( SKI-73 ) as a CARM1 chemical probe and of 6b ( SKI-73N ) as its control compound .","After demonstrating the utility of SKI-73 ( 6a ) as a chemical probe for CARM1 , we examined whether chemical inhibition of CARM1 can recapitulate biological outcomes that are associated with CARM1 knockout ( CARM1-KO ) ( Wang et al . , 2014a ) .","Our prior work showed that CARM1’s methyltransferase activity is required for invasion of MDA-MB-231 cells ( Wang et al . , 2014a ) .","We thus conducted a matrigel invasion assay with MDA-MB-231 cells in the presence of 6a .","Relative to the control treatment with DMSO , treatment with SKI-73 ( 6a ) but not its negative control compound SKI-73N ( 6b ) suppressed the invasion of MDA-MB-231 cells ( EC50 = 1 . 3 µM ) ( Figure 6c , d ) .","The treatment with ≥10 µM 6a produced the maximal 80% suppression of the invasion by MDA-MB-231 relative to the DMSO control , which is comparable with the phenotype of CARM1-KO ( Figure 6e ) .","Critically , no further inhibition by 6a on the invasiveness was observed upon 6a treatment ( in comparison with the treatment with DMSO or 6b treatment ) of MDA-MB-231 CARM1-KO cells ( Figure 6e ) .","Notably , treatment with 6a and 6b under the current condition has no apparent impact on the proliferation of parental or CARM1-KO MDA-MB-231 cells ( Figure 6—figure supplement 1 ) , consistent with the intact proliferation upon treatment with other CARM1 chemical probes ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018 ) .","These results suggest that SKI-73 ( 6a ) and CARM1 knockout perturb the common , proliferation-independent biological process and then suppresses 80% of the invasiveness of MDA-MB-231 cells .","We thus characterized SKI-73 ( 6a ) as a chemical probe that can be used to interrogate the CARM1-dependent invasion of breast cancer cells .","Because of the advancement of scRNA-seq technology , stunning subpopulation heterogeneity has been uncovered even for well-defined cellular types ( Tanay and Regev , 2017 ) .","In the context of tumor metastasis , including its initial invasion step , epigenetic plasticity is required to allow a small subset of tumor cells to adapt distinct transcriptional cues for neo-properties ( Chatterjee et al . , 2018; Flavahan et al . , 2017; Wu et al . , 2019 ) .","To explore the feasibility of dissecting the CARM1-dependent , invasion-prone subset of MDA-MB-231 breast cancer cells , we formulated a cell-cycle-aware algorithm of scRNA-seq analysis and dissected those subpopulations that were sensitive to CARM1 perturbation ( Figure 7a , see Materials and methods ) .","Here we conducted 10 × Genomics droplet-based scRNA-seq of 3232 , 3583 and 4099 individual cells ( a total of 10 , 914 cells ) exposed to 48 hr treatment with SKI-73 ( 6a ) , SKI-73N ( 6b ) and DMSO , respectively .","Guided by Silhouette analysis , cell-cycle-associated transcripts were identified as dominant signatures of subpopulations ( Figure 7—figure supplements 1–18 ) .","These signatures naturally exist for proliferative cells and are not expected to be specific for the invasive phenotype .","To dissect the subpopulation-associated transcriptomic signatures of invasive cells , we included one additional layer for hierarchical clustering by first classifying the individual cells into G0/G1 , S , and G2/M stages ( 6885 , 1520 and 2509 cells , respectively ) ( Figure 7—figure supplement 6 , Supplementary file 1-Table E ) , and then conducted the unsupervised clustering within each cell-cycle-aware subset ( Figure 7b , Figure 7—figure supplements 19–30 , Supplementary file 1-Table E ) .","To resolve the subpopulations without redundant clustering , we developed an entropy analysis method and relied on the Fisher Exact test ( see Materials and methods ) .","The optimal scores of the combined methods were implemented to determine the numbers of cluster for each subset ( Figure 7b , Figure 7—figure supplements 20 , 24 and 28 ) ( Butler et al . , 2018 ) .","The cell-cycle-aware algorithm allowed the clustering of these cells according to the three cell cycle stages under the three treatment conditions and resulted in 21 , 7 and 6 subpopulations in G0/G1 , S , and G2/M phases , respectively ( Figure 7b , Figure 7—figure supplements 21 , 25 and 29 , Supplementary file 1-Tables F–H ) .","Notably , the 48 hr treatments with SKI-73 ( 6a ) or SKI-73N ( 6b ) had no effect on the cell cycle , as indicated by their comparable cell-cycle distribution patterns ( Figure 7—figure supplement 6; Supplementary file 1-Table E ) , a finding that is consistent with intact proliferation after all three treatments ( Figure 6—figure supplement 1 ) .","With the 21 , 7 and 6 subpopulations clustered into the G0/G1 , S , and G2/M stages , respectively , we then conducted population analysis , comparing SKI-73 ( 6a ) and SKI-73N ( 6b ) against DMSO ( Figure 7c , Figure 7—figure supplements 31 and 32 , Supplementary file 1-Table F–H ) .","These subpopulations can be readily classified into five distinct categories according to their cell-cycle-aware responses to SKI-73 ( 6a ) and SKI-73N ( 6b ) treatment: commonly resistant/emerging/depleted versus differentially depleted/emerging ( SKI-73/SKI-73N- or 6a/6b-specific ) ( Figure 7c , Figure 7—figure supplements 31 and 32 , Supplementary file 1-Table F–H ) .","Here , we are particularly interested in the SKI-73 ( 6a ) -specific depleted subpopulations ( Subpopulation 0/2/8/11/13/14/17/19 of G0/G1-phase cells and 3 of S-phase cells ) as the potential invasion-associated subpopulations , given their sensitivity to SKI-73 ( 6a ) but not its control compound SKI-73N ( 6b ) .","The subpopulations that remain unchanged after the treatment with SKI-73 ( 6a ) or SKI-73N ( 6b ) ( Subpopulation 1/7/12/15 of G0/G1-phase cells; 2/4/5 of S-phase cells; 1 of G2/M-phase cells ) were defined as the common resistant subset .","SKI-73 ( 6a ) -specific emerging subpopulations ( Subpopulation 3/4/5/6/16 of G0/G1-phase cells; 6 of S-phase cells; 4 of G2/M-phase cells ) are expected to be suppressed by CARM1 but emerge upon its inhibition .","The remaining subpopulations are either associated with the effects of the small-molecule scaffold of SKI-73 ( 6a ) /SKI-73N ( 6b ) ( commonly emerging Subpopulation 9 of G0/G1-phase cells , 0/5 of G2/M-phase cells; commonly depleted Subpopulation 2/3 of G2/M-phase cells ) or with SKI-73N ( 6b ) -specific effects ( differentially depleted Subpopulation 10/18 of G0/G1-phase cells , 0 of S-phase cells; differentially emerging Subpopulation 20 of G0/G1-phase cells ) .","Interestingly , scRNA-seq analysis of CARM1-KO cells ( in comparison with SKI-73 ( 6a ) -treated cells ) suggests that CARM1 knockout has more profound effects on the overall landscape of epigenetic plasticity ( Figure 7d , Figure 7—figure supplement 33 ) .","Collectively , the chemical probe SKI-73 ( 6a ) alters the epigenetic plasticity of MDA-MB-231 breast cancer cells via the combined effects of SKI-73 ( 6a ) ’s molecular scaffold and specific inhibition of CARM1’s methyltransferase activity .","Given that SKI-73 ( 6a ) has no effect on the cell cycle and proliferation of MDA-MB-231 cells under the current treatment dose and duration , we envision that the invasion capability of MDA-MB-231 cells mainly arises from an invasion-prone subset , 80% of which is depleted by SKI-73 ( 6a ) treatment ( Figure 6c–e ) .","We thus focused on Subpopulations 0/2/8/11/13/14/17/19 of G0/G1-phase cells and Subpopulation 3 of S-phase cells: in total nine depleted subpopulations specific for SKI-73 ( 6a ) ( Figure 7c , Figure 7—figure supplements 31 and 32 , Supplementary file 1-Table F–H ) .","To identify invasion-prone subpopulation ( s ) among these candidates , we compared the transcriptional signature ( s ) of these subpopulations with those of cells that freshly invaded through Matrigel within 16 hr .","Strikingly , in comparison with the highly heterogenous scRNA-seq signature of the parental MDA-MB-231 cells , the freshly harvested invasive cells ( 3793 cells for scRNA-seq ) are relatively homogeneous , with their subpopulations mainly determined by the cell-cycle-related transcriptomic signatures ( Figure 7d , Figure 7—figure supplements 33–37 ) .","We then classified the freshly harvested invasive cells into G0/G1 , S and G2/M stages ( Figure 7—figure supplements 38–40 , Supplementary file 1-Table E ) .","Through correlation analysis comparing the invasive cells and the subpopulations within each cell-cycle stage ( Figure 7d , Figure 7—figure supplements 39–42 ) , we revealed the subsets whose transcriptional signatures closely relate to those of the invasive cells , including Subpopulations 6/7/8/9/14 in G0/G1-phase cells , 0/3 in S-phase cells and 1/2 of G2/M-phase cells ( Supplementary file 1-Table F–K ) .","In the context of population analysis for the nine SKI-73 ( 6a ) -specific depleted subpopulations , Subpopulations 8/14 of G0/G1-phase cells and Subpopulation 3 in S-phase are putative invasion-prone candidates .","Subpopulation 8 of G0/G1-phase cells is the most sensitive and the only subpopulation that can be depleted by around 80% with SKI-73 ( 6a ) treatment ( Figure 7c ) .","Given the ~80% suppression and ~20% residual invasion capability upon SKI-73 ( 6a ) treatment , we argue that the invasive phenotype of MDA-MB-231 cells predominantly arises from Subpopulation 8 G0/G1-phase cells , which only accounts for ~8% of the parental cells in G0/G1 phase ( ~5% without cell-cycle awareness ) .","Differential expression analysis further revealed the single-cell transcriptional signatures of metastasis-implicated genes ( e . g . MORC4 , S100A2 , RPL39 , IFI27 , ARF6 , CHD11 , SDPR and KRT18 ) that are specific for the G0/G1-phase Subpopulation 8 and invasive cells but not for other G0/G1-phase invasion-prone candidates such as Subpopulation 6/7/9/14 ( Figure 7e , Figure 7—figure supplement 43 , Supplementary file 1-Table L ) .","The remaining cells of G0/G1-phase Subpopulation 8 after SKI-73 ( 6a ) treatment ( Figure 7c , d ) together with others ( subpopulation-6/7/9/14 in G0/G1-phase cells , 0/3 in S-phase cells and 1/2 of G2/M-phase cells , Figure 7—figure supplements 31 , 32 , 41 , 42 ) may account for the 20% residual invasion capacity .","In the context of SKI-73 ( 6a ) -specific depletion of G0/G1-phase subpopulations , there are SKI-73 ( 6a ) -specific emerging G0/G1-phase subpopulations: Subpopulation 3/4/5/6/16 ( Figure 7c ) .","Population analysis of G0/G1-phase cells further revealed that Subpopulations 4 and 16 account for 90% of the emerging subset upon SKI-73 ( 6a ) treatment ( Supplementary file 1-Table F ) .","The transcriptional signatures and probably the associated invasion capability of Subpopulations 4 and 16 are dramatically different from those of the freshly harvested invasive cells and the bulk population of the parental cells , including the invasion-prone Subpopulation 8 ( Figure 7b , d ) .","Collectively , either CARM1 knockout or CARM1 inhibition with SKI-73 ( 6a ) alters the epigenetic plasticity in a proliferation-independent manner by replacing the most invasion-prone subpopulation with the non-invasive subpopulation ( s ) to suppress the invasive phenotype ( Figure 7f ) ."],["On the basis of a novel small-molecule scaffold , 6′-homosinefungin ( HSF ) , SKI-73 ( 6a ) was developed as a pro-drug-like chemical probe for CARM1 by cloaking the 9′-amine moiety of 5a with the TML moiety .","SKI-73N ( 6b ) was developed as a control compound for SKI-73 ( 6a ) .","The inhibitory activity of SKI-73 ( 6a ) against CARM1 was demonstrated by the ability of SKI-73 ( 6a ) but not SKI-73N ( 6b ) to abolish the cellular methylation marks of CARM1: the Arg1064 methylation of BAF155 and the Arg455/Arg460 methylation of PABP1 ( Lee and Bedford , 2002; Wang et al . , 2014a ) .","The ready intracellular cleavage of TML is expected for the conversion of SKI-73 and SKI-73N ( 6a and 6b ) into 5a and 5b , respectively , but it is remarkable that SKI-73 and SKI-73N ( 6a and 6b ) can also be efficiently processed into 2a and 2b inside cells .","In comparison , 6a showed decent metabolic stability with no production of 2a in the presence of microsomes .","Here , 2a and 5a are presented as potent and selective CARM1 inhibitors , whereas their control compounds 2b and 5b interact poorly with CARM1 .","Competitive assays with the SAM cofactor and the peptide substrate showed that 2a and 5a act on CARM1 in a SAM-competitive and substrate-noncompetitive manner .","The SAM-competitive mode is consistent with the ligand–complex structures of CARM1 , in which the SAM binding site is occupied by 2a and 5a .","Strikingly , as revealed by their ligand–CARM1 complex structures , 2a and 5a engage CARM1 through noncanonical modes , with their 6′-N-benzyl moieties in the binding pocket that is otherwise occupied by the α-amino carboxylate moiety of conventional SAM analogs such as SAH , SNF and 1 .","This observation is consistent with the 4 . 1‒6 . 5 °C increase in the in vitro and cellular Tm of CARM1 upon binding 2a and 5a , which contrasts with the smaller Tm changes with SAM as a ligand .","The distinct modes of interaction of CARM1 with 2a and 5a ( Figure 3c , g ) also rationalize the CARM1 selectivity of the two SAM analogs over other methyltransferases , including closely related PRMT homologs .","Through mathematic modeling using the inputs of the LC-MS/MS-quantified intracellular concentrations and CARM1-binding constants of relevant HSF derivatives and the SAM cofactor , we concluded that high intracellular concentrations of 5a and 2a , and thus efficient CARM1 occupancy , can be achieved rapidly and maintained for several days with a single low dose of SKI-73 ( 6a ) .","By contrast , the occupancy by 5a and 2a of SMYD2 , the next likely engaged target , is below 20% with the efficacious doses of 6a that affect cell invasion .","The polar α-amino acid zwitterion moiety of 2a and the polar α-amino moiety of 5a probably account for their accumulation and retention inside cells .","To the best of our knowledge , EZM2302 , TP-064 , and SKI-73 ( also 6a in this work , www . thesgc . org/chemical-probes/SKI-73 ) and their derivatives are the only selective and cell-active CARM1 inhibitors ( Drew et al . , 2017; Nakayama et al . , 2018 ) .","Although the potency , selectivity , on-target engagement and potential off-target effects associated with these compounds have been examined in vitro and in cellular contexts as chemical probes , EZM2302 , TP-064 , and SKI-73 ( 6a ) are differentiated by their molecular scaffolds and modes of interaction with CARM1 ( www . thesgc . org/chemical-probes/SKI-73 ) ( Drew et al . , 2017; Nakayama et al . , 2018 ) .","SKI-73 ( 6a ) is a cofactor analog inhibitor embedding a N6'-homosinefungin moiety to engage the SAM binding site of CARM1 in a cofactor-competitive , substrate-noncompetitive manner; EZM2302 and TP-064 occupy the substrate-binding pocket of CARM1 in a SAH-uncompetitive or SAM-noncompetitive manner ( Drew et al . , 2017; Nakayama et al . , 2018 ) .","In particular , the prodrug property of SKI-73 ( 6a ) allows its ready cellular uptake , followed by rapid conversion into its active forms inside cells .","The prolonged intracellular CARM1 inhibition further distinguishes SKI-73 ( 6a ) from EZM2302 and TP-064 .","With SKI-73 ( 6a ) as a CARM1 chemical probe and SKI-73N ( 6b ) as a control compound , we showed that pharmacological inhibition of CARM1 with SKI-73 ( 6a ) , but not SKI-73N ( 6b ) , suppressed 80% of the invasion capability of MDA-MB-231 cells .","By contrast , the pharmacological inhibition of CARM1 with SKI-73 ( 6a ) had no effect on the proliferation of MDA-MB-231 cells .","This result is consistent with the lack of anti-proliferation activities of the other two CARM1 chemical probes , EZM2302 and TP-064 , against breast cancer cell lines ( Drew et al . , 2017; Nakayama et al . , 2018 ) .","The anti-invasion efficiency of SKI-73 ( 6a ) is in good agreement with the intracellular occupancy and the resulting abolition of several methylation marks of CARM1 upon treatment with SKI-73 ( 6a ) .","Our prior work showed that the methyltransferase activity of CARM1 is required for breast cancer metastasis ( Wang et al . , 2014a ) .","Among the diverse cellular substrates of CARM1 ( Blanc and Richard , 2017 ) , BAF155—a key component of the SWI/SNF chromatin-remodeling complex–is essential for the invasion of MDA-MB-231 cells ( Wang et al . , 2014a ) .","Mechanistically , the CARM1-mediated Arg1064 methylation of BAF155 facilitates the recruitment of the SWI/SNF chromatin-remodeling complex to a specific subset of gene loci ( Wang et al . , 2014a ) .","Replacement of the native CARM1 with its catalytically dead mutant or with an Arg-to-Lys point mutation at the Arg1064 methylation site of BAF155 is sufficient to abolish the invasive capability of breast cancer cells ( Wang et al . , 2014a ) .","CARM1 inhibition with SKI-73 ( 6a ) , but not with its control compound SKI-73N ( 6b ) , recapitulates the anti-invasion phenotype associated with the genetic perturbation of CARM1 .","More importantly , there is no additive effect upon combining CARM1-KO with SKI-73 ( 6a ) treatment , underlying the fact that the two orthogonal approaches target the commonly shared pathway ( s ) that are essential for the invasion of breast cancer cells .","In comparison to SKI-73 ( 6a ) , the CARM1 inhibitors EZM2302 and TP-064 demonstrated anti-proliferation effects on hematopoietic cancer cells , in particular multiple myeloma ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018 ) .","Mechanistically , genetic perturbation of CARM1 in the context of leukemia impairs cell-cycle progression , promotes myeloid differentiation , and ultimately induces apoptosis , probably by targeting pathways of proliferation and cell-cycle progression , that is , E2F- , MYC- , and mTOR-regulated processes ( Greenblatt et al . , 2018 ) .","In comparison , CARM1 inhibition with EZM2302 led to a slightly different phenotype , which includes reduction of RNA stability , E2F target downregulation , and induction of a p53 response signature for senescence .","( Greenblatt et al . , 2018 ) .","Collectively , the effects of CARM1 chemical probes are highly context-dependent , with SKI-73 ( 6a ) having different uses in impairing the invasiveness of breast cancer cells , while TP-064 and EZM2302 have uses in preventing the proliferation of hematopoietic cancer cells .","Given the increased awareness of epigenetic plasticity ( Flavahan et al . , 2017 ) , we employed the scRNA-seq approach to examine MDA-MB-231 cells and their responses to chemical and genetic perturbation with CARM1 .","Because of the lack of a prior reference to define subpopulations of MDA-MB-231 cells , we developed a cell-cycle-aware algorithm to cluster the subpopulations with a resolution that was able to dissect subtle changes upon treatment with SKI-73 ( 6a ) versus its control compound SKI-73N ( 6b ) in each cell-cycle stage .","Guided by Silhouette analysis , the population entropy analysis and the Fisher Exact test , >10 , 000 MDA-MB-231 breast cancer cells were classified on the basis of their cell-cycle stages and then clustered into 34 subpopulations .","With further annotation of these subpopulations according to their different responses to treatment with SKI-73 ( 6a ) versus SKI-73N ( 6b ) , we readily dissected the subpopulations that were altered in a SKI-73 ( 6a ) -specific ( CARM1-dependent ) manner and then identified subsets with transcriptional signatures that are similar to that of the freshly isolated invasive cells .","Quantitative analysis of SKI-73 ( 6a ) -depleted subpopulations further revealed the most invasion-prone subpopulation , which accounts for only 5% of the total population but at least 80% of the invasive capability of the parental cells .","Collectively , we propose a model in which MDA-MB-231 cells consist of subpopulations , with their epigenetic plasticity ( Figure 7f ) determined by multiple factors including the CARM1-involved BAF155 methylation ( Wang et al . , 2014a ) .","SKI-73 ( 6a ) inhibits the methyltransferase activity of CARM1 , the Arg1064 methylation of BAF155 , and thus the target genes associated with the methylated BAF155 .","These effects alter the cellular epigenetic landscape by affecting certain subpopulations of MDA-MB-231 cells without any apparent effect on cell cycle and proliferation .","In the context of the invasion phenotype of MDA-MB-231 cells , the subset of invasion-prone cells is significantly suppressed upon the treatment with SKI-73 ( 6a ) .","Essential components that are used to dissect the invasion-prone population in this CARM1-dependent epigenetic plasticity model are the scRNA-seq analysis of sufficient MDA-MB-231 cells ( >10 , 000 cells here ) , the utility of the freshly isolated invasive cells as the reference , the timing and duration of treatment , and the use of SKI-73N ( 6b ) and DMSO as controls .","Interestingly , although the invasion-prone subpopulation is also abolished in the CARM1-KO strain , CARM1-KO reshapes the epigenetic plasticity in a much more profound manner , significantly reducing the subpopulation heterogeneity of MDA-MB-231 cells .","The distinct outcomes for the pharmacological and genetic perturbation could be due to their different modes of action: short-term treatment with SKI-73 ( 6a ) versus long-term clonal expansion of CARM1-KO cells .","The pharmacological inhibition captures the immediate response , whereas the genetic perturbation reports long-term and potential resistant outcomes .","This work thus presents a new paradigm to understand cancer metastasis in the context of epigenetic plasticity and provides guidance for similar analyses in broader contexts: other cell lines , patient-derived xenograft samples , and in vivo mouse models of breast cancer ."],["SAM- and substrate-dependent IC50 values were determined with the radiometric filter paper assay as described above except in the presence of the varied concentrations of the SAM cofactor and the peptide substrate .","The IC50 values of 2a and 5a were obtained in the presence of 0 . 09 , 0 . 18 , 0 . 37 , 0 . 75 , 1 . 50 , 2 . 25 , 3 . 00 , 5 . 62 , and 7 . 50 µM [3H-Me]-SAM or 0 . 3 , 1 . 5 , 7 . 5 , 15 , 30 , and 50 µM H3 peptide ( aa 1–40 ) .","The resultant IC50 values were plotted as a function of the concentrations of SAM and the peptide substrate using GraphPad Prism Software .","Given the SAM-competitive character of 2a and 5a , their SAM-dependent IC50 values were then fitted according to Equation","1 . Here [SAM] is the concentration of SAM , Km , SAM is the Michaelis-Menten constant , and Kd is the dissociation constant of the CARM1 inhibitor 2a or 5a .","Kd , Kd/Km , SAM and Km , SAM can therefore be obtained through the intercept of the y-axis , the slope , and their ratio , respectively , upon fitting Equation","1 . Given that 6a showed a much higher IC50 ( 1 . 1 ± 0 . 1 µM , Figure 5—figure supplement 3 ) , the Kd value of 6a ( Kd , 6a = 0 . 32 µM ) was directly obtained from the IC50 value , the Km , SAM derived above , and the assayed SAM concentration according to Equation 1 . ( 1 ) IC50=[SAM]×kd/km , SAM+kd Full-length CARM1 was used for SPR assay .","DNA fragment encoding the full-length CARM1 was cloned into pFB-N-flag-LIC donor plasmid .","The resulting plasmid was transformed into DH10Bac-competent E . coli cells ( Invitrogen ) and a recombinant Bacmid DNA was purified , followed by a recombinant baculovirus generation in Sf9 insect cells .","Sf9 cells grown in HyQ SFX insect serum-free medium ( ThermoScientific ) were infected with 10 mL of P3 viral stocks per 1 L of suspension cell culture and incubated at 27°C using a platform shaker set at 150 revolutions per minute .","The cells were collected when viability dropped to 70–80% ( ~72 hr after infection ) .","Harvested cells were re-suspended in PBS , 1X protease inhibitor cocktail ( 100X protease inhibitor stock in 70% ethanol containing 0 . 25 mg/mL aprotinin , 0 . 25 mg/ml leupeptin , 0 . 25 mg/mL pepstatin A and 0 . 25 mg/mL E-64 ) and 2X Roche complete EDTA-free protease inhibitor cocktail tablet .","The cells were lysed chemically by rotating 30 min with NP40 ( final concentration of 0 . 6% ) and 50 U/mL benzonase nuclease ( Sigma ) , 2 mM 2-mercaptoethanol and 10% glycerol followed by sonication at frequency of 7 ( 10′‘on/10′‘off ) for 2 min ( Sonicator 3000 , Misoni ) .","The crude extract was clarified by high-speed centrifugation ( 60 min at 36 , 000 × g at 4°C ) by Beckman Coulter centrifuge .","The recombinant protein was purified by incubating the cleared lysate with anti-FLAG M2 affinity agarose gel ( Sigma , Cat # A2220 ) and then rotating for 3 hr , followed by washing with 10 CV TBS ( 50 mM Tris-HCl , 150 mM NaCl , pH 7 . 4 ) containing 2 mM 2-mercaptoethanol , 1X protease inhibitor cocktail ( 100X protease inhibitor stock in 70% ethanol containing 0 . 25 mg/mL aprotinin , 0 . 25 mg/mL leupeptin , 0 . 25 mg/mL pepstatin A and 0 . 25 mg/mL E-64 ) and 1X Roche complete EDTA-free protease inhibitor cocktail tablet .","The recombinant protein was eluted by competitive elution with a solution containing 100 µg/mL FLAG peptide ( Sigma , Catalog # F4799 ) in 20 mM Tris pH:7 . 4 , 150 mM NaCl , 5% glycerol , 3 mM 2-mercaptoethanol .","Quality of CARM1 ( >95% ) was determined by SDS–PAGE .","The protein was then concentrated , flash frozen with liquid nitrogen , and stored at −80°C for future use .","SPR analysis was performed using a Biacore T200 ( GE Health Sciences Inc ) at 25°C .","Approximately 5500 response units of CARM1 ( amino acids 1–608 ) were amino-coupled onto a CM5 chip in one flow cell according to the protocol of the manufacturer .","Another flow cell was left empty for reference subtraction .","SPR analysis was conducted in the HBS-EP buffer ( 20 mM HEPES pH 7 . 4 , 150 mM NaCl , 3 mM EDTA , 0 . 05% Tween-20 ) containing 2% ( v/v ) DMSO .","The stock solutions of five concentrations of compound 2a ( 24 . 7 , 74 . 1 , 222 , 667 and 2000 nM ) and four concentrations of 5a ( 6 . 2 , 18 . 5 , 55 . 6 and 167 nM ) were prepared by serial dilution .","Binding kinetic experiments were performed with single cycle kinetics with the contact time of 60 s and off time of 300 s in a flow rate of 30 µL/min .","To facilitate complete dissociation of compound for the next cycle , a regeneration step ( 300 s , 40 µL/min of buffer ) , a period of stabilization ( 120 s ) and two blank cycles were included between each cycle .","Kinetic curve fittings were carried out with a 1:1 binding model or a heterogeneous ligand model using Biacore T200 Evaluation software ( GE Health Sciences Inc ) .","TSA was performed as described previously ( Blum et al . , 2014; Niesen et al . , 2007 ) to examine the melting temperature ( Tm ) of CARM1 in the presence or absence of ligands .","For each measurement ( triplicate of each data point ) , the assay solution containing 50 mM HEPES-HCl ( pH = 8 . 0 ) , Tween-20 0 . 005% ( v/v ) , 1 mM TCEP , 0 . 5 μM CARM1 , and 5 μM ligand ( SAM , 1 , 2a or 5a ) was mixed with 5 × SYPRO Orange Protein Gel Stain stock ( Sigma Aldrich ) in a 96-well PCR plate .","The mixture was equilibrated at 25°C in dark for 5 min and then loaded onto Bio-Rad CFX96 Real-Time PCR Detection System .","The fluorescence readouts were recorded upon increasing the heating temperature from 25°C to 100°C at a rate of 0 . 2 °C/s .","The raw data of the fluorescence readouts versus the temperatures were exported with CFX software and processed as the percentage of the fluorescent signal normalized between the lowest readout of 0% and the highest readout of 100% within the 25‒100°C region .","The melting curves were plotted as the normalized fluorescence ( % ) versus the heating temperature and fit with a sigmoid curve with GraphPad Prism .","The Tm corresponds to the temperature with the 50% relative fluorescent signal in the sigmoidal curve .","A DNA fragment encoding the methyltransferase domain of human CARM1 ( residues 140–480 ) was cloned into a baculovirus expression vector pFBOH-MHL ( http://www . thesgc . org/sites/default/files/toronto_vectors/pFBOH-MHL . pdf ) .","The protein was expressed in Sf9 cells as an N-terminal 6 × His tag fusion protein and purified by metal chelating affinity chromatography ( TALON resin , Clontech , Mountain View , CA , USA ) followed by size-exclusion chromatography ( Superdex 200 , GE Healthcare ) .","Pooled fractions containing CARM1 were subjected to the treatment of tobacco etch virus to remove the 6 × His tag .","The protein was purified to homogeneity by ion-exchange chromatography .","Purified CARM1 ( 6 . 5 mg/mL ) was crystallized with the sitting drop vapor diffusion method at 20°C .","For the CARM1–1 complex , CARM1 was mixed with 1 at a 1:5 molar ratio ( protein:ligand ) and crystallized with the sitting drop vapor diffusion method at 20°C by mixing 1 µL of the protein solution with 1 µL of the reservoir solution containing 20% PEG3350 and 0 . 2 M diammonium tartrate .","X-ray diffraction data for the CARM1–1 complex were collected at 100 K at beam line 23ID-B of Advanced Photon Source ( APS ) , Argonne National Laboratory .","Data sets were processed using the HKL-2000 suite ( Otwinowski and Minor , 1997 ) .","The structure of the CARM1–1 complex was solved by molecular replacement using MOLREP ( Otwinowski and Minor , 1997 ) with the PDB entry 2V74 as the search template .","REFMAC ( Emsley and Cowtan , 2004; Murshudov et al . , 1997 ) was used for the structure refinement .","Graphics program COOT was used for model building and visualization .","Crystal diffraction data and refinement statistics for the structure are displayed in Table","2 . To further confirm the electronic densities of the ligand , the total omission electron density map was calculated using SFCHECK from CCP4suite and contoured at 1 . 0 sigma ( Figure 4—figure supplement 1 ) ( Emsley and Cowtan , 2004; Murshudov et al . , 1997 ) .","For the CARM1-5a complex , CARM1 was crystallized with the sitting drop vapor diffusion method at 20°C by mixing 1 µL of protein solution with 1 µL of the reservoir solution containing 25% PEGG3350 , 0 . 1M ammonium sulfate and 0 . 1 M Hepes ( pH = 7 . 5 ) .","The compound 5a ( 0 . 2 μL of 10 μM in DMSO ) was added to the drops with apo crystals and incubated overnight .","X-ray diffraction data for the CARM1-5a complex were collected at 100 K at beamline 24ID-E of Advanced Photon Source ( APS ) , Argonne National Laboratory .","Data were processed using the HKL-3000 suite ( Minor et al . , 2006 ) .","The structure was isomorphous to PDB entry 4IKP , which was used as a starting model .","REFMAC ( Murshudov et al . , 1997 ) was used for structure refinement .","Geometric restraints for compound refinement were prepared with GRADE v . 1 . 102 developed at Global Phasing Ltd . ( Cambridge , UK ) .","The COOT graphics program ( Emsley et al . , 2010 ) was used for model building and visualization , and MOLPROBITY ( Williams et al . , 2018 ) was used for structure validation .","To further confirm the electronic densities of the ligand , the total omission electron density map was calculated and contoured at 2 . 5 sigma ( Figure 3b ) ( Emsley and Cowtan , 2004; Murshudov et al . , 1997 ) .","The ligands were docked into the binding site of CARM1 using the induced-fit docking ( IFD ) protocol ( Sherman et al . , 2006 ) implemented in the Schrodinger suite ( release 2016–4 ) .","The poses for SNF and 2a were selected according to the IFD scores .","Specifically , the results identified two distinct poses for 2a with similar scores but only one pose for SNF .","The poses were then further relaxed by all-atom , explicit solvent molecular dynamics ( MD ) simulations .","Herein , CARM1 models in complex with the two ligands were placed into explicit water boxes .","A simple point charge ( SPC ) water model ( Berendsen et al . , 1981 ) was used to solvate the system , charges were neutralized , and 0 . 15 M NaCl was added .","The total system size was ∼50 , 000 atoms .","Desmond MD systems ( D . E . Shaw Research , New York , NY ) with OPLS3 force field ( Harder et al . , 2016 ) were used .","The system was initially minimized and equilibrated with restraints on the ligand heavy atoms and protein backbone atoms , followed by production runs with all atoms unrestrained .","The isothermal-isobaric ensemble was used with constant temperature ( 310 K ) maintained with Langevin dynamics and 1 atm constant pressure achieved using the hybrid Nose-Hoover Langevin piston method ( Feller et al . , 1995 ) on a flexible periodic cell .","For each CARM1–ligand complex , a 600 ns trajectory was collected .","MCF-7 and MDA-MB-231 ( parental and CARM1-KO ) cell lines ( Wang et al . , 2014a ) were used after their quality was confirmed with STR profile and standard mycoplasma contamination testing .","Using the standard working curves generated above , the concentration ( CA ) of each analyte ( 6a , 5a , 2a , and SAM ) in the MeOH extraction of cell lysates was obtained through the ratio ( PA/PIS ) of the mass peak areas of each analyte ( PA ) versus each internal standard ( PIS ) , and the concentration of the internal standard ( CIS ) according to Equation","2 . For 6a , 5a , and 2a under each assay condition , three similar CA values ( CA-6b , CA-5b , and CA-2b ) were obtained with 6b , 5b and 2b as the internal standards , respectively .","An average concentration of the analyte ( C¯A ) was obtained on the basis of the three concentrations weighted by the mass peak areas of the three internal standards ( PIS , 6b , PIS , 5b and PIS , 2b ) according to Equation 3: ( 3 ) C¯A=CA−6b×PIS−6bPIS−6b+PIS−5b+PIS−2b+CA−5b×PIS−5bPIS−6b+PIS−5b+PIS−2b+CA−2b×PIS−2bPIS−6b+PIS−5b+PIS−2b On the basis of the C¯A values of 6a , 5a , and 2a ( C¯A , 6a , C¯A , 5a and C¯A , 2a ) in the MeOH extraction of cell lysates , the intracellular concentrations of the analyte 6a , 5a , and 2a ( Cintra , 6a Cintra , 5a and Cintra , 2a ) were calculated according to Equation 4 .","Here C¯A , analyte is the weighted average of the three concentrations in the MeOH extraction , N is the cell number , and V is the mean volume of cells ( µL ) .","The mean volume of MDA-MB-231 cells is 1 . 3 × 10−6 µL/cell as reported previously ( Coulter et al . , 2012 ) .","The cell number ( N ) was determined with a hemocytometer .","The concentration of SAM ( CA , SAM ) in the MeOH extraction was obtained according to Equation 2 , in which ‘X’ is the ratio ( PA , SAM/PIS , CD3-SAM ) of the mass peak areas of SAM ( PA , SAM ) versus the internal standard CD3-SAM ( PIS , CD3-SAM ) and ‘Y’ is the ratio ( CA , SAM/CIS , CD3-SAM ) of the concentrations of SAM versus CD3-SAM in the MeOH extraction .","Given the identical LC-MS properties of SAM and CD3-SAM , CA , SAM was obtained solely on the basis of the working curve with CD3-SAM as the internal standard .","The intracellular concentration of SAM ( Cintra , SAM ) was then calculated using Equation 4 , in which N is the cell number and V is the mean volume of MDA-MB-231 cells ( 1 . 3 × 10−6 µL/cell ) .","( 4 ) Cinter , analyte=C¯A , analyte×40μLN×V Similar experimental procedures were carried out to quantify the intracellular concentrations of 6b , 5b , 2b and SAM except that 6b , 5b , and 2b were the analytes and their structurally related 6a , 5a and 2a were used as the internal standards with their CIS values of 0 . 125 µM , 0 . 125 µM , and 0 . 0125 µM , respectively .","For each analyte ( 6b , 5b , and 2b ) , three working curves ( Figure 5c , e , Figure 5—figure supplement 2 ) were generated to cover the concentration range of 3 . 9 nM‒18 . 0 µM of the analyte ( CA , 6b , CA , 5b and CA , 2b ) with 6a , 5a and 2a ( CIS-6a , CIS-5a , and CIS-2a ) as the LC-MS/MS internal standards , respectively .","In a manner similar to that described above for the ligand occupancy of CARM1 , the SMYD2 occupancy by 2a was modeled with Kd , SAM = 60 nM , Kd , 2a = 150 nM and Equation 6 .","Here , the Kd , SAM value was reported previously upon developing the activity assay of SMYD2 ( Sweis et al . , 2015 ) .","The Kd , 2a value was obtained according to IC50 = [SAM]×Kd/Kd , SAM+Kd , in which [SAM]=Kd , SAM = 60 nM for the IC50 assay and IC50 = 0 . 3 µM was obtained ( Supplementary file 1-Table A ) .","The terms of SMYD2 occupancy by exogenous ligands were ignored given their low concentrations and low affinity relative to that of 2a ( Supplementary file 1-Table A ) .","Potential in vivo clearance of 6a was evaluated in the presence of liver microsomes as previously reported ( Hansen et al . , 2012 ) .","Briefly , 6a was dissolved into 50 mM potassium phosphate buffer ( pH 7 . 4 ) to a final concentration of 100 μM .","To a 0 . 5 mL Eppendorf vial containing ice-cold potassium phosphate buffer ( pH 7 . 4 ) , we added 5 μL of the 100 μM stock solution of 6a , 3 μL of a freshly prepared NADPH solution ( 5 mM , Sigma-Aldrich , N7785 ) , and 1 . 33 μL of freshly thawed rat liver microsomes ( Sigma-Aldrich , M9066 ) to yield assay samples with a final volume of 50 μL .","The resulting mixture was immediately vortexed and incubated in a 37°C shaker .","At the time interval of 0 , 20 , 40 , and 60 min in triplicates , respective assay samples were quenched by adding 50 μL of ice-cold methanol , vigorously vortexed , and centrifuged with 5 , 000 g at 4°C for 20 min .","From the methanol-quenched sample , 50 μL of supernatant was collected and mixed with 50 μL of 0 . 1% ( v/v ) TFA/MeOH solution containing 0 . 125 µM 6b , 0 . 125 µM 5b , and 16 µM 2b as the internal standards .","The resulting mixture was transferred into a 96-well plate ( Agilent , 5042–1386 ) and stored at −20°C for LC-MS/MS analysis .","The microsomal stability of 6a was evaluated via LC-MS/MS quantification of the residual 6a and of two potential microsome-processed products , 5a and 2a , upon quantification of intracellular concentrations of 6a , 5a and 2a ( as described above ) .","The amount of 6a was plotted as the percentage of the residual concentration at each time point against its initial concentration at 0 min with GraphPad Prism Software .","Potential production of 5a and 2a from 6a was examined in parallel .","Interestingly , despite robust accumulation of 5a , no 2a was identified with the detection threshold of 0 . 02 µM .","For negative controls , 3 μL of 50 mM potassium phosphate buffer ( pH 7 . 4 ) replaced the freshly prepared NADPH solution; two assay samples at the time interval of 0 and 60 min were collected .","Under the current microsomal condition , the concentration of 6a remained the same ( 10 ± 0 . 7 μM ) and there was no production of 5a or 2a .","MDA-MB-231 cells and MCF-7 parental and CARM1-KO ( Wang et al . , 2014a ) cells were maintained in DMEM ( Gibco , Gaithersburg , MD ) medium containing 10% FBS ( Gibco ) .","These cells were then treated with compounds or with DMSO for 48 hr .","The resultant cells were washed twice in phosphate-buffered saline ( PBS ) and the samples were sonicated in ice-cold RIPA buffer ( Thermo , Waltham , MA ) .","The lysates were centrifuged ( 15 , 000 g ) at 4°C for 15 min .","The supernatants were kept with the total protein amount determined with Bradford protein assay ( BioRad , Hercules , CA ) .","After quantification , 50 μg of protein from each sample was loaded onto 6% SDS-PAGE and transferred onto a nitrocellulose membrane ( PALL , Port Washington , NY ) .","For the Arg1064 methylation mark of BAF155 , the blots were blocked in 5% non-fat milk for 1 hr and incubated with anti-me-BAF155 ( Cancer Cell , 2014 ) , anti-BAF155 ( 1:1000; Santa Cruz Biotechnology , Dallas , TX ) , anti-CARM1 ( 1:1000; Genemed Synthesis , San Antonio , TX ) , and anti-β-actin ( 1:20000; Sigma-Aldrich , St . Louis , MO ) overnight at 4°C .","After washing three times in Tris-buffered saline with Tween 20 ( TBST ) , the blots were incubated with HRP-conjugated secondary antibody ( 1:3000; Jackson ImmunoResearch , West Grove , PA ) .","After washing blots with TBST , the membranes were developed using SuperSignal West Pico ECL solution ( Thermo ) .","For the Arg455/Arg460 methylation mark of PABP1 , anti-me-PABP1 ( 1:1000 ) and anti-PABP1 ( 1:1000 ) antibodies ( Genemed Synthesis , San Antonio , TX ) were used instead .","Here the antibodies against CARM1 , PABP1 , and me-PABP1 were custom generated by Genemed Synthesis ( San Antonio , TX ) ( Zeng et al . , 2013 ) .","The density of the protein bands was quantified using ImageJ software ( NIH , Bethesda , MD ) .","The EC50 values were obtained by fitting the methylation percentage ( % ) of BAF155 or PABP1 against the concentrations of the inhibitor using a sigmoidal equation with GraphPad Prism Software .","CETSA was performed as described previously ( Jafari et al . , 2014 ) to examine the intracellular engagement of 6a or 6b with CARM1 .","Briefly , 2 . 0 × 106 MDA-MB-231 cells were incubated with 15 μM 6a or 6b and DMSO for 48 hr .","The harvested cells were re-suspended in PBS buffer and divided into eight aliquots ( 30 µL/aliquot ) , and then heat-shocked at various temperatures ( 49 . 1 , 54 . 6 , 57 . 0 , 59 . 5 , 61 . 8 , 63 . 9 , 65 . 6 , and 67 . 0°C ) for 3 min with a Bio-Rad CFX96 Real-Time PCR Detection Instrument ( using a temperature gradient of 49‒67°C ) .","The heat-shocked cells were then lysed with the freeze-thaw method with a liquid nitrogen bath followed by a 25°C water bath for five cycles .","The cell lysate was centrifuged at 4°C at 18 , 000 g for 20 min .","The resultant supernatant containing the soluble protein fraction was collected and loaded onto an SDS-PAGE gel ( 20 µL ) .","Western blotting of CARM1 was performed with anti-CARM1 antibody ( Cell Signaling Technology , C31G9 ) .","For each sample , the band intensity of CARM1 was quantified with ImageJ .","The band intensity of CARM1 was normalized to the band intensity at 49 . 1°C ( the lowest heat-shock temperature ) .","Melting curves were obtained by plotting the normalized band intensity against the heat-shock temperatures and fit with a Boltzmann sigmoidal equation in GraphPad Prism .","The melting temperatures ( Tm ) of 6a , 6b , or DMSO were obtained as the heat-shock temperatures that correspond to the 50% normalized band intensity in the fitted sigmoidal curves .","MDA-MB-231 parental and CARM1-KO cells ( Wang et al . , 2014a ) were maintained in DMEM ( Gibco , Gaithersburg , MD ) medium containing 10% FBS ( Gibco ) .","Cell invasion assays were performed using 8 . 0 µm pore size Transwell inserts ( Greiner Bio-One , Kremsmünster , Austria ) .","MDA-MB-231 parental and CARM1-KO ( Cancer Cell , 2014 ) cells were harvested with trypsin/EDTA and washed twice with serum-free DMEM ( Gibco ) .","2 × 105 cells in 0 . 2 mL serum-free DMEM ( Gibco ) were seeded onto the upper chamber , which was pre-coated with a thin layer of 40 uL of 2 mg/mL Matrigel ( Corning , NY , USA ) for 2 hr incubation at 37°C .","To the lower chamber , we added 0 . 6 mL DMEM containing 10% FBS ( GIBCO ) and compounds or DMSO .","After 16 hr in the 37°C incubator , the cells on the inner side of the upper chamber together with the Matrigel layer were removed using cotton tips .","The residual invasive cells in the outer side of the upper chamber were fixed in 3 . 7% formaldehyde ( a weight percentage ) at an ambient temperature ( 22°C ) for 2 min and 100% methanol for 20 min , and then stained for 15 min with a solution containing 1% crystal violet and 2% ethanol in 100 mM borate buffer ( pH 9 . 0 ) .","The number of invasive cells was counted under a microscope by taking five independent fields .","Relative cell invasion was determined by the number of the invasive cells normalized to the total number of cells adhering to 0 . 8 µm transwell filters .","The EC50 was obtained with GraphPad Prism Software upon fitting Equation 8 , in which ‘%Inhibition’ is the percentage of the inhibition of invasiveness , ‘Maximal Inhibition%’ is the percentage of the maximal inhibition of invasiveness , and [Inhibitor] is the concentration of the inhibitor .","( 8 ) Inhibition%=MaximalInhibition%×[Inhibitor][Inhibitor]+EC50 To examine proliferation of MDA-MB-231 cells , 5000 cells of parental and CARM1-KO cells were seeded in a 96-well plate and incubated in 37°C overnight .","These cells were treated with various doses ( 0 . 0001‒10 μM ) of SKI-73 or SKI-73N ( 6a or 6b ) in DMSO and incubated for 72 hr .","An MTT assay was then performed to examine viability with DMSO-treated cells as the control .","The relative viability of compound-treated cells versus DMSO-treated parent cells were plotted against the concentrations of SKI-73 and SKI-73N ( 6a and 6b ) .","Cell preparation for scRNA-seq .","10 × Genomics droplet-based scRNA-seq was implemented to characterize five types of MDA-MB-231 cells: MDA-MB-231 cells treated with DMSO , SKI-73 ( 6a ) and SKI-73N ( 6b ) , MDA-MB-231 cells that freshly invaded through Matrigel , and CARM1-KO MDA-MB-231 cells .","For the first three samples , MDA-MB-231 cells were treated with DMSO , 10 μM SKI-73 ( 6a ) , and 10 μM SKI-73N ( 6b ) for 48 hr .","The resulting cells were trypsinized at 37°C for 3 min .","The harvested cells were washed twice with 1 × PBS ( phosphate-buffered saline ) containing 0 . 04% ( volume% ) bovine serum albumin ( BSA ) , gently dispersed to dissociate cells , and then filtered through a cell strainer ( 100 μm Nylon mesh , Fisherbrand ) to obtain single-cell suspensions for scRNA-seq .","For the CARM1-KO sample , the cells were trypsinized at 37°C for 3 min , washed twice with 1 × PBS containing 0 . 04% BSA , and filtered through a 100 μm Nylon-mesh cell strainer ( Fisherbrand ) to obtain single-cell suspensions for scRNA-seq .","To collect the cells that freshly invaded through Matrigel , the conditions described above for cell invasion assays were applied to allow approximately 5% of the 1 × 107 seeded MDA-MB-231 cells to invade through Matrigel .","Briefly , ten Transwell inserts with a diameter of 24 mm and a pore size of 8 . 0 μm were pre-coated with a thin layer of 54 μL of 2 mg/mL Matrigel ( Corning ) .","1 × 106 MDA-MB-231 cells in 1 . 0 mL serum-free DMEM ( Gibco ) were seeded into the upper chamber of each Matrigel-coated Transwell insert .","2 . 0 mL DMEM containing 10% FBS ( Gibco ) was added into the lower chambers .","After 16 hr incubation , the cells on the inner side of the upper chamber together with the Matrigel layer were removed using cotton tips .","The cells that freshly invaded–those attached on the outer side of the upper chamber–were subjected to 3 min trypsin digestion at 37°C for detachment .","The resulting cells were washed twice with 1 × PBS containing 0 . 04% BSA , gently dispersed to dissociate cells , and filtered through a 100 μm Nylon-mesh cell strainer ( Fisherbrand ) to obtain single-cell suspensions for scRNA-seq .","The scRNA-seq libraries were prepared following the user guide manual ( CG00052 Rev E ) provided by the 10 × Genomics and Chromium Single Cell 3' Reagent Kit ( v2 ) .","Briefly , samples containing approximately 8700 cells ( 93–97% viability ) were encapsulated in microfluidic droplets at a dilution of 66–70 cells/µL , which resulted in 4369–5457 recovered single-cells per sample with a multiplet rate ~3 . 9% .","The resultant emulsion droplets were then broken and barcoded cDNA was purified with DynaBeads , followed by 12-cycles of PCR-amplification: −98 °C for 180 s , 12× ( 98°C for 15 s , 67°C for 20 s , 72°C for 60 s ) , and 72°C for 60 s .","The 50 ng of PCR-amplified barcoded-cDNA was fragmented with the reagents provided in the kit and purified by SPRI beads with an averaged fragment size of 600 bp .","The DNA library was then ligated to the sequencing adapter followed by indexing PCR: −98 °C for 45 s; 12× ( 98°C for 20 s , 54°C for 30 s , 72°C for 20 s ) , and 72°C for 60 s .","The resulting DNA library was double-size purified ( 0 . 6–0 . 8× ) with SPRI beads and sequenced on Illumina NovaSeq platform ( R1 – 26 cycles , i7 – eight cycles , R2 – 96 cycles ) resulting in 70–79 million reads per sample with average reads per single-cell being 8075–10 , 342 and average reads per transcript 1 . 11–1 . 15 .","The fastq files containing the transcriptome and barcoding metadata were demultiplexed using the SEquence Quality Control ( SEQC ) pipeline ( http:github . com/ambrosejcarr/seqc . git ) resulting in around 8000 UMIs per one cell .","The table of UMI counts was used as the input and Seurat package v . 2 . 3 . 4 ( Butler et al . , 2018 ) was applied for scRNA-seq analysis .","Here , the raw UMI counts were normalized per cell by dividing the total number of UMIs in each individual cell , multiplying by a scale factor of 10 , 000 and transforming into natural logarithm values .","Cells with 1000~5000 genes and <20% mitochondrial RNA transcripts were kept for further analysis .","Dimensionality reduction was carried out by selecting a set of highly variable genes on the basis of the average expression and dispersion per gene .","The set of genes was used for principle component analysis ( PCA ) .","Top principle components were then chosen for cell clustering analysis and t-SNE projection .","Regression was performed to remove cell–cell variation in gene expression driven by the UMI number , mitochondrial gene content and ribosomal gene content using ‘ScaleData’ function in Seurat package ( Nestorowa et al . , 2016 ) .","Clusters of cells were identified by clustering algorithm based on a shared nearest neighbor ( SNN ) modularity optimization that was included in Seurat package v . 2 . 3 . 4 ( Butler et al . , 2018 ) .","Fisher’s Exact Test was implemented to evaluate the agreement of the clusters with the three cell origins ( DMSO- , SKI-73 ( 6a ) - or SKI-73N ( 6b ) -treated cells ) using R package ‘fisher . test’: http://mathworld . wolfram . com/FishersExactTest . html ( Mehta and Patel , 1983 ) .","Because of the significant computation cost of Fisher’s Exact Test , the cell population of each Fisher’s Exact Test was down-sampled to 150 cells and this process was repeated for 100 times to cover a majority of the cell population .","The p-value of Fisher’s Exact Test was then computed by Monte Carlo simulation .","Means and standard errors of p-values were calculated and reported as the outputs of Fisher’s Exact Test .","Three algorithmic scoring systems were used over a range of the resolution parameter that sets the corresponding ‘granularity’ of clustering , with higher values indicating a greater number of clusters .","Here , silhouette analysis was applied to determine the number of clusters in an unsupervised manner , without awareness of the three cell origins ( DMSO- , SKI-73 ( 6a ) - or SKI-73N ( 6b ) -treated cells ) .","The entropy scoring and Fisher’s Exact Test were implemented to evaluate the biological meaning of the clustering , using the minimal cluster number to resolve the cells between the three treatment conditions maximally .","Given the awareness of the three cell origins , the minimal number of clusters with the maximal resolution of cell origin guided by the entropy scoring and Fisher’s Exact Test is expected within the 1~3-fold range of the optimized number of clusters guided by Silhouette analysis .","Fisher’s Exact Test was used as the primary scoring method to determine the efficiency of clustering , given its higher resolution .","‘dj , i’ ( ‘dDMSO , i’ , ‘dSKI-73N , i’ and ‘dSKI-73 , i’ in Equation 9 are defined as the fraction of the cells with the ‘j’ origin ( j = 1 , 2 or three for the treatment with DMSO , SKI-73N/6b and SKI-73/6a , respectively ) within the ‘i’ subpopulation ( i = 0‒n for the clustered subpopulation ) .","‘di , total’ is defined as the fraction of the cells of the ‘i’ subpopulation within the total cell population in each cell-cycle stage .","‘d ( j , i ) , total =djj , i ×di , total’ represents the fraction of the cells with the ‘j’ origin ( DMSO , SKI-73/6a , and SKI-73N/6b ) and the ‘i’ subpopulation within the total cell population in each cell-cycle stage .","For a specific subpopulation ‘i’ , there are three ‘d ( j , i ) , total’ values ( d ( DMSO , i ) , total , d ( SKI-73N , i ) , total and d ( SKI-73 , i ) , total ) for the treatments with DMSO , SKI-73N ( 6b ) and SKI-73 ( 6a ) , respectively .","The alteration of subpopulations is defined as the ratios of d ( SKI-73N , i ) , total/d ( DMSO , i ) , total or d ( SKI-73 , i ) , total/d ( DMSO , i ) , total fall out of the range of 1 . 0 ± 0 . 2 .","Population analysis was conducted by classifying the SKI-73N/SKI-73 ( 6a/6b ) -treated subpopulations into the following five categories: commonly resistant ( 0 . 8 < ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ and ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’<1 . 2 ) , commonly emerging ( ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ and ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’≥1 . 2 ) , commonly depleted ( ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ and ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’≤0 . 8; ∣‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’−‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’∣<0 . 15 ) , differentially emerging ( either ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ or ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’>1 . 2 ) , and differentially depleted ( ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ or ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’<0 . 8; ∣‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’−‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’∣>0 . 15 ) .","Although additional combinations of differentially altered subpopulations could be possible , only those defined above were found under our treatment conditions .","In each cell cycle ( G0/G1 , S and G2/M ) of the cells treated with DMSO , SKI-73 ( 6a ) or SKI-73N ( 6b ) and ‘invasion cells’ , correlation analysis of subpopulations was conducted with ‘BuildClusterTree’ function in the Seurat package ( https://rdrr . io/cran/Seurat/man/BuildClusterTree . html ) ( Nestorowa et al . , 2016 ) .","The phylogenetic trees were constructed by averaging gene expressions across all cells in each subpopulation and then calculating distance on the basis of expressions averaged between different subpopulations .","Differentially expressed genes were identified by comparing two groups of cells using the Wilcox rank sumtest with the ‘FindMarkers’ function in the Seurat package ( Nestorowa et al . , 2016 ) .","In particular , the ‘invasion cells’ and their most correlated clusters ( which were revealed in the correlation analysis ) were selected as the ‘high’ group; the remaining remotely related clusters were selected as the ‘low’ group .","The differential expression analysis was then performed by comparing cells in the ‘high’ group with the cells in the ‘low’ group .","For the G0/G1-phase cells , the ‘invasion cells’ and Subpopulation 6 , 7 , 8 , 9 , 14 were selected as the ‘high’ group; Subpopulation 0 , 1 , 2 , 3 , 4 , 5 , 10 , 11 , 12 , 13 , 15 , 16 , 17 , 18 , 19 , 20 were selected as the ‘low’ group .","For the G2/M-phase cells , the ‘invasion cells’ and Subpopulations 1 , 2 were selected as the ‘high’ group; and Subpopulations 0 , 3 , 4 , 5 were selected as the ‘low’ group .","For the S-phase cells , the ‘invasion cells’ and Subpopulations 0 , 3 were selected as the ‘high’ group; and Subpopulations 1 , 2 , 4 , 5 , 6 were selected as the ‘low’ group .","Differentially expressed genes were ranked according to the average log2 fold change ‘avg_logFC’ and the adjusted p-values ‘p_val_adj’ with the Seurat package .","Top upregulated and downregulated genes were chosen by setting a cutoff on their ‘avg_logFC’ values ( >0 . 25 or <−0 . 25 ) .","Then , curated genes with potential functional relevance to cancer malignancy ( 30 upregulated and 10 downregulated genes ) were selected as representative genes for generating heat map plots using the ‘DoHeatmap’ function in the Seurat package ."]],"string":"[\n [\n \"Numerous biological events are orchestrated epigenetically upon defining cellular fates ( Atlasi and Stunnenberg , 2017; Berdasco and Esteller , 2019 ) .\",\n \"Among the key epigenetic regulators are protein methyltransferases ( PMTs ) , which can render downstream signals by modifying specific Arg or Lys residues of their substrates with S-adenosyl-L-methionine ( SAM ) as a methyl donor cofactor ( Luo , 2018 ) .\",\n \"Significant efforts have been made to identify the PMT-dependent epigenetic cues that are dysregulated or addicted under specific disease settings such as cancer ( Berdasco and Esteller , 2019 ) .\",\n \"Many PMTs are implicated as vulnerable targets against cancer malignancy ( Kaniskan et al . , 2018; Luo , 2018 ) .\",\n \"The pro-cancerous mechanism of these PMTs can be attributed to their methyltransferase activities , which act individually or in combination to upregulate oncogenes , downregulate tumor suppressors , and maintain cancer-cell-addicted homeostasis ( Berdasco and Esteller , 2019; Blanc and Richard , 2017 ) .\",\n \"Pharmacological inhibition of these epigenetic events thus presents promising anti-cancer strategies ( Berdasco and Esteller , 2019 ) , as exemplified by the development of the clinical inhibitors of DOT1L ( Bernt et al . , 2011; Daigle et al . , 2011 ) , EZH2 ( Kim et al . , 2013; Konze et al . , 2013; McCabe et al . , 2012; Qi et al . , 2012; Qi et al . , 2017 ) , and PRMT5 ( Bonday et al . , 2018; Chan-Penebre et al . , 2015 ) .\",\n \"Protein arginine methyltransferases ( PRMTs ) act on their substrates to yield three different forms of methylated arginine: asymmetric dimethylarginine ( ADMA ) , symmetric dimethylarginine ( SDMA ) , and monomethylarginine ( MMA ) , which are the terminal products of Type I , II and III PRMTs , respectively ( Blanc and Richard , 2017; Yang and Bedford , 2013 ) .\",\n \"Among the important Type I PRMTs is CARM1 ( PRMT4 ) , which regulates multiple aspects of transcription by methylating diverse targets including RNAPII , SRC3 , C/EBPβ , PAX3/7 , SOX2/9 , RUNX1 , Notch1 , p300 , CBP , p/CIP , Med12 , and BAF155 ( Blanc and Richard , 2017; Hein et al . , 2015; Vu et al . , 2013; Wang et al . , 2015; Wang et al . , 2014a; Yang and Bedford , 2013 ) .\",\n \"The physiological function of CARM1 has been linked to the differentiation and maturation of embryonic stem cells to form immune cells , adipocytes , chondrocytes , myocytes , and lung tissues ( Blanc and Richard , 2017; Yang and Bedford , 2013 ) .\",\n \"The requirement of CARM1 is implicated in multiple cancers , with its methyltransferase activity particularly addicted by hematopoietic malignancies and metastatic breast cancer ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018; Wang et al . , 2014a ) .\",\n \"Our prior efforts using in vivo mouse and in vitro cell models uncovered the role of CARM1 in promoting breast cancer metastasis ( Wang et al . , 2014a ) .\",\n \"Mechanistically , CARM1 methylates Arg1064 of BAF155 and thus facilitates the recruitment of the BAF155-containing SWI/SNF complex to a specific subset of gene loci that are essential for breast cancer metastasis .\",\n \"CARM1 thus emerges as a novel anti-cancer target ( Wang et al . , 2014a ) .\",\n \"Although this cancer relevance inspired the development of CARM1 inhibitors ( Kaniskan et al . , 2018; Scheer et al . , 2019 ) , many small-molecule CARM1 inhibitors lack target selectivity or cellular activity ( Kaniskan et al . , 2018 ) , two essential criteria of chemical probes ( Frye , 2010 ) .\",\n \"To the best of our knowledge , EZM2302 ( Drew et al . , 2017; Greenblatt et al . , 2018 ) , TP-064 ( Nakayama et al . , 2018 ) and SKI-73 ( 6a in this work , www . thesgc . org/chemical-probes/SKI-73 ) , which were developed by Epizyme , Takeda/SGC ( Structural Genomic Consortium ) , and our team , respectively , are the only selective and cell-active CARM1 chemical probes .\",\n \"EZM2302 and TP-064 were developed from conventional small-molecule scaffolds occupying the substrate-binding pocket of CARM1 ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018 ) .\",\n \"The potential utility of EZM2302 and TP-064 is implicated by their selective anti-proliferative effects on hematopoietic cancer cells , in particular multiple myeloma cells ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018 ) .\",\n \"However , definitive molecular mechanisms of the CARM1 addiction in these contexts remain elusive ( Greenblatt et al . , 2018 ) .\",\n \"Here , we report the characterization and novel utility of SKI-73 , a chemical probe of CARM1 with pro-drug properties .\",\n \"SKI-73 ( 6a in this work ) can readily penetrate cell membranes and then be processed into two active CARM1 inhibitors that contain 6′−homosinefungin ( HSF ) as their core scaffold ( Scheer et al . , 2019; Wu et al . , 2016 ) .\",\n \"Notably , the two inhibitors can accumulate inside cells to remarkably high concentrations and for a prolonged period .\",\n \"The potency , selectivity , modes of action , on-target engagement , and off-target effects of these compounds were characterized with multiple orthogonal assays in vitro and under cellular settings .\",\n \"The pharmacological inhibition of CARM1 by SKI-73 ( 6a ) recapitulates the anti-invasion effect of the genetic perturbation of CARM1 .\",\n \"In the context of cellular heterogeneity , we developed a cell-cycle-aware algorithm for single-cell RNA-seq ( scRNA-seq ) analysis and dissected the invasion-prone subset of breast cancer cells that is sensitive to SKI-73 ( 6a ) treatment .\",\n \"Our scRNA-seq analysis provides the unprecedented insight that pharmacological inhibition of CARM1 alters epigenetic plasticity and suppresses invasion by suppressing the most invasive subpopulation of breast cancer cells .\"\n ],\n [\n \"Upon developing cofactor-competitive PMT inhibitors ( Wu et al . , 2016; Zheng et al . , 2012 ) , we tailored the SAM analog sinefungin ( Figure 1a ) around its 6′-amino moiety to potentially engage CARM1’s substrate-binding pocket .\",\n \"6′-homosinefungin ( HSF , i . e . , 1 ) , a sinefungin analog with the insertion of 6′−methylene moiety , was discovered for its general high affinity to Type I PRMTs ( Figure 1a , b , Figure 1—figure supplement 1 , Supplementary file 1-Table A ) .\",\n \"As a SAM mimic , 1 binds to the Type I PRMTs ( namely PRMT1 , CARM1 , PRMT6 and PRMT8 ) with IC50 of 13‒300 nM ( Figure 1a , c , Supplementary file 1-Table A ) .\",\n \"Its relative affinity to Type I PRMTs aligns with that of the SAM mimics SAH and SNF ( around 20-fold lower IC50 of 1 versus SAH and SNF , Figure 1a , c , Supplementary file 1-Table A ) .\",\n \"This observation argues that 1 retains the structural features of SAH and SNF to engage PRMTs and meanwhile leverages its 6′-methyleneamine group for additional interaction .\",\n \"To further explore the 6′-region of HSF , we synthesized HSF derivatives from the same precursor 3 ( Figure 1b , Figure 1—figure supplement 2 ) , by further expanding the 6′-methylene amine moiety with different substituents .\",\n \"The HSF derivative 2a ( Figure 1b ) was identified for its preferential binding to CARM1 with IC50 = 30 ± 3 nM and >10 fold selectivity over other seven human PRMTs and 26 methyltransferases of other classes ( Figure 1c , Supplementary file 1-Table A ) .\",\n \"The structural difference between 2a and 1 ( Figure 1b ) suggests that the N-benzyl substituent enables 2a to engage CARM1 through a distinct mechanism ( see results below ) .\",\n \"With 2a as a lead , we then explored its α-amino carboxylate moiety with different amides from the common precursor three and then the intermediate 4 ( Figure 1—figure supplement 2 ) , which led to the discovery of 5a .\",\n \"This engagement of CARM1 with 2a is expected to be largely maintained by 5a .\",\n \"Here , 5a shows an IC50 of 43 ± 7 nM against CARM1 and a >10-fold selectivity over the panel of 33 diverse methyltransferases ( Figure 1c , Supplementary file 1-Table A ) .\",\n \"In comparison , the negative control compounds 2b ( Bn-SNF ) ( Zheng et al . , 2012 ) and 5b ( Figure 1b , Figure 1—figure supplement 3 ) , which differ from 2a and 5a only by the 6′-methylene group , poorly inhibit CARM1 ( IC50 = 22 ± 1 µM and 1 . 91 ± 0 . 03 µM ) ( Figure 1c , Supplementary file 1-Table A ) .\",\n \"The dramatic increase of the potency of 2a and 5a in contrast to 2b and 5b supports an essential role of the 6′-methylene moiety upon binding CARM1 .\",\n \"Distinguished from the SAM mimics SAH , SNF and 1 as nonspecific PMT inhibitors , 2a and 5a were developed as potent and selective SAM analogs .\",\n \"With 2a and 5a characterized as CARM1 inhibitors , we leveraged orthogonal in vitro assays to explore their modes of interaction ( Figure 2a ) .\",\n \"To examine whether 2a and 5a are SAM- or substrate-competitive , CARM1 inhibition by 2a and 5a was assessed in the presence of various concentrations of SAM cofactor and H3 peptide substrate ( Figure 2b , c ) .\",\n \"IC50 values of 2a and 5a showed a linear positive correlation with SAM concentrations , as expected for SAM-competitive inhibitors ( Daigle et al . , 2011; Luo , 2018; Zheng et al . , 2012 ) .\",\n \"The Kd values of 2a and 5a ( Kd , 2a = 17 ± 8 nM; Kd , 5a = 9 ± 5 nM ) were extrapolated from the y-axis intercepts upon fitting the equation IC50 = [SAM]×Kd/Km , SAM+Kd ( Figure 2b ) ( Segel , 1993 ) .\",\n \"Km , SAM of 0 . 21 ± 0 . 09 µM and 0 . 28 ± 0 . 14 µM ( an averaged Km , SAM = 0 . 25 µM ) for competition with 2a and 5a , respectively , can also be derived through the ratio of the y-axis intercepts to the slopes ( Figure 2b and Materials and methods ) ( Segel , 1993 ) .\",\n \"By contrast , the presence of the H3 peptide substrate had negligible effect on the binding of 2a and 5a , indicating their substrate-noncompetitive character ( Figure 2c ) .\",\n \"The SAM analogs 2a and 5a were thus characterized as SAM-competitive , substrate-noncompetitive inhibitors of CARM1 .\",\n \"For the direct binding of 2a and 5a to CARM1 , the CARM1-binding kinetics of 2a and 5a were examined using surface plasmon resonance ( SPR ) ( Figure 2d ) .\",\n \"The SPR signal progression of 2a and 5a fits with a biphasic rather than a mono-phasic binding mode , with the lower Ki1 , 2a = 0 . 06 ± 0 . 02 µM , Ki1 , 5a = 0 . 10 ± 0 . 01 µM , and the higher Ki2 , 5b = 0 . 54 ± 0 . 07 µM , Ki2 , 2a = 0 . 4 ± 0 . 1 µM , probably because of the multi-phase binding kinetics of 2a and 5a ( Figure 2d ) .\",\n \"To cross validate the binding of 2a and 5a to CARM1 , we conducted an in vitro thermal shift assay , for which ligand binding is expected to increase CARM1’s thermal stability ( Blum et al . , 2014 ) .\",\n \"The binding of 2a and 5a ( at 5 μM concentration ) increased the melting temperature ( Tm ) of CARM1 by 4 . 4°C and 6 . 5°C , respectively ( Figure 2e , Tm , 2a = 44 . 2 ± 0 . 4 °C and Tm , 5a = 46 . 3 ± 0 . 3 °C versus Tm , DMSO = 39 . 8 ± 0 . 3 °C as control ) .\",\n \"By contrast , the binding of SAM and 1 show much reduced effects on Tm of CARM1 ( Figure 2e , Tm , SAM = 40 . 1 ± 0 . 3 °C and Tm , 1 = 42 . 8 ± 0 . 4 °C versus Tm , DMSO = 39 . 8 ± 0 . 3 °C ) .\",\n \"Therefore , although the affinities of 1 , 2a and 5a to CARM1 are comparable ( IC50 = 13‒43 nM , Figure 1c ) , their well-separated effects on Tm suggest that these inhibitors engage CARM1 differentially ( see results below ) .\",\n \"The two orthogonal biochemical assays thus verified the tight binding of 2a and 5a with CARM1 .\",\n \"To further seek a structural rationale for 5a and 2a for CARM1 inhibition , we solved the X-ray structure of CARM1 in complex with 5a with resolution of 2 . 00 Å and modeled the CARM1 binding of 2a ( Figure 3 , Materials and methods ) .\",\n \"The overall topology of the CARM1–5a complex is indistinguishable with the V-shaped subunit of the CARM1 dimer in complex with SNF and 1 ( details in the next section ) , which is typical of the Rossmann fold of Class I methyltransferases ( Figure 3a , b , Table\",\n \"1 ) ( Luo , 2018 ) .\",\n \"However , 5a adopts a noncanonical pose with its 6′-N-benzyl moiety in a binding pocket that used to be occupied by the α-amino carboxylate moiety of canonical ligands such as SAH , SNF and 1 ( Figures 3c and 4 ) , while the α-amino methoxyphenethyl amide moiety of 5a protrudes into the substrate-binding pocket ( Boriack-Sjodin et al . , 2016; Sack et al . , 2011 ) .\",\n \"This noncanonical mode is consistent with the SAM-competitive character of 5a ( Figure 2b ) .\",\n \"In contrast to the noncanonical mode , Arg168 in the CARM1–5a complex adopts an alternative orientation ( two possible configurations ) , accompanied by an altered conformation of Glu257 , to accommodate the 6′-N-benzyl moiety of 5a ( Figure 3d ) .\",\n \"The α-amino amide moiety of 5a also engages CARM1 through the combined outcomes of a hydrogen-bond network and hydrophobic interactions with nearby resides ( Figure 3e ) .\",\n \"Interestingly , the overlaid structures of CARM1 in complex with 5a and a substrate peptide implicate a steric clash and thus a potential for binding competition between 5a and a CARM1 substrate ( Figure 3f ) .\",\n \"However , the apparent substrate-noncompetitive character of 5a ( Figure 2c ) suggests that this steric clash might be avoided if there is no significant energy penalty when the substrate Arg adopts alternative conformation ( s ) .\",\n \"The binding mode of the CARM1–2a complex was modeled via molecular docking followed by molecular dynamics ( MD ) simulation ( Materials and methods ) .\",\n \"Here , we uncovered two distinct poses of 2a ( Binding Pose 1/2 or BP1/2 , Figure 3g , Figure 3—figure supplement 1 ) .\",\n \"BP1 was characterized by the direct interaction between the α-amino carboxylate moiety of 2a and the guanidinium of Arg168 , whereas BP2 features a titled orientation of Arg168 to accommodate the 6′-N-benzyl moiety of 2a ( Figure 3g , Figure 3—figure supplement 1 ) .\",\n \"The BP1 and BP2 of 2a closely resemble those of 1 and 5a , respectively , in terms of the orientations of Arg168 and the α-amino carboxylate moiety of the ligands .\",\n \"When the same modeling protocol was applied to the CARM1–SNF complex , only the canonical pose was identified ( Materials and methods ) .\",\n \"Energy calculation indicated that both BP1 and BP2 are stable with comparable binding free energies .\",\n \"Interestingly , the side chain configurations of His414 in both BP1 and BP2 are different from those in the CARM1–5a complex and the CARM1–SNF complex ( Figure 3g ) .\",\n \"Collectively , 5a and 2a , though structurally related to the SAM analogs 1 and SNF , engage CARM1 via distinct modes of interaction .\",\n \"Upon comparing the CARM1 structure in complex with 5a and 2a , we observed the additional hydrogen-bond and hydrophobic interactions of 5a that involve its α-amino amide moiety ( Figure 3e ) .\",\n \"Interestingly , these interactions do not increase but rather decrease the affinity of 5a to CARM1 by two-fold ( Kd , 2a = 17 ± 8 nM versus Kd , 5a = 9 ± 5 nM , Figure 2b ) .\",\n \"By contrast , there is a significant 10-fold increase of affinity between 2b and 5b ( Figure 1c , Supplementary file 1-Table A ) .\",\n \"These observations suggest that , although 5b facilitates CARM1’s engagement better than 2b via the former’s α-amino amide moiety , such an effect is dispensed with in the presence of the 6′-methylene ( N-benzyl ) amine moiety of 5a and 2a .\",\n \"Given the tight CARM1 binding by 1 , we solved the X-ray structure of CARM1 in complex with 1 ( HSF ) with a resolution of 2 . 00 Å ( Figure 4 ) .\",\n \"The overall folding of the CARM1–1 complex ( PDB: 4IKP ) is similar to those of 1 in complex with SNF and SAH ( PDB: 2Y1X , 2Y1W ) with a V-shape subunit in a dimer of dimers ( Figure 4a , Table\",\n \"2 ) ( Sack et al . , 2011 ) .\",\n \"However , the CARM1–1 complex is distinct for multiple configurations of its ligand and interactions via its 6′-methyleneamine moiety ( Figure 4b–d , Figure 4—figure supplement 1 ) .\",\n \"In the CARM1–1 complex , 1 can adopt four alternative configurations ( Configuration I in Chains B , D; Configuration II in Chain C; Configuration III and IV in Chain A ) accompanied with the structural accommodation of the adjacent residues and water hydrogen bonds ( Figure 4b–d , Supplementary file 1-Table B–D ) .\",\n \"By contrast , only a single configuration of the ligands was observed in SAH- or SNF-bound CARM1 ( Figure 4c , d , PDB: 2Y1X , 2Y1W ) ( Sack et al . , 2011 ) .\",\n \"Detailed structural comparison of CARM1 in complex with SNF , SAH and 1 further revealed that 1 maintains common interactions observed in the CARM1–SNF and CARM1–SAH complexes with several exceptions ( Figure 4c , d , Supplementary file 1-Table B–D ) .\",\n \"Most noticeably , the CARM1–1 complex gains the strong hydrogen bonds via the 6′-methyleneamine moiety of the ligand with\",\n \"( i ) the backbone carbonyl of CARM1’s Glu258 ( Configuration I , II in Chains B , C , D and Configuration III in Chain A ) or\",\n \"( ii ) the side chain of Tyr154 ( Configuration IV in Chain A ) , together with several less conserved water hydrogen bonds ( Figure 4c , d , Supplementary file 1-Table B–D ) .\",\n \"By contrast , the 6′-amine of SNF in the CARM1–SNF complex forms weaker hydrogen bonds with Glu258 and may fewer water hydrogen bonds ( Figure 4c , Supplementary file 1-Table B , C , PDB: 2Y1W ) .\",\n \"Comparable interactions are completely absent from the CARM1–SAH complex ( Figure 4d , Supplementary file 1-Table B–D , PDB: 2Y1X ) .\",\n \"The desired hydrogen-bond networks of the 6′-methyleneamine moiety of 1 with CARM1 , which are present in the CARM1-1 complex but absent from the CARM1–SNF and CARM1–SAH complexes , can rationalize the significant decrease of IC50 from SNF and SAH to 1 .\",\n \"Another key difference among CARM1–1 , CARM1–SNF and CARM1–SAH complexes lies in the region around the carboxylate moiety of these ligands .\",\n \"In Chains A and C of the CARM1–1 complex , the carboxylate moiety of the ligand forms an ionic bond with Arg169 and a hydrogen bond with Gln160 ( Figure 4b , c , d , Supplementary file 1-Table B , C ) .\",\n \"Such interactions are absent from CARM1–SNF and CARM1–SAH complexes ( Figure 4d ) .\",\n \"By contrast , in Chains B and D of the CARM1–1 complex , the same carboxylate moiety forms the ionic bonds with Arg169 and a water hydrogen bond ( Figure 4b , c , d , Supplementary file 1-Table B , C ) .\",\n \"To accommodate the latter conformation , the Gln160 residue flips toward the 3′-ribosyl hydroxyl moiety of 1 to form a new hydrogen bond ( Figure 4b , c , Supplementary file 1-Table B , C ) .\",\n \"Similar interaction patterns can also be found in the CARM1–SNF and CARM1–SAH complexes ( Figure 4 , Supplementary file 1-Table B–D ) .\",\n \"With regards to the rest of the CARM1–ligand interactions , the CARM1 complexes with 1 , SAH and SNF are nearly identical except for slightly altered water hydrogen bonds ( Figure 4 , Supplementary file 1-Table S2–S4 ) .\",\n \"Here , the α-amino moiety of these ligands forms hydrogen bonds with the carbonyl backbone of Gly193 , as well as two water hydrogen bonds; their 2′ , 3′-ribosyl hydroxyl groups form two hydrogen bonds with the side chain of CARM1’s Glu215; adenine’s N1 and N6 form the hydrogen bonds with Asn243 and Glu244/Ser272 , respectively; and the adenine ring of these ligands is buried within a hydrophobic pocket .\",\n \"By contrast , there are fewer conserved water hydrogen bonds , such as those involved with the carboxylate and 3′-ribosyl hydroxyl moieties of 1 in Chain A of the CARM1–1 complex ( Figure 4 , Supplementary file 1-Table C ) .\",\n \"By contrast , adenine-N7 in the CARM1–SNF and CARM1–SAH complexes forms water hydrogen bonds bridged to Ser272 , which are absent from the CARM1–1 complex ( Figure 4 , Supplementary file 1-Table C , D ) .\",\n \"Collectively , the general high affinity of 1 , SNF and SAH ( Figure 4 , Supplementary file 1-Table B–D ) arises from the combined hydrophilic and hydrophobic interactions of these ligands with CARM1 .\",\n \"However , in comparison with SNF and SAH , 1 gains the extra interactions via its 6′-methyleneamine moiety ( Figure 4c , d , Supplementary file 1-Table B–D ) .\",\n \"In addition , 1 adopts the canonical pose with its α-amino carboxylate moiety interacting with Arg168 , which is similar to that of SNF and SAH but different from the noncanonical pose of 2a and 5a , upon binding CARM1 ( Figure 4d ) .\",\n \"Although the in vitro characterization demonstrated the potency and selectivity of 2a and 5a against CARM1 , we anticipated their poor membrane permeability as observed for structurally related analogs such as SAH and SNF ( Figure 1a ) ( Boriack-Sjodin et al . , 2016; Sack et al . , 2011 ) .\",\n \"The lack of membrane penetration is probably due to their primary amine moiety , which has pKa of ~10 and is fully protonated at a physiological pH of 7 . 4 .\",\n \"Given the essential roles of the 9′−amine moiety of 2a and 5a in CARM1 binding ( Figure 3e ) , we envisioned overcoming the membrane permeability issue through a pro-drug strategy by cloaking this amine moiety with a redox-triggered trimethyl-locked quinone butanoate moiety ( TML , Figure 5a ) ( Levine and Raines , 2012 ) .\",\n \"We thus prepared 6a and its control compound 6b by derivatizing 5a and 5b with the TML moiety ( Figure 1b , Figure 1—figure supplements 1 and 2 ) .\",\n \"To assess the cellular activity of 6a , we relied on our prior knowledge that CARM1 methylates the Arg1064 of BAF155 , a core component of the SWI/SNF chromatin-remodeling complex , and CARM1 knockout abolishes this posttranslational modification in MCF-7 cells ( Wang et al . , 2014a ) .\",\n \"Treatment of MCF-7 cells with 10 µM of 6a fully suppressed this methylation mark , whereas treatment with 2a and 5a did not affect this mark ( Figure 5b ) .\",\n \"We thus demonstrated the prodrug-like cellular activity of 6a .\",\n \"To further evaluate 6a as a chemical probe against CARM1 , we quantified the efficiency by which 6a engages CARM1 in a cellular context and thus suppresses the CARM1-dependent invasion by breast cancer cells .\",\n \"Because of the pro-drug character of 6a and its control compound 6b , we first developed quantitative LC-MS/MS methods to examine their cellular fates for CARM1 engagement ( see Materials and methods ) .\",\n \"Upon treatment of MDA-MB-231 cells with 6a , we observed its time- and dose-dependent intracellular accumulation ( Figure 5c , Figure 5—figure supplements 1 and 2 ) .\",\n \"We anticipated the conversion of the pro-drug 6a into 5a , but a striking finding is that 6a can also be readily processed into 2a inside cells ( Figure 5c , Figure 5—figure supplements 1 and 2 ) .\",\n \"Remarkably , >100 μM of 2a can be accumulated inside cells for 2 days after 6 hr treatment with a single dose of 5‒10 µM 6a .\",\n \"This observation probably reflects a slow efflux and thus effective intracellular retention of 2a due to its polar α-amino acid zwitterion moiety .\",\n \"Given that cellular CARM1 inhibition is involved with multiple species ( 2a , 5a and 6a ) in competition with SAM , we modeled the ligand occupancy of cellular CARM1 on the basis of their Kd values ( Kd , 2a=17 nM , Kd , 5a=9 nM , Kd , 6a=0 . 28 µM and Kd , SAM≈Km , SAM=0 . 25 µM ) and MS-quantified intracellular concentrations ( Figure 5d , Figure 5—figure supplement 3 , Equations 5–7 , see Materials and methods ) .\",\n \"The SAM cofactor , whose intracellular concentration was determined to be 89 ± 16 µM ( Figure 5c , d ) , is expected to occupy >99 . 5% CARM1 with residual <0 . 5% as the apo-enzyme under a native setting .\",\n \"With single doses of 6a of 2 . 5‒10 µM , the combined CARM1 occupancy by 2a , 5a and their pro-drug precursor 6a rapidly reached the plateau of >95% within 6 hr , and was maintained at this level for at least 48 hr ( Figure 5e , Figure 5—figure supplements 1 and 2 ) .\",\n \"Notably , treatment with 6a concentrations as low as 0 . 5 µM is sufficient to reach 60% target engagement within 10 hr and to maintain this occupancy for 48 hr ( Figure 5e , Figure 5—figure supplements 1 and 2 ) .\",\n \"The time- and dose-dependent progression of the CARM1 occupancy by these ligands thus provides quantitative guidance upon the treatment of MDA-MB-231 cells with 6a .\",\n \"Given that 2a is the predominant metabolic product of 6a within cells ( Figure 5c , e ) and also shows certain affinity to SMYD2 ( ~10 fold higher IC50 in comparison with CARM1 , Figure 1c , Supplementary file 1-Table A ) , we evaluated SMYD2 engagement of 2a for its potential off-target effect .\",\n \"In a similar manner to that described for the ligand occupancy of cellular CARM1 , we modeled the occupancy of cellular SMYD2 by 2a on the basis of Kd , 2a=150 nM and Kd , SAM=60 nM for SMYD2 ( Figure 5e , Figure 5—figure supplement 4 , Materials and methods ) .\",\n \"Largely because of the high affinity of 2a to SAM and thus a 37-fold larger Kd , 2a/Kd , SAM ratio of SMYD2 relative to CARM1 ( Kd , 2a/Kd , SAM of 2 . 5 and 0 . 068 for SMYD2 and CARM1 , respectively ) , the occupancy of SMYD2 by 2a is below 20% ( Figure 5e , Figure 5—figure supplement 4 ) under the efficacy doses of 6a ( see cellular data for 6a below ) .\",\n \"We were thus less concerned about the SMYD2-associated off-target effect under our assay conditions .\",\n \"The metabolic stability of 6a was also evaluated with a microsomal stability assay ( Figure 5c , e , Figure 5—figure supplement 5 , see Materials and methods ) .\",\n \"In the presence of rat liver microsomes , 6a showed decent stability with 24% residual 6a after one-hour incubation .\",\n \"Here , the conversion of 6a into 5a accounted for 40% of the microsome-processed 6a; no production of 2a was detected .\",\n \"Such observation suggests that NQO1 , the putative enzyme candidate to reduce the TML moiety in 6a or 6b , is present in microsomes as well as in tumor cells ( Dias et al . , 2018; Huang et al . , 2016 ) .\",\n \"By contrast , peptidase enzymes that are expected to process 5a into 2a are absent from microsomes but rich in tumor cells .\",\n \"We then conducted a cellular thermal shift assay ( CETSA ) , in which ligand binding is expected to increase CARM1’s thermal stability in a cellular context ( Jafari et al . , 2014 ) .\",\n \"Our data showed that the treatment of MDA-MB-231 cells with 6a but with not the control compound 6b increased cellular Tm and thus the thermal stability of CARM1 by 4 . 3 ± 0 . 6°C ( Figure 5f ) .\",\n \"The distinct effect of 6a in contrast to 6b on the cellular Tm of CARM1 aligns well with the 4 . 1‒6 . 2°C difference in the in vitro Tm of CARM1 upon binding 2a and 5a versus SAM ( Figure 2e ) .\",\n \"Here , 6b can penetrate cell membranes and be processed into 5b and 2b in a similar manner as 6a ( Figure 5—figure supplement 6 ) .\",\n \"These observations thus present the cellular evidence to show that CARM1 engages 2a and 5a .\",\n \"To further characterize 6a as a CARM1 chemical probe , we examined the Arg1064 methylation of BAF155 and the Arg455/Arg460 methylation of PABP1 , two well-characterized cellular methylation marks of CARM1 , upon treating MDA-MB-231 cells with 6a ( Lee and Bedford , 2002; Wang et al . , 2014a ) .\",\n \"These methylation marks can be fully suppressed by 6a in a dose-dependent manner ( Figure 6a ) .\",\n \"The resultant EC50 values of 0 . 45‒0 . 75 µM ( Figure 6b ) are well correlated with the modeled 60% cellular occupancy of CARM1 upon treatment with 0 . 5 µM 6a for 48 hr ( Figure 5e ) .\",\n \"By contrast , the treatment of the negative control compound 6b showed no effect on these methylation marks ( Figure 6a ) .\",\n \"We therefore demonstrated the robust use of 6a ( SKI-73 ) as a CARM1 chemical probe and of 6b ( SKI-73N ) as its control compound .\",\n \"After demonstrating the utility of SKI-73 ( 6a ) as a chemical probe for CARM1 , we examined whether chemical inhibition of CARM1 can recapitulate biological outcomes that are associated with CARM1 knockout ( CARM1-KO ) ( Wang et al . , 2014a ) .\",\n \"Our prior work showed that CARM1’s methyltransferase activity is required for invasion of MDA-MB-231 cells ( Wang et al . , 2014a ) .\",\n \"We thus conducted a matrigel invasion assay with MDA-MB-231 cells in the presence of 6a .\",\n \"Relative to the control treatment with DMSO , treatment with SKI-73 ( 6a ) but not its negative control compound SKI-73N ( 6b ) suppressed the invasion of MDA-MB-231 cells ( EC50 = 1 . 3 µM ) ( Figure 6c , d ) .\",\n \"The treatment with ≥10 µM 6a produced the maximal 80% suppression of the invasion by MDA-MB-231 relative to the DMSO control , which is comparable with the phenotype of CARM1-KO ( Figure 6e ) .\",\n \"Critically , no further inhibition by 6a on the invasiveness was observed upon 6a treatment ( in comparison with the treatment with DMSO or 6b treatment ) of MDA-MB-231 CARM1-KO cells ( Figure 6e ) .\",\n \"Notably , treatment with 6a and 6b under the current condition has no apparent impact on the proliferation of parental or CARM1-KO MDA-MB-231 cells ( Figure 6—figure supplement 1 ) , consistent with the intact proliferation upon treatment with other CARM1 chemical probes ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018 ) .\",\n \"These results suggest that SKI-73 ( 6a ) and CARM1 knockout perturb the common , proliferation-independent biological process and then suppresses 80% of the invasiveness of MDA-MB-231 cells .\",\n \"We thus characterized SKI-73 ( 6a ) as a chemical probe that can be used to interrogate the CARM1-dependent invasion of breast cancer cells .\",\n \"Because of the advancement of scRNA-seq technology , stunning subpopulation heterogeneity has been uncovered even for well-defined cellular types ( Tanay and Regev , 2017 ) .\",\n \"In the context of tumor metastasis , including its initial invasion step , epigenetic plasticity is required to allow a small subset of tumor cells to adapt distinct transcriptional cues for neo-properties ( Chatterjee et al . , 2018; Flavahan et al . , 2017; Wu et al . , 2019 ) .\",\n \"To explore the feasibility of dissecting the CARM1-dependent , invasion-prone subset of MDA-MB-231 breast cancer cells , we formulated a cell-cycle-aware algorithm of scRNA-seq analysis and dissected those subpopulations that were sensitive to CARM1 perturbation ( Figure 7a , see Materials and methods ) .\",\n \"Here we conducted 10 × Genomics droplet-based scRNA-seq of 3232 , 3583 and 4099 individual cells ( a total of 10 , 914 cells ) exposed to 48 hr treatment with SKI-73 ( 6a ) , SKI-73N ( 6b ) and DMSO , respectively .\",\n \"Guided by Silhouette analysis , cell-cycle-associated transcripts were identified as dominant signatures of subpopulations ( Figure 7—figure supplements 1–18 ) .\",\n \"These signatures naturally exist for proliferative cells and are not expected to be specific for the invasive phenotype .\",\n \"To dissect the subpopulation-associated transcriptomic signatures of invasive cells , we included one additional layer for hierarchical clustering by first classifying the individual cells into G0/G1 , S , and G2/M stages ( 6885 , 1520 and 2509 cells , respectively ) ( Figure 7—figure supplement 6 , Supplementary file 1-Table E ) , and then conducted the unsupervised clustering within each cell-cycle-aware subset ( Figure 7b , Figure 7—figure supplements 19–30 , Supplementary file 1-Table E ) .\",\n \"To resolve the subpopulations without redundant clustering , we developed an entropy analysis method and relied on the Fisher Exact test ( see Materials and methods ) .\",\n \"The optimal scores of the combined methods were implemented to determine the numbers of cluster for each subset ( Figure 7b , Figure 7—figure supplements 20 , 24 and 28 ) ( Butler et al . , 2018 ) .\",\n \"The cell-cycle-aware algorithm allowed the clustering of these cells according to the three cell cycle stages under the three treatment conditions and resulted in 21 , 7 and 6 subpopulations in G0/G1 , S , and G2/M phases , respectively ( Figure 7b , Figure 7—figure supplements 21 , 25 and 29 , Supplementary file 1-Tables F–H ) .\",\n \"Notably , the 48 hr treatments with SKI-73 ( 6a ) or SKI-73N ( 6b ) had no effect on the cell cycle , as indicated by their comparable cell-cycle distribution patterns ( Figure 7—figure supplement 6; Supplementary file 1-Table E ) , a finding that is consistent with intact proliferation after all three treatments ( Figure 6—figure supplement 1 ) .\",\n \"With the 21 , 7 and 6 subpopulations clustered into the G0/G1 , S , and G2/M stages , respectively , we then conducted population analysis , comparing SKI-73 ( 6a ) and SKI-73N ( 6b ) against DMSO ( Figure 7c , Figure 7—figure supplements 31 and 32 , Supplementary file 1-Table F–H ) .\",\n \"These subpopulations can be readily classified into five distinct categories according to their cell-cycle-aware responses to SKI-73 ( 6a ) and SKI-73N ( 6b ) treatment: commonly resistant/emerging/depleted versus differentially depleted/emerging ( SKI-73/SKI-73N- or 6a/6b-specific ) ( Figure 7c , Figure 7—figure supplements 31 and 32 , Supplementary file 1-Table F–H ) .\",\n \"Here , we are particularly interested in the SKI-73 ( 6a ) -specific depleted subpopulations ( Subpopulation 0/2/8/11/13/14/17/19 of G0/G1-phase cells and 3 of S-phase cells ) as the potential invasion-associated subpopulations , given their sensitivity to SKI-73 ( 6a ) but not its control compound SKI-73N ( 6b ) .\",\n \"The subpopulations that remain unchanged after the treatment with SKI-73 ( 6a ) or SKI-73N ( 6b ) ( Subpopulation 1/7/12/15 of G0/G1-phase cells; 2/4/5 of S-phase cells; 1 of G2/M-phase cells ) were defined as the common resistant subset .\",\n \"SKI-73 ( 6a ) -specific emerging subpopulations ( Subpopulation 3/4/5/6/16 of G0/G1-phase cells; 6 of S-phase cells; 4 of G2/M-phase cells ) are expected to be suppressed by CARM1 but emerge upon its inhibition .\",\n \"The remaining subpopulations are either associated with the effects of the small-molecule scaffold of SKI-73 ( 6a ) /SKI-73N ( 6b ) ( commonly emerging Subpopulation 9 of G0/G1-phase cells , 0/5 of G2/M-phase cells; commonly depleted Subpopulation 2/3 of G2/M-phase cells ) or with SKI-73N ( 6b ) -specific effects ( differentially depleted Subpopulation 10/18 of G0/G1-phase cells , 0 of S-phase cells; differentially emerging Subpopulation 20 of G0/G1-phase cells ) .\",\n \"Interestingly , scRNA-seq analysis of CARM1-KO cells ( in comparison with SKI-73 ( 6a ) -treated cells ) suggests that CARM1 knockout has more profound effects on the overall landscape of epigenetic plasticity ( Figure 7d , Figure 7—figure supplement 33 ) .\",\n \"Collectively , the chemical probe SKI-73 ( 6a ) alters the epigenetic plasticity of MDA-MB-231 breast cancer cells via the combined effects of SKI-73 ( 6a ) ’s molecular scaffold and specific inhibition of CARM1’s methyltransferase activity .\",\n \"Given that SKI-73 ( 6a ) has no effect on the cell cycle and proliferation of MDA-MB-231 cells under the current treatment dose and duration , we envision that the invasion capability of MDA-MB-231 cells mainly arises from an invasion-prone subset , 80% of which is depleted by SKI-73 ( 6a ) treatment ( Figure 6c–e ) .\",\n \"We thus focused on Subpopulations 0/2/8/11/13/14/17/19 of G0/G1-phase cells and Subpopulation 3 of S-phase cells: in total nine depleted subpopulations specific for SKI-73 ( 6a ) ( Figure 7c , Figure 7—figure supplements 31 and 32 , Supplementary file 1-Table F–H ) .\",\n \"To identify invasion-prone subpopulation ( s ) among these candidates , we compared the transcriptional signature ( s ) of these subpopulations with those of cells that freshly invaded through Matrigel within 16 hr .\",\n \"Strikingly , in comparison with the highly heterogenous scRNA-seq signature of the parental MDA-MB-231 cells , the freshly harvested invasive cells ( 3793 cells for scRNA-seq ) are relatively homogeneous , with their subpopulations mainly determined by the cell-cycle-related transcriptomic signatures ( Figure 7d , Figure 7—figure supplements 33–37 ) .\",\n \"We then classified the freshly harvested invasive cells into G0/G1 , S and G2/M stages ( Figure 7—figure supplements 38–40 , Supplementary file 1-Table E ) .\",\n \"Through correlation analysis comparing the invasive cells and the subpopulations within each cell-cycle stage ( Figure 7d , Figure 7—figure supplements 39–42 ) , we revealed the subsets whose transcriptional signatures closely relate to those of the invasive cells , including Subpopulations 6/7/8/9/14 in G0/G1-phase cells , 0/3 in S-phase cells and 1/2 of G2/M-phase cells ( Supplementary file 1-Table F–K ) .\",\n \"In the context of population analysis for the nine SKI-73 ( 6a ) -specific depleted subpopulations , Subpopulations 8/14 of G0/G1-phase cells and Subpopulation 3 in S-phase are putative invasion-prone candidates .\",\n \"Subpopulation 8 of G0/G1-phase cells is the most sensitive and the only subpopulation that can be depleted by around 80% with SKI-73 ( 6a ) treatment ( Figure 7c ) .\",\n \"Given the ~80% suppression and ~20% residual invasion capability upon SKI-73 ( 6a ) treatment , we argue that the invasive phenotype of MDA-MB-231 cells predominantly arises from Subpopulation 8 G0/G1-phase cells , which only accounts for ~8% of the parental cells in G0/G1 phase ( ~5% without cell-cycle awareness ) .\",\n \"Differential expression analysis further revealed the single-cell transcriptional signatures of metastasis-implicated genes ( e . g . MORC4 , S100A2 , RPL39 , IFI27 , ARF6 , CHD11 , SDPR and KRT18 ) that are specific for the G0/G1-phase Subpopulation 8 and invasive cells but not for other G0/G1-phase invasion-prone candidates such as Subpopulation 6/7/9/14 ( Figure 7e , Figure 7—figure supplement 43 , Supplementary file 1-Table L ) .\",\n \"The remaining cells of G0/G1-phase Subpopulation 8 after SKI-73 ( 6a ) treatment ( Figure 7c , d ) together with others ( subpopulation-6/7/9/14 in G0/G1-phase cells , 0/3 in S-phase cells and 1/2 of G2/M-phase cells , Figure 7—figure supplements 31 , 32 , 41 , 42 ) may account for the 20% residual invasion capacity .\",\n \"In the context of SKI-73 ( 6a ) -specific depletion of G0/G1-phase subpopulations , there are SKI-73 ( 6a ) -specific emerging G0/G1-phase subpopulations: Subpopulation 3/4/5/6/16 ( Figure 7c ) .\",\n \"Population analysis of G0/G1-phase cells further revealed that Subpopulations 4 and 16 account for 90% of the emerging subset upon SKI-73 ( 6a ) treatment ( Supplementary file 1-Table F ) .\",\n \"The transcriptional signatures and probably the associated invasion capability of Subpopulations 4 and 16 are dramatically different from those of the freshly harvested invasive cells and the bulk population of the parental cells , including the invasion-prone Subpopulation 8 ( Figure 7b , d ) .\",\n \"Collectively , either CARM1 knockout or CARM1 inhibition with SKI-73 ( 6a ) alters the epigenetic plasticity in a proliferation-independent manner by replacing the most invasion-prone subpopulation with the non-invasive subpopulation ( s ) to suppress the invasive phenotype ( Figure 7f ) .\"\n ],\n [\n \"On the basis of a novel small-molecule scaffold , 6′-homosinefungin ( HSF ) , SKI-73 ( 6a ) was developed as a pro-drug-like chemical probe for CARM1 by cloaking the 9′-amine moiety of 5a with the TML moiety .\",\n \"SKI-73N ( 6b ) was developed as a control compound for SKI-73 ( 6a ) .\",\n \"The inhibitory activity of SKI-73 ( 6a ) against CARM1 was demonstrated by the ability of SKI-73 ( 6a ) but not SKI-73N ( 6b ) to abolish the cellular methylation marks of CARM1: the Arg1064 methylation of BAF155 and the Arg455/Arg460 methylation of PABP1 ( Lee and Bedford , 2002; Wang et al . , 2014a ) .\",\n \"The ready intracellular cleavage of TML is expected for the conversion of SKI-73 and SKI-73N ( 6a and 6b ) into 5a and 5b , respectively , but it is remarkable that SKI-73 and SKI-73N ( 6a and 6b ) can also be efficiently processed into 2a and 2b inside cells .\",\n \"In comparison , 6a showed decent metabolic stability with no production of 2a in the presence of microsomes .\",\n \"Here , 2a and 5a are presented as potent and selective CARM1 inhibitors , whereas their control compounds 2b and 5b interact poorly with CARM1 .\",\n \"Competitive assays with the SAM cofactor and the peptide substrate showed that 2a and 5a act on CARM1 in a SAM-competitive and substrate-noncompetitive manner .\",\n \"The SAM-competitive mode is consistent with the ligand–complex structures of CARM1 , in which the SAM binding site is occupied by 2a and 5a .\",\n \"Strikingly , as revealed by their ligand–CARM1 complex structures , 2a and 5a engage CARM1 through noncanonical modes , with their 6′-N-benzyl moieties in the binding pocket that is otherwise occupied by the α-amino carboxylate moiety of conventional SAM analogs such as SAH , SNF and 1 .\",\n \"This observation is consistent with the 4 . 1‒6 . 5 °C increase in the in vitro and cellular Tm of CARM1 upon binding 2a and 5a , which contrasts with the smaller Tm changes with SAM as a ligand .\",\n \"The distinct modes of interaction of CARM1 with 2a and 5a ( Figure 3c , g ) also rationalize the CARM1 selectivity of the two SAM analogs over other methyltransferases , including closely related PRMT homologs .\",\n \"Through mathematic modeling using the inputs of the LC-MS/MS-quantified intracellular concentrations and CARM1-binding constants of relevant HSF derivatives and the SAM cofactor , we concluded that high intracellular concentrations of 5a and 2a , and thus efficient CARM1 occupancy , can be achieved rapidly and maintained for several days with a single low dose of SKI-73 ( 6a ) .\",\n \"By contrast , the occupancy by 5a and 2a of SMYD2 , the next likely engaged target , is below 20% with the efficacious doses of 6a that affect cell invasion .\",\n \"The polar α-amino acid zwitterion moiety of 2a and the polar α-amino moiety of 5a probably account for their accumulation and retention inside cells .\",\n \"To the best of our knowledge , EZM2302 , TP-064 , and SKI-73 ( also 6a in this work , www . thesgc . org/chemical-probes/SKI-73 ) and their derivatives are the only selective and cell-active CARM1 inhibitors ( Drew et al . , 2017; Nakayama et al . , 2018 ) .\",\n \"Although the potency , selectivity , on-target engagement and potential off-target effects associated with these compounds have been examined in vitro and in cellular contexts as chemical probes , EZM2302 , TP-064 , and SKI-73 ( 6a ) are differentiated by their molecular scaffolds and modes of interaction with CARM1 ( www . thesgc . org/chemical-probes/SKI-73 ) ( Drew et al . , 2017; Nakayama et al . , 2018 ) .\",\n \"SKI-73 ( 6a ) is a cofactor analog inhibitor embedding a N6'-homosinefungin moiety to engage the SAM binding site of CARM1 in a cofactor-competitive , substrate-noncompetitive manner; EZM2302 and TP-064 occupy the substrate-binding pocket of CARM1 in a SAH-uncompetitive or SAM-noncompetitive manner ( Drew et al . , 2017; Nakayama et al . , 2018 ) .\",\n \"In particular , the prodrug property of SKI-73 ( 6a ) allows its ready cellular uptake , followed by rapid conversion into its active forms inside cells .\",\n \"The prolonged intracellular CARM1 inhibition further distinguishes SKI-73 ( 6a ) from EZM2302 and TP-064 .\",\n \"With SKI-73 ( 6a ) as a CARM1 chemical probe and SKI-73N ( 6b ) as a control compound , we showed that pharmacological inhibition of CARM1 with SKI-73 ( 6a ) , but not SKI-73N ( 6b ) , suppressed 80% of the invasion capability of MDA-MB-231 cells .\",\n \"By contrast , the pharmacological inhibition of CARM1 with SKI-73 ( 6a ) had no effect on the proliferation of MDA-MB-231 cells .\",\n \"This result is consistent with the lack of anti-proliferation activities of the other two CARM1 chemical probes , EZM2302 and TP-064 , against breast cancer cell lines ( Drew et al . , 2017; Nakayama et al . , 2018 ) .\",\n \"The anti-invasion efficiency of SKI-73 ( 6a ) is in good agreement with the intracellular occupancy and the resulting abolition of several methylation marks of CARM1 upon treatment with SKI-73 ( 6a ) .\",\n \"Our prior work showed that the methyltransferase activity of CARM1 is required for breast cancer metastasis ( Wang et al . , 2014a ) .\",\n \"Among the diverse cellular substrates of CARM1 ( Blanc and Richard , 2017 ) , BAF155—a key component of the SWI/SNF chromatin-remodeling complex–is essential for the invasion of MDA-MB-231 cells ( Wang et al . , 2014a ) .\",\n \"Mechanistically , the CARM1-mediated Arg1064 methylation of BAF155 facilitates the recruitment of the SWI/SNF chromatin-remodeling complex to a specific subset of gene loci ( Wang et al . , 2014a ) .\",\n \"Replacement of the native CARM1 with its catalytically dead mutant or with an Arg-to-Lys point mutation at the Arg1064 methylation site of BAF155 is sufficient to abolish the invasive capability of breast cancer cells ( Wang et al . , 2014a ) .\",\n \"CARM1 inhibition with SKI-73 ( 6a ) , but not with its control compound SKI-73N ( 6b ) , recapitulates the anti-invasion phenotype associated with the genetic perturbation of CARM1 .\",\n \"More importantly , there is no additive effect upon combining CARM1-KO with SKI-73 ( 6a ) treatment , underlying the fact that the two orthogonal approaches target the commonly shared pathway ( s ) that are essential for the invasion of breast cancer cells .\",\n \"In comparison to SKI-73 ( 6a ) , the CARM1 inhibitors EZM2302 and TP-064 demonstrated anti-proliferation effects on hematopoietic cancer cells , in particular multiple myeloma ( Drew et al . , 2017; Greenblatt et al . , 2018; Nakayama et al . , 2018 ) .\",\n \"Mechanistically , genetic perturbation of CARM1 in the context of leukemia impairs cell-cycle progression , promotes myeloid differentiation , and ultimately induces apoptosis , probably by targeting pathways of proliferation and cell-cycle progression , that is , E2F- , MYC- , and mTOR-regulated processes ( Greenblatt et al . , 2018 ) .\",\n \"In comparison , CARM1 inhibition with EZM2302 led to a slightly different phenotype , which includes reduction of RNA stability , E2F target downregulation , and induction of a p53 response signature for senescence .\",\n \"( Greenblatt et al . , 2018 ) .\",\n \"Collectively , the effects of CARM1 chemical probes are highly context-dependent , with SKI-73 ( 6a ) having different uses in impairing the invasiveness of breast cancer cells , while TP-064 and EZM2302 have uses in preventing the proliferation of hematopoietic cancer cells .\",\n \"Given the increased awareness of epigenetic plasticity ( Flavahan et al . , 2017 ) , we employed the scRNA-seq approach to examine MDA-MB-231 cells and their responses to chemical and genetic perturbation with CARM1 .\",\n \"Because of the lack of a prior reference to define subpopulations of MDA-MB-231 cells , we developed a cell-cycle-aware algorithm to cluster the subpopulations with a resolution that was able to dissect subtle changes upon treatment with SKI-73 ( 6a ) versus its control compound SKI-73N ( 6b ) in each cell-cycle stage .\",\n \"Guided by Silhouette analysis , the population entropy analysis and the Fisher Exact test , >10 , 000 MDA-MB-231 breast cancer cells were classified on the basis of their cell-cycle stages and then clustered into 34 subpopulations .\",\n \"With further annotation of these subpopulations according to their different responses to treatment with SKI-73 ( 6a ) versus SKI-73N ( 6b ) , we readily dissected the subpopulations that were altered in a SKI-73 ( 6a ) -specific ( CARM1-dependent ) manner and then identified subsets with transcriptional signatures that are similar to that of the freshly isolated invasive cells .\",\n \"Quantitative analysis of SKI-73 ( 6a ) -depleted subpopulations further revealed the most invasion-prone subpopulation , which accounts for only 5% of the total population but at least 80% of the invasive capability of the parental cells .\",\n \"Collectively , we propose a model in which MDA-MB-231 cells consist of subpopulations , with their epigenetic plasticity ( Figure 7f ) determined by multiple factors including the CARM1-involved BAF155 methylation ( Wang et al . , 2014a ) .\",\n \"SKI-73 ( 6a ) inhibits the methyltransferase activity of CARM1 , the Arg1064 methylation of BAF155 , and thus the target genes associated with the methylated BAF155 .\",\n \"These effects alter the cellular epigenetic landscape by affecting certain subpopulations of MDA-MB-231 cells without any apparent effect on cell cycle and proliferation .\",\n \"In the context of the invasion phenotype of MDA-MB-231 cells , the subset of invasion-prone cells is significantly suppressed upon the treatment with SKI-73 ( 6a ) .\",\n \"Essential components that are used to dissect the invasion-prone population in this CARM1-dependent epigenetic plasticity model are the scRNA-seq analysis of sufficient MDA-MB-231 cells ( >10 , 000 cells here ) , the utility of the freshly isolated invasive cells as the reference , the timing and duration of treatment , and the use of SKI-73N ( 6b ) and DMSO as controls .\",\n \"Interestingly , although the invasion-prone subpopulation is also abolished in the CARM1-KO strain , CARM1-KO reshapes the epigenetic plasticity in a much more profound manner , significantly reducing the subpopulation heterogeneity of MDA-MB-231 cells .\",\n \"The distinct outcomes for the pharmacological and genetic perturbation could be due to their different modes of action: short-term treatment with SKI-73 ( 6a ) versus long-term clonal expansion of CARM1-KO cells .\",\n \"The pharmacological inhibition captures the immediate response , whereas the genetic perturbation reports long-term and potential resistant outcomes .\",\n \"This work thus presents a new paradigm to understand cancer metastasis in the context of epigenetic plasticity and provides guidance for similar analyses in broader contexts: other cell lines , patient-derived xenograft samples , and in vivo mouse models of breast cancer .\"\n ],\n [\n \"SAM- and substrate-dependent IC50 values were determined with the radiometric filter paper assay as described above except in the presence of the varied concentrations of the SAM cofactor and the peptide substrate .\",\n \"The IC50 values of 2a and 5a were obtained in the presence of 0 . 09 , 0 . 18 , 0 . 37 , 0 . 75 , 1 . 50 , 2 . 25 , 3 . 00 , 5 . 62 , and 7 . 50 µM [3H-Me]-SAM or 0 . 3 , 1 . 5 , 7 . 5 , 15 , 30 , and 50 µM H3 peptide ( aa 1–40 ) .\",\n \"The resultant IC50 values were plotted as a function of the concentrations of SAM and the peptide substrate using GraphPad Prism Software .\",\n \"Given the SAM-competitive character of 2a and 5a , their SAM-dependent IC50 values were then fitted according to Equation\",\n \"1 . Here [SAM] is the concentration of SAM , Km , SAM is the Michaelis-Menten constant , and Kd is the dissociation constant of the CARM1 inhibitor 2a or 5a .\",\n \"Kd , Kd/Km , SAM and Km , SAM can therefore be obtained through the intercept of the y-axis , the slope , and their ratio , respectively , upon fitting Equation\",\n \"1 . Given that 6a showed a much higher IC50 ( 1 . 1 ± 0 . 1 µM , Figure 5—figure supplement 3 ) , the Kd value of 6a ( Kd , 6a = 0 . 32 µM ) was directly obtained from the IC50 value , the Km , SAM derived above , and the assayed SAM concentration according to Equation 1 . ( 1 ) IC50=[SAM]×kd/km , SAM+kd Full-length CARM1 was used for SPR assay .\",\n \"DNA fragment encoding the full-length CARM1 was cloned into pFB-N-flag-LIC donor plasmid .\",\n \"The resulting plasmid was transformed into DH10Bac-competent E . coli cells ( Invitrogen ) and a recombinant Bacmid DNA was purified , followed by a recombinant baculovirus generation in Sf9 insect cells .\",\n \"Sf9 cells grown in HyQ SFX insect serum-free medium ( ThermoScientific ) were infected with 10 mL of P3 viral stocks per 1 L of suspension cell culture and incubated at 27°C using a platform shaker set at 150 revolutions per minute .\",\n \"The cells were collected when viability dropped to 70–80% ( ~72 hr after infection ) .\",\n \"Harvested cells were re-suspended in PBS , 1X protease inhibitor cocktail ( 100X protease inhibitor stock in 70% ethanol containing 0 . 25 mg/mL aprotinin , 0 . 25 mg/ml leupeptin , 0 . 25 mg/mL pepstatin A and 0 . 25 mg/mL E-64 ) and 2X Roche complete EDTA-free protease inhibitor cocktail tablet .\",\n \"The cells were lysed chemically by rotating 30 min with NP40 ( final concentration of 0 . 6% ) and 50 U/mL benzonase nuclease ( Sigma ) , 2 mM 2-mercaptoethanol and 10% glycerol followed by sonication at frequency of 7 ( 10′‘on/10′‘off ) for 2 min ( Sonicator 3000 , Misoni ) .\",\n \"The crude extract was clarified by high-speed centrifugation ( 60 min at 36 , 000 × g at 4°C ) by Beckman Coulter centrifuge .\",\n \"The recombinant protein was purified by incubating the cleared lysate with anti-FLAG M2 affinity agarose gel ( Sigma , Cat # A2220 ) and then rotating for 3 hr , followed by washing with 10 CV TBS ( 50 mM Tris-HCl , 150 mM NaCl , pH 7 . 4 ) containing 2 mM 2-mercaptoethanol , 1X protease inhibitor cocktail ( 100X protease inhibitor stock in 70% ethanol containing 0 . 25 mg/mL aprotinin , 0 . 25 mg/mL leupeptin , 0 . 25 mg/mL pepstatin A and 0 . 25 mg/mL E-64 ) and 1X Roche complete EDTA-free protease inhibitor cocktail tablet .\",\n \"The recombinant protein was eluted by competitive elution with a solution containing 100 µg/mL FLAG peptide ( Sigma , Catalog # F4799 ) in 20 mM Tris pH:7 . 4 , 150 mM NaCl , 5% glycerol , 3 mM 2-mercaptoethanol .\",\n \"Quality of CARM1 ( >95% ) was determined by SDS–PAGE .\",\n \"The protein was then concentrated , flash frozen with liquid nitrogen , and stored at −80°C for future use .\",\n \"SPR analysis was performed using a Biacore T200 ( GE Health Sciences Inc ) at 25°C .\",\n \"Approximately 5500 response units of CARM1 ( amino acids 1–608 ) were amino-coupled onto a CM5 chip in one flow cell according to the protocol of the manufacturer .\",\n \"Another flow cell was left empty for reference subtraction .\",\n \"SPR analysis was conducted in the HBS-EP buffer ( 20 mM HEPES pH 7 . 4 , 150 mM NaCl , 3 mM EDTA , 0 . 05% Tween-20 ) containing 2% ( v/v ) DMSO .\",\n \"The stock solutions of five concentrations of compound 2a ( 24 . 7 , 74 . 1 , 222 , 667 and 2000 nM ) and four concentrations of 5a ( 6 . 2 , 18 . 5 , 55 . 6 and 167 nM ) were prepared by serial dilution .\",\n \"Binding kinetic experiments were performed with single cycle kinetics with the contact time of 60 s and off time of 300 s in a flow rate of 30 µL/min .\",\n \"To facilitate complete dissociation of compound for the next cycle , a regeneration step ( 300 s , 40 µL/min of buffer ) , a period of stabilization ( 120 s ) and two blank cycles were included between each cycle .\",\n \"Kinetic curve fittings were carried out with a 1:1 binding model or a heterogeneous ligand model using Biacore T200 Evaluation software ( GE Health Sciences Inc ) .\",\n \"TSA was performed as described previously ( Blum et al . , 2014; Niesen et al . , 2007 ) to examine the melting temperature ( Tm ) of CARM1 in the presence or absence of ligands .\",\n \"For each measurement ( triplicate of each data point ) , the assay solution containing 50 mM HEPES-HCl ( pH = 8 . 0 ) , Tween-20 0 . 005% ( v/v ) , 1 mM TCEP , 0 . 5 μM CARM1 , and 5 μM ligand ( SAM , 1 , 2a or 5a ) was mixed with 5 × SYPRO Orange Protein Gel Stain stock ( Sigma Aldrich ) in a 96-well PCR plate .\",\n \"The mixture was equilibrated at 25°C in dark for 5 min and then loaded onto Bio-Rad CFX96 Real-Time PCR Detection System .\",\n \"The fluorescence readouts were recorded upon increasing the heating temperature from 25°C to 100°C at a rate of 0 . 2 °C/s .\",\n \"The raw data of the fluorescence readouts versus the temperatures were exported with CFX software and processed as the percentage of the fluorescent signal normalized between the lowest readout of 0% and the highest readout of 100% within the 25‒100°C region .\",\n \"The melting curves were plotted as the normalized fluorescence ( % ) versus the heating temperature and fit with a sigmoid curve with GraphPad Prism .\",\n \"The Tm corresponds to the temperature with the 50% relative fluorescent signal in the sigmoidal curve .\",\n \"A DNA fragment encoding the methyltransferase domain of human CARM1 ( residues 140–480 ) was cloned into a baculovirus expression vector pFBOH-MHL ( http://www . thesgc . org/sites/default/files/toronto_vectors/pFBOH-MHL . pdf ) .\",\n \"The protein was expressed in Sf9 cells as an N-terminal 6 × His tag fusion protein and purified by metal chelating affinity chromatography ( TALON resin , Clontech , Mountain View , CA , USA ) followed by size-exclusion chromatography ( Superdex 200 , GE Healthcare ) .\",\n \"Pooled fractions containing CARM1 were subjected to the treatment of tobacco etch virus to remove the 6 × His tag .\",\n \"The protein was purified to homogeneity by ion-exchange chromatography .\",\n \"Purified CARM1 ( 6 . 5 mg/mL ) was crystallized with the sitting drop vapor diffusion method at 20°C .\",\n \"For the CARM1–1 complex , CARM1 was mixed with 1 at a 1:5 molar ratio ( protein:ligand ) and crystallized with the sitting drop vapor diffusion method at 20°C by mixing 1 µL of the protein solution with 1 µL of the reservoir solution containing 20% PEG3350 and 0 . 2 M diammonium tartrate .\",\n \"X-ray diffraction data for the CARM1–1 complex were collected at 100 K at beam line 23ID-B of Advanced Photon Source ( APS ) , Argonne National Laboratory .\",\n \"Data sets were processed using the HKL-2000 suite ( Otwinowski and Minor , 1997 ) .\",\n \"The structure of the CARM1–1 complex was solved by molecular replacement using MOLREP ( Otwinowski and Minor , 1997 ) with the PDB entry 2V74 as the search template .\",\n \"REFMAC ( Emsley and Cowtan , 2004; Murshudov et al . , 1997 ) was used for the structure refinement .\",\n \"Graphics program COOT was used for model building and visualization .\",\n \"Crystal diffraction data and refinement statistics for the structure are displayed in Table\",\n \"2 . To further confirm the electronic densities of the ligand , the total omission electron density map was calculated using SFCHECK from CCP4suite and contoured at 1 . 0 sigma ( Figure 4—figure supplement 1 ) ( Emsley and Cowtan , 2004; Murshudov et al . , 1997 ) .\",\n \"For the CARM1-5a complex , CARM1 was crystallized with the sitting drop vapor diffusion method at 20°C by mixing 1 µL of protein solution with 1 µL of the reservoir solution containing 25% PEGG3350 , 0 . 1M ammonium sulfate and 0 . 1 M Hepes ( pH = 7 . 5 ) .\",\n \"The compound 5a ( 0 . 2 μL of 10 μM in DMSO ) was added to the drops with apo crystals and incubated overnight .\",\n \"X-ray diffraction data for the CARM1-5a complex were collected at 100 K at beamline 24ID-E of Advanced Photon Source ( APS ) , Argonne National Laboratory .\",\n \"Data were processed using the HKL-3000 suite ( Minor et al . , 2006 ) .\",\n \"The structure was isomorphous to PDB entry 4IKP , which was used as a starting model .\",\n \"REFMAC ( Murshudov et al . , 1997 ) was used for structure refinement .\",\n \"Geometric restraints for compound refinement were prepared with GRADE v . 1 . 102 developed at Global Phasing Ltd . ( Cambridge , UK ) .\",\n \"The COOT graphics program ( Emsley et al . , 2010 ) was used for model building and visualization , and MOLPROBITY ( Williams et al . , 2018 ) was used for structure validation .\",\n \"To further confirm the electronic densities of the ligand , the total omission electron density map was calculated and contoured at 2 . 5 sigma ( Figure 3b ) ( Emsley and Cowtan , 2004; Murshudov et al . , 1997 ) .\",\n \"The ligands were docked into the binding site of CARM1 using the induced-fit docking ( IFD ) protocol ( Sherman et al . , 2006 ) implemented in the Schrodinger suite ( release 2016–4 ) .\",\n \"The poses for SNF and 2a were selected according to the IFD scores .\",\n \"Specifically , the results identified two distinct poses for 2a with similar scores but only one pose for SNF .\",\n \"The poses were then further relaxed by all-atom , explicit solvent molecular dynamics ( MD ) simulations .\",\n \"Herein , CARM1 models in complex with the two ligands were placed into explicit water boxes .\",\n \"A simple point charge ( SPC ) water model ( Berendsen et al . , 1981 ) was used to solvate the system , charges were neutralized , and 0 . 15 M NaCl was added .\",\n \"The total system size was ∼50 , 000 atoms .\",\n \"Desmond MD systems ( D . E . Shaw Research , New York , NY ) with OPLS3 force field ( Harder et al . , 2016 ) were used .\",\n \"The system was initially minimized and equilibrated with restraints on the ligand heavy atoms and protein backbone atoms , followed by production runs with all atoms unrestrained .\",\n \"The isothermal-isobaric ensemble was used with constant temperature ( 310 K ) maintained with Langevin dynamics and 1 atm constant pressure achieved using the hybrid Nose-Hoover Langevin piston method ( Feller et al . , 1995 ) on a flexible periodic cell .\",\n \"For each CARM1–ligand complex , a 600 ns trajectory was collected .\",\n \"MCF-7 and MDA-MB-231 ( parental and CARM1-KO ) cell lines ( Wang et al . , 2014a ) were used after their quality was confirmed with STR profile and standard mycoplasma contamination testing .\",\n \"Using the standard working curves generated above , the concentration ( CA ) of each analyte ( 6a , 5a , 2a , and SAM ) in the MeOH extraction of cell lysates was obtained through the ratio ( PA/PIS ) of the mass peak areas of each analyte ( PA ) versus each internal standard ( PIS ) , and the concentration of the internal standard ( CIS ) according to Equation\",\n \"2 . For 6a , 5a , and 2a under each assay condition , three similar CA values ( CA-6b , CA-5b , and CA-2b ) were obtained with 6b , 5b and 2b as the internal standards , respectively .\",\n \"An average concentration of the analyte ( C¯A ) was obtained on the basis of the three concentrations weighted by the mass peak areas of the three internal standards ( PIS , 6b , PIS , 5b and PIS , 2b ) according to Equation 3: ( 3 ) C¯A=CA−6b×PIS−6bPIS−6b+PIS−5b+PIS−2b+CA−5b×PIS−5bPIS−6b+PIS−5b+PIS−2b+CA−2b×PIS−2bPIS−6b+PIS−5b+PIS−2b On the basis of the C¯A values of 6a , 5a , and 2a ( C¯A , 6a , C¯A , 5a and C¯A , 2a ) in the MeOH extraction of cell lysates , the intracellular concentrations of the analyte 6a , 5a , and 2a ( Cintra , 6a Cintra , 5a and Cintra , 2a ) were calculated according to Equation 4 .\",\n \"Here C¯A , analyte is the weighted average of the three concentrations in the MeOH extraction , N is the cell number , and V is the mean volume of cells ( µL ) .\",\n \"The mean volume of MDA-MB-231 cells is 1 . 3 × 10−6 µL/cell as reported previously ( Coulter et al . , 2012 ) .\",\n \"The cell number ( N ) was determined with a hemocytometer .\",\n \"The concentration of SAM ( CA , SAM ) in the MeOH extraction was obtained according to Equation 2 , in which ‘X’ is the ratio ( PA , SAM/PIS , CD3-SAM ) of the mass peak areas of SAM ( PA , SAM ) versus the internal standard CD3-SAM ( PIS , CD3-SAM ) and ‘Y’ is the ratio ( CA , SAM/CIS , CD3-SAM ) of the concentrations of SAM versus CD3-SAM in the MeOH extraction .\",\n \"Given the identical LC-MS properties of SAM and CD3-SAM , CA , SAM was obtained solely on the basis of the working curve with CD3-SAM as the internal standard .\",\n \"The intracellular concentration of SAM ( Cintra , SAM ) was then calculated using Equation 4 , in which N is the cell number and V is the mean volume of MDA-MB-231 cells ( 1 . 3 × 10−6 µL/cell ) .\",\n \"( 4 ) Cinter , analyte=C¯A , analyte×40μLN×V Similar experimental procedures were carried out to quantify the intracellular concentrations of 6b , 5b , 2b and SAM except that 6b , 5b , and 2b were the analytes and their structurally related 6a , 5a and 2a were used as the internal standards with their CIS values of 0 . 125 µM , 0 . 125 µM , and 0 . 0125 µM , respectively .\",\n \"For each analyte ( 6b , 5b , and 2b ) , three working curves ( Figure 5c , e , Figure 5—figure supplement 2 ) were generated to cover the concentration range of 3 . 9 nM‒18 . 0 µM of the analyte ( CA , 6b , CA , 5b and CA , 2b ) with 6a , 5a and 2a ( CIS-6a , CIS-5a , and CIS-2a ) as the LC-MS/MS internal standards , respectively .\",\n \"In a manner similar to that described above for the ligand occupancy of CARM1 , the SMYD2 occupancy by 2a was modeled with Kd , SAM = 60 nM , Kd , 2a = 150 nM and Equation 6 .\",\n \"Here , the Kd , SAM value was reported previously upon developing the activity assay of SMYD2 ( Sweis et al . , 2015 ) .\",\n \"The Kd , 2a value was obtained according to IC50 = [SAM]×Kd/Kd , SAM+Kd , in which [SAM]=Kd , SAM = 60 nM for the IC50 assay and IC50 = 0 . 3 µM was obtained ( Supplementary file 1-Table A ) .\",\n \"The terms of SMYD2 occupancy by exogenous ligands were ignored given their low concentrations and low affinity relative to that of 2a ( Supplementary file 1-Table A ) .\",\n \"Potential in vivo clearance of 6a was evaluated in the presence of liver microsomes as previously reported ( Hansen et al . , 2012 ) .\",\n \"Briefly , 6a was dissolved into 50 mM potassium phosphate buffer ( pH 7 . 4 ) to a final concentration of 100 μM .\",\n \"To a 0 . 5 mL Eppendorf vial containing ice-cold potassium phosphate buffer ( pH 7 . 4 ) , we added 5 μL of the 100 μM stock solution of 6a , 3 μL of a freshly prepared NADPH solution ( 5 mM , Sigma-Aldrich , N7785 ) , and 1 . 33 μL of freshly thawed rat liver microsomes ( Sigma-Aldrich , M9066 ) to yield assay samples with a final volume of 50 μL .\",\n \"The resulting mixture was immediately vortexed and incubated in a 37°C shaker .\",\n \"At the time interval of 0 , 20 , 40 , and 60 min in triplicates , respective assay samples were quenched by adding 50 μL of ice-cold methanol , vigorously vortexed , and centrifuged with 5 , 000 g at 4°C for 20 min .\",\n \"From the methanol-quenched sample , 50 μL of supernatant was collected and mixed with 50 μL of 0 . 1% ( v/v ) TFA/MeOH solution containing 0 . 125 µM 6b , 0 . 125 µM 5b , and 16 µM 2b as the internal standards .\",\n \"The resulting mixture was transferred into a 96-well plate ( Agilent , 5042–1386 ) and stored at −20°C for LC-MS/MS analysis .\",\n \"The microsomal stability of 6a was evaluated via LC-MS/MS quantification of the residual 6a and of two potential microsome-processed products , 5a and 2a , upon quantification of intracellular concentrations of 6a , 5a and 2a ( as described above ) .\",\n \"The amount of 6a was plotted as the percentage of the residual concentration at each time point against its initial concentration at 0 min with GraphPad Prism Software .\",\n \"Potential production of 5a and 2a from 6a was examined in parallel .\",\n \"Interestingly , despite robust accumulation of 5a , no 2a was identified with the detection threshold of 0 . 02 µM .\",\n \"For negative controls , 3 μL of 50 mM potassium phosphate buffer ( pH 7 . 4 ) replaced the freshly prepared NADPH solution; two assay samples at the time interval of 0 and 60 min were collected .\",\n \"Under the current microsomal condition , the concentration of 6a remained the same ( 10 ± 0 . 7 μM ) and there was no production of 5a or 2a .\",\n \"MDA-MB-231 cells and MCF-7 parental and CARM1-KO ( Wang et al . , 2014a ) cells were maintained in DMEM ( Gibco , Gaithersburg , MD ) medium containing 10% FBS ( Gibco ) .\",\n \"These cells were then treated with compounds or with DMSO for 48 hr .\",\n \"The resultant cells were washed twice in phosphate-buffered saline ( PBS ) and the samples were sonicated in ice-cold RIPA buffer ( Thermo , Waltham , MA ) .\",\n \"The lysates were centrifuged ( 15 , 000 g ) at 4°C for 15 min .\",\n \"The supernatants were kept with the total protein amount determined with Bradford protein assay ( BioRad , Hercules , CA ) .\",\n \"After quantification , 50 μg of protein from each sample was loaded onto 6% SDS-PAGE and transferred onto a nitrocellulose membrane ( PALL , Port Washington , NY ) .\",\n \"For the Arg1064 methylation mark of BAF155 , the blots were blocked in 5% non-fat milk for 1 hr and incubated with anti-me-BAF155 ( Cancer Cell , 2014 ) , anti-BAF155 ( 1:1000; Santa Cruz Biotechnology , Dallas , TX ) , anti-CARM1 ( 1:1000; Genemed Synthesis , San Antonio , TX ) , and anti-β-actin ( 1:20000; Sigma-Aldrich , St . Louis , MO ) overnight at 4°C .\",\n \"After washing three times in Tris-buffered saline with Tween 20 ( TBST ) , the blots were incubated with HRP-conjugated secondary antibody ( 1:3000; Jackson ImmunoResearch , West Grove , PA ) .\",\n \"After washing blots with TBST , the membranes were developed using SuperSignal West Pico ECL solution ( Thermo ) .\",\n \"For the Arg455/Arg460 methylation mark of PABP1 , anti-me-PABP1 ( 1:1000 ) and anti-PABP1 ( 1:1000 ) antibodies ( Genemed Synthesis , San Antonio , TX ) were used instead .\",\n \"Here the antibodies against CARM1 , PABP1 , and me-PABP1 were custom generated by Genemed Synthesis ( San Antonio , TX ) ( Zeng et al . , 2013 ) .\",\n \"The density of the protein bands was quantified using ImageJ software ( NIH , Bethesda , MD ) .\",\n \"The EC50 values were obtained by fitting the methylation percentage ( % ) of BAF155 or PABP1 against the concentrations of the inhibitor using a sigmoidal equation with GraphPad Prism Software .\",\n \"CETSA was performed as described previously ( Jafari et al . , 2014 ) to examine the intracellular engagement of 6a or 6b with CARM1 .\",\n \"Briefly , 2 . 0 × 106 MDA-MB-231 cells were incubated with 15 μM 6a or 6b and DMSO for 48 hr .\",\n \"The harvested cells were re-suspended in PBS buffer and divided into eight aliquots ( 30 µL/aliquot ) , and then heat-shocked at various temperatures ( 49 . 1 , 54 . 6 , 57 . 0 , 59 . 5 , 61 . 8 , 63 . 9 , 65 . 6 , and 67 . 0°C ) for 3 min with a Bio-Rad CFX96 Real-Time PCR Detection Instrument ( using a temperature gradient of 49‒67°C ) .\",\n \"The heat-shocked cells were then lysed with the freeze-thaw method with a liquid nitrogen bath followed by a 25°C water bath for five cycles .\",\n \"The cell lysate was centrifuged at 4°C at 18 , 000 g for 20 min .\",\n \"The resultant supernatant containing the soluble protein fraction was collected and loaded onto an SDS-PAGE gel ( 20 µL ) .\",\n \"Western blotting of CARM1 was performed with anti-CARM1 antibody ( Cell Signaling Technology , C31G9 ) .\",\n \"For each sample , the band intensity of CARM1 was quantified with ImageJ .\",\n \"The band intensity of CARM1 was normalized to the band intensity at 49 . 1°C ( the lowest heat-shock temperature ) .\",\n \"Melting curves were obtained by plotting the normalized band intensity against the heat-shock temperatures and fit with a Boltzmann sigmoidal equation in GraphPad Prism .\",\n \"The melting temperatures ( Tm ) of 6a , 6b , or DMSO were obtained as the heat-shock temperatures that correspond to the 50% normalized band intensity in the fitted sigmoidal curves .\",\n \"MDA-MB-231 parental and CARM1-KO cells ( Wang et al . , 2014a ) were maintained in DMEM ( Gibco , Gaithersburg , MD ) medium containing 10% FBS ( Gibco ) .\",\n \"Cell invasion assays were performed using 8 . 0 µm pore size Transwell inserts ( Greiner Bio-One , Kremsmünster , Austria ) .\",\n \"MDA-MB-231 parental and CARM1-KO ( Cancer Cell , 2014 ) cells were harvested with trypsin/EDTA and washed twice with serum-free DMEM ( Gibco ) .\",\n \"2 × 105 cells in 0 . 2 mL serum-free DMEM ( Gibco ) were seeded onto the upper chamber , which was pre-coated with a thin layer of 40 uL of 2 mg/mL Matrigel ( Corning , NY , USA ) for 2 hr incubation at 37°C .\",\n \"To the lower chamber , we added 0 . 6 mL DMEM containing 10% FBS ( GIBCO ) and compounds or DMSO .\",\n \"After 16 hr in the 37°C incubator , the cells on the inner side of the upper chamber together with the Matrigel layer were removed using cotton tips .\",\n \"The residual invasive cells in the outer side of the upper chamber were fixed in 3 . 7% formaldehyde ( a weight percentage ) at an ambient temperature ( 22°C ) for 2 min and 100% methanol for 20 min , and then stained for 15 min with a solution containing 1% crystal violet and 2% ethanol in 100 mM borate buffer ( pH 9 . 0 ) .\",\n \"The number of invasive cells was counted under a microscope by taking five independent fields .\",\n \"Relative cell invasion was determined by the number of the invasive cells normalized to the total number of cells adhering to 0 . 8 µm transwell filters .\",\n \"The EC50 was obtained with GraphPad Prism Software upon fitting Equation 8 , in which ‘%Inhibition’ is the percentage of the inhibition of invasiveness , ‘Maximal Inhibition%’ is the percentage of the maximal inhibition of invasiveness , and [Inhibitor] is the concentration of the inhibitor .\",\n \"( 8 ) Inhibition%=MaximalInhibition%×[Inhibitor][Inhibitor]+EC50 To examine proliferation of MDA-MB-231 cells , 5000 cells of parental and CARM1-KO cells were seeded in a 96-well plate and incubated in 37°C overnight .\",\n \"These cells were treated with various doses ( 0 . 0001‒10 μM ) of SKI-73 or SKI-73N ( 6a or 6b ) in DMSO and incubated for 72 hr .\",\n \"An MTT assay was then performed to examine viability with DMSO-treated cells as the control .\",\n \"The relative viability of compound-treated cells versus DMSO-treated parent cells were plotted against the concentrations of SKI-73 and SKI-73N ( 6a and 6b ) .\",\n \"Cell preparation for scRNA-seq .\",\n \"10 × Genomics droplet-based scRNA-seq was implemented to characterize five types of MDA-MB-231 cells: MDA-MB-231 cells treated with DMSO , SKI-73 ( 6a ) and SKI-73N ( 6b ) , MDA-MB-231 cells that freshly invaded through Matrigel , and CARM1-KO MDA-MB-231 cells .\",\n \"For the first three samples , MDA-MB-231 cells were treated with DMSO , 10 μM SKI-73 ( 6a ) , and 10 μM SKI-73N ( 6b ) for 48 hr .\",\n \"The resulting cells were trypsinized at 37°C for 3 min .\",\n \"The harvested cells were washed twice with 1 × PBS ( phosphate-buffered saline ) containing 0 . 04% ( volume% ) bovine serum albumin ( BSA ) , gently dispersed to dissociate cells , and then filtered through a cell strainer ( 100 μm Nylon mesh , Fisherbrand ) to obtain single-cell suspensions for scRNA-seq .\",\n \"For the CARM1-KO sample , the cells were trypsinized at 37°C for 3 min , washed twice with 1 × PBS containing 0 . 04% BSA , and filtered through a 100 μm Nylon-mesh cell strainer ( Fisherbrand ) to obtain single-cell suspensions for scRNA-seq .\",\n \"To collect the cells that freshly invaded through Matrigel , the conditions described above for cell invasion assays were applied to allow approximately 5% of the 1 × 107 seeded MDA-MB-231 cells to invade through Matrigel .\",\n \"Briefly , ten Transwell inserts with a diameter of 24 mm and a pore size of 8 . 0 μm were pre-coated with a thin layer of 54 μL of 2 mg/mL Matrigel ( Corning ) .\",\n \"1 × 106 MDA-MB-231 cells in 1 . 0 mL serum-free DMEM ( Gibco ) were seeded into the upper chamber of each Matrigel-coated Transwell insert .\",\n \"2 . 0 mL DMEM containing 10% FBS ( Gibco ) was added into the lower chambers .\",\n \"After 16 hr incubation , the cells on the inner side of the upper chamber together with the Matrigel layer were removed using cotton tips .\",\n \"The cells that freshly invaded–those attached on the outer side of the upper chamber–were subjected to 3 min trypsin digestion at 37°C for detachment .\",\n \"The resulting cells were washed twice with 1 × PBS containing 0 . 04% BSA , gently dispersed to dissociate cells , and filtered through a 100 μm Nylon-mesh cell strainer ( Fisherbrand ) to obtain single-cell suspensions for scRNA-seq .\",\n \"The scRNA-seq libraries were prepared following the user guide manual ( CG00052 Rev E ) provided by the 10 × Genomics and Chromium Single Cell 3' Reagent Kit ( v2 ) .\",\n \"Briefly , samples containing approximately 8700 cells ( 93–97% viability ) were encapsulated in microfluidic droplets at a dilution of 66–70 cells/µL , which resulted in 4369–5457 recovered single-cells per sample with a multiplet rate ~3 . 9% .\",\n \"The resultant emulsion droplets were then broken and barcoded cDNA was purified with DynaBeads , followed by 12-cycles of PCR-amplification: −98 °C for 180 s , 12× ( 98°C for 15 s , 67°C for 20 s , 72°C for 60 s ) , and 72°C for 60 s .\",\n \"The 50 ng of PCR-amplified barcoded-cDNA was fragmented with the reagents provided in the kit and purified by SPRI beads with an averaged fragment size of 600 bp .\",\n \"The DNA library was then ligated to the sequencing adapter followed by indexing PCR: −98 °C for 45 s; 12× ( 98°C for 20 s , 54°C for 30 s , 72°C for 20 s ) , and 72°C for 60 s .\",\n \"The resulting DNA library was double-size purified ( 0 . 6–0 . 8× ) with SPRI beads and sequenced on Illumina NovaSeq platform ( R1 – 26 cycles , i7 – eight cycles , R2 – 96 cycles ) resulting in 70–79 million reads per sample with average reads per single-cell being 8075–10 , 342 and average reads per transcript 1 . 11–1 . 15 .\",\n \"The fastq files containing the transcriptome and barcoding metadata were demultiplexed using the SEquence Quality Control ( SEQC ) pipeline ( http:github . com/ambrosejcarr/seqc . git ) resulting in around 8000 UMIs per one cell .\",\n \"The table of UMI counts was used as the input and Seurat package v . 2 . 3 . 4 ( Butler et al . , 2018 ) was applied for scRNA-seq analysis .\",\n \"Here , the raw UMI counts were normalized per cell by dividing the total number of UMIs in each individual cell , multiplying by a scale factor of 10 , 000 and transforming into natural logarithm values .\",\n \"Cells with 1000~5000 genes and <20% mitochondrial RNA transcripts were kept for further analysis .\",\n \"Dimensionality reduction was carried out by selecting a set of highly variable genes on the basis of the average expression and dispersion per gene .\",\n \"The set of genes was used for principle component analysis ( PCA ) .\",\n \"Top principle components were then chosen for cell clustering analysis and t-SNE projection .\",\n \"Regression was performed to remove cell–cell variation in gene expression driven by the UMI number , mitochondrial gene content and ribosomal gene content using ‘ScaleData’ function in Seurat package ( Nestorowa et al . , 2016 ) .\",\n \"Clusters of cells were identified by clustering algorithm based on a shared nearest neighbor ( SNN ) modularity optimization that was included in Seurat package v . 2 . 3 . 4 ( Butler et al . , 2018 ) .\",\n \"Fisher’s Exact Test was implemented to evaluate the agreement of the clusters with the three cell origins ( DMSO- , SKI-73 ( 6a ) - or SKI-73N ( 6b ) -treated cells ) using R package ‘fisher . test’: http://mathworld . wolfram . com/FishersExactTest . html ( Mehta and Patel , 1983 ) .\",\n \"Because of the significant computation cost of Fisher’s Exact Test , the cell population of each Fisher’s Exact Test was down-sampled to 150 cells and this process was repeated for 100 times to cover a majority of the cell population .\",\n \"The p-value of Fisher’s Exact Test was then computed by Monte Carlo simulation .\",\n \"Means and standard errors of p-values were calculated and reported as the outputs of Fisher’s Exact Test .\",\n \"Three algorithmic scoring systems were used over a range of the resolution parameter that sets the corresponding ‘granularity’ of clustering , with higher values indicating a greater number of clusters .\",\n \"Here , silhouette analysis was applied to determine the number of clusters in an unsupervised manner , without awareness of the three cell origins ( DMSO- , SKI-73 ( 6a ) - or SKI-73N ( 6b ) -treated cells ) .\",\n \"The entropy scoring and Fisher’s Exact Test were implemented to evaluate the biological meaning of the clustering , using the minimal cluster number to resolve the cells between the three treatment conditions maximally .\",\n \"Given the awareness of the three cell origins , the minimal number of clusters with the maximal resolution of cell origin guided by the entropy scoring and Fisher’s Exact Test is expected within the 1~3-fold range of the optimized number of clusters guided by Silhouette analysis .\",\n \"Fisher’s Exact Test was used as the primary scoring method to determine the efficiency of clustering , given its higher resolution .\",\n \"‘dj , i’ ( ‘dDMSO , i’ , ‘dSKI-73N , i’ and ‘dSKI-73 , i’ in Equation 9 are defined as the fraction of the cells with the ‘j’ origin ( j = 1 , 2 or three for the treatment with DMSO , SKI-73N/6b and SKI-73/6a , respectively ) within the ‘i’ subpopulation ( i = 0‒n for the clustered subpopulation ) .\",\n \"‘di , total’ is defined as the fraction of the cells of the ‘i’ subpopulation within the total cell population in each cell-cycle stage .\",\n \"‘d ( j , i ) , total =djj , i ×di , total’ represents the fraction of the cells with the ‘j’ origin ( DMSO , SKI-73/6a , and SKI-73N/6b ) and the ‘i’ subpopulation within the total cell population in each cell-cycle stage .\",\n \"For a specific subpopulation ‘i’ , there are three ‘d ( j , i ) , total’ values ( d ( DMSO , i ) , total , d ( SKI-73N , i ) , total and d ( SKI-73 , i ) , total ) for the treatments with DMSO , SKI-73N ( 6b ) and SKI-73 ( 6a ) , respectively .\",\n \"The alteration of subpopulations is defined as the ratios of d ( SKI-73N , i ) , total/d ( DMSO , i ) , total or d ( SKI-73 , i ) , total/d ( DMSO , i ) , total fall out of the range of 1 . 0 ± 0 . 2 .\",\n \"Population analysis was conducted by classifying the SKI-73N/SKI-73 ( 6a/6b ) -treated subpopulations into the following five categories: commonly resistant ( 0 . 8 < ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ and ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’<1 . 2 ) , commonly emerging ( ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ and ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’≥1 . 2 ) , commonly depleted ( ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ and ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’≤0 . 8; ∣‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’−‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’∣<0 . 15 ) , differentially emerging ( either ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ or ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’>1 . 2 ) , and differentially depleted ( ‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’ or ‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’<0 . 8; ∣‘d ( SKI-73N , i ) , total/d ( DMSO , i ) , total’−‘d ( SKI-73 , i ) , total/d ( DMSO , i ) , total’∣>0 . 15 ) .\",\n \"Although additional combinations of differentially altered subpopulations could be possible , only those defined above were found under our treatment conditions .\",\n \"In each cell cycle ( G0/G1 , S and G2/M ) of the cells treated with DMSO , SKI-73 ( 6a ) or SKI-73N ( 6b ) and ‘invasion cells’ , correlation analysis of subpopulations was conducted with ‘BuildClusterTree’ function in the Seurat package ( https://rdrr . io/cran/Seurat/man/BuildClusterTree . html ) ( Nestorowa et al . , 2016 ) .\",\n \"The phylogenetic trees were constructed by averaging gene expressions across all cells in each subpopulation and then calculating distance on the basis of expressions averaged between different subpopulations .\",\n \"Differentially expressed genes were identified by comparing two groups of cells using the Wilcox rank sumtest with the ‘FindMarkers’ function in the Seurat package ( Nestorowa et al . , 2016 ) .\",\n \"In particular , the ‘invasion cells’ and their most correlated clusters ( which were revealed in the correlation analysis ) were selected as the ‘high’ group; the remaining remotely related clusters were selected as the ‘low’ group .\",\n \"The differential expression analysis was then performed by comparing cells in the ‘high’ group with the cells in the ‘low’ group .\",\n \"For the G0/G1-phase cells , the ‘invasion cells’ and Subpopulation 6 , 7 , 8 , 9 , 14 were selected as the ‘high’ group; Subpopulation 0 , 1 , 2 , 3 , 4 , 5 , 10 , 11 , 12 , 13 , 15 , 16 , 17 , 18 , 19 , 20 were selected as the ‘low’ group .\",\n \"For the G2/M-phase cells , the ‘invasion cells’ and Subpopulations 1 , 2 were selected as the ‘high’ group; and Subpopulations 0 , 3 , 4 , 5 were selected as the ‘low’ group .\",\n \"For the S-phase cells , the ‘invasion cells’ and Subpopulations 0 , 3 were selected as the ‘high’ group; and Subpopulations 1 , 2 , 4 , 5 , 6 were selected as the ‘low’ group .\",\n \"Differentially expressed genes were ranked according to the average log2 fold change ‘avg_logFC’ and the adjusted p-values ‘p_val_adj’ with the Seurat package .\",\n \"Top upregulated and downregulated genes were chosen by setting a cutoff on their ‘avg_logFC’ values ( >0 . 25 or <−0 . 25 ) .\",\n \"Then , curated genes with potential functional relevance to cancer malignancy ( 30 upregulated and 10 downregulated genes ) were selected as representative genes for generating heat map plots using the ‘DoHeatmap’ function in the Seurat package .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["CARM1 is a cancer-relevant protein arginine methyltransferase that regulates many aspects of transcription .","Its pharmacological inhibition is a promising anti-cancer strategy .","Here SKI-73 ( 6a in this work ) is presented as a CARM1 chemical probe with pro-drug properties .","SKI-73 ( 6a ) can rapidly penetrate cell membranes and then be processed into active inhibitors , which are retained intracellularly with 10-fold enrichment for several days .","These compounds were characterized for their potency , selectivity , modes of action , and on-target engagement .","SKI-73 ( 6a ) recapitulates the effect of CARM1 knockout against breast cancer cell invasion .","Single-cell RNA-seq analysis revealed that the SKI-73 ( 6a ) -associated reduction of invasiveness acts by altering epigenetic plasticity and suppressing the invasion-prone subpopulation .","Interestingly , SKI-73 ( 6a ) and CARM1 knockout alter the epigenetic plasticity with remarkable difference , suggesting distinct modes of action for small-molecule and genetic perturbations .","We therefore discovered a CARM1-addiction mechanism of cancer metastasis and developed a chemical probe to target this process ."],"string":"[\n \"CARM1 is a cancer-relevant protein arginine methyltransferase that regulates many aspects of transcription .\",\n \"Its pharmacological inhibition is a promising anti-cancer strategy .\",\n \"Here SKI-73 ( 6a in this work ) is presented as a CARM1 chemical probe with pro-drug properties .\",\n \"SKI-73 ( 6a ) can rapidly penetrate cell membranes and then be processed into active inhibitors , which are retained intracellularly with 10-fold enrichment for several days .\",\n \"These compounds were characterized for their potency , selectivity , modes of action , and on-target engagement .\",\n \"SKI-73 ( 6a ) recapitulates the effect of CARM1 knockout against breast cancer cell invasion .\",\n \"Single-cell RNA-seq analysis revealed that the SKI-73 ( 6a ) -associated reduction of invasiveness acts by altering epigenetic plasticity and suppressing the invasion-prone subpopulation .\",\n \"Interestingly , SKI-73 ( 6a ) and CARM1 knockout alter the epigenetic plasticity with remarkable difference , suggesting distinct modes of action for small-molecule and genetic perturbations .\",\n \"We therefore discovered a CARM1-addiction mechanism of cancer metastasis and developed a chemical probe to target this process .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Drugs that are small molecules have the potential to block the individual proteins that drive the spread of cancer , but their design is a challenge .","This is because they need to get inside the cell and find their target without binding to other proteins on the way .","However , small molecule drugs often have an electric charge , which makes it hard for them to cross the cell membrane .","Additionally , most proteins are not completely unique , making it harder for the drugs to find the correct target .","CARM1 is a protein that plays a role in the spread of breast cancer cells , and scientists are currently looking for a small molecule that will inhibit its action .","The group of enzymes that CARM1 belongs to act by taking a small chemical group , called a methyl group , from a molecule called SAM , and transferring it to proteins that switch genes on and off .","In the case of CARM1 , this changes cell behavior by turning on genes involved in cell movement .","Genetically modifying cells so they will not produce any CARM1 stops the spread of breast cancer cells , but developing a drug with the same effects has proved difficult .","Existing drugs that can inhibit CARM1 in a test tube struggle to get inside cells and to distinguish between CARM1 and its related enzymes .","Now , Cai et al . have modified and tested a CARM1 inhibitor to address these problems , and find out how these small molecules work .","At its core , the inhibitor has a structure very similar to a SAM molecule , so it can fit into the SAM binding pocket of CARM1 and its related enzymes .","To stop the inhibitor from binding to other proteins , Cai et al . made small changes to its structure until it only interacted with CARM1 . Then , to get the inhibitor inside breast cancer cells , Cai et al . cloaked its charged area with a chemical shield , allowing it to cross the cell membrane .","Inside the cell , the chemical shield broke away , allowing the inhibitor to attach to CARM1 .","Analysis of cells showed that this inhibition only affected the cancer cells most likely to spread .","Blocking CARM1 switched off genes involved in cell movement and stopped cancer cells from travelling through 3D gels .","This work is a step towards making a drug that can block CARM1 in cancer cells , but there is still further work to be done .","The next stages will be to test whether the new inhibitor works in other types of cancer cells , in living animals , and in human patient samples ."],"string":"[\n \"Drugs that are small molecules have the potential to block the individual proteins that drive the spread of cancer , but their design is a challenge .\",\n \"This is because they need to get inside the cell and find their target without binding to other proteins on the way .\",\n \"However , small molecule drugs often have an electric charge , which makes it hard for them to cross the cell membrane .\",\n \"Additionally , most proteins are not completely unique , making it harder for the drugs to find the correct target .\",\n \"CARM1 is a protein that plays a role in the spread of breast cancer cells , and scientists are currently looking for a small molecule that will inhibit its action .\",\n \"The group of enzymes that CARM1 belongs to act by taking a small chemical group , called a methyl group , from a molecule called SAM , and transferring it to proteins that switch genes on and off .\",\n \"In the case of CARM1 , this changes cell behavior by turning on genes involved in cell movement .\",\n \"Genetically modifying cells so they will not produce any CARM1 stops the spread of breast cancer cells , but developing a drug with the same effects has proved difficult .\",\n \"Existing drugs that can inhibit CARM1 in a test tube struggle to get inside cells and to distinguish between CARM1 and its related enzymes .\",\n \"Now , Cai et al . have modified and tested a CARM1 inhibitor to address these problems , and find out how these small molecules work .\",\n \"At its core , the inhibitor has a structure very similar to a SAM molecule , so it can fit into the SAM binding pocket of CARM1 and its related enzymes .\",\n \"To stop the inhibitor from binding to other proteins , Cai et al . made small changes to its structure until it only interacted with CARM1 . Then , to get the inhibitor inside breast cancer cells , Cai et al . cloaked its charged area with a chemical shield , allowing it to cross the cell membrane .\",\n \"Inside the cell , the chemical shield broke away , allowing the inhibitor to attach to CARM1 .\",\n \"Analysis of cells showed that this inhibition only affected the cancer cells most likely to spread .\",\n \"Blocking CARM1 switched off genes involved in cell movement and stopped cancer cells from travelling through 3D gels .\",\n \"This work is a step towards making a drug that can block CARM1 in cancer cells , but there is still further work to be done .\",\n \"The next stages will be to test whether the new inhibitor works in other types of cancer cells , in living animals , and in human patient samples .\"\n]","isConversational":false},"year":{"kind":"string","value":"2019"}}},{"rowIdx":93,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["structural biology and molecular biophysics","microbiology and infectious disease"],"string":"[\n \"structural biology and molecular biophysics\",\n \"microbiology and infectious disease\"\n]","isConversational":false},"title":{"kind":"string","value":"Massive antibody discovery used to probe structure–function relationships of the essential outer membrane protein LptD"},"id":{"kind":"string","value":"elife-46258-v1"},"sections":{"kind":"list like","value":[["The outer membrane ( OM ) of Gram-negative bacteria is a permeability barrier to antibiotics and other cytotoxic agents , such as detergents ( Nikaido , 2003 ) .","A key feature of the OM is its distinctive asymmetrical bilayer populated with lipopolysaccharide ( LPS ) in the outer leaflet and phospholipids in the inner leaflet ( Funahara and Nikaido , 1980; Kamio and Nikaido , 1976 ) .","Lateral interactions between LPS molecules mediated by divalent cations on the surface and packing of hydrocarbon chains in the membrane impede the passage of both hydrophilic and large hydrophobic molecules ( Nikaido , 2003 ) .","LPS is composed of a conserved lipid A molecule , a core oligosaccharide , and a variable O-antigen glycan .","Lipid A is synthesized in the cytoplasm where the core oligosaccharide is attached to form core-LPS ( Whitfield and Trent , 2014 ) .","This molecule is flipped into the periplasmic space by the inner membrane ATPase MsbA ( Ho et al . , 2018; Mi et al . , 2017; Zhou et al . , 1998 ) .","The O-antigen is appended onto core-LPS in the periplasm and LPS is transported to the outer leaflet of the OM by the LPS transport ( Lpt ) system ( Abeyrathne et al . , 2005; Han et al . , 2012; Okuda et al . , 2016 ) .","In E . coli , all 7 components of the Lpt pathway , LptABCDEFG , are essential and shuttle LPS molecules from the IM across the periplasm to the OM ( Okuda et al . , 2016; Okuda et al . , 2012; Ruiz et al . , 2008; Wu et al . , 2006 ) .","LptD is a β-barrel outer membrane protein ( OMP ) responsible for the final unidirectional insertion of LPS into the OM outer leaflet ( Botos et al . , 2016; Gu et al . , 2015; Li et al . , 2015 ) .","The LptD β-barrel is comprised of 26 anti-parallel β-strands with a proposed lateral gate formed where the first and last β-strands interact ( Dong et al . , 2014; Qiao et al . , 2014 ) .","Like other Gram-negative β-barrel OMPs , LptD possess long extracellular , environmentally-exposed loops and short periplasmic loops connecting adjacent strands ( Franklin and Slusky , 2018; Rollauer et al . , 2015; Schulz , 2002 ) .","The roles of the extracellular loops ( ECLs ) in folding or activity of LptD have not been determined .","LptD forms an obligate interaction with the OM lipoprotein LptE ( Chimalakonda et al . , 2011; Freinkman et al . , 2011 ) .","An N-terminal jelly roll domain is located on the periplasmic side of the β-barrel below the lateral gate .","Structural , biochemical , and genetic analyses suggest that the hydrophobic lipid A portion of LPS is embedded in the hydrophobic jelly roll domain .","The presumed model for LPS transport across the OM posits that lipid A is inserted into the hydrophobic OM through a lateral gate of LptD ( Botos et al . , 2016; Gu et al . , 2015; Li et al . , 2015 ) .","By an unknown mechanism , the polar sugars of the LPS core oligosaccharide and O-antigen are presumably moved through the LptD β-barrel lumen and flipped to the exterior of the cell .","Due to its essential role in OM biogenesis , high conservation among Gram-negative bacteria , and exposure to the extracellular environment , LptD represents a potential new antibacterial target .","Genetic experiments show that crosslinking LPS to LptD in the jelly roll domain or the lateral gate , or introduction of disulfide bonds that disrupt the proposed LPS path through LptD , inhibit growth ( Gu et al . , 2015; Li et al . , 2015 ) .","Indeed , two antibacterial peptides targeting LptD have been described , both presumably targeting the periplasmic jelly roll domain ( Srinivas et al . , 2010; Vetterli et al . , 2018 ) .","Critical roles for ECLs in OMP function have been described ( Browning et al . , 2013; Rigel et al . , 2013 ) .","For example , removal of the LptD extracellular loop 4 ( L4 ) , imp4213 , is known to disrupt the OM barrier leading to sensitization to antibiotics ( Sampson et al . , 1989 ) .","However , the roles of the extracellular LptD loops remain largely underexplored due to a lack of robust tools to systematically assess structure-function relationships in the context of a native OM .","Monoclonal antibodies ( mAbs ) represent useful tools to probe protein structure and function because they typically have high target specificity and affinity .","Accordingly , here we pursued four diverse strategies to produce thousands of mAbs that bound to LptD .","The breadth and depth of mAb discovery that we report here is unusual , especially for an integral membrane protein , and also takes advantage of an efficient and novel targeted boost-and-sort α-OMP discovery approach recently reported for another OMP , BamA ( Vij et al . , 2018 ) .","Characterization of the α-LptD mAbs discovered herein allowed us to map the environmentally-exposed surfaces of LptD and to systematically explore the structure-function relationship of the large ECLs of this essential OMP within its native OM environment .","Through the construction and characterization of LptD loop deletion strains , we identified non-exposed loops that are critical for LptD function under defined conditions , which are all inaccessible to antibody interference .","Our data provide unique insights into the structure and function of LptD and suggest that non-antibody approaches for interfering with the function of this essential OMP in E . coli should be prioritized to develop novel antibiotics ."],["To probe the functional relevance of the ECLs of the essential E . coli OMP LptD , we set out to discover α-LptD antibodies to these surface-exposed structures .","We generated a diverse mAb library against E . coli LptD by running four independent discovery campaigns .","Specifically , we immunized ( 1 ) rats with cyclic and linear peptides derived from extracellular LptD loop sequences , ( 2 ) rats and ( 3 ) mice with purified LptDE protein in detergent ( n-octyl-β-D-glucopyranoside ) and non-detergent polymer ( amphipol A8-35 ) , and ( 4 ) rats with a targeted boost-and-sort strategy that uses whole bacterial cell immunizations followed by immune boosts with purified LptDE protein ( Figure 1A and Figure 1B ) .","Within each campaign , the adjuvants and boosting strategies were varied to increase potential antibody diversity ( see Materials and methods for details ) .","Cumulatively , the discovery campaigns generated more than 7000 hybridomas and more than 3 , 000 ELISA+ LptD-specific mAbs ( Figure 1C ) .","In total , 774 ELISA+ α-LptD mAbs from the peptide and protein rat immunizations , 952 ELISA+ α-LptD mAbs from the mouse immunizations , and 1494 α-LptD mAbs from the targeted boost-and-sort immunizations were purified ( Figure 1C ) .","To explore epitope coverage and accessibility of the α-LptD ELISA+ mAbs from three of the antibody discovery campaigns , we first measured their extracellular binding by a fluorescence-assisted cell sorting ( FACS ) -based assay to two strains of E . coli ( Figure 2A ) .","The first strain was a laboratory E . coli strain , K-12 , that produces an intact core-LPS but no O-antigen .","The second strain was an E . coli ΔwaaD mutant , which produces a truncated form of LPS on the cell surface ( Kneidinger et al . , 2002 ) .","This truncation of LPS increases the exposed epitopes accessible to mAbs ( Bentley and Klebba , 1988; Storek et al . , 2018 ) .","Depending on the immunization campaign , 9–30% of the ELISA+ α-LptD mAbs were FACS+ on E . coli ΔwaaD , while <1% of the same antibodies were FACS+ on E . coli K-12 ( Figure 2A ) .","The E . coli K-12 FACS+ α-LptD mAbs were a subset of all the FACS+ mAbs , binding equally well to E . coli K-12 and E . coli ΔwaaD cells ( Figure 2—figure supplement 1 ) , and were discovered in both mouse and rat immunization campaigns .","These results reinforce the finding that LPS presents a significant barrier to accessing surface epitopes on E . coli ( Bentley and Klebba , 1988; Storek et al . , 2018 ) and suggest that the ECLs of OMPs may be masked or intimately associated with the core sugars of the LPS .","The diversity of epitopes bound by α-LptD mAbs in our library was next assessed by monitoring binding to LptD from different species .","ELISAs were performed with 134 FACS+ and 233 FACS- α-LptD mAbs using purified LptDE protein from two closely related Enterobacteriaceae species , Klebsiella pneumoniae and Enterobacter cloacae , which share 82% and 84% overall protein identity and 71% and 76% identify between the ECLs only when compared with E . coli LptD , respectively ( Figure 2B and Figure 2—figure supplement 2A ) .","The majority of FACS+ ( 61% ) and FACS- ( 69% ) α-LptD mAbs bound only E . coli LptDE by ELISA ( Figure 2C ) .","This result was not surprising considering that all protein immunization campaigns utilized only purified E . coli LptDE protein .","Analysis of the surface-exposed portion of LptD predicted from experimental x-ray crystallographic structural models identified regions of >5 amino acids that were conserved among all three Enterobacteriaceae species , the size of a typical hot spot of residues that contribute to antibody binding ( Stave and Lindpaintner , 2013 ) ( Figure 2B and Figure 2—figure supplement 2B ) .","Indeed , we identified 10 ( 7% ) α-LptD mAbs that bound to LptDE from all three species and 42 that bound either E . coli plus K . pneumoniae ( 22% ) or E . coli plus E . cloacae ( 9% ) LptDE ( Figure 2C ) .","Thus , our immunization campaigns using E . coli LptD-derived peptides and E . coli LptDE protein yielded diverse antibodies that bound to both unique and conserved extracellular accessible ( i . e . , FACS+ ) and extracellular-inaccessible or periplasmic ( i . e . , FACS- ) epitopes .","The differences in cross-species reactivity of the α-LptD mAbs suggested that the FACS+ antibodies bound multiple surface-exposed LptD epitopes .","To determine if these antibodies bound unique extracellular epitopes , we tested α-LptD mAbs with the highest FACS+ signal for their abilities to compete with each other for binding to purified E . coli LptDE .","Every possible FACS+ pair of α-LptD mAbs was screened for the ability to either bind LptDE protein simultaneously ( i . e . , different epitopes ) or to compete for binding ( i . e . , same or nearby epitopes ) using a SPR-based binding assay ( Abdiche et al . , 2017 ) .","From these results , we constructed a surface epitope map for the FACS+ α-LptD mAbs ( Figure 2D ) .","We identified two major epitope bins and a possible third bin with limited overlap ( Figure 2D ) .","When the species cross-reactivity data were considered , most three-species cross-reactive α-LptD mAbs were tightly clustered , while those that bound only E . coli LptDE formed a distinct cluster ( Figure 2D ) .","There were some outliers to these trends with the dual-species binders and E . coli-only binders binning with the cross-reactive mAbs suggesting distinct , but overlapping epitopes .","Thus , the α-LptD mAbs we discovered covered multiple distinct and overlapping extracellular LptD epitopes .","In order to map the extracellular epitopes of the α-LptD mAbs in a native asymmetrical OM environment , we first generated and characterized a panel of lptD mutants each lacking one ECL .","This strain panel also allowed us to assess the essentiality of each ECL in strains with different OM compositions .","We constructed 13 lptD mutants , each lacking the region encoding one of the 13 ECLs ( Figure 3A ) .","These loop deletions were based on available x-ray crystal structures of LptD and were constructed as complete loop deletions without replacing the removed sequence .","In the case of loop 4 ( L4 ) , which encompassed the classical imp4213 allele ( Sampson et al . , 1989 ) and in the case of loop 8 , which is an extensive loop that interacts with LptE in the LptD lumen ( Freinkman et al . , 2011 ) , we only removed a portion of the loop .","Each lptD loop mutant was expressed in a lptD-conditional E . coli strain .","In the presence of arabinose , which induces a chromosomal wild-type lptD , all strains grew ( Figure 3B ) .","In the absence of arabinose , all lptD loop mutants supported growth in the K-12 strain with the exception of loop 2 ( L2 ) removal , which showed reduced growth ( Figure 3B ) .","In all cases when growth was observed , the levels of the LptD loop deletion protein produced were similar to the protein level of wild-type LptD ( Figure 3—figure supplement 1A ) .","Thus , LptD is highly tolerable to genetic manipulation of its ECLs .","When the lptD loop mutants were expressed in an lptD-conditional E . coli ΔwaaD strain , which produces LPS with a truncated core oligosaccharide , the observed loop requirements were different .","In the E . coli ΔwaaD background , LptD lacking either loop 2 or loop 10 were unable to support bacterial growth on solid growth agar ( Figure 3B ) .","The inability of LptD lacking L10 ( LptDΔL10 ) to support growth of the E . coli ΔwaaD strain suggested that this ECL is important for LptD folding or activity in this strain .","In contrast , the LptDΔL10 variant was produced and did support growth of E . coli K-12 ( Figure 3B and Figure 3—figure supplement 1A ) , indicating that LptDΔL10 was folded and sufficiently functional in the strain producing core-LPS but not the minimal LPS of E . coli ΔwaaD strain .","To determine if the other LptD loops had differential activities in these two strains , we monitored the growth of all of the LptD loop variants in both the E . coli K-12 and E . coli ΔwaaD strain backgrounds over time in liquid growth media .","In the presence of arabinose , which induces a chromosomal wild-type lptD , all strains had similar growth patterns ( Figure 3—figure supplement 1C ) .","In the absence of arabinose , the E . coli K-12 strain producing LptDΔL10 grew , but at a decreased rate , and the strain with LptDΔL2 exhibited a dramatically prolonged lag phase and severe growth phenotype ( Figure 3C and Supplementary file 1 ) .","In contrast , E . coli ΔwaaD expressing LptDΔL10 did not grow at all , but expression of LptDΔL2 supported growth in liquid media ( Figure 3D and Supplementary file 1 ) .","Thus , L2 and L10 are important for LptD folding , activity , interaction with the β-barrel assembly machinery component BamA , or a combination of these .","Also , the LptDΔL10 loop deletion mutant was better tolerated in E . coli K-12 compared to the LPS-truncated ΔwaaD strain while in liquid growth media the LptDΔL2 mutant is better tolerated in the E . coli ΔwaaD strain .","This result is consistent with the synthetic lethality of other mutations that affect OM biogenesis , specifically bam101 , ΔbamB , ΔsurA , and ΔlpxL , in LPS-truncated E . coli strain backgrounds ( Klein et al . , 2009; Storek et al . , 2019 ) highlighting the redundancy that exists to maintain this important barrier structure .","The lptDΔL4 mutant used in this study encompassed the imp4213 allele .","This particular loop mutant is viable but increases membrane permeability and sensitivity to OM-excluded antibiotics ( Sampson et al . , 1989 ) .","We tested whether the strains producing other LptD loop mutants had increased membrane permeability by measuring sensitivities to rifampicin and vancomycin , two antibiotics excluded by the Gram-negative bacteria OM .","The minimal inhibitory concentrations ( MICs ) of both rifampicin and vancomycin were lower for strains producing LptDΔL4 , as expected , and also for strains producing LptDΔL2 , LptDΔL8 , and LptDΔL10 indicating an OM defect ( Table 1 ) .","Increases in OM permeability were observed in both E . coli K-12 and E . coli ΔwaaD strain backgrounds .","These results suggest that extracellular L2 , L4 , L8 and L10 are important for LptD function and potentially influence LptD folding , interaction with BamA , or activity .","We utilized our lptD loop deletion panel to map the α-LptD mAb surface coverage in a native E . coli membrane environment .","We characterized binding of 52 FACS+ α-LptD mAbs to the E . coli ΔwaaD loop deletion panel .","In the absence of arabinose driving expression of the chromosomal wild-type lptD , production of each mutant LptD was detected and growth of the E . coli ΔwaaD conditional lptD strain was supported by all LptD loop deletion mutants with the exception of L10 , so this construct was not tested in this assay ( Figure 3B and Figure 3—figure supplement 1B ) .","The mAbs in the panel were confirmed to bind to an E . coli ΔwaaD strain expressing wild-type lptD ( Figure 4A ) .","We expected that if an extracellular LptD loop was critical for binding , removing the loop would reduce α-LptD mAb binding .","The lack of binding by any particular mAb to a LptD loop mutant could be due to removal of the mAb binding site , in part or in full , or a structural change that alters the mAb binding site when a loop is removed .","The majority of the α-LptD mAbs could be categorized into one of 7 general binding patterns based on the mutants they were unable to bind with multiple representatives in each grouping: ( 1 ) Loop 4 ( L4 ) , ( 2 ) Loop 6/7 ( L6/L7 ) , ( 3 ) Loop 6/7/8 ( L6/L7/L8 ) , ( 4 ) Loop 8/9 ( L8/L9 ) , ( 5 ) Loop 9 ( L9 ) , ( 6 ) Loop 11 ( L11 ) , and ( 7 ) Loop 13 ( L13 ) ( Figure 4B and Figure 4—figure supplement 1 ) .","Thus , our immunization campaigns produced antibodies that allow systematic probing of accessible ELCs around the LptD structure .","In some cases , for example , a representative E . coli K-12 FACS+ α-LptD mAb ( Figure 2—figure supplement 1B ) , no loop mutant dramatically altered antibody binding .","This pattern suggests a surface-exposed epitope was present that was not altered in these particular loop mutants or the mAb bound a composite epitope that was not sufficiently disrupted by loss of a portion of the epitope encompassed by one individual loop deletion ( Figure 4B ) .","This could be especially true for L8 as the deletion encompassed only part of the loop that , based on structural models , is likely to occupy the LptD lumen while additional exposed portions were left unchanged ( Figure 4C ) .","For every loop deletion , at least one FACS+ α-LptD mAb had reduced cell binding by >90% .","The notable exceptions were L1 , L2 , L5 and L12 , but mAbs could be found that exhibited >80% decreased binding for L1 , L5 and L12 .","Comparing these data with the species cross-reactivity and in vitro binding competition epitope map provided a complete picture of the mAb accessible surface of LptD .","Specifically , a majority of the two- and three-species cross-reactive α-LptD mAbs were dependent upon L8 , L9 , or both for binding ( Figure 2B , Figure 2D , Figure 2—figure supplement 2 , Figure 4B , and Figure 4—figure supplement 1 ) .","The cross-reactive mAbs 5A3 and 3D4 exhibited slightly different binding requirements , L6/L8/L9/L11 and L7/L8 , respectively , consistent with these mAbs recognizing different epitopes in our in vitro binning experiment ( Figure 2D ) .","Additionally , E . coli-specific α-LptD mAbs , but not the three-species cross-reactive mAbs , were defective for binding upon removal of L3 , L7 , or L13 ( Figure 2B , Figure 2D , Figure 2—figure supplement 2 , Figure 4B , and Figure 4—figure supplement 1 ) .","These α-LptD mAbs and knowledge of their binding sites are valuable tools for probing the accessible LptD surface and suggest that LptD can tolerate extensive alteration to its extracellular surface with little or no effect on folding or function .","Because distinct antibody discovery approaches were utilized , we were also able to evaluate the potential correlation of LptD targeting with regard to the immunization campaign .","Partitioning the mAbs with respect to the host animal used for immunization highlighted a trend for L6 , L7 , and L8 to be critical from the rat immunization campaign , while mAbs from the mouse immunization campaign tended to yield FACS+ antibodies that required L8 and L9 .","Because these animals are likely not tolerized to the foreign bacterial LptD protein , this difference could reflect differences in the types of binding sites presented for rat versus mouse antibodies .","Thus , even within these immunization campaigns , diverse binding profiles to the loop mutants were seen ( Figure 4B and Figure 4—figure supplement 1 ) .","A potential application of antibodies raised against extracellularly-exposed bacterial proteins is as direct-acting antibacterial molecules ( LaRocca et al . , 2009; Storek et al . , 2018 ) .","Disruption of L2 or L10 resulted in growth defects for E . coli ΔwaaD ( Figure 3B ) , however , we did not identify any α-LptD mAbs that bound L2 , and both L2 and L10 appear to be inaccessible in available structural models of LptD ( Figure 4C ) ( Botos et al . , 2016; Dong et al . , 2014; Qiao et al . , 2014 ) .","Indeed , when we screened our entire antibody catalog using growth inhibition as a readout for mAb activity , we did not identify any α-LptD mAbs able to inhibit growth of E . coli K-12 or E . coli ΔwaaD by >50% ( Table 2 and Figure 4—figure supplement 2 ) .","Combined with our LptD loop deletion analysis , the lack of ability to identify an inhibitory α-LptD mAb suggests that LptD may be specifically evolved to withstand extracellular modulation by antibodies while protecting critical structural elements of LptD function ."],["The OM is a critical feature of Gram-negative bacteria and presents a barrier to the discovery of new antibiotics ( Lewis , 2013 ) .","There are two essential OMPs embedded in the OM and involved in construction of the OM barrier .","BamA is part of the β-barrel assembly machine that folds and inserts β-barrel OMPs , including BamA and LptD , into the OM ( Konovalova et al . , 2017; Ricci and Silhavy , 2012 ) .","LptD is the last protein in the Lpt pathway that establishes and maintains OM asymmetry by inserting LPS exclusively into the outer leaflet of the OM ( Okuda et al . , 2016 ) .","Despite the importance of these β-barrel OMPs , understanding of how their structures contributes to their folding and function is still incomplete .","Similar to other Gram-negative β-barrel OMPs , both BamA and LptD possess short periplasmic loops but long , variable , and environmentally-exposed ECLs ( Franklin and Slusky , 2018; Rollauer et al . , 2015; Schulz , 2002 ) .","A deeper examination of the roles of these ECLs in a native OM context could contribute to a greater understanding of how they contribute to the structure and function of OMPs in general .","To systematically interrogate the LptD ECLs , we performed four antibody discovery campaigns to generate thousands of antibodies targeting LptD to probe structure-function relationships of this essential OMP in its native OM environment .","The unusual breadth and depth of our approach represents a powerful way to probe the accessible surfaces of the ECLs of LptD and to tease apart the changes that accompany LPS insertion by LptD in the native OM environment of the bacterial cell .","By characterizing hundreds of these α-LptD mAbs , we discovered that extensive portions of the extracellular surface of LptD are dispensable in E . coli and critical features of LptD are likely occluded and protected by non-essential loops ( Figure 5 ) .","Little is known about the role and dynamics of the ECLs of LptD in the function of LPS insertion into the OM and there are few available tools to study this process .","In an effort to probe as many extracellular epitopes as possible , we generated a large , diverse α-LptD mAb library by undertaking multiple immunization campaigns utilizing different antigens , adjuvants , and hosts .","Varying antigens systematically presented LptD in increasingly complex arrangements to potentially capture different available epitopes .","First , peptide immunizations utilized both linear peptides derived from conserved and non-conserved LptD ECL sequences and cyclic variants in an effort to constrain particular conformations .","Next , purified full-length LptDE immunizations aimed to capture a more native state of the ECLs .","Because the matrix in which the protein is reconstituted can alter the conformational flexibility and exposure of membrane-proximal epitopes , we utilized multiple reconstitution matrices ( i . e . , detergents and amphipols ) .","Finally , using an approach that we have recently leveraged to find rare , inhibitory α-BamA antibodies , a targeted boost-and-sort strategy , we immunized rats with sub-lethal E . coli infections and then boosted them with purified LptDE protein to enrich for any LptD antibody response ( Vij et al . , 2018 ) .","Taken together , our effort to maximize the epitope coverage in our α-LptD mAb library led to distinct clusters of antibodies that possessed differing abilities to cross-react with LptD species variants , required different ECLs for binding , and exhibited varying levels of access to LptD through the protective LPS layer .","Although the full breadth of antibody diversity in our library required the multiple immunization approaches described here , we propose a streamlined workflow for future OMP efforts .","An accessible approach is to immunize SD rats with recombinant purified OMP either with or without whole bacterial cells ( strategies 3 and 4 , Figure 1 ) .","Both of these workflows returned the highest percentages of FACS+ antibodies coupled with recognition of diverse epitopes ( Figure 2 , Figure 4B , and Figure 4—figure supplement 1 ) .","Given the ease of manipulating bacterial strains , it could even be possible to steer immunizations towards particular ECLs by initially immunizing with a ΔECL-OMP strain followed by boosting with recombinant protein .","Other scaffolds or antibodies from different host species might be able to recognize distinct epitopes on the exposed portions of LptD .","For example , recognition of partially buried epitopes could be achieved using species , such as rabbits , chickens , llamas , or camels , that contain longer CDRH3s that are better suited to reach into cavities on a particular antigen ( Lavinder et al . , 2014; Muyldermans and Smider , 2016 ) .","In the case of the mouse and rat antibodies described here , we found that seven different LptD ECLs , out of a total of 13 , could affect binding of the mAbs and the remaining six loops are likely partially or fully buried based on models from crystal structures ( Botos et al . , 2016; Dong et al . , 2014; Qiao et al . , 2014 ) ( Figure 5 ) .","When this α-LptD mAb campaign is combined with of our α-BamA mAb effort ( Storek et al . , 2018; Vij et al . , 2018 ) , we now have the ability to probe the structure-function relationships of the only two essential OMPs in their native OM environment in E . coli at a level not previously possible .","Importantly , we note the antibody discovery methods that we describe herein are efficient , robust , and a streamlined approach can be readily applied within a standard laboratory setting .","We have observed that in LptD there appear to be differential requirements for particular ECLs , L2 and L10 , depending on the structure of LPS , specifically wild-type LPS versus the truncated ΔwaaD LPS .","Notably , BamA appears to be defective in the highly fluid OM of an LPS-truncated strain grown under low salt and high temperature conditions ( Storek et al . , 2018 ) .","This is manifest by an increased sensitivity to a recently discovered inhibitory α-BamA antibody MAB1 and a requirement for elevated BamA levels .","In these cases , the defects can be suppressed by altering the conditions to increase membrane rigidity ( Storek et al . , 2019 ) .","The LPS-dependent requirement for LptD loops suggests that effects of the OM structure on OMPs might be more general .","Among the 13 extracellular LptD loops , both L2 and L10 are short and lack large insertions or deletions across LptD homologs when compared to the longer loops in highly divergent LptD proteins ( Figure 5—figure supplement 1A ) .","One possible reason for this is that these loops are uniquely positioned to make interactions with the LPS in the OM to allow for proper folding or functioning of this essential OMP .","Although these interpretations are speculative and complicated by the fact that LPS is the substrate for LptD , it is tempting to hypothesize that the structure and function of OMPs evolved to be maximal when embedded in the constrained , liquid crystalline , OM matrix ( Paracini et al . , 2018 ) and altering the nature of this membrane leads to defects .","Thus , membrane environment is a critical consideration when interpreting biochemical and structural data of LptD and other OMPs in detergent micelles and must be considered when attempting to design or discover inhibitors of these essential processes in Gram-negative bacteria .","Genetic , biochemical , and structural data support a model for LPS insertion into the OM whereby the acyl chains of lipid A are inserted into the hydrophobic bilayer through a lateral gate created between the first and last strands of the LptD β-barrel ( Botos et al . , 2016; Dong et al . , 2014; Qiao et al . , 2014 ) .","It is less clear how the polar sugars of the core oligosaccharide and O-antigen emerge from the periplasm to the cell exterior as it would be unfavorable for them to pass directly through the membrane .","Presumably the sugars must pass through the LptD transmembrane β-barrel lumen , which is also occupied by an essential lipoprotein , LptE , and then through a proposed opening in the surface-exposed cap of LptD .","Thus , consistent with prior molecular dynamic simulations , there are likely major structural surface re-arrangements in LptD during LPS insertion ( Dong et al . , 2014 ) .","Given this model , we hypothesized that an antibody could have interfered with this process in a number of ways .","For example , a mAb could staple the lateral gate in either a closed or open conformation , prevent the movements associated with expelling the sugars to the outside , or allosterically alter the protein so that LPS movement is thwarted .","It is informative , therefore , that none of our identified α-LptD mAbs were able to inhibit the activity of LptD .","One explanation is that binding to a single epitope is insufficient to disrupt the required movements of LptD .","Alternatively , the positions on LptD critical for these presumed structural changes might be inaccessible to mAbs .","Indeed , few or no α-LptD mAbs bound to the extracellular LptD loops that were found genetically to be important for viability or OM maintenance under specific conditions .","Consistent with this possibility , in crystal structures of LptDE , the two loops we found most critical were occluded by other parts of the protein: L2 by L3 , and L10 by L9 and L11 ( Dong et al . , 2014; Qiao et al . , 2014 ) .","By burying critical regions , LptD appears to avoid direct interference by antibodies , providing bacteria an evolutionary advantage in evading functional antagonism by the host immune system .","Accordingly , the antibody-accessible LptD loops also are the most diverse in sequence , suggesting that these regions may be subject to selective pressures from the immune system ( Botos et al . , 2016 ) ( Figure 5—figure supplement 1B ) .","Moreover , the proximity of a particular epitope to the membrane can sterically or electrostatically block binding .","However , in rare cases , recognition can be achieved via stabilizing hydrophobic residues in an antibody CDR that transiently interact with the membrane , as observed for several α-HIV or α-GPCR antibodies ( Ishchenko et al . , 2017; Scherer et al . , 2010 ) .","It remains to be seen whether immobilizing these loops in LptD , for example , with a small molecule or peptide macrocycle that can access these positions , would inhibit LptD activity .","While our observations make direct interference with LptD function by antibodies unlikely , they do not rule out the potential value of LptD as a vaccine candidate ( Zha et al . , 2016 ) .","Overall , our study represents an exhaustive α-LptD antibody discovery campaign , and potentially the most exhaustive on any membrane protein described to date .","For the first time , we have been able to map essential and non-essential structural features of this intriguing and important potential antibiotic target .","We have observed extreme tolerability of the ECLs to interference , suggesting that the architecture of the exposed ECLs of LptD play a role in protecting critical functions of this essential OMP .","More generally , we provide a robust template for future efforts to dissect the structure-function relationships of complex membrane proteins from bacteria and mammals in their native environments , and a method to rapidly assess the therapeutic potential of antibody targeting ."],["Luria-Bertani broth ( ThermoFisher Scientific 12795027 ) and Mueller Hinton II cation-adjusted broth ( BBL 212322 ) was prepared according to manufacturer's instructions .","Bacterial cultures were grown at 37°C .","When appropriate , media was supplemented with kanamycin ( 50 μg/mL ) , carbenicillin ( 50 μg/mL ) , chloramphenicol ( 12 . 5 μg/mL ) , hygromycin ( 200 μg/mL ) , gentamicin ( 10 μg/mL ) and arabinose ( 0 . 2% vol/vol ) .","Bacterial strains and relevant primers are listed in Supplementary file","2 . Kanamycin deletion-insertion mutations of waaD was obtained from the Keio collection ( Baba et al . , 2006 ) .","Briefly , pKD4 or pKD3 was amplified with primers containing 50 bp nucleotide homology extensions ( Supplementary file 2 ) to the gene of interest .","The linear product was transformed into the appropriate background strain containing pSIM18 ( Chan et al . , 2007 ) recovered for 4 hr and selected on media containing 50 μg/mL kanamycin or 12 . 5 μg/mL chloramphenicol as appropriate .","Mutations were confirmed by PCR .","To make sequential mutants , the antibiotic marker was flipped out using pCP20 ( Datsenko and Wanner , 2000 ) .","The conditional lptD strain , ΔlptD::PBAD-lptD , was created by inserting PBAD-lptD at the attB site in MG1655 followed by deletion of the native copy of lptD ( Datsenko and Wanner , 2000; Diederich et al . , 1992 ) .","Briefly , lptD was cloned into pBAD24 using standard methods .","PBAD-lptD was amplified from pBAD24-lptD and sub-cloned into pLDR9 .","pLDR9-PBAD-lptD was digested with XbaI , ligated , and transformed into MG1655 expressing pLDR8 .","PCR and DNA sequencing confirmed insertion of PBAD-lptD at the attB site .","After integration of PBAD-lptD , the native copy of lptD was deleted using λ Red recombination as described above .","In the absence of arabinose , the conditional DlptD::PBAD-lptD strain did not grow .","To obtain a conditional lptD strain truncated for LPS , waaD was deleted using λ Red recombination .","To clone the loop mutants into the lptD conditional strains , pLMG18 was modified to replace the original antibiotic cassette providing chloramphenicol resistance with a gentamicin resistance cassette using Gibson Assembly .","The lptD gene and mutant constructs were codon optimized and ordered as gBlocks ( Integrated DNA technologies ) .","The genes were amplified and cloned into pLMG18gm using Gibson Assembly .","Clones were confirmed by sequencing .","The E . coli LptDE expression plasmid was constructed by cloning E . coli LptD ( amino acids 1–784 ) and E . coli LptE ( amino acids 1–193 ) into pCDFDuet1 vector at the first and second multiple cloning sites .","An AviTag followed by an 8x histidine tag was inserted into N-terminal of E . coli LptD .","The expression plasmid was expressed in E . coli C41 ( DE3 ) .","When the OD600 reached 0 . 9 , cells were induced with 0 . 5 mM isopropyl-β-D-1 thiogalactopyranoside ( IPTG ) overnight at 18°C .","Cells were harvested by centrifugation at 4500 rpm for 15 min and the cell pellets were resuspended in lysis buffer ( 50 mM Tris , pH 8 . 0 , 200 mM NaCl , 1x complete protease inhibitor mixture ( Roche ) ) .","The resuspended cells were disrupted by passing through a microfluidizer at 10 , 000 psi three times .","The cell lysate was added with 2% Zwittergent 3–14 and rocked overnight at 4°C .","The suspension was ultracentrifuged ( 45Ti rotor , 40 , 000 rpm ) for 1 hr at 4°C .","The supernatant was incubated with pre-equilibrated Ni-NTA resins ( Qiagen ) and rocked for 1 hr at 4°C .","The resins were washed by the buffer containing 50 mM Tris , pH 8 . 0 , 200 mM NaCl , 1% Zwittergent 3–14 , 15 mM imidazole , 1x complete protease inhibitor mixture and eluted with the buffer consisting of 50 mM Tris , pH 8 . 0 , 200 mM NaCl , 1% Zwittergent 3–14 , 300 mM imidazole and 1x complete protease inhibitor mixture .","The eluate was concentrated and purified by a Superdex 200 16/60 column ( GE Healthcare ) using 20 mM HEPES , pH 8 . 0 , 100 mM NaCl , 1 . 5% n-Octyl-β-D-Glucopyranoside ( OG; Anatrace ) , 1x complete protease inhibitor mixture as the running buffer .","The obtained LptD/E protein was diluted 4-fold into the buffer of 20 mM HEPES , pH 8 . 0 , 1 . 5% OG and further polished using Q HP anion exchange 5 ml column ( GE Healthcare ) .","20 mM HEPES , pH 8 . 0 , 25 mM NaCl , 1 . 5% OG was used as the start buffer and elution was achieved with linear gradient of 0 . 025–1 M NaCl .","When protein was required in amphipol , LptD/E protein was incubated with biotinylated amphipol ( Anatrace ) overnight at 4°C and then purified over a Superdex 200 16/60 column in a buffer consisting of 20 mM HEPES , pH 8 . 0 , 100 mM NaCl .","Peptide pools are listed in Supplementary file","3 . Peptides were designed based on sequence conservation , surface exposure and loop location .","When LptDE protein was required for ELISA or antigen-based sorting , an in vitro biotinylation reaction using BirA enzyme targeting the N-terminal Avi tag was first carried out according to the manufacturer's suggestions ( Avidity ) ; LptDE was then rerun over the Superdex 200 ( 26/60 ) in buffer E as described above in order to remove free biotin and other reaction components .","When protein was required in amphipol ( either non-biotinylated or biotinylated ) , LptDE at 1 mg/mL in buffer E was incubated with a stock solution of 2 mg/mL A8-35 amphipol ( Anatrace ) at 22°C for 1 hr and then applied over a Superdex 200 ( 26/60 ) column in buffer F ( 50 mM Tris pH 8 , 100 mM NaCl ) .","Protein reconstituted in A8-35 amphipol for immunization was concentrated to 1 mg/mL using a centrifugal device ( 10 K MWKO; Millipore ) .","When PE-labeled protein was required for antigen-specific sorting , LptDE at 1 mg/mL ( biotinyated and reconstituted in A8-35 amphipol in buffer F ) was mixed 1:1 ( vol/vol ) with PE-streptavidin ( Jackson ImmunoResearch ) reconstituted in buffer F; the PE-streptavidin-biotin-LptDE-amphipol complex was incubated at 22°C for at least 15 min prior to use .","All animal study designs for the mouse and rat immunizations were reviewed and approved by the Genentech Institutional Animal Care and Use Committee prior to the start of this work .","All animal work was performed in accordance with relevant guidelines and regulations .","LptD peptide immunization campaign immunized Balb/c mice ( Charles River Laboratories , Hollister , CA ) with cyclic and linear LptD peptide pools ( Supplementary file 3 ) .","The initial immunization contained 100 μl Complete Freund’s Adjuvant ( CFA ) and subsequent dosing included 100 μl Incomplete Freund’s Adjuvant ( IFA ) ( Sigma ) .","pAbs were purified by Protein A and assayed by ELISA , FACS , and inhibition of bacterial growth as described below .","Hybridoma fusions were performed as previously described and supernatants were screened for protein binding by ELISA ( Hazen et al . , 2014 ) .","All ELISA positive clones were purified and screened for inhibition of bacterial growth .","LptDE protein immunization campaign immunized Balb/c mice ( Charles River Laboratories , Hollister , CA ) with detergent-solubilized or amphipol-reconstituted E . coli LptDE protein and either CFA or Ribi adjuvants ( Sigma ) .","pAbs were purified and screened as described above .","LptDE protein immunization campaign immunized Sprague Dawley rats ( Charles River Laboratories , Hollister , CA ) with detergent-solubilized or amphipol-reconstituted E . coli LptDE protein and either CFA or Ribi adjuvants ( Sigma ) .","pAbs were purified and screened as described above .","Bacterial immunization campaign immunized Sprague Dawley rats ( Charles River Laboratories , Hollister , CA ) with E . coli K-12 cells in PBS ( 107–109 colony forming units via intravenous injection ) .","The rats were boosted four times with E . coli K-12 cells followed by two protein boosts with 10 μg E . coli LptDE protein in amphipol .","To generate monoclonal antibodies , hybridoma fusions were performed as previously described except with a myeloma partner SP2ab that enables surface display of IgG cell ( Hazen et al . , 2014; Price et al . , 2009 ) .","After HAT selection in ClonaCell-HY Medium C ( StemCell Technologies ) for 4 days , hybridomas were stained with a cocktail of FITC-conjugated anti-rat IgG1/IgG2a/IgG2b mAbs ( 1:100 dilution; Bethyl Laboratories ) and PE-conjugated LptDE protein antigen ( 5 μg/μl ) .","Samples were sorted on a FACSAria II cell sorter ( BD Biosciences ) .","Gating strategy to identify hybridoma cells was first based on size ( FSC/SSC ) .","After dead-cell exclusion with Propidium iodide ( Sigma-Aldrich P-4864 ) , IgG +LptDE + single cell hybridomas were sorted into 96-well tissue-culture plates containing 200 μl of ClonaCell-HY Medium E ( StemCell Technologies ) .","Profiles were analyzed by FlowJo v . 9 . 7 . 7 software .","Sorted cells were cultured for seven days .","Hybridomas were cultured using a previously described semi-automated high throughput process for hybridoma culturing and antibody purification ( Vij et al . , 2018 ) .","Epitope bins were determined by 96 × 96 array-based SPR imaging system ( Carterra , USA ) using classical sandwich method .","Purified antibodies were diluted at 20 μg/ml in 10 mM sodium acetate buffer pH 4 . 5 .","Using amine coupling , antibodies were directly immobilized onto a SPR sensorprism CMD 200 M chip ( XanTec Bioanalytics , Germany ) using a Continuous Flow Microspotter ( Carterra , USA ) to create an array of 96 antibodies .","For antibody binning , printed chip was loaded on IBIS MX96 SPRi ( Carterra USA ) and E . coli LptDE protein , diluted to 50 nM in 1 . 5% bOG HBS-P buffer , was injected over the chip at 25°C followed by a second injection of purified antibody , diluted at 20 μg/ml in 1 . 5% bOG HBS-P buffer , to make a sandwich .","The epitope binning data were processed using Carterra binning software tool .","Epitope binning experiments were performed as part of our high throughput antibody screening workflow and are purely descriptive .","The test strain was grown to log phase in MHB II supplemented with 0 . 002% Tween-80 , diluted to a final OD600 0 . 01 in sterile round-bottom 96-well plates ( Costar ) .","Antibodies were added at 10 μg/mL , incubated statically at 37°C and monitored for bacterial growth after 4 hr of static incubation .","A plate reader at OD600 measured optical density of bacterial growth after shaking the plate for 25 s .","Bacterial inhibition was calculated by subtracting the OD600 value from the media control , followed by dividing that value by the no antibody control OD600 value .","Antibody activity experiments were performed as part of our high throughput antibody screening workflow and are purely descriptive .","Antibodies were screened by capture ELISA .","Briefly , 50 µl of biotinylated LptDE protein , diluted in assay buffer ( PBS + 1 . 5% OG +0 . 5% BSA ) was added to streptavidin coated 384 well plates ( Thermo Scientific USA , Cat # 15504 ) and incubated at RT for 1 hr , while shaking .","The plates were washed 3X with wash buffer ( PBS + 1 . 5% OG ) .","50 µl of supernatants or purified antibody , neat or diluted in assay buffer , were added to the wells and incubated at RT for 1 hr , while shaking .","The plates were washed 3X with wash buffer .","The captured antibody was detected with goat anti-rat HRP or goal anti-mouse HRP secondary antibody as appropriate ( 50 μl per well diluted at 1:10K in assay buffer , Bethyl Laboratories USA , Cat # A110-236P ) .","The plates were incubated at RT for 1 hr , washed 3X with wash buffer and 50 µl of substrate , TMB solution ( Surmodics , USA Cat # TMBW-1000 ) , was added to each well .","The plates were incubated for 5 min at RT followed by addition of 50 μl of TMB stop solution ( Surmodics , USA Cat # LPSP-1000 ) .","Plates were read at 630 nm .","ELISA assays were performed as part of our high throughput antibody screening workflow and are purely descriptive .","Bacterial cells were grown to log phase , pelleted , re-suspended in ice-cold PBS with 1% BSA , and normalized to an OD600 of 0 . 5 .","Bacterial suspensions were diluted 1:2 . 5 in ice-cold PBS , 1% BSA and 2 × 106 CFU were incubated with 10 µg/ml mAb at 4°C for 1 hr under gentle agitation .","Bacterial cells were then washed three times in ice-cold PBS , 1% BSA and labeled using Alexa 488 anti-mouse IgG ( H + L ) ( 1:1000; Invitrogen ) or Alexa 488 anti-rat IgG ( H + L ) ( 1:1000; Invitrogen ) for 30 min 4°C .","After three washes in ice-cold PBS , bacterial cells were fixed by adding one vol .","of 2% w/v paraformaldehyde in PBS .","Samples were run on a BD FACSCalibur and data from 10 , 000 events were analyzed using FlowJo software .","FACS experiments were performed as part of our high-throughput antibody screening workflow and are purely descriptive .","An isotype matched antibody was used as a negative control .","Bacterial cells were grown to log phase , normalized to OD600 , and pelleted .","Samples were resuspended in 1x LDS sample buffer ( ThermoFisher Scientific ) and boiled for 5 min prior to loading on a 4–12% Bis-Tris SDS-PAGE gel .","Proteins were transferred onto cellulose membranes using the iBlot 2 Dry Blotting System ( ThermoFisher Scientific ) .","Membranes were blocked for 1 hr in Blocking Buffer ( TBS containing 5% nonfat milk and 0 . 05% Tween-20 ) , washed , then incubated either overnight at 4°C or room temperature for 1 hr with the following primary Abs: human α-LptD 3D11 ( 1 μg/mL , Genentech ) and rabbit α-GroEL ( 1:25 , 000 , Enzo ) .","Appropriate HRP-linked secondary antibodies ( GE Healthcare ) were diluted 1:20 , 000 in TBST and incubated with the membrane for 1 hr at RT .","Blots were developed using ECL Prime Western Blotting Detection Reagent ( Amersham ) .","The displayed Western blot experiment shows a single biological replicate .","Growth curves were performed in biological triplicate .","Bacterial strains were grown overnight on LB agar containing gentamicin and arabinose .","Cells were scraped from the plate into fresh media .","OD600 was measured and subsequently diluted to 0 . 001 ( OD600 ) .","100 μl transferred to a 96-well plate ( Corning ) and monitored for growth by measuring OD600 ( EnVision Multimode Plate Reader , PerkinElmer ) .","Bacterial growth curves were analyzed by determining the doubling time ( dt ) during exponential growth phase and compared via the unpaired Student’s t test using Prism 6 . 0 ( GraphPad Software ) ( Supplementary file 1 ) .","The Bonferroni correction was applied to control for multiple comparisons .","Minimum inhibitory concentration ( MIC ) assays were performed by inoculating 105 colony forming units into medium with only test antibiotics ( rifampicin and vancomycin ) .","Biological duplicates were performed .","MIC defined as the lowest concentration of drug to inhibit bacterial growth ."]],"string":"[\n [\n \"The outer membrane ( OM ) of Gram-negative bacteria is a permeability barrier to antibiotics and other cytotoxic agents , such as detergents ( Nikaido , 2003 ) .\",\n \"A key feature of the OM is its distinctive asymmetrical bilayer populated with lipopolysaccharide ( LPS ) in the outer leaflet and phospholipids in the inner leaflet ( Funahara and Nikaido , 1980; Kamio and Nikaido , 1976 ) .\",\n \"Lateral interactions between LPS molecules mediated by divalent cations on the surface and packing of hydrocarbon chains in the membrane impede the passage of both hydrophilic and large hydrophobic molecules ( Nikaido , 2003 ) .\",\n \"LPS is composed of a conserved lipid A molecule , a core oligosaccharide , and a variable O-antigen glycan .\",\n \"Lipid A is synthesized in the cytoplasm where the core oligosaccharide is attached to form core-LPS ( Whitfield and Trent , 2014 ) .\",\n \"This molecule is flipped into the periplasmic space by the inner membrane ATPase MsbA ( Ho et al . , 2018; Mi et al . , 2017; Zhou et al . , 1998 ) .\",\n \"The O-antigen is appended onto core-LPS in the periplasm and LPS is transported to the outer leaflet of the OM by the LPS transport ( Lpt ) system ( Abeyrathne et al . , 2005; Han et al . , 2012; Okuda et al . , 2016 ) .\",\n \"In E . coli , all 7 components of the Lpt pathway , LptABCDEFG , are essential and shuttle LPS molecules from the IM across the periplasm to the OM ( Okuda et al . , 2016; Okuda et al . , 2012; Ruiz et al . , 2008; Wu et al . , 2006 ) .\",\n \"LptD is a β-barrel outer membrane protein ( OMP ) responsible for the final unidirectional insertion of LPS into the OM outer leaflet ( Botos et al . , 2016; Gu et al . , 2015; Li et al . , 2015 ) .\",\n \"The LptD β-barrel is comprised of 26 anti-parallel β-strands with a proposed lateral gate formed where the first and last β-strands interact ( Dong et al . , 2014; Qiao et al . , 2014 ) .\",\n \"Like other Gram-negative β-barrel OMPs , LptD possess long extracellular , environmentally-exposed loops and short periplasmic loops connecting adjacent strands ( Franklin and Slusky , 2018; Rollauer et al . , 2015; Schulz , 2002 ) .\",\n \"The roles of the extracellular loops ( ECLs ) in folding or activity of LptD have not been determined .\",\n \"LptD forms an obligate interaction with the OM lipoprotein LptE ( Chimalakonda et al . , 2011; Freinkman et al . , 2011 ) .\",\n \"An N-terminal jelly roll domain is located on the periplasmic side of the β-barrel below the lateral gate .\",\n \"Structural , biochemical , and genetic analyses suggest that the hydrophobic lipid A portion of LPS is embedded in the hydrophobic jelly roll domain .\",\n \"The presumed model for LPS transport across the OM posits that lipid A is inserted into the hydrophobic OM through a lateral gate of LptD ( Botos et al . , 2016; Gu et al . , 2015; Li et al . , 2015 ) .\",\n \"By an unknown mechanism , the polar sugars of the LPS core oligosaccharide and O-antigen are presumably moved through the LptD β-barrel lumen and flipped to the exterior of the cell .\",\n \"Due to its essential role in OM biogenesis , high conservation among Gram-negative bacteria , and exposure to the extracellular environment , LptD represents a potential new antibacterial target .\",\n \"Genetic experiments show that crosslinking LPS to LptD in the jelly roll domain or the lateral gate , or introduction of disulfide bonds that disrupt the proposed LPS path through LptD , inhibit growth ( Gu et al . , 2015; Li et al . , 2015 ) .\",\n \"Indeed , two antibacterial peptides targeting LptD have been described , both presumably targeting the periplasmic jelly roll domain ( Srinivas et al . , 2010; Vetterli et al . , 2018 ) .\",\n \"Critical roles for ECLs in OMP function have been described ( Browning et al . , 2013; Rigel et al . , 2013 ) .\",\n \"For example , removal of the LptD extracellular loop 4 ( L4 ) , imp4213 , is known to disrupt the OM barrier leading to sensitization to antibiotics ( Sampson et al . , 1989 ) .\",\n \"However , the roles of the extracellular LptD loops remain largely underexplored due to a lack of robust tools to systematically assess structure-function relationships in the context of a native OM .\",\n \"Monoclonal antibodies ( mAbs ) represent useful tools to probe protein structure and function because they typically have high target specificity and affinity .\",\n \"Accordingly , here we pursued four diverse strategies to produce thousands of mAbs that bound to LptD .\",\n \"The breadth and depth of mAb discovery that we report here is unusual , especially for an integral membrane protein , and also takes advantage of an efficient and novel targeted boost-and-sort α-OMP discovery approach recently reported for another OMP , BamA ( Vij et al . , 2018 ) .\",\n \"Characterization of the α-LptD mAbs discovered herein allowed us to map the environmentally-exposed surfaces of LptD and to systematically explore the structure-function relationship of the large ECLs of this essential OMP within its native OM environment .\",\n \"Through the construction and characterization of LptD loop deletion strains , we identified non-exposed loops that are critical for LptD function under defined conditions , which are all inaccessible to antibody interference .\",\n \"Our data provide unique insights into the structure and function of LptD and suggest that non-antibody approaches for interfering with the function of this essential OMP in E . coli should be prioritized to develop novel antibiotics .\"\n ],\n [\n \"To probe the functional relevance of the ECLs of the essential E . coli OMP LptD , we set out to discover α-LptD antibodies to these surface-exposed structures .\",\n \"We generated a diverse mAb library against E . coli LptD by running four independent discovery campaigns .\",\n \"Specifically , we immunized ( 1 ) rats with cyclic and linear peptides derived from extracellular LptD loop sequences , ( 2 ) rats and ( 3 ) mice with purified LptDE protein in detergent ( n-octyl-β-D-glucopyranoside ) and non-detergent polymer ( amphipol A8-35 ) , and ( 4 ) rats with a targeted boost-and-sort strategy that uses whole bacterial cell immunizations followed by immune boosts with purified LptDE protein ( Figure 1A and Figure 1B ) .\",\n \"Within each campaign , the adjuvants and boosting strategies were varied to increase potential antibody diversity ( see Materials and methods for details ) .\",\n \"Cumulatively , the discovery campaigns generated more than 7000 hybridomas and more than 3 , 000 ELISA+ LptD-specific mAbs ( Figure 1C ) .\",\n \"In total , 774 ELISA+ α-LptD mAbs from the peptide and protein rat immunizations , 952 ELISA+ α-LptD mAbs from the mouse immunizations , and 1494 α-LptD mAbs from the targeted boost-and-sort immunizations were purified ( Figure 1C ) .\",\n \"To explore epitope coverage and accessibility of the α-LptD ELISA+ mAbs from three of the antibody discovery campaigns , we first measured their extracellular binding by a fluorescence-assisted cell sorting ( FACS ) -based assay to two strains of E . coli ( Figure 2A ) .\",\n \"The first strain was a laboratory E . coli strain , K-12 , that produces an intact core-LPS but no O-antigen .\",\n \"The second strain was an E . coli ΔwaaD mutant , which produces a truncated form of LPS on the cell surface ( Kneidinger et al . , 2002 ) .\",\n \"This truncation of LPS increases the exposed epitopes accessible to mAbs ( Bentley and Klebba , 1988; Storek et al . , 2018 ) .\",\n \"Depending on the immunization campaign , 9–30% of the ELISA+ α-LptD mAbs were FACS+ on E . coli ΔwaaD , while <1% of the same antibodies were FACS+ on E . coli K-12 ( Figure 2A ) .\",\n \"The E . coli K-12 FACS+ α-LptD mAbs were a subset of all the FACS+ mAbs , binding equally well to E . coli K-12 and E . coli ΔwaaD cells ( Figure 2—figure supplement 1 ) , and were discovered in both mouse and rat immunization campaigns .\",\n \"These results reinforce the finding that LPS presents a significant barrier to accessing surface epitopes on E . coli ( Bentley and Klebba , 1988; Storek et al . , 2018 ) and suggest that the ECLs of OMPs may be masked or intimately associated with the core sugars of the LPS .\",\n \"The diversity of epitopes bound by α-LptD mAbs in our library was next assessed by monitoring binding to LptD from different species .\",\n \"ELISAs were performed with 134 FACS+ and 233 FACS- α-LptD mAbs using purified LptDE protein from two closely related Enterobacteriaceae species , Klebsiella pneumoniae and Enterobacter cloacae , which share 82% and 84% overall protein identity and 71% and 76% identify between the ECLs only when compared with E . coli LptD , respectively ( Figure 2B and Figure 2—figure supplement 2A ) .\",\n \"The majority of FACS+ ( 61% ) and FACS- ( 69% ) α-LptD mAbs bound only E . coli LptDE by ELISA ( Figure 2C ) .\",\n \"This result was not surprising considering that all protein immunization campaigns utilized only purified E . coli LptDE protein .\",\n \"Analysis of the surface-exposed portion of LptD predicted from experimental x-ray crystallographic structural models identified regions of >5 amino acids that were conserved among all three Enterobacteriaceae species , the size of a typical hot spot of residues that contribute to antibody binding ( Stave and Lindpaintner , 2013 ) ( Figure 2B and Figure 2—figure supplement 2B ) .\",\n \"Indeed , we identified 10 ( 7% ) α-LptD mAbs that bound to LptDE from all three species and 42 that bound either E . coli plus K . pneumoniae ( 22% ) or E . coli plus E . cloacae ( 9% ) LptDE ( Figure 2C ) .\",\n \"Thus , our immunization campaigns using E . coli LptD-derived peptides and E . coli LptDE protein yielded diverse antibodies that bound to both unique and conserved extracellular accessible ( i . e . , FACS+ ) and extracellular-inaccessible or periplasmic ( i . e . , FACS- ) epitopes .\",\n \"The differences in cross-species reactivity of the α-LptD mAbs suggested that the FACS+ antibodies bound multiple surface-exposed LptD epitopes .\",\n \"To determine if these antibodies bound unique extracellular epitopes , we tested α-LptD mAbs with the highest FACS+ signal for their abilities to compete with each other for binding to purified E . coli LptDE .\",\n \"Every possible FACS+ pair of α-LptD mAbs was screened for the ability to either bind LptDE protein simultaneously ( i . e . , different epitopes ) or to compete for binding ( i . e . , same or nearby epitopes ) using a SPR-based binding assay ( Abdiche et al . , 2017 ) .\",\n \"From these results , we constructed a surface epitope map for the FACS+ α-LptD mAbs ( Figure 2D ) .\",\n \"We identified two major epitope bins and a possible third bin with limited overlap ( Figure 2D ) .\",\n \"When the species cross-reactivity data were considered , most three-species cross-reactive α-LptD mAbs were tightly clustered , while those that bound only E . coli LptDE formed a distinct cluster ( Figure 2D ) .\",\n \"There were some outliers to these trends with the dual-species binders and E . coli-only binders binning with the cross-reactive mAbs suggesting distinct , but overlapping epitopes .\",\n \"Thus , the α-LptD mAbs we discovered covered multiple distinct and overlapping extracellular LptD epitopes .\",\n \"In order to map the extracellular epitopes of the α-LptD mAbs in a native asymmetrical OM environment , we first generated and characterized a panel of lptD mutants each lacking one ECL .\",\n \"This strain panel also allowed us to assess the essentiality of each ECL in strains with different OM compositions .\",\n \"We constructed 13 lptD mutants , each lacking the region encoding one of the 13 ECLs ( Figure 3A ) .\",\n \"These loop deletions were based on available x-ray crystal structures of LptD and were constructed as complete loop deletions without replacing the removed sequence .\",\n \"In the case of loop 4 ( L4 ) , which encompassed the classical imp4213 allele ( Sampson et al . , 1989 ) and in the case of loop 8 , which is an extensive loop that interacts with LptE in the LptD lumen ( Freinkman et al . , 2011 ) , we only removed a portion of the loop .\",\n \"Each lptD loop mutant was expressed in a lptD-conditional E . coli strain .\",\n \"In the presence of arabinose , which induces a chromosomal wild-type lptD , all strains grew ( Figure 3B ) .\",\n \"In the absence of arabinose , all lptD loop mutants supported growth in the K-12 strain with the exception of loop 2 ( L2 ) removal , which showed reduced growth ( Figure 3B ) .\",\n \"In all cases when growth was observed , the levels of the LptD loop deletion protein produced were similar to the protein level of wild-type LptD ( Figure 3—figure supplement 1A ) .\",\n \"Thus , LptD is highly tolerable to genetic manipulation of its ECLs .\",\n \"When the lptD loop mutants were expressed in an lptD-conditional E . coli ΔwaaD strain , which produces LPS with a truncated core oligosaccharide , the observed loop requirements were different .\",\n \"In the E . coli ΔwaaD background , LptD lacking either loop 2 or loop 10 were unable to support bacterial growth on solid growth agar ( Figure 3B ) .\",\n \"The inability of LptD lacking L10 ( LptDΔL10 ) to support growth of the E . coli ΔwaaD strain suggested that this ECL is important for LptD folding or activity in this strain .\",\n \"In contrast , the LptDΔL10 variant was produced and did support growth of E . coli K-12 ( Figure 3B and Figure 3—figure supplement 1A ) , indicating that LptDΔL10 was folded and sufficiently functional in the strain producing core-LPS but not the minimal LPS of E . coli ΔwaaD strain .\",\n \"To determine if the other LptD loops had differential activities in these two strains , we monitored the growth of all of the LptD loop variants in both the E . coli K-12 and E . coli ΔwaaD strain backgrounds over time in liquid growth media .\",\n \"In the presence of arabinose , which induces a chromosomal wild-type lptD , all strains had similar growth patterns ( Figure 3—figure supplement 1C ) .\",\n \"In the absence of arabinose , the E . coli K-12 strain producing LptDΔL10 grew , but at a decreased rate , and the strain with LptDΔL2 exhibited a dramatically prolonged lag phase and severe growth phenotype ( Figure 3C and Supplementary file 1 ) .\",\n \"In contrast , E . coli ΔwaaD expressing LptDΔL10 did not grow at all , but expression of LptDΔL2 supported growth in liquid media ( Figure 3D and Supplementary file 1 ) .\",\n \"Thus , L2 and L10 are important for LptD folding , activity , interaction with the β-barrel assembly machinery component BamA , or a combination of these .\",\n \"Also , the LptDΔL10 loop deletion mutant was better tolerated in E . coli K-12 compared to the LPS-truncated ΔwaaD strain while in liquid growth media the LptDΔL2 mutant is better tolerated in the E . coli ΔwaaD strain .\",\n \"This result is consistent with the synthetic lethality of other mutations that affect OM biogenesis , specifically bam101 , ΔbamB , ΔsurA , and ΔlpxL , in LPS-truncated E . coli strain backgrounds ( Klein et al . , 2009; Storek et al . , 2019 ) highlighting the redundancy that exists to maintain this important barrier structure .\",\n \"The lptDΔL4 mutant used in this study encompassed the imp4213 allele .\",\n \"This particular loop mutant is viable but increases membrane permeability and sensitivity to OM-excluded antibiotics ( Sampson et al . , 1989 ) .\",\n \"We tested whether the strains producing other LptD loop mutants had increased membrane permeability by measuring sensitivities to rifampicin and vancomycin , two antibiotics excluded by the Gram-negative bacteria OM .\",\n \"The minimal inhibitory concentrations ( MICs ) of both rifampicin and vancomycin were lower for strains producing LptDΔL4 , as expected , and also for strains producing LptDΔL2 , LptDΔL8 , and LptDΔL10 indicating an OM defect ( Table 1 ) .\",\n \"Increases in OM permeability were observed in both E . coli K-12 and E . coli ΔwaaD strain backgrounds .\",\n \"These results suggest that extracellular L2 , L4 , L8 and L10 are important for LptD function and potentially influence LptD folding , interaction with BamA , or activity .\",\n \"We utilized our lptD loop deletion panel to map the α-LptD mAb surface coverage in a native E . coli membrane environment .\",\n \"We characterized binding of 52 FACS+ α-LptD mAbs to the E . coli ΔwaaD loop deletion panel .\",\n \"In the absence of arabinose driving expression of the chromosomal wild-type lptD , production of each mutant LptD was detected and growth of the E . coli ΔwaaD conditional lptD strain was supported by all LptD loop deletion mutants with the exception of L10 , so this construct was not tested in this assay ( Figure 3B and Figure 3—figure supplement 1B ) .\",\n \"The mAbs in the panel were confirmed to bind to an E . coli ΔwaaD strain expressing wild-type lptD ( Figure 4A ) .\",\n \"We expected that if an extracellular LptD loop was critical for binding , removing the loop would reduce α-LptD mAb binding .\",\n \"The lack of binding by any particular mAb to a LptD loop mutant could be due to removal of the mAb binding site , in part or in full , or a structural change that alters the mAb binding site when a loop is removed .\",\n \"The majority of the α-LptD mAbs could be categorized into one of 7 general binding patterns based on the mutants they were unable to bind with multiple representatives in each grouping: ( 1 ) Loop 4 ( L4 ) , ( 2 ) Loop 6/7 ( L6/L7 ) , ( 3 ) Loop 6/7/8 ( L6/L7/L8 ) , ( 4 ) Loop 8/9 ( L8/L9 ) , ( 5 ) Loop 9 ( L9 ) , ( 6 ) Loop 11 ( L11 ) , and ( 7 ) Loop 13 ( L13 ) ( Figure 4B and Figure 4—figure supplement 1 ) .\",\n \"Thus , our immunization campaigns produced antibodies that allow systematic probing of accessible ELCs around the LptD structure .\",\n \"In some cases , for example , a representative E . coli K-12 FACS+ α-LptD mAb ( Figure 2—figure supplement 1B ) , no loop mutant dramatically altered antibody binding .\",\n \"This pattern suggests a surface-exposed epitope was present that was not altered in these particular loop mutants or the mAb bound a composite epitope that was not sufficiently disrupted by loss of a portion of the epitope encompassed by one individual loop deletion ( Figure 4B ) .\",\n \"This could be especially true for L8 as the deletion encompassed only part of the loop that , based on structural models , is likely to occupy the LptD lumen while additional exposed portions were left unchanged ( Figure 4C ) .\",\n \"For every loop deletion , at least one FACS+ α-LptD mAb had reduced cell binding by >90% .\",\n \"The notable exceptions were L1 , L2 , L5 and L12 , but mAbs could be found that exhibited >80% decreased binding for L1 , L5 and L12 .\",\n \"Comparing these data with the species cross-reactivity and in vitro binding competition epitope map provided a complete picture of the mAb accessible surface of LptD .\",\n \"Specifically , a majority of the two- and three-species cross-reactive α-LptD mAbs were dependent upon L8 , L9 , or both for binding ( Figure 2B , Figure 2D , Figure 2—figure supplement 2 , Figure 4B , and Figure 4—figure supplement 1 ) .\",\n \"The cross-reactive mAbs 5A3 and 3D4 exhibited slightly different binding requirements , L6/L8/L9/L11 and L7/L8 , respectively , consistent with these mAbs recognizing different epitopes in our in vitro binning experiment ( Figure 2D ) .\",\n \"Additionally , E . coli-specific α-LptD mAbs , but not the three-species cross-reactive mAbs , were defective for binding upon removal of L3 , L7 , or L13 ( Figure 2B , Figure 2D , Figure 2—figure supplement 2 , Figure 4B , and Figure 4—figure supplement 1 ) .\",\n \"These α-LptD mAbs and knowledge of their binding sites are valuable tools for probing the accessible LptD surface and suggest that LptD can tolerate extensive alteration to its extracellular surface with little or no effect on folding or function .\",\n \"Because distinct antibody discovery approaches were utilized , we were also able to evaluate the potential correlation of LptD targeting with regard to the immunization campaign .\",\n \"Partitioning the mAbs with respect to the host animal used for immunization highlighted a trend for L6 , L7 , and L8 to be critical from the rat immunization campaign , while mAbs from the mouse immunization campaign tended to yield FACS+ antibodies that required L8 and L9 .\",\n \"Because these animals are likely not tolerized to the foreign bacterial LptD protein , this difference could reflect differences in the types of binding sites presented for rat versus mouse antibodies .\",\n \"Thus , even within these immunization campaigns , diverse binding profiles to the loop mutants were seen ( Figure 4B and Figure 4—figure supplement 1 ) .\",\n \"A potential application of antibodies raised against extracellularly-exposed bacterial proteins is as direct-acting antibacterial molecules ( LaRocca et al . , 2009; Storek et al . , 2018 ) .\",\n \"Disruption of L2 or L10 resulted in growth defects for E . coli ΔwaaD ( Figure 3B ) , however , we did not identify any α-LptD mAbs that bound L2 , and both L2 and L10 appear to be inaccessible in available structural models of LptD ( Figure 4C ) ( Botos et al . , 2016; Dong et al . , 2014; Qiao et al . , 2014 ) .\",\n \"Indeed , when we screened our entire antibody catalog using growth inhibition as a readout for mAb activity , we did not identify any α-LptD mAbs able to inhibit growth of E . coli K-12 or E . coli ΔwaaD by >50% ( Table 2 and Figure 4—figure supplement 2 ) .\",\n \"Combined with our LptD loop deletion analysis , the lack of ability to identify an inhibitory α-LptD mAb suggests that LptD may be specifically evolved to withstand extracellular modulation by antibodies while protecting critical structural elements of LptD function .\"\n ],\n [\n \"The OM is a critical feature of Gram-negative bacteria and presents a barrier to the discovery of new antibiotics ( Lewis , 2013 ) .\",\n \"There are two essential OMPs embedded in the OM and involved in construction of the OM barrier .\",\n \"BamA is part of the β-barrel assembly machine that folds and inserts β-barrel OMPs , including BamA and LptD , into the OM ( Konovalova et al . , 2017; Ricci and Silhavy , 2012 ) .\",\n \"LptD is the last protein in the Lpt pathway that establishes and maintains OM asymmetry by inserting LPS exclusively into the outer leaflet of the OM ( Okuda et al . , 2016 ) .\",\n \"Despite the importance of these β-barrel OMPs , understanding of how their structures contributes to their folding and function is still incomplete .\",\n \"Similar to other Gram-negative β-barrel OMPs , both BamA and LptD possess short periplasmic loops but long , variable , and environmentally-exposed ECLs ( Franklin and Slusky , 2018; Rollauer et al . , 2015; Schulz , 2002 ) .\",\n \"A deeper examination of the roles of these ECLs in a native OM context could contribute to a greater understanding of how they contribute to the structure and function of OMPs in general .\",\n \"To systematically interrogate the LptD ECLs , we performed four antibody discovery campaigns to generate thousands of antibodies targeting LptD to probe structure-function relationships of this essential OMP in its native OM environment .\",\n \"The unusual breadth and depth of our approach represents a powerful way to probe the accessible surfaces of the ECLs of LptD and to tease apart the changes that accompany LPS insertion by LptD in the native OM environment of the bacterial cell .\",\n \"By characterizing hundreds of these α-LptD mAbs , we discovered that extensive portions of the extracellular surface of LptD are dispensable in E . coli and critical features of LptD are likely occluded and protected by non-essential loops ( Figure 5 ) .\",\n \"Little is known about the role and dynamics of the ECLs of LptD in the function of LPS insertion into the OM and there are few available tools to study this process .\",\n \"In an effort to probe as many extracellular epitopes as possible , we generated a large , diverse α-LptD mAb library by undertaking multiple immunization campaigns utilizing different antigens , adjuvants , and hosts .\",\n \"Varying antigens systematically presented LptD in increasingly complex arrangements to potentially capture different available epitopes .\",\n \"First , peptide immunizations utilized both linear peptides derived from conserved and non-conserved LptD ECL sequences and cyclic variants in an effort to constrain particular conformations .\",\n \"Next , purified full-length LptDE immunizations aimed to capture a more native state of the ECLs .\",\n \"Because the matrix in which the protein is reconstituted can alter the conformational flexibility and exposure of membrane-proximal epitopes , we utilized multiple reconstitution matrices ( i . e . , detergents and amphipols ) .\",\n \"Finally , using an approach that we have recently leveraged to find rare , inhibitory α-BamA antibodies , a targeted boost-and-sort strategy , we immunized rats with sub-lethal E . coli infections and then boosted them with purified LptDE protein to enrich for any LptD antibody response ( Vij et al . , 2018 ) .\",\n \"Taken together , our effort to maximize the epitope coverage in our α-LptD mAb library led to distinct clusters of antibodies that possessed differing abilities to cross-react with LptD species variants , required different ECLs for binding , and exhibited varying levels of access to LptD through the protective LPS layer .\",\n \"Although the full breadth of antibody diversity in our library required the multiple immunization approaches described here , we propose a streamlined workflow for future OMP efforts .\",\n \"An accessible approach is to immunize SD rats with recombinant purified OMP either with or without whole bacterial cells ( strategies 3 and 4 , Figure 1 ) .\",\n \"Both of these workflows returned the highest percentages of FACS+ antibodies coupled with recognition of diverse epitopes ( Figure 2 , Figure 4B , and Figure 4—figure supplement 1 ) .\",\n \"Given the ease of manipulating bacterial strains , it could even be possible to steer immunizations towards particular ECLs by initially immunizing with a ΔECL-OMP strain followed by boosting with recombinant protein .\",\n \"Other scaffolds or antibodies from different host species might be able to recognize distinct epitopes on the exposed portions of LptD .\",\n \"For example , recognition of partially buried epitopes could be achieved using species , such as rabbits , chickens , llamas , or camels , that contain longer CDRH3s that are better suited to reach into cavities on a particular antigen ( Lavinder et al . , 2014; Muyldermans and Smider , 2016 ) .\",\n \"In the case of the mouse and rat antibodies described here , we found that seven different LptD ECLs , out of a total of 13 , could affect binding of the mAbs and the remaining six loops are likely partially or fully buried based on models from crystal structures ( Botos et al . , 2016; Dong et al . , 2014; Qiao et al . , 2014 ) ( Figure 5 ) .\",\n \"When this α-LptD mAb campaign is combined with of our α-BamA mAb effort ( Storek et al . , 2018; Vij et al . , 2018 ) , we now have the ability to probe the structure-function relationships of the only two essential OMPs in their native OM environment in E . coli at a level not previously possible .\",\n \"Importantly , we note the antibody discovery methods that we describe herein are efficient , robust , and a streamlined approach can be readily applied within a standard laboratory setting .\",\n \"We have observed that in LptD there appear to be differential requirements for particular ECLs , L2 and L10 , depending on the structure of LPS , specifically wild-type LPS versus the truncated ΔwaaD LPS .\",\n \"Notably , BamA appears to be defective in the highly fluid OM of an LPS-truncated strain grown under low salt and high temperature conditions ( Storek et al . , 2018 ) .\",\n \"This is manifest by an increased sensitivity to a recently discovered inhibitory α-BamA antibody MAB1 and a requirement for elevated BamA levels .\",\n \"In these cases , the defects can be suppressed by altering the conditions to increase membrane rigidity ( Storek et al . , 2019 ) .\",\n \"The LPS-dependent requirement for LptD loops suggests that effects of the OM structure on OMPs might be more general .\",\n \"Among the 13 extracellular LptD loops , both L2 and L10 are short and lack large insertions or deletions across LptD homologs when compared to the longer loops in highly divergent LptD proteins ( Figure 5—figure supplement 1A ) .\",\n \"One possible reason for this is that these loops are uniquely positioned to make interactions with the LPS in the OM to allow for proper folding or functioning of this essential OMP .\",\n \"Although these interpretations are speculative and complicated by the fact that LPS is the substrate for LptD , it is tempting to hypothesize that the structure and function of OMPs evolved to be maximal when embedded in the constrained , liquid crystalline , OM matrix ( Paracini et al . , 2018 ) and altering the nature of this membrane leads to defects .\",\n \"Thus , membrane environment is a critical consideration when interpreting biochemical and structural data of LptD and other OMPs in detergent micelles and must be considered when attempting to design or discover inhibitors of these essential processes in Gram-negative bacteria .\",\n \"Genetic , biochemical , and structural data support a model for LPS insertion into the OM whereby the acyl chains of lipid A are inserted into the hydrophobic bilayer through a lateral gate created between the first and last strands of the LptD β-barrel ( Botos et al . , 2016; Dong et al . , 2014; Qiao et al . , 2014 ) .\",\n \"It is less clear how the polar sugars of the core oligosaccharide and O-antigen emerge from the periplasm to the cell exterior as it would be unfavorable for them to pass directly through the membrane .\",\n \"Presumably the sugars must pass through the LptD transmembrane β-barrel lumen , which is also occupied by an essential lipoprotein , LptE , and then through a proposed opening in the surface-exposed cap of LptD .\",\n \"Thus , consistent with prior molecular dynamic simulations , there are likely major structural surface re-arrangements in LptD during LPS insertion ( Dong et al . , 2014 ) .\",\n \"Given this model , we hypothesized that an antibody could have interfered with this process in a number of ways .\",\n \"For example , a mAb could staple the lateral gate in either a closed or open conformation , prevent the movements associated with expelling the sugars to the outside , or allosterically alter the protein so that LPS movement is thwarted .\",\n \"It is informative , therefore , that none of our identified α-LptD mAbs were able to inhibit the activity of LptD .\",\n \"One explanation is that binding to a single epitope is insufficient to disrupt the required movements of LptD .\",\n \"Alternatively , the positions on LptD critical for these presumed structural changes might be inaccessible to mAbs .\",\n \"Indeed , few or no α-LptD mAbs bound to the extracellular LptD loops that were found genetically to be important for viability or OM maintenance under specific conditions .\",\n \"Consistent with this possibility , in crystal structures of LptDE , the two loops we found most critical were occluded by other parts of the protein: L2 by L3 , and L10 by L9 and L11 ( Dong et al . , 2014; Qiao et al . , 2014 ) .\",\n \"By burying critical regions , LptD appears to avoid direct interference by antibodies , providing bacteria an evolutionary advantage in evading functional antagonism by the host immune system .\",\n \"Accordingly , the antibody-accessible LptD loops also are the most diverse in sequence , suggesting that these regions may be subject to selective pressures from the immune system ( Botos et al . , 2016 ) ( Figure 5—figure supplement 1B ) .\",\n \"Moreover , the proximity of a particular epitope to the membrane can sterically or electrostatically block binding .\",\n \"However , in rare cases , recognition can be achieved via stabilizing hydrophobic residues in an antibody CDR that transiently interact with the membrane , as observed for several α-HIV or α-GPCR antibodies ( Ishchenko et al . , 2017; Scherer et al . , 2010 ) .\",\n \"It remains to be seen whether immobilizing these loops in LptD , for example , with a small molecule or peptide macrocycle that can access these positions , would inhibit LptD activity .\",\n \"While our observations make direct interference with LptD function by antibodies unlikely , they do not rule out the potential value of LptD as a vaccine candidate ( Zha et al . , 2016 ) .\",\n \"Overall , our study represents an exhaustive α-LptD antibody discovery campaign , and potentially the most exhaustive on any membrane protein described to date .\",\n \"For the first time , we have been able to map essential and non-essential structural features of this intriguing and important potential antibiotic target .\",\n \"We have observed extreme tolerability of the ECLs to interference , suggesting that the architecture of the exposed ECLs of LptD play a role in protecting critical functions of this essential OMP .\",\n \"More generally , we provide a robust template for future efforts to dissect the structure-function relationships of complex membrane proteins from bacteria and mammals in their native environments , and a method to rapidly assess the therapeutic potential of antibody targeting .\"\n ],\n [\n \"Luria-Bertani broth ( ThermoFisher Scientific 12795027 ) and Mueller Hinton II cation-adjusted broth ( BBL 212322 ) was prepared according to manufacturer's instructions .\",\n \"Bacterial cultures were grown at 37°C .\",\n \"When appropriate , media was supplemented with kanamycin ( 50 μg/mL ) , carbenicillin ( 50 μg/mL ) , chloramphenicol ( 12 . 5 μg/mL ) , hygromycin ( 200 μg/mL ) , gentamicin ( 10 μg/mL ) and arabinose ( 0 . 2% vol/vol ) .\",\n \"Bacterial strains and relevant primers are listed in Supplementary file\",\n \"2 . Kanamycin deletion-insertion mutations of waaD was obtained from the Keio collection ( Baba et al . , 2006 ) .\",\n \"Briefly , pKD4 or pKD3 was amplified with primers containing 50 bp nucleotide homology extensions ( Supplementary file 2 ) to the gene of interest .\",\n \"The linear product was transformed into the appropriate background strain containing pSIM18 ( Chan et al . , 2007 ) recovered for 4 hr and selected on media containing 50 μg/mL kanamycin or 12 . 5 μg/mL chloramphenicol as appropriate .\",\n \"Mutations were confirmed by PCR .\",\n \"To make sequential mutants , the antibiotic marker was flipped out using pCP20 ( Datsenko and Wanner , 2000 ) .\",\n \"The conditional lptD strain , ΔlptD::PBAD-lptD , was created by inserting PBAD-lptD at the attB site in MG1655 followed by deletion of the native copy of lptD ( Datsenko and Wanner , 2000; Diederich et al . , 1992 ) .\",\n \"Briefly , lptD was cloned into pBAD24 using standard methods .\",\n \"PBAD-lptD was amplified from pBAD24-lptD and sub-cloned into pLDR9 .\",\n \"pLDR9-PBAD-lptD was digested with XbaI , ligated , and transformed into MG1655 expressing pLDR8 .\",\n \"PCR and DNA sequencing confirmed insertion of PBAD-lptD at the attB site .\",\n \"After integration of PBAD-lptD , the native copy of lptD was deleted using λ Red recombination as described above .\",\n \"In the absence of arabinose , the conditional DlptD::PBAD-lptD strain did not grow .\",\n \"To obtain a conditional lptD strain truncated for LPS , waaD was deleted using λ Red recombination .\",\n \"To clone the loop mutants into the lptD conditional strains , pLMG18 was modified to replace the original antibiotic cassette providing chloramphenicol resistance with a gentamicin resistance cassette using Gibson Assembly .\",\n \"The lptD gene and mutant constructs were codon optimized and ordered as gBlocks ( Integrated DNA technologies ) .\",\n \"The genes were amplified and cloned into pLMG18gm using Gibson Assembly .\",\n \"Clones were confirmed by sequencing .\",\n \"The E . coli LptDE expression plasmid was constructed by cloning E . coli LptD ( amino acids 1–784 ) and E . coli LptE ( amino acids 1–193 ) into pCDFDuet1 vector at the first and second multiple cloning sites .\",\n \"An AviTag followed by an 8x histidine tag was inserted into N-terminal of E . coli LptD .\",\n \"The expression plasmid was expressed in E . coli C41 ( DE3 ) .\",\n \"When the OD600 reached 0 . 9 , cells were induced with 0 . 5 mM isopropyl-β-D-1 thiogalactopyranoside ( IPTG ) overnight at 18°C .\",\n \"Cells were harvested by centrifugation at 4500 rpm for 15 min and the cell pellets were resuspended in lysis buffer ( 50 mM Tris , pH 8 . 0 , 200 mM NaCl , 1x complete protease inhibitor mixture ( Roche ) ) .\",\n \"The resuspended cells were disrupted by passing through a microfluidizer at 10 , 000 psi three times .\",\n \"The cell lysate was added with 2% Zwittergent 3–14 and rocked overnight at 4°C .\",\n \"The suspension was ultracentrifuged ( 45Ti rotor , 40 , 000 rpm ) for 1 hr at 4°C .\",\n \"The supernatant was incubated with pre-equilibrated Ni-NTA resins ( Qiagen ) and rocked for 1 hr at 4°C .\",\n \"The resins were washed by the buffer containing 50 mM Tris , pH 8 . 0 , 200 mM NaCl , 1% Zwittergent 3–14 , 15 mM imidazole , 1x complete protease inhibitor mixture and eluted with the buffer consisting of 50 mM Tris , pH 8 . 0 , 200 mM NaCl , 1% Zwittergent 3–14 , 300 mM imidazole and 1x complete protease inhibitor mixture .\",\n \"The eluate was concentrated and purified by a Superdex 200 16/60 column ( GE Healthcare ) using 20 mM HEPES , pH 8 . 0 , 100 mM NaCl , 1 . 5% n-Octyl-β-D-Glucopyranoside ( OG; Anatrace ) , 1x complete protease inhibitor mixture as the running buffer .\",\n \"The obtained LptD/E protein was diluted 4-fold into the buffer of 20 mM HEPES , pH 8 . 0 , 1 . 5% OG and further polished using Q HP anion exchange 5 ml column ( GE Healthcare ) .\",\n \"20 mM HEPES , pH 8 . 0 , 25 mM NaCl , 1 . 5% OG was used as the start buffer and elution was achieved with linear gradient of 0 . 025–1 M NaCl .\",\n \"When protein was required in amphipol , LptD/E protein was incubated with biotinylated amphipol ( Anatrace ) overnight at 4°C and then purified over a Superdex 200 16/60 column in a buffer consisting of 20 mM HEPES , pH 8 . 0 , 100 mM NaCl .\",\n \"Peptide pools are listed in Supplementary file\",\n \"3 . Peptides were designed based on sequence conservation , surface exposure and loop location .\",\n \"When LptDE protein was required for ELISA or antigen-based sorting , an in vitro biotinylation reaction using BirA enzyme targeting the N-terminal Avi tag was first carried out according to the manufacturer's suggestions ( Avidity ) ; LptDE was then rerun over the Superdex 200 ( 26/60 ) in buffer E as described above in order to remove free biotin and other reaction components .\",\n \"When protein was required in amphipol ( either non-biotinylated or biotinylated ) , LptDE at 1 mg/mL in buffer E was incubated with a stock solution of 2 mg/mL A8-35 amphipol ( Anatrace ) at 22°C for 1 hr and then applied over a Superdex 200 ( 26/60 ) column in buffer F ( 50 mM Tris pH 8 , 100 mM NaCl ) .\",\n \"Protein reconstituted in A8-35 amphipol for immunization was concentrated to 1 mg/mL using a centrifugal device ( 10 K MWKO; Millipore ) .\",\n \"When PE-labeled protein was required for antigen-specific sorting , LptDE at 1 mg/mL ( biotinyated and reconstituted in A8-35 amphipol in buffer F ) was mixed 1:1 ( vol/vol ) with PE-streptavidin ( Jackson ImmunoResearch ) reconstituted in buffer F; the PE-streptavidin-biotin-LptDE-amphipol complex was incubated at 22°C for at least 15 min prior to use .\",\n \"All animal study designs for the mouse and rat immunizations were reviewed and approved by the Genentech Institutional Animal Care and Use Committee prior to the start of this work .\",\n \"All animal work was performed in accordance with relevant guidelines and regulations .\",\n \"LptD peptide immunization campaign immunized Balb/c mice ( Charles River Laboratories , Hollister , CA ) with cyclic and linear LptD peptide pools ( Supplementary file 3 ) .\",\n \"The initial immunization contained 100 μl Complete Freund’s Adjuvant ( CFA ) and subsequent dosing included 100 μl Incomplete Freund’s Adjuvant ( IFA ) ( Sigma ) .\",\n \"pAbs were purified by Protein A and assayed by ELISA , FACS , and inhibition of bacterial growth as described below .\",\n \"Hybridoma fusions were performed as previously described and supernatants were screened for protein binding by ELISA ( Hazen et al . , 2014 ) .\",\n \"All ELISA positive clones were purified and screened for inhibition of bacterial growth .\",\n \"LptDE protein immunization campaign immunized Balb/c mice ( Charles River Laboratories , Hollister , CA ) with detergent-solubilized or amphipol-reconstituted E . coli LptDE protein and either CFA or Ribi adjuvants ( Sigma ) .\",\n \"pAbs were purified and screened as described above .\",\n \"LptDE protein immunization campaign immunized Sprague Dawley rats ( Charles River Laboratories , Hollister , CA ) with detergent-solubilized or amphipol-reconstituted E . coli LptDE protein and either CFA or Ribi adjuvants ( Sigma ) .\",\n \"pAbs were purified and screened as described above .\",\n \"Bacterial immunization campaign immunized Sprague Dawley rats ( Charles River Laboratories , Hollister , CA ) with E . coli K-12 cells in PBS ( 107–109 colony forming units via intravenous injection ) .\",\n \"The rats were boosted four times with E . coli K-12 cells followed by two protein boosts with 10 μg E . coli LptDE protein in amphipol .\",\n \"To generate monoclonal antibodies , hybridoma fusions were performed as previously described except with a myeloma partner SP2ab that enables surface display of IgG cell ( Hazen et al . , 2014; Price et al . , 2009 ) .\",\n \"After HAT selection in ClonaCell-HY Medium C ( StemCell Technologies ) for 4 days , hybridomas were stained with a cocktail of FITC-conjugated anti-rat IgG1/IgG2a/IgG2b mAbs ( 1:100 dilution; Bethyl Laboratories ) and PE-conjugated LptDE protein antigen ( 5 μg/μl ) .\",\n \"Samples were sorted on a FACSAria II cell sorter ( BD Biosciences ) .\",\n \"Gating strategy to identify hybridoma cells was first based on size ( FSC/SSC ) .\",\n \"After dead-cell exclusion with Propidium iodide ( Sigma-Aldrich P-4864 ) , IgG +LptDE + single cell hybridomas were sorted into 96-well tissue-culture plates containing 200 μl of ClonaCell-HY Medium E ( StemCell Technologies ) .\",\n \"Profiles were analyzed by FlowJo v . 9 . 7 . 7 software .\",\n \"Sorted cells were cultured for seven days .\",\n \"Hybridomas were cultured using a previously described semi-automated high throughput process for hybridoma culturing and antibody purification ( Vij et al . , 2018 ) .\",\n \"Epitope bins were determined by 96 × 96 array-based SPR imaging system ( Carterra , USA ) using classical sandwich method .\",\n \"Purified antibodies were diluted at 20 μg/ml in 10 mM sodium acetate buffer pH 4 . 5 .\",\n \"Using amine coupling , antibodies were directly immobilized onto a SPR sensorprism CMD 200 M chip ( XanTec Bioanalytics , Germany ) using a Continuous Flow Microspotter ( Carterra , USA ) to create an array of 96 antibodies .\",\n \"For antibody binning , printed chip was loaded on IBIS MX96 SPRi ( Carterra USA ) and E . coli LptDE protein , diluted to 50 nM in 1 . 5% bOG HBS-P buffer , was injected over the chip at 25°C followed by a second injection of purified antibody , diluted at 20 μg/ml in 1 . 5% bOG HBS-P buffer , to make a sandwich .\",\n \"The epitope binning data were processed using Carterra binning software tool .\",\n \"Epitope binning experiments were performed as part of our high throughput antibody screening workflow and are purely descriptive .\",\n \"The test strain was grown to log phase in MHB II supplemented with 0 . 002% Tween-80 , diluted to a final OD600 0 . 01 in sterile round-bottom 96-well plates ( Costar ) .\",\n \"Antibodies were added at 10 μg/mL , incubated statically at 37°C and monitored for bacterial growth after 4 hr of static incubation .\",\n \"A plate reader at OD600 measured optical density of bacterial growth after shaking the plate for 25 s .\",\n \"Bacterial inhibition was calculated by subtracting the OD600 value from the media control , followed by dividing that value by the no antibody control OD600 value .\",\n \"Antibody activity experiments were performed as part of our high throughput antibody screening workflow and are purely descriptive .\",\n \"Antibodies were screened by capture ELISA .\",\n \"Briefly , 50 µl of biotinylated LptDE protein , diluted in assay buffer ( PBS + 1 . 5% OG +0 . 5% BSA ) was added to streptavidin coated 384 well plates ( Thermo Scientific USA , Cat # 15504 ) and incubated at RT for 1 hr , while shaking .\",\n \"The plates were washed 3X with wash buffer ( PBS + 1 . 5% OG ) .\",\n \"50 µl of supernatants or purified antibody , neat or diluted in assay buffer , were added to the wells and incubated at RT for 1 hr , while shaking .\",\n \"The plates were washed 3X with wash buffer .\",\n \"The captured antibody was detected with goat anti-rat HRP or goal anti-mouse HRP secondary antibody as appropriate ( 50 μl per well diluted at 1:10K in assay buffer , Bethyl Laboratories USA , Cat # A110-236P ) .\",\n \"The plates were incubated at RT for 1 hr , washed 3X with wash buffer and 50 µl of substrate , TMB solution ( Surmodics , USA Cat # TMBW-1000 ) , was added to each well .\",\n \"The plates were incubated for 5 min at RT followed by addition of 50 μl of TMB stop solution ( Surmodics , USA Cat # LPSP-1000 ) .\",\n \"Plates were read at 630 nm .\",\n \"ELISA assays were performed as part of our high throughput antibody screening workflow and are purely descriptive .\",\n \"Bacterial cells were grown to log phase , pelleted , re-suspended in ice-cold PBS with 1% BSA , and normalized to an OD600 of 0 . 5 .\",\n \"Bacterial suspensions were diluted 1:2 . 5 in ice-cold PBS , 1% BSA and 2 × 106 CFU were incubated with 10 µg/ml mAb at 4°C for 1 hr under gentle agitation .\",\n \"Bacterial cells were then washed three times in ice-cold PBS , 1% BSA and labeled using Alexa 488 anti-mouse IgG ( H + L ) ( 1:1000; Invitrogen ) or Alexa 488 anti-rat IgG ( H + L ) ( 1:1000; Invitrogen ) for 30 min 4°C .\",\n \"After three washes in ice-cold PBS , bacterial cells were fixed by adding one vol .\",\n \"of 2% w/v paraformaldehyde in PBS .\",\n \"Samples were run on a BD FACSCalibur and data from 10 , 000 events were analyzed using FlowJo software .\",\n \"FACS experiments were performed as part of our high-throughput antibody screening workflow and are purely descriptive .\",\n \"An isotype matched antibody was used as a negative control .\",\n \"Bacterial cells were grown to log phase , normalized to OD600 , and pelleted .\",\n \"Samples were resuspended in 1x LDS sample buffer ( ThermoFisher Scientific ) and boiled for 5 min prior to loading on a 4–12% Bis-Tris SDS-PAGE gel .\",\n \"Proteins were transferred onto cellulose membranes using the iBlot 2 Dry Blotting System ( ThermoFisher Scientific ) .\",\n \"Membranes were blocked for 1 hr in Blocking Buffer ( TBS containing 5% nonfat milk and 0 . 05% Tween-20 ) , washed , then incubated either overnight at 4°C or room temperature for 1 hr with the following primary Abs: human α-LptD 3D11 ( 1 μg/mL , Genentech ) and rabbit α-GroEL ( 1:25 , 000 , Enzo ) .\",\n \"Appropriate HRP-linked secondary antibodies ( GE Healthcare ) were diluted 1:20 , 000 in TBST and incubated with the membrane for 1 hr at RT .\",\n \"Blots were developed using ECL Prime Western Blotting Detection Reagent ( Amersham ) .\",\n \"The displayed Western blot experiment shows a single biological replicate .\",\n \"Growth curves were performed in biological triplicate .\",\n \"Bacterial strains were grown overnight on LB agar containing gentamicin and arabinose .\",\n \"Cells were scraped from the plate into fresh media .\",\n \"OD600 was measured and subsequently diluted to 0 . 001 ( OD600 ) .\",\n \"100 μl transferred to a 96-well plate ( Corning ) and monitored for growth by measuring OD600 ( EnVision Multimode Plate Reader , PerkinElmer ) .\",\n \"Bacterial growth curves were analyzed by determining the doubling time ( dt ) during exponential growth phase and compared via the unpaired Student’s t test using Prism 6 . 0 ( GraphPad Software ) ( Supplementary file 1 ) .\",\n \"The Bonferroni correction was applied to control for multiple comparisons .\",\n \"Minimum inhibitory concentration ( MIC ) assays were performed by inoculating 105 colony forming units into medium with only test antibiotics ( rifampicin and vancomycin ) .\",\n \"Biological duplicates were performed .\",\n \"MIC defined as the lowest concentration of drug to inhibit bacterial growth .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Outer membrane proteins ( OMPs ) in Gram-negative bacteria dictate permeability of metabolites , antibiotics , and toxins .","Elucidating the structure-function relationships governing OMPs within native membrane environments remains challenging .","We constructed a diverse library of >3000 monoclonal antibodies to assess the roles of extracellular loops ( ECLs ) in LptD , an essential OMP that inserts lipopolysaccharide into the outer membrane of Escherichia coli .","Epitope binning and mapping experiments with LptD-loop-deletion mutants demonstrated that 7 of the 13 ECLs are targeted by antibodies .","Only ECLs inaccessible to antibodies were required for the structure or function of LptD .","Our results suggest that antibody-accessible loops evolved to protect key extracellular regions of LptD , but are themselves dispensable .","Supporting this hypothesis , no α-LptD antibody interfered with essential functions of LptD .","Our experimental workflow enables structure-function studies of OMPs in native cellular environments , provides unexpected insight into LptD , and presents a method to assess the therapeutic potential of antibody targeting ."],"string":"[\n \"Outer membrane proteins ( OMPs ) in Gram-negative bacteria dictate permeability of metabolites , antibiotics , and toxins .\",\n \"Elucidating the structure-function relationships governing OMPs within native membrane environments remains challenging .\",\n \"We constructed a diverse library of >3000 monoclonal antibodies to assess the roles of extracellular loops ( ECLs ) in LptD , an essential OMP that inserts lipopolysaccharide into the outer membrane of Escherichia coli .\",\n \"Epitope binning and mapping experiments with LptD-loop-deletion mutants demonstrated that 7 of the 13 ECLs are targeted by antibodies .\",\n \"Only ECLs inaccessible to antibodies were required for the structure or function of LptD .\",\n \"Our results suggest that antibody-accessible loops evolved to protect key extracellular regions of LptD , but are themselves dispensable .\",\n \"Supporting this hypothesis , no α-LptD antibody interfered with essential functions of LptD .\",\n \"Our experimental workflow enables structure-function studies of OMPs in native cellular environments , provides unexpected insight into LptD , and presents a method to assess the therapeutic potential of antibody targeting .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["The overuse and misuse of antibiotics has led to the rise of multi-drug resistant bacteria which threaten global public health .","Antibiotics interfere with essential processes in bacteria so they are unable to divide or survive , but over time , the microbes have found ways to become immune to the drugs .","New antibiotics are now desperately needed .","Gram-negative bacteria are wrapped in an outer membrane made of large molecules called lipopolysaccharides .","This structure is an extra barrier to molecules ( such as drugs ) that try to enter the cell , but it could also hold new targets for antibiotics to exploit .","A protein called LptD is embedded in the outer membrane , where it inserts new lipopolysaccharides .","It is critical for bacteria to grow and survive , and is a relatively new potential target for antibiotic development .","The protein has a number of ‘extracellular loops’ that extend into the environment , but their roles in the structure and the activity of LptD are still largely unknown .","This is partly due to a lack of tools to investigate these elements .","In response , Storek et al . built a library of over 3 , 000 custom antibodies , which are small Y-shaped proteins that can each recognise a specific portion in one of the extracellular loops and potentially incapacitate LptD .","The antibodies were used to target LptD in its native environment , when it is embedded in the bacteria .","In parallel , mutant bacteria were created in which the loops were genetically removed one by one to assess their importance for LptD activity .","The experiments revealed that although the antibodies could target most extracellular loops , they could not target the few loops that were essential for LptD to work properly .","This suggests that antibody-accessible loops are expendable and that these structures could serve to shield other regions of LptD which are critical for survival .","The findings will help to prioritise research that develops other approaches to inhibit LptD .","Finally , the antibody workflow designed by Storek et al . can serve as a road map to study other membrane proteins in their native cellular environment ."],"string":"[\n \"The overuse and misuse of antibiotics has led to the rise of multi-drug resistant bacteria which threaten global public health .\",\n \"Antibiotics interfere with essential processes in bacteria so they are unable to divide or survive , but over time , the microbes have found ways to become immune to the drugs .\",\n \"New antibiotics are now desperately needed .\",\n \"Gram-negative bacteria are wrapped in an outer membrane made of large molecules called lipopolysaccharides .\",\n \"This structure is an extra barrier to molecules ( such as drugs ) that try to enter the cell , but it could also hold new targets for antibiotics to exploit .\",\n \"A protein called LptD is embedded in the outer membrane , where it inserts new lipopolysaccharides .\",\n \"It is critical for bacteria to grow and survive , and is a relatively new potential target for antibiotic development .\",\n \"The protein has a number of ‘extracellular loops’ that extend into the environment , but their roles in the structure and the activity of LptD are still largely unknown .\",\n \"This is partly due to a lack of tools to investigate these elements .\",\n \"In response , Storek et al . built a library of over 3 , 000 custom antibodies , which are small Y-shaped proteins that can each recognise a specific portion in one of the extracellular loops and potentially incapacitate LptD .\",\n \"The antibodies were used to target LptD in its native environment , when it is embedded in the bacteria .\",\n \"In parallel , mutant bacteria were created in which the loops were genetically removed one by one to assess their importance for LptD activity .\",\n \"The experiments revealed that although the antibodies could target most extracellular loops , they could not target the few loops that were essential for LptD to work properly .\",\n \"This suggests that antibody-accessible loops are expendable and that these structures could serve to shield other regions of LptD which are critical for survival .\",\n \"The findings will help to prioritise research that develops other approaches to inhibit LptD .\",\n \"Finally , the antibody workflow designed by Storek et al . can serve as a road map to study other membrane proteins in their native cellular environment .\"\n]","isConversational":false},"year":{"kind":"string","value":"2019"}}},{"rowIdx":94,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["developmental biology","neuroscience"],"string":"[\n \"developmental biology\",\n \"neuroscience\"\n]","isConversational":false},"title":{"kind":"string","value":"The Drosophila Sp8 transcription factor Buttonhead prevents premature differentiation of intermediate neural progenitors"},"id":{"kind":"string","value":"elife-03596-v2"},"sections":{"kind":"list like","value":[["Intermediate neural progenitor cells ( INPs ) play a critical role in increasing the brain size and complexity .","Transient amplification of INPs dramatically boosts the neural output from neural stem cells ( NSCs ) ( Kriegstein and Alvarez-Buylla , 2009; Florio and Huttner , 2014 ) .","Recent studies in developing human brains as well as other mammalian brains suggest that an expansion of the number of transiently amplifying INPs , the outer sub-ventricular zone radial glia-like cells ( oRGs ) , likely contributes to the increased cortical size and complexity in humans and other gyrencephalic animals ( Fietz et al . , 2010; Hansen et al . , 2010; Lui et al . , 2011; Wang et al . , 2011 ) .","On the other hand , accumulating body of evidence suggests that brain tumors could originate from dedifferentiation and unrestricted proliferation of INPs ( Holland et al . , 2000; Dai et al . , 2001; Walton et al . , 2009; Persson et al . , 2010; Zong et al . , 2012 ) .","Therefore , it is fundamentally important to understand how the generation and proliferation of INPs are regulated .","The recently discovered type II neuroblasts ( NBs , the Drosophila NSCs ) in developing Drosophila larval brains provide an excellent model system for studying mechanisms regulating the generation and proliferation of INPs ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .","There are 8 type II NBs in each brain lobe .","Like mammalian NSCs , Drosophila type II NBs produce neurons and glia indirectly by generating INPs .","Individual INPs undergo 4–6 rounds of asymmetric divisions to produce a new INP to self-renew and a ganglion mother cell ( GMC ) , which divides terminally to produce neurons and/or glia ( Bayraktar et al . , 2010; Viktorin et al . , 2011; Yang et al . , 2013 ) .","Meanwhile , individual INPs produce distinct types of neurons by sequentially expressing a set of distinct transcription factors to specify the identity of their progeny ( Bayraktar and Doe , 2013; Wang et al . , 2014 ) .","Through self-renewing divisions , INPs not only amplify the number but also increase the diversity of neural progeny generated from type II NBs .","Therefore , the neurogenesis pattern in type II NB lineages is remarkably similar to that in mammalian brains and the Drosophila INPs are functionally analogous to mammalian INPs , particularly oRGs .","The generation of INPs in type II NB lineages involves multiple steps ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .","Newly generated INPs are immature and do not express any NB markers , such as the proneural protein Asense ( Ase ) or the bHLH protein Deadpan ( Dpn ) , except for Miranda ( Mira ) .","The Ase− immature INPs first turn on the expression of Ase to become Ase+ immature INPs .","Ase+ immature INPs then further differentiate to become mature INPs , which express both Ase and Dpn .","INPs do not divide until they are fully mature .","The maturation of INPs requires Numb , the NHL family protein Brain tumor ( Brat ) , the transcription factor Earmuff ( Erm ) , as well as the BAP and Histone deacetylase 3 ( HDAC3 ) chromatin remodeling complexes ( Bowman et al . , 2008; Weng et al . , 2010; Eroglu et al . , 2014; Koe et al . , 2014 ) .","Both Numb and Brat are segregated into Ase− immature INPs during the division of type II NBs to prevent them from dedifferentiating into NB fate , but they function through independent pathways .","Numb inhibits Notch activity in Ase− immature INPs , whereas Brat likely antagonizes the activity of the EGR family transcription factor Klumpfuss ( Klu ) and Armadillo/β-Catenin in Ase− immature INPs ( Bowman et al . , 2008; Komori et al . , 2014 ) .","Erm functions together with BAP and HDAC3 chromatin remodeling complexes after Brat and Numb to further restrict the developmental potential of INPs by attenuating the response of INPs to self-renewal factors such as Klu and Dpn ( Janssens et al . , 2014; Koe et al . , 2014 ) .","In addition , the BAP chromatin remodeling complex limits the self-renewal of INPs by activating the expression of Prdm protein Hamlet ( Eroglu et al . , 2014 ) .","In the absence of Numb , Brat , Erm , or chromatin remodeling complexes , INPs dedifferentiate into type II NBs and initiate tumorigenic overproliferation ( Bowman et al . , 2008; Weng et al . , 2010; Eroglu et al . , 2014; Koe et al . , 2014 ) .","Therefore , these proteins are critical to prevent dedifferentiation of INPs .","However , despite the significant progress on elucidating mechanisms that promote maturation and prevent dedifferentiation of INPs in the past few years , much less is known about why only type II NBs produce self-renewing INPs but not the type I NBs , which produce neurons by generating terminally dividing GMCs .","One major difference between INPs and GMCs is that INPs divide to self-renew whereas GMCs divide terminally ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008; Yang et al . , 2013 ) .","Therefore , in addition to avoiding dedifferentiation and unrestricted tumorigenic overproliferation , INPs need to overcome another challenge–to avoid over-differentiation and cell cycle exit–in order to maintain their progenitor state and self-renewal while they differentiate to mature and undergo self-renewing divisions .","Type II NBs and newly born Ase− immature INPs differ from type I NBs and GMCs by the lack of the expression of Ase and the homeodomain protein Prospero ( Pros ) ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .","In type I NB lineages , Pros is expressed in the cytoplasm in the NBs and translocates to the nucleus in GMCs to promote differentiation and cell cycle exit by inhibiting NB self-renewing genes and activating neural differentiation genes ( Li and Vaessin , 2000; Choksi et al . , 2006 ) .","In type II NB lineages , Pros is not expressed in the NB or Ase− immature INPs .","In Ase+ immature INPs and mature INPs , Pros is expressed at low levels in the cytoplasm ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .","It has been demonstrated that the lack of Ase and Pros in type II NBs and Ase− immature INPs is essential for the generation of self-renewing INPs in type II NB lineages .","Forced expression of Ase or Pros in type II NBs and their progeny is sufficient to eliminate INPs although removing Ase or Pros in type I NBs does not change the identity of type I NBs or induce the generation of INPs ( Bowman et al . , 2008; Bayraktar et al . , 2010; Zhu et al . , 2012 ) .","Our recent studies demonstrated that the Ets family transcription factor Pointed P1 ( PntP1 ) suppresses Ase in type II NBs and is required for the generation of INPs ( Zhu et al . , 2011 ) .","However , mechanisms that prevent premature differentiation of INPs and/or inhibit Pros expression in type II NBs and immature INPs are not known .","In this study , we investigate the role of the Sp family transcription factor Buttonhead ( Btd ) in type II NB lineage development .","Btd is a homolog of mammalian Sp8 ( Treichel et al . , 2003; Estella and Mann , 2010 ) .","In developing mammalian brains , Sp8 is expressed in neural progenitor cells to regulate forebrain patterning and interneuron development ( Griesel et al . , 2006; Waclaw et al . , 2006; Sahara et al . , 2007; Li et al . , 2011 ) .","In Drosophila embryos , Btd is required for the formation of specific head segments and NB formation ( Wimmer et al . , 1993; Younossi-Hartenstein et al . , 1997 ) .","In addition , Sp8/Btd also promotes the growth of limbs and other appendages ( Estella et al . , 2003; Treichel et al . , 2003; Kawakami et al . , 2004; Estella and Mann , 2010 ) .","In this study , we report that Btd is expressed in type II NB lineages to prevent premature differentiation of INPs into GMCs by suppressing Pros expression in immature INPs .","We also demonstrate that PntP1 and Btd function cooperatively to specify type II NB lineages and promote the generation of INPs ."],["Our recent studies demonstrated that the Ets family transcription factor PntP1 is specifically expressed in type II NB lineages to promote the generation of INPs .","However , although forced expression of PntP1 suppress Ase expression in nearly all type I NB lineages , it induces the generation of INP-like cells only in a subset of type I NBs ( Zhu et al . , 2011 ) .","Therefore , it is likely that other protein ( s ) may function together with PntP1 to specify type II NB lineages and promote the generation of INPs .","A recent functional genomic study showed that in addition to PntP1 , there are other nine genes that are highly expressed in brain tumors derived from type II NB lineages ( Carney et al . , 2012 ) .","We wondered whether any of these genes could function together with PntP1 to promote INP generation .","To test this idea , we first examined how knockdown of these genes would affect INP generation in type II NB lineages .","A normal type II NB lineage contains 2–3 Ase− immature INPs , 2–3 Ase+ immature INP , and about 20–30 ( 26 . 9 ± 4 . 1 , mean ± SD ) Ase+ Dpn+ mature INPs ( Figure 1A–A′ , C–C′ , G ) .","Interestingly , RNAi knockdown of Btd using the type II NB lineage-specific pntP1-GAL4 ( named as GAL414−94 previously ) ( Zhu et al . , 2011 ) as a driver led to a complete elimination of mature INPs in about 50% of type II NB lineages ( Figure 1B–B′ , G–H ) .","Instead , only a few ( 3 . 7 ± 1 . 2 ) Ase+ Dpn− cells were observed next to the Ase− immature INPs ( Figure 1B–B′ ) .","However , type II NBs remain Ase− as normal type II NBs ( Figure 1B–B′ ) , suggesting that the identity of the type II NBs was not affected by Btd RNAi knockdown . 10 . 7554/eLife . 03596 . 003Figure 1 . Loss of Btd eliminates mature INPs in type II NB lineages .","( A–A′ , C–C′ )","Wild-type type II NB lineages in a third instar larval brain .","mCD8-GFP driven by pntP1-GAL4 labels all type II NB lineages ( A–A′ ) or a single type II NB clone ( C–C′ ) .","Ase− immature INPs , Ase+ immature INPs , and mature INPs are indicated by open arrows , solid arrows , and arrowheads , respectively .","( B–B′ , D–F′ )","Btd RNAi knockdown type II NB lineages ( B–B′ ) or type II NB clones homozygous mutant for btdXG81 ( D–E′ ) or btdXA ( F–F′ ) in 3rd instar larval brains produce Ase− immature INPs ( open arrows ) and a few Ase+ daughter cells ( arrows ) but no mature INPs .","Only 3 out of total 8 type II NB lineages are shown in ( A–A′ ) and ( B–B′ ) .","In this and all other figures , asterisks indicate type II NBs and scale bars equal to 20 µm .","Dpn staining alone shows the NB and mature INPs .","( G–H )","Quantifications of the number of mature INPs ( G ) and the percentage of type II NB lineages with mature INPs ( H ) in the wild type , Btd RNAi knockdown , and btd mutant type II NB lineages .","The numbers on top of each bar are the numbers of type II NB lineages analyzed except for the numbers for the wt and btd RNAi in ( H ) , which are the number of brain lobes examined .","The mean and stdev for btdXG81 and btdXA in ( H ) are calculated by bootstrapping .","**p < 0 . 01 , *p < 0 . 05 ( Student t test ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 00310 . 7554/eLife . 03596 . 004Figure 1—figure supplement 1 . Expression of mouse Sp8 ( mSp8 ) .","( A–A′′ ) and Drosophila Btd ( B–B′′ ) rescues the loss of mature INPs in btd mutant type II NB clones .","Multiple mature INPs are observed in btd mutant type II NB clone that expresses UAS-mSp8 ( A–A′′ ) or UAS-btd ( B–B′′ ) .","Type II NBs , Ase− immature INPs , Ase+ immature INPs , and mature INPs are indicated by asterisks , open arrows , solid arrows , and arrowheads , respectively . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 004 To confirm that the loss of INPs indeed results from the knockdown of Btd rather than off-target effects of UAS-Btd RNAi , we generated btd mutant type II NB clones using two loss-of-function alleles , btdXA and btdXG81 ( Wimmer et al . , 1993; Estella and Mann , 2010 ) .","Consistent with the Btd RNAi knockdown , all btdXG81 mutant and 90% of btdXA mutant type II NB clones failed to generate any mature INPs except for 4–6 Ase+ Dpn− cells ( Figure 1D–F′ , G–H ) .","Moreover , about 40% of btd mutant type II NBs ectopically express Ase , making them appear as type I NB lineages ( Figure 1E ) .","The loss of INPs resulting from the Btd RNAi knockdown and btd loss-of-function mutations suggests that Btd is required for the generation of INPs .","Remarkably , the loss of INPs in btd mutant clones can be similarly rescued by the expression of mouse Sp8 or Drosophila Btd ( Figure 1—figure supplement 1 ) , suggesting that mammalian Sp8 could have a conserved role in promoting the generation of transient amplifying INPs .","Since the loss of mature INPs occurred even when Ase was not ectopically expressed in btd mutant type II NBs , the loss of INPs is not primarily due to the ectopic Ase expression or transformation of type II NBs into type I NBs .","Therefore , we first focused our phenotypic analyses on lineages without the ectopic Ase expression in the NB .","We also used the btdXG81 allele for further mutant phenotypic analyses below , given that btdXG81 shows slightly stronger phenotypes than btdXA .","Why does the loss of Btd lead to the elimination of mature INPs ?","When mature INPs are eliminated in the absence of Btd , the type II NBs without the ectopic Ase expression still produce Ase− immature INPs and a few Ase+ Dpn− daughter cells .","In normal type II NB lineages , Ase+ Dpn− cells can be either Ase+ immature INPs or GMCs .","Therefore , three possible scenarios could happen when mature INPs are eliminated in the absence of Btd:","1 ) Ase− immature INPs differentiate into GMCs instead of Ase+ immature INPs;","2 ) Ase− immature INPs differentiate into Ase+ immature INPs , which then directly differentiate into neurons/glia without further dividing;","3 ) Ase− immature INPs differentiate into Ase+ immature INPs , which in turn differentiate into terminally dividing GMCs .","To distinguish these possibilities , we first wanted to determine if Ase− immature INPs still differentiate into Ase+ immature INP in the absence of Btd by examining the expression of INP specific marker R9D11-CD4-tdTomato and progenitor marker Miranda ( Mira ) in the Ase+ cells next to the Ase− immature INPs .","R9D11-CD4-tdTomato utilizes a DNA fragment R9D11 from the erm promoter to drive the expression of CD4-tdTomato ( Han et al . , 2011 ) .","In normal type II NB lineages , R9D11-CD4-tdTomato is first turned on in Ase+ immature INPs and becomes stronger as INPs mature ( Figure 2A–A′ ) , which is similar to R9D11-mCD8-GFP ( Zhu et al . , 2011 ) .","Mira is expressed in all NBs as well as INPs but not ( or very weakly ) in GMCs ( Figure 2—–figure supplement 1 ) .","In Btd RNAi knockdown type II NB lineages without mature INPs , we found that R9D11-CD4-tdTomato was expressed in Ase+ daughter cells next to the Ase− immature INPs but its overall expression was much weaker than that in normal type II NB lineages ( Figure 2B–B′ , I ) .","Consistently , Mira is also expressed in those Ase+ cells next to the Ase− immature INPs in the Btd RNAi knockdown type II NB lineages ( Figure 2—figure supplement 1 ) .","The expression of R9D11-CD4-tdTomato and Mira suggests that Ase− immature INPs still differentiate into Ase+ immature INP in the absence of Btd as in wild-type type II NB lineages ( Figure 2J ) . 10 . 7554/eLife . 03596 . 005Figure 2 . Loss of Btd results in ectopic nuclear Pros in immature INPs and premature differentiation of Ase+ immature INPs into GMCs .","( A–A′ )","R9D11-CD4-tdTomato is expressed in Ase+ immature INPs ( solid arrows ) and mature INPs ( arrowheads ) but not in Ase− immature INPs ( open arrows ) in a wild-type type II NB lineage .","( B–B′ )","R9D11-CD4-tdTomato remains expressed in Ase+ daughter cells ( solide arrows ) next to the Ase− immature INPs ( open arrows ) when mature INPs are eliminated by Btd RNAi knockdown .","( C–C′ ) pH3 is not detected in Ase+ immature INPs ( solid arrows ) in a wild-type type II NB lineage .","( D–E′ ) pH3 is expressed in Ase+ daughter cells ( solid arrows ) that are the furthest from the Ase− immature INPs ( open arrows ) in a Btd RNAi knockdown type II NB lineage without mature INPs ( D–D′ ) or btd mutant type II NB lineages ( E–E′ ) .","( F–F′ )","Nuclear Pros is not expressed in Ase− immature INPs ( open arrows ) in a wild-type type II NB lineage .","( G-H′ )","Nuclear Pros is ectopically expressed in both Ase− immature INPs ( open arrows ) and Ase+ cells ( solid arrows ) in Btd RNAi knockdown ( G–G′ ) or btd mutant ( H–H′ ) type II NB lineages .","( I ) Quantifications of relative overall expression levels of R9D11-CD4-tdTomato in wild-type ( A–A′ ) and Btd RNAi knockdown ( B–B′ ) type II NB lineages .","( J ) A diagram of neurogenesis patterns in type II NB lineages in the presence or absence of Btd . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 00510 . 7554/eLife . 03596 . 006Figure 2—figure supplement 1 . Mira expression in Btd RNAi knockdown type II NB clone .","( A–A′ )","Mira is expressed in INPs ( yellow arrows ) in a wild-type type II NB lineage and forms a basal crescent ( white arrows ) at metaphase .","( B–B′ )","In a Btd RNAi knockdown type II NB lineage without mature INPs , Mira is expressed in Ase+ daughter cells ( yellow arrows ) next to the Ase− immature INP ( open arrows ) but does not form a basal crescent at metaphase in the dividing Ase+ cell ( white arrows ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 006 Next we asked if Ase+ immature INPs differentiate into neurons/glia directly or GMCs in the absence of Btd .","GMCs express both Ase and nuclear Pros and divide terminally but do not form a Mira crescent while dividing .","If Ase+ immature INPs directly differentiate into neurons/glia , then all the Ase+ daughter cells should be Ase+ immature INPs and none of them should be dividing .","In contrast , if Ase+ immature INPs differentiate into GMCs , then some Ase+ daughter cells will become mitotically active and express nuclear Pros but will not form a Mira crescent at the metaphase or telophase .","Immunostaining with the mitotic marker phospho-histone 3 ( pH3 ) showed that unlike Ase+ immature INPs , which never become pH3-positive ( Figure 2C–C′ ) , some Ase+ daughter cells generated in both Btd RNAi knockdown type II NB lineages without mature INPs ( Figure 2D–D′ ) and btd mutant type II NB lineages became mitotically active ( Figure 2E–E′ ) .","However , the pH3 positive cells were always the furthest from the Ase− immature INPs among the Ase+ daughter cells ( Figure 2D–D′ ) , suggesting that the Ase+ daughter cells divide terminally like GMCs .","Consistently , unlike in mature INPs , which form a Mira crescent at metaphase ( Figure 2—figure supplement 1 ) , we did not observe any Mira crescents in the Ase+ daughter cells at metaphase ( Figure 2—figure supplement 1 ) .","The terminal division and the lack of the Mira crescent strongly argue that Ase+ immature INPs differentiate into GMCs in the absence of Btd .","To further confirm that late immature INPs differentiate into GMCs in the absence of Btd , we then examined the expression of nuclear Pros in the Ase+ daughter cells .","Nuclear Pros is a cell fate determinant of GMCs .","In normal type II NBs lineages , Pros is expressed in the cytoplasm of Ase+ immature INPs and mature INPs and in the nucleus of GMCs and post-mitotic neurons , but not in type II NBs or Ase− immature INPs ( Figure 2F–F′ ) .","If Ase+ immature INPs differentiate into GMCs in the absence of Btd , we expected that some of the Ase+ daughter cells express nuclear Pros .","Interestingly , immunostaining of Pros showed that nuclear Pros was expressed not only in all Ase+ daughter cells but also in Ase− immature INPs generated in Btd RNAi knockdown or btd mutant type II NB lineages ( Figure 2G–H′ ) .","Given that Pros promotes cell cycle exit and GMC differentiation and that forced expression of Pros is sufficient to eliminate INPs in type II NB lineages ( Li and Vaessin , 2000; Choksi et al . , 2006; Bayraktar et al . , 2010 ) , the ectopic expression of nuclear Pros in immature INPs resulting from the loss of Btd very likely promotes the premature differentiation of Ase+ immature INPs into GMCs and cell cycle exit , leading to the loss of mature INPs ( Figure 2J ) .","These results also reveal that it is Btd that is responsible for the suppression of Pros in immature INPs .","To determine if the ectopic expression of nuclear Pros in immature INPs is indeed responsible for the elimination of mature INPs in the absence of Btd , we next examined if reducing Pros expression was able to rescue the elimination of INPs in Btd RNAi knockdown or btd mutant type II NB lineages .","To reduce Pros expression , we either removed one wild-type copy of pros or knocked down Pros by RNAi in type II NB lineages .","Remarkably , the elimination of mature INPs resulting from the Btd RNAi knockdown was nearly fully rescued even just by removing one wild-type copy of pros ( Figure 3A–D′ , I–J ) .","Unlike Btd RNAi knockdown in wild-type background , which resulted in a completely elimination of mature INPs in about 50% of type II NB lineages ( Figure 1B–B′ , G–H , Figure 3B–B′ , I–J ) , knockdown of Btd in pros17 or pros10419 heterozygous mutant animals no longer led to an obvious loss of mature INPs ( Figure 3D–D′ , I–J , and data not shown ) , although type II NB lineages develop normally in pros17 or pros10419 heterozygous mutant animals ( Figure 3C–C′ , I–J , and data not shown ) .","Similarly , the loss of INPs in btd mutant clones was also largely rescued when btd mutant type II NB clones were generated in pros17 heterozygous mutant background ( Figure 3—figure supplement 1 ) . 10 . 7554/eLife . 03596 . 007Figure 3 . Reducing Pros rescues the elimination of INPs resulting from the loss of Btd .","( A–D′ )","The loss of INPs resulting from Btd RNAi knockdown is rescued in pros17 heterozygous mutant background .","Only two lineages are shown in each brain .","( A–A′ )","Wild-type type II NB lineages have multiple mature INPs .","( B–B′ )","Btd RNAi knockdown causes a loss of mature INPs .","( C–C′ )","Type II NB lineages in pros17 heterozygous mutant larvae produce a similar number of mature INPs as in wild-type larvae .","( D–D′ )","Btd RNAi knockdown no long leads to the loss of mature INPs in pros17 heterozygous mutant type II NB lineages .","( E–H′ )","Pros RNAi knockdown rescues the loss of INPs in btd mutant type II NB clones .","( E–E′ )","A wild-type type II NB clone has multiple mature INPs .","( F–F′ )","A btd mutant type II NB clone contains no mature INPs .","( G–G′ )","Pros RNAi knockdown causes overproliferation of mature INPs in a type II NB clone .","( H–H′ )","Pros RNAi knockdown rescues the loss of mature INPs in a btd mutant type II clone .","Arrowheads point to mature INPs in all images .","( I–L )","Quantifications of the number of mature INPs ( I–K ) and the percentage of lineages with mature INPs ( J–L ) for the rescue of Btd RNAi knockdown phenotypes in pros17/+ larvae ( I–J ) or the rescue of btd mutant phenotypes by Pros RNAi knockdown ( K–L ) .","The samples sizes on top of each bar represent the number of type II NB lineages ( I , K , L ) or the number of brain lobes ( J ) .","The mean and stdev in ( L ) are calculated by bootstrapping .","**p < 0 . 01 , *p < 0 . 05 ( Student t test ) .","NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 00710 . 7554/eLife . 03596 . 008Figure 3—figure supplement 1 . Reducing Pros rescues the elimination of INPs resulting from the loss of Btd .","( A–D′ )","The loss of INPs resulting from Btd RNAi knockdown is rescued by Pros RNAi knockdown .","( A–A′ )","Wide-type type II lineages labeled with mCD8-GFP driven by insc-GAL4 have multiple mature INPs .","( B–B′ )","Btd RNAi knockdown driven by insc-GAL4 eliminates mature INPs in a subset of type II NB lineages .","( C–C′ )","RNAi knockdown of Pros results in overproliferation of mature INPs .","( D–D′ )","Simultaneous knockdown of Btd and Pros leads to a similar overproliferation of mature INPs in type II NB lineages as Pros RNAi knockdown alone .","( E–H′ )","The loss of mature INPs in btd mutant clones is rescued in pros17 heterozygous mutant larvae .","( E–E′ )","A wild-type type II NB clone has multiple mature INPs .","( F–F′ )","A btd mutant type II NB clone contains no mature INPs .","( G–G′ ) pros17 heterozygous mutant type II NB clone contains multiple mature INPs as the wild-type type II NB clone ( A–A′ ) .","( H–H′ )","A btd mutant type II NB clone generated in pros17 heterozygous mutant larvae has multiple mature INPs .","Asterisks indicate NBs and arrowheads point to mature INPs in all images .","( I–L )","Quantifications of the number of mature INPs ( I–K ) or the percentage of type II NB lineages with mature INPs ( J–L ) for the rescue of Btd RNAi knockdown phenotypes by Pros RNAi knockdown ( I–J ) or the rescue of btd mutant phenotypes in pros17 heterozygous mutant larvae ( K–L ) .","The sample size on top of each bar represents the number of lineages ( I , K , L ) or the number of brain lobes ( J ) .","The mean and stdev in ( L ) are calculated by bootstrapping .","** , p < 0 . 01; NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 008 Consistent with the rescue in pros heterozygous mutant animals , Pros RNAi knockdown also rescued the loss of INPs resulting from the loss of Btd ( Figure 3E–H′ , K–L , Figure 3—figure supplement 1 ) .","Pros RNAi knockdown led to overproliferation of mature INPs as observed in pros mutant type II NB clones ( Figure 3G–G′ , K , Figure 3—figure supplement","1 ) ( Bowman et al . , 2008 ) .","When Pros was knocked down in btd mutant type II NB clones , mature INPs were rescued in all btd mutant clones ( Figure 3H–H′ , L ) .","In about 70% of btd mutant type II NB clones , Pros RNAi knockdown led to a similar mature INP overproliferation as in wild-type clones .","In other 30% of btd mutant clones , Pros RNAi knockdown partially or fully rescued mature INPs without causing the overproliferation of mature INPs .","Similarly , Pros RNAi knockdown also rescued the loss of mature INPs resulting from Btd RNAi knockdown ( Figure 3—figure supplement 1 ) .","Results from these rescue experiments demonstrate that the ectopic nuclear Pros in immature INPs is indeed responsible for the loss of mature INPs .","Interestingly , removing one wild-type copy of pros or knocking down Pros not only rescued the loss of mature INPs but also suppressed the ectopic Ase expression in all btd mutant type II NBs ( Figure 3H , Figure 3—figure supplement 1 ) , suggesting that the ectopic Ase expression in btd mutant type II NBs also results from the ectopic expression of nuclear Pros in immature INPs .","Our results showed that Btd is required to suppress Pros in Ase− immature INPs .","We next asked if Btd is also required to partially suppress Pros at later stages of INP development .","In normal type II NB lineages , Pros is absent in Ase− immature INPs but is expressed at low levels in the cytoplasm of Ase+ immature INPs and mature INPs ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .","Maintaining the expression of Pros at low levels is essential for the self-renewal of INPs ( Bayraktar et al . , 2010 ) .","However , the complete elimination of mature INPs makes it difficult to assess the role of Btd in mature INPs .","Therefore , we used erm-GAL4 ( III ) and erm-GAL4 ( II ) to knock down Btd .","Both erm-GAL4 ( III ) and erm-GAL4 ( II ) are expressed in Ase+ immature INPs and mature INPs , whereas erm-GAL4 ( II ) is also expressed in Ase− immature INPs except for the newly born Ase− immature INPs ( Xiao et al . , 2012 ) .","However , knockdown of Btd using either erm-GAL4 ( III ) ( Figure 4A–B′′ , E ) or erm-GAL4 ( II ) ( Figure 4C–D′′ , F ) did not result in any obvious loss of mature INPs in type II NB lineages .","In line with these RNAi knockdown results , we were able to recover multicellular btd mutant INP clones that were comparable to wild-type INP clones ( Figure 4—figure supplement","1 ) while we generated btd mutant type II NB clones , indicating that btd mutant INPs were still able to divide multiple rounds like wild-type INPs and did not prematurely differentiate into GMCs .","These data suggest that Btd likely suppresses Pros expression only in newly born Ase− immature INPs but not in immature INPs at later developmental stages or mature INPs . 10 . 7554/eLife . 03596 . 009Figure 4 . Knockdown of Btd in immature or mature INPs by erm-GAL4 lines does not lead to the loss of mature INPs .","( A–A′′ , C–C′′ )","Wild-type type II NB lineages are labeled with mCD8-GFP driven by erm-GAL4 ( III ) ( A–A′′ ) or erm-GAL4 ( II ) ( C–C′′ ) .","( B–B′′ , D–D′′ )","Knockdown of Btd in Ase+ immature INPs and mature INPs by erm-GAL4 ( III ) ( B–B′′ ) or in Ase− immature INPs as well as Ase+ immature INP and mature INPs by erm-GAL4 ( II ) ( D–D′′ ) does not cause a reduction of the number of mature INPs .","Only two lineages are shown in each brain .","( E–F )","Quantifications of the number of mature INPs in type II NB lineages in which Btd is knocked down by erm-GAL4 ( III ) ( E ) or erm-GAL4 ( II ) ( F ) .","NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 00910 . 7554/eLife . 03596 . 010Figure 4—figure supplement 1 . Btd likely does not function in mature INPs .","( A ) A wild-type INP clone with 4 post-mitotic cells .","( B ) A btd mutant INP clone with 6 post-mitotic cells .","INP clones labeled with mCD8-GFP are outlined by dashed circles .","INP clones are identified as clones that have more than two cells but no NBs . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 010 Our btd mutant MARCM analyses showed that in addition to the loss of INPs , Ase was ectopically expressed in about 40% of btd mutant type II NBs .","We showed previously that PntP1 is expressed in type II NBs as well as Ase− and Ase+ immature INPs ( Figure 5A–A′′ ) ( Zhu et al . , 2011 ) .","Inhibiting PntP1 activity results in ectopic Ase expression in type II NBs and elimination of INPs ( Zhu et al . , 2011 ) .","Therefore , we wondered if PntP1 expression was reduced or even lost in btd mutant type II NBs .","Immunostaining of PntP1 showed that PntP1 was expressed at reduced levels in most btd mutant type II NB lineages without the ectopic Ase expression ( Figure 5B–B′ ) .","The reduction is about 10% in the NBs and 50% in the immature INPs ( Figure 5E–F ) .","In those btd mutant clones , PntP1 was also detected in the Ase+ daughter cells next to the Ase− immature INPs ( Figure 5B–B′ ) , providing additional evidence to support that Ase− immature INPs still differentiate into Ase+ immature INPs in the absence of Btd .","However , in btd mutant type II NB lineages with the ectopic Ase expression in the NB , PntP1 was largely abolished in both the NBs and their progeny ( Figure 5C–C′ , E–F ) .","The correlation of the ectopic Ase expression in the NB and the severe reduction or loss of PntP1 suggesting that the ectopic Ase expression in btd mutant type II NBs could result from the severe reduction or loss of PntP1 expression . 10 . 7554/eLife . 03596 . 011Figure 5 . PntP1 expression is reduced in btd mutant type II NB clones .","( A–A′′ )","PntP1 is expressed in the NB ( * ) , Ase− immature INPs ( open arrows ) , as well as Ase+ immature INPs ( solid arrows ) in a wild-type type II NB clone .","( B–B′′ )","PntP1 expression is much lower in a btd mutant type II NB clone without the ectopic Ase expression in the NB than that in a neighboring btd heterozygous type II NB lineage .","The reduction is particularly obvious in the Ase− immature INPs ( open arrows ) .","Note that PntP1 remains expressed in Ase+ daughter cells ( arrows ) next to the Ase− immature INPs .","( C–C′′ )","PntP1 expression is largely abolished in a btd mutant type II NB clone with the ectopic Ase expression in the NB .","In a neighboring btd heterozygous type II NB lineages , PntP1 is still detected in the NBs ( * ) , Ase- immature INPs ( open arrows ) and Ase+ immature INPs ( arrows ) .","( D–D′′ )","Knocking down Pros restores the expression of PntP1 in the NB ( * ) , Ase− immature INPs ( open arrows ) , and Ase+ immature INPs ( arrows ) in a btd mutant clone to levels comparable to those in a neighboring btd heterozygous type II NB lineage .","Wild-type ( A–A′′ ) or btd mutant type II NB clones ( B–D′′ ) are outlined by dashed lines and neighboring btd heterozygous type II NB lineages ( B–D′′ ) are marked with dotted lines .","( E–F )","Quantifications of PntP1 expression levels in type II NBs ( E ) and Ase− immature INPs ( F ) in btd mutant type II NB clones relative to neighboring type II NB lineages in the same brains .","**p < 0 . 01; NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01110 . 7554/eLife . 03596 . 012Figure 5—figure supplement 1 . The reduction of PntP1 is unlikely responsible for the loss of INPs in btd mutant type II NB clones .","( A–A′′ )","A btd mutant type II NB clone has similar expression levels of PntP1 as a neighboring btd heterozygous mutant type II NB lineage but fails to produce mature INPs .","( B–B′′ )","Expressing UAS-PntP1 does not rescue the loss of mature INPs in a btd mutant clone , even though the PntP1 expression in the btd mutant clone is much higher than that in a neighboring btd heterozygous mutant type II NB lineage .","( C–C′′ )","Nuclear Pros remains ectopically expressed in Ase− immature INPs and Ase+ daughter cells in a btd mutant type II NB clone expressing UAS-PntP1 .","( D–D′′ )","Mature INPs are partially rescued by UAS-PntP1 in a btd mutant type II NB clone .","In all images , btd mutant clones are outlined by dashed lines and their neighboring btd heterozygous mutant type II NB lineages are marked by dotted lines .","Open arrows: Ase− immature INPs; solid arrows: Ase+ immature INPs; arrowheads: mature INPs . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 012 To determine if the reduction/loss of PntP1 is responsible for the ectopic Ase expression and/or the loss of INPs in btd mutant type II NB lineages , we examined if restoring PntP1 expression was sufficient to suppress ectopic Ase expression and/or rescue the loss of INPs resulting from the loss of Btd by expressing UAS-pntP1 in btd mutant type II NB clones .","Our results showed that expressing UAS-pntP1 resulted in higher expression of PntP1 in btd mutant type II NB clones than that in neighboring btd heterozygous mutant type II NB lineages and suppressed the ectopic Ase expression in all btd mutant type II NBs ( n = 12 ) ( Figure 5—figure supplement 1 ) .","However , unlike reducing Pros expression , which rescued mature INPs in nearly all btd mutant type II NB clones ( Figure 3 , Figure 3—figure supplement 1 ) , expressing UAS-pntP1 failed to rescue mature INPs or suppress the ectopic nuclear Pros in Ase− immature INPs or Ase+ daughter cells in the majority of btd mutant clones ( Figure 5—figure supplement 1 ) .","Only in 3 out of total 10 btd mutant clones expressing UAS-PntP1 , we observed that mature INPs were partially rescued to 9 . 3 ± 3 . 2 per lineages ( Figure 5—figure supplement 1 ) , which is still much fewer than the number of mature INPs ( 20–30 per lineages ) in normal type II NB lineages .","Consistent with the inability of UAS-pntP1 to fully rescue the loss of mature INPs , we found occasionally that btd mutant clones that did not show an obvious reduction of PntP1 in either the NBs or early immature INPs still failed to generate any mature INPs ( Figure 5—figure supplement 1 ) .","Therefore , these results demonstrate that the severe reduction/loss of PntP1 accounts for the ectopic Ase expression in btd mutant type II NBs but is not the primary reason for the loss of mature INPs .","Given that reducing the expression of Pros suppressed the ectopic Ase expression in btd mutant type II NBs , we then asked if reducing Pros expression could also rescue the reduction/loss of PntP1 in btd mutant type II NB clones .","Indeed , consistent with the suppression of the ectopic Ase expression by Pros RNAi knockdown , PntP1 expression in both the NBs and immature INPs returned to normal levels when the loss of INPs was rescued by Pros RNAi knockdown in btd mutant type II NB clones ( Figure 5D–F ) .","These results suggest that the reduction/loss of PntP1 in btd mutant type II NB clones is due to the ectopic Pros expression in immature INPs .","However , given that ectopic nuclear Pros is only observed in immature INPs but not in the NBs in btd mutant type II NB clones , the reduction/loss of PntP1 and the subsequent ectopic Ase expression in the NB is most likely a secondary effect of the ectopic nuclear Pros expression in immature INPs .","Our Btd loss of function analyses demonstrated that Btd is critical for the generation of INPs in type II NBs lineages .","We next examined if Btd is only expressed type II NB lineages .","Since Btd antibodies are not available and our in situ hybridization signals of btd mRNAs in the central brain were barely detectable ( data now shown ) , we used the btd-GAL4 as a reporter for btd expression .","btd-GAL4 is a GAL4 enhancer trap line , in which the GAL4 transgene is inserted at 753bp upstream of the transcription start site of btd ( Estella et al . , 2003 ) .","btd-GAL4 shows similar expression patterns as endogenous Btd in ventral imaginal discs ( Estella et al . , 2003 ) .","We found that mCD8-GFP driven by the btd-GAL4 is expressed in all type II NB lineages but not type I NB lineages on the dorsal side of larval brains ( Figure 6A–A′ ) .","In type II NB lineages , the expression of mCD8-GFP driven by btd-GAL4 is detected in the NB but becomes much stronger in immature INPs next to the NBs ( Figure 6A ) .","The expression of mCD8-GFP then becomes progressively weak in cells away from the NBs and is barely detectable in some mature INPs distal from the NB ( Figure 6A–A′ ) .","The expression pattern of btd-GAL4 in type II NB lineages is similar to that of pntP1-GAL4 ( e . g . Figure 1A–A′ ) and is consistent with our results that Btd mainly functions in immature INPs . 10 . 7554/eLife . 03596 . 013Figure 6 . Btd is expressed in type II NB lineages and a subset of type I NB lineages .","( A–A′ ) mCD8-GFP driven by btd-Gal4 is expressed in all type II NB lineages ( outlined by dashed lines ) but not type I NB lineages ( e . g . arrows ) on the dorsal side of a 3rd instar larval brain .","The expression of mCD8-GFP becomes progressively weak in cell away from the NB .","Some mature INPs ( e . g . arrowheads ) distant from the NB have no obvious expression of mCD8-GFP .","Only seven out of total eight type II NB lineages are shown in this particular focal plane .","( B–B′ )","Two type I NB lineages are labeled by mCD8-GFP driven by btd-GAL4 on the ventral side of a 3rd instar larval brain .","( C–C′ ) mCD8-GFP driven by btd-Gal4 labels a subset of type I NB lineages ( e . g . arrows ) in the ventral nerve cord ( VNC ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01310 . 7554/eLife . 03596 . 014Figure 6—figure supplement 1 . The GAL4 insertion in the btd-GAL4 line does not affect type II NB lineage development .","( A–A′′ )","A wild-type type II NB clone contains multiple mature INP .","( B–B′′ )","A btd-GAL4 mutant type II NB clone has a similar number of mature INPs as wild-type type II NB clone .","( C ) Quantifications of the number of mature INPs in wild-type and btd-GAL4 mutant type II NB clones .","Arrowheads: mature INPs; NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01410 . 7554/eLife . 03596 . 015Figure 6—figure supplement 2 . The loss of INP phenotype resulting from Btd RNAi knockdown is rescued by the expression of mouse Sp8 ( mSp8 ) or Drosophila Btd .","( A–A′′ )","Btd RNAi knockdown driven by btd-GAL4 completely eliminates mature INPs in all type II NB lineages in a 3rd instar larval brain .","( B–B′′ )","The expression of UAS-mSp8 driven by btd-GAL4 fully rescues the loss of mature INPs in all type II NB lineages resulting from Btd RNAi knockdown .","( C–C′′ )","The expression of UAS-btd driven by btd-GAL4 partially rescues the loss of mature INPs in type II NB lineages resulting from Btd RNAi knockdown .","Arrowheads point to mature INPs in all images .","( D–E )","Quantifications of the percentage of type II NB lineages with mature INPs ( D ) and the number of mature INPs ( E ) in 3rd instar larvae with indicated genotypes .","Sample sizes represent the number of brain lobes ( D ) or the number of type II NB lineages ( E ) .","* , p < 0 . 05; ** , p < 0 . 01; NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 015 In addition to type II NB lineages , mCD8-GFP driven by btd-GAL4 is also expressed in two type I NB lineages on the ventral side of larval brains as well as about 31 type I NB lineages in the ventral nerve cord ( VNC ) ( Figure 6B–C′ ) .","In those type I NB lineages , mCD8-GFP driven by btd-GAL4 is expressed in the NBs , GMCs , and newly born neurons .","The expression pattern of btd-GAL4 in single type I NB lineages is similar to those of other NB GAL4 lines such as insc-GAL4 ( e . g . Figure 7D ) , suggesting that Btd likely functions in the NB in these type I NB lineages . 10 . 7554/eLife . 03596 . 016Figure 7 . Btd and PntP1 function cooperatively to promote the generation of INPs .","( A–C′ )","Ectopic expression of PntP1 consistently promotes the generation of INP-like cells in Btd-positive type I NB lineages in larval brains .","( A–A′ )","A wild-type type I NB lineage labeled by btd-GAL4 has no INP-like cells .","( B–C′ )","The ectopic expression of PntP1 driven by btd-GAL4 suppresses Ase in the NB ( * ) and induces the generation of Ase+ Dpn+ INP-like cells ( arrowheads ) ( B–B′ ) , which also express INP-specific marker R9D11-CD4-tdTomato ( C–C′ ) .","( D–E′ )","The expression of PntP1 driven by tub-GAL4 suppresses Ase in both wild-type ( D ) and btd mutant ( E ) type I NBs ( * ) but only induced the generation of INP-like cells in the wild-type type I NB clone ( D–D′ ) but not in the btd mutant clone ( E–E′ ) .","( F–I′ ) insc-GAL4 drives the expression of UAS-mCD8-GFP alone ( F–F′ ) or together with UAS-btd ( G–G′ ) , UAS-pntP1 ( H–H′ ) , or UAS-btd plus UAS-pntP1 ( I–I′ ) in type I NB lineages .","Images are from the ventral side of larval brains , where only type I NB lineages are observed in wild-type animals ( F–F′ ) .","The expression of UAS-btd ( G–G′ ) or UAS-pntP1 ( H–H′ ) alone promotes the generation of INP-like cells only in small subset of type I NB lineages ( dashed circles ) .","The expression of Btd only suppresses/reduces Ase expression in type I NBs ( arrows ) that produce INP-like cells ( G–G′ ) but not in other type I NBs ( arrowheads ) , where PntP1 suppresses Ase in nearly all type I NBs ( e . g . arrows ) regardless of the generation of INP-like cells ( H–H′ ) .","Co-expression of Btd and PntP1 promotes the generation of INP-like cells nearly in all type I NB lineages ( dashed circles ) ( I–I′ ) .","( J–J′ )","Quantifications of the percentage of type I NB lineages that produce INP-like cells in larvae with indicated genotypes .","The number on top of each bar represents the number of brain lobes except for numbers for the expression of UAS-pntP1 driven by tub-GAL4 , which are the number of clones .","The mean and stdev for the percentage of wild-type or btd mutant clones expressing UAS-pntP1 driven by tub-GAL4 are calculated by bootstrapping .","**p < 0 . 01 . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01610 . 7554/eLife . 03596 . 017Figure 7—figure supplement 1 . Btd and PntP1 function cooperatively to induce the generation of INP-like cells in type I NB lineages in the VNC .","( A–A′ )","Type I NB lineages labeled by mCD8-GFP driven by btd-GAL4 in the VNCs have no INPs ( e . g . dashed circles ) .","( B–B′ )","Expression of PntP1 driven by btd-GAL4 induces the generation of INP-like cells ( small Dpn+ cells ) in nearly all Btd-positive lineages ( dashed circles ) .","( C–C′ )","Type I NB lineages labeled by mCD8-GFP driven by insc-GAL4 in the VNC have no INPs .","( D–D′ )","Expression of UAS-btd driven by insc-GAL4 neither suppresses Ase in the NB nor induces the generation of INP-like cells in type I NB lineages ( e . g . dashed circles ) in the VNC .","( E–E′ )","Expression of UAS-pntP1 driven by insc-GAL4 suppresses Ase in nearly all NBs ( * ) and induces the generation of INP-like cells in a subset of type I NB lineages ( dashed circles ) in the VNC .","( F–F′ )","Co-expression of UAS-btd and UAS-pntP1 driven by insc-GAL4 induces INP-like cells in almost all type I NB lineages ( dashed circles ) in the VNC .","( G ) Quantifications of the percentage of type I NB lineages with INP-like cells in brain lobes with indicated genotypes .","**p < 0 . 01 . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01710 . 7554/eLife . 03596 . 018Figure 7—figure supplement 2 . Overexpression of Btd promotes the generation of INP-like cells .","( A–A′′ )","INP specific marker R9D11-CD4-tdTomato is not expressed in any type I NB lineages ( e . g . arrows ) on the ventral side of a larval brain .","( B–B′′ )","Overexpression of Btd induces the generation of INP-like cells in a subset of type I NB lineages on the ventral side of a larval brain as indicated by the expression of R9D11-CD4-tdTomato ( e . g . arrows ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 018 In order to determine if the btd-GAL4 reflects the endogenous Btd expression pattern in the central brain , we tried to rescue the btd loss-of-function phenotypes by expressing UAS-mSp8 or UAS-btd driven by btd-GAL4 .","However , the GAL4 insertion in the btd-GAL4 line does not affect the type II NB lineage development although it causes a lethal mutation of btd ( Figure 6—figure supplement 1 ) .","Therefore , we tried to rescue Btd RNAi knockdown phenotypes by the expression of UAS-mSp8 or UAS-btd driven by btd-GAL4 instead .","The expression of UAS-btd RNAi driven by btd-GAL4 completely eliminated mature INPs in nearly all type II NB lineages ( Figure 6—figure supplement 2 ) , which was much stronger than the phenotype of Btd RNAi knockdown driven by pntP1-GAL4 .","As expected , the expression of UAS-mSp8 driven by btd-GAL4 fully rescued the loss of INPs resulting from Btd RNAi knockdown in all type II NB lineages ( Figure 6—figure supplement 2 ) .","Similarly , the expression of UAS-btd driven by btd-GAL4 partially rescued the loss of INPs in about 67% of lineages ( Figure 6—figure supplement 2 ) .","The incomplete rescue by UAS-btd is likely because UAS-btd contains the sequence targeted by UAS-btd RNAi .","The rescue of Btd RNAi knockdown phenotypes by the expression of UAS-mSp8 and UAS-btd driven by btd-GAL4 together with the strong loss of INP phenotypes in all btd mutant type II NB clones strongly argue that btd-GAL4 expression is likely consistent with the endogenous Btd expression pattern in the central brain .","Our previous studies showed that ectopic expression of PntP1 could induce the generation of INP-like cells in more type I NB lineages in VNCs than in larval brains ( Figure 7H–H′ , J , Figure 7—figure supplement","1 ) ( Zhu et al . , 2011 ) .","Since Btd is expressed in much more type I NB lineages in VNCs than in larval brains and Btd is required to prevent the premature differentiation of INPs , we wondered if ectopic PntP1-induced generation of INP-like cells requires Btd activity and occurs mostly in Btd-positive type I NB lineages .","To test this idea , we examined if co-expression of PntP1 and Btd in type I NB lineages was sufficient to induce the generation of INPs and if the ectopic PntP1-induced generation of INP-like cells would be impaired in the absence of Btd .","To coexpress PntP1 and Btd in type I NB lineages , we used either btd-GAL4 to drive the expression of UAS-pntP1 in Btd-positive type I NB lineages or insc-GAL4 to drive the expression of UAS-pntP1 and UAS-btd simultaneously in all type I NB lineages .","INP-like cells were identified by their expression of Ase and Dpn as well as INP-specific marker R9D11-CD4-tdTomato .","Since the GAL4 insertion in the btd-GAL4 line causes a lethal mutation in btd ( Estella et al . , 2003 ) , we examined the phenotype of the expression of UAS-pntP1 driven by btd-GAL4 only in btd-GAL4 heterozygous female larvae .","As shown in Figure 6A–A′ , type II NB lineages in btd-GAL4 heterozygous mutant larvae are indistinguishable from those in wild-type animals ( e . g . Figure 1A–A′ ) .","Furthermore , as mentioned above , btd-GAL4 homozygous mutant type II NB clones develop normally ( Figure 6—figure supplement 1 ) .","Therefore , the generation of INPs is not affected in the btd-GAL4 line .","Interestingly , ectopic expression of PntP1 using btd-GAL4 as a driver induced the generation of INP-like cells in about 90% Btd-positive type I NB lineages in both larval brains ( Figure 7A–C′ , J ) and VNCs ( Figure 7—figure supplement 1 ) .","Consistently , the co-expression of UAS-pntP1 and UAS-btd driven by insc-GAL4 induced INP-like cells in about 95% of type I NB lineages in both larval brains and VNCs ( Figure 7F–F′ , I–I′ , J , Figure 7—figure supplement 1 ) .","In contrast , the expression of UAS-pntP1 alone driven by insc-GAL4 only induced INP-like cells in about 10% and 46% of type I NB lineages in larval brains ( Figure 7H–H′ , J ) and VNCs ( Figure 7—figure supplement 1 ) , respectively , although ectopic PntP1 expression suppressed Ase in nearly all type I NBs ( Figure 7H , Figure 7—figure supplement 1 ) .","The expression of UAS-btd alone neither suppressed Ase nor induced the generation of INP-like cells in VNCs ( Figure 7—figure supplement 1 ) , whereas in larval brains , the expression of UAS-btd driven by insc-GAL4 suppressed/reduced the expression of Ase in the NB and promoted the generation of INP-like cells in about 20% of type I NB lineages ( Figure 7G–G′ , J ) .","These ectopic INP-like cells induced by the expression of UAS-btd also expressed INP-specific marker R9D11-CD4-tdTomato ( Figure 7—figure supplement 2 ) .","These results indicate that the generation of INP-like cells induced by the ectopic PntP1 expression requires Btd activity , whereas the expression of either PntP1 or Btd alone has limited ability to induce the generation of INP-like cells in type I NB lineages .","To further confirm if Btd is required for ectopic PntP1 to induce the generation of INP-like cells , we then examined if the generation of INP-like cells induced by ectopic PntP1 expression would be impaired in the absence of Btd .","To this end , we expressed UAS-pntP1 in wild-type or btd mutant type I NB clones in VNCs in that there are more Btd-positive type I NB lineages and the expression of UAS-pntP1 can induce INP-like cells in much more type I NB lineages in VNCs than in larval brains .","Consistent with the induction of INP-like cells in nearly all type I NB lineages when PntP1 and Btd were coexpressed , the efficiency of PntP1 to induce the generation of INP-like cell was drastically reduced in the absence of Btd .","Our results showed that the expression of UAS-pntP1 driven by tub-GAL4 could have induced the generation of INP-like cells in about 50% of wild-type type I NB clones but not in btd mutant type I NB clones , although the expression of PntP1 equally suppressed the expression of Ase in both wild-type and btd mutant type I NBs ( Figure 7D–E′ , J ) .","These results provide additional evidence to support that only in the presence of Btd could PntP1 induce the generation of INP-like cells in type I NB lineages .","Taken together , these results suggest that the generation of INPs requires the cooperative action of PntP1 and Btd .","Thus this study together with our previous work ( Zhu et al . , 2011 ) identified two key factors , PntP1 and Btd , a combination of which is sufficient to specify type II NB lineages and promote INP generation ."],["The most striking phenotype resulting from the loss of Btd is the elimination of mature INPs .","In addition , about 40% of btd mutant type II NB lineages ectopically express Ase in the NB and become type I-like NB lineages .","However , although forced expression of Ase in type II NBs is sufficient to eliminate INPs in type II NB lineages ( Bowman et al . , 2008; Zhu et al . , 2012 ) , the loss of INPs is obviously not primarily due to the ectopic Ase expression or the transformation of type II NB lineages into type I-like NB lineage in that the loss of mature INPs occurs independently of the ectopic Ase expression in most btd mutant or Btd RNAi knockdown type II NB lineages .","Instead , we demonstrate that the loss of mature INPs in the absence of Btd is due to the premature differentiation of Ase+ immature INPs into GMCs .","We show that in Btd RNAi knockdown or btd mutant type II NB lineages without the ectopic Ase expression , Ase− immature INPs differentiate into Ase+ immature INPs normally as indicated by the expression of R9D11-mCD8-GFP , Mira , as well as PntP1 in Ase+ daughter cells next to the Ase− immature INPs .","However , instead of differentiating into mature INPs , we argue that Ase+ immature INPs prematurely differentiate into GMCs based on the following two pieces of evidence .","First , Ase+ daughter cells eventually undergo terminal divisions as indicated by the positive pH3 staining and the position of the pH3-positive cells .","Second , unlike mature INPs , the dividing Ase+ daughter cells do not form basal Mira crescent at metaphase .","The terminal division and the lack of Mira crescent during the division are two unique features that distinguish GMCs from INPs in addition to the expression of nuclear Pros ( Ikeshima-Kataoka et al . , 1997; Matsuzaki et al . , 1998; Schuldt et al . , 1998 ) .","Therefore , the elimination of mature INPs resulting from the loss of Btd is due to the premature differentiation of Ase+ immature INPs into GMCs .","Why does the loss of Btd lead to premature differentiation of INPs ?","Our results show that the loss of Btd results in a reduction or loss of PntP1 in type II NBs and immature INPs as well as ectopic expression of Pros in early immature INPs .","Our previous studies show that PntP1 suppresses Ase in type II NBs and that inhibiting PntP1 activity leads to ectopic expression of Ase in type II NBs and elimination of INPs ( Zhu et al . , 2011 ) .","Given that the ectopic Ase expression in btd mutant type II NBs is closely associated with the severe reduction or complete loss of PntP1 and that expression of UAS-pntP1 largely suppresses the ectopic Ase expression in btd mutant type II NBs , the severe reduction or loss of PntP1 most likely accounts for the ectopic Ase expression in btd mutant type II NBs .","However , although the loss of PntP1 could lead to the loss of INPs , we provide several lines of evidence to demonstrate that the elimination of INPs in btd mutant or Btd RNAi knockdown type II NB lineages is primarily due to the ectopic activation of Pros in immature INPs rather than the reduction or loss of PntP1 .","First , ectopic nuclear Pros is consistently expressed in Ase− immature INPs when mature INPs are eliminated .","Second , the loss of mature INPs can be fully rescued by Pros RNAi knockdown or even just by removing one wild-type copy of pros .","Third , Pros RNAi knockdown also rescues the reduction of PntP1 and suppresses the ectopic Ase expression in btd mutant type II NBs .","In contrast , the expression of UAS-pntP1 fails to rescue mature INPs in most btd mutant type II NB lineages although it largely suppresses the ectopic Ase expression in the NBs .","Furthermore , the complete elimination of mature INPs is also observed occasionally in btd mutant type II NB lineages without the reduction of PntP1 .","Therefore , the elimination of mature INPs resulting from the loss of Btd is primarily due to the ectopic Pros expression , which likely promotes premature differentiation of INPs into GMCs and cell cycle exit .","The severe reduction or loss of PntP1 is responsible for the ectopic Ase expression in btd mutant type II NBs and is more likely a secondary effect due to the ectopic Pros expression and/or the loss of INPs .","INPs and/or other progeny may provide feedback signals to the NBs as has been demonstrated in other systems ( Yoon et al . , 2008; Hsu et al . , 2014 ) .","The ectopic expression of Pros in Ase− immature INPs resulting from the loss of Btd suggests that Btd is critical for suppressing Pros expression in Ase− immature INPs .","Btd was known as a head gap gene .","It has been suggested that gap factors act largely as transcriptional repressors ( Schroeder et al . , 2004 ) .","Btd could directly suppress Pros by binding to the pros promoter as a transcriptional repressor .","Alternatively , Btd could suppress Pros indirectly by regulating the expression or antagonizing the activity of factor ( s ) that activate ( s ) pros expression .","Our results show that ectopic/overly expression of Btd in type I NB lineages or mature INPs does not lead to overproliferation of type I NBs as observed in pros mutant type I NB lineages .","Instead , ectopic expression of Btd promotes the generation of INP-like cells from type I NBs and transforms some type I NB lineages into type II-like NB lineages .","Therefore , it is more likely that Btd suppresses Pros indirectly by regulating the expression or antagonizing the activity of pros activator ( s ) .","Previous studies have suggested that Ase , Daughterless , Numb , and Erm could activate pros expression ( Reddy and Rodrigues , 1999; Southall and Brand , 2009; Weng et al . , 2010; Yasugi et al . , 2014 ) .","Since Ase and R9D11-Cd4-tdTomato , which are under the control of erm promoter , are not expressed in Ase− immature INPs in the absence of Btd , it is unlikely that they are involved in the activation of pros in immature INPs .","It would be interesting to investigate in the future if Numb or Daughterless could activate pros in immature INPs in the absence of Btd .","In this study , we provided several lines of evidence to demonstrate that Btd and PntP1 function cooperatively to specify type II NB lineages and promote the generation of INPs .","Results from this study as well as our previous study ( Zhu et al . , 2011 ) show that ectopic expression of UAS-pntP1 or UAS-btd alone can only promote the generation of INP-like cells in a subset of type I NB lineage , whereas ectopic expression of UAS-pntP1 in Btd-positive type I NB lineages or co-expression of UAS-btd and UAS-pntP1 can promote the generation of INP-like cells in nearly all type I NB lineages and transforms all these lineages into type II-like NB lineages .","Consistently , the ability of PntP1 to promote the generation of INP-like cells in btd mutant type I NB lineages is largely impaired .","These results suggest that the specification of type II NB lineages and the generation of INPs requires both PntP1 and Btd and that the combinatorial PntP1 and Btd is sufficient to promote the generation of INPs .","We propose that PntP1 and Btd function cooperatively but through different mechanisms to promote INP generation .","PntP1 is responsible for the suppression of Ase in type II NBs .","Meanwhile , PntP1 must be regulating the expression of other unknown target gene ( s ) that are/is essential for the generation of INPs , such as specification of immature INPs , because loss of Ase is not sufficient to promote the generation of INP-like cells in any type I NB lineages .","Btd likely acts after PntP1 to mainly prevent premature differentiation of INPs into GMCs by indirectly suppressing pros in immature INPs .","The role of Btd in suppressing Ase in type II NBs is minimal if there is any because unlike PntP1 , which suppresses ase in nearly all type I NBs when it is ectopically expressed , overexpression of Btd only suppresses Ase in a small subset of type I NBs that produce INP-like cells in larval brains .","Furthermore , Ase is expressed in Btd+ type I NBs , indicating that Btd does not suppress Ase in type I NBs when it is expressed at normal levels .","Studies in mammals as well as in Drosophila suggest that the Btd/Sp8 could function downstream of Wnt signaling to regulate the expression of Fgf8 as well as Distal-less ( Dll ) and Headcase ( Hdc ) during the forebrain patterning as well as limb development ( Estella et al . , 2003; Treichel et al . , 2003; Kawakami et al . , 2004; Sahara et al . , 2007; Estella and Mann , 2010 ) .","However , inhibiting Wnt signaling alone in type II NB lineages does not have any obvious phenotypes ( Komori et al . , 2014 ) , indicating that Btd unlikely functions downstream of Wnt signaling in type II NB lineages .","Whether Fgf8 , Dll , or Hdc could function downstream of Btd to regulate INP generation remains to be investigated in the future .","In mammals , the Btd homolog Sp8 plays important roles in brain development .","In the developing mouse forebrain , Sp8 is expressed in cortical progenitors in a mediolateral gradient across the ventricular zone as well as in the lateral ganglionic eminence ( LGE ) and medial ganglionic eminence ( MGE ) ( Sahara et al . , 2007; Zembrzycki et al . , 2007 ) .","In developing human brains , Sp8 is abundantly expressed in the ventricular zone and the outer sub-ventricular zone where RGs and oRGs reside ( Ma et al . , 2013 ) .","In addition to its roles in interneuron development and the patterning of developing mammalian brains and spinal cords ( Griesel et al . , 2006; Waclaw et al . , 2006; Sahara et al . , 2007; Li et al . , 2011 ) , it was also shown that the loss of Sp8 led to the reduction of the progenitor pool ( Zembrzycki et al . , 2007 ) .","Our results show that mammalian Sp8 can rescue the loss of mature INPs resulting from the loss of Btd in Drosophila , suggesting that Btd/Sp8 could have conserved functions across different species .","It would be interesting to investigate if Sp8 has similar roles in promoting the generation of transient amplifying INPs , such as oRGs , in developing mammalian brains ."],["For btd loss-of-function analyses , yw btdXA FRT19A/FM7c and yw btdXG81 FRT19A/FM7c ( Wimmer et al . , 1993; Estella and Mann , 2010 ) were used for generating btd mutant clones and UAS-btd RNAi ( #29453; Bloomington Drosophila stock Center , Bloomington , Indiana ) for Btd RNAi knockdown .","Type II NB lineage-specific pntP1-GAL4 ( also named as GAL414−94 ) ( Zhu et al . , 2011 ) and erm-GAL4 ( II ) or ( III ) ( Pfeiffer et al . , 2008; Xiao et al . , 2012 ) were used to drive the expression of UAS-transgenes in type II NB lineages or in immature as well as mature INPs , respectively .","insc-Gal4 ( Gal41407 inserted in inscuteable promoter ) ( Luo et al . , 1994 ) was used to drive UAS-transgenes in all NB lineages .","UAS-mCD8-GFP driven by Btd-GAL4 was utilized as reporter for btd expression .","The R9D11-CD4-tdTomato transgenic line ( Han et al . , 2011 ) was used for labeling Ase+ immature INPs and mature INPs .","Other fly stocks include: hs-Flpase tub-GAL80 FRT19A; UAS-mCD8-GFP; pntP1-GAL4 for generating type II NB clones; pros17/TM6 Tb , pros10419/Tm3 Sb , and UAS-pros RNAi ( #26745; Bloomington stock ) for rescuing Btd loss-of-function phenotypes .","For RNAi knockdown analyses of Btd and Pros , larvae were raised at 29°C to increase the expression of UAS-RNAi transgenes after hatching .","Furthermore , UAS-Dcr2 was expressed together with UAS-RNAi transgenes to enhance the efficiency of RNAi knockdown .","For clonal analyses , MARCM ( Lee and Luo , 1999 ) clones were induced by 1 hr heat shock at 38°C for 1 day after larval hatching .","Larval brains were dissected at third instar larval stages for the examination of phenotypes .","Larval brains were dissected , fixed , and stained as described before ( Lee and Luo , 1999 ) .","Primary antibodies used in this study include: rabbit anti-Mira ( 1:500 ) , guinea pig anti-Ase ( 1:5000 ) , rabbit anti-Dpn ( 1:500 ) ( a gift from Y . N . Jan ) , rat anti-mCD8 ( Life Technologies , Grand Island , New York , 1:100 ) , mouse anti-Pros ( Developmental Studies Hybridoma Bank , Iowa City , Iowa , 1:20 ) , mouse monoclonal anti-α-tubulin ( Sigma , St . Louis , Missouri , 1:1000 ) , rabbit anti-dsRed ( Clontech , Mountain View , California , 1:500 ) , rabbit anti-PntP1 ( 1:500 , a gift from JS Skeath ) .","Secondary antibodies conjugated to Cy2 , Cy3 , Cy5 , or DyLight 647 ( Jackson ImmunoResearch , West Grove , Pennsylvania ) were used at 1:100 , 1:500 , or 1:500 , respectively .","Images were taken with a Zeiss LSM510 confocal microscopy and processed with Adobe Photoshop .","For quantifications of the number of mature INPs or the percentage of type II NB lineages with mature INPs , we focus on the medial group of type II NB lineages ( lineages DM1–DM6 ) .","Two-tailed t-tests were used for statistics analyses ."]],"string":"[\n [\n \"Intermediate neural progenitor cells ( INPs ) play a critical role in increasing the brain size and complexity .\",\n \"Transient amplification of INPs dramatically boosts the neural output from neural stem cells ( NSCs ) ( Kriegstein and Alvarez-Buylla , 2009; Florio and Huttner , 2014 ) .\",\n \"Recent studies in developing human brains as well as other mammalian brains suggest that an expansion of the number of transiently amplifying INPs , the outer sub-ventricular zone radial glia-like cells ( oRGs ) , likely contributes to the increased cortical size and complexity in humans and other gyrencephalic animals ( Fietz et al . , 2010; Hansen et al . , 2010; Lui et al . , 2011; Wang et al . , 2011 ) .\",\n \"On the other hand , accumulating body of evidence suggests that brain tumors could originate from dedifferentiation and unrestricted proliferation of INPs ( Holland et al . , 2000; Dai et al . , 2001; Walton et al . , 2009; Persson et al . , 2010; Zong et al . , 2012 ) .\",\n \"Therefore , it is fundamentally important to understand how the generation and proliferation of INPs are regulated .\",\n \"The recently discovered type II neuroblasts ( NBs , the Drosophila NSCs ) in developing Drosophila larval brains provide an excellent model system for studying mechanisms regulating the generation and proliferation of INPs ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .\",\n \"There are 8 type II NBs in each brain lobe .\",\n \"Like mammalian NSCs , Drosophila type II NBs produce neurons and glia indirectly by generating INPs .\",\n \"Individual INPs undergo 4–6 rounds of asymmetric divisions to produce a new INP to self-renew and a ganglion mother cell ( GMC ) , which divides terminally to produce neurons and/or glia ( Bayraktar et al . , 2010; Viktorin et al . , 2011; Yang et al . , 2013 ) .\",\n \"Meanwhile , individual INPs produce distinct types of neurons by sequentially expressing a set of distinct transcription factors to specify the identity of their progeny ( Bayraktar and Doe , 2013; Wang et al . , 2014 ) .\",\n \"Through self-renewing divisions , INPs not only amplify the number but also increase the diversity of neural progeny generated from type II NBs .\",\n \"Therefore , the neurogenesis pattern in type II NB lineages is remarkably similar to that in mammalian brains and the Drosophila INPs are functionally analogous to mammalian INPs , particularly oRGs .\",\n \"The generation of INPs in type II NB lineages involves multiple steps ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .\",\n \"Newly generated INPs are immature and do not express any NB markers , such as the proneural protein Asense ( Ase ) or the bHLH protein Deadpan ( Dpn ) , except for Miranda ( Mira ) .\",\n \"The Ase− immature INPs first turn on the expression of Ase to become Ase+ immature INPs .\",\n \"Ase+ immature INPs then further differentiate to become mature INPs , which express both Ase and Dpn .\",\n \"INPs do not divide until they are fully mature .\",\n \"The maturation of INPs requires Numb , the NHL family protein Brain tumor ( Brat ) , the transcription factor Earmuff ( Erm ) , as well as the BAP and Histone deacetylase 3 ( HDAC3 ) chromatin remodeling complexes ( Bowman et al . , 2008; Weng et al . , 2010; Eroglu et al . , 2014; Koe et al . , 2014 ) .\",\n \"Both Numb and Brat are segregated into Ase− immature INPs during the division of type II NBs to prevent them from dedifferentiating into NB fate , but they function through independent pathways .\",\n \"Numb inhibits Notch activity in Ase− immature INPs , whereas Brat likely antagonizes the activity of the EGR family transcription factor Klumpfuss ( Klu ) and Armadillo/β-Catenin in Ase− immature INPs ( Bowman et al . , 2008; Komori et al . , 2014 ) .\",\n \"Erm functions together with BAP and HDAC3 chromatin remodeling complexes after Brat and Numb to further restrict the developmental potential of INPs by attenuating the response of INPs to self-renewal factors such as Klu and Dpn ( Janssens et al . , 2014; Koe et al . , 2014 ) .\",\n \"In addition , the BAP chromatin remodeling complex limits the self-renewal of INPs by activating the expression of Prdm protein Hamlet ( Eroglu et al . , 2014 ) .\",\n \"In the absence of Numb , Brat , Erm , or chromatin remodeling complexes , INPs dedifferentiate into type II NBs and initiate tumorigenic overproliferation ( Bowman et al . , 2008; Weng et al . , 2010; Eroglu et al . , 2014; Koe et al . , 2014 ) .\",\n \"Therefore , these proteins are critical to prevent dedifferentiation of INPs .\",\n \"However , despite the significant progress on elucidating mechanisms that promote maturation and prevent dedifferentiation of INPs in the past few years , much less is known about why only type II NBs produce self-renewing INPs but not the type I NBs , which produce neurons by generating terminally dividing GMCs .\",\n \"One major difference between INPs and GMCs is that INPs divide to self-renew whereas GMCs divide terminally ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008; Yang et al . , 2013 ) .\",\n \"Therefore , in addition to avoiding dedifferentiation and unrestricted tumorigenic overproliferation , INPs need to overcome another challenge–to avoid over-differentiation and cell cycle exit–in order to maintain their progenitor state and self-renewal while they differentiate to mature and undergo self-renewing divisions .\",\n \"Type II NBs and newly born Ase− immature INPs differ from type I NBs and GMCs by the lack of the expression of Ase and the homeodomain protein Prospero ( Pros ) ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .\",\n \"In type I NB lineages , Pros is expressed in the cytoplasm in the NBs and translocates to the nucleus in GMCs to promote differentiation and cell cycle exit by inhibiting NB self-renewing genes and activating neural differentiation genes ( Li and Vaessin , 2000; Choksi et al . , 2006 ) .\",\n \"In type II NB lineages , Pros is not expressed in the NB or Ase− immature INPs .\",\n \"In Ase+ immature INPs and mature INPs , Pros is expressed at low levels in the cytoplasm ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .\",\n \"It has been demonstrated that the lack of Ase and Pros in type II NBs and Ase− immature INPs is essential for the generation of self-renewing INPs in type II NB lineages .\",\n \"Forced expression of Ase or Pros in type II NBs and their progeny is sufficient to eliminate INPs although removing Ase or Pros in type I NBs does not change the identity of type I NBs or induce the generation of INPs ( Bowman et al . , 2008; Bayraktar et al . , 2010; Zhu et al . , 2012 ) .\",\n \"Our recent studies demonstrated that the Ets family transcription factor Pointed P1 ( PntP1 ) suppresses Ase in type II NBs and is required for the generation of INPs ( Zhu et al . , 2011 ) .\",\n \"However , mechanisms that prevent premature differentiation of INPs and/or inhibit Pros expression in type II NBs and immature INPs are not known .\",\n \"In this study , we investigate the role of the Sp family transcription factor Buttonhead ( Btd ) in type II NB lineage development .\",\n \"Btd is a homolog of mammalian Sp8 ( Treichel et al . , 2003; Estella and Mann , 2010 ) .\",\n \"In developing mammalian brains , Sp8 is expressed in neural progenitor cells to regulate forebrain patterning and interneuron development ( Griesel et al . , 2006; Waclaw et al . , 2006; Sahara et al . , 2007; Li et al . , 2011 ) .\",\n \"In Drosophila embryos , Btd is required for the formation of specific head segments and NB formation ( Wimmer et al . , 1993; Younossi-Hartenstein et al . , 1997 ) .\",\n \"In addition , Sp8/Btd also promotes the growth of limbs and other appendages ( Estella et al . , 2003; Treichel et al . , 2003; Kawakami et al . , 2004; Estella and Mann , 2010 ) .\",\n \"In this study , we report that Btd is expressed in type II NB lineages to prevent premature differentiation of INPs into GMCs by suppressing Pros expression in immature INPs .\",\n \"We also demonstrate that PntP1 and Btd function cooperatively to specify type II NB lineages and promote the generation of INPs .\"\n ],\n [\n \"Our recent studies demonstrated that the Ets family transcription factor PntP1 is specifically expressed in type II NB lineages to promote the generation of INPs .\",\n \"However , although forced expression of PntP1 suppress Ase expression in nearly all type I NB lineages , it induces the generation of INP-like cells only in a subset of type I NBs ( Zhu et al . , 2011 ) .\",\n \"Therefore , it is likely that other protein ( s ) may function together with PntP1 to specify type II NB lineages and promote the generation of INPs .\",\n \"A recent functional genomic study showed that in addition to PntP1 , there are other nine genes that are highly expressed in brain tumors derived from type II NB lineages ( Carney et al . , 2012 ) .\",\n \"We wondered whether any of these genes could function together with PntP1 to promote INP generation .\",\n \"To test this idea , we first examined how knockdown of these genes would affect INP generation in type II NB lineages .\",\n \"A normal type II NB lineage contains 2–3 Ase− immature INPs , 2–3 Ase+ immature INP , and about 20–30 ( 26 . 9 ± 4 . 1 , mean ± SD ) Ase+ Dpn+ mature INPs ( Figure 1A–A′ , C–C′ , G ) .\",\n \"Interestingly , RNAi knockdown of Btd using the type II NB lineage-specific pntP1-GAL4 ( named as GAL414−94 previously ) ( Zhu et al . , 2011 ) as a driver led to a complete elimination of mature INPs in about 50% of type II NB lineages ( Figure 1B–B′ , G–H ) .\",\n \"Instead , only a few ( 3 . 7 ± 1 . 2 ) Ase+ Dpn− cells were observed next to the Ase− immature INPs ( Figure 1B–B′ ) .\",\n \"However , type II NBs remain Ase− as normal type II NBs ( Figure 1B–B′ ) , suggesting that the identity of the type II NBs was not affected by Btd RNAi knockdown . 10 . 7554/eLife . 03596 . 003Figure 1 . Loss of Btd eliminates mature INPs in type II NB lineages .\",\n \"( A–A′ , C–C′ )\",\n \"Wild-type type II NB lineages in a third instar larval brain .\",\n \"mCD8-GFP driven by pntP1-GAL4 labels all type II NB lineages ( A–A′ ) or a single type II NB clone ( C–C′ ) .\",\n \"Ase− immature INPs , Ase+ immature INPs , and mature INPs are indicated by open arrows , solid arrows , and arrowheads , respectively .\",\n \"( B–B′ , D–F′ )\",\n \"Btd RNAi knockdown type II NB lineages ( B–B′ ) or type II NB clones homozygous mutant for btdXG81 ( D–E′ ) or btdXA ( F–F′ ) in 3rd instar larval brains produce Ase− immature INPs ( open arrows ) and a few Ase+ daughter cells ( arrows ) but no mature INPs .\",\n \"Only 3 out of total 8 type II NB lineages are shown in ( A–A′ ) and ( B–B′ ) .\",\n \"In this and all other figures , asterisks indicate type II NBs and scale bars equal to 20 µm .\",\n \"Dpn staining alone shows the NB and mature INPs .\",\n \"( G–H )\",\n \"Quantifications of the number of mature INPs ( G ) and the percentage of type II NB lineages with mature INPs ( H ) in the wild type , Btd RNAi knockdown , and btd mutant type II NB lineages .\",\n \"The numbers on top of each bar are the numbers of type II NB lineages analyzed except for the numbers for the wt and btd RNAi in ( H ) , which are the number of brain lobes examined .\",\n \"The mean and stdev for btdXG81 and btdXA in ( H ) are calculated by bootstrapping .\",\n \"**p < 0 . 01 , *p < 0 . 05 ( Student t test ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 00310 . 7554/eLife . 03596 . 004Figure 1—figure supplement 1 . Expression of mouse Sp8 ( mSp8 ) .\",\n \"( A–A′′ ) and Drosophila Btd ( B–B′′ ) rescues the loss of mature INPs in btd mutant type II NB clones .\",\n \"Multiple mature INPs are observed in btd mutant type II NB clone that expresses UAS-mSp8 ( A–A′′ ) or UAS-btd ( B–B′′ ) .\",\n \"Type II NBs , Ase− immature INPs , Ase+ immature INPs , and mature INPs are indicated by asterisks , open arrows , solid arrows , and arrowheads , respectively . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 004 To confirm that the loss of INPs indeed results from the knockdown of Btd rather than off-target effects of UAS-Btd RNAi , we generated btd mutant type II NB clones using two loss-of-function alleles , btdXA and btdXG81 ( Wimmer et al . , 1993; Estella and Mann , 2010 ) .\",\n \"Consistent with the Btd RNAi knockdown , all btdXG81 mutant and 90% of btdXA mutant type II NB clones failed to generate any mature INPs except for 4–6 Ase+ Dpn− cells ( Figure 1D–F′ , G–H ) .\",\n \"Moreover , about 40% of btd mutant type II NBs ectopically express Ase , making them appear as type I NB lineages ( Figure 1E ) .\",\n \"The loss of INPs resulting from the Btd RNAi knockdown and btd loss-of-function mutations suggests that Btd is required for the generation of INPs .\",\n \"Remarkably , the loss of INPs in btd mutant clones can be similarly rescued by the expression of mouse Sp8 or Drosophila Btd ( Figure 1—figure supplement 1 ) , suggesting that mammalian Sp8 could have a conserved role in promoting the generation of transient amplifying INPs .\",\n \"Since the loss of mature INPs occurred even when Ase was not ectopically expressed in btd mutant type II NBs , the loss of INPs is not primarily due to the ectopic Ase expression or transformation of type II NBs into type I NBs .\",\n \"Therefore , we first focused our phenotypic analyses on lineages without the ectopic Ase expression in the NB .\",\n \"We also used the btdXG81 allele for further mutant phenotypic analyses below , given that btdXG81 shows slightly stronger phenotypes than btdXA .\",\n \"Why does the loss of Btd lead to the elimination of mature INPs ?\",\n \"When mature INPs are eliminated in the absence of Btd , the type II NBs without the ectopic Ase expression still produce Ase− immature INPs and a few Ase+ Dpn− daughter cells .\",\n \"In normal type II NB lineages , Ase+ Dpn− cells can be either Ase+ immature INPs or GMCs .\",\n \"Therefore , three possible scenarios could happen when mature INPs are eliminated in the absence of Btd:\",\n \"1 ) Ase− immature INPs differentiate into GMCs instead of Ase+ immature INPs;\",\n \"2 ) Ase− immature INPs differentiate into Ase+ immature INPs , which then directly differentiate into neurons/glia without further dividing;\",\n \"3 ) Ase− immature INPs differentiate into Ase+ immature INPs , which in turn differentiate into terminally dividing GMCs .\",\n \"To distinguish these possibilities , we first wanted to determine if Ase− immature INPs still differentiate into Ase+ immature INP in the absence of Btd by examining the expression of INP specific marker R9D11-CD4-tdTomato and progenitor marker Miranda ( Mira ) in the Ase+ cells next to the Ase− immature INPs .\",\n \"R9D11-CD4-tdTomato utilizes a DNA fragment R9D11 from the erm promoter to drive the expression of CD4-tdTomato ( Han et al . , 2011 ) .\",\n \"In normal type II NB lineages , R9D11-CD4-tdTomato is first turned on in Ase+ immature INPs and becomes stronger as INPs mature ( Figure 2A–A′ ) , which is similar to R9D11-mCD8-GFP ( Zhu et al . , 2011 ) .\",\n \"Mira is expressed in all NBs as well as INPs but not ( or very weakly ) in GMCs ( Figure 2—–figure supplement 1 ) .\",\n \"In Btd RNAi knockdown type II NB lineages without mature INPs , we found that R9D11-CD4-tdTomato was expressed in Ase+ daughter cells next to the Ase− immature INPs but its overall expression was much weaker than that in normal type II NB lineages ( Figure 2B–B′ , I ) .\",\n \"Consistently , Mira is also expressed in those Ase+ cells next to the Ase− immature INPs in the Btd RNAi knockdown type II NB lineages ( Figure 2—figure supplement 1 ) .\",\n \"The expression of R9D11-CD4-tdTomato and Mira suggests that Ase− immature INPs still differentiate into Ase+ immature INP in the absence of Btd as in wild-type type II NB lineages ( Figure 2J ) . 10 . 7554/eLife . 03596 . 005Figure 2 . Loss of Btd results in ectopic nuclear Pros in immature INPs and premature differentiation of Ase+ immature INPs into GMCs .\",\n \"( A–A′ )\",\n \"R9D11-CD4-tdTomato is expressed in Ase+ immature INPs ( solid arrows ) and mature INPs ( arrowheads ) but not in Ase− immature INPs ( open arrows ) in a wild-type type II NB lineage .\",\n \"( B–B′ )\",\n \"R9D11-CD4-tdTomato remains expressed in Ase+ daughter cells ( solide arrows ) next to the Ase− immature INPs ( open arrows ) when mature INPs are eliminated by Btd RNAi knockdown .\",\n \"( C–C′ ) pH3 is not detected in Ase+ immature INPs ( solid arrows ) in a wild-type type II NB lineage .\",\n \"( D–E′ ) pH3 is expressed in Ase+ daughter cells ( solid arrows ) that are the furthest from the Ase− immature INPs ( open arrows ) in a Btd RNAi knockdown type II NB lineage without mature INPs ( D–D′ ) or btd mutant type II NB lineages ( E–E′ ) .\",\n \"( F–F′ )\",\n \"Nuclear Pros is not expressed in Ase− immature INPs ( open arrows ) in a wild-type type II NB lineage .\",\n \"( G-H′ )\",\n \"Nuclear Pros is ectopically expressed in both Ase− immature INPs ( open arrows ) and Ase+ cells ( solid arrows ) in Btd RNAi knockdown ( G–G′ ) or btd mutant ( H–H′ ) type II NB lineages .\",\n \"( I ) Quantifications of relative overall expression levels of R9D11-CD4-tdTomato in wild-type ( A–A′ ) and Btd RNAi knockdown ( B–B′ ) type II NB lineages .\",\n \"( J ) A diagram of neurogenesis patterns in type II NB lineages in the presence or absence of Btd . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 00510 . 7554/eLife . 03596 . 006Figure 2—figure supplement 1 . Mira expression in Btd RNAi knockdown type II NB clone .\",\n \"( A–A′ )\",\n \"Mira is expressed in INPs ( yellow arrows ) in a wild-type type II NB lineage and forms a basal crescent ( white arrows ) at metaphase .\",\n \"( B–B′ )\",\n \"In a Btd RNAi knockdown type II NB lineage without mature INPs , Mira is expressed in Ase+ daughter cells ( yellow arrows ) next to the Ase− immature INP ( open arrows ) but does not form a basal crescent at metaphase in the dividing Ase+ cell ( white arrows ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 006 Next we asked if Ase+ immature INPs differentiate into neurons/glia directly or GMCs in the absence of Btd .\",\n \"GMCs express both Ase and nuclear Pros and divide terminally but do not form a Mira crescent while dividing .\",\n \"If Ase+ immature INPs directly differentiate into neurons/glia , then all the Ase+ daughter cells should be Ase+ immature INPs and none of them should be dividing .\",\n \"In contrast , if Ase+ immature INPs differentiate into GMCs , then some Ase+ daughter cells will become mitotically active and express nuclear Pros but will not form a Mira crescent at the metaphase or telophase .\",\n \"Immunostaining with the mitotic marker phospho-histone 3 ( pH3 ) showed that unlike Ase+ immature INPs , which never become pH3-positive ( Figure 2C–C′ ) , some Ase+ daughter cells generated in both Btd RNAi knockdown type II NB lineages without mature INPs ( Figure 2D–D′ ) and btd mutant type II NB lineages became mitotically active ( Figure 2E–E′ ) .\",\n \"However , the pH3 positive cells were always the furthest from the Ase− immature INPs among the Ase+ daughter cells ( Figure 2D–D′ ) , suggesting that the Ase+ daughter cells divide terminally like GMCs .\",\n \"Consistently , unlike in mature INPs , which form a Mira crescent at metaphase ( Figure 2—figure supplement 1 ) , we did not observe any Mira crescents in the Ase+ daughter cells at metaphase ( Figure 2—figure supplement 1 ) .\",\n \"The terminal division and the lack of the Mira crescent strongly argue that Ase+ immature INPs differentiate into GMCs in the absence of Btd .\",\n \"To further confirm that late immature INPs differentiate into GMCs in the absence of Btd , we then examined the expression of nuclear Pros in the Ase+ daughter cells .\",\n \"Nuclear Pros is a cell fate determinant of GMCs .\",\n \"In normal type II NBs lineages , Pros is expressed in the cytoplasm of Ase+ immature INPs and mature INPs and in the nucleus of GMCs and post-mitotic neurons , but not in type II NBs or Ase− immature INPs ( Figure 2F–F′ ) .\",\n \"If Ase+ immature INPs differentiate into GMCs in the absence of Btd , we expected that some of the Ase+ daughter cells express nuclear Pros .\",\n \"Interestingly , immunostaining of Pros showed that nuclear Pros was expressed not only in all Ase+ daughter cells but also in Ase− immature INPs generated in Btd RNAi knockdown or btd mutant type II NB lineages ( Figure 2G–H′ ) .\",\n \"Given that Pros promotes cell cycle exit and GMC differentiation and that forced expression of Pros is sufficient to eliminate INPs in type II NB lineages ( Li and Vaessin , 2000; Choksi et al . , 2006; Bayraktar et al . , 2010 ) , the ectopic expression of nuclear Pros in immature INPs resulting from the loss of Btd very likely promotes the premature differentiation of Ase+ immature INPs into GMCs and cell cycle exit , leading to the loss of mature INPs ( Figure 2J ) .\",\n \"These results also reveal that it is Btd that is responsible for the suppression of Pros in immature INPs .\",\n \"To determine if the ectopic expression of nuclear Pros in immature INPs is indeed responsible for the elimination of mature INPs in the absence of Btd , we next examined if reducing Pros expression was able to rescue the elimination of INPs in Btd RNAi knockdown or btd mutant type II NB lineages .\",\n \"To reduce Pros expression , we either removed one wild-type copy of pros or knocked down Pros by RNAi in type II NB lineages .\",\n \"Remarkably , the elimination of mature INPs resulting from the Btd RNAi knockdown was nearly fully rescued even just by removing one wild-type copy of pros ( Figure 3A–D′ , I–J ) .\",\n \"Unlike Btd RNAi knockdown in wild-type background , which resulted in a completely elimination of mature INPs in about 50% of type II NB lineages ( Figure 1B–B′ , G–H , Figure 3B–B′ , I–J ) , knockdown of Btd in pros17 or pros10419 heterozygous mutant animals no longer led to an obvious loss of mature INPs ( Figure 3D–D′ , I–J , and data not shown ) , although type II NB lineages develop normally in pros17 or pros10419 heterozygous mutant animals ( Figure 3C–C′ , I–J , and data not shown ) .\",\n \"Similarly , the loss of INPs in btd mutant clones was also largely rescued when btd mutant type II NB clones were generated in pros17 heterozygous mutant background ( Figure 3—figure supplement 1 ) . 10 . 7554/eLife . 03596 . 007Figure 3 . Reducing Pros rescues the elimination of INPs resulting from the loss of Btd .\",\n \"( A–D′ )\",\n \"The loss of INPs resulting from Btd RNAi knockdown is rescued in pros17 heterozygous mutant background .\",\n \"Only two lineages are shown in each brain .\",\n \"( A–A′ )\",\n \"Wild-type type II NB lineages have multiple mature INPs .\",\n \"( B–B′ )\",\n \"Btd RNAi knockdown causes a loss of mature INPs .\",\n \"( C–C′ )\",\n \"Type II NB lineages in pros17 heterozygous mutant larvae produce a similar number of mature INPs as in wild-type larvae .\",\n \"( D–D′ )\",\n \"Btd RNAi knockdown no long leads to the loss of mature INPs in pros17 heterozygous mutant type II NB lineages .\",\n \"( E–H′ )\",\n \"Pros RNAi knockdown rescues the loss of INPs in btd mutant type II NB clones .\",\n \"( E–E′ )\",\n \"A wild-type type II NB clone has multiple mature INPs .\",\n \"( F–F′ )\",\n \"A btd mutant type II NB clone contains no mature INPs .\",\n \"( G–G′ )\",\n \"Pros RNAi knockdown causes overproliferation of mature INPs in a type II NB clone .\",\n \"( H–H′ )\",\n \"Pros RNAi knockdown rescues the loss of mature INPs in a btd mutant type II clone .\",\n \"Arrowheads point to mature INPs in all images .\",\n \"( I–L )\",\n \"Quantifications of the number of mature INPs ( I–K ) and the percentage of lineages with mature INPs ( J–L ) for the rescue of Btd RNAi knockdown phenotypes in pros17/+ larvae ( I–J ) or the rescue of btd mutant phenotypes by Pros RNAi knockdown ( K–L ) .\",\n \"The samples sizes on top of each bar represent the number of type II NB lineages ( I , K , L ) or the number of brain lobes ( J ) .\",\n \"The mean and stdev in ( L ) are calculated by bootstrapping .\",\n \"**p < 0 . 01 , *p < 0 . 05 ( Student t test ) .\",\n \"NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 00710 . 7554/eLife . 03596 . 008Figure 3—figure supplement 1 . Reducing Pros rescues the elimination of INPs resulting from the loss of Btd .\",\n \"( A–D′ )\",\n \"The loss of INPs resulting from Btd RNAi knockdown is rescued by Pros RNAi knockdown .\",\n \"( A–A′ )\",\n \"Wide-type type II lineages labeled with mCD8-GFP driven by insc-GAL4 have multiple mature INPs .\",\n \"( B–B′ )\",\n \"Btd RNAi knockdown driven by insc-GAL4 eliminates mature INPs in a subset of type II NB lineages .\",\n \"( C–C′ )\",\n \"RNAi knockdown of Pros results in overproliferation of mature INPs .\",\n \"( D–D′ )\",\n \"Simultaneous knockdown of Btd and Pros leads to a similar overproliferation of mature INPs in type II NB lineages as Pros RNAi knockdown alone .\",\n \"( E–H′ )\",\n \"The loss of mature INPs in btd mutant clones is rescued in pros17 heterozygous mutant larvae .\",\n \"( E–E′ )\",\n \"A wild-type type II NB clone has multiple mature INPs .\",\n \"( F–F′ )\",\n \"A btd mutant type II NB clone contains no mature INPs .\",\n \"( G–G′ ) pros17 heterozygous mutant type II NB clone contains multiple mature INPs as the wild-type type II NB clone ( A–A′ ) .\",\n \"( H–H′ )\",\n \"A btd mutant type II NB clone generated in pros17 heterozygous mutant larvae has multiple mature INPs .\",\n \"Asterisks indicate NBs and arrowheads point to mature INPs in all images .\",\n \"( I–L )\",\n \"Quantifications of the number of mature INPs ( I–K ) or the percentage of type II NB lineages with mature INPs ( J–L ) for the rescue of Btd RNAi knockdown phenotypes by Pros RNAi knockdown ( I–J ) or the rescue of btd mutant phenotypes in pros17 heterozygous mutant larvae ( K–L ) .\",\n \"The sample size on top of each bar represents the number of lineages ( I , K , L ) or the number of brain lobes ( J ) .\",\n \"The mean and stdev in ( L ) are calculated by bootstrapping .\",\n \"** , p < 0 . 01; NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 008 Consistent with the rescue in pros heterozygous mutant animals , Pros RNAi knockdown also rescued the loss of INPs resulting from the loss of Btd ( Figure 3E–H′ , K–L , Figure 3—figure supplement 1 ) .\",\n \"Pros RNAi knockdown led to overproliferation of mature INPs as observed in pros mutant type II NB clones ( Figure 3G–G′ , K , Figure 3—figure supplement\",\n \"1 ) ( Bowman et al . , 2008 ) .\",\n \"When Pros was knocked down in btd mutant type II NB clones , mature INPs were rescued in all btd mutant clones ( Figure 3H–H′ , L ) .\",\n \"In about 70% of btd mutant type II NB clones , Pros RNAi knockdown led to a similar mature INP overproliferation as in wild-type clones .\",\n \"In other 30% of btd mutant clones , Pros RNAi knockdown partially or fully rescued mature INPs without causing the overproliferation of mature INPs .\",\n \"Similarly , Pros RNAi knockdown also rescued the loss of mature INPs resulting from Btd RNAi knockdown ( Figure 3—figure supplement 1 ) .\",\n \"Results from these rescue experiments demonstrate that the ectopic nuclear Pros in immature INPs is indeed responsible for the loss of mature INPs .\",\n \"Interestingly , removing one wild-type copy of pros or knocking down Pros not only rescued the loss of mature INPs but also suppressed the ectopic Ase expression in all btd mutant type II NBs ( Figure 3H , Figure 3—figure supplement 1 ) , suggesting that the ectopic Ase expression in btd mutant type II NBs also results from the ectopic expression of nuclear Pros in immature INPs .\",\n \"Our results showed that Btd is required to suppress Pros in Ase− immature INPs .\",\n \"We next asked if Btd is also required to partially suppress Pros at later stages of INP development .\",\n \"In normal type II NB lineages , Pros is absent in Ase− immature INPs but is expressed at low levels in the cytoplasm of Ase+ immature INPs and mature INPs ( Bello et al . , 2008; Boone and Doe , 2008; Bowman et al . , 2008 ) .\",\n \"Maintaining the expression of Pros at low levels is essential for the self-renewal of INPs ( Bayraktar et al . , 2010 ) .\",\n \"However , the complete elimination of mature INPs makes it difficult to assess the role of Btd in mature INPs .\",\n \"Therefore , we used erm-GAL4 ( III ) and erm-GAL4 ( II ) to knock down Btd .\",\n \"Both erm-GAL4 ( III ) and erm-GAL4 ( II ) are expressed in Ase+ immature INPs and mature INPs , whereas erm-GAL4 ( II ) is also expressed in Ase− immature INPs except for the newly born Ase− immature INPs ( Xiao et al . , 2012 ) .\",\n \"However , knockdown of Btd using either erm-GAL4 ( III ) ( Figure 4A–B′′ , E ) or erm-GAL4 ( II ) ( Figure 4C–D′′ , F ) did not result in any obvious loss of mature INPs in type II NB lineages .\",\n \"In line with these RNAi knockdown results , we were able to recover multicellular btd mutant INP clones that were comparable to wild-type INP clones ( Figure 4—figure supplement\",\n \"1 ) while we generated btd mutant type II NB clones , indicating that btd mutant INPs were still able to divide multiple rounds like wild-type INPs and did not prematurely differentiate into GMCs .\",\n \"These data suggest that Btd likely suppresses Pros expression only in newly born Ase− immature INPs but not in immature INPs at later developmental stages or mature INPs . 10 . 7554/eLife . 03596 . 009Figure 4 . Knockdown of Btd in immature or mature INPs by erm-GAL4 lines does not lead to the loss of mature INPs .\",\n \"( A–A′′ , C–C′′ )\",\n \"Wild-type type II NB lineages are labeled with mCD8-GFP driven by erm-GAL4 ( III ) ( A–A′′ ) or erm-GAL4 ( II ) ( C–C′′ ) .\",\n \"( B–B′′ , D–D′′ )\",\n \"Knockdown of Btd in Ase+ immature INPs and mature INPs by erm-GAL4 ( III ) ( B–B′′ ) or in Ase− immature INPs as well as Ase+ immature INP and mature INPs by erm-GAL4 ( II ) ( D–D′′ ) does not cause a reduction of the number of mature INPs .\",\n \"Only two lineages are shown in each brain .\",\n \"( E–F )\",\n \"Quantifications of the number of mature INPs in type II NB lineages in which Btd is knocked down by erm-GAL4 ( III ) ( E ) or erm-GAL4 ( II ) ( F ) .\",\n \"NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 00910 . 7554/eLife . 03596 . 010Figure 4—figure supplement 1 . Btd likely does not function in mature INPs .\",\n \"( A ) A wild-type INP clone with 4 post-mitotic cells .\",\n \"( B ) A btd mutant INP clone with 6 post-mitotic cells .\",\n \"INP clones labeled with mCD8-GFP are outlined by dashed circles .\",\n \"INP clones are identified as clones that have more than two cells but no NBs . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 010 Our btd mutant MARCM analyses showed that in addition to the loss of INPs , Ase was ectopically expressed in about 40% of btd mutant type II NBs .\",\n \"We showed previously that PntP1 is expressed in type II NBs as well as Ase− and Ase+ immature INPs ( Figure 5A–A′′ ) ( Zhu et al . , 2011 ) .\",\n \"Inhibiting PntP1 activity results in ectopic Ase expression in type II NBs and elimination of INPs ( Zhu et al . , 2011 ) .\",\n \"Therefore , we wondered if PntP1 expression was reduced or even lost in btd mutant type II NBs .\",\n \"Immunostaining of PntP1 showed that PntP1 was expressed at reduced levels in most btd mutant type II NB lineages without the ectopic Ase expression ( Figure 5B–B′ ) .\",\n \"The reduction is about 10% in the NBs and 50% in the immature INPs ( Figure 5E–F ) .\",\n \"In those btd mutant clones , PntP1 was also detected in the Ase+ daughter cells next to the Ase− immature INPs ( Figure 5B–B′ ) , providing additional evidence to support that Ase− immature INPs still differentiate into Ase+ immature INPs in the absence of Btd .\",\n \"However , in btd mutant type II NB lineages with the ectopic Ase expression in the NB , PntP1 was largely abolished in both the NBs and their progeny ( Figure 5C–C′ , E–F ) .\",\n \"The correlation of the ectopic Ase expression in the NB and the severe reduction or loss of PntP1 suggesting that the ectopic Ase expression in btd mutant type II NBs could result from the severe reduction or loss of PntP1 expression . 10 . 7554/eLife . 03596 . 011Figure 5 . PntP1 expression is reduced in btd mutant type II NB clones .\",\n \"( A–A′′ )\",\n \"PntP1 is expressed in the NB ( * ) , Ase− immature INPs ( open arrows ) , as well as Ase+ immature INPs ( solid arrows ) in a wild-type type II NB clone .\",\n \"( B–B′′ )\",\n \"PntP1 expression is much lower in a btd mutant type II NB clone without the ectopic Ase expression in the NB than that in a neighboring btd heterozygous type II NB lineage .\",\n \"The reduction is particularly obvious in the Ase− immature INPs ( open arrows ) .\",\n \"Note that PntP1 remains expressed in Ase+ daughter cells ( arrows ) next to the Ase− immature INPs .\",\n \"( C–C′′ )\",\n \"PntP1 expression is largely abolished in a btd mutant type II NB clone with the ectopic Ase expression in the NB .\",\n \"In a neighboring btd heterozygous type II NB lineages , PntP1 is still detected in the NBs ( * ) , Ase- immature INPs ( open arrows ) and Ase+ immature INPs ( arrows ) .\",\n \"( D–D′′ )\",\n \"Knocking down Pros restores the expression of PntP1 in the NB ( * ) , Ase− immature INPs ( open arrows ) , and Ase+ immature INPs ( arrows ) in a btd mutant clone to levels comparable to those in a neighboring btd heterozygous type II NB lineage .\",\n \"Wild-type ( A–A′′ ) or btd mutant type II NB clones ( B–D′′ ) are outlined by dashed lines and neighboring btd heterozygous type II NB lineages ( B–D′′ ) are marked with dotted lines .\",\n \"( E–F )\",\n \"Quantifications of PntP1 expression levels in type II NBs ( E ) and Ase− immature INPs ( F ) in btd mutant type II NB clones relative to neighboring type II NB lineages in the same brains .\",\n \"**p < 0 . 01; NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01110 . 7554/eLife . 03596 . 012Figure 5—figure supplement 1 . The reduction of PntP1 is unlikely responsible for the loss of INPs in btd mutant type II NB clones .\",\n \"( A–A′′ )\",\n \"A btd mutant type II NB clone has similar expression levels of PntP1 as a neighboring btd heterozygous mutant type II NB lineage but fails to produce mature INPs .\",\n \"( B–B′′ )\",\n \"Expressing UAS-PntP1 does not rescue the loss of mature INPs in a btd mutant clone , even though the PntP1 expression in the btd mutant clone is much higher than that in a neighboring btd heterozygous mutant type II NB lineage .\",\n \"( C–C′′ )\",\n \"Nuclear Pros remains ectopically expressed in Ase− immature INPs and Ase+ daughter cells in a btd mutant type II NB clone expressing UAS-PntP1 .\",\n \"( D–D′′ )\",\n \"Mature INPs are partially rescued by UAS-PntP1 in a btd mutant type II NB clone .\",\n \"In all images , btd mutant clones are outlined by dashed lines and their neighboring btd heterozygous mutant type II NB lineages are marked by dotted lines .\",\n \"Open arrows: Ase− immature INPs; solid arrows: Ase+ immature INPs; arrowheads: mature INPs . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 012 To determine if the reduction/loss of PntP1 is responsible for the ectopic Ase expression and/or the loss of INPs in btd mutant type II NB lineages , we examined if restoring PntP1 expression was sufficient to suppress ectopic Ase expression and/or rescue the loss of INPs resulting from the loss of Btd by expressing UAS-pntP1 in btd mutant type II NB clones .\",\n \"Our results showed that expressing UAS-pntP1 resulted in higher expression of PntP1 in btd mutant type II NB clones than that in neighboring btd heterozygous mutant type II NB lineages and suppressed the ectopic Ase expression in all btd mutant type II NBs ( n = 12 ) ( Figure 5—figure supplement 1 ) .\",\n \"However , unlike reducing Pros expression , which rescued mature INPs in nearly all btd mutant type II NB clones ( Figure 3 , Figure 3—figure supplement 1 ) , expressing UAS-pntP1 failed to rescue mature INPs or suppress the ectopic nuclear Pros in Ase− immature INPs or Ase+ daughter cells in the majority of btd mutant clones ( Figure 5—figure supplement 1 ) .\",\n \"Only in 3 out of total 10 btd mutant clones expressing UAS-PntP1 , we observed that mature INPs were partially rescued to 9 . 3 ± 3 . 2 per lineages ( Figure 5—figure supplement 1 ) , which is still much fewer than the number of mature INPs ( 20–30 per lineages ) in normal type II NB lineages .\",\n \"Consistent with the inability of UAS-pntP1 to fully rescue the loss of mature INPs , we found occasionally that btd mutant clones that did not show an obvious reduction of PntP1 in either the NBs or early immature INPs still failed to generate any mature INPs ( Figure 5—figure supplement 1 ) .\",\n \"Therefore , these results demonstrate that the severe reduction/loss of PntP1 accounts for the ectopic Ase expression in btd mutant type II NBs but is not the primary reason for the loss of mature INPs .\",\n \"Given that reducing the expression of Pros suppressed the ectopic Ase expression in btd mutant type II NBs , we then asked if reducing Pros expression could also rescue the reduction/loss of PntP1 in btd mutant type II NB clones .\",\n \"Indeed , consistent with the suppression of the ectopic Ase expression by Pros RNAi knockdown , PntP1 expression in both the NBs and immature INPs returned to normal levels when the loss of INPs was rescued by Pros RNAi knockdown in btd mutant type II NB clones ( Figure 5D–F ) .\",\n \"These results suggest that the reduction/loss of PntP1 in btd mutant type II NB clones is due to the ectopic Pros expression in immature INPs .\",\n \"However , given that ectopic nuclear Pros is only observed in immature INPs but not in the NBs in btd mutant type II NB clones , the reduction/loss of PntP1 and the subsequent ectopic Ase expression in the NB is most likely a secondary effect of the ectopic nuclear Pros expression in immature INPs .\",\n \"Our Btd loss of function analyses demonstrated that Btd is critical for the generation of INPs in type II NBs lineages .\",\n \"We next examined if Btd is only expressed type II NB lineages .\",\n \"Since Btd antibodies are not available and our in situ hybridization signals of btd mRNAs in the central brain were barely detectable ( data now shown ) , we used the btd-GAL4 as a reporter for btd expression .\",\n \"btd-GAL4 is a GAL4 enhancer trap line , in which the GAL4 transgene is inserted at 753bp upstream of the transcription start site of btd ( Estella et al . , 2003 ) .\",\n \"btd-GAL4 shows similar expression patterns as endogenous Btd in ventral imaginal discs ( Estella et al . , 2003 ) .\",\n \"We found that mCD8-GFP driven by the btd-GAL4 is expressed in all type II NB lineages but not type I NB lineages on the dorsal side of larval brains ( Figure 6A–A′ ) .\",\n \"In type II NB lineages , the expression of mCD8-GFP driven by btd-GAL4 is detected in the NB but becomes much stronger in immature INPs next to the NBs ( Figure 6A ) .\",\n \"The expression of mCD8-GFP then becomes progressively weak in cells away from the NBs and is barely detectable in some mature INPs distal from the NB ( Figure 6A–A′ ) .\",\n \"The expression pattern of btd-GAL4 in type II NB lineages is similar to that of pntP1-GAL4 ( e . g . Figure 1A–A′ ) and is consistent with our results that Btd mainly functions in immature INPs . 10 . 7554/eLife . 03596 . 013Figure 6 . Btd is expressed in type II NB lineages and a subset of type I NB lineages .\",\n \"( A–A′ ) mCD8-GFP driven by btd-Gal4 is expressed in all type II NB lineages ( outlined by dashed lines ) but not type I NB lineages ( e . g . arrows ) on the dorsal side of a 3rd instar larval brain .\",\n \"The expression of mCD8-GFP becomes progressively weak in cell away from the NB .\",\n \"Some mature INPs ( e . g . arrowheads ) distant from the NB have no obvious expression of mCD8-GFP .\",\n \"Only seven out of total eight type II NB lineages are shown in this particular focal plane .\",\n \"( B–B′ )\",\n \"Two type I NB lineages are labeled by mCD8-GFP driven by btd-GAL4 on the ventral side of a 3rd instar larval brain .\",\n \"( C–C′ ) mCD8-GFP driven by btd-Gal4 labels a subset of type I NB lineages ( e . g . arrows ) in the ventral nerve cord ( VNC ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01310 . 7554/eLife . 03596 . 014Figure 6—figure supplement 1 . The GAL4 insertion in the btd-GAL4 line does not affect type II NB lineage development .\",\n \"( A–A′′ )\",\n \"A wild-type type II NB clone contains multiple mature INP .\",\n \"( B–B′′ )\",\n \"A btd-GAL4 mutant type II NB clone has a similar number of mature INPs as wild-type type II NB clone .\",\n \"( C ) Quantifications of the number of mature INPs in wild-type and btd-GAL4 mutant type II NB clones .\",\n \"Arrowheads: mature INPs; NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01410 . 7554/eLife . 03596 . 015Figure 6—figure supplement 2 . The loss of INP phenotype resulting from Btd RNAi knockdown is rescued by the expression of mouse Sp8 ( mSp8 ) or Drosophila Btd .\",\n \"( A–A′′ )\",\n \"Btd RNAi knockdown driven by btd-GAL4 completely eliminates mature INPs in all type II NB lineages in a 3rd instar larval brain .\",\n \"( B–B′′ )\",\n \"The expression of UAS-mSp8 driven by btd-GAL4 fully rescues the loss of mature INPs in all type II NB lineages resulting from Btd RNAi knockdown .\",\n \"( C–C′′ )\",\n \"The expression of UAS-btd driven by btd-GAL4 partially rescues the loss of mature INPs in type II NB lineages resulting from Btd RNAi knockdown .\",\n \"Arrowheads point to mature INPs in all images .\",\n \"( D–E )\",\n \"Quantifications of the percentage of type II NB lineages with mature INPs ( D ) and the number of mature INPs ( E ) in 3rd instar larvae with indicated genotypes .\",\n \"Sample sizes represent the number of brain lobes ( D ) or the number of type II NB lineages ( E ) .\",\n \"* , p < 0 . 05; ** , p < 0 . 01; NS: not significant . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 015 In addition to type II NB lineages , mCD8-GFP driven by btd-GAL4 is also expressed in two type I NB lineages on the ventral side of larval brains as well as about 31 type I NB lineages in the ventral nerve cord ( VNC ) ( Figure 6B–C′ ) .\",\n \"In those type I NB lineages , mCD8-GFP driven by btd-GAL4 is expressed in the NBs , GMCs , and newly born neurons .\",\n \"The expression pattern of btd-GAL4 in single type I NB lineages is similar to those of other NB GAL4 lines such as insc-GAL4 ( e . g . Figure 7D ) , suggesting that Btd likely functions in the NB in these type I NB lineages . 10 . 7554/eLife . 03596 . 016Figure 7 . Btd and PntP1 function cooperatively to promote the generation of INPs .\",\n \"( A–C′ )\",\n \"Ectopic expression of PntP1 consistently promotes the generation of INP-like cells in Btd-positive type I NB lineages in larval brains .\",\n \"( A–A′ )\",\n \"A wild-type type I NB lineage labeled by btd-GAL4 has no INP-like cells .\",\n \"( B–C′ )\",\n \"The ectopic expression of PntP1 driven by btd-GAL4 suppresses Ase in the NB ( * ) and induces the generation of Ase+ Dpn+ INP-like cells ( arrowheads ) ( B–B′ ) , which also express INP-specific marker R9D11-CD4-tdTomato ( C–C′ ) .\",\n \"( D–E′ )\",\n \"The expression of PntP1 driven by tub-GAL4 suppresses Ase in both wild-type ( D ) and btd mutant ( E ) type I NBs ( * ) but only induced the generation of INP-like cells in the wild-type type I NB clone ( D–D′ ) but not in the btd mutant clone ( E–E′ ) .\",\n \"( F–I′ ) insc-GAL4 drives the expression of UAS-mCD8-GFP alone ( F–F′ ) or together with UAS-btd ( G–G′ ) , UAS-pntP1 ( H–H′ ) , or UAS-btd plus UAS-pntP1 ( I–I′ ) in type I NB lineages .\",\n \"Images are from the ventral side of larval brains , where only type I NB lineages are observed in wild-type animals ( F–F′ ) .\",\n \"The expression of UAS-btd ( G–G′ ) or UAS-pntP1 ( H–H′ ) alone promotes the generation of INP-like cells only in small subset of type I NB lineages ( dashed circles ) .\",\n \"The expression of Btd only suppresses/reduces Ase expression in type I NBs ( arrows ) that produce INP-like cells ( G–G′ ) but not in other type I NBs ( arrowheads ) , where PntP1 suppresses Ase in nearly all type I NBs ( e . g . arrows ) regardless of the generation of INP-like cells ( H–H′ ) .\",\n \"Co-expression of Btd and PntP1 promotes the generation of INP-like cells nearly in all type I NB lineages ( dashed circles ) ( I–I′ ) .\",\n \"( J–J′ )\",\n \"Quantifications of the percentage of type I NB lineages that produce INP-like cells in larvae with indicated genotypes .\",\n \"The number on top of each bar represents the number of brain lobes except for numbers for the expression of UAS-pntP1 driven by tub-GAL4 , which are the number of clones .\",\n \"The mean and stdev for the percentage of wild-type or btd mutant clones expressing UAS-pntP1 driven by tub-GAL4 are calculated by bootstrapping .\",\n \"**p < 0 . 01 . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01610 . 7554/eLife . 03596 . 017Figure 7—figure supplement 1 . Btd and PntP1 function cooperatively to induce the generation of INP-like cells in type I NB lineages in the VNC .\",\n \"( A–A′ )\",\n \"Type I NB lineages labeled by mCD8-GFP driven by btd-GAL4 in the VNCs have no INPs ( e . g . dashed circles ) .\",\n \"( B–B′ )\",\n \"Expression of PntP1 driven by btd-GAL4 induces the generation of INP-like cells ( small Dpn+ cells ) in nearly all Btd-positive lineages ( dashed circles ) .\",\n \"( C–C′ )\",\n \"Type I NB lineages labeled by mCD8-GFP driven by insc-GAL4 in the VNC have no INPs .\",\n \"( D–D′ )\",\n \"Expression of UAS-btd driven by insc-GAL4 neither suppresses Ase in the NB nor induces the generation of INP-like cells in type I NB lineages ( e . g . dashed circles ) in the VNC .\",\n \"( E–E′ )\",\n \"Expression of UAS-pntP1 driven by insc-GAL4 suppresses Ase in nearly all NBs ( * ) and induces the generation of INP-like cells in a subset of type I NB lineages ( dashed circles ) in the VNC .\",\n \"( F–F′ )\",\n \"Co-expression of UAS-btd and UAS-pntP1 driven by insc-GAL4 induces INP-like cells in almost all type I NB lineages ( dashed circles ) in the VNC .\",\n \"( G ) Quantifications of the percentage of type I NB lineages with INP-like cells in brain lobes with indicated genotypes .\",\n \"**p < 0 . 01 . DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 01710 . 7554/eLife . 03596 . 018Figure 7—figure supplement 2 . Overexpression of Btd promotes the generation of INP-like cells .\",\n \"( A–A′′ )\",\n \"INP specific marker R9D11-CD4-tdTomato is not expressed in any type I NB lineages ( e . g . arrows ) on the ventral side of a larval brain .\",\n \"( B–B′′ )\",\n \"Overexpression of Btd induces the generation of INP-like cells in a subset of type I NB lineages on the ventral side of a larval brain as indicated by the expression of R9D11-CD4-tdTomato ( e . g . arrows ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 03596 . 018 In order to determine if the btd-GAL4 reflects the endogenous Btd expression pattern in the central brain , we tried to rescue the btd loss-of-function phenotypes by expressing UAS-mSp8 or UAS-btd driven by btd-GAL4 .\",\n \"However , the GAL4 insertion in the btd-GAL4 line does not affect the type II NB lineage development although it causes a lethal mutation of btd ( Figure 6—figure supplement 1 ) .\",\n \"Therefore , we tried to rescue Btd RNAi knockdown phenotypes by the expression of UAS-mSp8 or UAS-btd driven by btd-GAL4 instead .\",\n \"The expression of UAS-btd RNAi driven by btd-GAL4 completely eliminated mature INPs in nearly all type II NB lineages ( Figure 6—figure supplement 2 ) , which was much stronger than the phenotype of Btd RNAi knockdown driven by pntP1-GAL4 .\",\n \"As expected , the expression of UAS-mSp8 driven by btd-GAL4 fully rescued the loss of INPs resulting from Btd RNAi knockdown in all type II NB lineages ( Figure 6—figure supplement 2 ) .\",\n \"Similarly , the expression of UAS-btd driven by btd-GAL4 partially rescued the loss of INPs in about 67% of lineages ( Figure 6—figure supplement 2 ) .\",\n \"The incomplete rescue by UAS-btd is likely because UAS-btd contains the sequence targeted by UAS-btd RNAi .\",\n \"The rescue of Btd RNAi knockdown phenotypes by the expression of UAS-mSp8 and UAS-btd driven by btd-GAL4 together with the strong loss of INP phenotypes in all btd mutant type II NB clones strongly argue that btd-GAL4 expression is likely consistent with the endogenous Btd expression pattern in the central brain .\",\n \"Our previous studies showed that ectopic expression of PntP1 could induce the generation of INP-like cells in more type I NB lineages in VNCs than in larval brains ( Figure 7H–H′ , J , Figure 7—figure supplement\",\n \"1 ) ( Zhu et al . , 2011 ) .\",\n \"Since Btd is expressed in much more type I NB lineages in VNCs than in larval brains and Btd is required to prevent the premature differentiation of INPs , we wondered if ectopic PntP1-induced generation of INP-like cells requires Btd activity and occurs mostly in Btd-positive type I NB lineages .\",\n \"To test this idea , we examined if co-expression of PntP1 and Btd in type I NB lineages was sufficient to induce the generation of INPs and if the ectopic PntP1-induced generation of INP-like cells would be impaired in the absence of Btd .\",\n \"To coexpress PntP1 and Btd in type I NB lineages , we used either btd-GAL4 to drive the expression of UAS-pntP1 in Btd-positive type I NB lineages or insc-GAL4 to drive the expression of UAS-pntP1 and UAS-btd simultaneously in all type I NB lineages .\",\n \"INP-like cells were identified by their expression of Ase and Dpn as well as INP-specific marker R9D11-CD4-tdTomato .\",\n \"Since the GAL4 insertion in the btd-GAL4 line causes a lethal mutation in btd ( Estella et al . , 2003 ) , we examined the phenotype of the expression of UAS-pntP1 driven by btd-GAL4 only in btd-GAL4 heterozygous female larvae .\",\n \"As shown in Figure 6A–A′ , type II NB lineages in btd-GAL4 heterozygous mutant larvae are indistinguishable from those in wild-type animals ( e . g . Figure 1A–A′ ) .\",\n \"Furthermore , as mentioned above , btd-GAL4 homozygous mutant type II NB clones develop normally ( Figure 6—figure supplement 1 ) .\",\n \"Therefore , the generation of INPs is not affected in the btd-GAL4 line .\",\n \"Interestingly , ectopic expression of PntP1 using btd-GAL4 as a driver induced the generation of INP-like cells in about 90% Btd-positive type I NB lineages in both larval brains ( Figure 7A–C′ , J ) and VNCs ( Figure 7—figure supplement 1 ) .\",\n \"Consistently , the co-expression of UAS-pntP1 and UAS-btd driven by insc-GAL4 induced INP-like cells in about 95% of type I NB lineages in both larval brains and VNCs ( Figure 7F–F′ , I–I′ , J , Figure 7—figure supplement 1 ) .\",\n \"In contrast , the expression of UAS-pntP1 alone driven by insc-GAL4 only induced INP-like cells in about 10% and 46% of type I NB lineages in larval brains ( Figure 7H–H′ , J ) and VNCs ( Figure 7—figure supplement 1 ) , respectively , although ectopic PntP1 expression suppressed Ase in nearly all type I NBs ( Figure 7H , Figure 7—figure supplement 1 ) .\",\n \"The expression of UAS-btd alone neither suppressed Ase nor induced the generation of INP-like cells in VNCs ( Figure 7—figure supplement 1 ) , whereas in larval brains , the expression of UAS-btd driven by insc-GAL4 suppressed/reduced the expression of Ase in the NB and promoted the generation of INP-like cells in about 20% of type I NB lineages ( Figure 7G–G′ , J ) .\",\n \"These ectopic INP-like cells induced by the expression of UAS-btd also expressed INP-specific marker R9D11-CD4-tdTomato ( Figure 7—figure supplement 2 ) .\",\n \"These results indicate that the generation of INP-like cells induced by the ectopic PntP1 expression requires Btd activity , whereas the expression of either PntP1 or Btd alone has limited ability to induce the generation of INP-like cells in type I NB lineages .\",\n \"To further confirm if Btd is required for ectopic PntP1 to induce the generation of INP-like cells , we then examined if the generation of INP-like cells induced by ectopic PntP1 expression would be impaired in the absence of Btd .\",\n \"To this end , we expressed UAS-pntP1 in wild-type or btd mutant type I NB clones in VNCs in that there are more Btd-positive type I NB lineages and the expression of UAS-pntP1 can induce INP-like cells in much more type I NB lineages in VNCs than in larval brains .\",\n \"Consistent with the induction of INP-like cells in nearly all type I NB lineages when PntP1 and Btd were coexpressed , the efficiency of PntP1 to induce the generation of INP-like cell was drastically reduced in the absence of Btd .\",\n \"Our results showed that the expression of UAS-pntP1 driven by tub-GAL4 could have induced the generation of INP-like cells in about 50% of wild-type type I NB clones but not in btd mutant type I NB clones , although the expression of PntP1 equally suppressed the expression of Ase in both wild-type and btd mutant type I NBs ( Figure 7D–E′ , J ) .\",\n \"These results provide additional evidence to support that only in the presence of Btd could PntP1 induce the generation of INP-like cells in type I NB lineages .\",\n \"Taken together , these results suggest that the generation of INPs requires the cooperative action of PntP1 and Btd .\",\n \"Thus this study together with our previous work ( Zhu et al . , 2011 ) identified two key factors , PntP1 and Btd , a combination of which is sufficient to specify type II NB lineages and promote INP generation .\"\n ],\n [\n \"The most striking phenotype resulting from the loss of Btd is the elimination of mature INPs .\",\n \"In addition , about 40% of btd mutant type II NB lineages ectopically express Ase in the NB and become type I-like NB lineages .\",\n \"However , although forced expression of Ase in type II NBs is sufficient to eliminate INPs in type II NB lineages ( Bowman et al . , 2008; Zhu et al . , 2012 ) , the loss of INPs is obviously not primarily due to the ectopic Ase expression or the transformation of type II NB lineages into type I-like NB lineage in that the loss of mature INPs occurs independently of the ectopic Ase expression in most btd mutant or Btd RNAi knockdown type II NB lineages .\",\n \"Instead , we demonstrate that the loss of mature INPs in the absence of Btd is due to the premature differentiation of Ase+ immature INPs into GMCs .\",\n \"We show that in Btd RNAi knockdown or btd mutant type II NB lineages without the ectopic Ase expression , Ase− immature INPs differentiate into Ase+ immature INPs normally as indicated by the expression of R9D11-mCD8-GFP , Mira , as well as PntP1 in Ase+ daughter cells next to the Ase− immature INPs .\",\n \"However , instead of differentiating into mature INPs , we argue that Ase+ immature INPs prematurely differentiate into GMCs based on the following two pieces of evidence .\",\n \"First , Ase+ daughter cells eventually undergo terminal divisions as indicated by the positive pH3 staining and the position of the pH3-positive cells .\",\n \"Second , unlike mature INPs , the dividing Ase+ daughter cells do not form basal Mira crescent at metaphase .\",\n \"The terminal division and the lack of Mira crescent during the division are two unique features that distinguish GMCs from INPs in addition to the expression of nuclear Pros ( Ikeshima-Kataoka et al . , 1997; Matsuzaki et al . , 1998; Schuldt et al . , 1998 ) .\",\n \"Therefore , the elimination of mature INPs resulting from the loss of Btd is due to the premature differentiation of Ase+ immature INPs into GMCs .\",\n \"Why does the loss of Btd lead to premature differentiation of INPs ?\",\n \"Our results show that the loss of Btd results in a reduction or loss of PntP1 in type II NBs and immature INPs as well as ectopic expression of Pros in early immature INPs .\",\n \"Our previous studies show that PntP1 suppresses Ase in type II NBs and that inhibiting PntP1 activity leads to ectopic expression of Ase in type II NBs and elimination of INPs ( Zhu et al . , 2011 ) .\",\n \"Given that the ectopic Ase expression in btd mutant type II NBs is closely associated with the severe reduction or complete loss of PntP1 and that expression of UAS-pntP1 largely suppresses the ectopic Ase expression in btd mutant type II NBs , the severe reduction or loss of PntP1 most likely accounts for the ectopic Ase expression in btd mutant type II NBs .\",\n \"However , although the loss of PntP1 could lead to the loss of INPs , we provide several lines of evidence to demonstrate that the elimination of INPs in btd mutant or Btd RNAi knockdown type II NB lineages is primarily due to the ectopic activation of Pros in immature INPs rather than the reduction or loss of PntP1 .\",\n \"First , ectopic nuclear Pros is consistently expressed in Ase− immature INPs when mature INPs are eliminated .\",\n \"Second , the loss of mature INPs can be fully rescued by Pros RNAi knockdown or even just by removing one wild-type copy of pros .\",\n \"Third , Pros RNAi knockdown also rescues the reduction of PntP1 and suppresses the ectopic Ase expression in btd mutant type II NBs .\",\n \"In contrast , the expression of UAS-pntP1 fails to rescue mature INPs in most btd mutant type II NB lineages although it largely suppresses the ectopic Ase expression in the NBs .\",\n \"Furthermore , the complete elimination of mature INPs is also observed occasionally in btd mutant type II NB lineages without the reduction of PntP1 .\",\n \"Therefore , the elimination of mature INPs resulting from the loss of Btd is primarily due to the ectopic Pros expression , which likely promotes premature differentiation of INPs into GMCs and cell cycle exit .\",\n \"The severe reduction or loss of PntP1 is responsible for the ectopic Ase expression in btd mutant type II NBs and is more likely a secondary effect due to the ectopic Pros expression and/or the loss of INPs .\",\n \"INPs and/or other progeny may provide feedback signals to the NBs as has been demonstrated in other systems ( Yoon et al . , 2008; Hsu et al . , 2014 ) .\",\n \"The ectopic expression of Pros in Ase− immature INPs resulting from the loss of Btd suggests that Btd is critical for suppressing Pros expression in Ase− immature INPs .\",\n \"Btd was known as a head gap gene .\",\n \"It has been suggested that gap factors act largely as transcriptional repressors ( Schroeder et al . , 2004 ) .\",\n \"Btd could directly suppress Pros by binding to the pros promoter as a transcriptional repressor .\",\n \"Alternatively , Btd could suppress Pros indirectly by regulating the expression or antagonizing the activity of factor ( s ) that activate ( s ) pros expression .\",\n \"Our results show that ectopic/overly expression of Btd in type I NB lineages or mature INPs does not lead to overproliferation of type I NBs as observed in pros mutant type I NB lineages .\",\n \"Instead , ectopic expression of Btd promotes the generation of INP-like cells from type I NBs and transforms some type I NB lineages into type II-like NB lineages .\",\n \"Therefore , it is more likely that Btd suppresses Pros indirectly by regulating the expression or antagonizing the activity of pros activator ( s ) .\",\n \"Previous studies have suggested that Ase , Daughterless , Numb , and Erm could activate pros expression ( Reddy and Rodrigues , 1999; Southall and Brand , 2009; Weng et al . , 2010; Yasugi et al . , 2014 ) .\",\n \"Since Ase and R9D11-Cd4-tdTomato , which are under the control of erm promoter , are not expressed in Ase− immature INPs in the absence of Btd , it is unlikely that they are involved in the activation of pros in immature INPs .\",\n \"It would be interesting to investigate in the future if Numb or Daughterless could activate pros in immature INPs in the absence of Btd .\",\n \"In this study , we provided several lines of evidence to demonstrate that Btd and PntP1 function cooperatively to specify type II NB lineages and promote the generation of INPs .\",\n \"Results from this study as well as our previous study ( Zhu et al . , 2011 ) show that ectopic expression of UAS-pntP1 or UAS-btd alone can only promote the generation of INP-like cells in a subset of type I NB lineage , whereas ectopic expression of UAS-pntP1 in Btd-positive type I NB lineages or co-expression of UAS-btd and UAS-pntP1 can promote the generation of INP-like cells in nearly all type I NB lineages and transforms all these lineages into type II-like NB lineages .\",\n \"Consistently , the ability of PntP1 to promote the generation of INP-like cells in btd mutant type I NB lineages is largely impaired .\",\n \"These results suggest that the specification of type II NB lineages and the generation of INPs requires both PntP1 and Btd and that the combinatorial PntP1 and Btd is sufficient to promote the generation of INPs .\",\n \"We propose that PntP1 and Btd function cooperatively but through different mechanisms to promote INP generation .\",\n \"PntP1 is responsible for the suppression of Ase in type II NBs .\",\n \"Meanwhile , PntP1 must be regulating the expression of other unknown target gene ( s ) that are/is essential for the generation of INPs , such as specification of immature INPs , because loss of Ase is not sufficient to promote the generation of INP-like cells in any type I NB lineages .\",\n \"Btd likely acts after PntP1 to mainly prevent premature differentiation of INPs into GMCs by indirectly suppressing pros in immature INPs .\",\n \"The role of Btd in suppressing Ase in type II NBs is minimal if there is any because unlike PntP1 , which suppresses ase in nearly all type I NBs when it is ectopically expressed , overexpression of Btd only suppresses Ase in a small subset of type I NBs that produce INP-like cells in larval brains .\",\n \"Furthermore , Ase is expressed in Btd+ type I NBs , indicating that Btd does not suppress Ase in type I NBs when it is expressed at normal levels .\",\n \"Studies in mammals as well as in Drosophila suggest that the Btd/Sp8 could function downstream of Wnt signaling to regulate the expression of Fgf8 as well as Distal-less ( Dll ) and Headcase ( Hdc ) during the forebrain patterning as well as limb development ( Estella et al . , 2003; Treichel et al . , 2003; Kawakami et al . , 2004; Sahara et al . , 2007; Estella and Mann , 2010 ) .\",\n \"However , inhibiting Wnt signaling alone in type II NB lineages does not have any obvious phenotypes ( Komori et al . , 2014 ) , indicating that Btd unlikely functions downstream of Wnt signaling in type II NB lineages .\",\n \"Whether Fgf8 , Dll , or Hdc could function downstream of Btd to regulate INP generation remains to be investigated in the future .\",\n \"In mammals , the Btd homolog Sp8 plays important roles in brain development .\",\n \"In the developing mouse forebrain , Sp8 is expressed in cortical progenitors in a mediolateral gradient across the ventricular zone as well as in the lateral ganglionic eminence ( LGE ) and medial ganglionic eminence ( MGE ) ( Sahara et al . , 2007; Zembrzycki et al . , 2007 ) .\",\n \"In developing human brains , Sp8 is abundantly expressed in the ventricular zone and the outer sub-ventricular zone where RGs and oRGs reside ( Ma et al . , 2013 ) .\",\n \"In addition to its roles in interneuron development and the patterning of developing mammalian brains and spinal cords ( Griesel et al . , 2006; Waclaw et al . , 2006; Sahara et al . , 2007; Li et al . , 2011 ) , it was also shown that the loss of Sp8 led to the reduction of the progenitor pool ( Zembrzycki et al . , 2007 ) .\",\n \"Our results show that mammalian Sp8 can rescue the loss of mature INPs resulting from the loss of Btd in Drosophila , suggesting that Btd/Sp8 could have conserved functions across different species .\",\n \"It would be interesting to investigate if Sp8 has similar roles in promoting the generation of transient amplifying INPs , such as oRGs , in developing mammalian brains .\"\n ],\n [\n \"For btd loss-of-function analyses , yw btdXA FRT19A/FM7c and yw btdXG81 FRT19A/FM7c ( Wimmer et al . , 1993; Estella and Mann , 2010 ) were used for generating btd mutant clones and UAS-btd RNAi ( #29453; Bloomington Drosophila stock Center , Bloomington , Indiana ) for Btd RNAi knockdown .\",\n \"Type II NB lineage-specific pntP1-GAL4 ( also named as GAL414−94 ) ( Zhu et al . , 2011 ) and erm-GAL4 ( II ) or ( III ) ( Pfeiffer et al . , 2008; Xiao et al . , 2012 ) were used to drive the expression of UAS-transgenes in type II NB lineages or in immature as well as mature INPs , respectively .\",\n \"insc-Gal4 ( Gal41407 inserted in inscuteable promoter ) ( Luo et al . , 1994 ) was used to drive UAS-transgenes in all NB lineages .\",\n \"UAS-mCD8-GFP driven by Btd-GAL4 was utilized as reporter for btd expression .\",\n \"The R9D11-CD4-tdTomato transgenic line ( Han et al . , 2011 ) was used for labeling Ase+ immature INPs and mature INPs .\",\n \"Other fly stocks include: hs-Flpase tub-GAL80 FRT19A; UAS-mCD8-GFP; pntP1-GAL4 for generating type II NB clones; pros17/TM6 Tb , pros10419/Tm3 Sb , and UAS-pros RNAi ( #26745; Bloomington stock ) for rescuing Btd loss-of-function phenotypes .\",\n \"For RNAi knockdown analyses of Btd and Pros , larvae were raised at 29°C to increase the expression of UAS-RNAi transgenes after hatching .\",\n \"Furthermore , UAS-Dcr2 was expressed together with UAS-RNAi transgenes to enhance the efficiency of RNAi knockdown .\",\n \"For clonal analyses , MARCM ( Lee and Luo , 1999 ) clones were induced by 1 hr heat shock at 38°C for 1 day after larval hatching .\",\n \"Larval brains were dissected at third instar larval stages for the examination of phenotypes .\",\n \"Larval brains were dissected , fixed , and stained as described before ( Lee and Luo , 1999 ) .\",\n \"Primary antibodies used in this study include: rabbit anti-Mira ( 1:500 ) , guinea pig anti-Ase ( 1:5000 ) , rabbit anti-Dpn ( 1:500 ) ( a gift from Y . N . Jan ) , rat anti-mCD8 ( Life Technologies , Grand Island , New York , 1:100 ) , mouse anti-Pros ( Developmental Studies Hybridoma Bank , Iowa City , Iowa , 1:20 ) , mouse monoclonal anti-α-tubulin ( Sigma , St . Louis , Missouri , 1:1000 ) , rabbit anti-dsRed ( Clontech , Mountain View , California , 1:500 ) , rabbit anti-PntP1 ( 1:500 , a gift from JS Skeath ) .\",\n \"Secondary antibodies conjugated to Cy2 , Cy3 , Cy5 , or DyLight 647 ( Jackson ImmunoResearch , West Grove , Pennsylvania ) were used at 1:100 , 1:500 , or 1:500 , respectively .\",\n \"Images were taken with a Zeiss LSM510 confocal microscopy and processed with Adobe Photoshop .\",\n \"For quantifications of the number of mature INPs or the percentage of type II NB lineages with mature INPs , we focus on the medial group of type II NB lineages ( lineages DM1–DM6 ) .\",\n \"Two-tailed t-tests were used for statistics analyses .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Intermediate neural progenitor cells ( INPs ) need to avoid differentiation and cell cycle exit while maintaining restricted developmental potential , but mechanisms preventing differentiation and cell cycle exit of INPs are not well understood .","In this study , we report that the Drosophila homolog of mammalian Sp8 transcription factor Buttonhead ( Btd ) prevents premature differentiation and cell cycle exit of INPs in Drosophila larval type II neuroblast ( NB ) lineages .","We show that the loss of Btd leads to elimination of mature INPs due to premature differentiation of INPs into terminally dividing ganglion mother cells .","We provide evidence to demonstrate that Btd prevents the premature differentiation by suppressing the expression of the homeodomain protein Prospero in immature INPs .","We further show that Btd functions cooperatively with the Ets transcription factor Pointed P1 to promote the generation of INPs .","Thus , our work reveals a critical mechanism that prevents premature differentiation and cell cycle exit of Drosophila INPs ."],"string":"[\n \"Intermediate neural progenitor cells ( INPs ) need to avoid differentiation and cell cycle exit while maintaining restricted developmental potential , but mechanisms preventing differentiation and cell cycle exit of INPs are not well understood .\",\n \"In this study , we report that the Drosophila homolog of mammalian Sp8 transcription factor Buttonhead ( Btd ) prevents premature differentiation and cell cycle exit of INPs in Drosophila larval type II neuroblast ( NB ) lineages .\",\n \"We show that the loss of Btd leads to elimination of mature INPs due to premature differentiation of INPs into terminally dividing ganglion mother cells .\",\n \"We provide evidence to demonstrate that Btd prevents the premature differentiation by suppressing the expression of the homeodomain protein Prospero in immature INPs .\",\n \"We further show that Btd functions cooperatively with the Ets transcription factor Pointed P1 to promote the generation of INPs .\",\n \"Thus , our work reveals a critical mechanism that prevents premature differentiation and cell cycle exit of Drosophila INPs .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Whereas the majority of cells in the brain are unable to divide to produce new cells , neural stem cells can divide numerous times and have the potential to become many different types of brain cells .","However , in between these two extremes there is another group of cells called neural progenitors .","These cells can give rise to multiple types of neurons but , in contrast to stem cells , they can undergo only a limited number of divisions .","Many of the molecular mechanisms by which stem cells give rise to progenitors are similar in mammals and in the fruit fly Drosophila .","In the brains of fruit fly larvae , neural stem cells called neuroblasts give rise to ‘intermediate neural progenitors' , each of which can divide between four and six times . Every division generates a replacement intermediate progenitor and a cell called a GMC , which divides one last time to produce two brain cells . Intermediate progenitors must be tightly regulated to ensure that they undergo an appropriate number of divisions: too few divisions will result in a shortage of cells , disrupting brain development , whereas too many divisions will result in the formation of tumors . Now , using Drosophila brains in the laboratory , Xie et al . —and , independently , Komori et al . —have shown that a protein called ‘Buttonhead’ is responsible for maintaining this balance .","Xie et al . show that deletion of the gene for Buttonhead gene caused the progenitor cells to become GMCs before they had undergone the correct number of divisions .","Further experiments revealed that Buttonhead prevents this problem by suppressing a protein called Prospero .","The mammalian equivalent of Buttonhead—a protein called Sp8—can substitute for Buttonhead in Drosophila neural progenitors , suggesting that the observed mechanisms may also apply to mammals .","Further work is required to test this possibility directly and to examine the involvement of Sp8 in brain development and tumor formation ."],"string":"[\n \"Whereas the majority of cells in the brain are unable to divide to produce new cells , neural stem cells can divide numerous times and have the potential to become many different types of brain cells .\",\n \"However , in between these two extremes there is another group of cells called neural progenitors .\",\n \"These cells can give rise to multiple types of neurons but , in contrast to stem cells , they can undergo only a limited number of divisions .\",\n \"Many of the molecular mechanisms by which stem cells give rise to progenitors are similar in mammals and in the fruit fly Drosophila .\",\n \"In the brains of fruit fly larvae , neural stem cells called neuroblasts give rise to ‘intermediate neural progenitors' , each of which can divide between four and six times . Every division generates a replacement intermediate progenitor and a cell called a GMC , which divides one last time to produce two brain cells . Intermediate progenitors must be tightly regulated to ensure that they undergo an appropriate number of divisions: too few divisions will result in a shortage of cells , disrupting brain development , whereas too many divisions will result in the formation of tumors . Now , using Drosophila brains in the laboratory , Xie et al . —and , independently , Komori et al . —have shown that a protein called ‘Buttonhead’ is responsible for maintaining this balance .\",\n \"Xie et al . show that deletion of the gene for Buttonhead gene caused the progenitor cells to become GMCs before they had undergone the correct number of divisions .\",\n \"Further experiments revealed that Buttonhead prevents this problem by suppressing a protein called Prospero .\",\n \"The mammalian equivalent of Buttonhead—a protein called Sp8—can substitute for Buttonhead in Drosophila neural progenitors , suggesting that the observed mechanisms may also apply to mammals .\",\n \"Further work is required to test this possibility directly and to examine the involvement of Sp8 in brain development and tumor formation .\"\n]","isConversational":false},"year":{"kind":"string","value":"2014"}}},{"rowIdx":95,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["structural biology and molecular biophysics"],"string":"[\n \"structural biology and molecular biophysics\"\n]","isConversational":false},"title":{"kind":"string","value":"Molecular architecture of the yeast Mediator complex"},"id":{"kind":"string","value":"elife-08719-v2"},"sections":{"kind":"list like","value":[["Mediator is a complex of at least 21 subunits , with a mass of greater than 1 million Daltons , conserved from yeast to humans ( reviewed in Kornberg , 2005; Conaway and Conaway , 2011; Ansari and Morse , 2013 ) .","Genetic studies have shown that Mediator is essential for RNA polymerase II ( pol II ) transcription and required for positive and negative regulation of transcription .","Biochemical studies have confirmed the importance of Mediator for both basal and regulated transcription , and have demonstrated interactions of Mediator with pol II , with general transcription factors , and with transcriptional activator proteins .","Through these interactions , Mediator is thought to promote assembly of the entire machinery for pol II transcription .","Electron microscopy ( EM ) has shown a division of Mediator in three modules , termed Head , Middle , and Tail ( Asturias et al . , 1999 ) .","EM studies further showed that Mediator structure is conserved from yeast to humans , that Mediator is compact when free in solution , and that a major structural rearrangement occurs upon interaction with pol II , which is surrounded by the Head and Middle modules in the so-called ‘holoenzyme’ ( Asturias et al . , 1999; Dotson et al . , 2000 ) .","Details of subunit interaction networks within Mediator modules have come from co-expression studies ( Koh et al . , 1998; Lee et al . , 1998; Kang et al . , 2001; Koschubs et al . , 2010 ) and from Mediator-specific ( Guglielmi et al . , 2004 ) and genome-wide ( Uetz et al . , 2000; Ito et al . , 2001 ) two-hybrid approaches .","The currently accepted subunit assignments are Med 6-8-11-17 ( Srb4 ) -18 ( Srb5 ) -20 ( Srb2 ) -22 ( Srb6 ) in the Head module , Med 1-4-7-9 ( Cse2 ) -10 ( Nut2 ) -19 ( Rox3 ) -21 ( Sbr7 ) -31 ( Soh1 ) in the Middle module , and Med 2-3 ( Pgd1 ) -5 ( Nut1 ) -14 ( Rgr1 ) -15 ( Gal11 ) -16 ( Sin4 ) in the Tail module ( old naming convention in parentheses ) .","X-ray crystal structures of Head modules from Saccharomyces cerevisiae and Schizosaccharomyces pombe have been determined , with ( Robinson et al . , 2012 ) and without ( Imasaki et al . , 2011; Lariviere et al . , 2012 ) the carboxy-terminal domain ( CTD ) of the largest pol II subunit , Rpb1 , bound to a conserved site on the surface .","CTD interaction with Head and Middle modules plays an important role in the association of Mediator with pol II and the pre-initiation complex .","Beyond the X-ray crystallography of the Head module , structural information on Mediator is limited .","Only small portions of the Middle ( Baumli et al . , 2005; Koschubs et al . , 2009 ) and Tail modules ( Bontems et al . , 2011; Eletsky et al . , 2011; Vojnic et al . , 2011 ) have been solved at near atomic resolution .","EM at low resolution with subunit-labeling has given an approximate idea of the spatial organization of the modules ( Tsai et al . , 2014; Wang et al . , 2014 ) , and an architectural model for a subcomplex of the Middle module ( lacking Med 1 and 19 ) has been proposed on the basis of cross-linking data and homology modeling ( Lariviere et al . , 2013 ) .","Recently , reconstituted yeast Head and Middle Mediator modules were visualized by EM in a complex with pol II and a subset of general transcription factors ( Plaschka et al . , 2015 ) .","Mediator is representative of a large number of macromolecular complexes that are refractory to classical high-resolution structural approaches , due to low native abundance and to conformational and compositional heterogeneity .","For the study of such systems , data from chemical cross-linking and mass spectrometry can provide multi-protein interaction maps at residue-level resolution ( Murakami et al . , 2013; Erzberger et al . , 2014; Shi et al . , 2014 ) .","The approach is particularly effective when combined with other structural information , such as electron density maps from EM ( Murakami et al . , 2013 ) and subunit structures from X-ray crystallography ( Robinson et al . , 2012 ) .","To construct structural models of macromolecular assemblies by combining such diverse types of information , the open-source Integrative Modeling Platform ( IMP ) program was developed ( http://integrativemodeling . org ) ( Russel et al . , 2012 ) .","IMP translates the data into spatial restraints , computes an ensemble of models by satisfying these restraints as well as possible , and finally assesses the ensemble against the data used or not used in its construction ( Schneidman-Duhovny et al . , 2014 ) .","IMP has been successfully applied to a number of protein complexes ( Lasker et al . , 2012; Erzberger et al . , 2014; Shi et al . , 2014 ) .","Here we determined the molecular architecture of Mediator by integrative structure determination , based on chemical cross-links , X-ray crystal structures , homology models , and a cryo-EM electron density map , with the use of IMP .","Possible configurations of the Mediator subunits were exhaustively sampled to identify those that best satisfied the experimental restraints .","The resulting model was validated by re-computing it with random subsets of chemical cross-links ( i . e . , ‘jackknifing’ ) ( Brunger et al . , 1993 ) as well as by comparison with known protein structures and protein interaction networks ( Guglielmi et al . , 2004 ) .","The Mediator model will serve as a basis for studies of Mediator interaction with pol II , general transcription factors , gene activators , and gene repressors ."],["Mediator was isolated from yeast as originally described ( Kim et al . , 1994 ) in the form of a complex with pol II ( holoenzyme ) .","The complex is not only a functionally relevant form of Mediator , but is also more soluble than free Mediator under the conditions used for lysine directed cross-linking .","We employed a double affinity-tagging strategy ( Figure 1—figure supplement","1 ) and obtained 20-fold more stoichiometric holoenzyme than from previous purification procedures .","The native complex was stable , persisting throughout purification , whereas a complex formed from separately isolated Mediator and pol II was disrupted by the same purification procedures .","Cross-linking and mass spectrometry were performed on the native holoenzyme , with a mixture of D12-labelled and unlabeled BS3 cross-linking reagents , or with unlabeled BS3 and enrichment of cross-linked trypsin fragments by gel filtration .","The BS3 reagent bridges free-amine moieties at surface lysines and N-termini with Cα spacing within approximately 25 Å .","Cross-links were assigned with the Protein Prospector platform and a refined scoring function ( Trnka et al . , 2014 ) .","The 320 , 767 initial product ion spectra were searched against a database containing the sequences of the yeast transcriptional machinery as well as 520 randomized decoy sequences , identifying 20 , 388 potential cross-linked spectra ( Figure 1A ) .","Classification of these matches and the application of quality filters resulted in the identification of 3196 holoenzyme cross-linked spectra at a false discovery rate ( FDR , fraction of decoy hits above score cutoff ) of 0 . 7% ( Figure 1A ) .","These spectra defined 402 unique pairs of linked amino acid residues ( ‘cross-links’ ) with a FDR of 4 . 1% ( Figure 1—source data 1 , Figure 1—figure supplement 2 , Supplementary file 1 ) .","The majority ( 260 cross-linked fragments ) were between Mediator proteins , 124 were between pol II subunits , and 18 were between Mediator and pol II . 10 . 7554/eLife . 08719 . 003Figure 1 . Holoenzyme cross-linking and modeling results and methodology .","( A ) Cross-links were identified by searching MS2 product ion spectra against concatenated target and decoy databases containing 52 Saccharomyces cerevisiae transcription proteins +520 sequence randomized versions , respectively .","Confidence in the spectral assignment is represented by an support vector machine ( SVM ) classifier ( Trnka et al . , 2014 ) .","The distribution of the target and decoy spectral matches in relation to the acceptance threshold ( red line ) is shown with an overall spectral false discovery rate ( FDR ) of 0 . 7% .","( B ) Mapping identified cross-linked spectra onto regions of known structure such as the Mediator Head module and RNA polymerase II yields distance distributions that reflect the accuracy of cross-link identification .","The Cα violation distance of 35 Å is indicated with a dashed line .","( C ) Demonstration of RNA pol II structural recapitulation from a random starting configuration ( top panel ) of individual pol II subunits represented as rigid bead models and restrained by cross-links ( green links ) and electron microscopy ( EM ) density map ( mesh ) .","A representation of the converged modeling solution is presented alongside the 12-subunit pol II X-ray structure ( PDB: 1WCM ) for reference .","( D ) Schematic representation of the Mediator complex cross-linking network .","Mediator subunits are colored according to their location within the Head , Middle and Tail modules .","All Inter-subunit cross-links and selected intra-subunit cross-links from the current study as well from ( Lariviere et al . , 2013 ) are represented , colored according to their origin . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00310 . 7554/eLife . 08719 . 004Figure 1—source data 1 . Categorization of cross-links with respect to module and data source . Numbers of unique cross-linked residue pairs ( ‘cross-links’ ) and cross-linked spectral matches used for integrative modeling of Mediator apo-complex ( ‘Total Mediator’ ) as well as total number of holoenzyme cross-links identified in this study ( ‘Total [this study]’ ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00410 . 7554/eLife . 08719 . 005Figure 1—figure supplement 1 . Purification of Native S . cerevisiae Holoenzyme complex .","( A ) Schematic of the affinity capture and chromatographic fractionation of S . cerevisiae holoenzyme complex from whole-cell extract using a double-affinity tag and stepwise cleavage protocol .","In the procedure , a mixture of holoenzyme and free Mediator and RNA Pol II are affinity captured , with release of free Mediator through TEV cleavage ( panel B ) and subsequent release of holoenzyme and free Mediator upon 3C protease cleavage ( panel C ) .","Following the enrichment of Mediator-bound pol II using ion-exchange chromatography , the holoenzyme was purified to homogeneity using size-exclusion chromatography ( panel D ) .","Mass spectrometry analysis confirmed the presence of all 21 Mediator and 12 pol II subunits ( panel E ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00510 . 7554/eLife . 08719 . 006Figure 1—figure supplement 2 . Annotated product ion spectra for selected cross-linked peptides . HCD product ion spectra of BS3 cross-linked Mediator tryptic peptides classified by Protein Prospector .","Spectra representing highly confident ( SVM score 5 . 1 ) to less confident ( SVM score 0 . 0 ) classifications are shown .","The modified lysine position is denoted with a star .","N-terminal acetylation and carbamidomethylation of cysteine residues ( Cam ) are denoted .","P , PL , and PLK ions refer to ions formed from dissociation of the cross-linking reagent to lysine bond ( Trnka et al . , 2014 ) .","The full list of cross-linked Mediator peptides may be viewed online using search key = qzpxihtngx at: http://prospector2 . ucsf . edu/prospector/cgi-bin/msform . cgi ?","form=msviewer . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 006 The pol II component of the holoenzyme provided an internal control , enabling validation of the cross-link dataset by two approaches .","First , the X-ray crystal structure of pol II and also that of the Mediator Head module could be used to assess the validity of cross-links , as 130 of the 402 high confidence cross-links were between residues in these structures .","Only five of these 130 cross-links occurred between residues further apart than the 35 Å threshold ( Figure 1B ) .","This number of violations was therefore consistent with the FDR of 4 . 1% , estimated from the number of decoy database hits .","Second , we constructed an integrative model of pol II based on the cross-link dataset , a cryo-EM electron density map ( EMD-1883 , 20 . 9 Å resolution using FSC 0 . 5 ) , and X-ray crystal structures of the pol II subunits , using IMP .","The X-ray crystal structures of the subunits were represented as independent rigid bodies , and the regions not observed in the structure were represented by flexible strings of beads .","A low resolution EM map was used for comparison with the EM data available for Mediator ( see below ) .","Starting from scrambled configurations , the positions and orientations of the pol II subunits were optimized with a Monte Carlo procedure ( which applies a large number of random movements guided by the input data ) , resulting in configurations that satisfied the input data well .","This ensemble of good-scoring solutions recapitulated the known architecture of pol II ( Figure 1C ) .","The average root-mean-square deviation ( RMSD ) of the solutions with respect to the crystal structure of pol II was 14 Å .","All subunits were correctly placed , except for Rpb8 and Rpb9 , whose locations were undetermined , due to a lack of cross-link restraints .","We augmented our Mediator cross-link dataset with 38 Mediator Middle module cross-links from a dataset collected on a recombinant preparation of S . cerevisiae Middle module containing 6 subunits ( Lariviere et al . , 2013 ) .","The 38 cross-links were selected according to our minimal length requirement of four amino acid residues per cross-linked peptide; three cross-links were common to both datasets .","This modest overlap is likely due to the use of a different chemical cross-linker , as well as to the presence of additional subunits in our full native Middle module .","The complete dataset of 294 inter-subunit cross-links defined an interaction map of 19 of the 21 Mediator proteins ( Figure 1D ) .","The two subunits for which there were no inter-subunit cross-links have been crystallized in complexes with other Mediator subunits: Med6 as part of the Head module ( Robinson et al . , 2012 ) , and Med31 in a complex with the N-terminal region of Med7 ( Koschubs et al . , 2009 ) .","Both proteins were present in our native holoenzyme preparations , as shown by SDS PAGE and mass spectrometry ( Figure 1—figure supplement 1 ) .","Two subunits , Med17 ( Head module ) and Med14 ( Tail module ) showed patterns of cross-links split between two or more different Mediator modules .","The majority of Med17 ( residues 123–687 ) formed an extensive cross-link network with other subunits of the Head module , as expected from the X-ray crystal structure ( Robinson et al . , 2012 ) .","The N-terminal region of Med17 ( 1–122 ) , not observed in the Head module crystal structure , formed many cross-links with Med7 and Med21 , subunits of the Middle module ( Figure 1D ) .","Similarly , Med14 , a presumptive Tail module subunit , formed cross-links to other Tail module proteins ( Med2 and 15 ) , but almost exclusively in its C-terminal region ( residues 712–1082 ) , while the N-terminal region and the majority of the protein ( residues 1–711 ) cross-linked instead to Middle module subunits ( Med4 , Med9 , Med10 , and Med21 ) .","The N-terminal regions of Med17 and Med14 not only extended into the Middle module but also formed many cross-links with each other .","For these reasons , we redefine the Mediator modules to include the N-terminal domains of Med17 ( residues 1–122 ) and Med14 ( residues 1–711 ) in the Middle module rather than the Head and Tail modules ( Figure 1D , Med14 and Med17 colored by module localization ) .","Although cross-linking was performed on the holoenzyme , about 90% of the cross-links in our dataset were within Mediator modules ( intra-modular ) or within pol II .","Due to the paucity of cross-links between Mediator and pol II , and because a cryo-EM map at sufficient resolution for molecular modeling was only available for Mediator without pol II , modeling calculations were performed for Mediator alone .","The four-step modeling ( ‘integrative structure determination’ ) procedure ( Figure 2—figure supplement","1 ) entailed: ( 1 ) gathering data; ( 2 ) representing and translating the data into spatial restraints; ( 3 ) sampling the conformational space and identifying good scoring solutions; and ( 4 ) analyzing and assessing the ensemble of solutions .","All IMP input files , scripts , and output models for this study are available at http://salilab . org/mediator/ .","In step 1 , the data comprised our holoenzyme cross-link dataset , X-ray crystal structures of some subunits and subunit regions , homology models for some subunits , and the highest resolution cryo-EM map of free Mediator available ( Tsai et al . , 2014 ) .","In step 2 , we constructed a set of 21 Mediator subunit model representations ( Figure 2—figure supplement 2 ) .","Atomic models from X-ray crystallography or homology models were represented by rigid bodies ( Figure 2—figure supplement 2 , blue subunit shading ) , and unmodeled subunit regions were represented as flexible strings of beads ( Figure 2—figure supplement 2 , yellow subunit shading ) .","The entire Head module was represented as a rigid body ( based on the X-ray structure PDB 4GWP ) , with the exception of the unmodeled N-terminus of Med17 ( 1–181 ) , the last 103 residues of Med6 , and other missing regions , all of which were represented as flexible strings of beads .","Manual docking and the mapping of Med 8 , 18 , and 20 onto 2-D EM projections previously supported a single docking solution for the Head module in the EM density ( Tsai et al . , 2014 ) .","We performed an exhaustive docking calculation , comparing all possible Head module locations with a cross-correlation scoring function ( Materials and methods ) .","We found a docking solution that is consistent with the solution proposed previously , with a normalized cross-correlation coefficient approximately 40% better than the next best solution .","We modeled the Mediator complex with the Head module fixed in a single position corresponding to the highest scoring docking solution .","Other regions defined by atomic models were Med7C-Med21 and Med7N-Med31 , represented by rigid bodies based on their X-ray structures ( Baumli et al . , 2005; Koschubs et al . , 2009 ) , and Med4-Med9 , based on a homology model ( Lariviere et al . , 2013 ) .","In addition , we created a homology model of the first 540 residues of Med16 , also in a rigid body representation , based on a seven-bladed β-propeller fold ( proposed and justified below ) .","The EM map was divided into segments to reflect the modular organization of the complex .","Since the Head module was docked and fixed in a portion of the density map , the remaining density was segmented into regions corresponding to the Middle and Tail modules on the basis of previous approximate localization of Middle and Tail module subunits within two distinct regions of the EM map ( Tsai et al . , 2014; Wang et al . , 2014 ) .","During modeling , each Middle or Tail subunit was restrained to the corresponding Middle or Tail EM density .","EM spatial restraints were based on a Gaussian decomposition of the electron density ( Materials and methods ) .","The cross-linking data was encoded in a Bayesian scoring function ( Erzberger et al . , 2014; Shi et al . , 2014 ) .","The scoring function allowed the independent optimization of the relative weights of arbitrary subsets of the cross-linking data .","We split the cross-link dataset into two subsets composed of inter-module ( 8% of the total ) and intra-module ( 92% of the total ) cross-links ( Figure 1—source data 1 ) .","In step 3 , we conducted a large number of independent modeling runs .","The positions and orientations of rigid bodies and flexible strings of beads were repeatedly perturbed in an effort to satisfy the scoring function consisting of the excluded volume , sequence connectivity , EM , and cross-linking restraints .","A total of 165 , 523 Mediator model configurations were produced .","The 500 best-scoring models ( solutions ) were grouped on the basis of RMSD into four clusters ( C1-4 , Figure 2—figure supplement 3 ) , the minimum number required to fully represent the main structural differences between the best scoring models .","To confirm that we had sampled conformational space sufficiently to reach model convergence , we compared two independent halves of the solutions to each other and to the entire set , showing they were all similar to one another ( Figure 2—figure supplement 4 ) .","The solutions satisfied all excluded volume , sequence connectivity , and EM restraints .","Intra-module cross-links were satisfied to a high degree ( approximately 95% ) , and inter-modular cross-links to a much lesser degree ( about 10% ) ( Figure 2—figure supplement 3 ) .","In fact , the Bayesian scoring function automatically assigned a low scoring weight to the inter-modular cross-links , pointing to an inconsistency of the inter-modular cross-links in the holoenzyme dataset with the EM density map of the free Mediator .","The satisfaction of intra-modular cross-links suggests that the arrangement of Mediator subunits within modules of the holoenzyme is similar to that in the free Mediator , whereas the inconsistency of the inter-modular cross-links with the EM map suggests that motions of the modules relative to one another occur upon association with pol II in the holoenzyme .","Indeed , EM studies have indicated large motions about hinges between the modules , opening the compact structure of free Mediator for wrapping around pol II in the holoenzyme ( Asturias et al . , 1999; Davis et al . , 2002 ) .","Based on these observations , the inter-modular cross-links were not considered in our analysis of the free Mediator .","The four clusters of best-scoring solutions differed in two respects .","Cluster 3 , which represented a minor subset of best-scoring models ( ∼8% ) , differed from the other clusters by an inversion of orientation of the Middle module .","This cluster was disregarded because the inverted orientation was inconsistent with prior EM localization experiments ( Tsai et al . , 2014 ) and because of significantly worse intra-module cross-link violation statistics ( Figure 2—figure supplement 3D–E ) .","The remaining three clusters were virtually identical in the locations of all subunits except for Med5 , Med15 , and Med16 in the Tail module ( Figure 2—figure supplement 3C , G ) .","Only Cluster 1 showed an arrangement of these three subunits consistent with previous EM localization ( Figure 2—figure supplement 3H ) .","Cluster 1 also showed the best cross-link satisfaction statistics and had the best average score .","Cluster 1 was therefore taken to represent the true Mediator subunit architecture .","The overall precision of our integrative model , computed as the average RMSD of the solutions in Cluster 1 with respect to the cluster center , is 19 Å .","The precision of the Middle and Tail modules is 24 and 33 Å , respectively .","These values represent the average fluctuations of the individual residues or beads in 3-D space across the ensemble of solutions comprising the highest scoring cluster .","Precision defined in this way is not directly comparable to the resolution of an EM map , which refers to the uncertainty of the overall electron density , without regard to the placement of components within the density .","Although the precision of the Tail module is lower than that of the Middle or Head modules , it is sufficient to position the Tail subunits within the Tail EM density as features with high cross-link density , and consequently high local precision values ( such as the Middle-Tail junction ) , provide strong spatial restraints .","The overall precision of the Mediator model sufficed to determine the positions and orientations of all Mediator subunits , and 20–40 residue beads could be localized with precisions of 10–50 Å ( Figure 2—figure supplement 5 ) , depending on the density of cross-linking restraints .","We represented the cluster of solutions by a localization density ( Figure 2 ) , defined as the probability of observing a particular subunit at a specific point in space in the cluster of solutions .","Because the model is three-dimensional , it depicts the contact surfaces between subunits , with residue-level resolution at many points where cross-links are formed between them .","This contact information is especially rich and reliable in regions where multiple cross-links are formed and where contacts occur in multiple models in the cluster , as depicted by the darkest boxes in the domain interaction map ( Figure 3 ) .","Some cross-links anticipated from the domain interaction map were not observed , most likely because cross-linking depends on the occurrence of unprotonated , appropriately oriented , solvent–accessible pairs of lysine residues , and on the formation of tryptic peptides with physicochemical properties amenable to liquid chromatography and mass spectrometry . 10 . 7554/eLife . 08719 . 007Figure 2 . Mediator complex architecture .","( A ) Mediator subunit localization density map colored by individual subunit .","( B ) Mediator localization density map ( solid grey ) calculated from the highest scoring model cluster and shown at a threshold level ( T = 0 . 1 ) that most closely matches the volume of the EM density map used as 3D spatial restraint during modeling ( EMD-2634 , T = 0 . 35: Blue mesh ) .","The position of the three Mediator modules ( Head , Middle , and Tail ) is indicated .","( C ) Schematics of the architecture of the Mediator complex obtained from the localization density analysis .","Med14 and Med17-NTD are also schematically represented by splines .","( D ) Protein–protein interactions derived from published Yeast-two hybrid , immunoprecipitation and sub-complex isolation data ( Uetz et al . , 2000; Ito et al . , 2001; Kang et al . , 2001; Guglielmi et al . , 2004; Zhang et al . , 2004; Beve et al . , 2005; Koschubs et al . , 2010 ) .","The data is represented by a graph superposed on the EM density map of the complex ( Tsai et al . , 2014 ) .","Nodes are Mediator subunits , the edges are the observed proteomic interactions , and the green circles are isolated sub-complexes .","Green and red edges are interactions that are in agreement and in disagreement with the Mediator model , respectively .","( E ) Localization by labeling ( blue triangle ) and domain deletion ( areas encircled by light red closed lines ) mapped on the EM density map ( Tsai et al . , 2014; Wang et al . , 2014 ) .","Straight light-red lines represent domain deletion assays that split the complex . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00710 . 7554/eLife . 08719 . 008Figure 2—figure supplement 1 . Schematic of integrative structure determination highlighting the individual data inputs and the four stages in our approach . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00810 . 7554/eLife . 08719 . 009Figure 2—figure supplement 2 . Input Model Representations . Individual Mediator subunit model representations are grouped according by membership within the Head , Middle , and Tail modules .","For each subunit , a schematic representation of the model substructure is shown below the corresponding bead representation .","The subunit substructure is mapped onto the protein sequence and coloured according to atomic model coverage ( blue fill ) or bead structure ( yellow ) where each bead in the subunit model is represented by a single box in the schematic representation .","The multi-protein complex structures from which most subunit atomic models are derived are represented on the right . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00910 . 7554/eLife . 08719 . 010Figure 2—figure supplement 3 . Cluster analysis of the solution ensemble .","( A ) Number of satisfied intra-module cross-links plotted against the score for all 165 , 523 models .","The vertical brown line represents the score threshold used to select the best scoring 500 models .","( B ) RMSD distance matrix for the 500 best-scoring models .","The distance matrix is represented by a heat map , where the bin color is proportional to the RMSD distance of the corresponding solution pair .","( C ) Difference between the four clusters represented by a hierarchical graph .","Clusters 1 , 2 , and 4 have the middle module oriented in the ‘normal’ , expected orientation ( i . e . , Med1 and Med19 are at the bottom and at the top of the module , respectively ) .","The three clusters differ in the arrangement of the tail subunits .","Cluster 3 has the middle module oriented in a ‘flipped’ orientation ( i . e . , Med1 and Med19 are at the top and at the bottom of the module , respectively ) .","( D ) Detailed statistics of cross-link satisfaction for the whole dataset ( top ) , intra-module cross links ( middle ) , and inter-module cross-links ( bottom ) .","( E ) Student t-test analysis of the intra-module cross-link violation distance .","( F ) Precision of the four clusters .","( G ) Statistical properties of the four clusters .","( H ) Agreement between subunit localization in Cluster 1 ( left ) and the interpretation of localization data mapped on the EM density map of the Tail module ( right ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01010 . 7554/eLife . 08719 . 011Figure 2—figure supplement 4 . Exhaustiveness of sampling and robustness of cross-link data .","( A ) Comparison of localization density maps calculated from the ensemble of solutions ( 500 best-scoring models ) for the entire sample of models , the first half , the second half , and for jackknifing modeling runs ( where 10% of cross-links were randomly removed ) .","( B ) Comparison of localization density maps calculated for Cluster 1 for the entire sample of models , the first half , the second half , and for jackknifing runs .","( C , D , E )","Precision of Cluster 1 solutions ( diagonal values ) and average RMSD between Cluster 1 solutions ( off-diagonal values ) computed for the four ensembles , considering the whole Mediator complex , the Middle module , and the Tail module . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01110 . 7554/eLife . 08719 . 012Figure 2—figure supplement 5 . Representation of subunit position precision .","( A ) Root-mean-square fluctuation ( RMSF ) plots for each subunit of the Middle and Tail modules show how precisely the residues of all proteins were localized within the best cluster of models .","For each subunit , RMSF values are mapped onto a stick representation of their centroid model as a color gradient with low and high values of RMSF colored blue and red , respectively .","Each subunit model is annotated with several residue numbers for reference .","( B ) Precision plots show the relative localization precision of each subunit of the Mediator complex , grouped by membership in the Head , Middle , and Tail domains .","The high localization precision of Head module subunits is due to the Head module X-ray structure being fixed at a single position in the EM map .","Within the Middle module , the subunits comprising the backbone scaffold ( Med7-21-4-9 ) are all precisely localized . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01210 . 7554/eLife . 08719 . 013Figure 3 . Mediator subunit domain interaction matrix . Average domain–domain contact map calculated for the cluster of best-scoring solutions .","Long sequences are divided into domains of 200 residues .","The intensity of the color for each box is proportional to the fraction of models for which the contact between corresponding domains is formed .","Two domains are in contact when the surface of the beads of one domain is within 10 Å from the surface of any of the beads of the other domain . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 013 The Mediator model ( Figure 2A ) fits the EM map used as a modeling restraint ( Figure 2B and Figure 7—figure supplement 1 ) , satisfied most of the cross-link restraints ( Figure 2—figure supplement 3 ) , and was consistent with almost all data from previous subunit interaction and subunit localization studies ( Figure 2C–E ) .","The model explains subunit interactions deduced from co-expression trials ( Kang et al . , 2001; Beve et al . , 2005; Koschubs et al . , 2010 ) , pulldowns/immunoprecipitation ( Kang et al . , 2001; Zhang et al . , 2004 ) , and yeast two-hybrid analyses ( Uetz et al . , 2000; Ito et al . , 2001; Guglielmi et al . , 2004 ) ( Figure 2D , green lines ) .","There were only three discrepancies ( Figure 2D , red lines ) , corresponding to the Med3-Med21 , Med1-Med7 , and Med1-Med5 interactions .","The Med3-Med21 interaction , from two-hybrid data , is clearly spurious , because the proteins are found in different Mediator modules ( Head and Tail ) and our model , consistent with an EM localization study ( Tsai et al . , 2014 ) , places the two subunits more than 100 Å apart .","A second subunit interaction that is unsupported by our model , between the C-terminal domain of Med1 and Med5 , also comes from two-hybrid data ( Guglielmi et al . , 2004 ) .","Although it also relates proteins located in different modules ( Middle and Tail ) , the two proteins are less than 30 Å apart in our model .","The holoenzyme cross-link dataset includes one cross-link of Med5 ( Tail ) to Med1 ( Middle ) and two cross-links of Med5 ( Tail ) to Med15 ( Tail ) .","Because the Med1-Med5 cross-links are inter-modular , they were assigned lower weights by our scoring function and were not satisfied in the final model .","It is possible that the interactions occur in alternative conformational states , where Med1 is in close proximity to Med5 .","A third subunit interaction that is unsupported by our model , between Med1-Med7 , comes from immunoprecipitation data ( Kang et al . , 2001 ) , and cannot be ruled out , because the N-terminus of Med1 is in the vicinity of the 16 C-terminal residues of Med7 , which are unmodeled in the crystal structure and not well localized in our study ( Figure 2—figure supplement 5 ) .","A lack of lysine residues in the 16 C-terminal residues of Med7 might explain why this putative interaction was not detected by cross-linking .","Our Mediator model was also validated by comparison with results of previous EM studies that mapped subunit terminal tags or subunit deletions onto 2-D class averages ( Tsai et al . , 2014; Wang et al . , 2014 ) .","Our detailed 3-D model was consistent with all 2-D mapping results ( compare Figure 2C , E ) .","It also confirms a proposal that a Med7C-Med21-Med4-Med9 tetramer serves as a structural backbone within the central portion of the Middle module ( Lariviere et al . , 2013 ) ( Figure 4 ) . 10 . 7554/eLife . 08719 . 014Figure 4 . Architecture of the Mediator Middle module .","( A ) Blow-out views of the Middle module subunit localization maps with the first view and coloring identical to Figure 2A and the second view related by a 180° rotation around the y-axis .","( B ) Average contact maps calculated for the cluster of best scoring solutions .","Each square is a contact map calculated between a given pair of Middle module proteins with border length proportional to the length of corresponding subunit sequences .","The grey-shaded areas indicate observed interactions , with the grey-scale proportional to the fraction of models observing the corresponding interaction .","The colored circles are observed cross-links , where the green and orange colors represent respectively satisfied and violated cross-links for the cluster center ( i . e . , the solution that has the minimal root-mean-square deviation ( RMSD ) from all the other solutions in the cluster ) .","( C ) Top view of Middle module with individual panels showing the full localization map of each Middle subunit within its surrounding density , made semi-transparent for visual clarity .","In the central portion of the Middle module the end-to-end stacking of Med7C-21 ( PDB 1YKH ) and Med4-9 ( Homology Model [Lariviere et al . , 2013] ) heterodimers forms a backbone scaffold ( top panel ) , upon which Med10 and Med19 associate at the Med7C-21 extreme ( left panels ) while Med1 associates closely with the N-terminal portions of Med4-9 ( bottom right panel ) .","The Med7N-31 complex ( PDB 3FBI ) is located proximal to the Med7C-21 heterodimer and the unmodeled carboxy-terminal domain ( CTD ) of Med4 ( top panel ) .","The centrally located Med17 NTD ( bottom panel ) is wedged between the Med7C-21-4-9 backbone and Med14 , which forms contacts with Med10 and Med21 at its extreme NTD , while the bulk of its density localizes to Med4-9 and Med1 ( top right panel ) .","Subunit localization density and ribbon models are colored according to Figure 2A . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 014 The proposed structure of the Med7C-Med21-Med4-Med9 tetramer was based on a set of 40 cross-links , an X-ray crystal structure of Med7C-Med21 , and Med4-Med9 modeled as a structural homolog of Med7C-Med21 ( Lariviere et al . , 2013 ) .","Med7C and Med21 form a four-helix bundle , proposed to stack end-to-end on a four-helix bundle of Med4 and Med9 , in a manner analogous to the packing observed in Med7C-21 crystals .","In our modeling , we did not impose end-to-end stacking , but rather represented the Med7C-21 crystal structure and the Med4-Med9 homology model as two distinct rigid bodies .","In our final Mediator model , the ends of the two four-helix bundles are in close proximity ( Figure 4B , C ) , although without enforcing end-to-end helical stacking .","Satisfaction of the EM restraints in other regions of the Middle module appears to require a distortion of the putative end-to-end interaction ( Figure 4C ) , and many solutions displayed a twist of one four-helix bundle with respect to the other , reflecting the asymmetrical nature of the Middle module EM density .","The Med7C-Med21-Med4-Med9 tetramer provides a structural scaffold for the association of all remaining Middle module subunits ( Figure 4B , C ) .","The Middle module can be subdivided into two halves: Med7C-Med21 , with interacting proteins Med10 , Med19 , and Med31; and Med4-Med9 , with interacting protein Med1 .","The only two proteins that interact extensively with both halves of the Middle module are Med14 and Med17 , which perform special architectural roles in the Mediator complex , discussed in more detail below .","A Med7N-Med31 heterodimer , connected to the helical Med7C domain by an unstructured 27 amino acid residue linker , is localized to a distinct protrusion within the central portion of the EM map .","The flexible C-terminal domain of Med4 is localized to the same region .","The Med4-9 half of the Middle module is defined by a close association of the two large subunits Med1 and Med14 with opposite faces of the helical backbone .","The N-terminus of Med1 forms an extensive interaction surface with Med4-9 ( Figures 3 , 4 ) , with eight cross-links between residues in the first 418 residues of Med1 and residues of Med4-9 located on one face of the N-terminal four-helix bundle ( Figures 1D , 4B and Supplementary file 1 ) .","This high density of cross-links leads to a precise localization of the first half of Med1 , with Cluster 1 showing low root-mean-square fluctuations ( RMSFs ) for residues within this portion of the protein .","In contrast , the C-terminus of Med1 lacks cross-links ( Figures 1D , 4B ) and is restrained only by the EM density , resulting in a lower precision of the model and higher RMSF values ( Figure 2—figure supplement 5 ) .","The region of the Tail module that abuts the Middle module is a rich subunit interaction hub , composed of residues from the N-termini of Med2 and Med3 and the C-termini of Med15 and Med16 ( Figures 3 , 5 , 6C ) .","The C-terminus of Med14 also contributes to the structure of this region , consistent with the finding that a C-terminal deletion of Med14 results in the loss of all Tail subunits from Mediator ( Li et al . , 1995 ) .","The Med2 and Med3 NTDs are modeled with relatively high precision ( Figure 2—figure supplement 5 ) , and structural similarity searches with these NTDs show a number of high similarity hits to coiled-coil proteins , in particular to multiple chains within the highly elongated Fibrinogen trimeric coiled-coil structure ( Figure 5—source data 1 ) .","We extended this analysis to calculate the odds that each of these NTDs forms a coiled-coil structure , with the use of two independent prediction servers ( Figure 5D ) .","We found a good correspondence between regions of similarity to coiled-coil proteins and regions predicted with high confidence ( p > 0 . 8 ) to form coiled-coil structures .","These analyses , together with the co-localization of these domains in our Mediator model and the previous observation that Med2 and Med3 can be isolated as a heterodimer ( Beve et al . , 2005 ) , suggest that the N-termini of Med2 and 3 form a coiled-coil motif in the Middle-Tail contact region . 10 . 7554/eLife . 08719 . 015Figure 5 . Architecture of the Mediator Tail module .","( A ) Subunit localization within the Tail module .","For each Tail subunit , the corresponding localization density is shown within a semi-transparent Tail density ( grey ) and colored according to Figure 2A .","( B ) Average contact maps calculated for the cluster of best-scoring solutions and represented as described in Figure 4B .","( C ) β-propeller model of the Med16 NTD .","The location of 7 WD domains within the N-terminus of Med16 , predicted by MSA analysis ( Bourbon , 2008 ) , is shown in the context of the predicted Med16 N-term secondary structure ( bottom panel ) .","Similarity searches identified a high-confidence match to the 7-bladed β-propeller of the S . cerevisiae vesicle coat protein Sec31 ( PDB 2PM9 ) .","( D ) Schematic showing the predicted coiled-coil regions of three interacting Tail proteins , Med2 , Med3 and Med15 .","The co-localization of the Med2/3 N-termini and Med15 C-term ( panel A ) supports the formation of a coiled-coil at this tail locus . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01510 . 7554/eLife . 08719 . 016Figure 5—source data 1 . HHpred comparative modeling results for Tail module proteins . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01610 . 7554/eLife . 08719 . 017Figure 6 . Novel structural insights into Mediator complex architecture and module connectivity .","( A ) Subunit localization density for Med17 ( blue ) reveals the extension of the unmodeled N-terminal domain to contact Middle module subunits .","( Left A panel )","The N-terminus of the Head subunit Med17 ( blue density and beads ) forms an extensive cross-linking network ( green links ) within the Middle module , acting as an inter-modular bridge ( see right A panel and Supplementary file 1 for detailed cross-link information ) .","Head and Middle module ribbon models are colored light grey and according to their subunit color in Figure 2A , respectively , and unmodeled residues cross-linked to Med17 are represented with light grey beads .","Only Med17 beads with cross-links are shown for clarity and coordinates are derived from the full-complex centroid model .","( Right A panel )","A detailed stick representation of the cross-linking network shown in the left panel .","( B ) Subunit localization density for Med14 ( salmon ) highlights its unique function as a Mediator scaffold protein spanning ∼220 Å to connect all three modules .","( Left B panel )","The N-terminus of Med14 forms an extensive cross-linking network across the full Middle module .","The structural elements of Med14 and other subunits are represented as in left panel of ( A ) .","( Right B panel )","A detailed stick representation of the cross-linking network shown in left panel .","( C ) The localization of the Med14 CTD to the Middle-Tail junction is observed through a pair of cross-links to a Tail protein located in the region ( Left C panel ) .","( Right C panel )","A detailed stick representation of the subunit localization and cross-linking network within the Middle-Tail junction .","( D ) Subunit localization density for the subunits localized in the central portion of the Tail module .","( Left D panel )","Bead representation of the cross-links involving the N- and C-terminal portions of Med16 ( yellow ) with Med5 ( brown ) and Med15 ( green ) , respectively .","The Med16 N-terminal β-propeller is centrally located in the Tail module and forms extensive interactions with Med5 .","The Med16 CTD links Med16-Med5 with the remainder of the Tail .","( Right D panel )","A detailed stick representation of the cross-linking network shown in left panel .","Coordinates for both panels of ( C ) and ( D ) are derived for the Tail module centroid model . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 017 Med15 is co-localized in our Mediator model with Med2 and Med3 and can be isolated with Med2 and Med3 in a trimeric complex ( Zhang et al . , 2004 ) .","Central residues of Med15 exhibit homology to Fibrinogen ( Figure 5—source data 1 ) , and may participate in a coiled-coil structure , but it is the C-terminal domain of Med15 ( 850–1081 ) that interacts with Med2 and Med3 in the Middle-Tail contact region .","The Med15 CTD also interacts with the Med16 CTD , apparently stabilizing the association of Med16 within the Tail , and that of Med5 as well , because truncation of the Med16 CTD results in the dissociation of both Med16 and Med5 from the Mediator ( Tsai et al . , 2014 ) .","Previous sequence analysis of Mediator subunits identified seven WD repeats in the N-terminal domain of Med16 ( Figure 5C ) ( Bourbon , 2008 ) .","We extended this analysis and identified a number of high-confidence hits to seven-bladed β-propeller structures .","The most complete match to a full β-propeller was to Sec31 , a component of the S . cerevisiae coat protein complex ( COPII ) involved in ER vesicle transport ( Figure 5C and Figure 5—source data 1 ) .","We used a homology model , in which the first 540 residues of Med16 were mapped onto the Sec31 β-propeller structure , in our modeling of the Tail module architecture ( Figure 2—figure supplement 2 ) .","Two intra-protein cross-links of Med16 could be mapped onto the homology model without violation ( Supplementary file 1 ) .","In our model , the Med16 β-propeller was located in a central position at the base of the Tail module , making extensive contacts with Med5 ( Figure 6D ) .","The central location of Med16 within the Tail module and a large interaction surface with Med5 are consistent with the finding that the entire Tail module is destabilized and lost in Mediator preparations from a strain from which the MED16 gene was deleted ( Li et al . , 1995 ) .","In contrast , Med5 occupies a terminal position within the Tail module , and the MED5 gene can be deleted without disrupting the remainder of the module ( Tsai et al . , 2014 ) .","Although the majority of Med5 occupies a single location within the Tail module , the N-domain of Med5 follows a path around the outside edge of the Med16 β-propeller domain before interacting with Med15 at a more internal locus ( Figure 6D ) .","This interaction is consistent with an earlier finding that Med5 can form a tetrameric complex with the Med2-Med3-Med15 trimer in the absence of Med16 ( Beve et al . , 2005 ) .","Especially noteworthy in our Mediator model are the special roles performed by Med17 and Med14 .","An integral component of the Mediator Head module , Med17 is of particular interest for a temperature sensitive allele ( srb4-138ts ) in which all pol II-dependent transcription is abolished at the restrictive temperature ( Thompson and Young , 1995; Holstege et al . , 1998 ) .","N-terminal truncations of Med17 in S . cerevisiae are inviable ( data not shown ) .","The N-terminal 181 residues of Med17 are disordered in Mediator Head module crystals .","Our Mediator model reveals a central role of the Med17 N-terminal domain in the connection between the Head and Middle modules .","The Med17 NTD is strongly localized at a central site in the Middle module , wedged between Med7-Med21-Med4-Med9 backbone density on one side and Med14 density on the other ( Figures 2 , 4 ) .","The entire path of Med17 from this Middle interaction site to its first modeled residue in the Head module is clearly resolved in the model ( Figure 6A ) .","The essential Med14 subunit was originally identified as a repressor protein in yeast ( Sakai et al . , 1990 ) .","Reports of the location of Med14 in the Mediator have been contradictory .","Med14 truncations uncouple the Tail module , suggesting that Med14 is primarily a Middle module protein ( Li et al . , 1995 ) , but in yeast strains in which the Middle module has been disrupted by subunit deletion , Med14 remains associated with the Tail ( Baidoobonso et al . , 2007 ) .","Furthermore , Med14 is required along with reconstituted Head and Middle modules for basal transcription in vitro in Mediator-depleted cell extracts ( Cevher et al . , 2014 ) .","Med14 formed cross-links with components of the reconstituted Head and Middle modules , as well as with Med17 ( Cevher et al . , 2014 ) .","In our Mediator model , Med14 makes extensive contacts with proteins from all three modules , and is the only Mediator subunit that does so ( Figure 1D ) .","It spans almost the entire Mediator complex , extending a distance of about 220 Å from N-terminal interactions with Med10-Med19 at the tip of the Middle module to C-terminal interactions at the Middle-Tail contact region ( Figure 6B , C ) .","Med14 residues 264–600 interact with the Med4-Med9 four-helix bundle ( Figures 3 , 4 , 6B ) .","Med14 residues 247–350 participate in bridging the Head and Middle modules through multiple contacts with the Med17 NTD .","Additional interactions with the Head module proteins are suggested by contacts of Med14 residues 601–1082 with Med20 and the C-terminal region of Med17 ( residues 401–687 ) ( Figure 3 ) .","Plaschka et al . ( 2015 ) recently reported a 9 . 7 Å structure of a yeast ‘core initiation complex’ , comprising the Mediator Head module , a minimal Middle module , pol II , a nucleic acid scaffold , and the general transcription factors TBP , TFIIB and TFIIF , obtained by cryo-EM and chemical cross-linking ( Plaschka et al , 2015 ) .","The Mediator portion of the cryo-EM map represents only about 50% of the mass of Mediator , as the ‘core initiation complex’ does not include the 480 kDa Tail module or the Middle module protein Med1 .","The structure shows the ‘neck’ of the Head module binding to pol II near subunits Rpb 4 and Rpb7 , with the ‘mobile jaw’ regions of Med 18 and Med20 extending to contact the TFIIB-binding site and subunits Rpb 3 , 10 , 11 , and 12 .","Cross-links between subunits within the Head ( 30 unique cross-links ) and Middle ( 71 unique cross-links ) modules are in complete agreement with our three-dimensional architectural model of the free Mediator complex ( Figure 7—figure supplement 1B ) .","Docking our Mediator model to the EM map of Plaschka et al . ( 2015 ) resulted in a holoenzyme model that was also largely consistent with Mediator to polymerase cross-links ( 12 non-redundant and within the modeled sequence ) identified by Plaschka et al . ( 2015 ) ( Figure 7A , C ) .","However , a set of 17 Mediator to polymerase cross-links from our study of holoenzyme showed agreement only in the region where the Head module contacts Rpb4 and Rpb7 ( Figure 7B ) .","Cross-links between the Tail and Middle modules to pol II in our study were not consistent with the map of Plaschka et al . ( 2015 ) ( Figure 7C ) .","As mentioned above , cross-links between Mediator modules in our study were also poorly satisfied by the EM density for free Mediator used in our integrative modeling .","Together , these findings suggest a different conformation of the Mediator-polymerase holoenzyme in the presence of the Tail module ( not included in the map of Plaschka et al . ( 2015 ) ) and in the absence of the nucleic acid scaffold and general factors .","Our results point to a similar contact site between the Head module and pol II to that in the model of the ‘core initiation complex’ , but suggest that a conformational change in the Mediator brings the Tail module in contact with pol II near the DNA binding cleft of pol II in the holoenzyme , a suggestion that is consistent with our own preliminary cryo-EM data on the holoenzyme ( P Robinson , unpublished ) ( Figure 7D ) .","Such conformational dynamics are in keeping with the idea that interaction between Mediator and pol II in the absence of other factors is largely determined by interactions between the Rpb1 CTD and the neck of the Head module ( Robinson et al . , 2012 ) . 10 . 7554/eLife . 08719 . 018Figure 7 . Holoenzyme cross-linking data indicate different conformational states between the holoenzyme and the core Mediator initiation complex .","( A ) Docking of our Mediator model to the EM density of Plaschka et al . ( 2015 ) provides a subunit architectural map that is highly consistent with the 12 Mediator to pol II cross-links of Plaschka et al . ( 2015 ) ( non-redundant and within modeled sequence ) , collected on a core initiation complex containing the Mediator Head and Middle modules with pol II , nucleic acid scaffold and the general factors TBP , TFIIB and TFIIF .","( B ) In contrast , the 17 equivalent Mediator to pol II cross-links of the current study are largely inconsistent with the position of Mediator modules found in the core initiation complex .","( C ) Distributions of cross-link Cα distances from the two datasets measured in the context of the docked Mediator model .","( D ) The pattern of holoenzyme cross-links implies a major conformational rearrangement in the presence of the Tail module and absence of DNA scaffold and general transcription factors . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01810 . 7554/eLife . 08719 . 019Figure 7—figure supplement 1 . Further validation of the Mediator model . Agreement between the structural model of the Mediator complex and the data for the core initiation complex collected by Plaschka et al . ( 2015 ) .","( A ) Comparison between our model ( left ) , the EM map of the apo-mediator complex ( center ) ( Tsai et al . , 2014 ) , and the Mediator Head-Middle EM density from the core initiation complex ( right ) ( Plaschka et al . , 2015 ) .","( B ) Agreement between our model and the CX-MS data from the core initiation complex .","The histogram is obtained from cross-linked residue–residue distances calculated on our model .","In the inset , the box plots are calculated on the inter-module , intra-head module , and intra-middle module cross-links .","The solid box represents the interquartile range of the distance distribution .","The median is represented by the red line , and the cross-link distances are represented with green crosses .","To make the comparison more stringent , we removed distances of intra-molecular cross-links calculated on residues pairs that were less than 50 residues apart in the sequence . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 019"],["Through integration of data from diverse sources , we have arrived at a 3-D model of Mediator , which explains virtually all previous findings , and which provides a complete picture , including internal details , of the complex .","One source of data was cross-linking , performed on a native holoenzyme , comprising pol II and Mediator , for several reasons: Mediator structure was thereby analyzed in a more functionally relevant state; the native holoenzyme , isolated from yeast , was much more soluble and stable than a reconstituted holoenzyme formed by mixing pol II and Mediator in vitro; and the X-ray crystal structure of the pol II component of the holoenzyme provided an internal control for the modeling process .","We restricted our analysis to the internal organization of the three Mediator modules , because their relationship to one another is believed to differ between free Mediator and the holoenzyme , and free Mediator was the source of our EM information .","Notable features of our 3-D Mediator model include the following:The definition of the three Mediator modules .","Whereas Med17 and Med14 were previously classified as Head and Tail proteins , the N-terminal domains of both proteins are now seen to form part of the Middle module .","The two NTDs not only interact with core components of the Middle module but also with each other . The special roles of Med17 and Med14 in the architecture of the entire Mediator .","Med17 connects Head and Middle modules , while Med14 extends the entire length of Mediator , connecting all three modules .","The path of the NTD of Med17 , not revealed in X-ray crystal structures of the Head module , can now be followed from a site of interaction in the Middle module to its connection with the body of the protein in the Head module . The complete 3-D arrangement of subunits of the Middle module .","A Med7-Med21-Med4-Med9 tetramer is seen to form the core of the module , with a cap of Med10-Med19 at one end and a cap of Med1 at the other . The complete 3-D arrangement of components of the Tail module .","Of particular note are the extensive interactions of the NTDs of Med2 and Med3 and the CTDs of Med 15 and Med16 , which form the region of the Tail module that contacts the Middle module , and which is important for Mediator-activator protein interaction .","The Med2-Med3-Med15 triad is a binding target of multiple yeast transcriptional activator proteins , including Gcn4 and Gal4 ( Lee et al . , 1999; Myers et al . , 1999; Zhang et al . , 2004 ) .","Yeast harboring a Med3Δ Med15Δ double knockout exhibits decreased levels of Mediator and of transcription pre-initiation complexes at inducible gene promoters ( Ansari et al . , 2012 ) .","Secondary structural features of Tail module proteins .","The NTDs of Med2 and Med3 interact in a coiled-coil motif in the Middle-Tail junction .","A 7-bladed β-propeller encompassing the N-terminal half of Med16 interacts extensively with Med5 at the base of the Tail .","Both proteins play roles in transcriptional repression , and expression analysis in yeast Tail mutants has established a functional link between them ( van de Peppel et al . , 2005 ) .","The 3-D Mediator model not only unifies and resolves discrepancies in the extensive literature on Mediator organization , but also forms a starting point for future investigation .","High-resolution structures can be docked and additional structural information incorporated in the model , using the same integrative approach we applied here ( Russel et al . , 2012 ) .","Interactions with pol II , general transcription factors , and transcriptional regulatory proteins can be mapped to the surface .","Changes associated with activity or due to mutation can be identified and their functional significance understood .","Our Mediator model could be docked into the Mediator portion of the EM map of Plaschka et al . ( 2015 ) , as association with pol II appears to cause little change in the structure and relative position of the Head and Middle Mediator modules compared to those in the free Mediator EM map used as restraint in our modeling study ( Tsai et al . , 2014 ) ( Figure 7—figure supplement 1A ) .","Our docked model provides a more detailed map of the locations of Middle module subunits across the pol II surface , and therefore allows a more complete interpretation of the Mediator portion of the map of Plaschka et al . ( 2015 ) .","Our model provides a structural basis for the Mediator to pol II cross-links reported by Plaschka et al . ( 2015 ) , which can be mapped with precision onto both the Mediator and pol II surfaces to yield a structurally consistent set of cross-link distances ( Figure 7A , C ) .","Furthermore , mapping novel Middle module cross-links from Plaschka et al . ( 2015 ) onto our Mediator model , we find additional strong support for the validity of our subunit localizations ( Figure 7—figure supplement 1B ) .","Also as mentioned above , the cross-links between Mediator modules that we obtained with holoenzyme are inconsistent with the result of docking our Mediator model into the map of Plaschka et al . ( 2015 ) ( Figure 7B , C ) .","We found multiple cross-links between the Tail module and the region of pol II around the active center cleft , whereas docking our Mediator model into the map of Plaschka et al . ( 2015 ) places the Tail module on the opposite side of pol II .","We have noted the Mediator conformational change required to satisfy our Tail module-pol II cross-links ( Figure 7D ) .","Tail module interactions with pol II in the vicinity of the active center cleft may be relevant to the role of Tail module subunits Med14 , 15 and 16 in transcriptional repression at multiple loci in yeast ( Covitz et al . , 1994; Sussel et al . , 1995 ) .","The success of the integrative modeling approach in the derivation of Middle and Tail module structures warrants application to the issues raised here and to other challenging problems .","It is of immediate interest to apply IMP to the holoenzyme with an expanded cross-link dataset and a high resolution holoenzyme EM map .","Further studies may be directed towards a complete RNA polymerase II pre-initiation complex including Mediator , and to additional biological particles of great size and complexity ."],["The S . cerevisiae Holoenzyme was purified as an intact complex using a yeast strain in which both endogenous Mediator and RNA Pol II were affinity tagged .","To achieve double tagging the following genetic manipulations were performed in the CB010 background ( Matα pep4::HIS3 prb1::LEU2 prc1::HISG can1 ade2 trp1 ura3 his3 leu2–3 leu2–112 ) .","First , the N-terminus of the essential Mediator Head module subunit , Med17p , was TAP-tagged using the N-terminus tagging cassette , including Kluyveromyces lactis TRP1 marker gene , amplified from vector pBS1761 as described previously ( Puig et al . , 2001 ) .","The MED17 gene was returned to endogenous promoter regulation following Cre-mediated LoxP recombination driven from pSH47 ( Ura+ ) .","The C-terminus of the largest RNA Pol II subunit , Rpb1p , was tagged with a custom protein G affinity tag with 3C PreScission protease cleavage site .","The vector for C-terminal protein G tag cassette amplification [pCEMM-CTAP ( KanMX ) ] was first generated by sub-cloning a 1633 bp fragment containing the Tn903 KanMX marker gene with flanking LoxP sites from pUG6 between the XhoI ( 597 ) and NotI ( 616 ) sites of mammalian expression vector pCEMM-CTAP ( SG ) , positioned immediately C-terminal to the protein G domain sequence ( Forward primer 5′-3′: actgtcactgaatagctcgaggaacgcggccgccagctgaagcttcgtacg–Reverse primer 5′-3′: ggcggaatttacgtagcggccgcccgcggccgcataggccactagtggatctg ) .","The Rpb1 C-terminal protein G tagging cassette including the protein G and KanMX elements was PCR amplified from pCEMM-CTAP ( KanMX ) using primers with flanking sequence including 45 bp complementary to Rpb1 genomic DNA immediately upstream/downstream of the STOP codon and sequence encoding the 3C PreScission cleavage site ( 5′-3′: ctggaagttctgttccaaggtcca ) .","The endogenous yeast complex was purified from whole-cell extracts using an affinity capture method described previously with minor modifications ( Robinson et al . , 2012 ) .","Double-tagged CB010 were grown in 200L YPAD to 9 . 0 A600 and 3 . 0 Kg cells were lysed by continuous-flow bead beating in A25 buffer ( 25 mM ammonium sulfate , 100 mM HEPES pH 7 . 8 , 1 mM EDTA , 5% glycerol , 5 mM DTT , and 1× protease inhibitor cocktail [0 . 6 µM leupeptin hemisulfate , 2 µM pepstatin A , 1 mM PMSF and 2 . 1 mM benzamidine hydrochloride] ) with 25 µg/ml RNAse A ( Qiagen , Hilden , Germany ) .","The cell debris was pelleted by centrifugation at 12 , 250×g for 1 hr at 4°C and the supernatant made up to 300 mM ammonium sulfate with the addition of cold 3 . 9 M ammonium sulfate and gentle stirring .","Nucleic acids were depleted from the lysate by adding 500 ml DEAE ( GE healthcare , Little Chalfont , UK ) pre-equilibrated in A300 buffer .","After 30 min of stirring the DEAE was pelleted by centrifugation ( 12 , 250×g , 30 min at 4°C ) and the supernatant loaded onto an IgG column before washing in A500 buffer .","The use of two distinct protease cleavage sites allowed the isolation of stable holoenzyme complex using a two-stage elution approach ( Figure 1—figure supplement 1 ) .","First-stage TEV cleavage led to the release of free Mediator , with Mediator retention occurring through complex formation with uncleaved RNA pol II .","Subsequent cleavage with 3C PreScission protease led to the release of both free RNA pol II and holoenzyme complex .","The order of protease cleavage could be reversed with equivalent yields of holoenzyme complex ( ∼5 mg/3 Kg dry cell mass ) .","The holoenzyme was further purified from free RNA pol II ( TEV-3C ) or Mediator ( 3C-TEV ) using ion exchange ( 75–600 mM ( NH4 ) 2SO4; HiTrap FFQ ) and size-exclusion ( 200 mM ( NH4 ) 2SO4 , 25 mM HEPES , pH 7 . 8 , 5% glycerol , 2 mM DTT; Superose 6 ) chromatography steps .","In the case of the TEV-3C cleavage order , the holoenzyme eluted from the IE column as a leading shoulder in the large peak due to free RNA pol II .","Using the 3C-TEV cleavage order , the holoenzyme was retained longer in IE than free Mediator and could be resolved as an independent peak .","Prior to cross-linking , holoenzyme samples were exchanged into phosphate buffer containing: 150 mM NaPO4 , 150 mM KOAc , pH 7 . 5 , 5% glycerol and 2 mM DTT by dialyzing against three changes of buffer .","Cross-linking was then performed by one of two procedures: MS Peaklists were generated such that all monoisotopic ions detected within the selection window of the linear ion trap ( 4 m/z units ) were annotated as possible precursors to a given product ion spectrum , as described previously ( Robinson et al . , 2012; Trnka et al . , 2014 ) .","Peaklists were first searched against the full SwissProt database to check the composition of the sample .","Cross-link searches were then carried out against a target database consisting of the 52 components of the S . cerevisiae Preinitiation Complex .","For each polypeptide sequence , 10 randomized versions were generated and concatenated to the target database for a total of 572 sequences ( 52 target and 520 decoy ) .","Decoy protein sequences were the same length as their corresponding target sequences and sampled from the natural distribution of S . cerevisiae amino acid frequency .","Peaklists were then searched against this database using Protein Prospector ( version 5 . 13 . 1 ) to assign putative cross-linked peptides ( Trnka et al . , 2014 ) .","Carbamidomethylation of cysteine was considered as a fixed modification .","N-terminal methionine loss with and without acetylation , peptide N-terminal glutamine conversion to pyroglutamate , oxidation of methionine , and ‘dead-end’ modification of lysine and the protein N-terminus by semi-hydrolyzed BS3 were considered as variable modifications in addition to cross-linking by BS3 .","Up to 3 variable modifications per peptide were considered .","Mass tolerances of 13 ppm and 20 ppm were used for precursor and product ions respectively .","Trypsin specificity with 3 missed cleavages was used to generate theoretical peptides .","85 product ion signals from each MS2 spectrum were used for assessment .","For experiments involving isotopic mixtures of cross-linking reagents , the searches were performed twice , once specifying heavy BS3 and once with light BS3 .","The search results were later merged .","Both light and heavy dead-end modifications were considered in each of these database searches .","For all experiments , cross-link spectral matches ( CSM ) were discarded if the following Protein Prospector parameters fell outside the threshold values: peptide score below 20 , peptide/protein/worse_peptide expectation values above 50 , and score_difference below 0 .","A linear support vector machine ( SVM ) model was constructed to classify CSMs between decoy and target classes ( Trnka et al . , 2014 ) .","Briefly , the CSMs were split evenly into two sets .","SVM models were trained on half of the data and evaluated for their ability to correctly classify true negatives ( hits to the decoy database ) on the test set .","CSMs were considered decoy matches if either of the two constituent peptides came from the decoy database .","The final SVM model included a metric reflecting the number of product ions matched to the less fragmented of the two peptides ( score_difference ) and a metric for the overall match of the spectrum ( percentage of the total ion current annotated ) .","The inclusion of additional metrics did not increase the predictive power of the model .","An SVM decision value of 0 was taken as the acceptance threshold for reporting cross-linked peptides .","FDRs were calculated by dividing the number of decoy hits at this threshold by 10 ( to account for the larger decoy database size ) and dividing by the number of hits to the target database .","CSMs where either peptide was less than 4-amino acids long were discarded , with a few manually validated exceptions .","When spectra could be interpreted by multiple cross-linked peptide-pairs , CSMs were considered to be unambiguous if the second best match was further than 0 . 3 SVM decision value units from the top match .","Intra-protein cross-links were selected over inter-protein cross-links if there were still conflicts .","The remaining decoy matches were also removed from consideration .","Following these selection criteria , most CSMs identified a single cross-linked peptide-pair .","The remaining spectral redundancy came exclusively from ambiguous site localizations , where there was sufficient evidence to identify both peptides , but the position of the modified amino acid remained ambiguous .","In these cases , all possible sites were reported using a spectral identifier and used in subsequent modelling steps .","Cross-linking results and annotated HCD spectra may be viewed online using Protein Prospector's MS-Viewer program: http://prospector2 . ucsf . edu/prospector/cgi-bin/msform . cgi ?","=msviewer .","Search key: qzpxihtngx All raw mass spectrometry files and peaklists used in the study have been deposited to the MassIVE proteomics repository and are available through the ProteomeXchange consortium .","MassIVE accession: MSV000079237 ProteomeXchange accession: PXD002723 Our integrative structure modeling of RNA Pol II and Mediator complexes proceeds through four stages ( Lasker et al . , 2010; Fernandez-Martinez et al . , 2012; Lasker et al . , 2012 ) ( Figure 2 ) : ( 1 ) gathering of data , ( 2 ) representation of subunits and translation of the data into spatial restraints , ( 3 ) configurational sampling to produce an ensemble of models that satisfies the restraints , and ( 4 ) analysis and assessment of the ensemble .","The modelling protocol ( i . e . , stages 2 , 3 , and","4 ) was scripted using the Python Modelling Interface ( https://github . com/salilab/pmi ) , version be72c15 , a library to model macromolecular complexes based on our open source IMP package ( http://salilab . org/imp/ ) , version 829c3f0 ( Russel et al . , 2012 ) .","Files containing the input data , scripts , and output models are available at http://salilab . org/mediator .","In the IMP modeling procedure we perform a large number of independent modeling runs , each starting from a random configuration and undergoing thousands of rounds of small random model perturbations , in order to achieve exhaustive sampling of conformational space .","As we are interested only in models that best satisfy the restraints , we select a small fraction of the models representing the very best scoring solutions ( Figure 2—figure supplement 3A ) .","These models then undergo pairwise-RMSD clustering to separate out divergent structural states ( Figure 2—figure supplement 3B ) .","In our study , most of the best scoring models ( 92% , n = 500 ) gave a unique subunit arrangement for 18 of the 21 Mediator subunits .","The divergent clusters differed only in the relative locations of Med5 , Med15 and Med16 within the Tail module ( Figure 2—figure supplement 3C ) .","We compared these three clusters in terms of crosslink violation statistics ( Figure 2—figure supplement 3E ) and consistency with prior EM localization experiments ( Figure 2—figure supplement 3H ) .","A single cluster showed both consistency with earlier localizations and significantly better violation statistics and was therefore deemed to be the Mediator architecture with the highest confidence .","Modeling studies of RNA pol II employed cross-link data from two sources: 156 cross-links from previous experiments conducted with the free pol II fraction isolated during our holoenzyme SEC purification ( Trnka et al . , 2014 ) , and 108 cross-links from a previous study ( Chen et al . , 2010 ) .","These sources had 63 cross-links in common .","Hence , the current study used 201 unique cross-links .","The atomic structures for the twelve-subunit yeast pol II have been previously determined by X-ray crystallography ( PDB accession ID: 1WCM , [Armache et al . , 2005] ) .","The crystallographic structure covers 89% of the residues in the pol II complex .","A 20 . 9 Å resolution density map from single particle EM reconstruction of the pol II complex was also considered ( EMDB accession id: 1883 , [Czeko et al . , 2011] ) .","The 12 pol II subunits were represented as 15 domains , where Rpb1 was decomposed into 3 domains ( residues: 1–1140 , 1141–1274 , 1275–1733 ) , Rpb2 was decomposed into 2 domains ( residues 1–1102 , 1103–1224 ) , and the remaining subunits were represented as single domains .","Regions without a crystal structure were modeled as beads containing up to 5 amino acid residues .","We followed the same 4-stage procedure as described for the Mediator complex above ."]],"string":"[\n [\n \"Mediator is a complex of at least 21 subunits , with a mass of greater than 1 million Daltons , conserved from yeast to humans ( reviewed in Kornberg , 2005; Conaway and Conaway , 2011; Ansari and Morse , 2013 ) .\",\n \"Genetic studies have shown that Mediator is essential for RNA polymerase II ( pol II ) transcription and required for positive and negative regulation of transcription .\",\n \"Biochemical studies have confirmed the importance of Mediator for both basal and regulated transcription , and have demonstrated interactions of Mediator with pol II , with general transcription factors , and with transcriptional activator proteins .\",\n \"Through these interactions , Mediator is thought to promote assembly of the entire machinery for pol II transcription .\",\n \"Electron microscopy ( EM ) has shown a division of Mediator in three modules , termed Head , Middle , and Tail ( Asturias et al . , 1999 ) .\",\n \"EM studies further showed that Mediator structure is conserved from yeast to humans , that Mediator is compact when free in solution , and that a major structural rearrangement occurs upon interaction with pol II , which is surrounded by the Head and Middle modules in the so-called ‘holoenzyme’ ( Asturias et al . , 1999; Dotson et al . , 2000 ) .\",\n \"Details of subunit interaction networks within Mediator modules have come from co-expression studies ( Koh et al . , 1998; Lee et al . , 1998; Kang et al . , 2001; Koschubs et al . , 2010 ) and from Mediator-specific ( Guglielmi et al . , 2004 ) and genome-wide ( Uetz et al . , 2000; Ito et al . , 2001 ) two-hybrid approaches .\",\n \"The currently accepted subunit assignments are Med 6-8-11-17 ( Srb4 ) -18 ( Srb5 ) -20 ( Srb2 ) -22 ( Srb6 ) in the Head module , Med 1-4-7-9 ( Cse2 ) -10 ( Nut2 ) -19 ( Rox3 ) -21 ( Sbr7 ) -31 ( Soh1 ) in the Middle module , and Med 2-3 ( Pgd1 ) -5 ( Nut1 ) -14 ( Rgr1 ) -15 ( Gal11 ) -16 ( Sin4 ) in the Tail module ( old naming convention in parentheses ) .\",\n \"X-ray crystal structures of Head modules from Saccharomyces cerevisiae and Schizosaccharomyces pombe have been determined , with ( Robinson et al . , 2012 ) and without ( Imasaki et al . , 2011; Lariviere et al . , 2012 ) the carboxy-terminal domain ( CTD ) of the largest pol II subunit , Rpb1 , bound to a conserved site on the surface .\",\n \"CTD interaction with Head and Middle modules plays an important role in the association of Mediator with pol II and the pre-initiation complex .\",\n \"Beyond the X-ray crystallography of the Head module , structural information on Mediator is limited .\",\n \"Only small portions of the Middle ( Baumli et al . , 2005; Koschubs et al . , 2009 ) and Tail modules ( Bontems et al . , 2011; Eletsky et al . , 2011; Vojnic et al . , 2011 ) have been solved at near atomic resolution .\",\n \"EM at low resolution with subunit-labeling has given an approximate idea of the spatial organization of the modules ( Tsai et al . , 2014; Wang et al . , 2014 ) , and an architectural model for a subcomplex of the Middle module ( lacking Med 1 and 19 ) has been proposed on the basis of cross-linking data and homology modeling ( Lariviere et al . , 2013 ) .\",\n \"Recently , reconstituted yeast Head and Middle Mediator modules were visualized by EM in a complex with pol II and a subset of general transcription factors ( Plaschka et al . , 2015 ) .\",\n \"Mediator is representative of a large number of macromolecular complexes that are refractory to classical high-resolution structural approaches , due to low native abundance and to conformational and compositional heterogeneity .\",\n \"For the study of such systems , data from chemical cross-linking and mass spectrometry can provide multi-protein interaction maps at residue-level resolution ( Murakami et al . , 2013; Erzberger et al . , 2014; Shi et al . , 2014 ) .\",\n \"The approach is particularly effective when combined with other structural information , such as electron density maps from EM ( Murakami et al . , 2013 ) and subunit structures from X-ray crystallography ( Robinson et al . , 2012 ) .\",\n \"To construct structural models of macromolecular assemblies by combining such diverse types of information , the open-source Integrative Modeling Platform ( IMP ) program was developed ( http://integrativemodeling . org ) ( Russel et al . , 2012 ) .\",\n \"IMP translates the data into spatial restraints , computes an ensemble of models by satisfying these restraints as well as possible , and finally assesses the ensemble against the data used or not used in its construction ( Schneidman-Duhovny et al . , 2014 ) .\",\n \"IMP has been successfully applied to a number of protein complexes ( Lasker et al . , 2012; Erzberger et al . , 2014; Shi et al . , 2014 ) .\",\n \"Here we determined the molecular architecture of Mediator by integrative structure determination , based on chemical cross-links , X-ray crystal structures , homology models , and a cryo-EM electron density map , with the use of IMP .\",\n \"Possible configurations of the Mediator subunits were exhaustively sampled to identify those that best satisfied the experimental restraints .\",\n \"The resulting model was validated by re-computing it with random subsets of chemical cross-links ( i . e . , ‘jackknifing’ ) ( Brunger et al . , 1993 ) as well as by comparison with known protein structures and protein interaction networks ( Guglielmi et al . , 2004 ) .\",\n \"The Mediator model will serve as a basis for studies of Mediator interaction with pol II , general transcription factors , gene activators , and gene repressors .\"\n ],\n [\n \"Mediator was isolated from yeast as originally described ( Kim et al . , 1994 ) in the form of a complex with pol II ( holoenzyme ) .\",\n \"The complex is not only a functionally relevant form of Mediator , but is also more soluble than free Mediator under the conditions used for lysine directed cross-linking .\",\n \"We employed a double affinity-tagging strategy ( Figure 1—figure supplement\",\n \"1 ) and obtained 20-fold more stoichiometric holoenzyme than from previous purification procedures .\",\n \"The native complex was stable , persisting throughout purification , whereas a complex formed from separately isolated Mediator and pol II was disrupted by the same purification procedures .\",\n \"Cross-linking and mass spectrometry were performed on the native holoenzyme , with a mixture of D12-labelled and unlabeled BS3 cross-linking reagents , or with unlabeled BS3 and enrichment of cross-linked trypsin fragments by gel filtration .\",\n \"The BS3 reagent bridges free-amine moieties at surface lysines and N-termini with Cα spacing within approximately 25 Å .\",\n \"Cross-links were assigned with the Protein Prospector platform and a refined scoring function ( Trnka et al . , 2014 ) .\",\n \"The 320 , 767 initial product ion spectra were searched against a database containing the sequences of the yeast transcriptional machinery as well as 520 randomized decoy sequences , identifying 20 , 388 potential cross-linked spectra ( Figure 1A ) .\",\n \"Classification of these matches and the application of quality filters resulted in the identification of 3196 holoenzyme cross-linked spectra at a false discovery rate ( FDR , fraction of decoy hits above score cutoff ) of 0 . 7% ( Figure 1A ) .\",\n \"These spectra defined 402 unique pairs of linked amino acid residues ( ‘cross-links’ ) with a FDR of 4 . 1% ( Figure 1—source data 1 , Figure 1—figure supplement 2 , Supplementary file 1 ) .\",\n \"The majority ( 260 cross-linked fragments ) were between Mediator proteins , 124 were between pol II subunits , and 18 were between Mediator and pol II . 10 . 7554/eLife . 08719 . 003Figure 1 . Holoenzyme cross-linking and modeling results and methodology .\",\n \"( A ) Cross-links were identified by searching MS2 product ion spectra against concatenated target and decoy databases containing 52 Saccharomyces cerevisiae transcription proteins +520 sequence randomized versions , respectively .\",\n \"Confidence in the spectral assignment is represented by an support vector machine ( SVM ) classifier ( Trnka et al . , 2014 ) .\",\n \"The distribution of the target and decoy spectral matches in relation to the acceptance threshold ( red line ) is shown with an overall spectral false discovery rate ( FDR ) of 0 . 7% .\",\n \"( B ) Mapping identified cross-linked spectra onto regions of known structure such as the Mediator Head module and RNA polymerase II yields distance distributions that reflect the accuracy of cross-link identification .\",\n \"The Cα violation distance of 35 Å is indicated with a dashed line .\",\n \"( C ) Demonstration of RNA pol II structural recapitulation from a random starting configuration ( top panel ) of individual pol II subunits represented as rigid bead models and restrained by cross-links ( green links ) and electron microscopy ( EM ) density map ( mesh ) .\",\n \"A representation of the converged modeling solution is presented alongside the 12-subunit pol II X-ray structure ( PDB: 1WCM ) for reference .\",\n \"( D ) Schematic representation of the Mediator complex cross-linking network .\",\n \"Mediator subunits are colored according to their location within the Head , Middle and Tail modules .\",\n \"All Inter-subunit cross-links and selected intra-subunit cross-links from the current study as well from ( Lariviere et al . , 2013 ) are represented , colored according to their origin . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00310 . 7554/eLife . 08719 . 004Figure 1—source data 1 . Categorization of cross-links with respect to module and data source . Numbers of unique cross-linked residue pairs ( ‘cross-links’ ) and cross-linked spectral matches used for integrative modeling of Mediator apo-complex ( ‘Total Mediator’ ) as well as total number of holoenzyme cross-links identified in this study ( ‘Total [this study]’ ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00410 . 7554/eLife . 08719 . 005Figure 1—figure supplement 1 . Purification of Native S . cerevisiae Holoenzyme complex .\",\n \"( A ) Schematic of the affinity capture and chromatographic fractionation of S . cerevisiae holoenzyme complex from whole-cell extract using a double-affinity tag and stepwise cleavage protocol .\",\n \"In the procedure , a mixture of holoenzyme and free Mediator and RNA Pol II are affinity captured , with release of free Mediator through TEV cleavage ( panel B ) and subsequent release of holoenzyme and free Mediator upon 3C protease cleavage ( panel C ) .\",\n \"Following the enrichment of Mediator-bound pol II using ion-exchange chromatography , the holoenzyme was purified to homogeneity using size-exclusion chromatography ( panel D ) .\",\n \"Mass spectrometry analysis confirmed the presence of all 21 Mediator and 12 pol II subunits ( panel E ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00510 . 7554/eLife . 08719 . 006Figure 1—figure supplement 2 . Annotated product ion spectra for selected cross-linked peptides . HCD product ion spectra of BS3 cross-linked Mediator tryptic peptides classified by Protein Prospector .\",\n \"Spectra representing highly confident ( SVM score 5 . 1 ) to less confident ( SVM score 0 . 0 ) classifications are shown .\",\n \"The modified lysine position is denoted with a star .\",\n \"N-terminal acetylation and carbamidomethylation of cysteine residues ( Cam ) are denoted .\",\n \"P , PL , and PLK ions refer to ions formed from dissociation of the cross-linking reagent to lysine bond ( Trnka et al . , 2014 ) .\",\n \"The full list of cross-linked Mediator peptides may be viewed online using search key = qzpxihtngx at: http://prospector2 . ucsf . edu/prospector/cgi-bin/msform . cgi ?\",\n \"form=msviewer . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 006 The pol II component of the holoenzyme provided an internal control , enabling validation of the cross-link dataset by two approaches .\",\n \"First , the X-ray crystal structure of pol II and also that of the Mediator Head module could be used to assess the validity of cross-links , as 130 of the 402 high confidence cross-links were between residues in these structures .\",\n \"Only five of these 130 cross-links occurred between residues further apart than the 35 Å threshold ( Figure 1B ) .\",\n \"This number of violations was therefore consistent with the FDR of 4 . 1% , estimated from the number of decoy database hits .\",\n \"Second , we constructed an integrative model of pol II based on the cross-link dataset , a cryo-EM electron density map ( EMD-1883 , 20 . 9 Å resolution using FSC 0 . 5 ) , and X-ray crystal structures of the pol II subunits , using IMP .\",\n \"The X-ray crystal structures of the subunits were represented as independent rigid bodies , and the regions not observed in the structure were represented by flexible strings of beads .\",\n \"A low resolution EM map was used for comparison with the EM data available for Mediator ( see below ) .\",\n \"Starting from scrambled configurations , the positions and orientations of the pol II subunits were optimized with a Monte Carlo procedure ( which applies a large number of random movements guided by the input data ) , resulting in configurations that satisfied the input data well .\",\n \"This ensemble of good-scoring solutions recapitulated the known architecture of pol II ( Figure 1C ) .\",\n \"The average root-mean-square deviation ( RMSD ) of the solutions with respect to the crystal structure of pol II was 14 Å .\",\n \"All subunits were correctly placed , except for Rpb8 and Rpb9 , whose locations were undetermined , due to a lack of cross-link restraints .\",\n \"We augmented our Mediator cross-link dataset with 38 Mediator Middle module cross-links from a dataset collected on a recombinant preparation of S . cerevisiae Middle module containing 6 subunits ( Lariviere et al . , 2013 ) .\",\n \"The 38 cross-links were selected according to our minimal length requirement of four amino acid residues per cross-linked peptide; three cross-links were common to both datasets .\",\n \"This modest overlap is likely due to the use of a different chemical cross-linker , as well as to the presence of additional subunits in our full native Middle module .\",\n \"The complete dataset of 294 inter-subunit cross-links defined an interaction map of 19 of the 21 Mediator proteins ( Figure 1D ) .\",\n \"The two subunits for which there were no inter-subunit cross-links have been crystallized in complexes with other Mediator subunits: Med6 as part of the Head module ( Robinson et al . , 2012 ) , and Med31 in a complex with the N-terminal region of Med7 ( Koschubs et al . , 2009 ) .\",\n \"Both proteins were present in our native holoenzyme preparations , as shown by SDS PAGE and mass spectrometry ( Figure 1—figure supplement 1 ) .\",\n \"Two subunits , Med17 ( Head module ) and Med14 ( Tail module ) showed patterns of cross-links split between two or more different Mediator modules .\",\n \"The majority of Med17 ( residues 123–687 ) formed an extensive cross-link network with other subunits of the Head module , as expected from the X-ray crystal structure ( Robinson et al . , 2012 ) .\",\n \"The N-terminal region of Med17 ( 1–122 ) , not observed in the Head module crystal structure , formed many cross-links with Med7 and Med21 , subunits of the Middle module ( Figure 1D ) .\",\n \"Similarly , Med14 , a presumptive Tail module subunit , formed cross-links to other Tail module proteins ( Med2 and 15 ) , but almost exclusively in its C-terminal region ( residues 712–1082 ) , while the N-terminal region and the majority of the protein ( residues 1–711 ) cross-linked instead to Middle module subunits ( Med4 , Med9 , Med10 , and Med21 ) .\",\n \"The N-terminal regions of Med17 and Med14 not only extended into the Middle module but also formed many cross-links with each other .\",\n \"For these reasons , we redefine the Mediator modules to include the N-terminal domains of Med17 ( residues 1–122 ) and Med14 ( residues 1–711 ) in the Middle module rather than the Head and Tail modules ( Figure 1D , Med14 and Med17 colored by module localization ) .\",\n \"Although cross-linking was performed on the holoenzyme , about 90% of the cross-links in our dataset were within Mediator modules ( intra-modular ) or within pol II .\",\n \"Due to the paucity of cross-links between Mediator and pol II , and because a cryo-EM map at sufficient resolution for molecular modeling was only available for Mediator without pol II , modeling calculations were performed for Mediator alone .\",\n \"The four-step modeling ( ‘integrative structure determination’ ) procedure ( Figure 2—figure supplement\",\n \"1 ) entailed: ( 1 ) gathering data; ( 2 ) representing and translating the data into spatial restraints; ( 3 ) sampling the conformational space and identifying good scoring solutions; and ( 4 ) analyzing and assessing the ensemble of solutions .\",\n \"All IMP input files , scripts , and output models for this study are available at http://salilab . org/mediator/ .\",\n \"In step 1 , the data comprised our holoenzyme cross-link dataset , X-ray crystal structures of some subunits and subunit regions , homology models for some subunits , and the highest resolution cryo-EM map of free Mediator available ( Tsai et al . , 2014 ) .\",\n \"In step 2 , we constructed a set of 21 Mediator subunit model representations ( Figure 2—figure supplement 2 ) .\",\n \"Atomic models from X-ray crystallography or homology models were represented by rigid bodies ( Figure 2—figure supplement 2 , blue subunit shading ) , and unmodeled subunit regions were represented as flexible strings of beads ( Figure 2—figure supplement 2 , yellow subunit shading ) .\",\n \"The entire Head module was represented as a rigid body ( based on the X-ray structure PDB 4GWP ) , with the exception of the unmodeled N-terminus of Med17 ( 1–181 ) , the last 103 residues of Med6 , and other missing regions , all of which were represented as flexible strings of beads .\",\n \"Manual docking and the mapping of Med 8 , 18 , and 20 onto 2-D EM projections previously supported a single docking solution for the Head module in the EM density ( Tsai et al . , 2014 ) .\",\n \"We performed an exhaustive docking calculation , comparing all possible Head module locations with a cross-correlation scoring function ( Materials and methods ) .\",\n \"We found a docking solution that is consistent with the solution proposed previously , with a normalized cross-correlation coefficient approximately 40% better than the next best solution .\",\n \"We modeled the Mediator complex with the Head module fixed in a single position corresponding to the highest scoring docking solution .\",\n \"Other regions defined by atomic models were Med7C-Med21 and Med7N-Med31 , represented by rigid bodies based on their X-ray structures ( Baumli et al . , 2005; Koschubs et al . , 2009 ) , and Med4-Med9 , based on a homology model ( Lariviere et al . , 2013 ) .\",\n \"In addition , we created a homology model of the first 540 residues of Med16 , also in a rigid body representation , based on a seven-bladed β-propeller fold ( proposed and justified below ) .\",\n \"The EM map was divided into segments to reflect the modular organization of the complex .\",\n \"Since the Head module was docked and fixed in a portion of the density map , the remaining density was segmented into regions corresponding to the Middle and Tail modules on the basis of previous approximate localization of Middle and Tail module subunits within two distinct regions of the EM map ( Tsai et al . , 2014; Wang et al . , 2014 ) .\",\n \"During modeling , each Middle or Tail subunit was restrained to the corresponding Middle or Tail EM density .\",\n \"EM spatial restraints were based on a Gaussian decomposition of the electron density ( Materials and methods ) .\",\n \"The cross-linking data was encoded in a Bayesian scoring function ( Erzberger et al . , 2014; Shi et al . , 2014 ) .\",\n \"The scoring function allowed the independent optimization of the relative weights of arbitrary subsets of the cross-linking data .\",\n \"We split the cross-link dataset into two subsets composed of inter-module ( 8% of the total ) and intra-module ( 92% of the total ) cross-links ( Figure 1—source data 1 ) .\",\n \"In step 3 , we conducted a large number of independent modeling runs .\",\n \"The positions and orientations of rigid bodies and flexible strings of beads were repeatedly perturbed in an effort to satisfy the scoring function consisting of the excluded volume , sequence connectivity , EM , and cross-linking restraints .\",\n \"A total of 165 , 523 Mediator model configurations were produced .\",\n \"The 500 best-scoring models ( solutions ) were grouped on the basis of RMSD into four clusters ( C1-4 , Figure 2—figure supplement 3 ) , the minimum number required to fully represent the main structural differences between the best scoring models .\",\n \"To confirm that we had sampled conformational space sufficiently to reach model convergence , we compared two independent halves of the solutions to each other and to the entire set , showing they were all similar to one another ( Figure 2—figure supplement 4 ) .\",\n \"The solutions satisfied all excluded volume , sequence connectivity , and EM restraints .\",\n \"Intra-module cross-links were satisfied to a high degree ( approximately 95% ) , and inter-modular cross-links to a much lesser degree ( about 10% ) ( Figure 2—figure supplement 3 ) .\",\n \"In fact , the Bayesian scoring function automatically assigned a low scoring weight to the inter-modular cross-links , pointing to an inconsistency of the inter-modular cross-links in the holoenzyme dataset with the EM density map of the free Mediator .\",\n \"The satisfaction of intra-modular cross-links suggests that the arrangement of Mediator subunits within modules of the holoenzyme is similar to that in the free Mediator , whereas the inconsistency of the inter-modular cross-links with the EM map suggests that motions of the modules relative to one another occur upon association with pol II in the holoenzyme .\",\n \"Indeed , EM studies have indicated large motions about hinges between the modules , opening the compact structure of free Mediator for wrapping around pol II in the holoenzyme ( Asturias et al . , 1999; Davis et al . , 2002 ) .\",\n \"Based on these observations , the inter-modular cross-links were not considered in our analysis of the free Mediator .\",\n \"The four clusters of best-scoring solutions differed in two respects .\",\n \"Cluster 3 , which represented a minor subset of best-scoring models ( ∼8% ) , differed from the other clusters by an inversion of orientation of the Middle module .\",\n \"This cluster was disregarded because the inverted orientation was inconsistent with prior EM localization experiments ( Tsai et al . , 2014 ) and because of significantly worse intra-module cross-link violation statistics ( Figure 2—figure supplement 3D–E ) .\",\n \"The remaining three clusters were virtually identical in the locations of all subunits except for Med5 , Med15 , and Med16 in the Tail module ( Figure 2—figure supplement 3C , G ) .\",\n \"Only Cluster 1 showed an arrangement of these three subunits consistent with previous EM localization ( Figure 2—figure supplement 3H ) .\",\n \"Cluster 1 also showed the best cross-link satisfaction statistics and had the best average score .\",\n \"Cluster 1 was therefore taken to represent the true Mediator subunit architecture .\",\n \"The overall precision of our integrative model , computed as the average RMSD of the solutions in Cluster 1 with respect to the cluster center , is 19 Å .\",\n \"The precision of the Middle and Tail modules is 24 and 33 Å , respectively .\",\n \"These values represent the average fluctuations of the individual residues or beads in 3-D space across the ensemble of solutions comprising the highest scoring cluster .\",\n \"Precision defined in this way is not directly comparable to the resolution of an EM map , which refers to the uncertainty of the overall electron density , without regard to the placement of components within the density .\",\n \"Although the precision of the Tail module is lower than that of the Middle or Head modules , it is sufficient to position the Tail subunits within the Tail EM density as features with high cross-link density , and consequently high local precision values ( such as the Middle-Tail junction ) , provide strong spatial restraints .\",\n \"The overall precision of the Mediator model sufficed to determine the positions and orientations of all Mediator subunits , and 20–40 residue beads could be localized with precisions of 10–50 Å ( Figure 2—figure supplement 5 ) , depending on the density of cross-linking restraints .\",\n \"We represented the cluster of solutions by a localization density ( Figure 2 ) , defined as the probability of observing a particular subunit at a specific point in space in the cluster of solutions .\",\n \"Because the model is three-dimensional , it depicts the contact surfaces between subunits , with residue-level resolution at many points where cross-links are formed between them .\",\n \"This contact information is especially rich and reliable in regions where multiple cross-links are formed and where contacts occur in multiple models in the cluster , as depicted by the darkest boxes in the domain interaction map ( Figure 3 ) .\",\n \"Some cross-links anticipated from the domain interaction map were not observed , most likely because cross-linking depends on the occurrence of unprotonated , appropriately oriented , solvent–accessible pairs of lysine residues , and on the formation of tryptic peptides with physicochemical properties amenable to liquid chromatography and mass spectrometry . 10 . 7554/eLife . 08719 . 007Figure 2 . Mediator complex architecture .\",\n \"( A ) Mediator subunit localization density map colored by individual subunit .\",\n \"( B ) Mediator localization density map ( solid grey ) calculated from the highest scoring model cluster and shown at a threshold level ( T = 0 . 1 ) that most closely matches the volume of the EM density map used as 3D spatial restraint during modeling ( EMD-2634 , T = 0 . 35: Blue mesh ) .\",\n \"The position of the three Mediator modules ( Head , Middle , and Tail ) is indicated .\",\n \"( C ) Schematics of the architecture of the Mediator complex obtained from the localization density analysis .\",\n \"Med14 and Med17-NTD are also schematically represented by splines .\",\n \"( D ) Protein–protein interactions derived from published Yeast-two hybrid , immunoprecipitation and sub-complex isolation data ( Uetz et al . , 2000; Ito et al . , 2001; Kang et al . , 2001; Guglielmi et al . , 2004; Zhang et al . , 2004; Beve et al . , 2005; Koschubs et al . , 2010 ) .\",\n \"The data is represented by a graph superposed on the EM density map of the complex ( Tsai et al . , 2014 ) .\",\n \"Nodes are Mediator subunits , the edges are the observed proteomic interactions , and the green circles are isolated sub-complexes .\",\n \"Green and red edges are interactions that are in agreement and in disagreement with the Mediator model , respectively .\",\n \"( E ) Localization by labeling ( blue triangle ) and domain deletion ( areas encircled by light red closed lines ) mapped on the EM density map ( Tsai et al . , 2014; Wang et al . , 2014 ) .\",\n \"Straight light-red lines represent domain deletion assays that split the complex . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00710 . 7554/eLife . 08719 . 008Figure 2—figure supplement 1 . Schematic of integrative structure determination highlighting the individual data inputs and the four stages in our approach . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00810 . 7554/eLife . 08719 . 009Figure 2—figure supplement 2 . Input Model Representations . Individual Mediator subunit model representations are grouped according by membership within the Head , Middle , and Tail modules .\",\n \"For each subunit , a schematic representation of the model substructure is shown below the corresponding bead representation .\",\n \"The subunit substructure is mapped onto the protein sequence and coloured according to atomic model coverage ( blue fill ) or bead structure ( yellow ) where each bead in the subunit model is represented by a single box in the schematic representation .\",\n \"The multi-protein complex structures from which most subunit atomic models are derived are represented on the right . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 00910 . 7554/eLife . 08719 . 010Figure 2—figure supplement 3 . Cluster analysis of the solution ensemble .\",\n \"( A ) Number of satisfied intra-module cross-links plotted against the score for all 165 , 523 models .\",\n \"The vertical brown line represents the score threshold used to select the best scoring 500 models .\",\n \"( B ) RMSD distance matrix for the 500 best-scoring models .\",\n \"The distance matrix is represented by a heat map , where the bin color is proportional to the RMSD distance of the corresponding solution pair .\",\n \"( C ) Difference between the four clusters represented by a hierarchical graph .\",\n \"Clusters 1 , 2 , and 4 have the middle module oriented in the ‘normal’ , expected orientation ( i . e . , Med1 and Med19 are at the bottom and at the top of the module , respectively ) .\",\n \"The three clusters differ in the arrangement of the tail subunits .\",\n \"Cluster 3 has the middle module oriented in a ‘flipped’ orientation ( i . e . , Med1 and Med19 are at the top and at the bottom of the module , respectively ) .\",\n \"( D ) Detailed statistics of cross-link satisfaction for the whole dataset ( top ) , intra-module cross links ( middle ) , and inter-module cross-links ( bottom ) .\",\n \"( E ) Student t-test analysis of the intra-module cross-link violation distance .\",\n \"( F ) Precision of the four clusters .\",\n \"( G ) Statistical properties of the four clusters .\",\n \"( H ) Agreement between subunit localization in Cluster 1 ( left ) and the interpretation of localization data mapped on the EM density map of the Tail module ( right ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01010 . 7554/eLife . 08719 . 011Figure 2—figure supplement 4 . Exhaustiveness of sampling and robustness of cross-link data .\",\n \"( A ) Comparison of localization density maps calculated from the ensemble of solutions ( 500 best-scoring models ) for the entire sample of models , the first half , the second half , and for jackknifing modeling runs ( where 10% of cross-links were randomly removed ) .\",\n \"( B ) Comparison of localization density maps calculated for Cluster 1 for the entire sample of models , the first half , the second half , and for jackknifing runs .\",\n \"( C , D , E )\",\n \"Precision of Cluster 1 solutions ( diagonal values ) and average RMSD between Cluster 1 solutions ( off-diagonal values ) computed for the four ensembles , considering the whole Mediator complex , the Middle module , and the Tail module . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01110 . 7554/eLife . 08719 . 012Figure 2—figure supplement 5 . Representation of subunit position precision .\",\n \"( A ) Root-mean-square fluctuation ( RMSF ) plots for each subunit of the Middle and Tail modules show how precisely the residues of all proteins were localized within the best cluster of models .\",\n \"For each subunit , RMSF values are mapped onto a stick representation of their centroid model as a color gradient with low and high values of RMSF colored blue and red , respectively .\",\n \"Each subunit model is annotated with several residue numbers for reference .\",\n \"( B ) Precision plots show the relative localization precision of each subunit of the Mediator complex , grouped by membership in the Head , Middle , and Tail domains .\",\n \"The high localization precision of Head module subunits is due to the Head module X-ray structure being fixed at a single position in the EM map .\",\n \"Within the Middle module , the subunits comprising the backbone scaffold ( Med7-21-4-9 ) are all precisely localized . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01210 . 7554/eLife . 08719 . 013Figure 3 . Mediator subunit domain interaction matrix . Average domain–domain contact map calculated for the cluster of best-scoring solutions .\",\n \"Long sequences are divided into domains of 200 residues .\",\n \"The intensity of the color for each box is proportional to the fraction of models for which the contact between corresponding domains is formed .\",\n \"Two domains are in contact when the surface of the beads of one domain is within 10 Å from the surface of any of the beads of the other domain . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 013 The Mediator model ( Figure 2A ) fits the EM map used as a modeling restraint ( Figure 2B and Figure 7—figure supplement 1 ) , satisfied most of the cross-link restraints ( Figure 2—figure supplement 3 ) , and was consistent with almost all data from previous subunit interaction and subunit localization studies ( Figure 2C–E ) .\",\n \"The model explains subunit interactions deduced from co-expression trials ( Kang et al . , 2001; Beve et al . , 2005; Koschubs et al . , 2010 ) , pulldowns/immunoprecipitation ( Kang et al . , 2001; Zhang et al . , 2004 ) , and yeast two-hybrid analyses ( Uetz et al . , 2000; Ito et al . , 2001; Guglielmi et al . , 2004 ) ( Figure 2D , green lines ) .\",\n \"There were only three discrepancies ( Figure 2D , red lines ) , corresponding to the Med3-Med21 , Med1-Med7 , and Med1-Med5 interactions .\",\n \"The Med3-Med21 interaction , from two-hybrid data , is clearly spurious , because the proteins are found in different Mediator modules ( Head and Tail ) and our model , consistent with an EM localization study ( Tsai et al . , 2014 ) , places the two subunits more than 100 Å apart .\",\n \"A second subunit interaction that is unsupported by our model , between the C-terminal domain of Med1 and Med5 , also comes from two-hybrid data ( Guglielmi et al . , 2004 ) .\",\n \"Although it also relates proteins located in different modules ( Middle and Tail ) , the two proteins are less than 30 Å apart in our model .\",\n \"The holoenzyme cross-link dataset includes one cross-link of Med5 ( Tail ) to Med1 ( Middle ) and two cross-links of Med5 ( Tail ) to Med15 ( Tail ) .\",\n \"Because the Med1-Med5 cross-links are inter-modular , they were assigned lower weights by our scoring function and were not satisfied in the final model .\",\n \"It is possible that the interactions occur in alternative conformational states , where Med1 is in close proximity to Med5 .\",\n \"A third subunit interaction that is unsupported by our model , between Med1-Med7 , comes from immunoprecipitation data ( Kang et al . , 2001 ) , and cannot be ruled out , because the N-terminus of Med1 is in the vicinity of the 16 C-terminal residues of Med7 , which are unmodeled in the crystal structure and not well localized in our study ( Figure 2—figure supplement 5 ) .\",\n \"A lack of lysine residues in the 16 C-terminal residues of Med7 might explain why this putative interaction was not detected by cross-linking .\",\n \"Our Mediator model was also validated by comparison with results of previous EM studies that mapped subunit terminal tags or subunit deletions onto 2-D class averages ( Tsai et al . , 2014; Wang et al . , 2014 ) .\",\n \"Our detailed 3-D model was consistent with all 2-D mapping results ( compare Figure 2C , E ) .\",\n \"It also confirms a proposal that a Med7C-Med21-Med4-Med9 tetramer serves as a structural backbone within the central portion of the Middle module ( Lariviere et al . , 2013 ) ( Figure 4 ) . 10 . 7554/eLife . 08719 . 014Figure 4 . Architecture of the Mediator Middle module .\",\n \"( A ) Blow-out views of the Middle module subunit localization maps with the first view and coloring identical to Figure 2A and the second view related by a 180° rotation around the y-axis .\",\n \"( B ) Average contact maps calculated for the cluster of best scoring solutions .\",\n \"Each square is a contact map calculated between a given pair of Middle module proteins with border length proportional to the length of corresponding subunit sequences .\",\n \"The grey-shaded areas indicate observed interactions , with the grey-scale proportional to the fraction of models observing the corresponding interaction .\",\n \"The colored circles are observed cross-links , where the green and orange colors represent respectively satisfied and violated cross-links for the cluster center ( i . e . , the solution that has the minimal root-mean-square deviation ( RMSD ) from all the other solutions in the cluster ) .\",\n \"( C ) Top view of Middle module with individual panels showing the full localization map of each Middle subunit within its surrounding density , made semi-transparent for visual clarity .\",\n \"In the central portion of the Middle module the end-to-end stacking of Med7C-21 ( PDB 1YKH ) and Med4-9 ( Homology Model [Lariviere et al . , 2013] ) heterodimers forms a backbone scaffold ( top panel ) , upon which Med10 and Med19 associate at the Med7C-21 extreme ( left panels ) while Med1 associates closely with the N-terminal portions of Med4-9 ( bottom right panel ) .\",\n \"The Med7N-31 complex ( PDB 3FBI ) is located proximal to the Med7C-21 heterodimer and the unmodeled carboxy-terminal domain ( CTD ) of Med4 ( top panel ) .\",\n \"The centrally located Med17 NTD ( bottom panel ) is wedged between the Med7C-21-4-9 backbone and Med14 , which forms contacts with Med10 and Med21 at its extreme NTD , while the bulk of its density localizes to Med4-9 and Med1 ( top right panel ) .\",\n \"Subunit localization density and ribbon models are colored according to Figure 2A . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 014 The proposed structure of the Med7C-Med21-Med4-Med9 tetramer was based on a set of 40 cross-links , an X-ray crystal structure of Med7C-Med21 , and Med4-Med9 modeled as a structural homolog of Med7C-Med21 ( Lariviere et al . , 2013 ) .\",\n \"Med7C and Med21 form a four-helix bundle , proposed to stack end-to-end on a four-helix bundle of Med4 and Med9 , in a manner analogous to the packing observed in Med7C-21 crystals .\",\n \"In our modeling , we did not impose end-to-end stacking , but rather represented the Med7C-21 crystal structure and the Med4-Med9 homology model as two distinct rigid bodies .\",\n \"In our final Mediator model , the ends of the two four-helix bundles are in close proximity ( Figure 4B , C ) , although without enforcing end-to-end helical stacking .\",\n \"Satisfaction of the EM restraints in other regions of the Middle module appears to require a distortion of the putative end-to-end interaction ( Figure 4C ) , and many solutions displayed a twist of one four-helix bundle with respect to the other , reflecting the asymmetrical nature of the Middle module EM density .\",\n \"The Med7C-Med21-Med4-Med9 tetramer provides a structural scaffold for the association of all remaining Middle module subunits ( Figure 4B , C ) .\",\n \"The Middle module can be subdivided into two halves: Med7C-Med21 , with interacting proteins Med10 , Med19 , and Med31; and Med4-Med9 , with interacting protein Med1 .\",\n \"The only two proteins that interact extensively with both halves of the Middle module are Med14 and Med17 , which perform special architectural roles in the Mediator complex , discussed in more detail below .\",\n \"A Med7N-Med31 heterodimer , connected to the helical Med7C domain by an unstructured 27 amino acid residue linker , is localized to a distinct protrusion within the central portion of the EM map .\",\n \"The flexible C-terminal domain of Med4 is localized to the same region .\",\n \"The Med4-9 half of the Middle module is defined by a close association of the two large subunits Med1 and Med14 with opposite faces of the helical backbone .\",\n \"The N-terminus of Med1 forms an extensive interaction surface with Med4-9 ( Figures 3 , 4 ) , with eight cross-links between residues in the first 418 residues of Med1 and residues of Med4-9 located on one face of the N-terminal four-helix bundle ( Figures 1D , 4B and Supplementary file 1 ) .\",\n \"This high density of cross-links leads to a precise localization of the first half of Med1 , with Cluster 1 showing low root-mean-square fluctuations ( RMSFs ) for residues within this portion of the protein .\",\n \"In contrast , the C-terminus of Med1 lacks cross-links ( Figures 1D , 4B ) and is restrained only by the EM density , resulting in a lower precision of the model and higher RMSF values ( Figure 2—figure supplement 5 ) .\",\n \"The region of the Tail module that abuts the Middle module is a rich subunit interaction hub , composed of residues from the N-termini of Med2 and Med3 and the C-termini of Med15 and Med16 ( Figures 3 , 5 , 6C ) .\",\n \"The C-terminus of Med14 also contributes to the structure of this region , consistent with the finding that a C-terminal deletion of Med14 results in the loss of all Tail subunits from Mediator ( Li et al . , 1995 ) .\",\n \"The Med2 and Med3 NTDs are modeled with relatively high precision ( Figure 2—figure supplement 5 ) , and structural similarity searches with these NTDs show a number of high similarity hits to coiled-coil proteins , in particular to multiple chains within the highly elongated Fibrinogen trimeric coiled-coil structure ( Figure 5—source data 1 ) .\",\n \"We extended this analysis to calculate the odds that each of these NTDs forms a coiled-coil structure , with the use of two independent prediction servers ( Figure 5D ) .\",\n \"We found a good correspondence between regions of similarity to coiled-coil proteins and regions predicted with high confidence ( p > 0 . 8 ) to form coiled-coil structures .\",\n \"These analyses , together with the co-localization of these domains in our Mediator model and the previous observation that Med2 and Med3 can be isolated as a heterodimer ( Beve et al . , 2005 ) , suggest that the N-termini of Med2 and 3 form a coiled-coil motif in the Middle-Tail contact region . 10 . 7554/eLife . 08719 . 015Figure 5 . Architecture of the Mediator Tail module .\",\n \"( A ) Subunit localization within the Tail module .\",\n \"For each Tail subunit , the corresponding localization density is shown within a semi-transparent Tail density ( grey ) and colored according to Figure 2A .\",\n \"( B ) Average contact maps calculated for the cluster of best-scoring solutions and represented as described in Figure 4B .\",\n \"( C ) β-propeller model of the Med16 NTD .\",\n \"The location of 7 WD domains within the N-terminus of Med16 , predicted by MSA analysis ( Bourbon , 2008 ) , is shown in the context of the predicted Med16 N-term secondary structure ( bottom panel ) .\",\n \"Similarity searches identified a high-confidence match to the 7-bladed β-propeller of the S . cerevisiae vesicle coat protein Sec31 ( PDB 2PM9 ) .\",\n \"( D ) Schematic showing the predicted coiled-coil regions of three interacting Tail proteins , Med2 , Med3 and Med15 .\",\n \"The co-localization of the Med2/3 N-termini and Med15 C-term ( panel A ) supports the formation of a coiled-coil at this tail locus . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01510 . 7554/eLife . 08719 . 016Figure 5—source data 1 . HHpred comparative modeling results for Tail module proteins . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01610 . 7554/eLife . 08719 . 017Figure 6 . Novel structural insights into Mediator complex architecture and module connectivity .\",\n \"( A ) Subunit localization density for Med17 ( blue ) reveals the extension of the unmodeled N-terminal domain to contact Middle module subunits .\",\n \"( Left A panel )\",\n \"The N-terminus of the Head subunit Med17 ( blue density and beads ) forms an extensive cross-linking network ( green links ) within the Middle module , acting as an inter-modular bridge ( see right A panel and Supplementary file 1 for detailed cross-link information ) .\",\n \"Head and Middle module ribbon models are colored light grey and according to their subunit color in Figure 2A , respectively , and unmodeled residues cross-linked to Med17 are represented with light grey beads .\",\n \"Only Med17 beads with cross-links are shown for clarity and coordinates are derived from the full-complex centroid model .\",\n \"( Right A panel )\",\n \"A detailed stick representation of the cross-linking network shown in the left panel .\",\n \"( B ) Subunit localization density for Med14 ( salmon ) highlights its unique function as a Mediator scaffold protein spanning ∼220 Å to connect all three modules .\",\n \"( Left B panel )\",\n \"The N-terminus of Med14 forms an extensive cross-linking network across the full Middle module .\",\n \"The structural elements of Med14 and other subunits are represented as in left panel of ( A ) .\",\n \"( Right B panel )\",\n \"A detailed stick representation of the cross-linking network shown in left panel .\",\n \"( C ) The localization of the Med14 CTD to the Middle-Tail junction is observed through a pair of cross-links to a Tail protein located in the region ( Left C panel ) .\",\n \"( Right C panel )\",\n \"A detailed stick representation of the subunit localization and cross-linking network within the Middle-Tail junction .\",\n \"( D ) Subunit localization density for the subunits localized in the central portion of the Tail module .\",\n \"( Left D panel )\",\n \"Bead representation of the cross-links involving the N- and C-terminal portions of Med16 ( yellow ) with Med5 ( brown ) and Med15 ( green ) , respectively .\",\n \"The Med16 N-terminal β-propeller is centrally located in the Tail module and forms extensive interactions with Med5 .\",\n \"The Med16 CTD links Med16-Med5 with the remainder of the Tail .\",\n \"( Right D panel )\",\n \"A detailed stick representation of the cross-linking network shown in left panel .\",\n \"Coordinates for both panels of ( C ) and ( D ) are derived for the Tail module centroid model . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 017 Med15 is co-localized in our Mediator model with Med2 and Med3 and can be isolated with Med2 and Med3 in a trimeric complex ( Zhang et al . , 2004 ) .\",\n \"Central residues of Med15 exhibit homology to Fibrinogen ( Figure 5—source data 1 ) , and may participate in a coiled-coil structure , but it is the C-terminal domain of Med15 ( 850–1081 ) that interacts with Med2 and Med3 in the Middle-Tail contact region .\",\n \"The Med15 CTD also interacts with the Med16 CTD , apparently stabilizing the association of Med16 within the Tail , and that of Med5 as well , because truncation of the Med16 CTD results in the dissociation of both Med16 and Med5 from the Mediator ( Tsai et al . , 2014 ) .\",\n \"Previous sequence analysis of Mediator subunits identified seven WD repeats in the N-terminal domain of Med16 ( Figure 5C ) ( Bourbon , 2008 ) .\",\n \"We extended this analysis and identified a number of high-confidence hits to seven-bladed β-propeller structures .\",\n \"The most complete match to a full β-propeller was to Sec31 , a component of the S . cerevisiae coat protein complex ( COPII ) involved in ER vesicle transport ( Figure 5C and Figure 5—source data 1 ) .\",\n \"We used a homology model , in which the first 540 residues of Med16 were mapped onto the Sec31 β-propeller structure , in our modeling of the Tail module architecture ( Figure 2—figure supplement 2 ) .\",\n \"Two intra-protein cross-links of Med16 could be mapped onto the homology model without violation ( Supplementary file 1 ) .\",\n \"In our model , the Med16 β-propeller was located in a central position at the base of the Tail module , making extensive contacts with Med5 ( Figure 6D ) .\",\n \"The central location of Med16 within the Tail module and a large interaction surface with Med5 are consistent with the finding that the entire Tail module is destabilized and lost in Mediator preparations from a strain from which the MED16 gene was deleted ( Li et al . , 1995 ) .\",\n \"In contrast , Med5 occupies a terminal position within the Tail module , and the MED5 gene can be deleted without disrupting the remainder of the module ( Tsai et al . , 2014 ) .\",\n \"Although the majority of Med5 occupies a single location within the Tail module , the N-domain of Med5 follows a path around the outside edge of the Med16 β-propeller domain before interacting with Med15 at a more internal locus ( Figure 6D ) .\",\n \"This interaction is consistent with an earlier finding that Med5 can form a tetrameric complex with the Med2-Med3-Med15 trimer in the absence of Med16 ( Beve et al . , 2005 ) .\",\n \"Especially noteworthy in our Mediator model are the special roles performed by Med17 and Med14 .\",\n \"An integral component of the Mediator Head module , Med17 is of particular interest for a temperature sensitive allele ( srb4-138ts ) in which all pol II-dependent transcription is abolished at the restrictive temperature ( Thompson and Young , 1995; Holstege et al . , 1998 ) .\",\n \"N-terminal truncations of Med17 in S . cerevisiae are inviable ( data not shown ) .\",\n \"The N-terminal 181 residues of Med17 are disordered in Mediator Head module crystals .\",\n \"Our Mediator model reveals a central role of the Med17 N-terminal domain in the connection between the Head and Middle modules .\",\n \"The Med17 NTD is strongly localized at a central site in the Middle module , wedged between Med7-Med21-Med4-Med9 backbone density on one side and Med14 density on the other ( Figures 2 , 4 ) .\",\n \"The entire path of Med17 from this Middle interaction site to its first modeled residue in the Head module is clearly resolved in the model ( Figure 6A ) .\",\n \"The essential Med14 subunit was originally identified as a repressor protein in yeast ( Sakai et al . , 1990 ) .\",\n \"Reports of the location of Med14 in the Mediator have been contradictory .\",\n \"Med14 truncations uncouple the Tail module , suggesting that Med14 is primarily a Middle module protein ( Li et al . , 1995 ) , but in yeast strains in which the Middle module has been disrupted by subunit deletion , Med14 remains associated with the Tail ( Baidoobonso et al . , 2007 ) .\",\n \"Furthermore , Med14 is required along with reconstituted Head and Middle modules for basal transcription in vitro in Mediator-depleted cell extracts ( Cevher et al . , 2014 ) .\",\n \"Med14 formed cross-links with components of the reconstituted Head and Middle modules , as well as with Med17 ( Cevher et al . , 2014 ) .\",\n \"In our Mediator model , Med14 makes extensive contacts with proteins from all three modules , and is the only Mediator subunit that does so ( Figure 1D ) .\",\n \"It spans almost the entire Mediator complex , extending a distance of about 220 Å from N-terminal interactions with Med10-Med19 at the tip of the Middle module to C-terminal interactions at the Middle-Tail contact region ( Figure 6B , C ) .\",\n \"Med14 residues 264–600 interact with the Med4-Med9 four-helix bundle ( Figures 3 , 4 , 6B ) .\",\n \"Med14 residues 247–350 participate in bridging the Head and Middle modules through multiple contacts with the Med17 NTD .\",\n \"Additional interactions with the Head module proteins are suggested by contacts of Med14 residues 601–1082 with Med20 and the C-terminal region of Med17 ( residues 401–687 ) ( Figure 3 ) .\",\n \"Plaschka et al . ( 2015 ) recently reported a 9 . 7 Å structure of a yeast ‘core initiation complex’ , comprising the Mediator Head module , a minimal Middle module , pol II , a nucleic acid scaffold , and the general transcription factors TBP , TFIIB and TFIIF , obtained by cryo-EM and chemical cross-linking ( Plaschka et al , 2015 ) .\",\n \"The Mediator portion of the cryo-EM map represents only about 50% of the mass of Mediator , as the ‘core initiation complex’ does not include the 480 kDa Tail module or the Middle module protein Med1 .\",\n \"The structure shows the ‘neck’ of the Head module binding to pol II near subunits Rpb 4 and Rpb7 , with the ‘mobile jaw’ regions of Med 18 and Med20 extending to contact the TFIIB-binding site and subunits Rpb 3 , 10 , 11 , and 12 .\",\n \"Cross-links between subunits within the Head ( 30 unique cross-links ) and Middle ( 71 unique cross-links ) modules are in complete agreement with our three-dimensional architectural model of the free Mediator complex ( Figure 7—figure supplement 1B ) .\",\n \"Docking our Mediator model to the EM map of Plaschka et al . ( 2015 ) resulted in a holoenzyme model that was also largely consistent with Mediator to polymerase cross-links ( 12 non-redundant and within the modeled sequence ) identified by Plaschka et al . ( 2015 ) ( Figure 7A , C ) .\",\n \"However , a set of 17 Mediator to polymerase cross-links from our study of holoenzyme showed agreement only in the region where the Head module contacts Rpb4 and Rpb7 ( Figure 7B ) .\",\n \"Cross-links between the Tail and Middle modules to pol II in our study were not consistent with the map of Plaschka et al . ( 2015 ) ( Figure 7C ) .\",\n \"As mentioned above , cross-links between Mediator modules in our study were also poorly satisfied by the EM density for free Mediator used in our integrative modeling .\",\n \"Together , these findings suggest a different conformation of the Mediator-polymerase holoenzyme in the presence of the Tail module ( not included in the map of Plaschka et al . ( 2015 ) ) and in the absence of the nucleic acid scaffold and general factors .\",\n \"Our results point to a similar contact site between the Head module and pol II to that in the model of the ‘core initiation complex’ , but suggest that a conformational change in the Mediator brings the Tail module in contact with pol II near the DNA binding cleft of pol II in the holoenzyme , a suggestion that is consistent with our own preliminary cryo-EM data on the holoenzyme ( P Robinson , unpublished ) ( Figure 7D ) .\",\n \"Such conformational dynamics are in keeping with the idea that interaction between Mediator and pol II in the absence of other factors is largely determined by interactions between the Rpb1 CTD and the neck of the Head module ( Robinson et al . , 2012 ) . 10 . 7554/eLife . 08719 . 018Figure 7 . Holoenzyme cross-linking data indicate different conformational states between the holoenzyme and the core Mediator initiation complex .\",\n \"( A ) Docking of our Mediator model to the EM density of Plaschka et al . ( 2015 ) provides a subunit architectural map that is highly consistent with the 12 Mediator to pol II cross-links of Plaschka et al . ( 2015 ) ( non-redundant and within modeled sequence ) , collected on a core initiation complex containing the Mediator Head and Middle modules with pol II , nucleic acid scaffold and the general factors TBP , TFIIB and TFIIF .\",\n \"( B ) In contrast , the 17 equivalent Mediator to pol II cross-links of the current study are largely inconsistent with the position of Mediator modules found in the core initiation complex .\",\n \"( C ) Distributions of cross-link Cα distances from the two datasets measured in the context of the docked Mediator model .\",\n \"( D ) The pattern of holoenzyme cross-links implies a major conformational rearrangement in the presence of the Tail module and absence of DNA scaffold and general transcription factors . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 01810 . 7554/eLife . 08719 . 019Figure 7—figure supplement 1 . Further validation of the Mediator model . Agreement between the structural model of the Mediator complex and the data for the core initiation complex collected by Plaschka et al . ( 2015 ) .\",\n \"( A ) Comparison between our model ( left ) , the EM map of the apo-mediator complex ( center ) ( Tsai et al . , 2014 ) , and the Mediator Head-Middle EM density from the core initiation complex ( right ) ( Plaschka et al . , 2015 ) .\",\n \"( B ) Agreement between our model and the CX-MS data from the core initiation complex .\",\n \"The histogram is obtained from cross-linked residue–residue distances calculated on our model .\",\n \"In the inset , the box plots are calculated on the inter-module , intra-head module , and intra-middle module cross-links .\",\n \"The solid box represents the interquartile range of the distance distribution .\",\n \"The median is represented by the red line , and the cross-link distances are represented with green crosses .\",\n \"To make the comparison more stringent , we removed distances of intra-molecular cross-links calculated on residues pairs that were less than 50 residues apart in the sequence . DOI: http://dx . doi . org/10 . 7554/eLife . 08719 . 019\"\n ],\n [\n \"Through integration of data from diverse sources , we have arrived at a 3-D model of Mediator , which explains virtually all previous findings , and which provides a complete picture , including internal details , of the complex .\",\n \"One source of data was cross-linking , performed on a native holoenzyme , comprising pol II and Mediator , for several reasons: Mediator structure was thereby analyzed in a more functionally relevant state; the native holoenzyme , isolated from yeast , was much more soluble and stable than a reconstituted holoenzyme formed by mixing pol II and Mediator in vitro; and the X-ray crystal structure of the pol II component of the holoenzyme provided an internal control for the modeling process .\",\n \"We restricted our analysis to the internal organization of the three Mediator modules , because their relationship to one another is believed to differ between free Mediator and the holoenzyme , and free Mediator was the source of our EM information .\",\n \"Notable features of our 3-D Mediator model include the following:The definition of the three Mediator modules .\",\n \"Whereas Med17 and Med14 were previously classified as Head and Tail proteins , the N-terminal domains of both proteins are now seen to form part of the Middle module .\",\n \"The two NTDs not only interact with core components of the Middle module but also with each other . The special roles of Med17 and Med14 in the architecture of the entire Mediator .\",\n \"Med17 connects Head and Middle modules , while Med14 extends the entire length of Mediator , connecting all three modules .\",\n \"The path of the NTD of Med17 , not revealed in X-ray crystal structures of the Head module , can now be followed from a site of interaction in the Middle module to its connection with the body of the protein in the Head module . The complete 3-D arrangement of subunits of the Middle module .\",\n \"A Med7-Med21-Med4-Med9 tetramer is seen to form the core of the module , with a cap of Med10-Med19 at one end and a cap of Med1 at the other . The complete 3-D arrangement of components of the Tail module .\",\n \"Of particular note are the extensive interactions of the NTDs of Med2 and Med3 and the CTDs of Med 15 and Med16 , which form the region of the Tail module that contacts the Middle module , and which is important for Mediator-activator protein interaction .\",\n \"The Med2-Med3-Med15 triad is a binding target of multiple yeast transcriptional activator proteins , including Gcn4 and Gal4 ( Lee et al . , 1999; Myers et al . , 1999; Zhang et al . , 2004 ) .\",\n \"Yeast harboring a Med3Δ Med15Δ double knockout exhibits decreased levels of Mediator and of transcription pre-initiation complexes at inducible gene promoters ( Ansari et al . , 2012 ) .\",\n \"Secondary structural features of Tail module proteins .\",\n \"The NTDs of Med2 and Med3 interact in a coiled-coil motif in the Middle-Tail junction .\",\n \"A 7-bladed β-propeller encompassing the N-terminal half of Med16 interacts extensively with Med5 at the base of the Tail .\",\n \"Both proteins play roles in transcriptional repression , and expression analysis in yeast Tail mutants has established a functional link between them ( van de Peppel et al . , 2005 ) .\",\n \"The 3-D Mediator model not only unifies and resolves discrepancies in the extensive literature on Mediator organization , but also forms a starting point for future investigation .\",\n \"High-resolution structures can be docked and additional structural information incorporated in the model , using the same integrative approach we applied here ( Russel et al . , 2012 ) .\",\n \"Interactions with pol II , general transcription factors , and transcriptional regulatory proteins can be mapped to the surface .\",\n \"Changes associated with activity or due to mutation can be identified and their functional significance understood .\",\n \"Our Mediator model could be docked into the Mediator portion of the EM map of Plaschka et al . ( 2015 ) , as association with pol II appears to cause little change in the structure and relative position of the Head and Middle Mediator modules compared to those in the free Mediator EM map used as restraint in our modeling study ( Tsai et al . , 2014 ) ( Figure 7—figure supplement 1A ) .\",\n \"Our docked model provides a more detailed map of the locations of Middle module subunits across the pol II surface , and therefore allows a more complete interpretation of the Mediator portion of the map of Plaschka et al . ( 2015 ) .\",\n \"Our model provides a structural basis for the Mediator to pol II cross-links reported by Plaschka et al . ( 2015 ) , which can be mapped with precision onto both the Mediator and pol II surfaces to yield a structurally consistent set of cross-link distances ( Figure 7A , C ) .\",\n \"Furthermore , mapping novel Middle module cross-links from Plaschka et al . ( 2015 ) onto our Mediator model , we find additional strong support for the validity of our subunit localizations ( Figure 7—figure supplement 1B ) .\",\n \"Also as mentioned above , the cross-links between Mediator modules that we obtained with holoenzyme are inconsistent with the result of docking our Mediator model into the map of Plaschka et al . ( 2015 ) ( Figure 7B , C ) .\",\n \"We found multiple cross-links between the Tail module and the region of pol II around the active center cleft , whereas docking our Mediator model into the map of Plaschka et al . ( 2015 ) places the Tail module on the opposite side of pol II .\",\n \"We have noted the Mediator conformational change required to satisfy our Tail module-pol II cross-links ( Figure 7D ) .\",\n \"Tail module interactions with pol II in the vicinity of the active center cleft may be relevant to the role of Tail module subunits Med14 , 15 and 16 in transcriptional repression at multiple loci in yeast ( Covitz et al . , 1994; Sussel et al . , 1995 ) .\",\n \"The success of the integrative modeling approach in the derivation of Middle and Tail module structures warrants application to the issues raised here and to other challenging problems .\",\n \"It is of immediate interest to apply IMP to the holoenzyme with an expanded cross-link dataset and a high resolution holoenzyme EM map .\",\n \"Further studies may be directed towards a complete RNA polymerase II pre-initiation complex including Mediator , and to additional biological particles of great size and complexity .\"\n ],\n [\n \"The S . cerevisiae Holoenzyme was purified as an intact complex using a yeast strain in which both endogenous Mediator and RNA Pol II were affinity tagged .\",\n \"To achieve double tagging the following genetic manipulations were performed in the CB010 background ( Matα pep4::HIS3 prb1::LEU2 prc1::HISG can1 ade2 trp1 ura3 his3 leu2–3 leu2–112 ) .\",\n \"First , the N-terminus of the essential Mediator Head module subunit , Med17p , was TAP-tagged using the N-terminus tagging cassette , including Kluyveromyces lactis TRP1 marker gene , amplified from vector pBS1761 as described previously ( Puig et al . , 2001 ) .\",\n \"The MED17 gene was returned to endogenous promoter regulation following Cre-mediated LoxP recombination driven from pSH47 ( Ura+ ) .\",\n \"The C-terminus of the largest RNA Pol II subunit , Rpb1p , was tagged with a custom protein G affinity tag with 3C PreScission protease cleavage site .\",\n \"The vector for C-terminal protein G tag cassette amplification [pCEMM-CTAP ( KanMX ) ] was first generated by sub-cloning a 1633 bp fragment containing the Tn903 KanMX marker gene with flanking LoxP sites from pUG6 between the XhoI ( 597 ) and NotI ( 616 ) sites of mammalian expression vector pCEMM-CTAP ( SG ) , positioned immediately C-terminal to the protein G domain sequence ( Forward primer 5′-3′: actgtcactgaatagctcgaggaacgcggccgccagctgaagcttcgtacg–Reverse primer 5′-3′: ggcggaatttacgtagcggccgcccgcggccgcataggccactagtggatctg ) .\",\n \"The Rpb1 C-terminal protein G tagging cassette including the protein G and KanMX elements was PCR amplified from pCEMM-CTAP ( KanMX ) using primers with flanking sequence including 45 bp complementary to Rpb1 genomic DNA immediately upstream/downstream of the STOP codon and sequence encoding the 3C PreScission cleavage site ( 5′-3′: ctggaagttctgttccaaggtcca ) .\",\n \"The endogenous yeast complex was purified from whole-cell extracts using an affinity capture method described previously with minor modifications ( Robinson et al . , 2012 ) .\",\n \"Double-tagged CB010 were grown in 200L YPAD to 9 . 0 A600 and 3 . 0 Kg cells were lysed by continuous-flow bead beating in A25 buffer ( 25 mM ammonium sulfate , 100 mM HEPES pH 7 . 8 , 1 mM EDTA , 5% glycerol , 5 mM DTT , and 1× protease inhibitor cocktail [0 . 6 µM leupeptin hemisulfate , 2 µM pepstatin A , 1 mM PMSF and 2 . 1 mM benzamidine hydrochloride] ) with 25 µg/ml RNAse A ( Qiagen , Hilden , Germany ) .\",\n \"The cell debris was pelleted by centrifugation at 12 , 250×g for 1 hr at 4°C and the supernatant made up to 300 mM ammonium sulfate with the addition of cold 3 . 9 M ammonium sulfate and gentle stirring .\",\n \"Nucleic acids were depleted from the lysate by adding 500 ml DEAE ( GE healthcare , Little Chalfont , UK ) pre-equilibrated in A300 buffer .\",\n \"After 30 min of stirring the DEAE was pelleted by centrifugation ( 12 , 250×g , 30 min at 4°C ) and the supernatant loaded onto an IgG column before washing in A500 buffer .\",\n \"The use of two distinct protease cleavage sites allowed the isolation of stable holoenzyme complex using a two-stage elution approach ( Figure 1—figure supplement 1 ) .\",\n \"First-stage TEV cleavage led to the release of free Mediator , with Mediator retention occurring through complex formation with uncleaved RNA pol II .\",\n \"Subsequent cleavage with 3C PreScission protease led to the release of both free RNA pol II and holoenzyme complex .\",\n \"The order of protease cleavage could be reversed with equivalent yields of holoenzyme complex ( ∼5 mg/3 Kg dry cell mass ) .\",\n \"The holoenzyme was further purified from free RNA pol II ( TEV-3C ) or Mediator ( 3C-TEV ) using ion exchange ( 75–600 mM ( NH4 ) 2SO4; HiTrap FFQ ) and size-exclusion ( 200 mM ( NH4 ) 2SO4 , 25 mM HEPES , pH 7 . 8 , 5% glycerol , 2 mM DTT; Superose 6 ) chromatography steps .\",\n \"In the case of the TEV-3C cleavage order , the holoenzyme eluted from the IE column as a leading shoulder in the large peak due to free RNA pol II .\",\n \"Using the 3C-TEV cleavage order , the holoenzyme was retained longer in IE than free Mediator and could be resolved as an independent peak .\",\n \"Prior to cross-linking , holoenzyme samples were exchanged into phosphate buffer containing: 150 mM NaPO4 , 150 mM KOAc , pH 7 . 5 , 5% glycerol and 2 mM DTT by dialyzing against three changes of buffer .\",\n \"Cross-linking was then performed by one of two procedures: MS Peaklists were generated such that all monoisotopic ions detected within the selection window of the linear ion trap ( 4 m/z units ) were annotated as possible precursors to a given product ion spectrum , as described previously ( Robinson et al . , 2012; Trnka et al . , 2014 ) .\",\n \"Peaklists were first searched against the full SwissProt database to check the composition of the sample .\",\n \"Cross-link searches were then carried out against a target database consisting of the 52 components of the S . cerevisiae Preinitiation Complex .\",\n \"For each polypeptide sequence , 10 randomized versions were generated and concatenated to the target database for a total of 572 sequences ( 52 target and 520 decoy ) .\",\n \"Decoy protein sequences were the same length as their corresponding target sequences and sampled from the natural distribution of S . cerevisiae amino acid frequency .\",\n \"Peaklists were then searched against this database using Protein Prospector ( version 5 . 13 . 1 ) to assign putative cross-linked peptides ( Trnka et al . , 2014 ) .\",\n \"Carbamidomethylation of cysteine was considered as a fixed modification .\",\n \"N-terminal methionine loss with and without acetylation , peptide N-terminal glutamine conversion to pyroglutamate , oxidation of methionine , and ‘dead-end’ modification of lysine and the protein N-terminus by semi-hydrolyzed BS3 were considered as variable modifications in addition to cross-linking by BS3 .\",\n \"Up to 3 variable modifications per peptide were considered .\",\n \"Mass tolerances of 13 ppm and 20 ppm were used for precursor and product ions respectively .\",\n \"Trypsin specificity with 3 missed cleavages was used to generate theoretical peptides .\",\n \"85 product ion signals from each MS2 spectrum were used for assessment .\",\n \"For experiments involving isotopic mixtures of cross-linking reagents , the searches were performed twice , once specifying heavy BS3 and once with light BS3 .\",\n \"The search results were later merged .\",\n \"Both light and heavy dead-end modifications were considered in each of these database searches .\",\n \"For all experiments , cross-link spectral matches ( CSM ) were discarded if the following Protein Prospector parameters fell outside the threshold values: peptide score below 20 , peptide/protein/worse_peptide expectation values above 50 , and score_difference below 0 .\",\n \"A linear support vector machine ( SVM ) model was constructed to classify CSMs between decoy and target classes ( Trnka et al . , 2014 ) .\",\n \"Briefly , the CSMs were split evenly into two sets .\",\n \"SVM models were trained on half of the data and evaluated for their ability to correctly classify true negatives ( hits to the decoy database ) on the test set .\",\n \"CSMs were considered decoy matches if either of the two constituent peptides came from the decoy database .\",\n \"The final SVM model included a metric reflecting the number of product ions matched to the less fragmented of the two peptides ( score_difference ) and a metric for the overall match of the spectrum ( percentage of the total ion current annotated ) .\",\n \"The inclusion of additional metrics did not increase the predictive power of the model .\",\n \"An SVM decision value of 0 was taken as the acceptance threshold for reporting cross-linked peptides .\",\n \"FDRs were calculated by dividing the number of decoy hits at this threshold by 10 ( to account for the larger decoy database size ) and dividing by the number of hits to the target database .\",\n \"CSMs where either peptide was less than 4-amino acids long were discarded , with a few manually validated exceptions .\",\n \"When spectra could be interpreted by multiple cross-linked peptide-pairs , CSMs were considered to be unambiguous if the second best match was further than 0 . 3 SVM decision value units from the top match .\",\n \"Intra-protein cross-links were selected over inter-protein cross-links if there were still conflicts .\",\n \"The remaining decoy matches were also removed from consideration .\",\n \"Following these selection criteria , most CSMs identified a single cross-linked peptide-pair .\",\n \"The remaining spectral redundancy came exclusively from ambiguous site localizations , where there was sufficient evidence to identify both peptides , but the position of the modified amino acid remained ambiguous .\",\n \"In these cases , all possible sites were reported using a spectral identifier and used in subsequent modelling steps .\",\n \"Cross-linking results and annotated HCD spectra may be viewed online using Protein Prospector's MS-Viewer program: http://prospector2 . ucsf . edu/prospector/cgi-bin/msform . cgi ?\",\n \"=msviewer .\",\n \"Search key: qzpxihtngx All raw mass spectrometry files and peaklists used in the study have been deposited to the MassIVE proteomics repository and are available through the ProteomeXchange consortium .\",\n \"MassIVE accession: MSV000079237 ProteomeXchange accession: PXD002723 Our integrative structure modeling of RNA Pol II and Mediator complexes proceeds through four stages ( Lasker et al . , 2010; Fernandez-Martinez et al . , 2012; Lasker et al . , 2012 ) ( Figure 2 ) : ( 1 ) gathering of data , ( 2 ) representation of subunits and translation of the data into spatial restraints , ( 3 ) configurational sampling to produce an ensemble of models that satisfies the restraints , and ( 4 ) analysis and assessment of the ensemble .\",\n \"The modelling protocol ( i . e . , stages 2 , 3 , and\",\n \"4 ) was scripted using the Python Modelling Interface ( https://github . com/salilab/pmi ) , version be72c15 , a library to model macromolecular complexes based on our open source IMP package ( http://salilab . org/imp/ ) , version 829c3f0 ( Russel et al . , 2012 ) .\",\n \"Files containing the input data , scripts , and output models are available at http://salilab . org/mediator .\",\n \"In the IMP modeling procedure we perform a large number of independent modeling runs , each starting from a random configuration and undergoing thousands of rounds of small random model perturbations , in order to achieve exhaustive sampling of conformational space .\",\n \"As we are interested only in models that best satisfy the restraints , we select a small fraction of the models representing the very best scoring solutions ( Figure 2—figure supplement 3A ) .\",\n \"These models then undergo pairwise-RMSD clustering to separate out divergent structural states ( Figure 2—figure supplement 3B ) .\",\n \"In our study , most of the best scoring models ( 92% , n = 500 ) gave a unique subunit arrangement for 18 of the 21 Mediator subunits .\",\n \"The divergent clusters differed only in the relative locations of Med5 , Med15 and Med16 within the Tail module ( Figure 2—figure supplement 3C ) .\",\n \"We compared these three clusters in terms of crosslink violation statistics ( Figure 2—figure supplement 3E ) and consistency with prior EM localization experiments ( Figure 2—figure supplement 3H ) .\",\n \"A single cluster showed both consistency with earlier localizations and significantly better violation statistics and was therefore deemed to be the Mediator architecture with the highest confidence .\",\n \"Modeling studies of RNA pol II employed cross-link data from two sources: 156 cross-links from previous experiments conducted with the free pol II fraction isolated during our holoenzyme SEC purification ( Trnka et al . , 2014 ) , and 108 cross-links from a previous study ( Chen et al . , 2010 ) .\",\n \"These sources had 63 cross-links in common .\",\n \"Hence , the current study used 201 unique cross-links .\",\n \"The atomic structures for the twelve-subunit yeast pol II have been previously determined by X-ray crystallography ( PDB accession ID: 1WCM , [Armache et al . , 2005] ) .\",\n \"The crystallographic structure covers 89% of the residues in the pol II complex .\",\n \"A 20 . 9 Å resolution density map from single particle EM reconstruction of the pol II complex was also considered ( EMDB accession id: 1883 , [Czeko et al . , 2011] ) .\",\n \"The 12 pol II subunits were represented as 15 domains , where Rpb1 was decomposed into 3 domains ( residues: 1–1140 , 1141–1274 , 1275–1733 ) , Rpb2 was decomposed into 2 domains ( residues 1–1102 , 1103–1224 ) , and the remaining subunits were represented as single domains .\",\n \"Regions without a crystal structure were modeled as beads containing up to 5 amino acid residues .\",\n \"We followed the same 4-stage procedure as described for the Mediator complex above .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["The 21-subunit Mediator complex transduces regulatory information from enhancers to promoters , and performs an essential role in the initiation of transcription in all eukaryotes .","Structural information on two-thirds of the complex has been limited to coarse subunit mapping onto 2-D images from electron micrographs .","We have performed chemical cross-linking and mass spectrometry , and combined the results with information from X-ray crystallography , homology modeling , and cryo-electron microscopy by an integrative modeling approach to determine a 3-D model of the entire Mediator complex .","The approach is validated by the use of X-ray crystal structures as internal controls and by consistency with previous results from electron microscopy and yeast two-hybrid screens .","The model shows the locations and orientations of all Mediator subunits , as well as subunit interfaces and some secondary structural elements .","Segments of 20–40 amino acid residues are placed with an average precision of 20 Å .","The model reveals roles of individual subunits in the organization of the complex ."],"string":"[\n \"The 21-subunit Mediator complex transduces regulatory information from enhancers to promoters , and performs an essential role in the initiation of transcription in all eukaryotes .\",\n \"Structural information on two-thirds of the complex has been limited to coarse subunit mapping onto 2-D images from electron micrographs .\",\n \"We have performed chemical cross-linking and mass spectrometry , and combined the results with information from X-ray crystallography , homology modeling , and cryo-electron microscopy by an integrative modeling approach to determine a 3-D model of the entire Mediator complex .\",\n \"The approach is validated by the use of X-ray crystal structures as internal controls and by consistency with previous results from electron microscopy and yeast two-hybrid screens .\",\n \"The model shows the locations and orientations of all Mediator subunits , as well as subunit interfaces and some secondary structural elements .\",\n \"Segments of 20–40 amino acid residues are placed with an average precision of 20 Å .\",\n \"The model reveals roles of individual subunits in the organization of the complex .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Inside a cell , proteins are made from instructions encoded by DNA .","To produce a particular protein , a section of DNA within a gene is copied into a molecule of messenger ribonucleic acid ( or mRNA ) .","This process is called transcription and is carried out by an enzyme known as RNA polymerase .","Transcription begins in a region of DNA called a promoter , which is found at the start of the gene .","RNA polymerase is brought to the DNA by many proteins , including the so-called Mediator complex .","Mediator receives signals from within the cell and from the environment , processes the information , and instructs RNA polymerase whether to transcribe the gene or not .","Mediator performs this important role in all organisms from yeast to humans , but it is not clear how it works .","A crucial step towards the solution of this problem is to understand the three-dimensional structure of the complex .","Previous research using a technique called ‘electron microscopy’ showed that Mediator is composed of three modules , referred to as Head , Middle and Tail .","The images from electron microscopy were not sufficiently detailed to reveal the organization of the proteins within these modules .","An open-source Integrative Modeling Platform ( IMP for short ) was recently developed to arrive at structural models of large protein complexes from a combination of experimental data and computer models .","Now , Robinson , Trnka , Pellarin et al . have used this platform to study the Mediator complex .","First , Robinson , Trnka , Pellarin et al . collected experimental data on the structure of the Mediator complex using two approaches called ‘chemical cross-linking’ and ‘mass spectrometry’ .","This data was combined with biochemical and structural information from previous studies to generate a three-dimensional model of the structure of the entire Mediator using IMP .","The model is detailed enough to show the location and orientation of all the proteins in the complex .","For example , a protein called Med17 connects the Head and Middle modules , while another subunit—known as Med14—spans the entire complex and makes extensive contacts with other proteins in all three modules ."],"string":"[\n \"Inside a cell , proteins are made from instructions encoded by DNA .\",\n \"To produce a particular protein , a section of DNA within a gene is copied into a molecule of messenger ribonucleic acid ( or mRNA ) .\",\n \"This process is called transcription and is carried out by an enzyme known as RNA polymerase .\",\n \"Transcription begins in a region of DNA called a promoter , which is found at the start of the gene .\",\n \"RNA polymerase is brought to the DNA by many proteins , including the so-called Mediator complex .\",\n \"Mediator receives signals from within the cell and from the environment , processes the information , and instructs RNA polymerase whether to transcribe the gene or not .\",\n \"Mediator performs this important role in all organisms from yeast to humans , but it is not clear how it works .\",\n \"A crucial step towards the solution of this problem is to understand the three-dimensional structure of the complex .\",\n \"Previous research using a technique called ‘electron microscopy’ showed that Mediator is composed of three modules , referred to as Head , Middle and Tail .\",\n \"The images from electron microscopy were not sufficiently detailed to reveal the organization of the proteins within these modules .\",\n \"An open-source Integrative Modeling Platform ( IMP for short ) was recently developed to arrive at structural models of large protein complexes from a combination of experimental data and computer models .\",\n \"Now , Robinson , Trnka , Pellarin et al . have used this platform to study the Mediator complex .\",\n \"First , Robinson , Trnka , Pellarin et al . collected experimental data on the structure of the Mediator complex using two approaches called ‘chemical cross-linking’ and ‘mass spectrometry’ .\",\n \"This data was combined with biochemical and structural information from previous studies to generate a three-dimensional model of the structure of the entire Mediator using IMP .\",\n \"The model is detailed enough to show the location and orientation of all the proteins in the complex .\",\n \"For example , a protein called Med17 connects the Head and Middle modules , while another subunit—known as Med14—spans the entire complex and makes extensive contacts with other proteins in all three modules .\"\n]","isConversational":false},"year":{"kind":"string","value":"2015"}}},{"rowIdx":96,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["neuroscience"],"string":"[\n \"neuroscience\"\n]","isConversational":false},"title":{"kind":"string","value":"Working memory capacity of crows and monkeys arises from similar neuronal computations"},"id":{"kind":"string","value":"elife-72783-v2"},"sections":{"kind":"list like","value":[["Working memory ( WM ) can hold information for a short period of time to allow further processing in the absence of sensory input ( Cowan , 2017; Oberauer et al . , 2018 ) .","By bridging this gap between the immediate sensory environment and behavior , WM is a cornerstone for complex cognition .","It is a very flexible memory system , yet severely limited in its capacity .","While this capacity is often seen as a general cognitive bottleneck , for simple stimuli , like colors , the capacity is very similar between humans , monkeys , and crows ( Balakhonov and Rose , 2017; Buschman et al . , 2011; Cowan , 2001; Luck and Vogel , 1997 ) .","Different models have been proposed to conceptualize how this capacity limit arises .","This work motivated many psychophysical and electrophysiological experiments that in turn led to a spectrum of more refined models of WM ( Ma et al . , 2014 ) .","‘Discrete models’ of WM argue that a fixed number of items can be stored .","Once this capacity is reached , an additional item can only be maintained if it replaces a previous item ( Awh et al . , 2007; Fukuda et al . , 2010; Luck and Vogel , 1997; Vogel and Machizawa , 2004; Zhang and Luck , 2008 ) .","‘Continuous models’ describe WM as a flexible resource that is allocated to individual items .","A minimum amount of this resource has to be allocated to each item for successful retention , thereby resulting in a capacity limit ( Bays and Husain , 2008; van den Berg et al . , 2012; Wilken and Ma , 2004 ) .","On the neurophysiological level , models of WM capacity suggest that interference between memory representations ( ‘items’ ) within the neuronal network is a source of information loss and capacity limitation ( Bouchacourt and Buschman , 2019; Lundqvist et al . , 2016; Lundqvist et al . , 2011; Schneegans et al . , 2020 ) .","Interference may arise due to divisive normalization that appears as competition between items , related to oscillatory dynamics ( Lundqvist et al . , 2016; Lundqvist et al . , 2011 ) , WM flexibility ( Bouchacourt and Buschman , 2019 ) , and neuronal information sampling ( Schneegans et al . , 2020 ) .","Divisive normalization is a computational principle that acts upon neurons when presenting multiple stimuli simultaneously , it normalizes neuronal responses by creating ‘a ratio between the response of an individual neuron and the summed activity of a pool of neurons’ ( Carandini and Heeger , 2011 , p . 51 ) .","An effect related to divisive normalization can be observed when two stimuli are presented either individually or simultaneously within the receptive field of a visual sensory neuron .","The neuron’s firing rate when the stimuli are presented simultaneously becomes normalized by the populations’ responses to each individual stimulus ( Carandini et al . , 1997; Heeger , 1992 ) .","This effect also occurs in relation to attentive processes and has been formalized into the ‘normalization model of attention’ ( Reynolds et al . , 1999; Reynolds and Heeger , 2009 ) .","Normalization of neuronal responses is commonly observed in many species throughout the animal kingdom , not just in sensory , but also in cognitive domains ( Carandini and Heeger , 2011 ) .","Investigations into WM capacity and model predictions focus mostly on humans and monkeys .","By extending this work to include birds , one can gain a unique comparative perspective .","Crows have a similar limit in WM capacity and neuronal correlates of WM are comparable to monkeys’ ( Balakhonov and Rose , 2017; Nieder , 2017 ) .","But while the neuronal architecture of sensory areas is similar between birds and mammals , higher associative areas , critical for WM , do not share a common architecture between the species ( Stacho et al . , 2020 ) .","Therefore , an outstanding question is whether modern models of WM such as the ‘flexible model’ capture WM capacity in general , or if their predictions ( e . g . , divisive normalization ) are confined to the mammalian neocortex .","To resolve this , it is crucial to investigate the avian brain to understand how its different organization can produce such similar behavioral and neurophysiological results .","While the neuronal correlates of WM maintenance in birds have been investigated in some detail ( Diekamp et al . , 2002a; Hartmann et al . , 2018; Rinnert et al . , 2019; Rose and Colombo , 2005; Veit et al . , 2014 ) , a neurophysiological investigation of WM capacity limitation is still lacking .","The avian forebrain structure , nidopallium caudolaterale ( NCL ) is a critical component of avian WM .","The NCL is considered functionally equivalent to the mammalian prefrontal cortex ( PFC ) ( Güntürkün and Bugnyar , 2016; Nieder , 2017 ) , as it receives projections from all sensory modalities ( Kröner and Güntürkün , 1999 ) , projects to premotor areas ( Kröner and Güntürkün , 1999 ) , and is a target of dopaminergic innervation ( Waldmann and Güntürkün , 1993 ) .","To investigate the neurophysiology of WM capacity in birds , we adopted a task design developed for monkeys ( Buschman et al . , 2011 ) to use it with carrion crows ( Corvus corone ) .","Our animals were trained to memorize an array of colors and to indicate which color had changed after a short memory delay , while we performed extracellular recordings of individual neurons in the NCL using multi-channel probes .","We expected to find a clear correlate of WM representations in NCL neurons and a load-dependent response modulation based on divisive normalization of neuronal responses .","This would allow us to evaluate if the behavioral WM capacity observations of crows fit a ‘discrete’ or ‘continuous’ WM resource model .","If the neuronal responses also fit the contemporary neurophysiological models of WM capacity limitations ( Bouchacourt and Buschman , 2019; Lundqvist et al . , 2016; Lundqvist et al . , 2011; Schneegans et al . , 2020 ) , it would further suggest that crows and monkeys have convergently evolved a similar neurophysiological basis for WM capacity despite a different architecture of the critical forebrain structures ."],["The behavioral performance was influenced by the number of colored squares on the screen .","It significantly decreased with an increasing number of ipsilateral squares ( median performances , load 1: 95 . 88% , load 2: 78 . 31% , load 3: 58 . 21%; Friedman test: Χ² = 92 . 00 , p < 0 . 001 , Figure 1B ) .","We ran a generalized linear model with ipsilateral load ( i . e . , load of hemifield where a color changed ) , contralateral load ( i . e . , load of hemifield without a color change ) and their interaction as predictors for performance ( R2adj 0 . 78 , F ( 460 , 456 ) = 555 . 00 , p < 0 . 001 ) .","We found that the number of ipsilateral colors significantly reduced performance ( βipsi –0 . 177 , t ( 459 ) = –18 . 00 , p < 0 . 001 ) , whereas the number of contralateral colors did not ( βcontra –0 . 021 , t ( 459 ) = –1 . 77 , p = 0 . 0772; Figure 1C ) .","There was also a significant interaction between ipsilateral and contralateral load ( β = –0 . 024 , t ( 458 ) = –4 . 28 , p < 0 . 001 ) .","We compared this model to a reduced model , where we omitted the non-significant βcontra and found that this reduced model ( R2adj 0 . 78 , F ( 460 , 457 ) = 828 . 00 , p < 0 . 001 ) explained the performance equally as well ( |ΔLLR| = 0 . 0102 ) .","Therefore , we conclude that contralateral load by itself did not significantly affect performance .","We calculated the capacity K ( see Materials and methods ) for all full WM loads ( i . e . , two to five items ) .","The capacity K peaks at four items ( mean ± SEM: 3 . 05± . 038 , Figure 1—figure supplement 1 ) .","These observations are very similar to observations made in primates ( Buschman et al . , 2011 ) and fully reproduce our earlier behavioral findings ( Balakhonov and Rose , 2017 ) .","We recorded 362 neurons from the NCL of two crows performing the WM task ( delayed change localization ) .","All reported effects were also present in each individual bird ( Figure 3—figure supplements 1–2 ) , we , therefore , pooled the data for population analysis .","A large subset of neurons responded to the presence of a color ( i . e . , at load 1 ) by substantially increasing or decreasing their firing rate relative to baseline .","This change in firing rate occurred selectively , depending on the presented color either in the sample ( Figure 2A ) or the delay period ( Figure 2—figure supplement 1 ) .","For most neurons , this difference in firing rate between the two possible colors became attenuated when the load increased from one to two colors , and it was further attenuated from two to three colors .","To quantify this effect , we calculated the amount of information about the color identity at a neuron's favorite location as the percent explained variance ( PEV ) during a memory load of one , two , or three items in bins of 200 ms ( see Materials and methods for details ) .","Most neurons did not sustain information about color ( measured as a significant PEV , henceforth ‘information’ ) throughout the entire sample or memory delay but rather had shorter periods in which the information was significant ( Figure 2A bottom ) .","To better capture the time points when the individual neurons carried information , we performed a hierarchical clustering analysis of the PEV values of the individual neurons at load 1 ( see Materials and methods for details ) .","We found a total of seven clusters that were organized into two overarching groups ( Figure 3A ) .","Group 1 contained neurons ( n = 227 ) that showed peak information during the sample and early delay phase , while group 2 contained neurons ( n = 135 ) that showed peak information during the delay phase .","For each neuron , we then calculated if it carried a significant amount of color information by applying a permutation test ( for all bins at load 1 , see Materials and methods ) .","The individual neurons were then further classified into three groups depending on the phase in which they had a significant amount of information ( Figure 3B ) .","Overall , 37 . 57% ( n = 136 ) of neurons were significant during the sample phase , 9 . 39% ( n = 34 ) of neurons were significant during the memory delay , and 14 . 64% ( n = 53 ) of neurons were significant during both the sample phase and the memory delay ( all proportions of neurons were significantly higher than expected by chance [binomial test , see Materials and methods , all p < 0 . 001] ) .","Refer to Figure 2A for an example neuron , significant at load 1 with a large differentiation in firing rate between color identities ( a large PEV ) and a loss of differentiation with increasing ipsilateral load .","Further inspection of individual neuronal activity revealed , however , that a substantial number of neurons responded differently .","Instead of losing information at higher loads , many neurons gained information ( e . g . , Figure 2B , Figure 2—figure supplement 2 ) .","Thus , we additionally performed the permutation testing for loads 2 and 3 to determine which neurons had significant information ( see Materials and methods ) .","We found that many of the neurons that did not have significant information at load 1 did have significant information at load 2 and load 3 ( Figure 3C ) .","For the memory delay , more than half of the significant neurons we detected were only significant for either load 2 or load 3 , compared to only 36% of neurons that were significant at load 1 ( Figure 3C middle ) .","By including the higher loads in our analysis , we found a total of 249 ( 68 . 78% ) sample neurons and 94 ( 25 . 97% ) delay neurons .","For the population analyses , we subsequently pooled all significant neurons into three groups ( one per load ) .","These pooled groups were then each subdivided into sample and delay neurons ( i . e . , ‘sample-load1’ , ‘delay-load1’ , ‘sample-load2’ , etc . , see Table 1 in the Materials and methods for an overview ) .","The clustering analysis indicated that the population of neurons as a whole did sustain the color information throughout the entire trial ( Figure 3A ) .","Plotting the information averages of each of the three ‘sample populations’ and the ‘delay populations’ over time confirmed this result ( Figure 4A , Figure 4—figure supplement 1 ) .","After the onset of the stimulus array , the average information exhibited a sharp increase that peaked roughly 400 ms after stimulus onset and remained at an elevated level throughout the memory delay , until the choice array appeared .","Results obtained from neurons of the lateral PFC of monkeys indicated distinct hemispheric independence of WM capacity ( Buschman et al . , 2011 ) .","This means that increasing ipsilateral load ( i . e . , load in the hemifield containing the target for which information is assessed ) should affect neuronal processing while increasing contralateral load should not .","This effect might be further emphasized in birds due to the full decussation of their optic nerve ( Husband and Shimizu , 2001 ) .","Parallel to the behavioral results and in line with the results from monkeys , we found a strong effect of ipsilateral load on the information maintenance , as there was a sharp drop in information when the load increased from one item to two items ( Figure 4A , blue and yellow curves ) .","The addition of a third item only slightly decreased the maintained information further ( Figure 4A , red curve ) .","The load dependence was much more pronounced during the sample period than during the memory delay where the information remained at a lower elevated level .","Notably , the load effect was only present for ipsilateral manipulations .","If the number of items on the contralateral side was increased , the information encoded about the colors at the favorite location did not change ( Figure 4A , right ) .","To compare our results to the results obtained in monkeys we also applied the method of Buschman et al . , 2011 for testing the ipsilateral load effect during the sample and delay phase , by splitting each phase into an early and a late portion ( first and second 400 ms of the sample , and first and second 500 ms for the delay ) .","We did find a significant drop in information with an increase in load from one through three in the early ( F ( 2 , 537 ) = 18 . 73 , p < 0 . 001 , ω² = 0 . 0616 ) and late ( F ( 2 , 536 ) = 20 . 07 , p < 0 . 001 , ω² = 0 . 0661 ) sample period and the early ( F ( 2 , 267 ) = 6 . 88 , p = 0 . 0012 , ω² = 0 . 0417 ) and late ( F ( 2 , 267 ) = 3 . 85 , p = 0 . 0225 , ω² = 0 . 0207 ) delay period ( Figure 4B ) .","There was a large and significant drop between one and two items ( post hoc Bonferroni corrected multiple comparisons: early and late sample p < 0 . 001 , early delay p < 0 . 001 , late delay p = 0 . 0198 ) and one and three items ( post hoc Bonferroni corrected multiple comparisons: early and late sample p < 0 . 001 , early delay p = 0 . 019 , late delay p > 0 . 05 ) but no difference between loads 2 and 3 ( post hoc Bonferroni corrected multiple comparisons: all p > 0 . 05 ) .","The maintenance of a significant amount of information at higher loads ( even for three items , early sample t ( 156 ) = 7 . 55 , p < 0 . 001; late sample t ( 156 ) = 8 . 73 , p < 0 . 001; early delay t ( 87 ) = 3 . 84 , p < 0 . 001; late delay t ( 87 ) = 4 . 73 , p < 0 . 001 ) and its gradual reduction when items were added to the corresponding hemifield are indicative of a flexible resource allocation and not an all-or-nothing slot-like WM .","Furthermore , if there is a flexible resource , in error trials a small but insufficient amount of resource might still be allocated to an item .","Indeed , error trial analysis ( applying correct trial sub-sampling , see Materials and methods ) for the load 2 and 3 conditions further supported this interpretation .","The amount of information in the early and late sample phase remained above zero ( load 2: early , t ( 186 ) = 3 . 25 , p = 0 . 0014; late , t ( 186 ) = 5 . 33 , p < 0 . 001; load 3: t ( 156 ) = 4 . 21 , p < 0 . 001; Figure 4B asterisks ) , and was significantly smaller than in correct trials ( load 2: late , t ( 186 ) = 2 . 81 , p = 0 . 0055 , d = 0 . 26; load 3: late t ( 156 ) = 2 . 55 , p = 0 . 0117 , d = 0 . 23 ) .","Additionally , there was no further maintenance throughout the memory delay at any load ( Figure 4B , PEV at loads 2 and 3 in error trials delay , all non-significant ) .","This indicates that a failure to report which color had changed at higher loads ( two and three ipsilateral items ) resulted from a smaller amount of information encoding during the sample phase that was not maintained throughout the delay .","A possible alternative ‘slot-model’ explanation would be that , on error trials , the color information was completely lost after the sample phase , because it was not successfully transferred into a slot ( or that a slot was not available to take on information ) .","The graded amount of information on correct trials is not compatible with the simple ( all or none ) slot model , but could fit the ‘slots and averaging model’ ( Zhang and Luck , 2008 ) .","We next wanted to understand the neuronal mechanisms behind the information loss at higher WM loads .","For that , we analyzed how the responses of individual sample and delay neurons changed when the load increased from one color to two colors .","For the ‘sample populations’ and the ‘delay populations’ , an increasing number of items reduced the amount of encoded information about the color identity ( Figure 4 ) .","This effect was due to neurons that had a large difference of firing rates between the color A and color B at load 1 ( high PEV , i . e . , information about color ) , and reduced differentiation at load 2 ( small PEV , no or little information about color , e . g . , Figure 2A ) .","‘Divisive-normalization-like regularization’ ( DNR; Carandini and Heeger , 2011 ) can explain this effect .","DNR describes the computation that takes place when two stimuli are presented simultaneously .","In a simplified case a neuronal response becomes normalized , analogous to vector normalization , with a normalization factor consisting of the simultaneous stimuli ( Carandini and Heeger , 2011 ) .","Applied to our context , a consequence of DNR would be a reduced differentiation between two color identities at load 2 because differences in firing rate ( for each stimulus by itself ) at load 1 would be normalized at load 2 ( resulting in information loss , e . g . , Figure 2A ) .","We , therefore , hypothesized that DNR was observable for neurons with significant information at load","1 . We tested for DNR in the NCL by calculating a selectivity index ( SE ) and a sensory interaction index ( SI ) for each neuron for the sample phase and the memory delay phase ( Reynolds et al . , 1999 , see Materials and methods for details ) .","SE indicates how strongly the neuronal response is driven by a color at the favorite location of the neuron ( reference ) in relation to a selected probe color ( when either is presented alone ) .","SI indicates how the probe color interacts with the reference color when both are presented simultaneously .","Values of both indices , SE and SI , lie between –1 and +1 .","The addition of a probe color influences the response to the reference color by either suppressing the firing rate of the reference color ( if the reference elicits a higher firing rate than the probe , i . e . , SE <0 ) , or increasing the firing rate for the reference color ( if the probe elicits a higher firing rate than the reference , i . e . , SE >0 ) .","If DNR was present , this influence to suppress or enhance neuronal responses should be an even mixture at the population level , resulting in a significant regression between SE and SI with a slope of around 0 . 5 ( Bouchacourt and Buschman , 2019 ) .","We compared regressions for the sample and delay phase ( each as one bin , see Materials and methods for details ) for two groups of neurons: information-carrying neurons ( significant information at load 1 ) , and non-informative neurons ( no information at load 1 or at load 2; Figure 5—figure supplement 1 ) .","We found that DNR was present in both sample and delay phases ( Figure 5A ) .","Information-carrying sample neurons had a fitted slope of 0 . 47 ( R2adj 0 . 39 , F ( 1 , 838 ) = 547 . 69 , p < 0 . 001 , CI = [0 . 43 0 . 51] ) and delay neurons had a slope of 0 . 50 ( R2adj 0 . 34 , F ( 1 , 342 ) = 175 . 60 , p < 0 . 001 , CI = [0 . 43 0 . 58] ) .","As the slopes were not significantly different from 0 . 5 , this indicates that reference and probe color had an equal influence on neuronal responses .","We thus show that DNR was observable in the neuronal population , and as a consequence of this computation , neurons had generally less information about the color identity at load","2 . Some neurons showed encoding of color identity at higher loads , instead of loss of information .","These neurons were abundant in both the sample phase and the delay phase ( Figure 3C ) .","For example , the neuron shown in Figure 2B did not differentiate between color identities at load 1 but did so for load 2 , thus , representing a case of information gain ( instead of loss ) at a higher load .","We wanted to understand if DNR , the mechanism that we found reduced color information at load 2 , could also produce color differentiation .","The ‘normalization model of attention’ ( Reynolds and Heeger , 2009 ) incorporates divisive normalization , and can explain how attention can modulate neuronal responses .","By attending a preferred ( or non-preferred ) second colored square in the load 2 condition the neuronal response of a neuron to the target location ( i . e . , to color A and to color B at the favorite location ) might be altered .","As a result a difference between color A and B may arise even though each color by itself elicited a similar response .","In other words , the interaction between the additional color and the target color is unequal .","Neurons without a color differentiation at load 1 that gained differentiation at load 2 through this process ( e . g . , if the interaction of probe color A with reference color A is larger than the interaction of probe color A with reference color B , see Figure 5—figure supplement 2 for an example ) should have a population regression slope smaller than 0 . 5 .","We thus hypothesized that the population of neurons showing information at load 2 , but not at load 1 ( e . g . , Figure 2B ) , would have a smaller slope than the neurons that lost information ( Figure 5A ) .","Sample neurons had a slope of 0 . 19 ( R2adj 0 . 05 , F ( 1 , 446 ) = 23 . 0 , p < 0 . 001 , CI = [0 . 11 0 . 27] , Figure 5B ) , and delay neurons had a slope of –0 . 04 ( R2adj –0 . 015 , F ( 1 , 62 ) = 0 . 08 , p = 0 . 78 , CI = [–0 . 29 0 . 22] , Figure 5B ) .","Both slopes were significantly smaller than 0 . 5 and smaller than the slopes of the non-informative neurons ( Figure 5—figure supplement 1 ) .","This indicates that these neurons were influenced more strongly by the reference color , and that the addition of the probe color at load 2 resulted in an unequal interaction .","Therefore , DNR was also computationally responsible for a gain of information at load 2 , in a specific subset of neurons ."],["Our results confirm behavioral findings that have been discussed in detail in an earlier study ( Balakhonov and Rose , 2017 ) .","In brief , we found that the WM capacity of crows is limited to about four items , and that the two visual hemifields are largely independent ( i . e . , the number of items on one side does not affect change detection performance on the other side ) .","Within each hemifield , performance dropped gradually with the addition of a second and third item but remained above chance .","Fittingly , on the neuronal level , we found a markedly reduced amount of color information when the number of colored squares was increased from one to two ( roughly 50% reduction in correct trials ) .","This suggests that WM could be conceptualized as a continuous resource that has to be divided between the two items ( Bays and Husain , 2008; van den Berg et al . , 2012; Wilken and Ma , 2004 ) , rather than two ‘simple’ slots that would each have the same amount of information irrespective of the memory load .","This is also consistent with results of human neuroimaging that report decreased signal amplitude and precision with increasing memory load ( Emrich et al . , 2013; Sprague et al . , 2014 ) .","In contrast , the hemispheric independence we observed would fit a slot-like model , in which the hemispheres as a whole act like discrete slots .","A more nuanced version of the slot model ( ‘slots and averaging’; Zhang and Luck , 2008 ) could also account for graded amounts of information within a limited number of slots ( Fukuda et al . , 2010; Zhang and Luck , 2008 ) , as we found here .","The mix of discrete and independent hemispheres with a graded allocation of information between items that we found is comparable to results by Buschman et al . , 2011 , observed in monkey PFC .","On the neuronal level , recurrent connections between neurons within a hemisphere may reduce item differentiation when multiple items are present simultaneously , creating capacity limitations within the hemisphere ( Matsushima and Tanaka , 2014 ) .","A lack of interhemispheric recurrent connections would make processing in the other hemisphere independent .","Like in monkeys , WM capacity in crows may therefore result from neuronal activity patterns governed by multiple individual items .","We probed the WM capacity of crows using colored squares , based on the task design of Buschman et al . , 2011 .","Using the identical task allowed us to directly compare our neuronal results of WM capacity from NCL to results from PFC of monkeys .","Task similarity is very important for such cross species comparisons as even small changes in task parameters may introduce substantial differences in neuronal responses , leading to potentially different conclusions .","In a task similar to the one used here , Lara and Wallis , 2014 , have found that neurons in the PFC of monkeys encoded nearly no information about color , but instead about location .","In their task monkeys had to memorize the color of squares at two locations on a screen , and were again confronted with a colored square at one of the two locations after a delay .","The monkeys then had to indicate if the color at that location had changed .","Lara and Wallis , 2014 , discuss the absence of color information in the neurons they recorded in relation to the task of Buschman et al . , 2011 , who like us , did find color information .","In brief , the exact task design may determine the neuronal encoding of task relevant information ( Lara and Wallis , 2014 ) .","Similar to the complex contribution of PFC neurons to WM , neurons of NCL can also encode a wide range of very different task relevant aspects , like color ( this study ) , spatial locations ( Rinnert et al . , 2019; Veit et al . , 2017 ) , and more abstract items like rules ( Veit and Nieder , 2013 ) and numerosities ( Ditz and Nieder , 2015 ) .","One way to circumvent WM failure when item load increases is to allocate attention .","Our results suggest that attention may play an important role in crow WM .","Capacity limitation became apparent during encoding , as the amount of information at the end of the sample period was affected by the stimulus load .","Adding a second and third item to the ipsilateral stimulus array reduced the amount of color information encoded by NCL neurons that carried over into the memory delay .","Furthermore , neuronal activity in trials in which the birds made an incorrect response showed only weak encoding during the sample phase without information maintenance during the memory delay .","This fits studies of human WM that have shown attentive filtering during encoding of stimuli influencing WM capacity ( Bays and Husain , 2008; Vogel et al . , 2005; Vogel and Machizawa , 2004 ) , and neuronal correlates of this have been reported for monkeys as well ( Buschman et al . , 2011 ) .","Beyond the domain of sensory signals , attention and WM may be directly linked .","Neuronal correlates of WM and attention overlap in PFC neurons , for example , Lebedev et al . , 2004 , found that a substantial amount of PFC neurons encode either an attentional signal , or a memory signal , and some ( hybrid ) neurons do both .","A purely mnemonic function of PFC thereby seems unlikely .","Indeed , very recently , Panichello and Buschman , 2021 , have reported that at the population level neurons of PFC encode ‘both the selection of items from working memory and attention to sensory inputs’ ( p . 2 ) , rather than just memory content .","The independence of hemifields that we observed on the behavioral level ( this study and Balakhonov and Rose , 2017 ) and found in the neuronal responses could also be related to attention .","Adding stimuli in the contralateral hemifield affected neither performance nor information maintained by NCL neurons , whereas additional ipsilateral stimuli strongly reduced both .","This fits the influence of attention on WM and hemifield independence , which is consistently accentuated in studies in which attention had to be divided between the two hemifields ( Alvarez and Cavanagh , 2005; Buschman et al . , 2011; Cavanagh and Alvarez , 2005; Delvenne , 2005; Delvenne et al . , 2011 ) .","Finally , the DNR computation may explain the responses of the neurons that gained information at load 2 through attentional processes predicted by the ‘normalization model of attention’ ( Reynolds and Heeger , 2009 ) .","This may appear counter-intuitive and contradictory , considering that the same process is also responsible for the loss of information .","However , when attention is overtly directed to a specific ( preferred or non-preferred ) item within the receptive field of a neuron , the DNR computation shifts its weighting of the normalized response toward the response of the attended item ( Reynolds et al . , 1999; Reynolds and Heeger , 2009 ) .","This weighted normalization can produce a difference in the neuronal response to both color identities at load 2 , even if the neuronal response was non-informative at load 1 .","At the population level we were able to observe such an effect as the reduced slope of the selectivity/interaction fit .","Thus , an attentive process might have enhanced information in WM at higher loads .","It is important to clarify that , as we did not use any form of attentional cueing in our study , we cannot explicitly test for such an attention effect .","However , we do know that the animals participating in this study can use attentional cues to enhance their WM ( Fongaro and Rose , 2020 ) .","The attention cues used by Fongaro and Rose , 2020 , positively affected not only encoding but also the maintenance and retrieval of the information held in WM , comparable to results from monkeys and humans ( Brady and Hampton , 2018; Souza and Oberauer , 2016 ) .","We , therefore , want to emphasize that our data is in line with the interpretation that the birds possibly attended a load 2 stimulus array differently than a load 1 stimulus array in order to enhance their performance in trials with higher loads .","Our neuronal recordings offer a mechanistic explanation for the behavioral effects , as we found clear evidence of DNR governing the neuronal responses tied to WM capacity that is in accordance with mammalian models of WM capacity ( Bouchacourt and Buschman , 2019; Lundqvist et al . , 2016; Lundqvist et al . , 2011 ) .","The loss of information about color identity ( i . e . , neuronal response differentiation between colors ) can be accounted for by DNR when an item is added to a neuron’s receptive field .","The normalization of neuronal firing rate diminishes the differentiation between color identities .","As such it is analogous to neurophysiological responses from visual areas ( Carandini et al . , 1997; Reynolds et al . , 1999 ) and to the PFC during spatial WM ( Matsushima and Tanaka , 2014 ) .","The WM model of Bouchacourt and Buschman , 2019 , is based solely on data from monkey electrophysiology , and thus implicitly tied to the layered columns of the neocortex .","The results we report here show that the model also fits the neurophysiology of WM in crows .","However , the picture is incomplete since important aspects of monkeys’ WM are still not investigated in crows .","Oscillations of local field potentials are relevant for how information enters WM and how it is maintained ( Miller et al . , 2018 ) , and have been tied to normalization and competition between items in WM ( Lundqvist et al . , 2018b ) .","Thus , the oscillatory interplay of the layers and different regions of the mammalian neocortex are important fields of research to further our understanding of WM .","Such aspects are so far completely unknown in crows and their non-layered associative areas .","This encourages further investigation into the neuronal circuits of WM in birds .","There is also ongoing debate about the role of sustained activity during delay periods and how it relates to WM ( Constantinidis et al . , 2018; Lundqvist et al . , 2018a; Miller et al . , 2018 ) .","We cannot report of any neuron that showed persistent activity comparable to those reported by classical WM studies in PFC ( e . g . , Funahashi et al . , 1989; Fuster and Alexander , 1971 ) , or in NCL ( Diekamp et al . , 2002b; Veit et al . , 2014; Veit and Nieder , 2013 ) .","This may be reconcilable with some other contemporary models of WM .","One major type of those models implements ‘synfire chains’ , where individual neurons fire sequentially ( and transiently ) to bridge temporal gaps and maintain task relevant contents ( Rajan et al . , 2016 ) .","This has , for example , been reported to be the case in posterior parietal cortex of mice performing a T-maze task that required WM for cued spatial locations to be maintained ( Harvey et al . , 2012 ) .","The transient activity of neurons that we report ( Figures 2 and 3A ) might fit into such models .","However , our results can only be compared very cautiously to this ( since even small changes in task design significantly alter neuronal responses , e . g . , Lara and Wallis , 2014 ) .","Therefore , while we cannot , yet , fully equate crow and monkey WM , our results raise two important questions about how WM is implemented on the level of neuronal networks that have implications for our comparative view of crow WM .","The first regards the neuronal computations underlying WM .","Is there a common canonical computation governing WM , or are there different solutions based on different neuronal architectures ?","Recent work has shown that the sensory areas of mammals and birds show remarkably similar circuit organization ( Stacho et al . , 2020 ) .","However , higher-order associative areas involved in WM , like the LPFC in mammals and the NCL in birds , have distinctly different architectures ( Stacho et al . , 2020 ) .","The fact that differently organized areas like LPFC and NCL produce strikingly similar physiological responses points to shared computational principles .","Modeling work already suggests that the competing WM capacity models can be accommodated into a unifying framework based on theoretical neuronal information sampling , where stochastic information sampling ( assumed for continuous resource models ) can account for item limitations better than fixed information sampling ( assumed by the slots and averaging models ) ( Schneegans et al . , 2020 ) .","Similarly , DNR is already considered to be a general , canonical computation of the nervous system , present in evolutionarily distant phyla , for example , fruit flies and monkeys ( Carandini and Heeger , 2011 ) .","The second question regards the tradeoff between WM flexibility and capacity ( Bouchacourt and Buschman , 2019 ) .","Is the WM of a crow as flexible as that of a monkey ?","Our results show that the computations by individual neurons that result in WM capacity limitations are virtually the same in crows and monkeys , highlighting a further aspect of WM that is similar between these animal groups ( Nieder , 2017 ) .","Ultimately , our results were in line with different modern models of WM that implement DNR to explain capacity ( Bouchacourt and Buschman , 2019; Lundqvist et al . , 2016; Lundqvist et al . , 2011; Schneegans et al . , 2020 ) .","However , the data we presented cannot carry a definitive conclusion about which of the different models fits best .","For example , a tradeoff between flexibility and capacity ( Bouchacourt and Buschman , 2019 ) might be present , but further investigation into the models’ predictions is required .","We do , however , show that mammalian models of WM are in line with WM in birds , which implies that fundamental aspects of WM are shared between these animal groups .","Together , all these facets of crow WM capacity suggest that the different intricate neuronal architectures that carry out the computations in monkeys and crows have likely been shaped by convergent evolution – into systems that yield similar cognitive performances .","The systems may share the same basic mechanisms and thus limitations .","Further investigation into the oscillatory dynamics of WM in the avian brain may elucidate if birds also share the prominent limitation of a tradeoff between flexibility and capacity ."],["Two hand-raised carrion crows ( C . corone ) of 2 years of age served as subjects in this study .","The birds were housed in spacious aviaries in social groups .","During the experimental procedures , the animals were held on a controlled food protocol with ad libitum access to water and grit .","All experimental procedures and housing conditions were carried out in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals and were authorized by the national authority ( LANUV ) .","We used operant training chambers ( 50 × 50 . 5 × 77 . 5 cm3 , width × depth × height ) equipped with an acoustic pulse touchscreen ( 22’’ , ELO 2200L APR , Elo Touch Solutions Inc , Milpitas , CA ) and an infrared camera ( Sygonix , Nürnberg , Germany ) for remote monitoring .","The birds sat on a wooden perch so that the distance between the bird’s eye and the touchscreen was 8 cm .","Food pellets were delivered as a reward via a custom-made automatic feeder ( plans available at http://www . jonasrose . net/ ) .","The position of the animal’s head was tracked online during the experiment by two open-source computer vision cameras ( ‘Pixy’ , CMUcam5 , Charmed Labs , Austin , TX ) that reported the location and angle between two LEDs .","For tracking , we surgically implanted a lightweight head-post and used a lightweight 3D-printed mount with LEDs that was removed after each experimental session .","The system reported the head location at a frame rate of 50 Hz and data was smoothed by integrating over two frames in MATLAB using custom programs on a control PC .","All experiments were controlled by custom programs in MATLAB using the Biopsychology ( Rose et al . , 2008 ) and Psychophysics toolboxes ( Brainard , 1997 ) .","Digital input and output of the control PC were handled by a microcontroller ( ODROID C1 , Hardkernel co . Ltd , Anyang , South Korea ) connected through a gigabit network running custom software ( available at http://www . jonasrose . net/ ) .","The behavioral protocol was identical to the one described in Balakhonov and Rose , 2017 .","We trained the birds to perform a delayed change localization task that had previously been used to test the performance under different WM loads in primates ( Buschman et al . , 2011 ) .","Each trial started after a 2 s inter-trial-interval , with the presentation of a red dot centered on the touchscreen ( for a maximum of 40 s ) .","The animals initiated the trial by centering their head in front of the red dot for 160 ms . This caused the red dot to disappear and a stimulus array of two to five colored squares to appear ( Figure 1A , ‘sample’ ) .","The colored squares were presented for a period of 800 ms , during which the animals had to hold their head still and center their gaze on the screen ( ‘hold gaze’ , no more than 2 cm horizontal or vertical displacement , and no more than 20 degrees horizontal or vertical rotation ) .","Failure to hold the head in this position resulted in an aborted trial .","This sample phase was followed by a memory delay of 1000 ms after which the stimulus array reappeared with one color exchanged .","The animal had to indicate the location of the color change by pecking the respective square .","Correct responses were rewarded probabilistically ( BEO special pellets , in 55% of correct trials , additional 2 s illumination of the food receptacle in 100% of correct trials ) .","Incorrect responses to colors that had not changed or a failure to respond within 4 s resulted in a brief screen flash and a 10 s timeout .","The stimuli were presented at six fixed locations on the screen ( 1–6 , Figure 1A ) .","In each session , one pair of colors was assigned to each of the six locations .","Each location had its own distinct pair .","These pairs were randomly chosen from a pool of 14 colors ( two color combinations were excluded since the animals did not discriminate them equally well during a pre-training ) .","Let us consider Figure 1A as an example .","The color change occurs in the middle-left where turquois ( T ) is presented during the sample and orange ( O ) during the choice .","In this particular session the middle-left could thus show either of the following colors during the sample and choice: T-O ( shown in Figure 1A ) ; O-T; O-O; T-T; None-None .","On the next session a new random pair of colors was displayed at this location .","For identification and analysis an arbitrary label was assigned to each of the randomly drawn colors at the start of each session ( i . e . , left middle location: ‘color A’ , or ‘color B’; left top location: ‘color A’ , or ‘color B’; etc . ) .","These indices do not refer to the order of presentation of the colors at any time , but were held constant for neuronal analysis .","The order of presentation of colors within a pair , the target location ( where the color change occurred ) , and the number of stimuli in the array ( two to five ) were randomized and balanced across trials so that each condition had an equal likelihood to appear .","The order of presentation of colors within a pair , the target location ( where the color change occurred ) , and the number of stimuli in the array ( two to five ) were randomized and balanced across trials so that each condition had an equal likelihood to appear .","The color squares had a width of 10 degrees of visual angle ( DVA ) and were placed on the horizontal meridian of the screen and at 45 . 8 DVA above or below the meridian at a distance of 54 and 55 . 4 DVA from the center .","This arrangement in combination with the head tracking ensured that all stimuli appeared outside of the binocular visual field of crows ( 37 . 6 DVA; Troscianko et al . , 2012 ) .","Both animals were chronically implanted with a lightweight head-post to attach a small LED holder during the experiments .","Before surgery , animals were deeply anesthetized with ketamine ( 50 mg/kg ) and xylazine ( 5 mg/kg ) .","Once deeply anesthetized , animals were placed in a stereotaxic frame .","After attaching the small head-post with dental acrylic , a microdrive with a multi-channel microelectrode was stereotactically implanted at the craniotomy ( Neuronexus Technologies Inc , Ann Arbor MI , DDrive ) .","The electrode was positioned in NCL ( AP 5 . 0 , ML 13 . 0 ) of the left hemisphere ( coordinates for the region based on histological studies on the localization of NCL in crows; Veit and Nieder , 2013 ) .","After the surgery , the crows received analgesics .","Extracellular single neuron recordings were performed using chronically implanted multi-channel microelectrodes .","The distance between recording sites was 50 µm .","The signal was amplified , filtered , and digitized using Intan RHD2000 headstages and a USB-Interface board ( Intan Technologies LLC , Los Angeles , CA ) .","The system also recorded digital event codes that were sent from the behavioral control PC using a custom IO device ( details available at http://www . jonasrose . net/ ) .","Before each recording session , the electrodes were advanced manually using the microdrive .","Recordings were started 20 min after the advancement , and each recording site was manually checked for neuronal signals .","The signals were recorded at a sampling rate of 30 kHz and filtered with a band-pass filter at recording ( 0 . 5–7 . 5 kHz ) .","The recorded neuronal signals were not pre-selected for task involvement .","We performed spike sorting using the semi-automatic Klusta-suite software ( Rossant et al . , 2016 ) , which uses the high electrode count and their close spacing to isolate signals of single neurons .","For spike sorting , we filtered with a high pass of 500 Hz and a low pass of 7125 Hz .","The software utilizes the spatial distribution of the recorded signal along the different recording sites to untangle overlapping signals and separate signals with similar waveforms but different recording depths .","All statistical analyses were performed in MATLAB ( 2018b , Mathworks Inc ) using commercially available toolboxes ( Curve Fitting Toolbox Version 3 . 5 . 3 , Statistics and Machine Learning Toolbox Version 10 . 2 ) and custom code .","For all statistical tests , we assumed a significance level of α = 0 . 05 , unless stated otherwise .","Trials were classified as error trials if the bird chose a location where no change of colors had appeared .","Trials in which the bird did not choose any location or failed to maintain head fixation were not analyzed .","All correct trials were included in the analysis of neural data .","Depending on the analysis we refer to different ‘load conditions’ relative to referential sides of the screen which are either ipsilateral ( same side ) or contralateral ( opposite side ) , each with a possible load between one and three items .","For the behavioral analyses the terms ipsilateral and contralateral refer to the location of the change that had to be detected .","For the neurophysiological analyses the terms ipsilateral and contralateral refer to the respective neuron’s favorite location ( described in the following section ) .","Because there were only very few error trials in the load one condition , we performed error trial analysis only for the load 2 and load 3 conditions .","The behavioral data were analyzed as described in our previous study ( Balakhonov and Rose , 2017 ) , estimating the WM capacity K for each load by Equation 1 . ( 1 ) K=n*p where p is the percentage correct and n is the number of items in WM .","This estimate has been applied to similar primate data and in studies with humans ( Johnson et al . , 2013; Kornblith et al . , 2016 ) .","Based on a one-way ANOVA of color identity at a given location , we calculated a PEV statistic to measure the effect size of neuronal modulation .","Its main parameter ω² is a measurement for the percentage to which the tested factor can explain the variance of the data , and it is calculated from the sum of squares of the effect ( SSeffect ) and the mean squares of the within-group ( error ) variance ( MSerror ) ( Equation 2 ) .","( 2 ) ω2=SSeffect-df*MSerrorSStotal+MSerror For each neuron , we determined a ‘favorite location’ , which was defined as the location with the highest cumulative PEV , of the three possible locations on the right half of the screen , that is , opposite to the implanted hemisphere , across four non-overlapping bins during the sample phase ( bin size 200 ms , advanced in steps of 200 ms , from start till the end of the sample phase ) .","The significance of calculated effect size values was determined by a permutation test .","We ran the permutation to calculate the likelihood of getting an explained variance value bigger than the one calculated from the actual distribution of the data by randomly permuting the color identity labels and calculating the PEV 1000 times .","The test thereby does not assume any distribution of the data and returns an unbiased estimate of the likelihood of generating an effect size within the data randomly .","The measured value of explained variance from the actual dataset was assumed to be significant if the likelihood of randomly generating a bigger value was below 5% .","We chose to not correct for the multiple comparisons at this level , as we reasoned that if we were to only include those neurons that had the most information at the individual loads ( i . e . , those with the highest PEV values that are significant even under very stringent statistical criteria ) we would have artificially inflated the amount of information present at each load .","By including neurons that encoded less information too ( i . e . , at the uncorrected p-value ) our analysis population was more resistant to such outlier effects .","We additionally performed our analyses using a more stringent statistical criterion ( significance of two consecutive non-overlapping bins ) and found the same results ( Figure 4—figure supplement 2 ) .","We tested the proportions of significant neurons we found for the different trial phases by performing a binomial test , assuming a significance level α = 0 . 05 ( Equation 3 ) .","( 3 ) PX=i=Bp0 , n=nkp0i1-p0n-i Calculating the probability p , of finding X significant neurons , given a total amount of i ( 362 ) neurons , and a probability p0 of 5% finding a significance by chance .","We considered neuronal significance ( i . e . , significant PEV as determined above ) for each load independently .","This means , we tested if the PEV of a neuron was significant three times with the permutation method described above: once for each of the three load conditions .","Therefore , we can report seven groups of significance ( Table 1 , Figure 3C ) .","Subsequently , we created three pooled groups ( Table","1 ) from all neurons with a significant PEV at each individual load .","We used these pooled groups for the population analyses ( Figures 4 and 5 ) .","Neurons of these pooled groups , with a significant PEV during the sample phase were assigned to the ‘sample population’ , and neurons with a significant amount of information during the memory-delay phase were assigned to the ‘delay population’ ( significance criterion: one significant 200 ms bin , at α = 0 . 05 , see above for our reasoning not to correct for multiple comparison at this point ) .","Thus , neurons with significant PEV during both the sample and delay phase were included in both subpopulations .","We corrected for the unequal amount of correct and error trials when comparing information about color ( PEV ) between the trial conditions , by sub-sampling correct trials with the number of error trials 1000 times for each neuron .","The resulting PEV values of correct trials were then averaged for each neuron , this population of averaged PEV values was then statistically tested against the PEV values of error trials ( of the same neurons ) using a dependent t-test .","We tested for the presence of divisive normalization using the method of Reynolds et al . , 1999 .","Three conditions were considered: ( 1 ) neuronal response to stimulus A , ( 2 ) neuronal response to stimulus B , and ( 3 ) neuronal response to the simultaneity of stimuli A and B . As we wanted to relate this to the information about color identity , we selected subsets of the favorite location and the additional two ipsilateral locations .","To test how the neurons altered their response when multiple stimuli were presented simultaneously , we calculated the color selectivity index ( SE ) and the sensory interaction index ( SI ) of each neuron .","SEi was calculated by subtracting the normalized firing rate for the chosen reference color i ( REFi ) at the neuron’s favorite location , from a second color j ( PROBEj ) at a different location ( ipsilateral to the favorite location , Equation 4 ) .","( 4 ) SEi=PROBEj-REFi The resulting selectivity index lies between –1 ( completely selective for the reference color ) and 1 ( completely selective for the probe color ) .","SI was calculated ( Equation 5 ) by subtracting the normalized firing rate for REFi from the normalized firing rate of the combination of REFi and PROBEj ( PAIRi , j ) .","( 5 ) SIi , j=PAIRi , j-REFi This interaction index also lies between –1 ( full suppression of reference stimulus by the probe stimulus ) and 1 ( full enhancement of the reference stimulus by the probe stimulus ) .","As each of the three locations had two possible colors , we calculated eight SE and SI indices per neuron and performed a linear regression for all indices .","This is required as each stimulus combination is informative about the normalization .","The effects of divisive normalization were compared between the sample and the delay phase .","Therefore , SE and SI indices were calculated across the entire sample ( 800 ms ) and memory delay ( 1000 ms ) phase .","Neurons with significant information were accordingly identified over the entire sample and delay as one bin , using the permutation test described in the section ‘information about color identity’ .","We considered the entire sample and delay phase because we wanted to analyze the population response as a whole , irrespective of highly diverse response profiles of individual neurons .","To visualize the different groups of neurons that encoded and maintained information about the color identity during different phases of the trial , we performed a hierarchical clustering analysis in MATLAB on the normalized PEV values of individual neurons throughout the trial .","We used a ( 1 − correlation ) distance metric and an average distance linkage function for a maximum of seven clusters .","The maximum number of clusters was first determined by calculating the clustering for different amounts of clusters ( 1–10 ) and subsequently calculating the within-cluster sum-of-squares .","This resulted in a graph that allowed us to visually inspect the tradeoff between cluster number and fit improvement , from which we estimated the inflection point ( elbow method ) .","A cluster number of seven presented the best tradeoff that allowed visualization of the different groups at an acceptable clustering success .","We then ordered the neuron clusters to minimize the average distance between the clusters in the dendrogram ."]],"string":"[\n [\n \"Working memory ( WM ) can hold information for a short period of time to allow further processing in the absence of sensory input ( Cowan , 2017; Oberauer et al . , 2018 ) .\",\n \"By bridging this gap between the immediate sensory environment and behavior , WM is a cornerstone for complex cognition .\",\n \"It is a very flexible memory system , yet severely limited in its capacity .\",\n \"While this capacity is often seen as a general cognitive bottleneck , for simple stimuli , like colors , the capacity is very similar between humans , monkeys , and crows ( Balakhonov and Rose , 2017; Buschman et al . , 2011; Cowan , 2001; Luck and Vogel , 1997 ) .\",\n \"Different models have been proposed to conceptualize how this capacity limit arises .\",\n \"This work motivated many psychophysical and electrophysiological experiments that in turn led to a spectrum of more refined models of WM ( Ma et al . , 2014 ) .\",\n \"‘Discrete models’ of WM argue that a fixed number of items can be stored .\",\n \"Once this capacity is reached , an additional item can only be maintained if it replaces a previous item ( Awh et al . , 2007; Fukuda et al . , 2010; Luck and Vogel , 1997; Vogel and Machizawa , 2004; Zhang and Luck , 2008 ) .\",\n \"‘Continuous models’ describe WM as a flexible resource that is allocated to individual items .\",\n \"A minimum amount of this resource has to be allocated to each item for successful retention , thereby resulting in a capacity limit ( Bays and Husain , 2008; van den Berg et al . , 2012; Wilken and Ma , 2004 ) .\",\n \"On the neurophysiological level , models of WM capacity suggest that interference between memory representations ( ‘items’ ) within the neuronal network is a source of information loss and capacity limitation ( Bouchacourt and Buschman , 2019; Lundqvist et al . , 2016; Lundqvist et al . , 2011; Schneegans et al . , 2020 ) .\",\n \"Interference may arise due to divisive normalization that appears as competition between items , related to oscillatory dynamics ( Lundqvist et al . , 2016; Lundqvist et al . , 2011 ) , WM flexibility ( Bouchacourt and Buschman , 2019 ) , and neuronal information sampling ( Schneegans et al . , 2020 ) .\",\n \"Divisive normalization is a computational principle that acts upon neurons when presenting multiple stimuli simultaneously , it normalizes neuronal responses by creating ‘a ratio between the response of an individual neuron and the summed activity of a pool of neurons’ ( Carandini and Heeger , 2011 , p . 51 ) .\",\n \"An effect related to divisive normalization can be observed when two stimuli are presented either individually or simultaneously within the receptive field of a visual sensory neuron .\",\n \"The neuron’s firing rate when the stimuli are presented simultaneously becomes normalized by the populations’ responses to each individual stimulus ( Carandini et al . , 1997; Heeger , 1992 ) .\",\n \"This effect also occurs in relation to attentive processes and has been formalized into the ‘normalization model of attention’ ( Reynolds et al . , 1999; Reynolds and Heeger , 2009 ) .\",\n \"Normalization of neuronal responses is commonly observed in many species throughout the animal kingdom , not just in sensory , but also in cognitive domains ( Carandini and Heeger , 2011 ) .\",\n \"Investigations into WM capacity and model predictions focus mostly on humans and monkeys .\",\n \"By extending this work to include birds , one can gain a unique comparative perspective .\",\n \"Crows have a similar limit in WM capacity and neuronal correlates of WM are comparable to monkeys’ ( Balakhonov and Rose , 2017; Nieder , 2017 ) .\",\n \"But while the neuronal architecture of sensory areas is similar between birds and mammals , higher associative areas , critical for WM , do not share a common architecture between the species ( Stacho et al . , 2020 ) .\",\n \"Therefore , an outstanding question is whether modern models of WM such as the ‘flexible model’ capture WM capacity in general , or if their predictions ( e . g . , divisive normalization ) are confined to the mammalian neocortex .\",\n \"To resolve this , it is crucial to investigate the avian brain to understand how its different organization can produce such similar behavioral and neurophysiological results .\",\n \"While the neuronal correlates of WM maintenance in birds have been investigated in some detail ( Diekamp et al . , 2002a; Hartmann et al . , 2018; Rinnert et al . , 2019; Rose and Colombo , 2005; Veit et al . , 2014 ) , a neurophysiological investigation of WM capacity limitation is still lacking .\",\n \"The avian forebrain structure , nidopallium caudolaterale ( NCL ) is a critical component of avian WM .\",\n \"The NCL is considered functionally equivalent to the mammalian prefrontal cortex ( PFC ) ( Güntürkün and Bugnyar , 2016; Nieder , 2017 ) , as it receives projections from all sensory modalities ( Kröner and Güntürkün , 1999 ) , projects to premotor areas ( Kröner and Güntürkün , 1999 ) , and is a target of dopaminergic innervation ( Waldmann and Güntürkün , 1993 ) .\",\n \"To investigate the neurophysiology of WM capacity in birds , we adopted a task design developed for monkeys ( Buschman et al . , 2011 ) to use it with carrion crows ( Corvus corone ) .\",\n \"Our animals were trained to memorize an array of colors and to indicate which color had changed after a short memory delay , while we performed extracellular recordings of individual neurons in the NCL using multi-channel probes .\",\n \"We expected to find a clear correlate of WM representations in NCL neurons and a load-dependent response modulation based on divisive normalization of neuronal responses .\",\n \"This would allow us to evaluate if the behavioral WM capacity observations of crows fit a ‘discrete’ or ‘continuous’ WM resource model .\",\n \"If the neuronal responses also fit the contemporary neurophysiological models of WM capacity limitations ( Bouchacourt and Buschman , 2019; Lundqvist et al . , 2016; Lundqvist et al . , 2011; Schneegans et al . , 2020 ) , it would further suggest that crows and monkeys have convergently evolved a similar neurophysiological basis for WM capacity despite a different architecture of the critical forebrain structures .\"\n ],\n [\n \"The behavioral performance was influenced by the number of colored squares on the screen .\",\n \"It significantly decreased with an increasing number of ipsilateral squares ( median performances , load 1: 95 . 88% , load 2: 78 . 31% , load 3: 58 . 21%; Friedman test: Χ² = 92 . 00 , p < 0 . 001 , Figure 1B ) .\",\n \"We ran a generalized linear model with ipsilateral load ( i . e . , load of hemifield where a color changed ) , contralateral load ( i . e . , load of hemifield without a color change ) and their interaction as predictors for performance ( R2adj 0 . 78 , F ( 460 , 456 ) = 555 . 00 , p < 0 . 001 ) .\",\n \"We found that the number of ipsilateral colors significantly reduced performance ( βipsi –0 . 177 , t ( 459 ) = –18 . 00 , p < 0 . 001 ) , whereas the number of contralateral colors did not ( βcontra –0 . 021 , t ( 459 ) = –1 . 77 , p = 0 . 0772; Figure 1C ) .\",\n \"There was also a significant interaction between ipsilateral and contralateral load ( β = –0 . 024 , t ( 458 ) = –4 . 28 , p < 0 . 001 ) .\",\n \"We compared this model to a reduced model , where we omitted the non-significant βcontra and found that this reduced model ( R2adj 0 . 78 , F ( 460 , 457 ) = 828 . 00 , p < 0 . 001 ) explained the performance equally as well ( |ΔLLR| = 0 . 0102 ) .\",\n \"Therefore , we conclude that contralateral load by itself did not significantly affect performance .\",\n \"We calculated the capacity K ( see Materials and methods ) for all full WM loads ( i . e . , two to five items ) .\",\n \"The capacity K peaks at four items ( mean ± SEM: 3 . 05± . 038 , Figure 1—figure supplement 1 ) .\",\n \"These observations are very similar to observations made in primates ( Buschman et al . , 2011 ) and fully reproduce our earlier behavioral findings ( Balakhonov and Rose , 2017 ) .\",\n \"We recorded 362 neurons from the NCL of two crows performing the WM task ( delayed change localization ) .\",\n \"All reported effects were also present in each individual bird ( Figure 3—figure supplements 1–2 ) , we , therefore , pooled the data for population analysis .\",\n \"A large subset of neurons responded to the presence of a color ( i . e . , at load 1 ) by substantially increasing or decreasing their firing rate relative to baseline .\",\n \"This change in firing rate occurred selectively , depending on the presented color either in the sample ( Figure 2A ) or the delay period ( Figure 2—figure supplement 1 ) .\",\n \"For most neurons , this difference in firing rate between the two possible colors became attenuated when the load increased from one to two colors , and it was further attenuated from two to three colors .\",\n \"To quantify this effect , we calculated the amount of information about the color identity at a neuron's favorite location as the percent explained variance ( PEV ) during a memory load of one , two , or three items in bins of 200 ms ( see Materials and methods for details ) .\",\n \"Most neurons did not sustain information about color ( measured as a significant PEV , henceforth ‘information’ ) throughout the entire sample or memory delay but rather had shorter periods in which the information was significant ( Figure 2A bottom ) .\",\n \"To better capture the time points when the individual neurons carried information , we performed a hierarchical clustering analysis of the PEV values of the individual neurons at load 1 ( see Materials and methods for details ) .\",\n \"We found a total of seven clusters that were organized into two overarching groups ( Figure 3A ) .\",\n \"Group 1 contained neurons ( n = 227 ) that showed peak information during the sample and early delay phase , while group 2 contained neurons ( n = 135 ) that showed peak information during the delay phase .\",\n \"For each neuron , we then calculated if it carried a significant amount of color information by applying a permutation test ( for all bins at load 1 , see Materials and methods ) .\",\n \"The individual neurons were then further classified into three groups depending on the phase in which they had a significant amount of information ( Figure 3B ) .\",\n \"Overall , 37 . 57% ( n = 136 ) of neurons were significant during the sample phase , 9 . 39% ( n = 34 ) of neurons were significant during the memory delay , and 14 . 64% ( n = 53 ) of neurons were significant during both the sample phase and the memory delay ( all proportions of neurons were significantly higher than expected by chance [binomial test , see Materials and methods , all p < 0 . 001] ) .\",\n \"Refer to Figure 2A for an example neuron , significant at load 1 with a large differentiation in firing rate between color identities ( a large PEV ) and a loss of differentiation with increasing ipsilateral load .\",\n \"Further inspection of individual neuronal activity revealed , however , that a substantial number of neurons responded differently .\",\n \"Instead of losing information at higher loads , many neurons gained information ( e . g . , Figure 2B , Figure 2—figure supplement 2 ) .\",\n \"Thus , we additionally performed the permutation testing for loads 2 and 3 to determine which neurons had significant information ( see Materials and methods ) .\",\n \"We found that many of the neurons that did not have significant information at load 1 did have significant information at load 2 and load 3 ( Figure 3C ) .\",\n \"For the memory delay , more than half of the significant neurons we detected were only significant for either load 2 or load 3 , compared to only 36% of neurons that were significant at load 1 ( Figure 3C middle ) .\",\n \"By including the higher loads in our analysis , we found a total of 249 ( 68 . 78% ) sample neurons and 94 ( 25 . 97% ) delay neurons .\",\n \"For the population analyses , we subsequently pooled all significant neurons into three groups ( one per load ) .\",\n \"These pooled groups were then each subdivided into sample and delay neurons ( i . e . , ‘sample-load1’ , ‘delay-load1’ , ‘sample-load2’ , etc . , see Table 1 in the Materials and methods for an overview ) .\",\n \"The clustering analysis indicated that the population of neurons as a whole did sustain the color information throughout the entire trial ( Figure 3A ) .\",\n \"Plotting the information averages of each of the three ‘sample populations’ and the ‘delay populations’ over time confirmed this result ( Figure 4A , Figure 4—figure supplement 1 ) .\",\n \"After the onset of the stimulus array , the average information exhibited a sharp increase that peaked roughly 400 ms after stimulus onset and remained at an elevated level throughout the memory delay , until the choice array appeared .\",\n \"Results obtained from neurons of the lateral PFC of monkeys indicated distinct hemispheric independence of WM capacity ( Buschman et al . , 2011 ) .\",\n \"This means that increasing ipsilateral load ( i . e . , load in the hemifield containing the target for which information is assessed ) should affect neuronal processing while increasing contralateral load should not .\",\n \"This effect might be further emphasized in birds due to the full decussation of their optic nerve ( Husband and Shimizu , 2001 ) .\",\n \"Parallel to the behavioral results and in line with the results from monkeys , we found a strong effect of ipsilateral load on the information maintenance , as there was a sharp drop in information when the load increased from one item to two items ( Figure 4A , blue and yellow curves ) .\",\n \"The addition of a third item only slightly decreased the maintained information further ( Figure 4A , red curve ) .\",\n \"The load dependence was much more pronounced during the sample period than during the memory delay where the information remained at a lower elevated level .\",\n \"Notably , the load effect was only present for ipsilateral manipulations .\",\n \"If the number of items on the contralateral side was increased , the information encoded about the colors at the favorite location did not change ( Figure 4A , right ) .\",\n \"To compare our results to the results obtained in monkeys we also applied the method of Buschman et al . , 2011 for testing the ipsilateral load effect during the sample and delay phase , by splitting each phase into an early and a late portion ( first and second 400 ms of the sample , and first and second 500 ms for the delay ) .\",\n \"We did find a significant drop in information with an increase in load from one through three in the early ( F ( 2 , 537 ) = 18 . 73 , p < 0 . 001 , ω² = 0 . 0616 ) and late ( F ( 2 , 536 ) = 20 . 07 , p < 0 . 001 , ω² = 0 . 0661 ) sample period and the early ( F ( 2 , 267 ) = 6 . 88 , p = 0 . 0012 , ω² = 0 . 0417 ) and late ( F ( 2 , 267 ) = 3 . 85 , p = 0 . 0225 , ω² = 0 . 0207 ) delay period ( Figure 4B ) .\",\n \"There was a large and significant drop between one and two items ( post hoc Bonferroni corrected multiple comparisons: early and late sample p < 0 . 001 , early delay p < 0 . 001 , late delay p = 0 . 0198 ) and one and three items ( post hoc Bonferroni corrected multiple comparisons: early and late sample p < 0 . 001 , early delay p = 0 . 019 , late delay p > 0 . 05 ) but no difference between loads 2 and 3 ( post hoc Bonferroni corrected multiple comparisons: all p > 0 . 05 ) .\",\n \"The maintenance of a significant amount of information at higher loads ( even for three items , early sample t ( 156 ) = 7 . 55 , p < 0 . 001; late sample t ( 156 ) = 8 . 73 , p < 0 . 001; early delay t ( 87 ) = 3 . 84 , p < 0 . 001; late delay t ( 87 ) = 4 . 73 , p < 0 . 001 ) and its gradual reduction when items were added to the corresponding hemifield are indicative of a flexible resource allocation and not an all-or-nothing slot-like WM .\",\n \"Furthermore , if there is a flexible resource , in error trials a small but insufficient amount of resource might still be allocated to an item .\",\n \"Indeed , error trial analysis ( applying correct trial sub-sampling , see Materials and methods ) for the load 2 and 3 conditions further supported this interpretation .\",\n \"The amount of information in the early and late sample phase remained above zero ( load 2: early , t ( 186 ) = 3 . 25 , p = 0 . 0014; late , t ( 186 ) = 5 . 33 , p < 0 . 001; load 3: t ( 156 ) = 4 . 21 , p < 0 . 001; Figure 4B asterisks ) , and was significantly smaller than in correct trials ( load 2: late , t ( 186 ) = 2 . 81 , p = 0 . 0055 , d = 0 . 26; load 3: late t ( 156 ) = 2 . 55 , p = 0 . 0117 , d = 0 . 23 ) .\",\n \"Additionally , there was no further maintenance throughout the memory delay at any load ( Figure 4B , PEV at loads 2 and 3 in error trials delay , all non-significant ) .\",\n \"This indicates that a failure to report which color had changed at higher loads ( two and three ipsilateral items ) resulted from a smaller amount of information encoding during the sample phase that was not maintained throughout the delay .\",\n \"A possible alternative ‘slot-model’ explanation would be that , on error trials , the color information was completely lost after the sample phase , because it was not successfully transferred into a slot ( or that a slot was not available to take on information ) .\",\n \"The graded amount of information on correct trials is not compatible with the simple ( all or none ) slot model , but could fit the ‘slots and averaging model’ ( Zhang and Luck , 2008 ) .\",\n \"We next wanted to understand the neuronal mechanisms behind the information loss at higher WM loads .\",\n \"For that , we analyzed how the responses of individual sample and delay neurons changed when the load increased from one color to two colors .\",\n \"For the ‘sample populations’ and the ‘delay populations’ , an increasing number of items reduced the amount of encoded information about the color identity ( Figure 4 ) .\",\n \"This effect was due to neurons that had a large difference of firing rates between the color A and color B at load 1 ( high PEV , i . e . , information about color ) , and reduced differentiation at load 2 ( small PEV , no or little information about color , e . g . , Figure 2A ) .\",\n \"‘Divisive-normalization-like regularization’ ( DNR; Carandini and Heeger , 2011 ) can explain this effect .\",\n \"DNR describes the computation that takes place when two stimuli are presented simultaneously .\",\n \"In a simplified case a neuronal response becomes normalized , analogous to vector normalization , with a normalization factor consisting of the simultaneous stimuli ( Carandini and Heeger , 2011 ) .\",\n \"Applied to our context , a consequence of DNR would be a reduced differentiation between two color identities at load 2 because differences in firing rate ( for each stimulus by itself ) at load 1 would be normalized at load 2 ( resulting in information loss , e . g . , Figure 2A ) .\",\n \"We , therefore , hypothesized that DNR was observable for neurons with significant information at load\",\n \"1 . We tested for DNR in the NCL by calculating a selectivity index ( SE ) and a sensory interaction index ( SI ) for each neuron for the sample phase and the memory delay phase ( Reynolds et al . , 1999 , see Materials and methods for details ) .\",\n \"SE indicates how strongly the neuronal response is driven by a color at the favorite location of the neuron ( reference ) in relation to a selected probe color ( when either is presented alone ) .\",\n \"SI indicates how the probe color interacts with the reference color when both are presented simultaneously .\",\n \"Values of both indices , SE and SI , lie between –1 and +1 .\",\n \"The addition of a probe color influences the response to the reference color by either suppressing the firing rate of the reference color ( if the reference elicits a higher firing rate than the probe , i . e . , SE <0 ) , or increasing the firing rate for the reference color ( if the probe elicits a higher firing rate than the reference , i . e . , SE >0 ) .\",\n \"If DNR was present , this influence to suppress or enhance neuronal responses should be an even mixture at the population level , resulting in a significant regression between SE and SI with a slope of around 0 . 5 ( Bouchacourt and Buschman , 2019 ) .\",\n \"We compared regressions for the sample and delay phase ( each as one bin , see Materials and methods for details ) for two groups of neurons: information-carrying neurons ( significant information at load 1 ) , and non-informative neurons ( no information at load 1 or at load 2; Figure 5—figure supplement 1 ) .\",\n \"We found that DNR was present in both sample and delay phases ( Figure 5A ) .\",\n \"Information-carrying sample neurons had a fitted slope of 0 . 47 ( R2adj 0 . 39 , F ( 1 , 838 ) = 547 . 69 , p < 0 . 001 , CI = [0 . 43 0 . 51] ) and delay neurons had a slope of 0 . 50 ( R2adj 0 . 34 , F ( 1 , 342 ) = 175 . 60 , p < 0 . 001 , CI = [0 . 43 0 . 58] ) .\",\n \"As the slopes were not significantly different from 0 . 5 , this indicates that reference and probe color had an equal influence on neuronal responses .\",\n \"We thus show that DNR was observable in the neuronal population , and as a consequence of this computation , neurons had generally less information about the color identity at load\",\n \"2 . Some neurons showed encoding of color identity at higher loads , instead of loss of information .\",\n \"These neurons were abundant in both the sample phase and the delay phase ( Figure 3C ) .\",\n \"For example , the neuron shown in Figure 2B did not differentiate between color identities at load 1 but did so for load 2 , thus , representing a case of information gain ( instead of loss ) at a higher load .\",\n \"We wanted to understand if DNR , the mechanism that we found reduced color information at load 2 , could also produce color differentiation .\",\n \"The ‘normalization model of attention’ ( Reynolds and Heeger , 2009 ) incorporates divisive normalization , and can explain how attention can modulate neuronal responses .\",\n \"By attending a preferred ( or non-preferred ) second colored square in the load 2 condition the neuronal response of a neuron to the target location ( i . e . , to color A and to color B at the favorite location ) might be altered .\",\n \"As a result a difference between color A and B may arise even though each color by itself elicited a similar response .\",\n \"In other words , the interaction between the additional color and the target color is unequal .\",\n \"Neurons without a color differentiation at load 1 that gained differentiation at load 2 through this process ( e . g . , if the interaction of probe color A with reference color A is larger than the interaction of probe color A with reference color B , see Figure 5—figure supplement 2 for an example ) should have a population regression slope smaller than 0 . 5 .\",\n \"We thus hypothesized that the population of neurons showing information at load 2 , but not at load 1 ( e . g . , Figure 2B ) , would have a smaller slope than the neurons that lost information ( Figure 5A ) .\",\n \"Sample neurons had a slope of 0 . 19 ( R2adj 0 . 05 , F ( 1 , 446 ) = 23 . 0 , p < 0 . 001 , CI = [0 . 11 0 . 27] , Figure 5B ) , and delay neurons had a slope of –0 . 04 ( R2adj –0 . 015 , F ( 1 , 62 ) = 0 . 08 , p = 0 . 78 , CI = [–0 . 29 0 . 22] , Figure 5B ) .\",\n \"Both slopes were significantly smaller than 0 . 5 and smaller than the slopes of the non-informative neurons ( Figure 5—figure supplement 1 ) .\",\n \"This indicates that these neurons were influenced more strongly by the reference color , and that the addition of the probe color at load 2 resulted in an unequal interaction .\",\n \"Therefore , DNR was also computationally responsible for a gain of information at load 2 , in a specific subset of neurons .\"\n ],\n [\n \"Our results confirm behavioral findings that have been discussed in detail in an earlier study ( Balakhonov and Rose , 2017 ) .\",\n \"In brief , we found that the WM capacity of crows is limited to about four items , and that the two visual hemifields are largely independent ( i . e . , the number of items on one side does not affect change detection performance on the other side ) .\",\n \"Within each hemifield , performance dropped gradually with the addition of a second and third item but remained above chance .\",\n \"Fittingly , on the neuronal level , we found a markedly reduced amount of color information when the number of colored squares was increased from one to two ( roughly 50% reduction in correct trials ) .\",\n \"This suggests that WM could be conceptualized as a continuous resource that has to be divided between the two items ( Bays and Husain , 2008; van den Berg et al . , 2012; Wilken and Ma , 2004 ) , rather than two ‘simple’ slots that would each have the same amount of information irrespective of the memory load .\",\n \"This is also consistent with results of human neuroimaging that report decreased signal amplitude and precision with increasing memory load ( Emrich et al . , 2013; Sprague et al . , 2014 ) .\",\n \"In contrast , the hemispheric independence we observed would fit a slot-like model , in which the hemispheres as a whole act like discrete slots .\",\n \"A more nuanced version of the slot model ( ‘slots and averaging’; Zhang and Luck , 2008 ) could also account for graded amounts of information within a limited number of slots ( Fukuda et al . , 2010; Zhang and Luck , 2008 ) , as we found here .\",\n \"The mix of discrete and independent hemispheres with a graded allocation of information between items that we found is comparable to results by Buschman et al . , 2011 , observed in monkey PFC .\",\n \"On the neuronal level , recurrent connections between neurons within a hemisphere may reduce item differentiation when multiple items are present simultaneously , creating capacity limitations within the hemisphere ( Matsushima and Tanaka , 2014 ) .\",\n \"A lack of interhemispheric recurrent connections would make processing in the other hemisphere independent .\",\n \"Like in monkeys , WM capacity in crows may therefore result from neuronal activity patterns governed by multiple individual items .\",\n \"We probed the WM capacity of crows using colored squares , based on the task design of Buschman et al . , 2011 .\",\n \"Using the identical task allowed us to directly compare our neuronal results of WM capacity from NCL to results from PFC of monkeys .\",\n \"Task similarity is very important for such cross species comparisons as even small changes in task parameters may introduce substantial differences in neuronal responses , leading to potentially different conclusions .\",\n \"In a task similar to the one used here , Lara and Wallis , 2014 , have found that neurons in the PFC of monkeys encoded nearly no information about color , but instead about location .\",\n \"In their task monkeys had to memorize the color of squares at two locations on a screen , and were again confronted with a colored square at one of the two locations after a delay .\",\n \"The monkeys then had to indicate if the color at that location had changed .\",\n \"Lara and Wallis , 2014 , discuss the absence of color information in the neurons they recorded in relation to the task of Buschman et al . , 2011 , who like us , did find color information .\",\n \"In brief , the exact task design may determine the neuronal encoding of task relevant information ( Lara and Wallis , 2014 ) .\",\n \"Similar to the complex contribution of PFC neurons to WM , neurons of NCL can also encode a wide range of very different task relevant aspects , like color ( this study ) , spatial locations ( Rinnert et al . , 2019; Veit et al . , 2017 ) , and more abstract items like rules ( Veit and Nieder , 2013 ) and numerosities ( Ditz and Nieder , 2015 ) .\",\n \"One way to circumvent WM failure when item load increases is to allocate attention .\",\n \"Our results suggest that attention may play an important role in crow WM .\",\n \"Capacity limitation became apparent during encoding , as the amount of information at the end of the sample period was affected by the stimulus load .\",\n \"Adding a second and third item to the ipsilateral stimulus array reduced the amount of color information encoded by NCL neurons that carried over into the memory delay .\",\n \"Furthermore , neuronal activity in trials in which the birds made an incorrect response showed only weak encoding during the sample phase without information maintenance during the memory delay .\",\n \"This fits studies of human WM that have shown attentive filtering during encoding of stimuli influencing WM capacity ( Bays and Husain , 2008; Vogel et al . , 2005; Vogel and Machizawa , 2004 ) , and neuronal correlates of this have been reported for monkeys as well ( Buschman et al . , 2011 ) .\",\n \"Beyond the domain of sensory signals , attention and WM may be directly linked .\",\n \"Neuronal correlates of WM and attention overlap in PFC neurons , for example , Lebedev et al . , 2004 , found that a substantial amount of PFC neurons encode either an attentional signal , or a memory signal , and some ( hybrid ) neurons do both .\",\n \"A purely mnemonic function of PFC thereby seems unlikely .\",\n \"Indeed , very recently , Panichello and Buschman , 2021 , have reported that at the population level neurons of PFC encode ‘both the selection of items from working memory and attention to sensory inputs’ ( p . 2 ) , rather than just memory content .\",\n \"The independence of hemifields that we observed on the behavioral level ( this study and Balakhonov and Rose , 2017 ) and found in the neuronal responses could also be related to attention .\",\n \"Adding stimuli in the contralateral hemifield affected neither performance nor information maintained by NCL neurons , whereas additional ipsilateral stimuli strongly reduced both .\",\n \"This fits the influence of attention on WM and hemifield independence , which is consistently accentuated in studies in which attention had to be divided between the two hemifields ( Alvarez and Cavanagh , 2005; Buschman et al . , 2011; Cavanagh and Alvarez , 2005; Delvenne , 2005; Delvenne et al . , 2011 ) .\",\n \"Finally , the DNR computation may explain the responses of the neurons that gained information at load 2 through attentional processes predicted by the ‘normalization model of attention’ ( Reynolds and Heeger , 2009 ) .\",\n \"This may appear counter-intuitive and contradictory , considering that the same process is also responsible for the loss of information .\",\n \"However , when attention is overtly directed to a specific ( preferred or non-preferred ) item within the receptive field of a neuron , the DNR computation shifts its weighting of the normalized response toward the response of the attended item ( Reynolds et al . , 1999; Reynolds and Heeger , 2009 ) .\",\n \"This weighted normalization can produce a difference in the neuronal response to both color identities at load 2 , even if the neuronal response was non-informative at load 1 .\",\n \"At the population level we were able to observe such an effect as the reduced slope of the selectivity/interaction fit .\",\n \"Thus , an attentive process might have enhanced information in WM at higher loads .\",\n \"It is important to clarify that , as we did not use any form of attentional cueing in our study , we cannot explicitly test for such an attention effect .\",\n \"However , we do know that the animals participating in this study can use attentional cues to enhance their WM ( Fongaro and Rose , 2020 ) .\",\n \"The attention cues used by Fongaro and Rose , 2020 , positively affected not only encoding but also the maintenance and retrieval of the information held in WM , comparable to results from monkeys and humans ( Brady and Hampton , 2018; Souza and Oberauer , 2016 ) .\",\n \"We , therefore , want to emphasize that our data is in line with the interpretation that the birds possibly attended a load 2 stimulus array differently than a load 1 stimulus array in order to enhance their performance in trials with higher loads .\",\n \"Our neuronal recordings offer a mechanistic explanation for the behavioral effects , as we found clear evidence of DNR governing the neuronal responses tied to WM capacity that is in accordance with mammalian models of WM capacity ( Bouchacourt and Buschman , 2019; Lundqvist et al . , 2016; Lundqvist et al . , 2011 ) .\",\n \"The loss of information about color identity ( i . e . , neuronal response differentiation between colors ) can be accounted for by DNR when an item is added to a neuron’s receptive field .\",\n \"The normalization of neuronal firing rate diminishes the differentiation between color identities .\",\n \"As such it is analogous to neurophysiological responses from visual areas ( Carandini et al . , 1997; Reynolds et al . , 1999 ) and to the PFC during spatial WM ( Matsushima and Tanaka , 2014 ) .\",\n \"The WM model of Bouchacourt and Buschman , 2019 , is based solely on data from monkey electrophysiology , and thus implicitly tied to the layered columns of the neocortex .\",\n \"The results we report here show that the model also fits the neurophysiology of WM in crows .\",\n \"However , the picture is incomplete since important aspects of monkeys’ WM are still not investigated in crows .\",\n \"Oscillations of local field potentials are relevant for how information enters WM and how it is maintained ( Miller et al . , 2018 ) , and have been tied to normalization and competition between items in WM ( Lundqvist et al . , 2018b ) .\",\n \"Thus , the oscillatory interplay of the layers and different regions of the mammalian neocortex are important fields of research to further our understanding of WM .\",\n \"Such aspects are so far completely unknown in crows and their non-layered associative areas .\",\n \"This encourages further investigation into the neuronal circuits of WM in birds .\",\n \"There is also ongoing debate about the role of sustained activity during delay periods and how it relates to WM ( Constantinidis et al . , 2018; Lundqvist et al . , 2018a; Miller et al . , 2018 ) .\",\n \"We cannot report of any neuron that showed persistent activity comparable to those reported by classical WM studies in PFC ( e . g . , Funahashi et al . , 1989; Fuster and Alexander , 1971 ) , or in NCL ( Diekamp et al . , 2002b; Veit et al . , 2014; Veit and Nieder , 2013 ) .\",\n \"This may be reconcilable with some other contemporary models of WM .\",\n \"One major type of those models implements ‘synfire chains’ , where individual neurons fire sequentially ( and transiently ) to bridge temporal gaps and maintain task relevant contents ( Rajan et al . , 2016 ) .\",\n \"This has , for example , been reported to be the case in posterior parietal cortex of mice performing a T-maze task that required WM for cued spatial locations to be maintained ( Harvey et al . , 2012 ) .\",\n \"The transient activity of neurons that we report ( Figures 2 and 3A ) might fit into such models .\",\n \"However , our results can only be compared very cautiously to this ( since even small changes in task design significantly alter neuronal responses , e . g . , Lara and Wallis , 2014 ) .\",\n \"Therefore , while we cannot , yet , fully equate crow and monkey WM , our results raise two important questions about how WM is implemented on the level of neuronal networks that have implications for our comparative view of crow WM .\",\n \"The first regards the neuronal computations underlying WM .\",\n \"Is there a common canonical computation governing WM , or are there different solutions based on different neuronal architectures ?\",\n \"Recent work has shown that the sensory areas of mammals and birds show remarkably similar circuit organization ( Stacho et al . , 2020 ) .\",\n \"However , higher-order associative areas involved in WM , like the LPFC in mammals and the NCL in birds , have distinctly different architectures ( Stacho et al . , 2020 ) .\",\n \"The fact that differently organized areas like LPFC and NCL produce strikingly similar physiological responses points to shared computational principles .\",\n \"Modeling work already suggests that the competing WM capacity models can be accommodated into a unifying framework based on theoretical neuronal information sampling , where stochastic information sampling ( assumed for continuous resource models ) can account for item limitations better than fixed information sampling ( assumed by the slots and averaging models ) ( Schneegans et al . , 2020 ) .\",\n \"Similarly , DNR is already considered to be a general , canonical computation of the nervous system , present in evolutionarily distant phyla , for example , fruit flies and monkeys ( Carandini and Heeger , 2011 ) .\",\n \"The second question regards the tradeoff between WM flexibility and capacity ( Bouchacourt and Buschman , 2019 ) .\",\n \"Is the WM of a crow as flexible as that of a monkey ?\",\n \"Our results show that the computations by individual neurons that result in WM capacity limitations are virtually the same in crows and monkeys , highlighting a further aspect of WM that is similar between these animal groups ( Nieder , 2017 ) .\",\n \"Ultimately , our results were in line with different modern models of WM that implement DNR to explain capacity ( Bouchacourt and Buschman , 2019; Lundqvist et al . , 2016; Lundqvist et al . , 2011; Schneegans et al . , 2020 ) .\",\n \"However , the data we presented cannot carry a definitive conclusion about which of the different models fits best .\",\n \"For example , a tradeoff between flexibility and capacity ( Bouchacourt and Buschman , 2019 ) might be present , but further investigation into the models’ predictions is required .\",\n \"We do , however , show that mammalian models of WM are in line with WM in birds , which implies that fundamental aspects of WM are shared between these animal groups .\",\n \"Together , all these facets of crow WM capacity suggest that the different intricate neuronal architectures that carry out the computations in monkeys and crows have likely been shaped by convergent evolution – into systems that yield similar cognitive performances .\",\n \"The systems may share the same basic mechanisms and thus limitations .\",\n \"Further investigation into the oscillatory dynamics of WM in the avian brain may elucidate if birds also share the prominent limitation of a tradeoff between flexibility and capacity .\"\n ],\n [\n \"Two hand-raised carrion crows ( C . corone ) of 2 years of age served as subjects in this study .\",\n \"The birds were housed in spacious aviaries in social groups .\",\n \"During the experimental procedures , the animals were held on a controlled food protocol with ad libitum access to water and grit .\",\n \"All experimental procedures and housing conditions were carried out in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals and were authorized by the national authority ( LANUV ) .\",\n \"We used operant training chambers ( 50 × 50 . 5 × 77 . 5 cm3 , width × depth × height ) equipped with an acoustic pulse touchscreen ( 22’’ , ELO 2200L APR , Elo Touch Solutions Inc , Milpitas , CA ) and an infrared camera ( Sygonix , Nürnberg , Germany ) for remote monitoring .\",\n \"The birds sat on a wooden perch so that the distance between the bird’s eye and the touchscreen was 8 cm .\",\n \"Food pellets were delivered as a reward via a custom-made automatic feeder ( plans available at http://www . jonasrose . net/ ) .\",\n \"The position of the animal’s head was tracked online during the experiment by two open-source computer vision cameras ( ‘Pixy’ , CMUcam5 , Charmed Labs , Austin , TX ) that reported the location and angle between two LEDs .\",\n \"For tracking , we surgically implanted a lightweight head-post and used a lightweight 3D-printed mount with LEDs that was removed after each experimental session .\",\n \"The system reported the head location at a frame rate of 50 Hz and data was smoothed by integrating over two frames in MATLAB using custom programs on a control PC .\",\n \"All experiments were controlled by custom programs in MATLAB using the Biopsychology ( Rose et al . , 2008 ) and Psychophysics toolboxes ( Brainard , 1997 ) .\",\n \"Digital input and output of the control PC were handled by a microcontroller ( ODROID C1 , Hardkernel co . Ltd , Anyang , South Korea ) connected through a gigabit network running custom software ( available at http://www . jonasrose . net/ ) .\",\n \"The behavioral protocol was identical to the one described in Balakhonov and Rose , 2017 .\",\n \"We trained the birds to perform a delayed change localization task that had previously been used to test the performance under different WM loads in primates ( Buschman et al . , 2011 ) .\",\n \"Each trial started after a 2 s inter-trial-interval , with the presentation of a red dot centered on the touchscreen ( for a maximum of 40 s ) .\",\n \"The animals initiated the trial by centering their head in front of the red dot for 160 ms . This caused the red dot to disappear and a stimulus array of two to five colored squares to appear ( Figure 1A , ‘sample’ ) .\",\n \"The colored squares were presented for a period of 800 ms , during which the animals had to hold their head still and center their gaze on the screen ( ‘hold gaze’ , no more than 2 cm horizontal or vertical displacement , and no more than 20 degrees horizontal or vertical rotation ) .\",\n \"Failure to hold the head in this position resulted in an aborted trial .\",\n \"This sample phase was followed by a memory delay of 1000 ms after which the stimulus array reappeared with one color exchanged .\",\n \"The animal had to indicate the location of the color change by pecking the respective square .\",\n \"Correct responses were rewarded probabilistically ( BEO special pellets , in 55% of correct trials , additional 2 s illumination of the food receptacle in 100% of correct trials ) .\",\n \"Incorrect responses to colors that had not changed or a failure to respond within 4 s resulted in a brief screen flash and a 10 s timeout .\",\n \"The stimuli were presented at six fixed locations on the screen ( 1–6 , Figure 1A ) .\",\n \"In each session , one pair of colors was assigned to each of the six locations .\",\n \"Each location had its own distinct pair .\",\n \"These pairs were randomly chosen from a pool of 14 colors ( two color combinations were excluded since the animals did not discriminate them equally well during a pre-training ) .\",\n \"Let us consider Figure 1A as an example .\",\n \"The color change occurs in the middle-left where turquois ( T ) is presented during the sample and orange ( O ) during the choice .\",\n \"In this particular session the middle-left could thus show either of the following colors during the sample and choice: T-O ( shown in Figure 1A ) ; O-T; O-O; T-T; None-None .\",\n \"On the next session a new random pair of colors was displayed at this location .\",\n \"For identification and analysis an arbitrary label was assigned to each of the randomly drawn colors at the start of each session ( i . e . , left middle location: ‘color A’ , or ‘color B’; left top location: ‘color A’ , or ‘color B’; etc . ) .\",\n \"These indices do not refer to the order of presentation of the colors at any time , but were held constant for neuronal analysis .\",\n \"The order of presentation of colors within a pair , the target location ( where the color change occurred ) , and the number of stimuli in the array ( two to five ) were randomized and balanced across trials so that each condition had an equal likelihood to appear .\",\n \"The order of presentation of colors within a pair , the target location ( where the color change occurred ) , and the number of stimuli in the array ( two to five ) were randomized and balanced across trials so that each condition had an equal likelihood to appear .\",\n \"The color squares had a width of 10 degrees of visual angle ( DVA ) and were placed on the horizontal meridian of the screen and at 45 . 8 DVA above or below the meridian at a distance of 54 and 55 . 4 DVA from the center .\",\n \"This arrangement in combination with the head tracking ensured that all stimuli appeared outside of the binocular visual field of crows ( 37 . 6 DVA; Troscianko et al . , 2012 ) .\",\n \"Both animals were chronically implanted with a lightweight head-post to attach a small LED holder during the experiments .\",\n \"Before surgery , animals were deeply anesthetized with ketamine ( 50 mg/kg ) and xylazine ( 5 mg/kg ) .\",\n \"Once deeply anesthetized , animals were placed in a stereotaxic frame .\",\n \"After attaching the small head-post with dental acrylic , a microdrive with a multi-channel microelectrode was stereotactically implanted at the craniotomy ( Neuronexus Technologies Inc , Ann Arbor MI , DDrive ) .\",\n \"The electrode was positioned in NCL ( AP 5 . 0 , ML 13 . 0 ) of the left hemisphere ( coordinates for the region based on histological studies on the localization of NCL in crows; Veit and Nieder , 2013 ) .\",\n \"After the surgery , the crows received analgesics .\",\n \"Extracellular single neuron recordings were performed using chronically implanted multi-channel microelectrodes .\",\n \"The distance between recording sites was 50 µm .\",\n \"The signal was amplified , filtered , and digitized using Intan RHD2000 headstages and a USB-Interface board ( Intan Technologies LLC , Los Angeles , CA ) .\",\n \"The system also recorded digital event codes that were sent from the behavioral control PC using a custom IO device ( details available at http://www . jonasrose . net/ ) .\",\n \"Before each recording session , the electrodes were advanced manually using the microdrive .\",\n \"Recordings were started 20 min after the advancement , and each recording site was manually checked for neuronal signals .\",\n \"The signals were recorded at a sampling rate of 30 kHz and filtered with a band-pass filter at recording ( 0 . 5–7 . 5 kHz ) .\",\n \"The recorded neuronal signals were not pre-selected for task involvement .\",\n \"We performed spike sorting using the semi-automatic Klusta-suite software ( Rossant et al . , 2016 ) , which uses the high electrode count and their close spacing to isolate signals of single neurons .\",\n \"For spike sorting , we filtered with a high pass of 500 Hz and a low pass of 7125 Hz .\",\n \"The software utilizes the spatial distribution of the recorded signal along the different recording sites to untangle overlapping signals and separate signals with similar waveforms but different recording depths .\",\n \"All statistical analyses were performed in MATLAB ( 2018b , Mathworks Inc ) using commercially available toolboxes ( Curve Fitting Toolbox Version 3 . 5 . 3 , Statistics and Machine Learning Toolbox Version 10 . 2 ) and custom code .\",\n \"For all statistical tests , we assumed a significance level of α = 0 . 05 , unless stated otherwise .\",\n \"Trials were classified as error trials if the bird chose a location where no change of colors had appeared .\",\n \"Trials in which the bird did not choose any location or failed to maintain head fixation were not analyzed .\",\n \"All correct trials were included in the analysis of neural data .\",\n \"Depending on the analysis we refer to different ‘load conditions’ relative to referential sides of the screen which are either ipsilateral ( same side ) or contralateral ( opposite side ) , each with a possible load between one and three items .\",\n \"For the behavioral analyses the terms ipsilateral and contralateral refer to the location of the change that had to be detected .\",\n \"For the neurophysiological analyses the terms ipsilateral and contralateral refer to the respective neuron’s favorite location ( described in the following section ) .\",\n \"Because there were only very few error trials in the load one condition , we performed error trial analysis only for the load 2 and load 3 conditions .\",\n \"The behavioral data were analyzed as described in our previous study ( Balakhonov and Rose , 2017 ) , estimating the WM capacity K for each load by Equation 1 . ( 1 ) K=n*p where p is the percentage correct and n is the number of items in WM .\",\n \"This estimate has been applied to similar primate data and in studies with humans ( Johnson et al . , 2013; Kornblith et al . , 2016 ) .\",\n \"Based on a one-way ANOVA of color identity at a given location , we calculated a PEV statistic to measure the effect size of neuronal modulation .\",\n \"Its main parameter ω² is a measurement for the percentage to which the tested factor can explain the variance of the data , and it is calculated from the sum of squares of the effect ( SSeffect ) and the mean squares of the within-group ( error ) variance ( MSerror ) ( Equation 2 ) .\",\n \"( 2 ) ω2=SSeffect-df*MSerrorSStotal+MSerror For each neuron , we determined a ‘favorite location’ , which was defined as the location with the highest cumulative PEV , of the three possible locations on the right half of the screen , that is , opposite to the implanted hemisphere , across four non-overlapping bins during the sample phase ( bin size 200 ms , advanced in steps of 200 ms , from start till the end of the sample phase ) .\",\n \"The significance of calculated effect size values was determined by a permutation test .\",\n \"We ran the permutation to calculate the likelihood of getting an explained variance value bigger than the one calculated from the actual distribution of the data by randomly permuting the color identity labels and calculating the PEV 1000 times .\",\n \"The test thereby does not assume any distribution of the data and returns an unbiased estimate of the likelihood of generating an effect size within the data randomly .\",\n \"The measured value of explained variance from the actual dataset was assumed to be significant if the likelihood of randomly generating a bigger value was below 5% .\",\n \"We chose to not correct for the multiple comparisons at this level , as we reasoned that if we were to only include those neurons that had the most information at the individual loads ( i . e . , those with the highest PEV values that are significant even under very stringent statistical criteria ) we would have artificially inflated the amount of information present at each load .\",\n \"By including neurons that encoded less information too ( i . e . , at the uncorrected p-value ) our analysis population was more resistant to such outlier effects .\",\n \"We additionally performed our analyses using a more stringent statistical criterion ( significance of two consecutive non-overlapping bins ) and found the same results ( Figure 4—figure supplement 2 ) .\",\n \"We tested the proportions of significant neurons we found for the different trial phases by performing a binomial test , assuming a significance level α = 0 . 05 ( Equation 3 ) .\",\n \"( 3 ) PX=i=Bp0 , n=nkp0i1-p0n-i Calculating the probability p , of finding X significant neurons , given a total amount of i ( 362 ) neurons , and a probability p0 of 5% finding a significance by chance .\",\n \"We considered neuronal significance ( i . e . , significant PEV as determined above ) for each load independently .\",\n \"This means , we tested if the PEV of a neuron was significant three times with the permutation method described above: once for each of the three load conditions .\",\n \"Therefore , we can report seven groups of significance ( Table 1 , Figure 3C ) .\",\n \"Subsequently , we created three pooled groups ( Table\",\n \"1 ) from all neurons with a significant PEV at each individual load .\",\n \"We used these pooled groups for the population analyses ( Figures 4 and 5 ) .\",\n \"Neurons of these pooled groups , with a significant PEV during the sample phase were assigned to the ‘sample population’ , and neurons with a significant amount of information during the memory-delay phase were assigned to the ‘delay population’ ( significance criterion: one significant 200 ms bin , at α = 0 . 05 , see above for our reasoning not to correct for multiple comparison at this point ) .\",\n \"Thus , neurons with significant PEV during both the sample and delay phase were included in both subpopulations .\",\n \"We corrected for the unequal amount of correct and error trials when comparing information about color ( PEV ) between the trial conditions , by sub-sampling correct trials with the number of error trials 1000 times for each neuron .\",\n \"The resulting PEV values of correct trials were then averaged for each neuron , this population of averaged PEV values was then statistically tested against the PEV values of error trials ( of the same neurons ) using a dependent t-test .\",\n \"We tested for the presence of divisive normalization using the method of Reynolds et al . , 1999 .\",\n \"Three conditions were considered: ( 1 ) neuronal response to stimulus A , ( 2 ) neuronal response to stimulus B , and ( 3 ) neuronal response to the simultaneity of stimuli A and B . As we wanted to relate this to the information about color identity , we selected subsets of the favorite location and the additional two ipsilateral locations .\",\n \"To test how the neurons altered their response when multiple stimuli were presented simultaneously , we calculated the color selectivity index ( SE ) and the sensory interaction index ( SI ) of each neuron .\",\n \"SEi was calculated by subtracting the normalized firing rate for the chosen reference color i ( REFi ) at the neuron’s favorite location , from a second color j ( PROBEj ) at a different location ( ipsilateral to the favorite location , Equation 4 ) .\",\n \"( 4 ) SEi=PROBEj-REFi The resulting selectivity index lies between –1 ( completely selective for the reference color ) and 1 ( completely selective for the probe color ) .\",\n \"SI was calculated ( Equation 5 ) by subtracting the normalized firing rate for REFi from the normalized firing rate of the combination of REFi and PROBEj ( PAIRi , j ) .\",\n \"( 5 ) SIi , j=PAIRi , j-REFi This interaction index also lies between –1 ( full suppression of reference stimulus by the probe stimulus ) and 1 ( full enhancement of the reference stimulus by the probe stimulus ) .\",\n \"As each of the three locations had two possible colors , we calculated eight SE and SI indices per neuron and performed a linear regression for all indices .\",\n \"This is required as each stimulus combination is informative about the normalization .\",\n \"The effects of divisive normalization were compared between the sample and the delay phase .\",\n \"Therefore , SE and SI indices were calculated across the entire sample ( 800 ms ) and memory delay ( 1000 ms ) phase .\",\n \"Neurons with significant information were accordingly identified over the entire sample and delay as one bin , using the permutation test described in the section ‘information about color identity’ .\",\n \"We considered the entire sample and delay phase because we wanted to analyze the population response as a whole , irrespective of highly diverse response profiles of individual neurons .\",\n \"To visualize the different groups of neurons that encoded and maintained information about the color identity during different phases of the trial , we performed a hierarchical clustering analysis in MATLAB on the normalized PEV values of individual neurons throughout the trial .\",\n \"We used a ( 1 − correlation ) distance metric and an average distance linkage function for a maximum of seven clusters .\",\n \"The maximum number of clusters was first determined by calculating the clustering for different amounts of clusters ( 1–10 ) and subsequently calculating the within-cluster sum-of-squares .\",\n \"This resulted in a graph that allowed us to visually inspect the tradeoff between cluster number and fit improvement , from which we estimated the inflection point ( elbow method ) .\",\n \"A cluster number of seven presented the best tradeoff that allowed visualization of the different groups at an acceptable clustering success .\",\n \"We then ordered the neuron clusters to minimize the average distance between the clusters in the dendrogram .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Complex cognition relies on flexible working memory , which is severely limited in its capacity .","The neuronal computations underlying these capacity limits have been extensively studied in humans and in monkeys , resulting in competing theoretical models .","We probed the working memory capacity of crows ( Corvus corone ) in a change detection task , developed for monkeys ( Macaca mulatta ) , while we performed extracellular recordings of the prefrontal-like area nidopallium caudolaterale .","We found that neuronal encoding and maintenance of information were affected by item load , in a way that is virtually identical to results obtained from monkey prefrontal cortex .","Contemporary neurophysiological models of working memory employ divisive normalization as an important mechanism that may result in the capacity limitation .","As these models are usually conceptualized and tested in an exclusively mammalian context , it remains unclear if they fully capture a general concept of working memory or if they are restricted to the mammalian neocortex .","Here , we report that carrion crows and macaque monkeys share divisive normalization as a neuronal computation that is in line with mammalian models .","This indicates that computational models of working memory developed in the mammalian cortex can also apply to non-cortical associative brain regions of birds ."],"string":"[\n \"Complex cognition relies on flexible working memory , which is severely limited in its capacity .\",\n \"The neuronal computations underlying these capacity limits have been extensively studied in humans and in monkeys , resulting in competing theoretical models .\",\n \"We probed the working memory capacity of crows ( Corvus corone ) in a change detection task , developed for monkeys ( Macaca mulatta ) , while we performed extracellular recordings of the prefrontal-like area nidopallium caudolaterale .\",\n \"We found that neuronal encoding and maintenance of information were affected by item load , in a way that is virtually identical to results obtained from monkey prefrontal cortex .\",\n \"Contemporary neurophysiological models of working memory employ divisive normalization as an important mechanism that may result in the capacity limitation .\",\n \"As these models are usually conceptualized and tested in an exclusively mammalian context , it remains unclear if they fully capture a general concept of working memory or if they are restricted to the mammalian neocortex .\",\n \"Here , we report that carrion crows and macaque monkeys share divisive normalization as a neuronal computation that is in line with mammalian models .\",\n \"This indicates that computational models of working memory developed in the mammalian cortex can also apply to non-cortical associative brain regions of birds .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Working memory is the brain’s ability to temporarily hold and manipulate information .","It is essential for carrying out complex cognitive tasks , such as reasoning , planning , following instructions or solving problems .","Unlike long-term memory , information is not stored and recalled , but held in an accessible state for brief periods .","However , the capacity of working memory is very limited .","Humans , for example , can only hold around four items of information simultaneously .","There are various competing theories about how this limitation arises from the network of neurons in the brain .","These models are based on studies of humans and other primates .","But memory limitations are not exclusive to mammals .","Indeed , the working memory of some birds , such as crows , has a similar capacity to humans despite the architecture of their brains being very different to mammals .","So , how do brains with such distinct structural differences produce working memories with similar capacities ?","To investigate , Hahn et al . probed the working memory of carrion crows in a change detection task developed for macaque monkeys .","Crows were trained to memorize varying numbers of colored squares and indicate which square had changed after a one second delay when the screen went blank .","While the crows performed the task , Hahn et al . measured the activity of neurons in an area of the brain equivalent to the prefrontal cortex , the central hub of cognition in mammals .","The experiments showed that neurons in the crow brain responded to the changing colors virtually the same way as neurons in monkeys .","Hahn et al . also noticed that increasing the number of items the crows had to remember affected individual neurons in a similar fashion as had previously been observed in monkeys .","This suggests that birds and monkeys share the same central mechanisms of , and limits to , working memory despite differences in brain architecture .","The similarities across distantly related species also validates core ideas about the limits of working memory developed from studies of mammals ."],"string":"[\n \"Working memory is the brain’s ability to temporarily hold and manipulate information .\",\n \"It is essential for carrying out complex cognitive tasks , such as reasoning , planning , following instructions or solving problems .\",\n \"Unlike long-term memory , information is not stored and recalled , but held in an accessible state for brief periods .\",\n \"However , the capacity of working memory is very limited .\",\n \"Humans , for example , can only hold around four items of information simultaneously .\",\n \"There are various competing theories about how this limitation arises from the network of neurons in the brain .\",\n \"These models are based on studies of humans and other primates .\",\n \"But memory limitations are not exclusive to mammals .\",\n \"Indeed , the working memory of some birds , such as crows , has a similar capacity to humans despite the architecture of their brains being very different to mammals .\",\n \"So , how do brains with such distinct structural differences produce working memories with similar capacities ?\",\n \"To investigate , Hahn et al . probed the working memory of carrion crows in a change detection task developed for macaque monkeys .\",\n \"Crows were trained to memorize varying numbers of colored squares and indicate which square had changed after a one second delay when the screen went blank .\",\n \"While the crows performed the task , Hahn et al . measured the activity of neurons in an area of the brain equivalent to the prefrontal cortex , the central hub of cognition in mammals .\",\n \"The experiments showed that neurons in the crow brain responded to the changing colors virtually the same way as neurons in monkeys .\",\n \"Hahn et al . also noticed that increasing the number of items the crows had to remember affected individual neurons in a similar fashion as had previously been observed in monkeys .\",\n \"This suggests that birds and monkeys share the same central mechanisms of , and limits to , working memory despite differences in brain architecture .\",\n \"The similarities across distantly related species also validates core ideas about the limits of working memory developed from studies of mammals .\"\n]","isConversational":false},"year":{"kind":"string","value":"2021"}}},{"rowIdx":97,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["epidemiology and global health","immunology and inflammation"],"string":"[\n \"epidemiology and global health\",\n \"immunology and inflammation\"\n]","isConversational":false},"title":{"kind":"string","value":"The complex relationship of exposure to new Plasmodium infections and incidence of clinical malaria in Papua New Guinea"},"id":{"kind":"string","value":"elife-23708-v2"},"sections":{"kind":"list like","value":[["Renewed emphasis on malaria control has resulted in substantial reductions in overall malaria prevalence and incidence in many endemic countries ( World Health Organization , 2015 ) .","However , where transmission persists , it is highly heterogeneous even on small spatial scales ( Bousema et al . , 2012 ) .","Individual exposure is further influenced by factors such as use of bednets , attractiveness to mosquitoes , or behavioural differences .","In Papua New Guinea ( PNG ) , malaria prevalence has sharply declined in the last decade , largely as a result of two nationwide distributions of long-lasting insecticide treated bednets ( LLIN ) ( Hetzel et al . , 2015; Hetzel et al . , 2014; Hetzel et al . , 2012 ) .","P . vivax and P . falciparum PCR-prevalence in the general population was reduced from 32% and 39% in 2006 to 13% and 18% in 2010 ( Koepfli et al . , 2015 ) .","Already before this decline in malaria prevalence , studies in PNG had reported significant heterogeneity in malaria transmission attributed to local population structure and geographical diversity ( Hetzel et al . , 2015; Cattani et al . , 1986; Genton et al . , 1995; Müller et al . , 2003; Mueller et al . , 2009a ) .","Prior to the up-scaling of malaria control , P . vivax endemicity in PNG was among the highest worldwide ( Hetzel et al . , 2015 ) .","Clinical immunity to P . vivax was acquired very rapidly in PNG children , and the incidence of P . vivax clinical episodes peaked in children younger than two years with only very few P . vivax clinical episodes reported in children older than 5 years or adults ( Genton et al . , 2008; Michon et al . , 2007; Lin et al . , 2010; Betuela et al . , 2012 ) .","In contrast , the risk for uncomplicated P . falciparum clinical episodes increased during early childhood ( Lin et al . , 2010 ) and significant reductions in incidence of clinical episodes or high-density infections were only observed in children aged 5 years and older ( Michon et al . , 2007 ) .","Compared to the incidence of clinical malaria , prevalence of P . falciparum and P . vivax peaked in older age groups , with asymptomatic infections remaining common until adulthood in PNG ( Koepfli et al . , 2015; Mueller et al . , 2009a ) .","Concordant with the species-specific pattern in the burden of clinical episodes , P . vivax prevalence peaked in younger age groups than P . falciparum prevalence ( Koepfli et al . , 2015; Mueller et al . , 2009a ) .","As malaria transmission declines , it is important to understand the resulting changes in malaria prevalence and clinical incidence patterns , as well as the extent of heterogeneity in transmission within malaria endemic regions so that high-risk areas can be identified and targeted ( Mosha et al . , 2014 ) .","Most attempts to delineate high and low transmission areas made to-date , by both researchers and control programs , have used passive case surveillance or cross-sectional malaria indicator surveys .","These surveillance strategies result in clinical incidence and prevalence estimates , both of which are surrogate markers for transmission .","A more accurate understanding of the relationship between exposure to new infections and malaria prevalence or clinical incidence is needed to determine how accurately these surrogate markers represent heterogeneity in transmission at local scales .","In addition , quantifying clinical incidence in relation to exposure to blood-stage infections can increase our insight into the development and maintenance of immunity to malaria in a setting of sustained malaria control ( Battle et al . , 2015; Cameron et al . , 2015 ) .","The molecular force of blood-stage infection ( molFOB ) describes the number of new genotypes observed in consecutive blood samples from cohort participants over time ( Mueller et al . , 2012; Koepfli et al . , 2013 ) .","Genotyping of highly polymorphic markers detects superinfecting parasite clones in asymptomatic ( but parasitaemic ) or symptomatic individuals .","molFOB thus provides a longitudinal , individual and quantitative measure for exposure to new blood-stage malaria infections ( Mueller et al . , 2012; Koepfli et al . , 2013 ) .","For P . falciparum , molFOB is closely linked to the number of infective mosquito bites and therefore is a direct proxy for the actual force of infection ( FOI ) and thus for transmission in endemic settings ( Smith et al . , 2010 ) .","For P . vivax , clones appearing in the blood-stream can either originate directly from an infective mosquito bite or from a relapsing liver hypnozoite ( Koepfli et al . , 2013 ) .","For P . vivax , molFOB is thus a compound measure of exposure to newly acquired infections from mosquito bites and relapsing blood-stage infections .","The usefulness of molFOB as a surrogate marker of individual exposure was validated originally in a cohort of young PNG children 1–4 years of age , in which species-specific molFOB was the most important predictor of clinical incidence for both species ( Mueller et al . , 2012; Koepfli et al . , 2013 ) .","Although some spatial heterogeneity of transmission was observed in that study for both species , due to the high level of transmission the species-specific difference in rate of immune acquisition was the predominant feature in that study .","While P . vivax molFOB ( Pv-molFOB ) did not change with age , the incidence of P . vivax clinical episodes decreased significantly with age , with a faster rate of decrease in children with high Pv-molFOB [12] .","P . falciparum molFOB ( Pf-molFOB ) in that cohort was lower compared to Pv-molFOB and the incidence of P . falciparum clinical episodes increased in parallel with an increasing Pf-molFOB in children 1–3 years , reaching a plateau thereafter ( Lin et al . , 2010; Mueller et al . , 2012 ) .","These earlier results indicated that","( i ) immunity to P . vivax is acquired more rapidly in children with higher cumulative exposure , that","( ii ) this developing immunity led to proportionally fewer clinical P . vivax episodes in older children despite similar exposure to new P . vivax blood-stage infections , and that","( iii ) higher exposure to P . vivax blood-stage infections , compared to P . falciparum , resulted in a more advanced immunity to P . vivax in this age group compared to P . falciparum ( Doolan et al . , 2009; Longley et al . , 2016 ) .","The major challenge to these cross-species comparisons lies within the intrinsic differences of Pf- and Pv-molFOB: whereas Pf-molFOB is a direct marker of mosquito-borne transmission , Pv-molFOB is a composite measure reflecting both newly acquired infections and those caused by relapses of previously acquired infections .","In this study , we extend the analysis of molFOB’s relationship with incidence of clinical malaria episodes to older PNG children and a lower transmission scenario .","In addition , given the unique study design that randomized blood-stage only or blood- plus liver-stage treatment at enrolment ( Robinson et al . , 2015 ) , we are now , for the first time , able to compare the incidence of newly acquired P . falciparum infections with both the incidence of newly acquired P . vivax infections and relapsing P . vivax infections .","Advancing on previous studies that investigated each species individually ( Mueller et al . , 2012; Koepfli et al . , 2013 ) , we now provide a comparative analysis of P . falciparum and P . vivax , directly exploring the role of exposure to multiple Plasmodium species in the development of clinical immunity to P . falciparum and P . vivax malaria .","We quantify in detail the extent of heterogeneity in molFOB on a small geographical scale and relate this to heterogeneity in clinical episode incidence to investigate effects of small-scale variation in malaria transmission on local malaria epidemiology .","By combining in-depth molecular parasitological data with demographic and clinical data , this study thus provides detailed insights into the changing epidemiology of malaria in PNG in response to intense malaria control efforts ."],["This study was conducted in six villages in Maprik district , East Sepik Province , PNG between August 2009 and May 2010 ( Robinson et al . , 2015 ) .","524 children aged 5–10 years were enrolled and randomized to receive either chloroquine ( CQ ) , artemeter-lumefantrine ( AL ) and primaquine ( PQ ) ; or CQ , AL and placebo .","Demographic parameters of the 466 children that completed the full course of randomized treatment with PQ/CQ/AL ( n = 233 ) or placebo/CQ/AL ( n = 233 ) , and were thereafter closely followed for 8 months , were comparable between the six villages ( Table 1 ) .","P . vivax was the most common infection at enrolment with 48% of children positive by quantitative PCR ( qPCR ) , followed by P . falciparum ( 24% ) , P . malariae ( 15% ) and P . ovale ( 3%; Table 2 ) .","39% of children were not infected with any Plasmodium species at enrolment .","The vast majority of P . malariae ( 75% ) and almost all P . ovale infections ( 93% ) occurred in children co-infected with either P . vivax and/or P . falciparum ( Table 2 ) .","Prevalence of each Plasmodium species varied between villages ( P .","falciparum , 9–71%; P . vivax , 38–67%; P . malariae , 8–40%; P . ovale , 0–11%; Table","2 ) and was highest in Bolumita for all species .","Accordingly , mixed-species infections were also most prevalent in Bolumita ( Table 2 ) .","The multiplicity of infection ( MOI ) , that is , the number of parasite genotypes per infection , also varied between villages for both species ( mean P . falciparum MOI , 1 . 1–2 . 2 clones/infection; mean P . vivax MOI , 1 . 6–2 . 9 clones/infection ) and children from Bolumita carried more multi-clone infections with P . vivax and P . falciparum than children in other villages ( Table 2 ) .","Mean P . falciparum parasite density was almost two- to six-fold higher in Bolumita ( 331 18S rRNA gene copies/µl ) than in other villages ( 56–192 18S rRNA gene copies/µl , Table 2 ) .","Children who had received PQ for clearance of P . vivax hypnozoites experienced similar numbers of new blood-stage infections with P . falciparum and P . vivax during follow-up ( mean Pf-molFOB = 1 . 5 CI95 [1 . 3–1 . 7] new blood-stage clones/year , Pv-molFOB = 1 . 6 [1 . 4–1 . 9] new blood-stage clones/year , Figure 1A , Figure 1—figure supplement 1 ) .","Pf-molFOB in the placebo arm was comparable to the PQ arm ( mean Pf-molFOB = 1 . 4 [1 . 2–1 . 6] new blood-stage clones/year ) , whereas due to the hypnozoite reservoir Pv-molFOB was more than three times higher in the placebo arm compared to the PQ arm ( mean Pv-molFOB = 5 . 4 [4 . 9–5 . 8] new blood-stage clones/year ) .","Pv-molFOB in the placebo arm showed a pronounced peak at months 2–3 of follow-up , which likely represents a wave of fast-relapsing hypnozoites in children who did not receive PQ ( Figure 1A ) .","P . vivax prevalence in the PQ arm was comparable to P . falciparum prevalence throughout the study and increased steadily , irrespective of the diagnostic method used ( Figure 1B and C ) .","P . vivax prevalence increased more rapidly in the placebo arm until month 3 of follow-up and dropped thereafter , similar to patterns in Pv-molFOB in the same arm .","Prevalence as measured by qPCR did not reach pre-treatment levels until the end of the study for any of the four Plasmodium species ( Figure 1B , Figure 1—figure supplement 2 ) .","At the end of follow-up , P . vivax prevalence by qPCR in the placebo arm was 25% [19–31%] , and therefore more than two-fold higher than in the PQ arm ( 9% [6–14%]; Figure 1B ) .","Also , throughout follow-up , P . vivax prevalence in the placebo arm was 2–3 fold higher compared to the PQ arm , suggesting that at least 50% of the overall P . vivax prevalence in this cohort can be attributed to the contribution of relapses .","Similarly , throughout and at the end of follow-up P . vivax prevalence in the placebo arm was 2–3 fold higher compared to P . falciparum ( irrespective of treatment arm; P . falciparum prevalence at end of follow-up , 10% [8–14%] ) , which is in agreement with the prevalence pattern at enrolment .","Assuming equal transmission from mosquitoes for both species , which was corroborated by a comparable Pf-molFOB and Pv-molFOB in the PQ arm , P . vivax relapses have contributed to a P . vivax prevalence twice as high as that of P . falciparum .","In the present study design , recurrent blood-stage infection can either originate from a new transmission event ( both arms and all species ) or for P . vivax and P . ovale also from a relapse of any previous infection ( placebo arm only ) .","After adjusting for the effect of PQ treatment ( Robinson et al . , 2015 ) , village of residence and infection status by qPCR at enrolment were the main predictors for the risk of recurrent Plasmodium spp .","during follow-up ( Table 3 ) .","Interestingly , in addition to a protective effect against recurrent P . vivax and P . ovale , the risk of recurrent P . falciparum was also reduced by 27% [0–48%] after PQ treatment ( p=0 . 064 ) .","The risk of a recurrent infection ( measured by qPCR ) with P . falciparum , P . vivax and P . ovale varied more than 7-fold between villages , with a higher risk observed in Bolumita ( 78% , 77% , and 15% with recurrent P . vivax , P . falciparum and P . ovale , respectively ) compared to the other villages ( recurrent P . vivax , range 25–73%; recurrent P . falciparum , range 12–44%; recurrent P . ovale , range 0–7% ) .","For P . falciparum and P . vivax , a mixed infection at enrolment as measured by qPCR was further associated with up to a two-fold increased risk of recurrent infection ( P . falciparum: AHR = 2 . 08 [1 . 25–3 . 48] , p=0 . 005; P . vivax: AHR = 1 . 74 [1 . 14–2 . 65] , p=0 . 010; Table 3 ) , supporting the idea that focal transmission within villages leads to the presence of high-risk and low-risk individuals .","For P . malariae , the infection status at enrolment was a stronger predictor of risk of recurrent infection than village of residence .","An infection with P . falciparum , P . malariae or a mixed infection at enrolment as measured by qPCR was associated with up to a 6-fold increase in risk of recurrent P . malariae ( AHRPf-enrol = 3 . 54 [0 . 85–14 . 72] , p=0 . 083; AHRPm-enrol = 6 . 35 [1 . 31–30 . 81] , p=0 . 022; AHRmixed = 3 . 37 [0 . 88–12 . 90] , p=0 . 076; Table 3 ) .","Reported use of a LLIN during the night previous to enrolment was associated with a reduced risk of recurrent P . vivax and P . falciparum in univariate analyses ( Supplementary file 1 - Table 1 ) but to a lesser extent in multivariable analyses ( P . vivax: AHR = 0 . 62 [0 . 39–0 . 98] , p=0 . 043 , P . falciparum: AHR = 0 . 84 [0 . 49–144] , p=0 . 531 ) .","Haemoglobin ( Hb ) level at enrolment was negatively associated with the risk of recurrent infection with P . vivax ( AHR = 0 . 88 [0 . 80–0 . 98] , p=0 . 019 ) and P . falciparum ( AHR = 0 . 90 [0 . 80–1 . 02] , p=0 . 099 ) .","Patterns in the risk of recurrent infections with P . falciparum and P . vivax as measured by light microscopy ( LM , Supplementary file","2 ) were similar to those observed for re-infection as measured by qPCR .","When based on LM observation ( but not as measured by qPCR ) , increasing age was associated with a reduced risk of recurrent P . vivax ( AHR = 0 . 85 [0 . 77–0 . 95] , p=0 . 004; Supplementary file","2 ) but an increased risk of recurrent P . falciparum ( AHR = 1 . 16 [1 . 01–1 . 33] , p=0 . 037; Supplementary file 2 ) .","The incidence of new P . falciparum and P . vivax blood-stage clones detected during follow-up , that is molFOB , was highly variable between individual children and ranged from 0 to 18 new clones/year for P . falciparum ( Figure 2A ) and 0 to 36 or 23 new blood-stage clones/year for P . vivax in the placebo or PQ arm , respectively ( Figure 2B ) .","Mean Pf- and Pv-molFOB varied significantly between villages and were higher in Bolumita ( Pf-molFOB = 4 . 9 new blood-stage clones/year , Pv-molFOBPQ arm=4 . 4 new blood-stage clones/year , Pv-molFOBplacebo arm=12 . 1 new blood-stage clones/year; Table 4; Figure","3 ) than in the other villages ( Pf-molFOB , range 0 . 7–1 . 8 new blood-stage clones/year; Pv-molFOBPQ arm , range 0 . 03–2 . 2 new blood-stage clones/year; Pv-molFOBplacebo arm , range 2 . 3–7 . 4 new blood-stage clones/year ) .","In univariate analyses , new P . vivax infections were strongly associated with new P . falciparum infections per sampling interval and vice versa , suggesting concurrent exposure to the two species ( Supplementary file 1 – Table 2 ) .","However , these effects were reduced when other variables of varying exposure such as village of residence or infection at enrolment were included in the multivariable model ( P . vivax: IRRPQ arm=1 . 32 [0 . 92–1 . 89] , p=0 . 134; IRRPlacebo arm=1 . 10 [0 . 85–1 . 42] , p=0 . 466; P . falciparum: IRR = 1 . 15 [0 . 97–1 . 36] , p=0 . 100 , Table 4 ) .","LLIN use , although strongly associated with lower Pf- and Pv-molFOB in univariate analyses ( Supplementary file 1 – Table 2 ) , remained significantly associated in multivariable models only for P . vivax in the placebo arm , where sleeping under a LLIN in the night previous to enrolment was associated with a 38% [9–57%] reduction in Pv-molFOB ( p=0 . 013 , Table 4 ) .","Each additional year of age was associated with a 14% [0–26%] reduction Pv-molFOB per sampling interval in the PQ arm ( p=0 . 059 ) , while no age effect was observed in the placebo arm or for P . falciparum ( Table 4 ) .","Hb level at enrolment was negatively associated with Pf- and Pv-molFOB ( P . vivax: IRRPQ arm=0 . 85 [0 . 72–1 . 01] , p=0 . 063; IRRPlacebo arm=0 . 91 [0 . 85–0 . 99] , p=0 . 025; P . falciparum: IRR = 0 . 85 [0 . 75–0 . 97] , p=0 . 013 ) , suggesting anaemia in individuals continuously exposed to blood-stage infections .","A total of 98 clinical malaria episodes , here defined as fever plus presence of LM-detectable parasites , were observed during the study period .","Of these , 64 ( 65% ) exceeded the previously established pyrogenic thresholds of 2500 and 500 parasites/µl per LM for P . falciparum and P . vivax , respectively ( Mueller et al . , 2009b ) .","P . falciparum was the most common cause of clinical malaria episodes ( P . falciparum , 64 clinical episodes; P . vivax , 31 clinical episodes; mixed P . falciparum/P . vivax by LM , 3 clinical episodes ) , despite lower incidence of new P . falciparum blood-stage clones compared with P . vivax ( P . falciparum , 342 new P . falciparum blood-stage clones; P . vivax , 849 new blood-stage clones ) .","Including clinical episodes with mixed infection as determined by LM in the estimates for both species , clinical incidence rate ( IR ) was 0 . 28 [0 . 21–0 . 35] P . falciparum episodes/year and 0 . 12 [0 . 08–0 . 17] P . vivax episodes/year .","At least one new blood-stage clone was detected in 70% ( 47/67 ) of samples from P . falciparum and 71% ( 24/34 ) of samples from P . vivax clinical episodes .","Of these clinical episodes with new blood-stage clones , 96% ( 45/47 ) and 83% ( 20/24 ) carried only the new but no persistent P . falciparum and P . vivax clones , respectively .","P . vivax clinical episodes occurred mainly in the placebo arm shortly after directly observed treatment ( DOT ) ( Robinson et al . , 2015 ) , the time of peak Pv-molFOB due to relapsing hypnozoites ( Figure 1A ) .","On an individual level , Pv-molFOB was positively associated with the risk of clinical episodes and each additional blood-stage P . vivax clone increased the risk of experiencing a P . vivax clinical episode slightly ( AHR = 1 . 07 [1 . 04–1 . 09] , p<0 . 001; Table 5 ) .","No significant differences in P . vivax clinical episode risk were observed between villages after adjusting for individual molFOB .","The risk for a P . vivax clinical episode decreased significantly with age ( AHR = 0 . 62 [0 . 46–0 . 84] , p=0 . 002; Table 5 ) .","This was paralleled by a decrease in P . vivax densities with age ( by qPCR , exp ( β ) =0 . 90 [0 . 83–0 . 98] , p=0 . 016; Supplementary file","3 ) indicative of more advanced immunity against P . vivax and thus better control of P . vivax densities in older children .","Patterns in the occurrence of P . falciparum clinical episodes during follow-up were more complex .","Between-village variation in P . falciparum clinical episode risk remained significant even after adjusting for individual exposure .","This effect was mainly apparent in Numangu , where children were at three- to six-fold higher risk for clinical episodes than children in other villages ( Numangu AHR = 4 . 29 [2 . 06–8 . 97] , other villages range AHR = 0 . 65 [0 . 20–2 . 08] to 1 . 39 [0 . 59–3 . 30]; Table 5 ) .","Overall , Pv-molFOB was positively associated with the risk of clinical episodes and each additional P . falciparum blood-stage clone slightly increased the risk for P . falciparum clinical episodes ( AHR = 1 . 15 [1 . 11–1 . 21] , p<0 . 001; Table 5 ) ; however , relative to the number of new blood-stage clones , P . falciparum clinical episodes were less frequent in highly exposed children compared to low-exposed children ( Figure 4 ) .","One clinical episode per three new blood-stage clones was detected in the least exposed children ( Pf-molFOB <4 new blood-stage clones/year ) , but only one clinical episode per 15 blood-stage clones in the highest exposed children ( Pf-molFOB >9 new blood-stage clones/year , Fisher’s exact test p<0 . 001 ) .","Age was not associated with the risk of P . falciparum clinical episodes .","LLIN use at enrolment was associated with a 56% [13–78%] reduced risk of P . falciparum clinical episodes ( p=0 . 018; Table 5 ) .","A higher risk for P . falciparum clinical episodes in children that had received PQ treatment for clearance of P . vivax liver stages was observed ( AHR = 1 . 79 [1 . 05–3 . 03] , p=0 . 031; Table 5 ) , suggesting a potential protective effect of P . vivax infections against P . falciparum clinical episodes .","Analysis to further explore this revealed that a concurrent or recent infection ( i . e . , at the same or preceding follow-up visit ) with P . vivax reduced the odds of a P . falciparum clinical episode by 65% [22-85%] ( p=0 . 011; Table 6 ) .","Further indications for a potential interaction between the two species was also observed when analyzing P . falciparum parasite densities , which were reduced by 55% [19–75%] ( p=0 . 008; Supplementary file","3 ) in mixed P . falciparum/P .","vivax infections compared to P . falciparum single infections , indicative of suppression of one of the species in mixed infections ."],["In the present study , we describe striking heterogeneity in malaria transmission not only between closely neighboring communities in Maprik district , PNG , but also substantial differences in exposure between individual children from the same village .","On village level this heterogeneity is apparent both when using traditional markers such as prevalence of infection , as well as when using the novel `reference standard` marker of individual exposure molFOB .","The increased resolution provided by molFOB further allows quantifying heterogeneity in exposure to new blood-stage infections between individual children .","Extending an earlier study in a neighboring area in which younger children had been enrolled , and that had identified molFOB as the most important predictor of malaria clinical episodes ( Mueller et al . , 2012; Koepfli et al . , 2013 ) , we confirmed that molFOB remains significantly associated with the risk for clinical episodes , but other factors such as age ( P . vivax ) , or a mixed Pf/Pv infection and village factors not captured by any of the other parameters assessed ( P . falciparum ) have a stronger effect on the risk for clinical malaria ( Table 5 and 6 ) .","Malaria transmission is often estimated by investigating the more accessible human host rather than the mosquito vector ( Tusting et al . , 2014 ) .","Because P . falciparum blood-stage infections are a direct outcome of mosquito-to-human transmission , infection parameters assessed in the human blood closely reflect P . falciparum transmission .","In contrast , relapses arising from dormant hypnozoites contribute substantially to P . vivax blood-stage infections ( Robinson et al . , 2015 ) , thus complicating the assessment of mosquito-to-human P . vivax transmission via infection parameters measured in the human blood .","The unique design of this study , that combined clearance of hypnozoites in half of the study participants with subsequent measurement of Pv-molFOB , allowed us to identify the burden of P . vivax infections due to mosquito-to-human transmission ( in hypnozoite-cleared individuals ) and compare it to the total burden of P . vivax infections .","We found a highly similar incidence and comparable temporal and spatial heterogeneity of P . falciparum and P . vivax infection acquired through renewed exposure to infected mosquito bites .","In children that experienced the full burden of relapses we found two-fold higher P . vivax infection prevalence and 4-times higher incidence ( molFOB ) compared to P . falciparum .","This first quantitative comparative assessment of P . falciparum and P . vivax transmission using non-entomological molecular parameters thus indicates that the observed differences in epidemiology between the two species are largely due to the high burden of relapsing P . vivax blood-stage infections ( Robinson et al . , 2015 ) .","Our molecular results thus support recent entomological data from Dreikikir district , 50 km from Maprik in East Sepik Province ( Reimer et al . , 2016 ) as well as earlier studies in East Sepik ( Hii et al . , 2001 ) that found similar sporozoite rates for P . falciparum and P . vivax .","Our previous analysis of this cohort had investigated the contribution of the hypnozoite reservoir to P . vivax infection and disease in order to inform strategies for achieving a sustained reduction of the P . vivax burden in PNG ( Robinson et al . , 2015 ) .","Here , we now describe the post-treatment re-infection dynamics in higher temporal detail .","Through a detailed comparison of these patterns for P . vivax and P . falciparum in PQ and placebo-treated children we further elucidate the contribution of relapses to P . vivax prevalence and clinical incidence , which are the most commonly used parameters for planning and monitoring of malaria control strategies .","P . vivax relapses accounted for more than half of the observed P . vivax prevalence in this cohort , which is lower than what was previously estimated as the contribution of relapses towards Pv-molFOB by comparison of treatment arms ( 77% , Robinson et al . , 2015 ) .","This difference can be accounted for by the higher number of P . vivax multiple clone infections that will accumulate more rapidly in the placebo-arm , where additional parasite clones from relapses and/or new infections may overlap with without a corresponding change in overall prevalence .","While we previously described a sustained effect of PQ treatment with significant reductions in Pv-molFOB observed up to eight months post treatment ( Robinson et al . , 2015 ) , here , we describe temporal variation in relapse rate with a rapid and wave-like recurrence of P . vivax in children from the placebo arm , who had retained their hypnozoites ( Figure 1A ) .","The concurrent , modest peaks in Pf-molFOB and Pv-molFOB in the PQ arm represent seasonal variation in transmission , which is highest in December and January in the study area ( Mueller et al . , 2012 ) ; corresponding to weeks 8–14 of follow-up ) .","A much higher peak and subsequent drop in appearance of new clones within three months after blood-stage only treatment , which was mirrored by a corresponding peak and drop in P . vivax prevalence ( Figure 1B andC ) , suggests that the incidence of relapse infections in the blood was not constant during follow-up .","P . vivax infections are often observed following treatment of P . falciparum malaria ( Douglas et al . , 2011 ) , and it can thus been hypothesized that the frequency of relapses may be temporarily increased after blood-stage antimalarial treatment ( White and Imwong , 2012 ) .","It is thus conceivable that the blood-stage antimalarial at baseline either triggered P . vivax relapses directly , or indirectly by allowing more hypnozoites to establish blood-stage infections in parasite-free hosts .","Alternatively , blood-stage P . vivax infections from hypnozoites relapsing shortly after baseline treatment ( during a period when antimalarial drugs were present at sub-curative levels ) may be suppressed to sub-detectable densities until complete waning of drug levels , resulting in simultaneous proliferation and detection of many new blood-stage clones within the first weeks after treatment ( Douglas et al . , 2011; Tarning et al . , 2014 ) .","More detailed modeling of the dynamics of individual P . vivax blood-stage infections and their association with potential triggers such as treatment or febrile illness will be required to determine the existence and importance of proposed relapse-triggers .","Malaria transmission showed high micro-spatial heterogeneity with more than 10-fold differences in Pf- and Pv-molFOB ( in the PQ arm ) between villages despite an overall high LLIN use by the study participants ( during follow-up; village average use , >90%; individual use , >50% ) .","Individual LLIN use at enrolment was nevertheless associated with a reduced risk of recurrent P . falciparum and P . vivax in univariate analyses .","However , after adjustment for other related variables ( i . e . , village of residence or infection status ) this association became non-significant .","Children living in Bolumita , where both P . falciparum and P . vivax molFOB and prevalence were highest , had a modestly lower LLIN use ( mean during follow-up , 92%; at enrolment , 77% ) compared to children from other villages ( mean during follow-up , 97–100%; at enrolment , 91–100% ) .","It is conceivable that LLIN use in the Bolumita community may be less effective in reducing malaria transmission ( Killeen et al . , 2007; Smith et al . , 2009 ) .","Potential differences between villages in mosquito density , behavior , sporozoite rate , proximity of house or play areas to mosquito breeding sites , or human behavioral factors ( related to LLIN use or other risk factors ) are however likely to be more important determinants for exposure to infective bites .","Small-scale variations in vector species and distribution between and within villages in PNG have been described previously ( Cattani et al . , 1986; Reimer et al . , 2016; Charlwood et al . , 1986; Hii et al . , 1997; Burkot et al . , 1988 ) and likely account in a large part for the micro-geographic heterogeneity in malariological parameters observed in this and other studies .","Assessing the incidence of new infections from consecutive blood samples using molecular methods ( as is necessary to determine molFOB ) , is complicated by fluctuating densities of clonal parasitemia that may temporarily fall below the limit of detection of the genotyping PCR , leading to imperfect detectability of clones ( Bretscher et al . , 2010; Felger et al . , 2012; Koepfli et al . , 2011 ) .","For P . falciparum , periodical sequestration of clones and absence from the peripheral blood at time of sampling may further contribute to imperfect detectability .","For P . vivax , generally low parasite densities aggravate the problem of imperfect detectability , and dis- and re-appearance of clones may be a result of imperfect detectability or relapsing hypnozoites .","The overall estimates of molFOB presented here may thus be biased .","Accurately assessing the effects of this imperfect detectability on parameters estimated from longitudinal genotyping data , such as molFOB , requires complex mathematical modeling ( Bretscher et al . , 2010; Felger et al . , 2012; Sama et al . , 2005; Sama et al . , 2006 ) .","However , although clonal detectability has been shown to decrease with age ( Felger et al . , 2012; Sama et al . , 2006 ) and MOI ( Koepfli et al . , 2011 ) , it is unlikely to vary substantially within our cohort’s age range and transmission setting .","Hence the observed differences in molFOB are likely to accurately reflect the relative differences in individual exposure as well as in population transmission levels within the study area .","Evaluating the impact of malaria control efforts requires monitoring changes in malariological metrics over extended periods of time .","Drawing comparisons between studies performed at different times in different age groups is particularly challenging because of the interplay of past and current exposure to infective bites and the resulting anti-malarial immunity in the study population of a certain age .","In our cohort , fewer P . vivax clinical episodes than P . falciparum clinical episodes were detected despite a higher incidence of P . vivax blood-stage infections , which is consistent with earlier studies in children of similar age ( Michon et al . , 2007 ) .","The very low incidence of clinical P . vivax episodes in our cohort , at 0 . 16 clinical episodes/year ( placebo arm [Robinson et al . , 2015] ) , contrasts drastically with that of 2 . 46 P . vivax clinical episodes/year observed in an earlier observational cohort of younger children aged 1–4 years from the same area ( Lin et al . , 2010 ) .","The 3-fold difference in Pv-molFOB between the two cohorts seems modest when compared to the 15-fold difference in the incidence of clinical Pv episodes ( Koepfli et al . , 2013 ) .","This suggests that the much lower incidence of P . vivax clinical illness in 5–10 years old children of this study is more likely explained by an advanced state of immunity to P . vivax compared to the younger children of the earlier cohort than the drop in P . vivax transmission .","Consistently , age emerged as the strongest factor associated with protection against P . vivax clinical episodes , Pv-molFOB and P . vivax parasite density .","Like in the previous cohort of younger children from neighboring villages ( Lin et al . , 2010 ) the incidence clinical P . vivax clinical episodes dropped significantly with age .","Unlike in the previous cohort of younger children ( Koepfli et al . , 2013 ) , in this cohort we additionally observed a drop in Pv-molFOB as well as P . vivax densities with age ( Table 4 and Supplementary file 3 ) .","This is a further indication of the substantial clinical immunity to P . vivax acquired during years of past exposure in the children of this study , which is still ongoing after the age of five .","In sharp contrast to the age-dependent decline of P . vivax clinical episode incidence , no age-dependent decrease in the incidence of clinical episodes was observed for P . falciparum .","In a previous cohort study conducted in 2004 in 5–14 year old children in an area from PNG with substantially higher transmission levels ( mean incidence risk 5 . 0 versus 0 . 8 infections/year , Michon et al . , 2007; Robinson et al . , 2015 ) , the risk of moderate- to high-density P . falciparum infections decreased significantly with age ( Michon et al . , 2007 ) .","Clinical immunity to P . falciparum in children of this earlier cohort was not only significantly further advanced compared to children of this cohort , but in addition , transmission was more homogeneous in the area of that study .","As a consequence age was a much better marker of life-time exposure and thus immune status compared to the present cohort .","In this cohort , exposure to P . falciparum infections was highly heterogeneous between study participants .","Mathematical modeling suggests that at heterogeneous transmission , changes of parasite prevalence and clinical episode incidence with age are less pronounced compared to settings with homogeneous transmission ( Ross and Smith , 2010 ) .","Although no age trends were observed for P . falciparum in this cohort , when children were stratified into groups ranging from low to high exposure we found that the proportion of P . falciparum clinical episodes relative to new infections decreased with increasing exposure .","This could either reflect the development of clinical immunity in highly exposed children , or premunition , a proposed mechanism by which established infections help to control superinfections by immunological cross-protection ( Sergent and Parrot , 1935; Smith et al . , 1999 ) .","In settings of decreasing and heterogeneous transmission , age alone may therefore not be a suitable marker of immunity to P . falciparum .","Instead , combining age and molFOB to estimate cumulative life-time exposure may provide a more accurate surrogate measure of the extent of acquired clinical immunity .","With P . falciparum transmission declining in PNG due to successful malaria control strategies ( Koepfli et al . , 2015 ) , it is conceivable that immunity against P . falciparum will develop more slowly , shifting the burden of disease towards older age groups or towards more complex , non-linear age patterns .","This delay in immune acquisition is however more than compensated by the overall much lower incidence of clinical malaria clinical episodes: in cohort studies in children younger than 4 years from Maprik district , clinical P . falciparum incidence had dropped from 2 . 56 clinical episodes/year before ( observational cohort , [Lin et al . , 2010; Mueller et al . , 2012] ) to 0 . 67 clinical episodes/year immediately after the free LLIN distribution campaign ( placebo arm , [Betuela et al . , 2012] ) .","Finally , given that four Plasmodium species co-exist in PNG , there has long been considerable interest in potential mechanisms of cross-species immunity and mixed species interactions ( Mueller et al . , 2009a; Bruce et al . , 2000; Smith et al . , 2001; Mehlotra et al . , 2000 ) .","However , there is so far no consistent evidence for the presence or absence of cross-protection among Plasmodium species .","P . vivax and P . falciparum infections in our study were concentrated in the same children and villages ( Figure 3 ) , likely due to overlapping focal transmission for P . falciparum and P . vivax and thus potentially high co-infection rates in mosquitoes .","Contrary to an earlier cohort in younger PNG children that found a decreased risk of P . falciparum clinical episodes after PQ radical cure ( Betuela et al . , 2012 ) , we found indications for an increased risk of P . falciparum illness after clearance of P . vivax hypnozoites using PQ .","Our data suggests that in individuals with substantial clinical immunity against P . vivax , a concurrent P . vivax infection may provide protection against P . falciparum clinical episodes by limiting P . falciparum densities ( Table 6 , Supplementary file 3 ) .","However , the comparably small number of clinical episodes in this and the earlier contrasting study does not allow an in-depth analysis of causal relationships and therefore does not allow firm conclusions on the potential effects and mechanisms of cross-species interactions in mixed infections .","In conclusion , this study provides detailed insight into the changing epidemiology of malaria in PNG children under sustained malaria control , by using molFOB as a powerful measure to quantitatively investigate patterns of new mosquito-derived P . falciparum and P . vivax infections versus those for P . vivax relapsing infections , as well as spatial and age trends in exposure to these infections .","Striking heterogeneity in malaria transmission between villages as well as in individual exposure to new P . falciparum and P . vivax infections persisted in our study area despite very high use of LLINs .","This presents a significant challenge for on-going malaria control efforts .","The comparable patterns of new mosquito-derived P . falciparum and P . vivax infections indicate that sustained use of LLINs does result in a comparable reduction in transmission of both species .","The higher incidence and prevalence of P . vivax infections observed in our data is thus directly linked to its ability to cause relapsing infections , highlighting the crucial role of hypnozoites for P . vivax epidemiology and the need to effectively intervene against these hidden stages .","Together , these insights provide a crucial link to evaluate the level of P . vivax mosquito-based transmission against that of P . falciparum and serve to calibrate other standard malaria indicators such as parasite prevalence or incidence of clinical episodes and to ultimately inform new approaches to surveillance and response systems ."],["This study was conducted in six villages in the Albinama and Balif areas , Maprik district , East Sepik Province , PNG between August 2009 and May 2010 .","The area is serviced by the Albinama health sub-center , Balif aid post and a network of health workers in all study villages .","The study design has been described in detail elsewhere ( Robinson et al . , 2015 ) .","Briefly , 524 children aged 5–10 years whose parents provided written informed consent for their participation were enrolled and randomized to receive either chloroquine ( CQ , days 1–3 , total dose 25 mg/kg ) , artemeter-lumefantrine ( Coartem , AL , days 11–13 , 2 mg/kg A , 12 mg/kg L ) and primaquine ( PQ , days 1–20 , 0 . 5 mg/kg/day ) ; or CQ ( days 1–3 ) , AL ( days 11–13 ) , and placebo ( days 1–20 ) over 20 days of directly observed treatment ( DOT1-20 ) in a double-blinded manner .","Children were actively visited and examined for signs and symptoms of malaria fortnightly at their schools for 8 months .","In addition , passive surveillance was provided by the local health centre , aid post and village health workers throughout the study period .","Finger-prick blood samples ( 250 µl ) were collected at fortnightly active-follow-up visits in the first 12 weeks and monthly thereafter , as well as from symptomatic children detected during active or passive morbidity surveillance .","Symptomatic children were tested for malaria infection with rapid diagnostic test ( RDT , CareStartMalaria pLDH/HRP2 Combo , AccessBio , USA ) , and only RDT and or LM-confirmed Plasmodium infections of any density were treated with a 3 day course of AL .","Household , village and health facility location data was collected using a handheld GPS receiver ( Garmin GPSmap62sc ) and maps were prepared using ArcGIS 10 . 2 ( Esri , USA ) .","The study received ethical clearance from the PNG IMR Institutional Review Board ( 0908 ) , the PNG Medical Advisory Committee ( 09 . 11 ) , the Ethics Committee of Basel 237/11 and was conducted in full concordance with the Declaration of Helsinki .","The study was registered on ClinicalTrials . gov ( NCT02143934 ) .","All blood samples were examined by LM and qPCR for detection and speciation of Plasmodium infections as described earlier ( Robinson et al . , 2015 ) .","Each blood slide was read independently by two skilled microscopists and re-read by an expert microscopist in case of discrepancies in positivity , speciation or density ( ≥2 x log10 difference ) .","Thick blood films were examined by LM for 200 fields ( 1000x magnification ) before being declared parasite-negative .","Parasite density was converted from the number of parasites per 200–500 white blood cells ( WBC ) to parasites/µl assuming 8000 WBC/µl ( WHO malaria microscopy training guide ) and calculated as the geometric mean of all positive reads .","DNA was extracted from the red blood cell pellet using the FavorPrep 96-well genomic DNA extraction kit ( Favorgen ) .","Samples carrying any Plasmodium spp .","infection were identified using a generic qPCR ( Wampfler et al . , 2013 ) and positives were subsequently tested in species-specific qPCRs ( Rosanas-Urgell et al . , 2010; Wampfler et al . , 2013 ) .","All qPCRs targeted the small subunit ( 18S ) ribosomal RNA gene and were performed as simplex ( P . vivax and P . falciparum ) or duplex qPCR ( P . malariae , P . ovale ) .","The concentration of target copies per µl of DNA was determined relative to a dilution row of standard plasmid as previously described ( Rosanas-Urgell et al . , 2010 ) .","The qPCR limit of detection ( LOD ) was determined using a standard plasmid dilution row and defined as the last point with more than 50% of replicates positive .","The LOD was 2 target copies/µl DNA , equaling 4 target copies/reaction , for all qPCRs .","All samples that crossed the fluorescence threshold were scored as positive for species-specific qPCRs .","In all samples positive in P . falciparum and/or P . vivax qPCRs , individual parasite clones were distinguished by genotyping the length-polymorphic Pf-msp2 or Pv-msp1F3 marker genes using capillary electrophoresis for highly precise fragment sizing ( Koepfli et al . , 2013; Koepfli et al . , 2011; Falk et al . , 2006; Schoepflin et al . , 2009 ) .","MOI was determined by counting the number of detected Pf-msp2 or Pv-msp1F3 alleles per sample .","molFOB was calculated from the number of new parasite clones detected per child or per sampling interval in the peripheral blood , divided by the individual time at risk or length of the interval .","A new infection was defined as a Pf-msp2 or Pv-msp1F3 allele not present in the two preceding genotyping-positive samples collected during active or passive surveillance ( Figure 1—figure supplement 1 ) .","Imperfect diagnostic detectability was not further adjusted for .","Children were considered at risk for clinical malaria clinical episodes until the end of the study or until they were censored ( on the last visit before two consecutively missed scheduled follow-up visits [Robinson et al . , 2015] ) .","For clinical endpoints , time-at-risk ( TAR ) was not further adjusted for interim missed follow-up visits because the intense active and passive case detection presumably led to detection of all malaria clinical episodes .","In contrast , TAR for analysis of molecular data ( e . g . , molFOB ) was reduced by the duration of the missed interval if a child was not seen by the study team for six weeks or more ( ≥42 days ) .","Children with a TAR of less than 3 months ( <84 days ) were excluded .","This resulted in an analyzed population of 466 children ( characterized in Table 1 ) of which 430 ( 92 . 3% ) completed the whole follow-up period , with a median of 15 ( IQR: 13–17 ) study contacts and mean TAR of 186 days ( IQR 168–223 days ) .","Time to first Plasmodium infection by qPCR and LM and its association with covariates were modeled using Cox regression , and the proportional hazards assumption was checked using the test based on the Schoenfeld residuals .","Multiple failure Cox regression was used to model the time to P . vivax and P . falciparum clinical episodes .","For statistical analysis , a malaria clinical episode was defined as fever ( >37 . 5°C axillary ) plus the presence of LM-detectable parasites , irrespective of RDT result or antimalarial treatment during the field visit .","Negative binomial generalized estimating equations ( GEE ) with log link function using an exchangeable correlation structure were used to model incidence of new infections with P . falciparum and P . vivax per sampling interval .","For these analyses , the time at risk was restricted to the intervals where molFOB could be estimated ( i . e . , starting in the third follow-up interval ) .","A binomial GEE with logit link function using an exchangeable correlation structure was used to model the odds of a P . falciparum clinical episode per interval .","Gaussian GEEs with log link function using an exchangeable correlation matrix were used to model log-transformed qPCR parasite densities in qPCR-positive samples , measured as 18S rRNA copy numbers/µl blood .","In the GEE and Cox models where molFOB was a covariate , it was included as a time-varying covariate .","When modelling molFOB ( Table 4 ) and the odds of clinical episodes ( Table 6 ) using GEEs , molFOB was calculated for each follow-up interval and used as predictor .","In Cox models investigating the risk of clinical episodes ( Table 5 ) , molFOB was calculated based on the new infections up to the time of failure and used as predictor .","In exploratory preliminary analyses we tested for a wide variety of interactions between covariates including interactions between all combinations of molFOB , enrolment infection status , age , village and bednet-usage .","All analyses were done using STATA v14 and R . Maps were drawn using Arcgis 10 . 1 ( Esri Inc . ) .","Ordinary kriging was used to generate the contour maps .","Semivariograms were used as the mathematical forms used to express autocorrelation .","Input variables for the spatial models were","( i ) molFOB ( prediction of independent variable , molFOB ) using the model presented in Table 4 ( negative binomial GEE ) for Pf and Supplementary file 4 for Pv ( same as that shown in Table 4 but with primaquine and placebo arms combined ) , resulting in Figure 3 Panels A and B;","( ii ) relative risk of clinical episodes as predicted by the model shown in Table 5 , resulting in Panels C and D of Figure 3 .","Relative standard error maps were generated by dividing the absolute standard error map by the model prediction map ."]],"string":"[\n [\n \"Renewed emphasis on malaria control has resulted in substantial reductions in overall malaria prevalence and incidence in many endemic countries ( World Health Organization , 2015 ) .\",\n \"However , where transmission persists , it is highly heterogeneous even on small spatial scales ( Bousema et al . , 2012 ) .\",\n \"Individual exposure is further influenced by factors such as use of bednets , attractiveness to mosquitoes , or behavioural differences .\",\n \"In Papua New Guinea ( PNG ) , malaria prevalence has sharply declined in the last decade , largely as a result of two nationwide distributions of long-lasting insecticide treated bednets ( LLIN ) ( Hetzel et al . , 2015; Hetzel et al . , 2014; Hetzel et al . , 2012 ) .\",\n \"P . vivax and P . falciparum PCR-prevalence in the general population was reduced from 32% and 39% in 2006 to 13% and 18% in 2010 ( Koepfli et al . , 2015 ) .\",\n \"Already before this decline in malaria prevalence , studies in PNG had reported significant heterogeneity in malaria transmission attributed to local population structure and geographical diversity ( Hetzel et al . , 2015; Cattani et al . , 1986; Genton et al . , 1995; Müller et al . , 2003; Mueller et al . , 2009a ) .\",\n \"Prior to the up-scaling of malaria control , P . vivax endemicity in PNG was among the highest worldwide ( Hetzel et al . , 2015 ) .\",\n \"Clinical immunity to P . vivax was acquired very rapidly in PNG children , and the incidence of P . vivax clinical episodes peaked in children younger than two years with only very few P . vivax clinical episodes reported in children older than 5 years or adults ( Genton et al . , 2008; Michon et al . , 2007; Lin et al . , 2010; Betuela et al . , 2012 ) .\",\n \"In contrast , the risk for uncomplicated P . falciparum clinical episodes increased during early childhood ( Lin et al . , 2010 ) and significant reductions in incidence of clinical episodes or high-density infections were only observed in children aged 5 years and older ( Michon et al . , 2007 ) .\",\n \"Compared to the incidence of clinical malaria , prevalence of P . falciparum and P . vivax peaked in older age groups , with asymptomatic infections remaining common until adulthood in PNG ( Koepfli et al . , 2015; Mueller et al . , 2009a ) .\",\n \"Concordant with the species-specific pattern in the burden of clinical episodes , P . vivax prevalence peaked in younger age groups than P . falciparum prevalence ( Koepfli et al . , 2015; Mueller et al . , 2009a ) .\",\n \"As malaria transmission declines , it is important to understand the resulting changes in malaria prevalence and clinical incidence patterns , as well as the extent of heterogeneity in transmission within malaria endemic regions so that high-risk areas can be identified and targeted ( Mosha et al . , 2014 ) .\",\n \"Most attempts to delineate high and low transmission areas made to-date , by both researchers and control programs , have used passive case surveillance or cross-sectional malaria indicator surveys .\",\n \"These surveillance strategies result in clinical incidence and prevalence estimates , both of which are surrogate markers for transmission .\",\n \"A more accurate understanding of the relationship between exposure to new infections and malaria prevalence or clinical incidence is needed to determine how accurately these surrogate markers represent heterogeneity in transmission at local scales .\",\n \"In addition , quantifying clinical incidence in relation to exposure to blood-stage infections can increase our insight into the development and maintenance of immunity to malaria in a setting of sustained malaria control ( Battle et al . , 2015; Cameron et al . , 2015 ) .\",\n \"The molecular force of blood-stage infection ( molFOB ) describes the number of new genotypes observed in consecutive blood samples from cohort participants over time ( Mueller et al . , 2012; Koepfli et al . , 2013 ) .\",\n \"Genotyping of highly polymorphic markers detects superinfecting parasite clones in asymptomatic ( but parasitaemic ) or symptomatic individuals .\",\n \"molFOB thus provides a longitudinal , individual and quantitative measure for exposure to new blood-stage malaria infections ( Mueller et al . , 2012; Koepfli et al . , 2013 ) .\",\n \"For P . falciparum , molFOB is closely linked to the number of infective mosquito bites and therefore is a direct proxy for the actual force of infection ( FOI ) and thus for transmission in endemic settings ( Smith et al . , 2010 ) .\",\n \"For P . vivax , clones appearing in the blood-stream can either originate directly from an infective mosquito bite or from a relapsing liver hypnozoite ( Koepfli et al . , 2013 ) .\",\n \"For P . vivax , molFOB is thus a compound measure of exposure to newly acquired infections from mosquito bites and relapsing blood-stage infections .\",\n \"The usefulness of molFOB as a surrogate marker of individual exposure was validated originally in a cohort of young PNG children 1–4 years of age , in which species-specific molFOB was the most important predictor of clinical incidence for both species ( Mueller et al . , 2012; Koepfli et al . , 2013 ) .\",\n \"Although some spatial heterogeneity of transmission was observed in that study for both species , due to the high level of transmission the species-specific difference in rate of immune acquisition was the predominant feature in that study .\",\n \"While P . vivax molFOB ( Pv-molFOB ) did not change with age , the incidence of P . vivax clinical episodes decreased significantly with age , with a faster rate of decrease in children with high Pv-molFOB [12] .\",\n \"P . falciparum molFOB ( Pf-molFOB ) in that cohort was lower compared to Pv-molFOB and the incidence of P . falciparum clinical episodes increased in parallel with an increasing Pf-molFOB in children 1–3 years , reaching a plateau thereafter ( Lin et al . , 2010; Mueller et al . , 2012 ) .\",\n \"These earlier results indicated that\",\n \"( i ) immunity to P . vivax is acquired more rapidly in children with higher cumulative exposure , that\",\n \"( ii ) this developing immunity led to proportionally fewer clinical P . vivax episodes in older children despite similar exposure to new P . vivax blood-stage infections , and that\",\n \"( iii ) higher exposure to P . vivax blood-stage infections , compared to P . falciparum , resulted in a more advanced immunity to P . vivax in this age group compared to P . falciparum ( Doolan et al . , 2009; Longley et al . , 2016 ) .\",\n \"The major challenge to these cross-species comparisons lies within the intrinsic differences of Pf- and Pv-molFOB: whereas Pf-molFOB is a direct marker of mosquito-borne transmission , Pv-molFOB is a composite measure reflecting both newly acquired infections and those caused by relapses of previously acquired infections .\",\n \"In this study , we extend the analysis of molFOB’s relationship with incidence of clinical malaria episodes to older PNG children and a lower transmission scenario .\",\n \"In addition , given the unique study design that randomized blood-stage only or blood- plus liver-stage treatment at enrolment ( Robinson et al . , 2015 ) , we are now , for the first time , able to compare the incidence of newly acquired P . falciparum infections with both the incidence of newly acquired P . vivax infections and relapsing P . vivax infections .\",\n \"Advancing on previous studies that investigated each species individually ( Mueller et al . , 2012; Koepfli et al . , 2013 ) , we now provide a comparative analysis of P . falciparum and P . vivax , directly exploring the role of exposure to multiple Plasmodium species in the development of clinical immunity to P . falciparum and P . vivax malaria .\",\n \"We quantify in detail the extent of heterogeneity in molFOB on a small geographical scale and relate this to heterogeneity in clinical episode incidence to investigate effects of small-scale variation in malaria transmission on local malaria epidemiology .\",\n \"By combining in-depth molecular parasitological data with demographic and clinical data , this study thus provides detailed insights into the changing epidemiology of malaria in PNG in response to intense malaria control efforts .\"\n ],\n [\n \"This study was conducted in six villages in Maprik district , East Sepik Province , PNG between August 2009 and May 2010 ( Robinson et al . , 2015 ) .\",\n \"524 children aged 5–10 years were enrolled and randomized to receive either chloroquine ( CQ ) , artemeter-lumefantrine ( AL ) and primaquine ( PQ ) ; or CQ , AL and placebo .\",\n \"Demographic parameters of the 466 children that completed the full course of randomized treatment with PQ/CQ/AL ( n = 233 ) or placebo/CQ/AL ( n = 233 ) , and were thereafter closely followed for 8 months , were comparable between the six villages ( Table 1 ) .\",\n \"P . vivax was the most common infection at enrolment with 48% of children positive by quantitative PCR ( qPCR ) , followed by P . falciparum ( 24% ) , P . malariae ( 15% ) and P . ovale ( 3%; Table 2 ) .\",\n \"39% of children were not infected with any Plasmodium species at enrolment .\",\n \"The vast majority of P . malariae ( 75% ) and almost all P . ovale infections ( 93% ) occurred in children co-infected with either P . vivax and/or P . falciparum ( Table 2 ) .\",\n \"Prevalence of each Plasmodium species varied between villages ( P .\",\n \"falciparum , 9–71%; P . vivax , 38–67%; P . malariae , 8–40%; P . ovale , 0–11%; Table\",\n \"2 ) and was highest in Bolumita for all species .\",\n \"Accordingly , mixed-species infections were also most prevalent in Bolumita ( Table 2 ) .\",\n \"The multiplicity of infection ( MOI ) , that is , the number of parasite genotypes per infection , also varied between villages for both species ( mean P . falciparum MOI , 1 . 1–2 . 2 clones/infection; mean P . vivax MOI , 1 . 6–2 . 9 clones/infection ) and children from Bolumita carried more multi-clone infections with P . vivax and P . falciparum than children in other villages ( Table 2 ) .\",\n \"Mean P . falciparum parasite density was almost two- to six-fold higher in Bolumita ( 331 18S rRNA gene copies/µl ) than in other villages ( 56–192 18S rRNA gene copies/µl , Table 2 ) .\",\n \"Children who had received PQ for clearance of P . vivax hypnozoites experienced similar numbers of new blood-stage infections with P . falciparum and P . vivax during follow-up ( mean Pf-molFOB = 1 . 5 CI95 [1 . 3–1 . 7] new blood-stage clones/year , Pv-molFOB = 1 . 6 [1 . 4–1 . 9] new blood-stage clones/year , Figure 1A , Figure 1—figure supplement 1 ) .\",\n \"Pf-molFOB in the placebo arm was comparable to the PQ arm ( mean Pf-molFOB = 1 . 4 [1 . 2–1 . 6] new blood-stage clones/year ) , whereas due to the hypnozoite reservoir Pv-molFOB was more than three times higher in the placebo arm compared to the PQ arm ( mean Pv-molFOB = 5 . 4 [4 . 9–5 . 8] new blood-stage clones/year ) .\",\n \"Pv-molFOB in the placebo arm showed a pronounced peak at months 2–3 of follow-up , which likely represents a wave of fast-relapsing hypnozoites in children who did not receive PQ ( Figure 1A ) .\",\n \"P . vivax prevalence in the PQ arm was comparable to P . falciparum prevalence throughout the study and increased steadily , irrespective of the diagnostic method used ( Figure 1B and C ) .\",\n \"P . vivax prevalence increased more rapidly in the placebo arm until month 3 of follow-up and dropped thereafter , similar to patterns in Pv-molFOB in the same arm .\",\n \"Prevalence as measured by qPCR did not reach pre-treatment levels until the end of the study for any of the four Plasmodium species ( Figure 1B , Figure 1—figure supplement 2 ) .\",\n \"At the end of follow-up , P . vivax prevalence by qPCR in the placebo arm was 25% [19–31%] , and therefore more than two-fold higher than in the PQ arm ( 9% [6–14%]; Figure 1B ) .\",\n \"Also , throughout follow-up , P . vivax prevalence in the placebo arm was 2–3 fold higher compared to the PQ arm , suggesting that at least 50% of the overall P . vivax prevalence in this cohort can be attributed to the contribution of relapses .\",\n \"Similarly , throughout and at the end of follow-up P . vivax prevalence in the placebo arm was 2–3 fold higher compared to P . falciparum ( irrespective of treatment arm; P . falciparum prevalence at end of follow-up , 10% [8–14%] ) , which is in agreement with the prevalence pattern at enrolment .\",\n \"Assuming equal transmission from mosquitoes for both species , which was corroborated by a comparable Pf-molFOB and Pv-molFOB in the PQ arm , P . vivax relapses have contributed to a P . vivax prevalence twice as high as that of P . falciparum .\",\n \"In the present study design , recurrent blood-stage infection can either originate from a new transmission event ( both arms and all species ) or for P . vivax and P . ovale also from a relapse of any previous infection ( placebo arm only ) .\",\n \"After adjusting for the effect of PQ treatment ( Robinson et al . , 2015 ) , village of residence and infection status by qPCR at enrolment were the main predictors for the risk of recurrent Plasmodium spp .\",\n \"during follow-up ( Table 3 ) .\",\n \"Interestingly , in addition to a protective effect against recurrent P . vivax and P . ovale , the risk of recurrent P . falciparum was also reduced by 27% [0–48%] after PQ treatment ( p=0 . 064 ) .\",\n \"The risk of a recurrent infection ( measured by qPCR ) with P . falciparum , P . vivax and P . ovale varied more than 7-fold between villages , with a higher risk observed in Bolumita ( 78% , 77% , and 15% with recurrent P . vivax , P . falciparum and P . ovale , respectively ) compared to the other villages ( recurrent P . vivax , range 25–73%; recurrent P . falciparum , range 12–44%; recurrent P . ovale , range 0–7% ) .\",\n \"For P . falciparum and P . vivax , a mixed infection at enrolment as measured by qPCR was further associated with up to a two-fold increased risk of recurrent infection ( P . falciparum: AHR = 2 . 08 [1 . 25–3 . 48] , p=0 . 005; P . vivax: AHR = 1 . 74 [1 . 14–2 . 65] , p=0 . 010; Table 3 ) , supporting the idea that focal transmission within villages leads to the presence of high-risk and low-risk individuals .\",\n \"For P . malariae , the infection status at enrolment was a stronger predictor of risk of recurrent infection than village of residence .\",\n \"An infection with P . falciparum , P . malariae or a mixed infection at enrolment as measured by qPCR was associated with up to a 6-fold increase in risk of recurrent P . malariae ( AHRPf-enrol = 3 . 54 [0 . 85–14 . 72] , p=0 . 083; AHRPm-enrol = 6 . 35 [1 . 31–30 . 81] , p=0 . 022; AHRmixed = 3 . 37 [0 . 88–12 . 90] , p=0 . 076; Table 3 ) .\",\n \"Reported use of a LLIN during the night previous to enrolment was associated with a reduced risk of recurrent P . vivax and P . falciparum in univariate analyses ( Supplementary file 1 - Table 1 ) but to a lesser extent in multivariable analyses ( P . vivax: AHR = 0 . 62 [0 . 39–0 . 98] , p=0 . 043 , P . falciparum: AHR = 0 . 84 [0 . 49–144] , p=0 . 531 ) .\",\n \"Haemoglobin ( Hb ) level at enrolment was negatively associated with the risk of recurrent infection with P . vivax ( AHR = 0 . 88 [0 . 80–0 . 98] , p=0 . 019 ) and P . falciparum ( AHR = 0 . 90 [0 . 80–1 . 02] , p=0 . 099 ) .\",\n \"Patterns in the risk of recurrent infections with P . falciparum and P . vivax as measured by light microscopy ( LM , Supplementary file\",\n \"2 ) were similar to those observed for re-infection as measured by qPCR .\",\n \"When based on LM observation ( but not as measured by qPCR ) , increasing age was associated with a reduced risk of recurrent P . vivax ( AHR = 0 . 85 [0 . 77–0 . 95] , p=0 . 004; Supplementary file\",\n \"2 ) but an increased risk of recurrent P . falciparum ( AHR = 1 . 16 [1 . 01–1 . 33] , p=0 . 037; Supplementary file 2 ) .\",\n \"The incidence of new P . falciparum and P . vivax blood-stage clones detected during follow-up , that is molFOB , was highly variable between individual children and ranged from 0 to 18 new clones/year for P . falciparum ( Figure 2A ) and 0 to 36 or 23 new blood-stage clones/year for P . vivax in the placebo or PQ arm , respectively ( Figure 2B ) .\",\n \"Mean Pf- and Pv-molFOB varied significantly between villages and were higher in Bolumita ( Pf-molFOB = 4 . 9 new blood-stage clones/year , Pv-molFOBPQ arm=4 . 4 new blood-stage clones/year , Pv-molFOBplacebo arm=12 . 1 new blood-stage clones/year; Table 4; Figure\",\n \"3 ) than in the other villages ( Pf-molFOB , range 0 . 7–1 . 8 new blood-stage clones/year; Pv-molFOBPQ arm , range 0 . 03–2 . 2 new blood-stage clones/year; Pv-molFOBplacebo arm , range 2 . 3–7 . 4 new blood-stage clones/year ) .\",\n \"In univariate analyses , new P . vivax infections were strongly associated with new P . falciparum infections per sampling interval and vice versa , suggesting concurrent exposure to the two species ( Supplementary file 1 – Table 2 ) .\",\n \"However , these effects were reduced when other variables of varying exposure such as village of residence or infection at enrolment were included in the multivariable model ( P . vivax: IRRPQ arm=1 . 32 [0 . 92–1 . 89] , p=0 . 134; IRRPlacebo arm=1 . 10 [0 . 85–1 . 42] , p=0 . 466; P . falciparum: IRR = 1 . 15 [0 . 97–1 . 36] , p=0 . 100 , Table 4 ) .\",\n \"LLIN use , although strongly associated with lower Pf- and Pv-molFOB in univariate analyses ( Supplementary file 1 – Table 2 ) , remained significantly associated in multivariable models only for P . vivax in the placebo arm , where sleeping under a LLIN in the night previous to enrolment was associated with a 38% [9–57%] reduction in Pv-molFOB ( p=0 . 013 , Table 4 ) .\",\n \"Each additional year of age was associated with a 14% [0–26%] reduction Pv-molFOB per sampling interval in the PQ arm ( p=0 . 059 ) , while no age effect was observed in the placebo arm or for P . falciparum ( Table 4 ) .\",\n \"Hb level at enrolment was negatively associated with Pf- and Pv-molFOB ( P . vivax: IRRPQ arm=0 . 85 [0 . 72–1 . 01] , p=0 . 063; IRRPlacebo arm=0 . 91 [0 . 85–0 . 99] , p=0 . 025; P . falciparum: IRR = 0 . 85 [0 . 75–0 . 97] , p=0 . 013 ) , suggesting anaemia in individuals continuously exposed to blood-stage infections .\",\n \"A total of 98 clinical malaria episodes , here defined as fever plus presence of LM-detectable parasites , were observed during the study period .\",\n \"Of these , 64 ( 65% ) exceeded the previously established pyrogenic thresholds of 2500 and 500 parasites/µl per LM for P . falciparum and P . vivax , respectively ( Mueller et al . , 2009b ) .\",\n \"P . falciparum was the most common cause of clinical malaria episodes ( P . falciparum , 64 clinical episodes; P . vivax , 31 clinical episodes; mixed P . falciparum/P . vivax by LM , 3 clinical episodes ) , despite lower incidence of new P . falciparum blood-stage clones compared with P . vivax ( P . falciparum , 342 new P . falciparum blood-stage clones; P . vivax , 849 new blood-stage clones ) .\",\n \"Including clinical episodes with mixed infection as determined by LM in the estimates for both species , clinical incidence rate ( IR ) was 0 . 28 [0 . 21–0 . 35] P . falciparum episodes/year and 0 . 12 [0 . 08–0 . 17] P . vivax episodes/year .\",\n \"At least one new blood-stage clone was detected in 70% ( 47/67 ) of samples from P . falciparum and 71% ( 24/34 ) of samples from P . vivax clinical episodes .\",\n \"Of these clinical episodes with new blood-stage clones , 96% ( 45/47 ) and 83% ( 20/24 ) carried only the new but no persistent P . falciparum and P . vivax clones , respectively .\",\n \"P . vivax clinical episodes occurred mainly in the placebo arm shortly after directly observed treatment ( DOT ) ( Robinson et al . , 2015 ) , the time of peak Pv-molFOB due to relapsing hypnozoites ( Figure 1A ) .\",\n \"On an individual level , Pv-molFOB was positively associated with the risk of clinical episodes and each additional blood-stage P . vivax clone increased the risk of experiencing a P . vivax clinical episode slightly ( AHR = 1 . 07 [1 . 04–1 . 09] , p<0 . 001; Table 5 ) .\",\n \"No significant differences in P . vivax clinical episode risk were observed between villages after adjusting for individual molFOB .\",\n \"The risk for a P . vivax clinical episode decreased significantly with age ( AHR = 0 . 62 [0 . 46–0 . 84] , p=0 . 002; Table 5 ) .\",\n \"This was paralleled by a decrease in P . vivax densities with age ( by qPCR , exp ( β ) =0 . 90 [0 . 83–0 . 98] , p=0 . 016; Supplementary file\",\n \"3 ) indicative of more advanced immunity against P . vivax and thus better control of P . vivax densities in older children .\",\n \"Patterns in the occurrence of P . falciparum clinical episodes during follow-up were more complex .\",\n \"Between-village variation in P . falciparum clinical episode risk remained significant even after adjusting for individual exposure .\",\n \"This effect was mainly apparent in Numangu , where children were at three- to six-fold higher risk for clinical episodes than children in other villages ( Numangu AHR = 4 . 29 [2 . 06–8 . 97] , other villages range AHR = 0 . 65 [0 . 20–2 . 08] to 1 . 39 [0 . 59–3 . 30]; Table 5 ) .\",\n \"Overall , Pv-molFOB was positively associated with the risk of clinical episodes and each additional P . falciparum blood-stage clone slightly increased the risk for P . falciparum clinical episodes ( AHR = 1 . 15 [1 . 11–1 . 21] , p<0 . 001; Table 5 ) ; however , relative to the number of new blood-stage clones , P . falciparum clinical episodes were less frequent in highly exposed children compared to low-exposed children ( Figure 4 ) .\",\n \"One clinical episode per three new blood-stage clones was detected in the least exposed children ( Pf-molFOB <4 new blood-stage clones/year ) , but only one clinical episode per 15 blood-stage clones in the highest exposed children ( Pf-molFOB >9 new blood-stage clones/year , Fisher’s exact test p<0 . 001 ) .\",\n \"Age was not associated with the risk of P . falciparum clinical episodes .\",\n \"LLIN use at enrolment was associated with a 56% [13–78%] reduced risk of P . falciparum clinical episodes ( p=0 . 018; Table 5 ) .\",\n \"A higher risk for P . falciparum clinical episodes in children that had received PQ treatment for clearance of P . vivax liver stages was observed ( AHR = 1 . 79 [1 . 05–3 . 03] , p=0 . 031; Table 5 ) , suggesting a potential protective effect of P . vivax infections against P . falciparum clinical episodes .\",\n \"Analysis to further explore this revealed that a concurrent or recent infection ( i . e . , at the same or preceding follow-up visit ) with P . vivax reduced the odds of a P . falciparum clinical episode by 65% [22-85%] ( p=0 . 011; Table 6 ) .\",\n \"Further indications for a potential interaction between the two species was also observed when analyzing P . falciparum parasite densities , which were reduced by 55% [19–75%] ( p=0 . 008; Supplementary file\",\n \"3 ) in mixed P . falciparum/P .\",\n \"vivax infections compared to P . falciparum single infections , indicative of suppression of one of the species in mixed infections .\"\n ],\n [\n \"In the present study , we describe striking heterogeneity in malaria transmission not only between closely neighboring communities in Maprik district , PNG , but also substantial differences in exposure between individual children from the same village .\",\n \"On village level this heterogeneity is apparent both when using traditional markers such as prevalence of infection , as well as when using the novel `reference standard` marker of individual exposure molFOB .\",\n \"The increased resolution provided by molFOB further allows quantifying heterogeneity in exposure to new blood-stage infections between individual children .\",\n \"Extending an earlier study in a neighboring area in which younger children had been enrolled , and that had identified molFOB as the most important predictor of malaria clinical episodes ( Mueller et al . , 2012; Koepfli et al . , 2013 ) , we confirmed that molFOB remains significantly associated with the risk for clinical episodes , but other factors such as age ( P . vivax ) , or a mixed Pf/Pv infection and village factors not captured by any of the other parameters assessed ( P . falciparum ) have a stronger effect on the risk for clinical malaria ( Table 5 and 6 ) .\",\n \"Malaria transmission is often estimated by investigating the more accessible human host rather than the mosquito vector ( Tusting et al . , 2014 ) .\",\n \"Because P . falciparum blood-stage infections are a direct outcome of mosquito-to-human transmission , infection parameters assessed in the human blood closely reflect P . falciparum transmission .\",\n \"In contrast , relapses arising from dormant hypnozoites contribute substantially to P . vivax blood-stage infections ( Robinson et al . , 2015 ) , thus complicating the assessment of mosquito-to-human P . vivax transmission via infection parameters measured in the human blood .\",\n \"The unique design of this study , that combined clearance of hypnozoites in half of the study participants with subsequent measurement of Pv-molFOB , allowed us to identify the burden of P . vivax infections due to mosquito-to-human transmission ( in hypnozoite-cleared individuals ) and compare it to the total burden of P . vivax infections .\",\n \"We found a highly similar incidence and comparable temporal and spatial heterogeneity of P . falciparum and P . vivax infection acquired through renewed exposure to infected mosquito bites .\",\n \"In children that experienced the full burden of relapses we found two-fold higher P . vivax infection prevalence and 4-times higher incidence ( molFOB ) compared to P . falciparum .\",\n \"This first quantitative comparative assessment of P . falciparum and P . vivax transmission using non-entomological molecular parameters thus indicates that the observed differences in epidemiology between the two species are largely due to the high burden of relapsing P . vivax blood-stage infections ( Robinson et al . , 2015 ) .\",\n \"Our molecular results thus support recent entomological data from Dreikikir district , 50 km from Maprik in East Sepik Province ( Reimer et al . , 2016 ) as well as earlier studies in East Sepik ( Hii et al . , 2001 ) that found similar sporozoite rates for P . falciparum and P . vivax .\",\n \"Our previous analysis of this cohort had investigated the contribution of the hypnozoite reservoir to P . vivax infection and disease in order to inform strategies for achieving a sustained reduction of the P . vivax burden in PNG ( Robinson et al . , 2015 ) .\",\n \"Here , we now describe the post-treatment re-infection dynamics in higher temporal detail .\",\n \"Through a detailed comparison of these patterns for P . vivax and P . falciparum in PQ and placebo-treated children we further elucidate the contribution of relapses to P . vivax prevalence and clinical incidence , which are the most commonly used parameters for planning and monitoring of malaria control strategies .\",\n \"P . vivax relapses accounted for more than half of the observed P . vivax prevalence in this cohort , which is lower than what was previously estimated as the contribution of relapses towards Pv-molFOB by comparison of treatment arms ( 77% , Robinson et al . , 2015 ) .\",\n \"This difference can be accounted for by the higher number of P . vivax multiple clone infections that will accumulate more rapidly in the placebo-arm , where additional parasite clones from relapses and/or new infections may overlap with without a corresponding change in overall prevalence .\",\n \"While we previously described a sustained effect of PQ treatment with significant reductions in Pv-molFOB observed up to eight months post treatment ( Robinson et al . , 2015 ) , here , we describe temporal variation in relapse rate with a rapid and wave-like recurrence of P . vivax in children from the placebo arm , who had retained their hypnozoites ( Figure 1A ) .\",\n \"The concurrent , modest peaks in Pf-molFOB and Pv-molFOB in the PQ arm represent seasonal variation in transmission , which is highest in December and January in the study area ( Mueller et al . , 2012 ) ; corresponding to weeks 8–14 of follow-up ) .\",\n \"A much higher peak and subsequent drop in appearance of new clones within three months after blood-stage only treatment , which was mirrored by a corresponding peak and drop in P . vivax prevalence ( Figure 1B andC ) , suggests that the incidence of relapse infections in the blood was not constant during follow-up .\",\n \"P . vivax infections are often observed following treatment of P . falciparum malaria ( Douglas et al . , 2011 ) , and it can thus been hypothesized that the frequency of relapses may be temporarily increased after blood-stage antimalarial treatment ( White and Imwong , 2012 ) .\",\n \"It is thus conceivable that the blood-stage antimalarial at baseline either triggered P . vivax relapses directly , or indirectly by allowing more hypnozoites to establish blood-stage infections in parasite-free hosts .\",\n \"Alternatively , blood-stage P . vivax infections from hypnozoites relapsing shortly after baseline treatment ( during a period when antimalarial drugs were present at sub-curative levels ) may be suppressed to sub-detectable densities until complete waning of drug levels , resulting in simultaneous proliferation and detection of many new blood-stage clones within the first weeks after treatment ( Douglas et al . , 2011; Tarning et al . , 2014 ) .\",\n \"More detailed modeling of the dynamics of individual P . vivax blood-stage infections and their association with potential triggers such as treatment or febrile illness will be required to determine the existence and importance of proposed relapse-triggers .\",\n \"Malaria transmission showed high micro-spatial heterogeneity with more than 10-fold differences in Pf- and Pv-molFOB ( in the PQ arm ) between villages despite an overall high LLIN use by the study participants ( during follow-up; village average use , >90%; individual use , >50% ) .\",\n \"Individual LLIN use at enrolment was nevertheless associated with a reduced risk of recurrent P . falciparum and P . vivax in univariate analyses .\",\n \"However , after adjustment for other related variables ( i . e . , village of residence or infection status ) this association became non-significant .\",\n \"Children living in Bolumita , where both P . falciparum and P . vivax molFOB and prevalence were highest , had a modestly lower LLIN use ( mean during follow-up , 92%; at enrolment , 77% ) compared to children from other villages ( mean during follow-up , 97–100%; at enrolment , 91–100% ) .\",\n \"It is conceivable that LLIN use in the Bolumita community may be less effective in reducing malaria transmission ( Killeen et al . , 2007; Smith et al . , 2009 ) .\",\n \"Potential differences between villages in mosquito density , behavior , sporozoite rate , proximity of house or play areas to mosquito breeding sites , or human behavioral factors ( related to LLIN use or other risk factors ) are however likely to be more important determinants for exposure to infective bites .\",\n \"Small-scale variations in vector species and distribution between and within villages in PNG have been described previously ( Cattani et al . , 1986; Reimer et al . , 2016; Charlwood et al . , 1986; Hii et al . , 1997; Burkot et al . , 1988 ) and likely account in a large part for the micro-geographic heterogeneity in malariological parameters observed in this and other studies .\",\n \"Assessing the incidence of new infections from consecutive blood samples using molecular methods ( as is necessary to determine molFOB ) , is complicated by fluctuating densities of clonal parasitemia that may temporarily fall below the limit of detection of the genotyping PCR , leading to imperfect detectability of clones ( Bretscher et al . , 2010; Felger et al . , 2012; Koepfli et al . , 2011 ) .\",\n \"For P . falciparum , periodical sequestration of clones and absence from the peripheral blood at time of sampling may further contribute to imperfect detectability .\",\n \"For P . vivax , generally low parasite densities aggravate the problem of imperfect detectability , and dis- and re-appearance of clones may be a result of imperfect detectability or relapsing hypnozoites .\",\n \"The overall estimates of molFOB presented here may thus be biased .\",\n \"Accurately assessing the effects of this imperfect detectability on parameters estimated from longitudinal genotyping data , such as molFOB , requires complex mathematical modeling ( Bretscher et al . , 2010; Felger et al . , 2012; Sama et al . , 2005; Sama et al . , 2006 ) .\",\n \"However , although clonal detectability has been shown to decrease with age ( Felger et al . , 2012; Sama et al . , 2006 ) and MOI ( Koepfli et al . , 2011 ) , it is unlikely to vary substantially within our cohort’s age range and transmission setting .\",\n \"Hence the observed differences in molFOB are likely to accurately reflect the relative differences in individual exposure as well as in population transmission levels within the study area .\",\n \"Evaluating the impact of malaria control efforts requires monitoring changes in malariological metrics over extended periods of time .\",\n \"Drawing comparisons between studies performed at different times in different age groups is particularly challenging because of the interplay of past and current exposure to infective bites and the resulting anti-malarial immunity in the study population of a certain age .\",\n \"In our cohort , fewer P . vivax clinical episodes than P . falciparum clinical episodes were detected despite a higher incidence of P . vivax blood-stage infections , which is consistent with earlier studies in children of similar age ( Michon et al . , 2007 ) .\",\n \"The very low incidence of clinical P . vivax episodes in our cohort , at 0 . 16 clinical episodes/year ( placebo arm [Robinson et al . , 2015] ) , contrasts drastically with that of 2 . 46 P . vivax clinical episodes/year observed in an earlier observational cohort of younger children aged 1–4 years from the same area ( Lin et al . , 2010 ) .\",\n \"The 3-fold difference in Pv-molFOB between the two cohorts seems modest when compared to the 15-fold difference in the incidence of clinical Pv episodes ( Koepfli et al . , 2013 ) .\",\n \"This suggests that the much lower incidence of P . vivax clinical illness in 5–10 years old children of this study is more likely explained by an advanced state of immunity to P . vivax compared to the younger children of the earlier cohort than the drop in P . vivax transmission .\",\n \"Consistently , age emerged as the strongest factor associated with protection against P . vivax clinical episodes , Pv-molFOB and P . vivax parasite density .\",\n \"Like in the previous cohort of younger children from neighboring villages ( Lin et al . , 2010 ) the incidence clinical P . vivax clinical episodes dropped significantly with age .\",\n \"Unlike in the previous cohort of younger children ( Koepfli et al . , 2013 ) , in this cohort we additionally observed a drop in Pv-molFOB as well as P . vivax densities with age ( Table 4 and Supplementary file 3 ) .\",\n \"This is a further indication of the substantial clinical immunity to P . vivax acquired during years of past exposure in the children of this study , which is still ongoing after the age of five .\",\n \"In sharp contrast to the age-dependent decline of P . vivax clinical episode incidence , no age-dependent decrease in the incidence of clinical episodes was observed for P . falciparum .\",\n \"In a previous cohort study conducted in 2004 in 5–14 year old children in an area from PNG with substantially higher transmission levels ( mean incidence risk 5 . 0 versus 0 . 8 infections/year , Michon et al . , 2007; Robinson et al . , 2015 ) , the risk of moderate- to high-density P . falciparum infections decreased significantly with age ( Michon et al . , 2007 ) .\",\n \"Clinical immunity to P . falciparum in children of this earlier cohort was not only significantly further advanced compared to children of this cohort , but in addition , transmission was more homogeneous in the area of that study .\",\n \"As a consequence age was a much better marker of life-time exposure and thus immune status compared to the present cohort .\",\n \"In this cohort , exposure to P . falciparum infections was highly heterogeneous between study participants .\",\n \"Mathematical modeling suggests that at heterogeneous transmission , changes of parasite prevalence and clinical episode incidence with age are less pronounced compared to settings with homogeneous transmission ( Ross and Smith , 2010 ) .\",\n \"Although no age trends were observed for P . falciparum in this cohort , when children were stratified into groups ranging from low to high exposure we found that the proportion of P . falciparum clinical episodes relative to new infections decreased with increasing exposure .\",\n \"This could either reflect the development of clinical immunity in highly exposed children , or premunition , a proposed mechanism by which established infections help to control superinfections by immunological cross-protection ( Sergent and Parrot , 1935; Smith et al . , 1999 ) .\",\n \"In settings of decreasing and heterogeneous transmission , age alone may therefore not be a suitable marker of immunity to P . falciparum .\",\n \"Instead , combining age and molFOB to estimate cumulative life-time exposure may provide a more accurate surrogate measure of the extent of acquired clinical immunity .\",\n \"With P . falciparum transmission declining in PNG due to successful malaria control strategies ( Koepfli et al . , 2015 ) , it is conceivable that immunity against P . falciparum will develop more slowly , shifting the burden of disease towards older age groups or towards more complex , non-linear age patterns .\",\n \"This delay in immune acquisition is however more than compensated by the overall much lower incidence of clinical malaria clinical episodes: in cohort studies in children younger than 4 years from Maprik district , clinical P . falciparum incidence had dropped from 2 . 56 clinical episodes/year before ( observational cohort , [Lin et al . , 2010; Mueller et al . , 2012] ) to 0 . 67 clinical episodes/year immediately after the free LLIN distribution campaign ( placebo arm , [Betuela et al . , 2012] ) .\",\n \"Finally , given that four Plasmodium species co-exist in PNG , there has long been considerable interest in potential mechanisms of cross-species immunity and mixed species interactions ( Mueller et al . , 2009a; Bruce et al . , 2000; Smith et al . , 2001; Mehlotra et al . , 2000 ) .\",\n \"However , there is so far no consistent evidence for the presence or absence of cross-protection among Plasmodium species .\",\n \"P . vivax and P . falciparum infections in our study were concentrated in the same children and villages ( Figure 3 ) , likely due to overlapping focal transmission for P . falciparum and P . vivax and thus potentially high co-infection rates in mosquitoes .\",\n \"Contrary to an earlier cohort in younger PNG children that found a decreased risk of P . falciparum clinical episodes after PQ radical cure ( Betuela et al . , 2012 ) , we found indications for an increased risk of P . falciparum illness after clearance of P . vivax hypnozoites using PQ .\",\n \"Our data suggests that in individuals with substantial clinical immunity against P . vivax , a concurrent P . vivax infection may provide protection against P . falciparum clinical episodes by limiting P . falciparum densities ( Table 6 , Supplementary file 3 ) .\",\n \"However , the comparably small number of clinical episodes in this and the earlier contrasting study does not allow an in-depth analysis of causal relationships and therefore does not allow firm conclusions on the potential effects and mechanisms of cross-species interactions in mixed infections .\",\n \"In conclusion , this study provides detailed insight into the changing epidemiology of malaria in PNG children under sustained malaria control , by using molFOB as a powerful measure to quantitatively investigate patterns of new mosquito-derived P . falciparum and P . vivax infections versus those for P . vivax relapsing infections , as well as spatial and age trends in exposure to these infections .\",\n \"Striking heterogeneity in malaria transmission between villages as well as in individual exposure to new P . falciparum and P . vivax infections persisted in our study area despite very high use of LLINs .\",\n \"This presents a significant challenge for on-going malaria control efforts .\",\n \"The comparable patterns of new mosquito-derived P . falciparum and P . vivax infections indicate that sustained use of LLINs does result in a comparable reduction in transmission of both species .\",\n \"The higher incidence and prevalence of P . vivax infections observed in our data is thus directly linked to its ability to cause relapsing infections , highlighting the crucial role of hypnozoites for P . vivax epidemiology and the need to effectively intervene against these hidden stages .\",\n \"Together , these insights provide a crucial link to evaluate the level of P . vivax mosquito-based transmission against that of P . falciparum and serve to calibrate other standard malaria indicators such as parasite prevalence or incidence of clinical episodes and to ultimately inform new approaches to surveillance and response systems .\"\n ],\n [\n \"This study was conducted in six villages in the Albinama and Balif areas , Maprik district , East Sepik Province , PNG between August 2009 and May 2010 .\",\n \"The area is serviced by the Albinama health sub-center , Balif aid post and a network of health workers in all study villages .\",\n \"The study design has been described in detail elsewhere ( Robinson et al . , 2015 ) .\",\n \"Briefly , 524 children aged 5–10 years whose parents provided written informed consent for their participation were enrolled and randomized to receive either chloroquine ( CQ , days 1–3 , total dose 25 mg/kg ) , artemeter-lumefantrine ( Coartem , AL , days 11–13 , 2 mg/kg A , 12 mg/kg L ) and primaquine ( PQ , days 1–20 , 0 . 5 mg/kg/day ) ; or CQ ( days 1–3 ) , AL ( days 11–13 ) , and placebo ( days 1–20 ) over 20 days of directly observed treatment ( DOT1-20 ) in a double-blinded manner .\",\n \"Children were actively visited and examined for signs and symptoms of malaria fortnightly at their schools for 8 months .\",\n \"In addition , passive surveillance was provided by the local health centre , aid post and village health workers throughout the study period .\",\n \"Finger-prick blood samples ( 250 µl ) were collected at fortnightly active-follow-up visits in the first 12 weeks and monthly thereafter , as well as from symptomatic children detected during active or passive morbidity surveillance .\",\n \"Symptomatic children were tested for malaria infection with rapid diagnostic test ( RDT , CareStartMalaria pLDH/HRP2 Combo , AccessBio , USA ) , and only RDT and or LM-confirmed Plasmodium infections of any density were treated with a 3 day course of AL .\",\n \"Household , village and health facility location data was collected using a handheld GPS receiver ( Garmin GPSmap62sc ) and maps were prepared using ArcGIS 10 . 2 ( Esri , USA ) .\",\n \"The study received ethical clearance from the PNG IMR Institutional Review Board ( 0908 ) , the PNG Medical Advisory Committee ( 09 . 11 ) , the Ethics Committee of Basel 237/11 and was conducted in full concordance with the Declaration of Helsinki .\",\n \"The study was registered on ClinicalTrials . gov ( NCT02143934 ) .\",\n \"All blood samples were examined by LM and qPCR for detection and speciation of Plasmodium infections as described earlier ( Robinson et al . , 2015 ) .\",\n \"Each blood slide was read independently by two skilled microscopists and re-read by an expert microscopist in case of discrepancies in positivity , speciation or density ( ≥2 x log10 difference ) .\",\n \"Thick blood films were examined by LM for 200 fields ( 1000x magnification ) before being declared parasite-negative .\",\n \"Parasite density was converted from the number of parasites per 200–500 white blood cells ( WBC ) to parasites/µl assuming 8000 WBC/µl ( WHO malaria microscopy training guide ) and calculated as the geometric mean of all positive reads .\",\n \"DNA was extracted from the red blood cell pellet using the FavorPrep 96-well genomic DNA extraction kit ( Favorgen ) .\",\n \"Samples carrying any Plasmodium spp .\",\n \"infection were identified using a generic qPCR ( Wampfler et al . , 2013 ) and positives were subsequently tested in species-specific qPCRs ( Rosanas-Urgell et al . , 2010; Wampfler et al . , 2013 ) .\",\n \"All qPCRs targeted the small subunit ( 18S ) ribosomal RNA gene and were performed as simplex ( P . vivax and P . falciparum ) or duplex qPCR ( P . malariae , P . ovale ) .\",\n \"The concentration of target copies per µl of DNA was determined relative to a dilution row of standard plasmid as previously described ( Rosanas-Urgell et al . , 2010 ) .\",\n \"The qPCR limit of detection ( LOD ) was determined using a standard plasmid dilution row and defined as the last point with more than 50% of replicates positive .\",\n \"The LOD was 2 target copies/µl DNA , equaling 4 target copies/reaction , for all qPCRs .\",\n \"All samples that crossed the fluorescence threshold were scored as positive for species-specific qPCRs .\",\n \"In all samples positive in P . falciparum and/or P . vivax qPCRs , individual parasite clones were distinguished by genotyping the length-polymorphic Pf-msp2 or Pv-msp1F3 marker genes using capillary electrophoresis for highly precise fragment sizing ( Koepfli et al . , 2013; Koepfli et al . , 2011; Falk et al . , 2006; Schoepflin et al . , 2009 ) .\",\n \"MOI was determined by counting the number of detected Pf-msp2 or Pv-msp1F3 alleles per sample .\",\n \"molFOB was calculated from the number of new parasite clones detected per child or per sampling interval in the peripheral blood , divided by the individual time at risk or length of the interval .\",\n \"A new infection was defined as a Pf-msp2 or Pv-msp1F3 allele not present in the two preceding genotyping-positive samples collected during active or passive surveillance ( Figure 1—figure supplement 1 ) .\",\n \"Imperfect diagnostic detectability was not further adjusted for .\",\n \"Children were considered at risk for clinical malaria clinical episodes until the end of the study or until they were censored ( on the last visit before two consecutively missed scheduled follow-up visits [Robinson et al . , 2015] ) .\",\n \"For clinical endpoints , time-at-risk ( TAR ) was not further adjusted for interim missed follow-up visits because the intense active and passive case detection presumably led to detection of all malaria clinical episodes .\",\n \"In contrast , TAR for analysis of molecular data ( e . g . , molFOB ) was reduced by the duration of the missed interval if a child was not seen by the study team for six weeks or more ( ≥42 days ) .\",\n \"Children with a TAR of less than 3 months ( <84 days ) were excluded .\",\n \"This resulted in an analyzed population of 466 children ( characterized in Table 1 ) of which 430 ( 92 . 3% ) completed the whole follow-up period , with a median of 15 ( IQR: 13–17 ) study contacts and mean TAR of 186 days ( IQR 168–223 days ) .\",\n \"Time to first Plasmodium infection by qPCR and LM and its association with covariates were modeled using Cox regression , and the proportional hazards assumption was checked using the test based on the Schoenfeld residuals .\",\n \"Multiple failure Cox regression was used to model the time to P . vivax and P . falciparum clinical episodes .\",\n \"For statistical analysis , a malaria clinical episode was defined as fever ( >37 . 5°C axillary ) plus the presence of LM-detectable parasites , irrespective of RDT result or antimalarial treatment during the field visit .\",\n \"Negative binomial generalized estimating equations ( GEE ) with log link function using an exchangeable correlation structure were used to model incidence of new infections with P . falciparum and P . vivax per sampling interval .\",\n \"For these analyses , the time at risk was restricted to the intervals where molFOB could be estimated ( i . e . , starting in the third follow-up interval ) .\",\n \"A binomial GEE with logit link function using an exchangeable correlation structure was used to model the odds of a P . falciparum clinical episode per interval .\",\n \"Gaussian GEEs with log link function using an exchangeable correlation matrix were used to model log-transformed qPCR parasite densities in qPCR-positive samples , measured as 18S rRNA copy numbers/µl blood .\",\n \"In the GEE and Cox models where molFOB was a covariate , it was included as a time-varying covariate .\",\n \"When modelling molFOB ( Table 4 ) and the odds of clinical episodes ( Table 6 ) using GEEs , molFOB was calculated for each follow-up interval and used as predictor .\",\n \"In Cox models investigating the risk of clinical episodes ( Table 5 ) , molFOB was calculated based on the new infections up to the time of failure and used as predictor .\",\n \"In exploratory preliminary analyses we tested for a wide variety of interactions between covariates including interactions between all combinations of molFOB , enrolment infection status , age , village and bednet-usage .\",\n \"All analyses were done using STATA v14 and R . Maps were drawn using Arcgis 10 . 1 ( Esri Inc . ) .\",\n \"Ordinary kriging was used to generate the contour maps .\",\n \"Semivariograms were used as the mathematical forms used to express autocorrelation .\",\n \"Input variables for the spatial models were\",\n \"( i ) molFOB ( prediction of independent variable , molFOB ) using the model presented in Table 4 ( negative binomial GEE ) for Pf and Supplementary file 4 for Pv ( same as that shown in Table 4 but with primaquine and placebo arms combined ) , resulting in Figure 3 Panels A and B;\",\n \"( ii ) relative risk of clinical episodes as predicted by the model shown in Table 5 , resulting in Panels C and D of Figure 3 .\",\n \"Relative standard error maps were generated by dividing the absolute standard error map by the model prediction map .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["The molecular force of blood-stage infection ( molFOB ) is a quantitative surrogate metric for malaria transmission at population level and for exposure at individual level .","Relationships between molFOB , parasite prevalence and clinical incidence were assessed in a treatment-to-reinfection cohort , where P . vivax ( Pv ) hypnozoites were eliminated in half the children by primaquine ( PQ ) .","Discounting relapses , children acquired equal numbers of new P . falciparum ( Pf ) and Pv blood-stage infections/year ( Pf-molFOB = 0–18 , Pv-molFOB = 0–23 ) resulting in comparable spatial and temporal patterns in incidence and prevalence of infections .","Including relapses , Pv-molFOB increased >3 fold ( relative to PQ-treated children ) showing greater heterogeneity at individual ( Pv-molFOB = 0–36 ) and village levels .","Pf- and Pv-molFOB were strongly associated with clinical episode risk .","Yearly Pf clinical incidence rate ( IR = 0 . 28 ) was higher than for Pv ( IR = 0 . 12 ) despite lower Pf-molFOB .","These relationships between molFOB , clinical incidence and parasite prevalence reveal a comparable decline in Pf and Pv transmission that is normally hidden by the high burden of Pv relapses .","Clinical trial registration: ClinicalTrials . gov NCT02143934"],"string":"[\n \"The molecular force of blood-stage infection ( molFOB ) is a quantitative surrogate metric for malaria transmission at population level and for exposure at individual level .\",\n \"Relationships between molFOB , parasite prevalence and clinical incidence were assessed in a treatment-to-reinfection cohort , where P . vivax ( Pv ) hypnozoites were eliminated in half the children by primaquine ( PQ ) .\",\n \"Discounting relapses , children acquired equal numbers of new P . falciparum ( Pf ) and Pv blood-stage infections/year ( Pf-molFOB = 0–18 , Pv-molFOB = 0–23 ) resulting in comparable spatial and temporal patterns in incidence and prevalence of infections .\",\n \"Including relapses , Pv-molFOB increased >3 fold ( relative to PQ-treated children ) showing greater heterogeneity at individual ( Pv-molFOB = 0–36 ) and village levels .\",\n \"Pf- and Pv-molFOB were strongly associated with clinical episode risk .\",\n \"Yearly Pf clinical incidence rate ( IR = 0 . 28 ) was higher than for Pv ( IR = 0 . 12 ) despite lower Pf-molFOB .\",\n \"These relationships between molFOB , clinical incidence and parasite prevalence reveal a comparable decline in Pf and Pv transmission that is normally hidden by the high burden of Pv relapses .\",\n \"Clinical trial registration: ClinicalTrials . gov NCT02143934\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Malaria is caused by five different species of parasites that are transmitted to humans by bites from parasite-carrying mosquitos .","Once in human blood , the parasites rapidly multiply .","People who live in countries where malaria is common may become infected and never show any symptoms because their immune systems are able to keep parasite numbers low .","Repeated infections , or infection with more than one species of malaria parasite also are common .","Some species of malaria , including Plasmodium vivax , can hibernate in the liver for weeks or months after the infection and only become active later .","Asymptomatic infections , multi-parasite infections , and reactivating parasites make it hard to measure how often new malaria infections occur .","One way scientists can determine if a new infection has occurred is by genotyping the parasites in a person’s blood .","Genotyping involves looking for small differences in the parasite DNA .","For example , a study in Papua New Guinea , where P . vivax is very common , showed that reactivations of hibernating parasites were more common than new infections .","Now , Hofmann et al . use the same study in Papua New Guinea to compare the frequency and consequences of new infections with P . vivax and another malaria parasite , Plasmodium falciparum .","In the study , 466 children from 6 villages were followed for 8 months with tests every 2 to 4 weeks to genotype the parasites in their blood .","Some of the children were treated with antimalarial drugs to help wipe out any existing parasites including hibernating ones .","While P . vivax was about twice as common in blood samples—likely due to reactivation—genotyping showed that new infections with the two parasites occur at equal rates and often at the same times and locations .","Hofmann et al . also showed that some villages and some children had much higher rates of infection than others .","This difference could not fully be explained by use of bednets or other preventive measures .","Children were more likely to become ill from P . falciparum than P . vivax even though P . vivax was more common .","But children with more frequent infections with P . falciparum seemed better able to manage the parasites and were less likely to develop symptoms that those with infrequent infections .","The experiments show that genotyping may help scientists better track new malaria infections and develop better strategies to prevent or treat malaria ."],"string":"[\n \"Malaria is caused by five different species of parasites that are transmitted to humans by bites from parasite-carrying mosquitos .\",\n \"Once in human blood , the parasites rapidly multiply .\",\n \"People who live in countries where malaria is common may become infected and never show any symptoms because their immune systems are able to keep parasite numbers low .\",\n \"Repeated infections , or infection with more than one species of malaria parasite also are common .\",\n \"Some species of malaria , including Plasmodium vivax , can hibernate in the liver for weeks or months after the infection and only become active later .\",\n \"Asymptomatic infections , multi-parasite infections , and reactivating parasites make it hard to measure how often new malaria infections occur .\",\n \"One way scientists can determine if a new infection has occurred is by genotyping the parasites in a person’s blood .\",\n \"Genotyping involves looking for small differences in the parasite DNA .\",\n \"For example , a study in Papua New Guinea , where P . vivax is very common , showed that reactivations of hibernating parasites were more common than new infections .\",\n \"Now , Hofmann et al . use the same study in Papua New Guinea to compare the frequency and consequences of new infections with P . vivax and another malaria parasite , Plasmodium falciparum .\",\n \"In the study , 466 children from 6 villages were followed for 8 months with tests every 2 to 4 weeks to genotype the parasites in their blood .\",\n \"Some of the children were treated with antimalarial drugs to help wipe out any existing parasites including hibernating ones .\",\n \"While P . vivax was about twice as common in blood samples—likely due to reactivation—genotyping showed that new infections with the two parasites occur at equal rates and often at the same times and locations .\",\n \"Hofmann et al . also showed that some villages and some children had much higher rates of infection than others .\",\n \"This difference could not fully be explained by use of bednets or other preventive measures .\",\n \"Children were more likely to become ill from P . falciparum than P . vivax even though P . vivax was more common .\",\n \"But children with more frequent infections with P . falciparum seemed better able to manage the parasites and were less likely to develop symptoms that those with infrequent infections .\",\n \"The experiments show that genotyping may help scientists better track new malaria infections and develop better strategies to prevent or treat malaria .\"\n]","isConversational":false},"year":{"kind":"string","value":"2017"}}},{"rowIdx":98,"cells":{"headings":{"kind":"list like","value":["Introduction","Results and discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results and discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["cell biology","physics of living systems"],"string":"[\n \"cell biology\",\n \"physics of living systems\"\n]","isConversational":false},"title":{"kind":"string","value":"Cortical flow aligns actin filaments to form a furrow"},"id":{"kind":"string","value":"elife-17807-v3"},"sections":{"kind":"list like","value":[["Cytokinesis begins in late anaphase , triggered by regulatory pathways with feedback from the mitotic spindle that ensure appropriate positioning of the molecular machinery ( reviewed in Fededa and Gerlich , 2012; Green et al . , 2012 ) .","As a consequence , the small regulatory GTPase RhoA ( Piekny et al . , 2005; Tse et al . , 2012 ) and the molecular motor myosin ( Shelton et al . , 1999; Werner and Glotzer , 2008 ) are enriched in an equatorial band ( Piekny et al . , 2005; Werner and Glotzer , 2008 ) .","Myosin can directly organize the actin network ( Miller et al . , 2012; Soares e Silva et al . , 2011; Vavylonis et al . , 2008 ) , and if myosin-based active alignment ( e . g . via search-and-capture [Vavylonis et al . , 2008] ) or flow-based compression as proposed by White and Borisy ( White and Borisy , 1983; Rappaport , 1996 ) drive furrow formation and cytokinesis in eukaryotes remains unclear ( Green et al . , 2012; Bray and White , 1988; Levayer and Lecuit , 2012; Mendes Pinto et al . , 2013; Cao and Wang , 1990; Murthy and Wadsworth , 2005; Zhou and Wang , 2008 ) ."],["We set out to address this question in the one cell embryo of the nematode C . elegans which forms two constricting ingressions , first a pseudocleavage furrow during polarity establishment , and then a cytokinetic furrow during cytokinesis ( Figure 1—figure supplement 1a , Video 1 ) ( Munro and Bowerman , 2009; Rose et al . , 1995 ) .","Notably , the pseudocleavage furrow is linked to cortical flow ( Mayer et al . , 2010; Munro et al . , 2004 ) and lacks regulatory control from the mitotic spindle ( Werner and Glotzer , 2008 ) .","We first characterized the actomyosin cortical network during pseudocleavage and cytokinesis .","For this , we generated a Lifeact::mKate2 fluorescent line of C . elegans by coupling the Lifeact peptide to the far-red fluorophore mKate2 that was codon optimized for C . elegans ( Redemann et al . , 2011 ) ( see Materials and methods and Video 2 ) .","This allows us to quantify cortical actin filament orientation in space and time , as well as cortical flow velocity fields by Particle Image Velocimetry ( PIV , Figures 1 and 2 , Figure 2—figure supplement 1 ) ( Mayer et al . , 2010 ) .","We find that actin filaments change from a randomly oriented and isotropic gel before cortical flows initiate , into a more circumferentially aligned organization at the future site of furrow ingression during both pseudocleavage and cytokinesis ( Figure 1b , Figure 1—figure supplement 1 and Videos 1 and 3 ) .","We studied how pseudocleavage and cytokinesis differ in the spatial distribution of molecular regulation and contractility within the equatorial region ( Figure 1 ) .","Non-muscle myosin II ( NMY-2 ) is the essential molecular motor that drives both cortical flow and cytokinesis ( Shelton et al . , 1999; Guo and Kemphues , 1996 ) , whereas the small GTPase RhoA ( Piekny et al . , 2005; Tse et al . , 2012 ) , when active , is responsible for its local activation .","At the onset of cytokinesis , active RhoA and NMY-2 locally increase at the position of the contractile ring , while for pseudocleavage we observed no such increase ( Figure 1 ) .","Instead , the equatorial region in pseudocleavage corresponds to a transition zone with a gradual decrease of RhoA and myosin between the anterior region of high and the posterior region of low contractility ( Figure 1 ) ( Mayer et al . , 2010 ) .","The actin nucleator formin CYK-1 is downstream of RhoA ( Piekny et al . , 2005 ) and follows a similar pattern of localization , and actin bundling via anillin ( Piekny and Glotzer , 2008 ) or plastin also does not appear to be increased in the equatorial region during pseudocleavage ( Figure 2—figure supplement 5e–f ) .","To conclude , the pseudocleavage furrow ingresses with a circumferential alignment of filaments in the gel , but without a local zone of RhoA activation and the corresponding local increase in myosin and formin density .","Since search-and-capture like mechanisms require a localized band of myosin and actin nucleators at the equator to drive myosin-based active alignment ( Vavylonis et al . , 2008 ) , this suggests that actin filaments in pseudocleavage align via flow-based compression and not via active alignment . 10 . 7554/eLife . 17807 . 003Video 1 . Actomyosin gel dynamics in the C . elegans zygote . Cortical and medial planes of an embryo expressing both Lifeact::mKate2 and endogenous NMY-2::GFP .","Left panel , medial plane Lifeact:mKate2 , center , cortical NMY-2::GFP , right panel , cortical Lifeact:mKate2 ( min:s ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00310 . 7554/eLife . 17807 . 004Video 2 . Three-dimensional reconstruction of an early one-cell embryo of C . elegans expressing Lifeact::mKate2 . The back of the embryo is dimmer due to imaging . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00410 . 7554/eLife . 17807 . 005Figure 1 . Flow and ingression during pseudocleavage and cytokinesis .","( a ) Average active RhoA biosensor intensity profiles along the AP axis ( 0 represents the embryo center; N = 14 embryos for pseudocleavage and N = 7 for cytokinesis; errors represent the standard error of the mean ) .","( b ) Actomyosin cortical organization at the onset of pseudocleavage and cytokinesis in embryos expressing both Lifeact::mKate2 and NMY-2::GFP .","( c–d )","Average actin , myosin and ingression profiles as a function of position along the AP axis ( N = 10 embryos for pseudocleavage and N = 22 for cytokinesis; errors represent the standard error of the mean ) .","Scale bars , 10 μm .","The following figure supplement is available for Figure 1:DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00510 . 7554/eLife . 17807 . 006Figure 1—figure supplement 1 . Details of actin organization during the polarization flow phase .","( a ) Actomyosin gel dynamics in the C . elegans zygote .","Single cortical and medial planes from a timelapse sequence of a representative embryo expressing both Lifeact::mKate2 and NMY-2::GFP .","Anterior is to the left , posterior to the right .","White arrows indicate the position of pseudocleavage and cytokinesis ingressions .","Red arrows indicate flow directions .","( b ) Close up look at the filaments dynamics near the site of flow initiation .","Cortical plane from a timelapse sequence of an embryo expressing both Lifeact::mKate2 and NMY-2::GFP .","( c ) Top: Actin network organization from an embryos expressing Lifeact::mKate2 .","Times rescaled to 0 at flow initiation .","Green arcs indicate the smooth cleared zone during the polarization process .","Red lines schematize filaments maximum vertical alignment positions .","White arrows underlie pseudocleavage ingression .","Bottom: color-coded representation of the orientation of filaments .","( d ) Example from an embryo for which the polarization process starts off centered from the physical pole of the embryo due to a mispositioning of the sperm pronucleus ( dashed circle on medial plane ) .","Alignment direction is observed to follow perpendicularly to the direction of the flow as it reorients along the long axis of the embryo ( also see Appendix ) .","Red arrows represent flow direction , red lines filaments main orientation .","Scale bars , 10 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00610 . 7554/eLife . 17807 . 007Figure 2 . Cortical flows compress the gel to generate alignment .","( a ) Filament orientation is quantified by a nematic order parameter; flow fields are obtained from Particle Image Velocimetry ( PIV ) analysis .","( b ) Gel flow , compression rate , filament orientation and cell ingression fields .","Top row , schematic representations .","Middle row , kymographs of the time evolution of the profiles along the AP axis of flow velocity , compression rate , nematic order parameter and outer ingression distance for pseudocleavage ( N = 10 ) .","Bottom row , cytokinesis ( N = 12 ) .","Orange dashed lines indicate the times where all four fields are stationary .","Flow , compression and nematic order were determined from bottom-section and ingression from mid-section views of Lifeact::mKate2 labelled embryos ( see Materials and methods and Appendix ) .","( c ) Spatial correlation between compression , alignment and ingression ( see Materials and methods ) .","Peak value positions ( estimated by a Gaussian fitting of the 1D correlation curve ) are indicated .","See Figure 2—figure supplement 2 for spatiotemporal correlation functions .","( d ) Table of the temporal delays obtained from the correlation peak value positions using the mean values of the velocity in the 10 μm central region ( mean values for pseudocleavage: v = 6 . 16 ±1 . 23 μm/min; for cytokinesis: v = 4 . 01±1 . 92 μm/min; mean ± SD ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00710 . 7554/eLife . 17807 . 008Figure 2—figure supplement 1 . Quantification methods developed .","( a ) Example of the filaments alignment quantification ( red lines ) and the flow velocity field ( yellow arrows ) , as well as nematic order parameters conventions .","( b ) Time evolution of flow ( light green ) and compression rate ( dark green ) as well as ingression ( grey ) measured at a fixed position along the AP axis ( −20 μm for velocity , 0 μm or central position of the embryo for compression and 3 μm for ingression , measured from the center ) .","Error bars represent the standard error of the mean .","Bottom , fit using the error function ( red line ) of the time evolution of the velocity from a single embryo ( blue dots ) used to determine the reference time point ( Time = 0 ) .","( c ) Time evolution of the mean velocity and compression rate along the AP axis and nematic order parameter , all taken at fixed positions along the AP axis .","Error bars represent the standard error of the mean .","In orange is represented the period chosen as stationary flow phase .","( d ) Mean profiles along the AP axis of the velocities ( vx light green and vy yellow ) , compression rate and nematic order parameters ( Q = −Qxx in dark blue and Qxy in cyan ) during the stationary flow phase .","( See Appendix for further explanations . ) DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00810 . 7554/eLife . 17807 . 009Figure 2—figure supplement 2 . Spatiotemporal correlations . Plots of the spatiotemporal correlations between the compression , alignment and ingression data sets taken during the pseudocleavage stationary phase","( a ) and cytokinesis onset stationary phase","( b ) .","Data is restricted to the 30 μm central region of the embryo . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00910 . 7554/eLife . 17807 . 010Figure 2—figure supplement 3 . Nematic order parameter quantification of artificial images with and without directional bias .","( a ) Example of a simulated isotropic actin meshwork ( see Appendix for details ) .","( b ) Example of a simulated actin meshwork for which the angle of the starting direction is biased toward vertical ( see Appendix ) .","( c ) Cumulative distribution function ( CDF ) used and example of an image and its nematic order map ( red lines indicate the preferential orientation inside a template or orientation vector ) obtained for different values of the starting parameter B that controls the bias of the filaments orientation .","B = 0 leads to an isotropic orientation .","( d ) Variation of the average nematic order parameter obtained for several synthetic meshworks while increasing the vertical bias of the filaments orientation .","The saturation is caused by the fact that we here bias only the initial direction of growth . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01010 . 7554/eLife . 17807 . 011Figure 2—figure supplement 4 . Impact of image quality on the nematic order parameter quantification .","( a ) An artificial filamentous network is obtained by a random iteration process with controllable parameters in Matlab ( see Appendix for more details ) .","( b-d )","We chose to maintain a vertical bias in the orientation of the filaments in all the following cases ( B = 1 ) .","Red lines indicate the preferential orientation inside a template or orientation vector .","A profile of the mean nematic order parameter along the x-axis as well as the angular distribution are shown in all cases .","Note that as we do not measure the directionality of filaments thus we obtain values modulo π .","( b ) Impact of the filament density ( ranging from 60 to 120 filaments per image ) on the nematic order parameter quantification .","The profile of nematic order is to first order independent of network density .","( c ) Impact of the signal to noise ratio .","High signal to noise ratio leads to higher values of the nematic order parameter , low signal to noise ratio leads to a poor detection of the direction of the filaments and artificially small values of the nematic order parameter . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01110 . 7554/eLife . 17807 . 012Figure 2—figure supplement 5 . Comparison of different actin labeling strains and actin-binding protein localization .","( a–c )","Average AP profiles of flow velocity ( light green ) and nematic order parameter ( blue ) at the time of stationary flow during pseudocleavage of respectively 22 embryos expressing Lifeact:GFP , NMY:RFP , 9 embryos expressing Lifeact:RFP , NMY-2:GFP and 16 embryos expressing Lifeact:mKate2 , NMY-2:GFP .","Overlaid in red is the velocity profile of worms expressing only labeled NMY-2 ( N = 5 ) .","Error bars represent the standard error of the mean .","( d ) Images of NMY-2:GFP in the early zygote .","Left strain without Lifeact , right strain expressing Lifeact:mKate2 .","The density , size and dynamics are similar in both strains .","Right panel , the average radial autocorrelation coefficient of the myosin intensity over a minute in early embryos gives an average myosin foci size .","( e ) Localization of the actin bundling protein plastin during pseudocleavage and cytokinesis .","Plastin localizes in dense bundles of actin filaments and is particularly abundant in the cytokenetic ring and could act to stabilize actin bundles in these regions .","( f ) Localization of the actin nucleating agent formin during pseudocleavage and cytokinesis .","CYK-1 is present throughout the cell cortex during pseudocleavage , with higher densities in foci regions .","It is recruited to the cell equator during cytokinetic ring formation and ingression .","Scale bars , 10 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01210 . 7554/eLife . 17807 . 013Video 3 . Cortical dynamics at flow initiation . Left panel , cortical NMY-2::GFP , center cortical Lifeact:mKate2 and a zoom is shown in the right panel ( min:s ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 013 We next sought to test if flow-based compression drives actin filament alignment .","To this end , we developed tools for the spatiotemporal quantification of compression by flow , of filament orientation , and of cortex ingression distance within the gel both at the onset of pseudocleavage and cytokinesis ( see Appendix as well as Figure 2a and Figure 2—figure supplement 1 ) .","We find that the flow compression rate along the antero-posterior direction ( AP axis or x axis ) , given by the spatial gradient of the velocity field −∂xv ( Figure 2b ) , increases with time during the very early stages of pseudocleavage .","The compression rate peaks in the central region of the embryo , with a maximum of ~0 . 4 min−1 .","Notably , this peak is stable for several minutes until pseudocleavage flows cease entirely ( Figure 2b ) .","For cytokinesis , we find that the early flow also proceeds in a unidirectional manner from the posterior toward the anterior ( Figure 2b ) .","Similar to pseudocleavage , the compression rate increases with time over the very early stages of cytokinesis , with the compression profile peaking at the center of the embryo with a maximum rate of ~0 . 7 min−1 ( Figure 2b ) .","We characterized the average actin filaments orientation by a nematic order tensor Q ( Figure 2—figure supplement 1a ) .","This analysis relies on the fact that the intensity of the Fourier transformed image encodes geometric characteristics such as its main orientation pattern ( see Appendix ) .","We find that vertical ( orthogonal to the direction of flow ) alignment appears coincidental in space and time with the local increase of the compression rate ( Figure 2b , compare column 2 and 3 ) .","Notably , changing the direction of flow also changes the direction of alignment: off-site sperm entry leads to flows along the short axis of the egg ( Goldstein et al . , 1993 ) and actin filaments still align in the direction determined by compression ( Figure 1—figure supplement 1d , Video 4 and Appendix ) .","Finally , the ingressing furrow forms in the region where filaments are aligned ( Figure 2b , column 4 ) .","Our quantifications reveal remarkable similarities between the compression rate fields , the pattern of alignment , and the ingression distance between pseudocleavage and cytokinesis ( Figure 2b ) , suggesting common mechanisms .","To conclude , actin filament alignment arises at locations of significant compression by flow . 10 . 7554/eLife . 17807 . 014Video 4 . Cortical dynamics of an embryo in which flow was initiated at the side ( bottom ) and far from the posterior pole of the embryo . Actin filament alignment follows flow and compression , thus transitions from a horizontal to a vertical direction ( min:s ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 014 If flow-based compression aligns actin filaments for forming an ingression , we would expect compression to precede or to be concomitant with alignment and ingression .","The 1D crosscorrelations between compression , filament alignment and ingression distance , calculated for the time that flows are essentially stationary ( Figure 2b and Figure 2—figure supplement 1 ) , show that for pseudocleavage both compression rate and alignment peak at about the same location , while the ingression field peaks ~ 3 µm further to the anterior ( Figure 2c and Figure 2—figure supplement 2 for the spatiotemporal crosscorrelations ) .","The situation is similar at cytokinesis , except that the compression field peaks furthest to the posterior .","We next translated the characteristic distances between peaks in stationary flow to temporal delays between events ( see Materials and methods for further explanation ) .","We find that for pseudocleavage , filament alignment is essentially concomitant with compression , while for cytokinesis compression precedes alignment by 0 . 40 ± 0 . 16 min ( Figure 2d ) .","Furthermore , ingression arises approximately 0 . 5 min after compression for pseudocleavage and 0 . 9 min after compression for cytokinesis .","To conclude , filament alignment appears to rapidly follow compression while ingression arises with a delay .","We next sought to provide a physical explanation for flow-based alignment ( Salbreux et al . , 2009 ) .","The cell cortex in this and other systems can be thought of as a gel that forms a thin layer underneath the plasma membrane .","Recent advances in physical theory show that this gel is active and generates the forces that drive cell shape changes ( Salbreux et al . , 2009; Gowrishankar , 2012; Kruse et al . , 2005; Turlier et al . , 2014 ) .","We describe the actin cortical layer as a thin film of a nematic gel under shear and compression by flow ( Salbreux et al . , 2009; Kruse et al . , 2005; Prost et al . , 2015 ) .","In the x-direction , the time evolution of the local nematic order Q depends on advection ( first term on the right hand side in Equation 1 ) , compression in gel flow ( second term ) , a process of relaxation to an isotropic configuration possibly via turnover ( third term ) , and a tendency of filaments to align with each other over space ( last term , see also Figure 3a ) ( 1 ) ∂tQ=−v∂xQ−β2∂xv−1τQ+ℓ2τ∂x2Q . 10 . 7554/eLife . 17807 . 015Figure 3 . Theory predicts an alignment peak .","( a ) Schematic representation of flow-based alignment .","For full theory refer to Appendix .","( b ) Average AP profiles of gel flow ( light green , smoothened ) , compression rates ( dark green ) and nematic order parameter ( blue ) at the time of stationary flow ( Figure 2C ) during pseudocleavage ( left , N = 16 ) and cytokinesis onset ( right , N = 32 ) .","Error bars represent the standard error of the mean .","Dashed red line indicates the respective best-fit theory predictions for the nematic order parameter ( blue ) .","( c ) Table of the material parameters .","See also Figure 3—figure supplement 1 and Table 1 , Table 2 .","( d ) Illustration of the cortical tension related to the ingression distance .","( e ) Outer ingression profiles at the onset of pseudocleavage ( brown ) and cytokinesis ( yellow ) .","The theoretical profiles ( thick dashed red lines: active tension depends on alignment and can be anisotropic; thin dashed lines: active tension is isotropic ) are determined by simultaneous fitting of both the pseudocleavage and cytokinesis datasets with a single fit parameter ( see Appendix ) using the myosin density , flow velocity and nematic order parameter fields . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01510 . 7554/eLife . 17807 . 016Figure 3—figure supplement 1 . Influence of the gel material parameters for the nematic order parameter .","( a ) Nematic order parameter profiles obtained when varying independently each of the fitting parameters τ , βτ , ℓ away from their the best fit values obtained for cytokinesis .","Dashed red line , best fit result ( τ = 2 . 34 min , βτ = 0 . 654 min , ℓ =1 . 69 μm ) .","For each set of parameters τ , βτ , ℓ , a fit is performed for the fitting parameters C1 and C2 only ( see Appendix ) .","( b ) Nematic order parameter profiles obtained when varying the fitting parameter τ while keeping βτ , ℓ fixed and equal to their best-fit values .","Appropriate fits to experimental curve are obtained for all values of τ < 0 . 5 min .","For each value of τ , a fit is performed for the fitting parameters C1 and C2 only . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01610 . 7554/eLife . 17807 . 017Figure 3—figure supplement 2 . Contribution of active alignment to the nematic order parameter profile .","( a ) Average AP profiles of gel flow ( light green ) , compression rates ( dashed dark green ) , nematic order parameter ( blue ) and myosin fluorescence ( magenta ) at the time of stationary flow during pseudocleavage ( left , N = 16 ) and cytokinesis onset ( right , N = 32 ) .","Error bars represent the standard error of the mean .","Dashed red lines indicate the best-fit theory predictions for the nematic order parameter in absence of active alignment , dashed black lines are the fits obtained while including the myosin-based active alignment , with λ′τ=4 . 1 10−8±0 . 54 for pseudocleavage and λ′τ=0 . 34± 0 . 45 for cytokinesis .","( b ) Contribution of the individual terms from Equation 3 in the main text ( red: −β2τ∂xv , green: −τv∂xQ , blue: ℓ2 ∂x2Q , magenta: λ′τ c ( x ) /c0Q ) .","Note that myosin-based active alignment ( magenta ) is insignificant for pseudocleavage and is low for cytokinesis compared to compression-based alignment ( red ) .","The theoretical alignment profile ( sum of all the terms shown in black ) overlays the experimental data points ( dots ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01710 . 7554/eLife . 17807 . 018Figure 3—figure supplement 3 . Ingression through anisotropic stress generation in the aligned gel .","( a ) Top , a representative embryo is used to visualize the time course of ingression .","Bottom , Kymograph obtained from the region specified by the dashed white box in the upper pannel .","The pseudocleavage and cytokinesis ingressions are visible .","Yellow arrow indicates the inner ingression distance during polarizing flow phase , magenta dashed line indicates the ingressing membrane and blue arrow indicates the outer ingression .","Red doted lines represent the position of the eggshell .","Scale bar 10 μm .","( b ) Top , illustration of the cortical tension during ingression .","Bottom , outer ingression profiles ( grey ) averaged over the time periods indicated by the thin horizontal white lines in ( a , bottom ) during pseudocleavage ( left ) and cytokinesis ( right ) .","The theoretical profiles ( dashed red lines ) are determined by simultaneous fitting of both the pseudocleavage and cytokinesis datasets with a single fit parameter ( see Appendix ) using the myosin density ( magenta ) , flow velocity ( green ) and nematic order parameter ( blue ) fields .","( c ) and","( d ) Illustration of the coordinate and reference systems used for this theoretical description .","( e ) Comparison of the theoretically obtained profile of the outer ingression without the anisotropic stress induced by nematic order ( dashed red , obtained by imposing p3 = 0 ) during pseudocleavage and cytokinesis phases with the experimental data ( black ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01810 . 7554/eLife . 17807 . 019Figure 3—figure supplement 4 . Shape and ingression distance quantification . Left column , plots of the cell shape ( mean radius for all embryos during the stationary flow phase and eggshell reference ) for pseudocleavage","( a ) and cytokinesis","( b ) .","Error bars , the standard error of the mean .","Right column , ingression distance for several embryos and mean value ( cyan ) during the stationary phase of pseudocleavage and cytokinesis . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 019 The characteristic time for relaxation to an isotropic state via turnover is determined by τ , while β is a dimensionless coefficient that describes how local compression −∂xv impacts nematic ordering .","Finally , ℓ is a length scale that determines the distance over which actin filaments tend to point in the same direction ( Salbreux et al . , 2009 ) .","In our experiments , we observe a stationary phase ( Figure 2b ) , and the profile of nematic order Q at steady-state is given by ( 2 ) Q=−τv∂xQ−β2τ∂xv+ℓ2 ∂x2Q .","We next determined the theoretical alignment field from the experimental flow field v and compression field ∂xv .","We used nonlinear least-square fitting to evaluate parameter values for which the theoretical alignment profile best matched the experimental one .","There are five unknowns ( three parameters that characterize emergent material properties , τ , ℓ , and β τ , as well as two boundary values of nematic order ) but three spatial fields , hence the fitting procedure is significantly constrained and the best fit parameters are uniquely defined in most cases ( see Appendix and Figure 3—figure supplement 1 ) .","For cytokinesis the calculated alignment profile ( dashed red lines in Figure 3b ) best matches the experimentally measured one for β τ =0 . 64 ( 0 . 548 , 0 . 713 ) min ( unless otherwise noted we indicate the median value together with the standard 68% confidence interval of the distribution of the bootstrap data , see Appendix ) , τ = 2 . 3 ( 1 . 89 , 2 . 71 ) min and ℓ = 1 . 7 ( 1 . 37 , 2 . 44 ) μm ( Table 1 and Table 2 ) .","For pseudocleavage , β τ was similar ( 0 . 6 ( 0 . 577 , 0 . 7 ) min ) , τ was determined to be smaller than 0 . 5 min , and ℓ = 4 . 7 ( 3 . 12 , 6 . 09 ) μm larger ( Table 1 and Table 2 ) .","We note that in both cases , we obtain good agreement between the theoretical and experimental alignment field , allowing us to conclude that compressive gel flow can account for alignment .","Further , we observe a small τ for pseudocleavage and a larger τ for cytokinesis , which is consistent with the observation that alignment appears concomitant with compression in pseudocleavage but arises with a short delay during cytokinesis ( Figure 2d ) .","We speculate that the changes in physical parameters of the actomyosin cortical layer between pseudocleavage and cytokinesis ( higher τ and smaller ℓ ) reflect the fact that the cell forms a strong and more pronounced ring of aligned filaments during cytokinesis .","To conclude , our results are consistent with a scenario in which actin filament alignment arises in a disordered network through compression by flow ( White and Borisy , 1983; Salbreux et al . , 2009; Turlier et al . , 2014 ) . 10 . 7554/eLife . 17807 . 020Table 1 . Best fit gel material parameters , bootstrapping .","The median values together with the standard 68% confidence bounds of the distribution of the bootstrap data are given . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 020Pseudocleavage Cytokinesis non RNAi non RNAi τ , min< 0 . 5 ( n . d . ) †2 . 26 ( 1 . 89 , 2 . 71 ) β τ , min0 . 626 ( 0 . 577 , 0 . 7 ) 0 . 646 ( 0 . 548 , 0 . 713 ) ℓ , μm4 . 72 ( 3 . 12 , 6 . 09 ) 1 . 74 ( 1 . 37 , 2 . 44 ) nop-1 RNAi nop-1 RNAi τ , minn . d .","*< 0 . 5 ( n . d . ) †β τ , minn . d .","*0 . 771 ( 0 . 694 , 2 . 75 ) ℓ , μmn . d .","*2 . 67 ( 1 . 7 , 11 . 5 ) ani-1 RNAi ani-1 RNAi τ , min< 0 . 5 ( n . d . ) †0 . 989 ( 0 . 555 , 1 . 79 ) β τ , min1 . 45 ( 0 . 969 , 18 ) ‡0 . 724 ( 0 . 636 , 0 . 929 ) ℓ , μm12 . 3 ( 6 . 39 , 58 . 4 ) ‡3 . 38 ( 2 . 97 , 6 . 16 ) mlc-4 ( 5–7 hr ) RNAi mlc-4 ( 5–7 hr ) RNAi τ , min< 0 . 5 ( n . d . ) †4 . 76 ( 3 . 67 , 6 . 6 ) β τ , min0 . 467 ( 0 . 451 , 0 . 503 ) 0 . 637 ( 0 . 567 , 0 . 768 ) ℓ , μm2 . 75 ( 2 . 38 , 4 . 1 ) 6 . 58 ( 5 . 31 , 6 . 85 ) mlc-4 ( 7–9 hr ) RNAi mlc-4 ( 7–9 hr ) RNAi τ , minn . d .","*< 0 . 5 ( n . d . ) †β τ , minn . d .","*0 . 916 ( 0 . 82 , 1 . 04 ) ℓ , μmn . d .","*3 . 12 ( 2 . 49 , 4 . 18 ) *n . d . denotes parameters which could not be determined , †< 0 . 5 ( n . d . ) denotes parameters which could not be determined , but could be determined to be below 0 . 5 ( see Figure 3—figure supplement 1b ) , ‡denotes values that could be determined but that were not well constrained . 10 . 7554/eLife . 17807 . 021Table 2 . Best fit gel material parameters , no bootstrapping .","The mean values together with its 95% confidence bounds are given . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 021PseudocleavageCytokinesisnon RNAi non RNAi τ , min< 0 . 5 ( n . d . ) †2 . 34 ( 1 . 74 , 3 . 14 ) β τ , min0 . 59 ( 0 . 513 , 0 . 806 ) 0 . 652 ( 0 . 541 , 0 . 764 ) ℓ , μm5 . 38 ( 3 . 27 , 8 . 85 ) 1 . 69 ( 1 . 04 , 2 . 75 ) nop-1 RNAi nop-1 RNAi τ , minn . d .","*< 0 . 5 ( n . d . ) †β τ , minn . d .","*0 . 738 ( 0 . 0 . 614 , 0 . 862 ) ℓ , μmn . d .","*1 . 79 ( 0 . 64 , 5 . 03 ) ani-1 RNAi ani-1 RNAi τ , min< 0 . 5 ( n . d . ) †1 . 01 ( 0 . 989 , 1 . 03 ) β τ , min1 . 31 ( 0 . 428 , 2 . 19 ) 0 . 74 ( 0 . 588 , 0 . 892 ) ℓ , μm10 . 4 ( 4 . 91 , 21 . 9 ) ‡4 . 08 ( 2 . 68 , 6 . 2 ) mlc-4 ( 5–7 hr ) RNAi mlc-4 ( 5–7 hr ) RNAi τ , min< 0 . 5 ( n . d . ) †**4 . 37 ( 3 . 08 , 6 . 21 ) β τ , min0 . 454 ( 0 . 437 , 0 . 47 ) 0 . 614 ( 0 . 501 , 0 . 727 ) ℓ , μm2 . 24 ( 1 . 87 , 2 . 67 ) 5 . 81 ( 5 . 02 , 6 . 72 ) mlc-4 ( 7–9 hr ) RNAi mlc-4 ( 7–9 hr ) RNAi τ , minn . d .","*< 0 . 5 ( n . d . ) †β τ , minn . d .","*0 . 97 ( 0 . 931 , 1 . 01 ) ℓ , μmn . d .","*2 . 98 ( 2 . 88 , 3 . 08 ) *n . d . denotes parameters which could not be determined , †< 0 . 5 ( n . d . ) denotes parameters which could not be determined , but could be determined to be below 0 . 5 ( see Figure 3—figure supplement 1b ) , ‡denotes values that could be determined but that were not well constrained .","Caption for Videos 1Videos 1 to 7 Until now , we have not considered processes of myosin-based active alignment in our theory ( Vavylonis et al . , 2008 ) .","This simplification is probably appropriate for pseudocleavage , but this is less clear for cytokinesis given that there is a clear band of myosin enrichment at the equator when the cell divides ( Figure 1c , d ) .","We next consider active alignment by myosin , and describe the cortical layer as a thin film of an active nematic gel .","For stationary flows the profile of nematic order Q is now given by ( 3 ) Q=−τv∂xQ−β2τ∂xv+ℓ2 ∂x2Q+λQ , where the right-most term describes active alignment by myosin molecular motors , with λ an alignment parameter dependent on the local myosin concentration .","In our modified fitting procedure , we now evaluate the respective contributions of myosin-based active and flow-based passive alignment to the stationary Q profile , under the assumption that λ is proportional to local myosin concentration .","We find that considering active alignment does not increase the overall agreement between calculated and measured alignment profiles for both pseudocleavage and cytokinesis ( Figure 3—figure supplement 2 ) .","Notably , the contribution of active alignment to pseudocleavage is insignificant , while active alignment contributes to enhancing compression-induced alignment during cytokinesis ( Figure 3—figure supplement 2b; see Appendix ) , albeit to a small degree .","We conclude that compression by flow and not myosin-based active alignment is the driving force of ring formation in both pseudocleavage and cytokinesis .","A key prediction from our model is that reducing flow speeds and compression rates should give rise to less filament alignment and a possible inhibition of pseudocleavage furrow formation .","To test this prediction , we performed weak-perturbation RNAi experiments ( Naganathan et al . , 2014 ) of the regulatory myosin light chain of non-muscle myosin ( mlc-4 ( RNAi ) ( Craig et al . , 1983 ) to mildly reduce actomyosin flow speeds without significant changes in overall actomyosin organization ( Figure 4—figure supplement 1 ) .","Short feeding times ( 5–7 hr ) lead to a small reduction in gel flow and compression rates ( Figure 4a ) , but actin filaments still aligned in the central region and the pseudocleavage furrow still formed and ingressed .","Consistent with our hypothesis , longer feeding times ( 7–9 hr ) lead to a complete abolishment of cortical flow and a loss of both actin alignment and pseudocleavage furrow ( Figure 4b ) .","Actomyosin foci are not required for compression-based alignment since anillin- depleted embryos ( ani-1 ( RNAi ) , 24 hr ) still show significant alignment under compression in the absence of these dense foci ( Figure 4c ) ( Maddox et al . , 2005; Piekny and Maddox , 2010; Tse et al . , 2011 ) .","Reduced flow speeds and compression rates can also account for the absence of alignment and pseudocleavage in nop-1 ( RNAi ) embryos , in which RhoA-dependent processes are affected and pseudocleavage formation abrogated ( Figure 4d , Figure 4—figure supplement 1 and 2 ) ( Tse et al . , 2012; Rose et al . , 1995; Zhang and Glotzer , 2015 ) .","Overall , these experiments lead us to conclude that reducing gel flow and compression rates affects alignment in a manner that is consistent with Equation 2 .","Finally , a complete loss of flow and compression abolishes pseudocleavage entirely ( Figure 4b;d and Figure 4—figure supplement 1 ) .","Taken together , our results suggest that the pseudocleavage furrow arises as a by-product of compressive flows , mechanically aligning actin filaments into a ring even in the absence of localized RhoA activation . 10 . 7554/eLife . 17807 . 022Figure 4 . Flows and compression are required to generate alignment .","( a-d )","Average AP profiles of gel flow ( light green , smoothened ) , compression rates ( dark green ) and nematic order parameter ( blue ) at the time of stationary flow during pseudocleavage onset for mlc-4 5–7 hr and 7–9 hr , nop-1 and ani-1 RNAi .","Error bars represent the standard error of the mean ( N = 17 , 14 , 10 , 22 for a-d , respectively ) .","Dashed red line indicate the respective best-fit theory predictions for the nematic order parameter ( blue ) .","For ani-1 and mlc-4 ( 5–7 hr ) , τ is small and is determined to be smaller than 0 . 5 min .","For nop-1 and mlc-4 ( 7–9 hr ) , theory profiles were generated using the parameters obtained for the unperturbed non-RNAi pseudocleavage , since insufficient compression rates did not constrain the physical parameters .","Scale bar , 10 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02210 . 7554/eLife . 17807 . 023Figure 4—figure supplement 1 . Cytoskeletal organization under RNAi perturbation .","( a-d )","Actin organization and myosin distribution under nop-1 ( RNAi ) , ani-1 ( RNAi ) and mlc-4 ( RNAi ) .","Inserts with fluorescence intensity profiles during pseudocleavage and cytokinesis onset are overlaid next to the corresponding images ( actin in blue , myosin in magenta ) .","Note that ANI-1 depleted embryos had more elongated and thin anterior halves , with no localized dense actin ring like structure and no clear and stable central furrow with membrane overlapping .","Also note that the asymmetry in the distribution of actin ( usually denser in the anterior half of the embryo ) during cytokinesis was reduced for both nop-1 and long mlc-4 ( RNAi ) .","Right , combined images and zoom ( green Lifeact , red NMY-2 ) .","Scale bar , 10 μm ( full images ) ; 5 μm ( zoom ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02310 . 7554/eLife . 17807 . 024Figure 4—figure supplement 2 . Flows , compression and alignment at cytokinesis onset .","( a-d )","Results of nop-1 ( RNAi ) , ani-1 ( RNAi ) and mlc-4 ( RNAi ) .","Average AP profiles of gel flow ( light green , smoothened ) , compression rates ( dark green ) and nematic order parameter ( blue ) at the time of stationary flow during cytokinesis onset .","Error bars represent the standard error of the mean ( for nop-1 , ani-1 , mlc-4 5–7 hr and 7–9 hr , N = 9 , 18 , 9 , 13 , respectively ) .","Dashed red line indicates the respective best-fit theory predictions for the nematic order parameter ( blue ) .","For cytokinesis mlc-4 ( 7–9 hr ) tau is small and is determined to be smaller than 0 . 5 min . ( e-f ) Overlay of the different profiles at pseudocleavage and cytokinesis onset , non-RNAi is shown in blue and compression rates in dashed lines .","Note that the level of outer ingression was similar to the non-RNAi case for ani-1 ( RNAi ) pseudocleavage , but with improper location and no membrane overlapping , thus no full pseudocleavage furrowing was obtained but a remaining initial anterior ingression was achieved in the presence of some filaments alignment . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 024 In this context we wondered if the cortex is specifically modified to favor ring formation in cytokinesis .","We used our fitting procedure to determine the parameters that characterize emergent material properties of the cortex under mlc-4 ( RNAi ) .","Notably , we observe that 7–9 hr of mlc-4 ( RNAi ) leads to a reduction of the alignment relaxation time scale τ during cytokinesis .","The value obtained is similar to the alignment relaxation time scale during pseudocleavage under non-RNAi conditions ( Table 1 and Table 2 ) , leading us to speculate that an increase in myosin activity via Rho signaling is responsible for the observed increase of the characteristic relaxation time τ in cytokinesis .","Importantly , increasing τ allows the cell to form a pronounced ring of aligned filaments by compressive cortical flow during cytokinesis , as predicted by theory ( Figure 3—figure supplement 1 ) .","Finally , we sought to find support from theory that aligned filaments in the equatorial region drive anisotropic stress generation ( Mayer et al . , 2010 ) to form an ingression ( White and Borisy , 1983; Salbreux et al . , 2009; Turlier et al . , 2014 ) .","By use of a theory that connects anisotropic active stress generation in the active nematic gel to changes in cell shape ( see Appendix and Figure 3d ) we show that we cannot appropriately account for the observed ingression profiles unless we consider anisotropic active stress generation in an aligned gel ( Figure 3e , Figure 3—figure supplement 3 ) .","This suggests that anisotropic active stress generation in a network of aligned actin filaments is important for forming an ingression .","To conclude , we find that compression by flow drives actomyosin ring formation in both pseudocleavage and cytokinesis , as originally proposed by White and Borisy ( White and Borisy , 1983 ) ( Figure 5 ) .","For this , we provide a general framework for determining emergent material parameters of the actomyosin gel .","This characterizes flow-alignment coupling and allows for capturing essential aspects of the mechanics of furrow generation .","Our analysis of the pseudocleavage ingression in C . elegans demonstrates that constricting rings can form in the absence of equatorial RhoA activation .","This reveals that cortical flow functions as a central organizer of network architecture , mechanically aligning filaments to form a constricting furrow in the equatorial region without a localized increase of actin nucleation ( Figure 2—figure supplement 5 , f ) and myosin contractility ( Figure 1b-d ) .","Finally , it will be interesting to investigate if a cortex that is aligned by compressive flow favors the recruitment of specific actin binding proteins such as bundling or motor proteins ( Robinson et al . , 2002 ) , comprising an interesting mechanism of positive feedback for stabilizing and enhancing the ring during cytokinesis .","Our work highlights that determining emergent material properties of actomyosin is important for understanding cytokinesis , and a challenge for the future is to link emergent material properties at the larger scale with molecular mechanisms ( Mendes Pinto et al . , 2013; Eggert et al . , 2006 ) ."],["Existing reagents did not allow for detailed and long-term imaging of actin filaments in embryos , which is critical for reliably quantifying their orientation .","Hence , we generated a new Lifeact transgenetic line with enhanced actin filament labeling ( SWG001 ) .","A codon optimized for C . elegans far-red fluorophore , mKate2 kindly shared by Henrik Bringman ( Redemann et al . , 2011 ) ( 26 kDa , 588/633 nm ) was added to the actin probe with a linker ( 66 bp ) and cloned into MOSCI vector containing unc119 rescue gene ( Hyman Lab ) under the control of the mex-5 promoter .","Stable integration was obtained by bombardment of this plasmid in DP38 strain .","This strain was then backcrossed with N2 .","In this line , we observed within the cortical plane bright actin filament labeling with a high signal-to-noise ratio and little photobleaching .","Importantly , overall cortical organization and flow dynamics appeared to not be affected by this reagent , foci lifetime , spacing , cortical flow velocities were similar to previous measurement with other fluorescent lines .","A dual colored transgenic line was obtained in order to image simultaneously actin and myosin dynamics by crossing Lifeact::mKate2 strain with LP133 ( NMY-2::GFP obtained by CRISPR ( Dickinson et al . , 2013 ) , SWG007 ) .","The RhoA biosensor used for Figure 1 was developed by the Glotzer lab ( GFP::AHPH , C . elegans strain MG617 , Tse et al . , 2012 ) .","This sensor consists of GFP fused to the C-terminal portion of C . elegans anillin , which contain its conserved region ( AH ) and pleckstrin homology ( PH ) domain .","It lacks the N-terminal myosin and actin-binding domains but retains its RhoA-binding domain .","The strains expressing CYK-1:GFP ( SWG004 ) and PLST-1:GFP ( SWG005 ) were obtained by the insertion of a GFP tag at the endogenous locus in their C-terminal region using CRISPR according to Dickinson et al ( Dickinson et al . , 2013 ) .","C . elegans worms were cultured on OP50-seeeded NGM agar plates as described ( Brenner , 1974 ) .","L4 mothers were maintained overnight at 20°C before dissection in M9 buffer .","For imaging , one-cell embryos were mounted on 2% agar pads , thereby slightly compressed to increase the cortical surface visible in a confocal plane .","Images were acquired with spinning disk confocal microscope ( Zeiss C-Apochromat , 63X/1 . 2 NA or 100X/1 . 42 NA , Yokogawa CSU-X1 scan head and Hamamatsu Orca-Flash4 . 0 camera ) every 2 s on two or three different planes ( one or two at the bottom for a cortical section , and one 12 µm above for a medial section of the embryo ) .","For 3D acquisitions , Z-stacks were acquired every 0 . 2 μm .","We assume the shape to be rotationally symmetrical to the longer axis , thus the outline of one medial section of the embryo reflects faithfully cell shape changes and ingression distance .","RNAi experiments were performed by feeding ( Timmons et al . , 2001 ) .","L4 worms were grown at 20°C on feeding plate ( NGM agar containing 1 mM isopropyl-β-D-thiogalactoside and 50 μg ml−1 ampicillin ) for the required number of hours before imaging .","Spatiotemporal correlation functions were calculated for all time points during the stationary period , in the 30 μm central region of the embryo .","The plot of the spatial correlation only ( 1D correlation function ) is shown in Figure 2c between the compression , alignment and ingression data sets .","Peak value positions are obtained by a Gaussian fitting procedure and allow the calculation of positional differences between the different data sets .","Because the flow field appears stationary , we expect this distance between peaks of compression , alignment and ingression to arise due to temporal delays between these events in the reference frame co-moving with the cortical flows .","We thus translated the characteristic distances between stationary peaks to temporal delays using the average flow velocity in the −10 μm to 0 central region ( 6 . 16 ± 1 . 23 µm/min for pseudocleavage and 4 . 01 ± 1 . 92 µm/min for cytokinesis ) ( Figure 2d ) . 10 . 7554/eLife . 17807 . 025Figure 5 . Compression drives alignment for furrowing . Schematic illustration of flows , cytoskeletal organization and cell shape changes during pseudocleavage and at the onset of cytokinesis . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02510 . 7554/eLife . 17807 . 026Video 5 . Actomyosin gel dynamics in the C . elegans zygote after 24 hr nop-1 ( RNAi ) .","Cortical and medial planes of an embryo expressing both Lifeact::mKate2 and endogenous NMY-2::GFP .","Left panel , medial plane Lifeact:mKate2 , center , cortical NMY-2::GFP , right panel , cortical Lifeact:mKate2 ( min:s ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02610 . 7554/eLife . 17807 . 027Video 6 . Actomyosin gel dynamics in the C . elegans zygote after 26 hr ani-1 ( RNAi ) .","Cortical and medial planes of an embryo expressing both Lifeact::mKate2 and endogenous NMY-2::GFP .","Left panel , medial plane Lifeact:mKate2 , center , cortical NMY-2::GFP , right panel , cortical Lifeact:mKate2 ( min:s ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02710 . 7554/eLife . 17807 . 028Video 7 . Actomyosin gel dynamics in the C . elegans zygote after 14 hr mlc-4 ( RNAi ) .","Cortical and medial planes of an embryo expressing both Lifeact::mKate2 and endogenous NMY-2::GFP .","Left panel , medial plane Lifeact:mKate2 , center , cortical NMY-2::GFP , right panel , cortical Lifeact:mKate2 ( min:s ) .","DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 028"]],"string":"[\n [\n \"Cytokinesis begins in late anaphase , triggered by regulatory pathways with feedback from the mitotic spindle that ensure appropriate positioning of the molecular machinery ( reviewed in Fededa and Gerlich , 2012; Green et al . , 2012 ) .\",\n \"As a consequence , the small regulatory GTPase RhoA ( Piekny et al . , 2005; Tse et al . , 2012 ) and the molecular motor myosin ( Shelton et al . , 1999; Werner and Glotzer , 2008 ) are enriched in an equatorial band ( Piekny et al . , 2005; Werner and Glotzer , 2008 ) .\",\n \"Myosin can directly organize the actin network ( Miller et al . , 2012; Soares e Silva et al . , 2011; Vavylonis et al . , 2008 ) , and if myosin-based active alignment ( e . g . via search-and-capture [Vavylonis et al . , 2008] ) or flow-based compression as proposed by White and Borisy ( White and Borisy , 1983; Rappaport , 1996 ) drive furrow formation and cytokinesis in eukaryotes remains unclear ( Green et al . , 2012; Bray and White , 1988; Levayer and Lecuit , 2012; Mendes Pinto et al . , 2013; Cao and Wang , 1990; Murthy and Wadsworth , 2005; Zhou and Wang , 2008 ) .\"\n ],\n [\n \"We set out to address this question in the one cell embryo of the nematode C . elegans which forms two constricting ingressions , first a pseudocleavage furrow during polarity establishment , and then a cytokinetic furrow during cytokinesis ( Figure 1—figure supplement 1a , Video 1 ) ( Munro and Bowerman , 2009; Rose et al . , 1995 ) .\",\n \"Notably , the pseudocleavage furrow is linked to cortical flow ( Mayer et al . , 2010; Munro et al . , 2004 ) and lacks regulatory control from the mitotic spindle ( Werner and Glotzer , 2008 ) .\",\n \"We first characterized the actomyosin cortical network during pseudocleavage and cytokinesis .\",\n \"For this , we generated a Lifeact::mKate2 fluorescent line of C . elegans by coupling the Lifeact peptide to the far-red fluorophore mKate2 that was codon optimized for C . elegans ( Redemann et al . , 2011 ) ( see Materials and methods and Video 2 ) .\",\n \"This allows us to quantify cortical actin filament orientation in space and time , as well as cortical flow velocity fields by Particle Image Velocimetry ( PIV , Figures 1 and 2 , Figure 2—figure supplement 1 ) ( Mayer et al . , 2010 ) .\",\n \"We find that actin filaments change from a randomly oriented and isotropic gel before cortical flows initiate , into a more circumferentially aligned organization at the future site of furrow ingression during both pseudocleavage and cytokinesis ( Figure 1b , Figure 1—figure supplement 1 and Videos 1 and 3 ) .\",\n \"We studied how pseudocleavage and cytokinesis differ in the spatial distribution of molecular regulation and contractility within the equatorial region ( Figure 1 ) .\",\n \"Non-muscle myosin II ( NMY-2 ) is the essential molecular motor that drives both cortical flow and cytokinesis ( Shelton et al . , 1999; Guo and Kemphues , 1996 ) , whereas the small GTPase RhoA ( Piekny et al . , 2005; Tse et al . , 2012 ) , when active , is responsible for its local activation .\",\n \"At the onset of cytokinesis , active RhoA and NMY-2 locally increase at the position of the contractile ring , while for pseudocleavage we observed no such increase ( Figure 1 ) .\",\n \"Instead , the equatorial region in pseudocleavage corresponds to a transition zone with a gradual decrease of RhoA and myosin between the anterior region of high and the posterior region of low contractility ( Figure 1 ) ( Mayer et al . , 2010 ) .\",\n \"The actin nucleator formin CYK-1 is downstream of RhoA ( Piekny et al . , 2005 ) and follows a similar pattern of localization , and actin bundling via anillin ( Piekny and Glotzer , 2008 ) or plastin also does not appear to be increased in the equatorial region during pseudocleavage ( Figure 2—figure supplement 5e–f ) .\",\n \"To conclude , the pseudocleavage furrow ingresses with a circumferential alignment of filaments in the gel , but without a local zone of RhoA activation and the corresponding local increase in myosin and formin density .\",\n \"Since search-and-capture like mechanisms require a localized band of myosin and actin nucleators at the equator to drive myosin-based active alignment ( Vavylonis et al . , 2008 ) , this suggests that actin filaments in pseudocleavage align via flow-based compression and not via active alignment . 10 . 7554/eLife . 17807 . 003Video 1 . Actomyosin gel dynamics in the C . elegans zygote . Cortical and medial planes of an embryo expressing both Lifeact::mKate2 and endogenous NMY-2::GFP .\",\n \"Left panel , medial plane Lifeact:mKate2 , center , cortical NMY-2::GFP , right panel , cortical Lifeact:mKate2 ( min:s ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00310 . 7554/eLife . 17807 . 004Video 2 . Three-dimensional reconstruction of an early one-cell embryo of C . elegans expressing Lifeact::mKate2 . The back of the embryo is dimmer due to imaging . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00410 . 7554/eLife . 17807 . 005Figure 1 . Flow and ingression during pseudocleavage and cytokinesis .\",\n \"( a ) Average active RhoA biosensor intensity profiles along the AP axis ( 0 represents the embryo center; N = 14 embryos for pseudocleavage and N = 7 for cytokinesis; errors represent the standard error of the mean ) .\",\n \"( b ) Actomyosin cortical organization at the onset of pseudocleavage and cytokinesis in embryos expressing both Lifeact::mKate2 and NMY-2::GFP .\",\n \"( c–d )\",\n \"Average actin , myosin and ingression profiles as a function of position along the AP axis ( N = 10 embryos for pseudocleavage and N = 22 for cytokinesis; errors represent the standard error of the mean ) .\",\n \"Scale bars , 10 μm .\",\n \"The following figure supplement is available for Figure 1:DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00510 . 7554/eLife . 17807 . 006Figure 1—figure supplement 1 . Details of actin organization during the polarization flow phase .\",\n \"( a ) Actomyosin gel dynamics in the C . elegans zygote .\",\n \"Single cortical and medial planes from a timelapse sequence of a representative embryo expressing both Lifeact::mKate2 and NMY-2::GFP .\",\n \"Anterior is to the left , posterior to the right .\",\n \"White arrows indicate the position of pseudocleavage and cytokinesis ingressions .\",\n \"Red arrows indicate flow directions .\",\n \"( b ) Close up look at the filaments dynamics near the site of flow initiation .\",\n \"Cortical plane from a timelapse sequence of an embryo expressing both Lifeact::mKate2 and NMY-2::GFP .\",\n \"( c ) Top: Actin network organization from an embryos expressing Lifeact::mKate2 .\",\n \"Times rescaled to 0 at flow initiation .\",\n \"Green arcs indicate the smooth cleared zone during the polarization process .\",\n \"Red lines schematize filaments maximum vertical alignment positions .\",\n \"White arrows underlie pseudocleavage ingression .\",\n \"Bottom: color-coded representation of the orientation of filaments .\",\n \"( d ) Example from an embryo for which the polarization process starts off centered from the physical pole of the embryo due to a mispositioning of the sperm pronucleus ( dashed circle on medial plane ) .\",\n \"Alignment direction is observed to follow perpendicularly to the direction of the flow as it reorients along the long axis of the embryo ( also see Appendix ) .\",\n \"Red arrows represent flow direction , red lines filaments main orientation .\",\n \"Scale bars , 10 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00610 . 7554/eLife . 17807 . 007Figure 2 . Cortical flows compress the gel to generate alignment .\",\n \"( a ) Filament orientation is quantified by a nematic order parameter; flow fields are obtained from Particle Image Velocimetry ( PIV ) analysis .\",\n \"( b ) Gel flow , compression rate , filament orientation and cell ingression fields .\",\n \"Top row , schematic representations .\",\n \"Middle row , kymographs of the time evolution of the profiles along the AP axis of flow velocity , compression rate , nematic order parameter and outer ingression distance for pseudocleavage ( N = 10 ) .\",\n \"Bottom row , cytokinesis ( N = 12 ) .\",\n \"Orange dashed lines indicate the times where all four fields are stationary .\",\n \"Flow , compression and nematic order were determined from bottom-section and ingression from mid-section views of Lifeact::mKate2 labelled embryos ( see Materials and methods and Appendix ) .\",\n \"( c ) Spatial correlation between compression , alignment and ingression ( see Materials and methods ) .\",\n \"Peak value positions ( estimated by a Gaussian fitting of the 1D correlation curve ) are indicated .\",\n \"See Figure 2—figure supplement 2 for spatiotemporal correlation functions .\",\n \"( d ) Table of the temporal delays obtained from the correlation peak value positions using the mean values of the velocity in the 10 μm central region ( mean values for pseudocleavage: v = 6 . 16 ±1 . 23 μm/min; for cytokinesis: v = 4 . 01±1 . 92 μm/min; mean ± SD ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00710 . 7554/eLife . 17807 . 008Figure 2—figure supplement 1 . Quantification methods developed .\",\n \"( a ) Example of the filaments alignment quantification ( red lines ) and the flow velocity field ( yellow arrows ) , as well as nematic order parameters conventions .\",\n \"( b ) Time evolution of flow ( light green ) and compression rate ( dark green ) as well as ingression ( grey ) measured at a fixed position along the AP axis ( −20 μm for velocity , 0 μm or central position of the embryo for compression and 3 μm for ingression , measured from the center ) .\",\n \"Error bars represent the standard error of the mean .\",\n \"Bottom , fit using the error function ( red line ) of the time evolution of the velocity from a single embryo ( blue dots ) used to determine the reference time point ( Time = 0 ) .\",\n \"( c ) Time evolution of the mean velocity and compression rate along the AP axis and nematic order parameter , all taken at fixed positions along the AP axis .\",\n \"Error bars represent the standard error of the mean .\",\n \"In orange is represented the period chosen as stationary flow phase .\",\n \"( d ) Mean profiles along the AP axis of the velocities ( vx light green and vy yellow ) , compression rate and nematic order parameters ( Q = −Qxx in dark blue and Qxy in cyan ) during the stationary flow phase .\",\n \"( See Appendix for further explanations . ) DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00810 . 7554/eLife . 17807 . 009Figure 2—figure supplement 2 . Spatiotemporal correlations . Plots of the spatiotemporal correlations between the compression , alignment and ingression data sets taken during the pseudocleavage stationary phase\",\n \"( a ) and cytokinesis onset stationary phase\",\n \"( b ) .\",\n \"Data is restricted to the 30 μm central region of the embryo . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 00910 . 7554/eLife . 17807 . 010Figure 2—figure supplement 3 . Nematic order parameter quantification of artificial images with and without directional bias .\",\n \"( a ) Example of a simulated isotropic actin meshwork ( see Appendix for details ) .\",\n \"( b ) Example of a simulated actin meshwork for which the angle of the starting direction is biased toward vertical ( see Appendix ) .\",\n \"( c ) Cumulative distribution function ( CDF ) used and example of an image and its nematic order map ( red lines indicate the preferential orientation inside a template or orientation vector ) obtained for different values of the starting parameter B that controls the bias of the filaments orientation .\",\n \"B = 0 leads to an isotropic orientation .\",\n \"( d ) Variation of the average nematic order parameter obtained for several synthetic meshworks while increasing the vertical bias of the filaments orientation .\",\n \"The saturation is caused by the fact that we here bias only the initial direction of growth . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01010 . 7554/eLife . 17807 . 011Figure 2—figure supplement 4 . Impact of image quality on the nematic order parameter quantification .\",\n \"( a ) An artificial filamentous network is obtained by a random iteration process with controllable parameters in Matlab ( see Appendix for more details ) .\",\n \"( b-d )\",\n \"We chose to maintain a vertical bias in the orientation of the filaments in all the following cases ( B = 1 ) .\",\n \"Red lines indicate the preferential orientation inside a template or orientation vector .\",\n \"A profile of the mean nematic order parameter along the x-axis as well as the angular distribution are shown in all cases .\",\n \"Note that as we do not measure the directionality of filaments thus we obtain values modulo π .\",\n \"( b ) Impact of the filament density ( ranging from 60 to 120 filaments per image ) on the nematic order parameter quantification .\",\n \"The profile of nematic order is to first order independent of network density .\",\n \"( c ) Impact of the signal to noise ratio .\",\n \"High signal to noise ratio leads to higher values of the nematic order parameter , low signal to noise ratio leads to a poor detection of the direction of the filaments and artificially small values of the nematic order parameter . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01110 . 7554/eLife . 17807 . 012Figure 2—figure supplement 5 . Comparison of different actin labeling strains and actin-binding protein localization .\",\n \"( a–c )\",\n \"Average AP profiles of flow velocity ( light green ) and nematic order parameter ( blue ) at the time of stationary flow during pseudocleavage of respectively 22 embryos expressing Lifeact:GFP , NMY:RFP , 9 embryos expressing Lifeact:RFP , NMY-2:GFP and 16 embryos expressing Lifeact:mKate2 , NMY-2:GFP .\",\n \"Overlaid in red is the velocity profile of worms expressing only labeled NMY-2 ( N = 5 ) .\",\n \"Error bars represent the standard error of the mean .\",\n \"( d ) Images of NMY-2:GFP in the early zygote .\",\n \"Left strain without Lifeact , right strain expressing Lifeact:mKate2 .\",\n \"The density , size and dynamics are similar in both strains .\",\n \"Right panel , the average radial autocorrelation coefficient of the myosin intensity over a minute in early embryos gives an average myosin foci size .\",\n \"( e ) Localization of the actin bundling protein plastin during pseudocleavage and cytokinesis .\",\n \"Plastin localizes in dense bundles of actin filaments and is particularly abundant in the cytokenetic ring and could act to stabilize actin bundles in these regions .\",\n \"( f ) Localization of the actin nucleating agent formin during pseudocleavage and cytokinesis .\",\n \"CYK-1 is present throughout the cell cortex during pseudocleavage , with higher densities in foci regions .\",\n \"It is recruited to the cell equator during cytokinetic ring formation and ingression .\",\n \"Scale bars , 10 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01210 . 7554/eLife . 17807 . 013Video 3 . Cortical dynamics at flow initiation . Left panel , cortical NMY-2::GFP , center cortical Lifeact:mKate2 and a zoom is shown in the right panel ( min:s ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 013 We next sought to test if flow-based compression drives actin filament alignment .\",\n \"To this end , we developed tools for the spatiotemporal quantification of compression by flow , of filament orientation , and of cortex ingression distance within the gel both at the onset of pseudocleavage and cytokinesis ( see Appendix as well as Figure 2a and Figure 2—figure supplement 1 ) .\",\n \"We find that the flow compression rate along the antero-posterior direction ( AP axis or x axis ) , given by the spatial gradient of the velocity field −∂xv ( Figure 2b ) , increases with time during the very early stages of pseudocleavage .\",\n \"The compression rate peaks in the central region of the embryo , with a maximum of ~0 . 4 min−1 .\",\n \"Notably , this peak is stable for several minutes until pseudocleavage flows cease entirely ( Figure 2b ) .\",\n \"For cytokinesis , we find that the early flow also proceeds in a unidirectional manner from the posterior toward the anterior ( Figure 2b ) .\",\n \"Similar to pseudocleavage , the compression rate increases with time over the very early stages of cytokinesis , with the compression profile peaking at the center of the embryo with a maximum rate of ~0 . 7 min−1 ( Figure 2b ) .\",\n \"We characterized the average actin filaments orientation by a nematic order tensor Q ( Figure 2—figure supplement 1a ) .\",\n \"This analysis relies on the fact that the intensity of the Fourier transformed image encodes geometric characteristics such as its main orientation pattern ( see Appendix ) .\",\n \"We find that vertical ( orthogonal to the direction of flow ) alignment appears coincidental in space and time with the local increase of the compression rate ( Figure 2b , compare column 2 and 3 ) .\",\n \"Notably , changing the direction of flow also changes the direction of alignment: off-site sperm entry leads to flows along the short axis of the egg ( Goldstein et al . , 1993 ) and actin filaments still align in the direction determined by compression ( Figure 1—figure supplement 1d , Video 4 and Appendix ) .\",\n \"Finally , the ingressing furrow forms in the region where filaments are aligned ( Figure 2b , column 4 ) .\",\n \"Our quantifications reveal remarkable similarities between the compression rate fields , the pattern of alignment , and the ingression distance between pseudocleavage and cytokinesis ( Figure 2b ) , suggesting common mechanisms .\",\n \"To conclude , actin filament alignment arises at locations of significant compression by flow . 10 . 7554/eLife . 17807 . 014Video 4 . Cortical dynamics of an embryo in which flow was initiated at the side ( bottom ) and far from the posterior pole of the embryo . Actin filament alignment follows flow and compression , thus transitions from a horizontal to a vertical direction ( min:s ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 014 If flow-based compression aligns actin filaments for forming an ingression , we would expect compression to precede or to be concomitant with alignment and ingression .\",\n \"The 1D crosscorrelations between compression , filament alignment and ingression distance , calculated for the time that flows are essentially stationary ( Figure 2b and Figure 2—figure supplement 1 ) , show that for pseudocleavage both compression rate and alignment peak at about the same location , while the ingression field peaks ~ 3 µm further to the anterior ( Figure 2c and Figure 2—figure supplement 2 for the spatiotemporal crosscorrelations ) .\",\n \"The situation is similar at cytokinesis , except that the compression field peaks furthest to the posterior .\",\n \"We next translated the characteristic distances between peaks in stationary flow to temporal delays between events ( see Materials and methods for further explanation ) .\",\n \"We find that for pseudocleavage , filament alignment is essentially concomitant with compression , while for cytokinesis compression precedes alignment by 0 . 40 ± 0 . 16 min ( Figure 2d ) .\",\n \"Furthermore , ingression arises approximately 0 . 5 min after compression for pseudocleavage and 0 . 9 min after compression for cytokinesis .\",\n \"To conclude , filament alignment appears to rapidly follow compression while ingression arises with a delay .\",\n \"We next sought to provide a physical explanation for flow-based alignment ( Salbreux et al . , 2009 ) .\",\n \"The cell cortex in this and other systems can be thought of as a gel that forms a thin layer underneath the plasma membrane .\",\n \"Recent advances in physical theory show that this gel is active and generates the forces that drive cell shape changes ( Salbreux et al . , 2009; Gowrishankar , 2012; Kruse et al . , 2005; Turlier et al . , 2014 ) .\",\n \"We describe the actin cortical layer as a thin film of a nematic gel under shear and compression by flow ( Salbreux et al . , 2009; Kruse et al . , 2005; Prost et al . , 2015 ) .\",\n \"In the x-direction , the time evolution of the local nematic order Q depends on advection ( first term on the right hand side in Equation 1 ) , compression in gel flow ( second term ) , a process of relaxation to an isotropic configuration possibly via turnover ( third term ) , and a tendency of filaments to align with each other over space ( last term , see also Figure 3a ) ( 1 ) ∂tQ=−v∂xQ−β2∂xv−1τQ+ℓ2τ∂x2Q . 10 . 7554/eLife . 17807 . 015Figure 3 . Theory predicts an alignment peak .\",\n \"( a ) Schematic representation of flow-based alignment .\",\n \"For full theory refer to Appendix .\",\n \"( b ) Average AP profiles of gel flow ( light green , smoothened ) , compression rates ( dark green ) and nematic order parameter ( blue ) at the time of stationary flow ( Figure 2C ) during pseudocleavage ( left , N = 16 ) and cytokinesis onset ( right , N = 32 ) .\",\n \"Error bars represent the standard error of the mean .\",\n \"Dashed red line indicates the respective best-fit theory predictions for the nematic order parameter ( blue ) .\",\n \"( c ) Table of the material parameters .\",\n \"See also Figure 3—figure supplement 1 and Table 1 , Table 2 .\",\n \"( d ) Illustration of the cortical tension related to the ingression distance .\",\n \"( e ) Outer ingression profiles at the onset of pseudocleavage ( brown ) and cytokinesis ( yellow ) .\",\n \"The theoretical profiles ( thick dashed red lines: active tension depends on alignment and can be anisotropic; thin dashed lines: active tension is isotropic ) are determined by simultaneous fitting of both the pseudocleavage and cytokinesis datasets with a single fit parameter ( see Appendix ) using the myosin density , flow velocity and nematic order parameter fields . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01510 . 7554/eLife . 17807 . 016Figure 3—figure supplement 1 . Influence of the gel material parameters for the nematic order parameter .\",\n \"( a ) Nematic order parameter profiles obtained when varying independently each of the fitting parameters τ , βτ , ℓ away from their the best fit values obtained for cytokinesis .\",\n \"Dashed red line , best fit result ( τ = 2 . 34 min , βτ = 0 . 654 min , ℓ =1 . 69 μm ) .\",\n \"For each set of parameters τ , βτ , ℓ , a fit is performed for the fitting parameters C1 and C2 only ( see Appendix ) .\",\n \"( b ) Nematic order parameter profiles obtained when varying the fitting parameter τ while keeping βτ , ℓ fixed and equal to their best-fit values .\",\n \"Appropriate fits to experimental curve are obtained for all values of τ < 0 . 5 min .\",\n \"For each value of τ , a fit is performed for the fitting parameters C1 and C2 only . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01610 . 7554/eLife . 17807 . 017Figure 3—figure supplement 2 . Contribution of active alignment to the nematic order parameter profile .\",\n \"( a ) Average AP profiles of gel flow ( light green ) , compression rates ( dashed dark green ) , nematic order parameter ( blue ) and myosin fluorescence ( magenta ) at the time of stationary flow during pseudocleavage ( left , N = 16 ) and cytokinesis onset ( right , N = 32 ) .\",\n \"Error bars represent the standard error of the mean .\",\n \"Dashed red lines indicate the best-fit theory predictions for the nematic order parameter in absence of active alignment , dashed black lines are the fits obtained while including the myosin-based active alignment , with λ′τ=4 . 1 10−8±0 . 54 for pseudocleavage and λ′τ=0 . 34± 0 . 45 for cytokinesis .\",\n \"( b ) Contribution of the individual terms from Equation 3 in the main text ( red: −β2τ∂xv , green: −τv∂xQ , blue: ℓ2 ∂x2Q , magenta: λ′τ c ( x ) /c0Q ) .\",\n \"Note that myosin-based active alignment ( magenta ) is insignificant for pseudocleavage and is low for cytokinesis compared to compression-based alignment ( red ) .\",\n \"The theoretical alignment profile ( sum of all the terms shown in black ) overlays the experimental data points ( dots ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01710 . 7554/eLife . 17807 . 018Figure 3—figure supplement 3 . Ingression through anisotropic stress generation in the aligned gel .\",\n \"( a ) Top , a representative embryo is used to visualize the time course of ingression .\",\n \"Bottom , Kymograph obtained from the region specified by the dashed white box in the upper pannel .\",\n \"The pseudocleavage and cytokinesis ingressions are visible .\",\n \"Yellow arrow indicates the inner ingression distance during polarizing flow phase , magenta dashed line indicates the ingressing membrane and blue arrow indicates the outer ingression .\",\n \"Red doted lines represent the position of the eggshell .\",\n \"Scale bar 10 μm .\",\n \"( b ) Top , illustration of the cortical tension during ingression .\",\n \"Bottom , outer ingression profiles ( grey ) averaged over the time periods indicated by the thin horizontal white lines in ( a , bottom ) during pseudocleavage ( left ) and cytokinesis ( right ) .\",\n \"The theoretical profiles ( dashed red lines ) are determined by simultaneous fitting of both the pseudocleavage and cytokinesis datasets with a single fit parameter ( see Appendix ) using the myosin density ( magenta ) , flow velocity ( green ) and nematic order parameter ( blue ) fields .\",\n \"( c ) and\",\n \"( d ) Illustration of the coordinate and reference systems used for this theoretical description .\",\n \"( e ) Comparison of the theoretically obtained profile of the outer ingression without the anisotropic stress induced by nematic order ( dashed red , obtained by imposing p3 = 0 ) during pseudocleavage and cytokinesis phases with the experimental data ( black ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 01810 . 7554/eLife . 17807 . 019Figure 3—figure supplement 4 . Shape and ingression distance quantification . Left column , plots of the cell shape ( mean radius for all embryos during the stationary flow phase and eggshell reference ) for pseudocleavage\",\n \"( a ) and cytokinesis\",\n \"( b ) .\",\n \"Error bars , the standard error of the mean .\",\n \"Right column , ingression distance for several embryos and mean value ( cyan ) during the stationary phase of pseudocleavage and cytokinesis . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 019 The characteristic time for relaxation to an isotropic state via turnover is determined by τ , while β is a dimensionless coefficient that describes how local compression −∂xv impacts nematic ordering .\",\n \"Finally , ℓ is a length scale that determines the distance over which actin filaments tend to point in the same direction ( Salbreux et al . , 2009 ) .\",\n \"In our experiments , we observe a stationary phase ( Figure 2b ) , and the profile of nematic order Q at steady-state is given by ( 2 ) Q=−τv∂xQ−β2τ∂xv+ℓ2 ∂x2Q .\",\n \"We next determined the theoretical alignment field from the experimental flow field v and compression field ∂xv .\",\n \"We used nonlinear least-square fitting to evaluate parameter values for which the theoretical alignment profile best matched the experimental one .\",\n \"There are five unknowns ( three parameters that characterize emergent material properties , τ , ℓ , and β τ , as well as two boundary values of nematic order ) but three spatial fields , hence the fitting procedure is significantly constrained and the best fit parameters are uniquely defined in most cases ( see Appendix and Figure 3—figure supplement 1 ) .\",\n \"For cytokinesis the calculated alignment profile ( dashed red lines in Figure 3b ) best matches the experimentally measured one for β τ =0 . 64 ( 0 . 548 , 0 . 713 ) min ( unless otherwise noted we indicate the median value together with the standard 68% confidence interval of the distribution of the bootstrap data , see Appendix ) , τ = 2 . 3 ( 1 . 89 , 2 . 71 ) min and ℓ = 1 . 7 ( 1 . 37 , 2 . 44 ) μm ( Table 1 and Table 2 ) .\",\n \"For pseudocleavage , β τ was similar ( 0 . 6 ( 0 . 577 , 0 . 7 ) min ) , τ was determined to be smaller than 0 . 5 min , and ℓ = 4 . 7 ( 3 . 12 , 6 . 09 ) μm larger ( Table 1 and Table 2 ) .\",\n \"We note that in both cases , we obtain good agreement between the theoretical and experimental alignment field , allowing us to conclude that compressive gel flow can account for alignment .\",\n \"Further , we observe a small τ for pseudocleavage and a larger τ for cytokinesis , which is consistent with the observation that alignment appears concomitant with compression in pseudocleavage but arises with a short delay during cytokinesis ( Figure 2d ) .\",\n \"We speculate that the changes in physical parameters of the actomyosin cortical layer between pseudocleavage and cytokinesis ( higher τ and smaller ℓ ) reflect the fact that the cell forms a strong and more pronounced ring of aligned filaments during cytokinesis .\",\n \"To conclude , our results are consistent with a scenario in which actin filament alignment arises in a disordered network through compression by flow ( White and Borisy , 1983; Salbreux et al . , 2009; Turlier et al . , 2014 ) . 10 . 7554/eLife . 17807 . 020Table 1 . Best fit gel material parameters , bootstrapping .\",\n \"The median values together with the standard 68% confidence bounds of the distribution of the bootstrap data are given . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 020Pseudocleavage Cytokinesis non RNAi non RNAi τ , min< 0 . 5 ( n . d . ) †2 . 26 ( 1 . 89 , 2 . 71 ) β τ , min0 . 626 ( 0 . 577 , 0 . 7 ) 0 . 646 ( 0 . 548 , 0 . 713 ) ℓ , μm4 . 72 ( 3 . 12 , 6 . 09 ) 1 . 74 ( 1 . 37 , 2 . 44 ) nop-1 RNAi nop-1 RNAi τ , minn . d .\",\n \"*< 0 . 5 ( n . d . ) †β τ , minn . d .\",\n \"*0 . 771 ( 0 . 694 , 2 . 75 ) ℓ , μmn . d .\",\n \"*2 . 67 ( 1 . 7 , 11 . 5 ) ani-1 RNAi ani-1 RNAi τ , min< 0 . 5 ( n . d . ) †0 . 989 ( 0 . 555 , 1 . 79 ) β τ , min1 . 45 ( 0 . 969 , 18 ) ‡0 . 724 ( 0 . 636 , 0 . 929 ) ℓ , μm12 . 3 ( 6 . 39 , 58 . 4 ) ‡3 . 38 ( 2 . 97 , 6 . 16 ) mlc-4 ( 5–7 hr ) RNAi mlc-4 ( 5–7 hr ) RNAi τ , min< 0 . 5 ( n . d . ) †4 . 76 ( 3 . 67 , 6 . 6 ) β τ , min0 . 467 ( 0 . 451 , 0 . 503 ) 0 . 637 ( 0 . 567 , 0 . 768 ) ℓ , μm2 . 75 ( 2 . 38 , 4 . 1 ) 6 . 58 ( 5 . 31 , 6 . 85 ) mlc-4 ( 7–9 hr ) RNAi mlc-4 ( 7–9 hr ) RNAi τ , minn . d .\",\n \"*< 0 . 5 ( n . d . ) †β τ , minn . d .\",\n \"*0 . 916 ( 0 . 82 , 1 . 04 ) ℓ , μmn . d .\",\n \"*3 . 12 ( 2 . 49 , 4 . 18 ) *n . d . denotes parameters which could not be determined , †< 0 . 5 ( n . d . ) denotes parameters which could not be determined , but could be determined to be below 0 . 5 ( see Figure 3—figure supplement 1b ) , ‡denotes values that could be determined but that were not well constrained . 10 . 7554/eLife . 17807 . 021Table 2 . Best fit gel material parameters , no bootstrapping .\",\n \"The mean values together with its 95% confidence bounds are given . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 021PseudocleavageCytokinesisnon RNAi non RNAi τ , min< 0 . 5 ( n . d . ) †2 . 34 ( 1 . 74 , 3 . 14 ) β τ , min0 . 59 ( 0 . 513 , 0 . 806 ) 0 . 652 ( 0 . 541 , 0 . 764 ) ℓ , μm5 . 38 ( 3 . 27 , 8 . 85 ) 1 . 69 ( 1 . 04 , 2 . 75 ) nop-1 RNAi nop-1 RNAi τ , minn . d .\",\n \"*< 0 . 5 ( n . d . ) †β τ , minn . d .\",\n \"*0 . 738 ( 0 . 0 . 614 , 0 . 862 ) ℓ , μmn . d .\",\n \"*1 . 79 ( 0 . 64 , 5 . 03 ) ani-1 RNAi ani-1 RNAi τ , min< 0 . 5 ( n . d . ) †1 . 01 ( 0 . 989 , 1 . 03 ) β τ , min1 . 31 ( 0 . 428 , 2 . 19 ) 0 . 74 ( 0 . 588 , 0 . 892 ) ℓ , μm10 . 4 ( 4 . 91 , 21 . 9 ) ‡4 . 08 ( 2 . 68 , 6 . 2 ) mlc-4 ( 5–7 hr ) RNAi mlc-4 ( 5–7 hr ) RNAi τ , min< 0 . 5 ( n . d . ) †**4 . 37 ( 3 . 08 , 6 . 21 ) β τ , min0 . 454 ( 0 . 437 , 0 . 47 ) 0 . 614 ( 0 . 501 , 0 . 727 ) ℓ , μm2 . 24 ( 1 . 87 , 2 . 67 ) 5 . 81 ( 5 . 02 , 6 . 72 ) mlc-4 ( 7–9 hr ) RNAi mlc-4 ( 7–9 hr ) RNAi τ , minn . d .\",\n \"*< 0 . 5 ( n . d . ) †β τ , minn . d .\",\n \"*0 . 97 ( 0 . 931 , 1 . 01 ) ℓ , μmn . d .\",\n \"*2 . 98 ( 2 . 88 , 3 . 08 ) *n . d . denotes parameters which could not be determined , †< 0 . 5 ( n . d . ) denotes parameters which could not be determined , but could be determined to be below 0 . 5 ( see Figure 3—figure supplement 1b ) , ‡denotes values that could be determined but that were not well constrained .\",\n \"Caption for Videos 1Videos 1 to 7 Until now , we have not considered processes of myosin-based active alignment in our theory ( Vavylonis et al . , 2008 ) .\",\n \"This simplification is probably appropriate for pseudocleavage , but this is less clear for cytokinesis given that there is a clear band of myosin enrichment at the equator when the cell divides ( Figure 1c , d ) .\",\n \"We next consider active alignment by myosin , and describe the cortical layer as a thin film of an active nematic gel .\",\n \"For stationary flows the profile of nematic order Q is now given by ( 3 ) Q=−τv∂xQ−β2τ∂xv+ℓ2 ∂x2Q+λQ , where the right-most term describes active alignment by myosin molecular motors , with λ an alignment parameter dependent on the local myosin concentration .\",\n \"In our modified fitting procedure , we now evaluate the respective contributions of myosin-based active and flow-based passive alignment to the stationary Q profile , under the assumption that λ is proportional to local myosin concentration .\",\n \"We find that considering active alignment does not increase the overall agreement between calculated and measured alignment profiles for both pseudocleavage and cytokinesis ( Figure 3—figure supplement 2 ) .\",\n \"Notably , the contribution of active alignment to pseudocleavage is insignificant , while active alignment contributes to enhancing compression-induced alignment during cytokinesis ( Figure 3—figure supplement 2b; see Appendix ) , albeit to a small degree .\",\n \"We conclude that compression by flow and not myosin-based active alignment is the driving force of ring formation in both pseudocleavage and cytokinesis .\",\n \"A key prediction from our model is that reducing flow speeds and compression rates should give rise to less filament alignment and a possible inhibition of pseudocleavage furrow formation .\",\n \"To test this prediction , we performed weak-perturbation RNAi experiments ( Naganathan et al . , 2014 ) of the regulatory myosin light chain of non-muscle myosin ( mlc-4 ( RNAi ) ( Craig et al . , 1983 ) to mildly reduce actomyosin flow speeds without significant changes in overall actomyosin organization ( Figure 4—figure supplement 1 ) .\",\n \"Short feeding times ( 5–7 hr ) lead to a small reduction in gel flow and compression rates ( Figure 4a ) , but actin filaments still aligned in the central region and the pseudocleavage furrow still formed and ingressed .\",\n \"Consistent with our hypothesis , longer feeding times ( 7–9 hr ) lead to a complete abolishment of cortical flow and a loss of both actin alignment and pseudocleavage furrow ( Figure 4b ) .\",\n \"Actomyosin foci are not required for compression-based alignment since anillin- depleted embryos ( ani-1 ( RNAi ) , 24 hr ) still show significant alignment under compression in the absence of these dense foci ( Figure 4c ) ( Maddox et al . , 2005; Piekny and Maddox , 2010; Tse et al . , 2011 ) .\",\n \"Reduced flow speeds and compression rates can also account for the absence of alignment and pseudocleavage in nop-1 ( RNAi ) embryos , in which RhoA-dependent processes are affected and pseudocleavage formation abrogated ( Figure 4d , Figure 4—figure supplement 1 and 2 ) ( Tse et al . , 2012; Rose et al . , 1995; Zhang and Glotzer , 2015 ) .\",\n \"Overall , these experiments lead us to conclude that reducing gel flow and compression rates affects alignment in a manner that is consistent with Equation 2 .\",\n \"Finally , a complete loss of flow and compression abolishes pseudocleavage entirely ( Figure 4b;d and Figure 4—figure supplement 1 ) .\",\n \"Taken together , our results suggest that the pseudocleavage furrow arises as a by-product of compressive flows , mechanically aligning actin filaments into a ring even in the absence of localized RhoA activation . 10 . 7554/eLife . 17807 . 022Figure 4 . Flows and compression are required to generate alignment .\",\n \"( a-d )\",\n \"Average AP profiles of gel flow ( light green , smoothened ) , compression rates ( dark green ) and nematic order parameter ( blue ) at the time of stationary flow during pseudocleavage onset for mlc-4 5–7 hr and 7–9 hr , nop-1 and ani-1 RNAi .\",\n \"Error bars represent the standard error of the mean ( N = 17 , 14 , 10 , 22 for a-d , respectively ) .\",\n \"Dashed red line indicate the respective best-fit theory predictions for the nematic order parameter ( blue ) .\",\n \"For ani-1 and mlc-4 ( 5–7 hr ) , τ is small and is determined to be smaller than 0 . 5 min .\",\n \"For nop-1 and mlc-4 ( 7–9 hr ) , theory profiles were generated using the parameters obtained for the unperturbed non-RNAi pseudocleavage , since insufficient compression rates did not constrain the physical parameters .\",\n \"Scale bar , 10 μm . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02210 . 7554/eLife . 17807 . 023Figure 4—figure supplement 1 . Cytoskeletal organization under RNAi perturbation .\",\n \"( a-d )\",\n \"Actin organization and myosin distribution under nop-1 ( RNAi ) , ani-1 ( RNAi ) and mlc-4 ( RNAi ) .\",\n \"Inserts with fluorescence intensity profiles during pseudocleavage and cytokinesis onset are overlaid next to the corresponding images ( actin in blue , myosin in magenta ) .\",\n \"Note that ANI-1 depleted embryos had more elongated and thin anterior halves , with no localized dense actin ring like structure and no clear and stable central furrow with membrane overlapping .\",\n \"Also note that the asymmetry in the distribution of actin ( usually denser in the anterior half of the embryo ) during cytokinesis was reduced for both nop-1 and long mlc-4 ( RNAi ) .\",\n \"Right , combined images and zoom ( green Lifeact , red NMY-2 ) .\",\n \"Scale bar , 10 μm ( full images ) ; 5 μm ( zoom ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02310 . 7554/eLife . 17807 . 024Figure 4—figure supplement 2 . Flows , compression and alignment at cytokinesis onset .\",\n \"( a-d )\",\n \"Results of nop-1 ( RNAi ) , ani-1 ( RNAi ) and mlc-4 ( RNAi ) .\",\n \"Average AP profiles of gel flow ( light green , smoothened ) , compression rates ( dark green ) and nematic order parameter ( blue ) at the time of stationary flow during cytokinesis onset .\",\n \"Error bars represent the standard error of the mean ( for nop-1 , ani-1 , mlc-4 5–7 hr and 7–9 hr , N = 9 , 18 , 9 , 13 , respectively ) .\",\n \"Dashed red line indicates the respective best-fit theory predictions for the nematic order parameter ( blue ) .\",\n \"For cytokinesis mlc-4 ( 7–9 hr ) tau is small and is determined to be smaller than 0 . 5 min . ( e-f ) Overlay of the different profiles at pseudocleavage and cytokinesis onset , non-RNAi is shown in blue and compression rates in dashed lines .\",\n \"Note that the level of outer ingression was similar to the non-RNAi case for ani-1 ( RNAi ) pseudocleavage , but with improper location and no membrane overlapping , thus no full pseudocleavage furrowing was obtained but a remaining initial anterior ingression was achieved in the presence of some filaments alignment . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 024 In this context we wondered if the cortex is specifically modified to favor ring formation in cytokinesis .\",\n \"We used our fitting procedure to determine the parameters that characterize emergent material properties of the cortex under mlc-4 ( RNAi ) .\",\n \"Notably , we observe that 7–9 hr of mlc-4 ( RNAi ) leads to a reduction of the alignment relaxation time scale τ during cytokinesis .\",\n \"The value obtained is similar to the alignment relaxation time scale during pseudocleavage under non-RNAi conditions ( Table 1 and Table 2 ) , leading us to speculate that an increase in myosin activity via Rho signaling is responsible for the observed increase of the characteristic relaxation time τ in cytokinesis .\",\n \"Importantly , increasing τ allows the cell to form a pronounced ring of aligned filaments by compressive cortical flow during cytokinesis , as predicted by theory ( Figure 3—figure supplement 1 ) .\",\n \"Finally , we sought to find support from theory that aligned filaments in the equatorial region drive anisotropic stress generation ( Mayer et al . , 2010 ) to form an ingression ( White and Borisy , 1983; Salbreux et al . , 2009; Turlier et al . , 2014 ) .\",\n \"By use of a theory that connects anisotropic active stress generation in the active nematic gel to changes in cell shape ( see Appendix and Figure 3d ) we show that we cannot appropriately account for the observed ingression profiles unless we consider anisotropic active stress generation in an aligned gel ( Figure 3e , Figure 3—figure supplement 3 ) .\",\n \"This suggests that anisotropic active stress generation in a network of aligned actin filaments is important for forming an ingression .\",\n \"To conclude , we find that compression by flow drives actomyosin ring formation in both pseudocleavage and cytokinesis , as originally proposed by White and Borisy ( White and Borisy , 1983 ) ( Figure 5 ) .\",\n \"For this , we provide a general framework for determining emergent material parameters of the actomyosin gel .\",\n \"This characterizes flow-alignment coupling and allows for capturing essential aspects of the mechanics of furrow generation .\",\n \"Our analysis of the pseudocleavage ingression in C . elegans demonstrates that constricting rings can form in the absence of equatorial RhoA activation .\",\n \"This reveals that cortical flow functions as a central organizer of network architecture , mechanically aligning filaments to form a constricting furrow in the equatorial region without a localized increase of actin nucleation ( Figure 2—figure supplement 5 , f ) and myosin contractility ( Figure 1b-d ) .\",\n \"Finally , it will be interesting to investigate if a cortex that is aligned by compressive flow favors the recruitment of specific actin binding proteins such as bundling or motor proteins ( Robinson et al . , 2002 ) , comprising an interesting mechanism of positive feedback for stabilizing and enhancing the ring during cytokinesis .\",\n \"Our work highlights that determining emergent material properties of actomyosin is important for understanding cytokinesis , and a challenge for the future is to link emergent material properties at the larger scale with molecular mechanisms ( Mendes Pinto et al . , 2013; Eggert et al . , 2006 ) .\"\n ],\n [\n \"Existing reagents did not allow for detailed and long-term imaging of actin filaments in embryos , which is critical for reliably quantifying their orientation .\",\n \"Hence , we generated a new Lifeact transgenetic line with enhanced actin filament labeling ( SWG001 ) .\",\n \"A codon optimized for C . elegans far-red fluorophore , mKate2 kindly shared by Henrik Bringman ( Redemann et al . , 2011 ) ( 26 kDa , 588/633 nm ) was added to the actin probe with a linker ( 66 bp ) and cloned into MOSCI vector containing unc119 rescue gene ( Hyman Lab ) under the control of the mex-5 promoter .\",\n \"Stable integration was obtained by bombardment of this plasmid in DP38 strain .\",\n \"This strain was then backcrossed with N2 .\",\n \"In this line , we observed within the cortical plane bright actin filament labeling with a high signal-to-noise ratio and little photobleaching .\",\n \"Importantly , overall cortical organization and flow dynamics appeared to not be affected by this reagent , foci lifetime , spacing , cortical flow velocities were similar to previous measurement with other fluorescent lines .\",\n \"A dual colored transgenic line was obtained in order to image simultaneously actin and myosin dynamics by crossing Lifeact::mKate2 strain with LP133 ( NMY-2::GFP obtained by CRISPR ( Dickinson et al . , 2013 ) , SWG007 ) .\",\n \"The RhoA biosensor used for Figure 1 was developed by the Glotzer lab ( GFP::AHPH , C . elegans strain MG617 , Tse et al . , 2012 ) .\",\n \"This sensor consists of GFP fused to the C-terminal portion of C . elegans anillin , which contain its conserved region ( AH ) and pleckstrin homology ( PH ) domain .\",\n \"It lacks the N-terminal myosin and actin-binding domains but retains its RhoA-binding domain .\",\n \"The strains expressing CYK-1:GFP ( SWG004 ) and PLST-1:GFP ( SWG005 ) were obtained by the insertion of a GFP tag at the endogenous locus in their C-terminal region using CRISPR according to Dickinson et al ( Dickinson et al . , 2013 ) .\",\n \"C . elegans worms were cultured on OP50-seeeded NGM agar plates as described ( Brenner , 1974 ) .\",\n \"L4 mothers were maintained overnight at 20°C before dissection in M9 buffer .\",\n \"For imaging , one-cell embryos were mounted on 2% agar pads , thereby slightly compressed to increase the cortical surface visible in a confocal plane .\",\n \"Images were acquired with spinning disk confocal microscope ( Zeiss C-Apochromat , 63X/1 . 2 NA or 100X/1 . 42 NA , Yokogawa CSU-X1 scan head and Hamamatsu Orca-Flash4 . 0 camera ) every 2 s on two or three different planes ( one or two at the bottom for a cortical section , and one 12 µm above for a medial section of the embryo ) .\",\n \"For 3D acquisitions , Z-stacks were acquired every 0 . 2 μm .\",\n \"We assume the shape to be rotationally symmetrical to the longer axis , thus the outline of one medial section of the embryo reflects faithfully cell shape changes and ingression distance .\",\n \"RNAi experiments were performed by feeding ( Timmons et al . , 2001 ) .\",\n \"L4 worms were grown at 20°C on feeding plate ( NGM agar containing 1 mM isopropyl-β-D-thiogalactoside and 50 μg ml−1 ampicillin ) for the required number of hours before imaging .\",\n \"Spatiotemporal correlation functions were calculated for all time points during the stationary period , in the 30 μm central region of the embryo .\",\n \"The plot of the spatial correlation only ( 1D correlation function ) is shown in Figure 2c between the compression , alignment and ingression data sets .\",\n \"Peak value positions are obtained by a Gaussian fitting procedure and allow the calculation of positional differences between the different data sets .\",\n \"Because the flow field appears stationary , we expect this distance between peaks of compression , alignment and ingression to arise due to temporal delays between these events in the reference frame co-moving with the cortical flows .\",\n \"We thus translated the characteristic distances between stationary peaks to temporal delays using the average flow velocity in the −10 μm to 0 central region ( 6 . 16 ± 1 . 23 µm/min for pseudocleavage and 4 . 01 ± 1 . 92 µm/min for cytokinesis ) ( Figure 2d ) . 10 . 7554/eLife . 17807 . 025Figure 5 . Compression drives alignment for furrowing . Schematic illustration of flows , cytoskeletal organization and cell shape changes during pseudocleavage and at the onset of cytokinesis . DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02510 . 7554/eLife . 17807 . 026Video 5 . Actomyosin gel dynamics in the C . elegans zygote after 24 hr nop-1 ( RNAi ) .\",\n \"Cortical and medial planes of an embryo expressing both Lifeact::mKate2 and endogenous NMY-2::GFP .\",\n \"Left panel , medial plane Lifeact:mKate2 , center , cortical NMY-2::GFP , right panel , cortical Lifeact:mKate2 ( min:s ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02610 . 7554/eLife . 17807 . 027Video 6 . Actomyosin gel dynamics in the C . elegans zygote after 26 hr ani-1 ( RNAi ) .\",\n \"Cortical and medial planes of an embryo expressing both Lifeact::mKate2 and endogenous NMY-2::GFP .\",\n \"Left panel , medial plane Lifeact:mKate2 , center , cortical NMY-2::GFP , right panel , cortical Lifeact:mKate2 ( min:s ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 02710 . 7554/eLife . 17807 . 028Video 7 . Actomyosin gel dynamics in the C . elegans zygote after 14 hr mlc-4 ( RNAi ) .\",\n \"Cortical and medial planes of an embryo expressing both Lifeact::mKate2 and endogenous NMY-2::GFP .\",\n \"Left panel , medial plane Lifeact:mKate2 , center , cortical NMY-2::GFP , right panel , cortical Lifeact:mKate2 ( min:s ) .\",\n \"DOI: http://dx . doi . org/10 . 7554/eLife . 17807 . 028\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Cytokinesis in eukaryotic cells is often accompanied by actomyosin cortical flow .","Over 30 years ago , Borisy and White proposed that cortical flow converging upon the cell equator compresses the actomyosin network to mechanically align actin filaments .","However , actin filaments also align via search-and-capture , and to what extent compression by flow or active alignment drive furrow formation remains unclear .","Here , we quantify the dynamical organization of actin filaments at the onset of ring assembly in the C . elegans zygote , and provide a framework for determining emergent actomyosin material parameters by the use of active nematic gel theory .","We characterize flow-alignment coupling , and verify at a quantitative level that compression by flow drives ring formation .","Finally , we find that active alignment enhances but is not required for ring formation .","Our work characterizes the physical mechanisms of actomyosin ring formation and highlights the role of flow as a central organizer of actomyosin network architecture ."],"string":"[\n \"Cytokinesis in eukaryotic cells is often accompanied by actomyosin cortical flow .\",\n \"Over 30 years ago , Borisy and White proposed that cortical flow converging upon the cell equator compresses the actomyosin network to mechanically align actin filaments .\",\n \"However , actin filaments also align via search-and-capture , and to what extent compression by flow or active alignment drive furrow formation remains unclear .\",\n \"Here , we quantify the dynamical organization of actin filaments at the onset of ring assembly in the C . elegans zygote , and provide a framework for determining emergent actomyosin material parameters by the use of active nematic gel theory .\",\n \"We characterize flow-alignment coupling , and verify at a quantitative level that compression by flow drives ring formation .\",\n \"Finally , we find that active alignment enhances but is not required for ring formation .\",\n \"Our work characterizes the physical mechanisms of actomyosin ring formation and highlights the role of flow as a central organizer of actomyosin network architecture .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Just under the surface of every animal cell , a thin and dynamic network of filaments called the cell cortex acts as a scaffold and determines the cell’s shape .","When the cell divides , this material re-organizes to make a ring of filaments – known as the cytokinetic ring – across the middle of the cell .","This ring then constricts to split the cell into two separate daughter cells .","The filaments are guided to form the ring by specific proteins around the middle of the cell .","A process called cortical flow – the mechanical compression of filaments towards the middle – also influences the shape of the ring .","However , it is not clear to what degree cortical flow actually helps the ring to form .","A tiny worm called Caenorhabditis elegans is often used to study how animal cells divide and grow .","When the C . elegans embryo is made of just a single cell , two rings of filaments form consecutively as this cell prepares to divide .","Reymann et al . used microscopy to investigate how filaments are arranged in C . elegans embryos as the rings assemble .","The experiments showed that filaments are arranged into rings in locations where the filaments are being mechanically compressed by cortical flow .","The first ring forms and partially constricts , and then relaxes once the cell is polarized; that is , once the cell has developed two distinct ends .","A second ring then forms during cytokinesis and constricts to divide the cell into two .","To understand the physical changes occurring , Reymann et al . compared the experimental data with a mathematical model of the cortical network .","This model assumed that the cortical network acts as a thin film in which the orientation of the filaments is coupled to the flow of the fluid .","Reymann et al . used this model to demonstrate that the observed arrangements of the filaments in both rings can be explained by cortical flow .","Together , the findings of Reymann et al . highlight the central role that cortical flow plays in organizing rings of filaments in C . elegans .","Future studies will explore whether cortical flow is linked to other mechanisms that affect the formation of the cytokinetic ring ."],"string":"[\n \"Just under the surface of every animal cell , a thin and dynamic network of filaments called the cell cortex acts as a scaffold and determines the cell’s shape .\",\n \"When the cell divides , this material re-organizes to make a ring of filaments – known as the cytokinetic ring – across the middle of the cell .\",\n \"This ring then constricts to split the cell into two separate daughter cells .\",\n \"The filaments are guided to form the ring by specific proteins around the middle of the cell .\",\n \"A process called cortical flow – the mechanical compression of filaments towards the middle – also influences the shape of the ring .\",\n \"However , it is not clear to what degree cortical flow actually helps the ring to form .\",\n \"A tiny worm called Caenorhabditis elegans is often used to study how animal cells divide and grow .\",\n \"When the C . elegans embryo is made of just a single cell , two rings of filaments form consecutively as this cell prepares to divide .\",\n \"Reymann et al . used microscopy to investigate how filaments are arranged in C . elegans embryos as the rings assemble .\",\n \"The experiments showed that filaments are arranged into rings in locations where the filaments are being mechanically compressed by cortical flow .\",\n \"The first ring forms and partially constricts , and then relaxes once the cell is polarized; that is , once the cell has developed two distinct ends .\",\n \"A second ring then forms during cytokinesis and constricts to divide the cell into two .\",\n \"To understand the physical changes occurring , Reymann et al . compared the experimental data with a mathematical model of the cortical network .\",\n \"This model assumed that the cortical network acts as a thin film in which the orientation of the filaments is coupled to the flow of the fluid .\",\n \"Reymann et al . used this model to demonstrate that the observed arrangements of the filaments in both rings can be explained by cortical flow .\",\n \"Together , the findings of Reymann et al . highlight the central role that cortical flow plays in organizing rings of filaments in C . elegans .\",\n \"Future studies will explore whether cortical flow is linked to other mechanisms that affect the formation of the cytokinetic ring .\"\n]","isConversational":false},"year":{"kind":"string","value":"2016"}}},{"rowIdx":99,"cells":{"headings":{"kind":"list like","value":["Introduction","Results","Discussion","Materials and methods"],"string":"[\n \"Introduction\",\n \"Results\",\n \"Discussion\",\n \"Materials and methods\"\n]","isConversational":false},"keywords":{"kind":"list like","value":["chromosomes and gene expression","computational and systems biology"],"string":"[\n \"chromosomes and gene expression\",\n \"computational and systems biology\"\n]","isConversational":false},"title":{"kind":"string","value":"Simple biochemical features underlie transcriptional activation domain diversity and dynamic, fuzzy binding to Mediator"},"id":{"kind":"string","value":"elife-68068-v3"},"sections":{"kind":"list like","value":[["Transcription factors ( TFs ) perform the last step in signal transduction pathways .","They thus serve key roles in central processes such as growth , stress response , and development , and their mutation or misregulation underlies many human diseases ( Spitz and Furlong , 2012 ) .","A TF includes a sequence-specific DNA-binding domain ( DBD ) and an effector domain that regulates nearby gene transcription .","Activation domains ( ADs ) – effector domains that increase transcription – have long been of particular interest due to their roles as oncogenic drivers and use as scientific tools ( Bradner et al . , 2017; Brückner et al . , 2009; Dominguez et al . , 2016 ) .","ADs were discovered as regions that could independently stimulate transcription when ectopically recruited to a gene promoter ( Brent and Ptashne , 1985 ) .","Early experiments showed that ADs were unlike structured domains because progressive truncations showed graded reductions in activity ( Hope and Struhl , 1986; Hope et al . , 1988 ) .","Subsequent studies showed that ADs were disordered and had few similarities in their primary sequence ( Mitchell and Tjian , 1989 ) .","Instead , ADs were classified based on their enrichment of certain residues , whether acidic , glutamine-rich , or proline-rich .","Acidic ADs are the most common and best characterized .","Acidic ADs retain activity when transferred between yeast and animals , pointing to a conserved eukaryotic mechanism ( Fischer et al . , 1988; Struhl , 1988 ) .","While some have found that acidic residues are necessary for activation , others have found that they are dispensable ( Brzovic et al . , 2011; Pacheco et al . , 2018; Staller et al . , 2018; Warfield et al . , 2014 ) .","Besides their negative charge , acidic ADs are rich in bulky hydrophobic residues .","Mutating these hydrophobic residues reduces activation , often in proportion to the number mutated .","Because AD sequences are highly diverse and poorly conserved , only a small fraction of all ADs in eukaryotic TFs have likely been annotated .","Sequence motifs have been proposed based on analysis of select ADs but have not been used for large-scale prediction ( Piskacek et al . , 2007 ) .","Screens of random sequences in yeast identified many activating sequences that represented as many as 1–4% of elements tested ( Hackett et al . , 2020; Ravarani et al . , 2018 ) .","Actual protein sequences are , however , highly non-random .","Direct screening of protein sequences has identified relatively few ADs at low resolution ( Arnold et al . , 2018; Tycko et al . , 2020 ) .","There is a need for methods to experimentally detect or computationally predict all ADs .","ADs stimulate transcription of genes by recruiting coactivator complexes , especially the Mediator complex , which interacts with RNA Polymerase II and regulates its transcription initiation ( Kornberg , 2005 ) .","Mediator is necessary for TF-dependent activation in vitro and in vivo and is required for regulation by enhancers in all eukaryotes ( Allen and Taatjes , 2015 ) .","Mediator and TFs are concentrated at strong enhancers , perhaps in a phase-separated state , which may play a role in gene activation ( Boija et al . , 2018; Chong et al . , 2018; Sabari et al . , 2018; Shrinivas et al . , 2019; Whyte et al . , 2013 ) .","Beyond Mediator , TFs have been suggested to recruit other conserved multiprotein complexes , including TFIID , SAGA , and SWI/SNF , that play roles in activation of various genes ( Hahn and Young , 2011; Mitchell and Tjian , 1989 ) .","While many such TF interactions have been described , their occurrence and roles remain to be determined .","Biochemical studies of TF-coactivator interaction have given insight into the mechanism of activation .","Nuclear magnetic resonance ( NMR ) studies revealed short alpha helices formed by acidic ADs of yeast Gcn4 protein , with hydrophobic faces contacting hydrophobic surfaces of the Med15 subunit of Mediator ( Brzovic et al . , 2011; Tuttle et al . , 2018 ) .","NMR constraints were consistent with multiple possible AD binding poses , suggesting a dynamic ‘fuzzy complex’ ( Tompa and Fuxreiter , 2008 ) .","Further consistent with these ideas , solvent exposure of aromatic residues was associated with activation in a screen of Gcn4 AD variants ( Staller et al . , 2018 ) .","Here , we combine quantitative , high-throughput measurements of in vivo activation and in vitro interaction with computational modeling to characterize ADs .","We identify 150 ADs in budding yeast and describe their shared sequence attributes .","We extend the analysis by training and validating a neural network that predicts new ADs across eukaryotes .","Guided by predictions , we design and measure activation of thousands of AD mutants to derive a deeper understanding of the principles underlying activation .","Correlating activation with measurements of binding of ADs to Mediator in vitro identifies the key protein interactions driving activation , for which we predict atomic structures using a peptide-docking algorithm .","We derive structural features common to AD-Mediator interactions and use them to explain measurements of Mediator recruitment kinetics in vitro .","In total , we develop an integrated model for function , sequence determinants , molecular interactions , and kinetic features of TF ADs ."],["We identified ADs across all TFs in budding yeast with the use of a quantitative , uniform , and high-throughput activation assay .","Variability due to protein expression , activation duration , and secondary genetic effects was minimized by fusing domains of interest to a three-part artificial TF ( aTF ) that ( 1 ) is tracked by an mCherry tag , ( 2 ) localizes to the nucleus only upon induction with estrogen , and ( 3 ) binds uniquely through its mouse DBD in the promoter of a chromosomally-integrated GFP reporter gene ( Figure 1A; McIsaac et al . , 2013; Staller et al . , 2018 ) .","In this assay , three known ADs stimulated GFP expression greater than 100-fold upon estrogen treatment ( Figure 1—figure supplement 1A ) .","We used a pooled screen to measure activation by 7460 protein segments , each 53 amino acids ( aa ) in length , that tiled all 164 TFs ( identified by Gene Ontology terms ) with a step size of 12–13 aa ( 3 . 8-fold average coverage; Figure 1B ) .","Outgrowth of transformed yeast for 5 days ensured that each cell contained a unique aTF expression plasmid due to mechanisms maintaining it at single-copy levels ( Materials and methods ) ( Scanlon et al . , 2009 ) .","After induction with estrogen , cells with a defined level of aTF expression were selected by mCherry signal and sorted into eight bins based on GFP expression ( Figure 1—figure supplement 1B ) .","By next-generation sequencing , the distribution of each protein tile across the bins was determined and the mean value was used to calculate activation—namely , the fold increase in GFP relative to background ( Figure 1—source data 1 , Materials and methods ) .","Activation was highly concordant between distinct DNA sequences encoding the same protein fragment ( Figure 1—figure supplement 1C ) .","GFP distributions of three known ADs in the pooled screen exactly reproduced measurements from individual pure populations ( Figure 1—figure supplement 1D ) .","Controlling by a defined aTF expression level was critical for precisely measuring activation ( Figure 1—figure supplement 1E–F ) .","Activation measurements of fragments shared between two independently cloned and assayed libraries were reproducible ( Pearson's r = 0 . 953 , Figure 1—figure supplement 1G ) .","Over 2 , 850 , 000 cells were assayed in total , giving excellent coverage of library elements ( approximately 99% of TF tiles were observed in at least 10 cells each ) .","Our assay exhibited high signal-to-noise: tiles in previously known and newly identified ADs activated nearly 200-fold while 88% of fragments activated less than twofold ( Figure 1C ) .","Across the library , 451 tiles showed significant activation ( p<0 . 0001 by Z-test ) .","When plotted by protein position , activating tiles clustered into discrete , well-defined ADs ( Figure 1D ) .","Using a positional activation score , we identified 150 ADs in 96 TFs ( Figure 1E and Figure 1—figure supplement 1H , Figure 1—source data 2 , Materials and methods ) .","These ADs overlapped 75% of all previously-reported ADs in TFs ( Figure 1—source data 2 ) .","Furthermore , the 53-aa tile length was not limiting , since our screen successfully identified ADs in over 85% of TFs that activated in a previous one-hybrid screen testing full-length proteins ( Figure 1—source data 2; Titz et al . , 2006 ) .","These results show that our screen is both sensitive and comprehensive .","Three-quarters ( 112 ) of ADs identified here were previously unknown ( Figure 1—source data 2 ) .","While 63 TFs contained just a single AD , 33 TFs had multiple , including up to seven distinct ADs in Adr1 ( Figure 1D and F ) .","C-terminal ADs were common—found in nearly half of all AD-containing TFs—and were stronger and shorter on average than other ADs ( Figure 1—figure supplement 1I ) .","Consistent with AD function in their native context , TFs that contained ADs upregulated a higher proportion of downstream genes than TFs without ADs ( Figure 1—figure supplement 1J; Hackett et al . , 2020 ) .","Consistent with the diverse functions of TFs beyond transcriptional activation ( e . g . repression ) , 68 TFs did not contain any ADs ( Figure 1F ) .","ADs were enriched in negative charge and hydrophobic content , and activation was strongest for fragments that were high in both ( Figure 2A and Figure 2—figure supplement 1A–B; Wimley-White hydrophobicity scale ) ( Wimley and White , 1996 ) .","Past experiments have clearly shown the importance of bulky hydrophobic residues but presented conflicting evidence for acidic residues in activation ( Brzovic et al . , 2011; Cress and Triezenberg , 1991; Drysdale et al . , 1995; Jackson et al . , 1996; Pacheco et al . , 2018; Staller et al . , 2018; Sullivan et al . , 1998; Warfield et al . , 2014 ) .","To determine the role of negative charge , we took the strongest-activating fragment from each AD and measured its activation with all acidic residues mutated .","Except for a single example that activated just as strongly at +7 net charge without its acidic residues ( the Gln3 C-terminal AD ) , activation was abolished in all ADs , showing definitively that acidic residues are necessary in budding yeast ADs ( Figure 2B and Figure 2—figure supplement 1C ) .","We also noticed that activation of wild-type tiles was more strongly correlated with the number of Asp residues than the number of Glu residues , and hypothesized that Asp promotes activation more strongly than Glu ( Figure 2—figure supplement 1D ) .","Indeed , ADs with all Glu residues mutated to Asp consistently activated as well as or better than ADs with all Asp residues mutated to Glu ( Figure 2—figure supplement 1E ) .","Having identified the important residue types , we looked for sequence motifs that could predict activation .","However , AD sequences appeared to share no common motifs upon inspection .","The 9aaTAD motif , originally proposed based on yeast ADs ( Piskacek et al . , 2007 ) , predicted ADs poorly: only 10% of 9aaTAD-containing tiles activated in our assay , which was not significantly more predictive than scrambled versions of the motif ( Figure 2C ) .","A de novo search for motifs using DREME also found none that were shared by more than two ADs ( Bailey , 2011 ) .","Evidently , ADs are diverse in sequence and not predictable from simple motifs .","The lack of motifs suggests that activation is determined not by positions of individual residues but by distributed or redundant features .","We pursued this point by measuring activation of a panel of designed mutants of eight ADs .","Indeed , 94% of scanning single-amino acid mutants affected activation by less than twofold , and all still activated more than 10-fold ( Figure 2—figure supplement 1F–G ) .","To produce larger effects , we varied the acidic and hydrophobic content of the ADs by mutating successively larger subsets of acidic or aromatic residues ( Figure 2D , Materials and methods ) .","Activation was gradually , and finally completely , eliminated as either characteristic was reduced .","It is noteworthy that the Gcn4 AD , a focus of many past studies , uniquely tolerated mutation of up to six acidic residues .","Most notably , activation by acidic and hydrophobic mutants were related to the number , rather than position , of mutated residues ( Figure 2D and Figure 2—figure supplement 1H ) .","Thus , activation strength is determined in large part by acidic and hydrophobic content , not by motifs .","Since motifs predicted ADs poorly , we turned to more sophisticated approaches using machine learning .","We first trained a simple neural network to predict activation based solely on tiles’ amino acid ( aa ) composition ( Materials and methods ) .","When tested on TF tiles withheld from training , this neural network performed remarkably well , explaining 66% of observed variation ( Figure 3—figure supplement 1A ) .","However , this algorithm was limited by its lack of information about residue positions .","To experimentally disentangle the contributions of aa positioning and aa composition , we measured activation of eight scrambled sequences from each of the eight ADs tested before .","Despite their shared abundance of acidic and hydrophobic residues , these mutants spanned a wide range of activation strengths , and some mutants activated stronger than wild-type while others failed to activate entirely ( Figure 3A ) .","Clearly , full sequence information beyond aa composition is needed to predict activation .","We therefore trained a deep convolutional neural network ( CNN ) to predict activation based on protein sequence , predicted secondary structure , and predicted disorder ( Materials and methods ) .","CNNs evaluate sequences by hierarchically integrating matches to a suite of learned patterns and have recently found great success in many genomic prediction tasks ( Ching et al . , 2018; Kelley et al . , 2016; Wang et al . , 2016 ) .","Our Predictor of Activation Domains using Deep Learning in Eukaryotes , or ‘PADDLE’ , explained 81% of observed variation in TF tiles withheld from training ( alternatively , area under the precision-recall curve of 0 . 805; Figure 3—figure supplement 1B–C , Figure 3—source data 1 ) , markedly better than the aa composition-based predictor .","De novo , PADDLE accurately predicted the activation strength of ( 1 ) new ADs within TFs omitted from training ( R2 score = 0 . 81 across tiles from 22 TFs; example predictions for Arg81 in Figure 3B ) ; ( 2 ) scrambled AD sequences , despite their identical amino acid composition ( R2 score = 0 . 61 ) ; and ( 3 ) 232 mutants and 178 orthologs of the Pdr1 AD ( R2 score = 0 . 90 ) ( Figure 3—figure supplement 1B ) .","Altogether , the performance of PADDLE was validated across a wide range of both wild-type and mutant sequences in yeast .","Classic experiments showed that many acidic ADs retained activity even when transferred between yeast and animals ( Fischer et al . , 1988; Struhl , 1988 ) , so we investigated whether PADDLE could identify ADs in human TFs that would activate in human cells .","Using PADDLE , we predicted 236 high-strength and 366 moderate-strength ADs , together spanning 462 ( 27% ) human TFs ( Figure 3C , Figure 3—source data 1 ) .","These predicted ADs overlapped many known ADs of TFs from diverse families , including p53 , NFkB , Myc , Klf4 , Fos , PPARA , SREBF1 , E2F proteins , and the glucocorticoid receptor ( Figure 3—figure supplement 1D ) .","PADDLE also predicted 41 high-strength and 45 moderate-strength ADs from among 419 transcription-regulating viral proteins ( Figure 3—source data 1; Liu et al . , 2020 ) .","We randomly selected 25 high-strength predicted ADs from human TFs and measured their activation individually using a luciferase reporter in HEK293T cells .","Remarkably , 23 domains ( 92% ) activated luciferase expression ( Figure 3D ) .","While other classes of human ADs , such as glutamine-rich or proline-rich ADs , are not predictable by PADDLE since they are not present in S . cerevisiae , acidic activation mechanisms are evidently conserved and the patterns learned by PADDLE are generalizable across eukaryotes .","To uncover the principles of acidic activation learned by PADDLE , we analyzed predictions on designed mutant sequences and developed hypotheses , which we then experimentally tested in a second library .","We noticed from predictions on sequences less than 30 aa that yeast ADs contained short , independently activating regions , which we term core ADs ( cADs ) ( Figure 3—figure supplement 1E ) .","As an experimental test , we measured in vivo activation of 13-aa-long fragments tiling ten ADs in 1-aa steps .","Predictions proved highly accurate ( R2 score = 0 . 84 , Figure 3—figure supplement 1F ) , and both predictions and experiments identified at least one significantly activating 13-aa cAD within each AD ( Figure 3E; p<0 . 001 ) .","Across all yeast TFs , PADDLE predicted that the minimal region for activation could be localized to within a 20-aa cAD in 85% of ADs ( Figure 3F , Figure 3—source data 1 ) .","These cADs still shared no motifs ( Figure 3—figure supplement 1G ) , but instead had an especially high density of acidic and hydrophobic residues , even more so than the full ADs ( Figure 3G ) .","To quantify the contribution of each residue to activation , we predicted the impact of mutating each position in each cAD to alanine ( Figure 3—figure supplement 1H ) .","Single-aa mutants in these short cADs had a large predicted impact on activation , unlike in the longer 53-aa ADs ( Figure 2—figure supplement 1F–G ) , and showed clear trends when grouped by residue identity .","Mutating the bulkiest hydrophobic residues led to the greatest decreases in predicted activation , and the three most important were Trp , Phe , and Leu ( Figure 3—figure supplement 1I ) .","Mutating acidic or basic residues had opposite but comparable effects on predicted activation .","Together , these analyses identify the regions and residues across ADs that contribute most to activation .","Having identified the core AD regions , we sought to understand the sequence properties that drive their activation .","We focused on the 28 strongest 13-aa cADs ( Figure 3—source data 1 ) , and first experimentally quantified the importance of their aa composition by comparing each cAD with 33 random scrambles of its sequence .","Remarkably , only nine cADs ( 32% ) showed activation by the wild-type sequence greater than threefold greater than the average activation by scrambled sequences ( Figure 4A ) .","In these nine cADs , maximal activation evidently requires a specific positioning of residues .","On the other hand , 18 cADs ( 64% ) showed activation by the wild-type sequence within twofold of the scrambled sequence average , indicating that their activation is primarily determined by aa composition and not by a unique positioning of residues .","The wild-type-to-scramble ratio was not correlated with wild-type activation strength ( Figure 4—figure supplement 1A , Pearson’s r = 0 . 32 , p=0 . 09 ) .","Instead , cADs with the greatest hydrophobic content had the lowest wild-type-to-scramble ratios ( Figure 4B , Pearson’s r = −0 . 51 , p=0 . 006 ) .","Thus , the majority of cADs studied here activate by a composition-driven mechanism that exploits an excess of bulky hydrophobic residues .","Alpha-helical folding is thought to be important for activation , which is at odds with the prevalence of composition-driven cADs .","We directly determined whether the 28 strongest 13-aa cADs require helical folding by measuring the impact of inserting a helix-breaking proline .","Within the seven central positions of each 13-aa cAD , we individually mutated each non-hydrophobic residue to alanine or proline and asked whether proline inhibited activation relative to alanine ( Figure 4C ) .","Proline mutations inhibited activation more than threefold over alanine in nine ADs ( 32% ) , including up to an 11-fold drop in activation in the Tda9 cAD .","However , proline mutations inhibited activation by less than twofold in 16 cADs ( 58% ) .","These effects were consistent between different positions within the same cAD ( Figure 4—figure supplement 1B ) .","Interestingly , all three cADs that contained a basic residue ( Pul4 , Tda9 , and Rsf2:586 ) were strongly inhibited by proline , suggesting that a helix is necessary to position their inhibitory positive charge away from the coactivator binding interface .","Most notably , the magnitude of proline disruption was tightly correlated with the wild-type-to-scramble ratio ( Figure 4D , Pearson’s r = 0 . 842 ) , showing that constraints on structure and sequence , or a lack thereof , were closely linked .","Thus , composition-driven cADs typically do not need to form a helix and likely sample disordered conformations , while cADs that require their wild-type sequence typically also require helical folding , employing a structure-driven mechanism .","To understand these two contrasting mechanisms , we studied them in a simplified system of artificial cADs that used only the four most activation-promoting residues .","We examined 9-aa sequences ( denoted ‘9mers’ ) consisting entirely of only two types of amino acids each—Asp and either Leu , Phe , or Trp—so that all possible such sequences ( 3 × 29 = 1536 ) could be systematically assayed ( Figure 4E ) .","As expected , activation required a balance of hydrophobic and acidic residues ( Figure 4—figure supplement 1C ) , so we focused analysis on aa compositions with the strongest median activation: 9mers with five or six hydrophobic residues .","Within these , nearly all Phe 9mers and Trp 9mers activated regardless of their sequence ( Figure 4F ) , characteristic of a composition-driven mechanism and consistent with the highly hydrophobic nature of Phe and Trp residues .","However , Leu 9mers spanned a wide range of activity with only a minority that were strongly activating , characteristic of structure-driven activation requiring a specific ordering of residues .","To see what allows only certain Leu 9mers to activate , we examined all possible short motifs containing two or three Leu residues .","The only motif strongly associated with activation was LxxLL ( Figure 4G and Figure 4—figure supplement 1D; p<10−8 by Kolmogorov-Smirnov test ) .","This motif was strictly required for activation in these 9mers , present in two copies in the three strongest 9mers , found in two wild-type structure-driven cADs ( Put3 and Crz1 ) , and previously described in ADs of many nuclear receptors and other TFs ( Plevin et al . , 2005 ) .","This motif also suggests an explanation for why helical folding was necessary for cADs with less hydrophobicity: by placing Leu residues along one face of an alpha helix , the motif efficiently forms a hydrophobic binding surface .","The analogous FxxFF and WxxWW motifs were also significantly but more weakly associated with activation of Phe 9mers and Trp 9mers ( Figure 4—figure supplement 1E ) .","However , their strongest predictor of activation was the average position of acidic residues , with N-terminal skew highly favored ( Figure 4H ) .","Most dramatically , DDDFFFFFF activated 20-fold while its reverse sequence FFFFFFDDD activated just 1 . 3-fold .","One role of N-terminal negative charge could be to neutralize the macroscopic dipole that arises from alpha helix backbone hydrogen bonds , stabilizing helical conformations ( Rocklin et al . , 2017 ) .","Even though a helix is not necessary for composition-driven activation , this effect could reduce the entropic cost of binding to coactivators by promoting transient folding .","Despite the importance of hydrophobic residues for activation , PADDLE also predicted that some TFs contained cADs whose activation was inhibited by nearby clusters of hydrophobic residues .","We confirmed this experimentally for Rds1 , in which residues 239–263 alone activated 6 . 8-fold but extending the sequence by eight aa , five of them hydrophobic , abolished activation ( Figure 4I ) .","Inhibition by hydrophobic residues also explains the weak activation of some scrambled 53-aa ADs ( Figure 3A , Figure 4—figure supplement 1F ) .","To systematically quantify inhibition , we measured the effect on activation when the Pdr1 cAD was placed next to a library of 2177 random 20-aa sequences chosen to span a wide range of net charge and hydrophobic content ( Figure 4J ) .","As expected , negative charge with moderate hydrophobicity bolstered activation .","Positive charge without hydrophobicity inhibited activation , but even variable regions with +8 charge ( +4 net charge ) still activated an average of 4 . 4-fold .","Therefore , the local clustering of negative charges near hydrophobic residues renders the Pdr1 cAD resistant to global changes in net charge , suggesting that interactions between opposite charges within the peptide do not fully prevent binding to coactivators .","In contrast , hydrophobic residues at high density inhibited activation completely .","This suggests that inhibitory hydrophobic clusters interact with hydrophobic residues in the cAD , completely preventing their interaction with coactivators .","Consistent with two independent mechanisms of inhibition , basic and hydrophobic residues together reduced activation additively .","In total , these experiments mapped the biochemical and structural properties that drive both activation and inhibition .","The surprising ability of simple biochemical features to explain a large and diverse set of ADs suggests that these features may also drive interactions with coactivator complexes .","To gain a mechanistic understanding of the AD-coactivator contacts that underlie activation , we mapped the interactions of wild-type and mutant ADs with Mediator .","We focused on Med15 , the primary subunit of Mediator targeted by many ADs in budding yeast , which contains four tandem activator-binding domains ( ABDs ) ( Herbig et al . , 2010; Jedidi et al . , 2010; Thakur et al . , 2008 ) .","We isolated the N-terminal portion of Med15 consisting of its four ABDs ( hereafter just called 'Med15' ) , confirmed its in vitro interaction with Gcn4 , and used it for binding experiments ( Figure 5—figure supplement 1A ) .","To measure binding in high-throughput , we used mRNA display ( Takahashi and Roberts , 2009 ) , expressing our library of TF tiles as a pool of protein fragments covalently tagged with their mRNA sequences , and using this pool in Med15 pull-down experiments ( Figure 5A ) .","Direct counts of Med15-bound and input protein molecules were obtained by amplifying and sequencing their mRNA tags and using unique molecular identifiers ( UMIs ) to remove PCR duplicates .","Finally , the fractional pull-down of each protein fragment was computed by its abundance in the Med15-bound sample relative to input , normalized to total library concentrations measured by qPCR ( Materials and methods ) .","Pull-down measurements were reproducible between replicates ( Pearson’s r > 0 . 90 ) and consistent between identical protein fragments encoded by distinct but synonymous mRNA sequences ( Figure 5—figure supplement 1B–C ) .","Fragments known to bind Med15 enriched up to 17-fold above random sequence controls and 1000-fold higher than in mock pull-downs with beads alone ( Figure 5—figure supplement 1D , Figure 5—source data 1 ) .","Binding to Med15 was widespread among ADs and coincident with activation activity .","In total , 324 tiles bound Med15 significantly more than negative controls ( p<0 . 001 by Z-test , Figure 5—figure supplement 1E , Materials and methods ) .","When plotted by protein position , Med15-binding tiles clustered into 153 discrete domains , the vast majority of which were not previously known to bind Mediator ( Figure 5B , Figure 5—source data 1 ) .","ADs and Med15-binding domains overlapped substantially: 73% of ADs bound Med15 , and 71% of Med15-binding domains functioned as ADs ( Figure 5C ) .","Moreover , ADs that did not bind Med15 tended to activate weakly ( Figure 5—figure supplement 1F ) , so it is possible that many bind Med15 at levels below the detection limit of our pull-down assay , which was less sensitive than our activation assay ( compare Figure 1C and Figure 5—figure supplement 1E ) .","Just as with activation , high acidic and hydrophobic content were associated with Med15 binding ( Figure 5—figure supplement 1G ) .","The Gln3 AD that did not require acidic residues also did not bind Med15 substantially ( 2 . 1-fold enrichment ) , suggesting that it activates through other mechanisms .","Most notably , activation strength and Med15 binding of tiles were directly proportional ( Figure 5D ) .","To test this relationship further , we measured Med15 binding of the set of AD mutants in which aromatic residues were systematically removed ( Figure 2D ) .","In every case , Med15 binding and activation were strongly correlated ( Figure 5E ) , showing that the hydrophobic features that drive activation similarly underlie Med15 binding .","Moreover , the ability for binding of a single protein in vitro to quantitatively explain activation in vivo suggests that Mediator recruitment is a key driver for gene activation and determined largely by AD binding to its Med15 subunit .","ADs can bind different coactivators , though the extent of such interaction is unknown .","To explore the possibility , we also examined binding of TF tiles to TFIID , a key factor in the transcription of nearly all genes ( Warfield et al . , 2017 ) .","Yeast TFIID contains three known AD-binding subunits , Taf4 , Taf5 , and Taf12 , which form a subcomplex with Taf6 and Taf9 ( Layer et al . , 2010; Wright et al . , 2006 ) .","We isolated this TFIID subcomplex and confirmed its interaction with the VP16 AD ( Figure 5—figure supplement 1H ) .","Pull-downs of the TF tiling library with the TFIID subcomplex enriched fragments up to 6 . 7-fold above random sequence controls ( Figure 5—figure supplement 1I–J , Figure 5—source data 1 ) .","In total , 27 ADs ( 18% ) across 26 TFs bound the TFIID subcomplex ( p<0 . 001 ) , greatly expanding its repertoire of known interactions ( Figure 5B–C ) .","However , all ADs that bound TFIID also bound Med15 , and despite the identical pull-down conditions , ADs bound more weakly to TFIID than to Med15 ( Figure 5—figure supplement 1K ) .","While more TFIID interactions may be discoverable if assay sensitivity can be further optimized , these results suggest that TFIID has lower affinity for most ADs and provide no evidence for its specific targeting as a coactivator .","Since TFIID binding by ADs is redundant and less frequent , its role in most TF-directed activation is secondary to that of Mediator .","What structural features of Med15 enable its promiscuous yet functional interaction with diverse AD sequences ?","In the best studied example , the Gcn4 AD interacts with hydrophobic patches and basic residues on multiple Med15 activator-binding domains ( ABDs ) in a large number of binding poses to form a ‘fuzzy’ complex ( Brzovic et al . , 2011; Tuttle et al . , 2018 ) .","We addressed the question by using FlexPepDock , a peptide docking algorithm from the Rosetta suite ( Leaver-Fay et al . , 2011; Raveh et al . , 2011 ) , to systematically build structural models of ABD-AD interactions , made computationally tractable by our identification of short cADs .","Our modeling focused on two ABDs with known structures: the KIX domain , which has homology to human Med15 and the p300/CBP coactivator family , and ABD1 ( Figure 6A; Brzovic et al . , 2011; Thakur et al . , 2008 ) .","To sample diverse sequences , aa composition , and secondary structure , we modeled interactions of the 28 13-aa cADs described above ( Figure 4A–D ) with the KIX domain and ABD1 .","For each interaction , 50 , 000 candidate structural models were sampled and ranked by the Rosetta energy score , and the 10 best-scoring models from each interaction were used in subsequent analyses ( Figure 6—source data 1 and Figure 6—source data 2 ) .","We validated our structural modeling by comparison to experimental data in two ways .","First , KIX domain residues previously shown by NMR to be important for interacting with the Pdr1 AD were recapitulated by the best-scoring model of this interaction ( Figure 6—figure supplement 1A; Thakur et al . , 2008 ) .","Second , the degree of alpha helix formation across the 28 cADs was consistent with the effect of proline insertion on in vivo activation ( Figure 6—figure supplement 1B ) .","As expected , cADs that were inhibited by proline predominantly bound both domains as an alpha helix , and conversely cADs that bound both domains in disordered conformations were minimally affected by proline insertion .","Some cADs that were minimally affected by proline insertion were modeled as alpha helical , possibly because disordered regions are underrepresented in protein structure databases on which Rosetta was trained .","A unifying feature of the interactions was the prominent role of hydrophobic contacts at the protein interface .","Hydrophobicity-driven interactions between proteins often employ high shape complementarity to maximize interface area and minimize solvent-exposed area .","However , the flat surface of the KIX domain was ineffective in burying hydrophobic cAD residues , which on average had as much surface area exposed to solvent as area contacting the KIX domain ( Figure 6B , top ) .","For example , the best-scoring Pdr1 AD model used Ile and Leu residues along one helical face to contact the KIX domain , but this left Trp and Tyr residues on the opposite face exposed to solvent ( Figure 6—figure supplement 1C ) .","The poor hydrophobic shape complementarity of the KIX domain was frequently mitigated by multiple salt bridges and cation-pi interactions ( Figure 6—figure supplement 1D ) .","ABD1 , which formed fewer ionic interactions , engaged cADs through a larger surface formed by its contoured hydrophobic ridge ( Figure 6B , bottom ) .","Nevertheless , hydrophobic residues of most cADs were incompletely buried , with those residues showing greater than 20% of area exposed to solvent in over half of all structures .","Thus , despite the central role of hydrophobic AD residues in activation and Med15 binding , hydrophobic shape complementarity is not an important feature of ABD-cAD interactions .","Consistent with weak constraints on the shape of the binding interface , all cADs bound in diverse poses .","For example , the 10 best-scoring models of the War1 cAD bound the KIX domain with its helix axis at different orientations and with different helical faces presented toward the interaction surface ( Figure 6C ) .","To quantify this diversity , we defined the unique binding poses of each ABD-cAD interaction by clustering similar structures based on cAD backbone root mean square distance ( 2 Å distance cutoff ) , and then counted the number of poses adopted by the 10 best-scoring models .","The 10 best-scoring models of nearly every cAD , when interacting with either domain , adopted six or more distinct poses ( Figure 6D ) .","To ensure that this did not result from insufficient sampling , we generated tenfold more candidate models ( 500 , 000 total ) for six cADs binding the KIX domain .","For all six cADs , the 10 best-scoring models still adopted seven or more distinct poses , despite spanning a narrower range of Rosetta scores ( Figure 6—figure supplement 1E–F ) .","Disordered cADs , lacking helical content , adopted especially many distinct poses; also , poses were frequently sampled only once , suggesting that disordered cADs inhabit an extremely large conformational space that may reduce the entropic cost of binding ( Figure 6—figure supplement 1G–H ) .","Altogether , these structures point to a shape-agnostic nature of Med15 ABD interaction surfaces and support the idea of a multi-pose , fuzzy mode of interaction .","To explore the consequence of fuzzy binding for amino acid sequence constraints , we took Leu , Phe , and Asp residues from models of all cADs and plotted their binding positions on the KIX and ABD1 interfaces ( Figure 6E ) .","As expected , the hydrophobic Leu and Phe residues occupied regions distinct from the negatively charged Asp residues .","However , the Leu and Phe distributions were similar to each other: there was no binding pocket that selectively preferred one residue over the other , despite large differences in their size and shape .","The diversity of AD sequences may thus be understood in terms of the fuzzy nature of the AD-ABD interaction , which only loosely constrains the characteristics of hydrophobic AD residues .","Pull-down experiments with individual ABDs showed no appreciable binding to any of nine ADs ( Figure 7—figure supplement 1A ) , consistent with the previous suggestion that multiple ABDs are required for strong binding to Med15 ( Brzovic et al . , 2011; Tuttle et al . , 2018 ) .","Multiple interaction sites were also common within ADs: PADDLE identified 42 ADs ( 28% ) that contained two or more non-overlapping cADs ( Figure 7—figure supplement 1B , Figure 3—source data 1 , Materials and methods ) .","We took 47 pairs of adjacent cADs and measured activation by the two cADs individually or in tandem .","For 40 of the pairs , the tandem cADs activated more strongly than expected from the combined activation of the individual cADs ( 4 . 0-fold median increase , Figure 7A ) .","Thus , an increased valence of interaction sites in both Med15 and ADs drives stronger binding and activation .","At larger scales , the number of interaction sites is further multiplied because some TFs contain multiple ADs , many TFs bind DNA as dimers , and many genes have several TF binding sites ( Hahn and Young , 2011; Spitz and Furlong , 2012 ) .","To understand the advantages conferred by the high valence of TF-Mediator interactions , we studied the kinetics in vitro of the initial step of activation: TF-driven recruitment of Mediator to DNA .","We performed these experiments with Gcn4 , which forms homodimers and has an extended AD consisting of three cADs .","We isolated budding yeast Mediator complex , coupled DNA containing a Gcn4 motif to a NeutrAvidin-coated surface , added Gcn4 , and measured real-time binding of Mediator by surface plasmon resonance ( Figure 7B , Figure 7—figure supplement 1C ) .","Mediator bound to the Gcn4-DNA complex with an apparent dissociation constant of KD = 14 nM ± 2 nM and an interaction half-life of 2 . 7 ± 0 . 2 min .","We repeated the experiment with DNA containing 2 , 4 , or 6 copies of the Gcn4 motif .","Mediator bound the resulting Gcn4-DNA complexes with identical on-rates .","Mediator was also recruited more efficiently , with more Mediator bound per mole of Gcn4 than on templates with only one motif ( Figure 7—figure supplement 1D ) .","The interaction half-life , however , increased with additional copies of the Gcn4 motif—up to 16 min when six motifs were present ( Figure 7B–C ) .","If all copies bound Mediator independently , then the half-life should be the same , regardless of the number of copies .","The presence of additional , neighboring ADs retards dissociation by enabling recapture of Mediator released from one AD through binding to another .","A slower dissociation rate corresponds to an increase in apparent affinity constant .","Conversely , we found that the multiplicity of ABDs in Mediator facilitated binding of Gcn4 ADs .","When the surface plasmon resonance experiment was repeated with large excess Gcn4 in solution , there was almost no change in the measured on-rate .","Even 160-fold excess of Gcn4 had only minimal effect on the rate of Mediator binding and final level achieved ( Figure 7—figure supplement 1E ) .","To maintain stable TF-DNA complexes without TFs in solution , we repeated these experiments using a fusion of the Gcn4 AD to nuclease deficient EcoRI ( E111Q ) , which resides on DNA for many hours and also forms homodimers ( Wright et al . , 1989 ) .","Still , Mediator with no Gcn4 competitor or with 250-fold excess of Gcn4 competitor bound similarly ( Figure 7D ) .","Excess competitor Gcn4-DNA complexes also failed to slow Mediator binding to the surface .","Evidently , binding of Gcn4 to Mediator does not appreciably block binding of other Gcn4 molecules .","This behavior may be explained by the occurrence of multiple Gcn4-binding sites on Mediator and weak interaction of a cAD with any one site .","Rapid association-dissociation equilibrium of a cAD allows a second Gcn4 molecule to interact with Gcn4-bound Mediator , increasing the frequency of exchange of one Gcn4 molecule by another ( Figure 7E ) .","Such facilitated exchange provides a rationale for both the fuzzy nature of the AD-ABD interaction and the multiplicity of ABDs .","PADDLE predictions and previous experiments ( Liu and Myers , 2015 ) identified two strong ADs in the Med2 subunit of Mediator , which is adjacent to Med15 in the Mediator complex , suggesting a broader role for ADs beyond TF proteins ( Figure 7—figure supplement 1F ) .","We used PADDLE to search for ADs across all nuclear-localized non-TF proteins in yeast and selected 1485 fragments that showed potential for activation ( Materials and methods ) .","Upon pooled in vivo testing , we found 290 fragments across 229 proteins that activated significantly ( p<0 . 001 by Z-test , Figure 1—source data 1 ) .","To see how often PADDLE fails to identify acidic ADs , we also included 291 control fragments that had extremely high acidic and hydrophobic content but were predicted not to activate .","Zero control fragments activated ( Figure 7—figure supplement 1G ) , showing that PADDLE likely identified all acidic ADs that exist and again demonstrating that AD function cannot be predicted from amino acid composition alone .","ADs were remarkably widespread and were found in proteins associated with diverse functions including transcriptional regulation , RNA splicing , DNA repair , ribosome biogenesis , and protein degradation .","We focused on the ADs in transcription-associated proteins .","Fourteen ADs were found in proteins that bind DNA and regulate pathway-specific genes; these proteins were not included in the original tiling library because they bind DNA only through a TF partner .","The two predicted Med2 ADs were also confirmed to activate , 36-fold and 25-fold in our assay .","Notably , ADs were identified in all major coactivator and chromatin modifying complexes , including Mediator , TFIID , cohesin , condensin , SAGA , RSC , SWI/SNF , ISWI , NuA4 , and FACT ( Figure 7F ) .","These complexes contained 2 . 9-fold more activating fragments than expected by chance ( p<10−12 compared to randomly shuffled controls ) , suggesting a functional role .","Domains that activate on their own could be inactive in their natural context if buried within a protein or protein complex .","To investigate this possibility , we cross-referenced coactivator ADs with all available structures in the Protein Data Bank ( PDB ) and with the Database of Disordered Protein Predictions ( D2P2 ) ( Berman et al . , 2000; Oates et al . , 2013 ) .","In fact , 23 ADs were predicted in D2P2 to be highly disordered ( more than 30% ) or were predominantly disordered or unresolved in PDB structures ( Figure 7F ) , and so may be exposed and active in vivo .","For example , cohesin , which binds Mediator and forms enhancer-promoter contacts through loop extrusion ( Davidson et al . , 2019; Kagey et al . , 2010; Kim et al . , 2019; Sanborn et al . , 2015 ) , contained strong and likely-disordered ADs in its subunits Scc1 ( RAD21 homolog ) , Scc3 ( STAG1/2 homolog ) , and Smc3 , as well as multiple ADs on the meiotic-specific subunit Rec8 .","These cohesin ADs , as well as ADs within other coactivators , may play a role in gene regulation by driving interactions with Mediator ."],["We obtained high-throughput measurements of activation strength that were both quantitative and precise .","This enabled us to detect—and predict with PADDLE—how activation strength depends on amino acid sequence , amino acid composition , and valence .","Quantitation was achieved by inducing artificial TF ( aTF ) binding for a brief , defined period so that GFP levels were representative of transcriptional activation ( McIsaac et al . , 2013 ) , and by partitioning the range of GFP signal into eight levels for sorting .","Others have measured activation only as an on-off binary system or screened for non-quantitative measures of activation , such as cell survival or pull-down of a cell surface marker ( Arnold et al . , 2018; Ravarani et al . , 2018; Tycko et al . , 2020 ) .","We achieved high precision by restricting measurements to cells with a defined level of aTF expression .","Controlling for this frequently overlooked confounding variable was critical because activation was strongly dependent on aTF abundance , especially at low levels of expression ( Figure 1—figure supplement 1E ) .","Another approach is to account for TF expression by sorting based on the GFP reporter signal divided by aTF expression ( Staller et al . , 2018; Staller et al . , 2021 ) ; this introduces substantial variability , because activation has a non-linear relationship with aTF expression ( Figure 1—figure supplement 1E ) , and because this ratio can be systematically biased by features that affect aTF expression ( Figure 1—figure supplement 1F ) .","This may be why it has been concluded that acidic residues are not necessary for Gcn4 activation , whereas using a similar reporter system we found definitive evidence to the contrary .","By this approach , we found that all ADs in S . cerevisiae employ a shared acidic and hydrophobic basis and trained a neural network termed PADDLE to predict sequences that activate on this basis .","PADDLE predicted ADs in human TFs with 92% accuracy , indicative of broad conservation of the acidic activation mechanism , and identified hundreds of new ADs in human TFs and virus proteins .","PADDLE could further be used to interpret the functional impact of cancer-associated and naturally occurring mutations in human ADs .","Predicted ADs from multiple species and instructions for running PADDLE are available online at paddle . stanford . edu .","During our development of PADDLE , a predictor of yeast activation named 'ADpred'’ was published ( Erijman et al . , 2020b ) .","ADpred , a shallow neural network trained on random protein sequences , achieved a prediction accuracy of 93% on random sequences .","However , its accuracy on wild-type sequences was much lower: only 33% of ADs predicted by ADpred activated in our experiments , and 29% of the ADs we identified were not predicted by ADpred , only 3 . 1-fold and 2 . 3-fold better than random predictions , respectively ( alternatively , tile-wise AUPRC of 0 . 294; Figure 3—figure supplement 1J–K , Materials and methods ) .","This discrepancy may be because the random sequences on which ADpred was trained do not sufficiently represent the vastly larger space of actual protein sequences; for example , random sequences rarely form significant secondary structure .","This failure to generalize highlights the importance of matching neural network training data to the prediction task and underscores the advantage of experiments on wild-type and related mutant sequences .","PADDLE predictions were also crucial for generating , refining , and testing hypotheses for activation mechanisms .","For example , we unexpectedly discovered that hydrophobic residues could inhibit activation by noticing that sub-domains of certain non-activating regions were predicted to activate on their own .","More generally , predictions on AD subsequences identified the core domains responsible for activation at single-amino acid resolution , focusing subsequent experiments on these domains .","By combining a high-throughput assay to measure and a machine learning algorithm to predict activity of protein sequences , we have demonstrated a design-build-test-learn cycle that could be applied to accelerate discovery of protein function in other areas of research .","This approach yielded the most detailed view to date of the principles governing acidic AD sequence , which we summarize as follows .","Activation arises from an abundance of acidic and bulky hydrophobic residues , especially Asp , Trp , Phe , and Leu .","The most potent activators cluster these residues densely , forming short core ADs .","In most ADs , and especially with high hydrophobic content or abundant Phe and Trp residues , this clustering is sufficient to activate in a composition-driven manner , regardless of sequence or secondary structure .","ADs with lower hydrophobic content , particularly those with many Leu residues , instead must position their hydrophobic residues along one face of an alpha helix to activate .","Overall , the loose constraints on sequence explain the remarkable diversity and lack of evolutionary conservation of ADs .","The unusual plasticity in AD sequence has several advantages .","It facilitates spontaneous evolution of new ADs , since an appreciable proportion of even random sequences can activate .","This flexibility would allow organisms to develop new transcriptional circuits responsive to their unique needs .","Furthermore , in contrast to classical interactions in which structure and function are dependent on a few key residues , activation strength can be adjusted gradually by mutation .","Indeed , the observation that nearly all core ADs could increase their strength with simple mutations demonstrates that proper regulation requires precisely adjusted rather than maximized activation .","Similarly , AD sequences can preserve their activity while evolving other features , such as in the Gal4 C-terminal AD , whose sequence is specifically bound and inhibited by Gal80 in the absence of galactose ( Johnston et al . , 1987; Ma and Ptashne , 1987 ) .","With the use of mRNA display , we found that 73% of ADs bound the Med15 subunit of Mediator and that binding strength was strongly predictive of activation for thousands of wild-type and mutant protein sequences .","In contrast , a five-protein TFIID subcomplex bound only 18% of ADs , all of which also bound Mediator .","We conclude that Mediator recruitment by TFs is a key driver for gene activation , and Mediator recruitment is largely determined by affinity of ADs for the Med15 subunit .","These findings are in keeping with reports that Mediator recruitment occurs independently of other coactivators and can stimulate the recruitment of other coactivators ( Ansari and Morse , 2013; Ansari et al . , 2014; Govind et al . , 2005 ) .","How is Med15 , through its four activator-binding domains ( ABDs ) , able to interact with such a large diversity of AD sequences ?","Despite the importance of hydrophobic interactions , our structural modeling showed that neither the Med15 KIX domain nor ABD1 provided enough shape complementarity to bury the hydrophobic residues of cADs .","Consistent with this , most cADs did not need to form an alpha helix to activate .","The lack of shape constraint has two important consequences .","First , all modeled cADs bound each ABD in many distinct , equally favored poses .","We therefore suggest that fuzzy binding to Med15 , previously demonstrated for the Gcn4 AD , is employed by all ADs .","Second , neither ABD showed favored binding positions of Leu versus Phe residues of the cAD , suggesting that the shapes of those residues were unimportant .","Thus , the loose constraints on binding shape explain the loose constraints on AD sequence .","Instead , each ABD can be approximated as a hydrophobic surface , which engages hydrophobic AD residues through simple hydrophobic forces , flanked by basic residues , which form salt bridges and cation-pi interactions with acidic and aromatic AD residues .","Clusters of hydrophobic residues adjacent to ADs inhibit activation , presumably because they compete for interaction with hydrophobic AD residues .","Many experiments have shown that transcriptional initiation involves a cycle in which ( 1 ) TFs recruit Mediator to enhancers or upstream activating sites , ( 2 ) Mediator contacts the promoter and orchestrates a series of events starting with chromatin rearrangement , followed by pre-initiation complex formation and finally RNA polymerase II recruitment , and ( 3 ) Mediator facilitates phosphorylation of the polymerase C-terminal domain , which releases polymerase into transcriptional elongation and triggers dissociation of Mediator ( Jeronimo and Robert , 2014; Knoll et al . , 2018; Robinson et al . , 2016; Whyte et al . , 2013; Wong et al . , 2014 ) .","Mediator release is apparently necessary because stably recruiting Mediator by fusing individual subunits to a DNA-binding domain failed to activate in nearly all subunits tested ( Wang et al . , 2010 ) .","The cycling of Mediator complexes increases the responsiveness of transcriptional regulation in at least two ways .","First , it requires certain steps to be repeated in order to continue initiating transcription , preventing the system from locking into an activated state .","Second , it frees Mediator complexes to participate in regulation of other genes , which all depend upon and compete for a relatively limited supply of Mediator ( Flanagan et al . , 1991; Gill and Ptashne , 1988 ) .","Thus , TFs must recruit Mediator in a manner that is specific and high-affinity but still dynamic .","Our kinetic experiments showed that Mediator binds Gcn4-DNA complexes with high affinity but that excess Gcn4 does not impede this interaction , showing that interacting Mediator and Gcn4 molecules can exchange rapidly .","If a bimolecular interaction occurs through a single site , a competing molecule cannot bind until the first molecule dissociates which , in a case of high affinity interaction , is a slow process .","Our observations are indicative of facilitated exchange , arising from the multiple sites and fuzzy nature of the Gcn4-Mediator interaction ( Figure 7E ) .","Fuzzy interactions allow for rapid association and dissociation , because a high proportion of random encounters lead to weak but productive binding ( Ferreira et al . , 2005; Sugase et al . , 2007; Tompa and Fuxreiter , 2008 ) .","One among multiple binding sites will frequently be available for a second molecule , which can invade and facilitate release of the first .","Similarly , accelerated dissociation or exchange of the E . coli sequence-flexible DNA-binding protein Fis in the presence of competitor proteins was proposed to occur by transitioning through a destabilizing Fis-DNA-Fis ternary complex ( Graham et al . , 2011; Kamar et al . , 2017 ) .","These advantageous kinetics provide a rationale for the prevalence of multivalent fuzzy interactions among transcription proteins .","ADs have primarily been characterized in TFs where they serve to recruit coactivators such as Mediator , TFIID , SAGA , SWI/SNF , and NuA4 ( Hahn and Young , 2011; Mitchell and Tjian , 1989 ) .","Our finding that functional ADs are also present in all coactivator complexes suggests that ADs have broader roles .","First , coactivator ADs could interact with ABDs within the same complex , limiting their weak or non-specific binding to TFs .","These auto-inhibited coactivators could still bind to ADs clustered on DNA through facilitated exchange .","For example , we found that two cADs activated more efficiently and two DNA-bound Gcn4 dimers recruited Mediator more efficiently in tandem than individually .","Second , coactivator ADs could amplify activation by mediating interactions between coactivators ( Ansari and Morse , 2013; Liu and Myers , 2015 ) .","For example , upon Mediator binding to TFs , the two ADs on Med2 would be liberated from auto-inhibition and could recruit other coactivators .","Alternatively , Mediator and other coactivators could cluster in the nucleoplasm through AD-ABD cross interactions and bind together to promoters .","Either mode of interaction would explain why recruitment of the Med2/Med3/Med15 subcomplex suffices to recruit SWI/SNF to the CHA1 promoter and suffices for activation of the ARG1 gene , while deletion of both Med2 ADs impairs transcription of galactose-inducible genes ( Ansari et al . , 2014; Liu and Myers , 2015; Zhang et al . , 2004 ) .","Phase separation of Mediator and TFs has recently been demonstrated in vitro and is proposed to underlie enhancer-promoter clustering and transcriptional activation in vivo ( Boija et al . , 2018; Chong et al . , 2018; Sabari et al . , 2018; Shrinivas et al . , 2019 ) .","Phase separation may occur when a protein makes dynamic self-interactions through two or more binding domains , ( Banani et al . , 2016 ) .","Interactions between the four Med15 ABDs and the two Med2 ADs make possible phase separation of the Mediator tail subcomplex; other coactivators , with both ADs and ABDs , might phase separate as well .","The functional role of such phase separation , if it occurs , is unclear .","A parsimonious interpretation suggests that phase separation of coactivators and TFs is simply a macroscopic consequence of the high valence and dynamism of coactivator-TF interactions , features which may serve an altogether different purpose .","ADs have long been enigmatic , due to the apparent mismatch of their structure and function: AD sequences are abundant among random polypeptides , and yet ADs bind their targets with high specificity; AD peptides are apparently disordered , and yet they bind with high affinity .","We now understand that these structural features of ADs are ideally suited to their functions—they render AD-target interaction dynamic , through rapid yet weak binding to single sites and strong binding yet rapid displacement from multiple sites .","Dynamic AD-target interaction may serve various purposes , such as a rapid response to changing conditions , and the recruitment of multiple AD-bearing proteins to a single target .","For example , Mediator bound to RNA polymerase II at a promoter might interact transiently with TFIID , SAGA , SWI/SNF complex and other proteins during transcription .","This mechanism may be employed with other unstructured sequences and high valence targets in other cellular processes ."],["Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact , Roger Kornberg ( kornberg@stanford . edu ) .","Raw sequencing data for in vivo activation and in vitro binding assays have been deposited in the Gene Expression Omnibus ( GEO ) database , accession number GSE173156 .","The S . cerevisiae strain BY4711 ( ATCC 200873 ) with genotype MATalpha trp1delta63 was used for activation experiments .","Standard cell growth conditions were 30°C incubation in YPD medium ( 1% w/v yeast extract , 2% w/v peptone , 2% w/v dextrose ) or synthetic complete ( SC ) medium ( Dunham , 2015 ) .","Growth in SC medium lacking tryptophan ( SC-Trp ) was used to select for transformants with the pRS414 plasmid backbone .","Selection for stable transformants harboring the natMX6 marker gene was performed by supplementing media with 100 µg/mL nourseothricin ( NAT ) .","Selection for stable transformants harboring the KanMX marker gene was performed by supplementing media with 200 µg/mL geneticin ( G418 ) .","In activation assays , overnight cultures were diluted to OD600 = 0 . 4 in SC-Trp-NAT-G418 with 10 nM beta-estradiol and grown for 4 hr at 30°C before fluorescence-activated cell sorting ( FACS ) .","HEK293T cells ( ATCC CRL-3216 ) were cultured in DMEM with high glucose and GlutaMAX ( Thermo 10566016 ) supplemented with 10% fetal bovine serum ( Millipore TMS-013-B ) at 37°C in 5% CO2 .","Yeast and human cells were used directly from ATCC , which authenticates all cell lines using STR profiling and tests for mycoplasma contamination .","The artificial TF ( aTF ) expression cassette , which contains a KanMX gene and an ACT1 promoter driving expression of an aTF ( comprising an mCherry tag , mouse Zif269 DBD , and estrogen binding domain ) with a multiple cloning site for insertion of protein fragments of interest , was derived from pMVS142_pACT1_mCherry_Zif268_EBD_MCS_KAN , which was a gift from Barak Cohen ( Addgene plasmid # 99049 ) ( Staller et al . , 2018 ) .","This cassette was cloned into the pRS414 backbone , which contains a TRP1 marker ( Sikorski and Hieter , 1989 ) .","pRS414 plasmid was digested with EcoRI-HF and SacI , the aTF cassette was PCR-amplified using Phusion polymerase ( NEB ) , and DNA fragments were assembled through Gibson assembly and transformed into XL10-Gold Ultracompetent Cells ( Agilent ) .","The final construct , called pRS414-TFchassis , was confirmed by Sanger sequencing .","Oligo pools encoding all the protein fragments to be tested in the screen were synthesized at 12K scale ( Twist Bioscience ) .","Each oligo was 200 nt in length and contained a 20-nt 5’ constant region and a 21-nt 3’ constant region which were used for PCR amplification .","Different sub-libraries had different 3’ constant regions , which allowed specific amplification of the sub-library from the oligo pool .","To enable cloning of the oligo pool into pRS414-TFchassis , 20 bp Gibson homology arms matching pRS414-TFchassis were added using PCR .","To amplify each sub-library , 2 . 5 ng of oligo template was PCR-amplified for nine cycles using Phusion polymerase , followed by column cleanup .","pRS414-TFchassis was digested with NheI-HF and AscI , the amplified oligo pool was cloned in by Gibson assembly , and the product was transformed in two replicates , each into 100 µL of Agilent XL10-Gold cells with 5 µL of Gibson mix .","Transformed cells were plated on lysogeny broth ( LB ) plates with Carbenicillin ( Carb ) to estimate efficiency and also grown in 50 mL LB Carb overnight at 37°C for next-day plasmid extraction using Qiagen Midiprep kit .","To generate positive control plasmids for the activation screen , IDT gBlocks containing genes expressing 53-aa regions from VP16 , Gcn4 , and Pho4 ADs were also cloned in the same manner as above .","The uncloned pRS414-TFchassis , which lacks an introduced insert , was used as a negative control .","The GFP reporter construct pMVS102_P3-GFP-NATMX6 was a gift from Barak Cohen ( Addgene plasmid # 99048 ) ( Staller et al . , 2018 ) .","To construct a yeast GFP reporter cell line , a cassette from pMVS102_P3-GFP-NATMX6 comprising a natMX6 marker and a P3 promoter ( a modified GAL1 promoter with six copies of the Zif268 binding site ) that drives expression of a fast maturing EGFP variant ( McIsaac et al . , 2013 ) , was PCR-amplified using Phusion polymerase .","PCR primers were designed to add 40 bp overhangs with homology to the dubious ORF YBR032W .","The PCR product was transformed into yeast using a high-efficiency method ( Gietz et al . , 1992 ) .","After overnight outgrowth at 30°C , transformed cells were plated onto YPD-NAT , colonies were picked , and integration into YBR032W was confirmed by PCR amplification of the integrant and Sanger sequencing .","The in vivo activation screen requires a yeast library in which each cell contains one pRS414-TFchassis plasmid with a unique insert .","The plasmid is maintained at single-copy or low levels due to its CEN/ARS origin of replication , so if a cell receives multiple different vectors upon transformation then repeated cell divisions will eventually separate the distinct vectors ( Scanlon et al . , 2009 ) .","To ensure generation of a cellular library in which a unique pRS414-TFchassis plasmid was present in each cell , an assay was developed to quantify the fraction of cells initially receiving multiple plasmids and the time required for cells to reach single-copy plasmid levels by repeated cell divisions .","First , using a high efficiency transformation protocol ( Gietz et al . , 1992 ) , approximately 50 million yeast cells in a volume of 360 µL were transformed with 2 µg total plasmid DNA , consisting of an equal mixture of aTF plasmids expressing one AD from VP16 , Gcn4 , or Pho4 .","Transformed cells were sparsely plated on SC-Trp-NAT plates either immediately or after a 24 hr or 48 hr outgrowth in SC-Trp-NAT liquid media .","For each plating condition , colonies were picked for DNA extraction , followed by qPCR to measure the abundance of VP16 , Gcn4 , and Pho4 plasmids .","After 0 hr , 24 hr , and 48 hr of growth , 2 ( of 15 ) , 2 ( of 15 ) , and 1 ( of 24 ) colonies had multiple vectors present , respectively .","Because this method cannot detect multiple transformation events of the same plasmid sequence in the same cell ( e . g . double VP16 transformation ) , a correction was applied , yielding multiple vector rates of 20% , 20% , and 6% for each outgrowth condition , respectively .","The cloned pRS414-TFchassis plasmid library was transformed into the GFP reporter yeast strain using the same method as in the multiple transformant outgrowth assay above but scaled up 2 . 5-fold .","To generate clonal control strains , cells were also separately transformed with a positive control plasmid expressing a VP16 , Gcn4 , or Pho4 AD or with a negative control plasmid , the pRS414-TFchassis plasmid without a protein fragment insert .","Transformed cells were grown in 200 mL SC-Trp-NAT-G418 for 6 days ( first library ) or 5 days ( second library ) ; cells were diluted down to OD600 = 0 . 05 every day upon reaching saturation .","After this period , cells were diluted to OD600 = 0 . 4 in SC-Trp-NAT-G418 containing 10 nM beta-estradiol and grown at 30°C for 4 hr .","Cells were then pelleted and washed once in FACS buffer ( PBS , 2 mM EDTA , 0 . 1% BSA; sterile filtered ) , and finally suspended in FACS buffer at roughly 40 million cells/mL .","Cells were analyzed and sorted on a BD Influx Cell Sorter at the Stanford Shared FACS Facility .","Cells from different sub-libraries were pooled just prior to sorting .","The 10–15% largest cells were excluded from sorting , and then an additional filter was applied to limit mCherry signal to within a twofold range .","For the remaining cells , the full range of GFP signal was divided into eight evenly spaced bins in log scale; cells were first sorted into the four low bins and then sorted into the four high bins .","In total , 2 . 85 million cells and 2 . 65 million cells were sorted for the first and second libraries , respectively .","To define the background GFP signal , cells containing the negative control plasmid were also induced with 10 nM beta-estradiol and their distribution across the eight GFP bins defined above was measured .","We found that these cells had slightly increased levels of GFP signal relative to untransformed cells , potentially because binding of the aTF alone slightly opens up chromatin around the GFP promoter .","Cells expressing just one plasmid containing the VP16 , Gcn4 , or Pho4 AD were also individually analyzed ( by similarly measuring their distribution across same eight GFP bins ) and compared to results obtained for identical sequences in the pooled activation screen .","Cells sorted into each of the 8 GFP bins and a sample of total input cells ( nine samples total ) were grown overnight and then pelleted .","Pellets were treated with lyticase and then plasmid DNA was extracted using a Qiagen Miniprep kit .","The region of the plasmid encoding the variable protein fragment was PCR-amplified for 11–20 cycles ( based on the OD600 of the culture ) using Phusion polymerase with PCR primers that added flanking sequences corresponding to the standard Illumina sequencing primers .","After a column cleanup , a second PCR was performed with primers that added sample-specific barcodes and the P5 and P7 sequences for Illumina sequencing .","This second PCR product was SPRI size-selected using AMPure XP beads ( Beckman Coulter ) at 1x dilution , and the different samples were pooled for paired-end sequencing on the Illumina NextSeq 550 .","Roughly 2 million reads of 2 × 76 bp length were sequenced per sample .","The pFN26A BIND plasmid ( Promega ) , which contains a Renilla luciferase gene used for normalization , was used to express domains of interest fused to a GAL4 DBD .","Genes expressing 50 randomly chosen putative human ADs predicted by PADDLE were synthesized in a single oPool ( IDT ) with Gibson homology arms , amplified using Phusion polymerase , and cloned into pFN26A BIND ( digested with AsiSI and PmeI ) using Gibson assembly .","The product was transformed into Agilent XL10 Gold cells , and then plated onto LB Carb plates .","Forty-three colonies were picked , which yielded clones of 25 correct unique ADs .","In addition , plasmids expressing a VP16 AD-positive control or three random protein sequences for negative controls were individually cloned in the same manner .","HEK293T cells were plated at 24 , 000 cells per well of a 96-well plate .","The next day , cells were transfected using Lipofectamine 3000 with 50 ng of cloned pFN26A BIND plasmid and 50 ng of pGL4 . 35 ( Promega ) , which contains a Firefly luciferase reporter gene under the control of a minimal promoter driven by nine upstream GAL4 DNA-binding sites .","Each plasmid was transfected into triplicate wells .","Seventy-two hours after transfection , luminescence from renilla and firefly luciferases were measured using the Dual-Glo Luciferase assay system ( Promega ) on a BioTek Synergy two plate reader .","Measured activity of each predicted AD or control sequence was computed by dividing the Firefly luciferase signal by the Renilla signal and then averaging the triplicate values .","Fold-activation was computed by dividing the activity of each predicted AD by the mean activity of the three negative controls , Z-scores were computed based on log-scale fold-activation by standardizing the negative controls to mean of 0 and variance of 1 , and p-values were computed with one tailed Z-tests .","An mRNA display library was made for each of sub-library A , sub-library B , and six pooled controls ( three libraries total ) , using the following method .","Sub-libraries of wild-type and mutant TF fragments were expressed as mRNA-tagged peptides using mRNA display with in vitro translation as described in Takahashi and Roberts , 2009 .","Specifically , sub-library pools were PCR-amplified using primers that added an upstream T7 promoter sequence and a downstream linker and HA-tag but no stop codon .","Thus , all encoded peptides were of the form MSGT[variable protein sequence]TSVGGSGSYPYDVPDYA and fused to mRNA at the C-terminus .","In addition to the two sub-libraries , similar DNA sequences encoding positive control ADs from VP16 , Gcn4 , and Pho4 and three negative control random protein sequences were synthesized as gBlocks ( IDT ) ( containing an upstream T7 promoter sequence and a downstream linker and HA-tag but no stop codon ) .","All subsequent work was performed in RNase-free conditions .","mRNA was in vitro transcribed for 4 hr at 37°C using the MEGAscript T7 transcription kit ( Thermo ) and purified with RNeasy mini columns ( Qiagen ) .","The PF30P oligo containing a poly-A sequence , spacers , and puromycin was synthesized by IDT and ligated to the 3’ end of the mRNA using Moore-sharp splinted ligation by T4 DNA ligase .","The ligation product was desalted by three rounds of concentrating in 0 . 5 mL Amicon Ultracel 30K MWCO columns and diluting in water .","The desalted product was run on a denaturing 4% polyacrylamide gel with 7 M urea .","The band corresponding to full-length ligated mRNA was excised from the gel and recovered using the Model 422 Electro-Eluter with 3500 MWCO caps ( Bio Rad ) .","After concentrating and desalting in the same manner , the ligated and purified mRNA was used in an in vitro translation reaction using the Retic Lysate IVT kit ( Thermo ) for 60 min at 30°C .","MgCl2 and KCl were added to final concentrations of 60 mM and 500 mM , respectively , and the sample was incubated for 15 min at room temperature to encourage puromycin fusion .","The sample was then diluted in binding buffer ( 20 mM Tris pH 8 . 0 , 100 mM KCl , 1 mM MgCl2 , 0 . 01% Tween-20 , 0 . 1 mg/mL BSA , 0 . 1 mM DTT ) and bound to Anti-HA magnetic beads ( Pierce ) for 30 min at room temperature with mixing .","After washing three times for 5 min with binding buffer without BSA , peptides were eluted by incubating in 1x SuperScript II buffer ( 50 mM Tris-HCl pH 8 . 3 , 75 mM KCl , 3 mM MgCl2; Thermo ) with 0 . 01% Tween-20 and 0 . 5 mg/mL synthetic HA peptide ( Thermo ) for 1 hr at room temperature with mixing .","The mRNA tag on the fused peptides was reverse transcribed using SuperScript II ( Thermo ) and a primer that contained ( 1 ) a sequence that hybridizes just downstream of variable protein sequence coding region , ( 2 ) a 12-nt random sequence that serves as a unique molecular identifier ( UMI ) , and ( 3 ) the Illumina read two sequencing primer sequence .","Enzyme was heat inactivated by adding EDTA to 10 mM and heating to 65°C for 10 min .","To remove aggregates , the library was fractionated in 20 mM Tris pH 8 . 0 , 100 mM KCl , 1 mM MgCl2 , 1 mM DTT over a 6 mL 10–30% sucrose gradient by spinning at 60 , 000 rpm in a SW60 rotor for 10 hr .","Fractions of 100–200 µL were collected and the abundance of peptide-mRNA-cDNA fusion in each fraction was assayed using qPCR .","The 3–4 peak fractions encompassing 80–90% of total input was pooled and used in pull-down experiments or flash frozen in aliquots and stored at −80°C .","Biotinylated bait proteins ( 12 µg of Med15 or 8 µg of TFIID subcomplex ) were bound to 50 µL Dynabeads MyOne Streptavidin T1 ( Thermo ) for 15 min at room temperature in binding buffer ( 20 mM Tris pH 8 . 0 , 100 mM KCl , 10 mM MgCl2 , 0 . 01% Tween-20 , 1 mM DTT ) and then washed three times for 5 min in binding buffer .","In the third wash , beads were blocked by also including 100 ng/µL salmon sperm DNA ( Thermo ) and 10 ng/µL BSA ( NEB ) in the buffer .","mRNA display library input was prepared by mixing sub-library A , sub-library B , and the three positive and three negative controls .","This input was then incubated with the bait-coated beads in binding buffer with salmon sperm DNA and BSA at 100 µL total volume for 8 min with mixing .","Afterward , the beads were separated on a magnet , briefly washed once with cold binding buffer , and then immediately separated on a magnet again .","cDNA from bound fragments was eluted by suspending the beads in 10 mM Tris pH 8 . 0 with proteinase K ( NEB ) and incubated for 30 min at room temperature .","Proteinase K was heat inactivated at 95°C for 10 min , beads were separated on a magnet , and the eluate was collected .","Total abundance of the library cDNA in the input , bound , and eluted samples was measured by qPCR .","To measure the total abundance of the whole library ( without regard to the identity of the library elements ) , qPCR primers were designed to hybridize to the constant sequences flanking the variable region of the library .","Additionally , this design enabled the use of different primers to separately track sub-library A and sub-library B , as well as each of the six control sequences individually .","The controls were used to provide an estimate of binding efficiency .","Binding percentage was calculated by dividing the abundance of each library in the bound sample by the abundance in the input sample .","cDNA of the input and bound samples were prepared for sequencing in the same manner as the purified plasmids in the activation assay .","Biotinylated 104 bp double-stranded DNA templates were assembled by annealing a 104-nt oligo to a 20-nt oligo with a 5’ biotin modification ( IDT ) and elongating using Klenow Fragment ( 3’-->5’ exo- ) ( NEB ) .","The oligos were designed such that the TF DNA binding sites were placed away from the biotinylated end .","SPR experiments were performed on the ProteOn XPR36 Surface Plasmon Resonance System using NeutrAvidin-coated NLC Sensor Chips ( Bio Rad ) .","Mass of biomolecules bound to the chip surface was measured in arbitrary units ( AU ) approximately every 1 s .","All plotted data were normalized to the first ligand channel and the first analyte channel which were used as negative controls in experiments as described below .","In SPR experiments with GCN4 , 104 bp DNA templates with 0 , 1 , 2 , 4 , or 6 Gcn4 motifs were bound to the chip along the ligand channels , except for the first ligand channel , which was left without DNA for normalization .","Binding buffer for all conditions was 20 mM Tris-HCl ( pH 7 . 5 ) , 100 mM KCl , 10 mM MgCl2 , 2 mM EDTA , 10 mM AmSO4 , 0 . 05% Tween-20 , with 1 mg/mL BSA ( NEB ) .","Bound DNA was washed twice with 1 M NaCl .","Gcn4 at 7 . 5 nM was then pre-bound to DNA for 480 s along the analyte channels , except for the first analyte channel , which was left without Gcn4 for normalization .","After Gcn4 binding , Gcn4 at 7 . 5 nM with Mediator at 0 , 1 . 25 , 2 . 5 , 5 , or 10 nM were flowed across analyte channels 2–6 respectively for 300 s , then dissociation was observed for 20 min by flowing in binding buffer .","Association , dissociation , and affinity constants of Mediator binding to Gcn4 on DNA were estimated by fitting to a Langmuir kinetic model using the ProteOn Manager Software ( Bio Rad ) .","Reported values are the mean and standard deviation across the four different Mediator concentrations .","Note that Mediator dissociation rates in conditions with multiple Gcn4 binding sites may be overestimated because it also includes the dissociation of Gcn4 from DNA .","For SPR experiments with EcoRI , biotinylated 60 bp DNA templates containing 0 or 1 EcoRI motif were annealed and elongated as described above , and then bound along the ligand channels and washed with 1 M NaCl .","Gcn4AD-EcoRI ( E111Q ) was bound along analyte channels at 27 nM monomer in 20 mM Tris-HCl ( pH 7 . 5 ) , 10 mM MgCl2 , 300 mM NaCl , 10 mM bME , 0 . 05% Tween-20 , 1 mg/mL BSA ( higher salt was used to reduce non-specific binding to DNA ) .","Minimal dissociation was observed after binding .","Then , Mediator at 5 nM with competitor Gcn4 protein ( in molar excess as specified in Figure 7D ) was flowed across the chip for 180 s in 20 mM Tris-HCl ( pH 7 . 5 ) , 10 mM MgCl2 , 100 mM NaCl , 10 mM AmSO4 , 10 mM bME , 0 . 05% Tween-20 , 1 mg/mL BSA .","All protein positions reported in the text are 1-indexed and inclusive .","The first library of protein sequences was composed of sub-library A , which contained wild-type tiles spanning all S . cerevisiae transcription factors , and sub-library B , which contains mutants of 8 known S . cerevisiae ADs .","The second library was composed of sub-library D and sub-library E , which contained multiple collections of wild-type , mutant , and designed sequences .","Sequences in sub-libraries A , B , and D were 53 amino acids in length and sequences in sub-library E were 30 amino acids in length .","Protein and DNA sequences , as well as activation and pull-down data where available , are listed in Tables S1 and S4 .","Libraries contained additional fragments which were included in experiments but are not reported here .","In the reverse translation design process , we aimed to optimize our library DNA fragments for compatibility and consistency with our in vitro and in vivo assays , as well as with standard RNA-seq protocols .","We also sought to build in redundancy for error-correcting reads .","In particular , we used the Python package dnachisel 1 . 4 . 1 ( Zulkower and Rosser , 2020 ) to optimize the following objectives: To allow paired-end sequencing to uniquely identify each element , we additionally enforced an edit distance of 6 among the first 48 bases and last 48 bases of any two sequences in the same sub-library .","This was performed in a randomized , brute-force , iterative approach , with each iteration consisting of the following steps: Finally , we verified that all sequences sharing the same sub-library primer ( e . g . all sequences in sub-library A , or all sequences in sub-library B ) had a paired-end edit distance ( sum of edit distance of 5’-most 50 bases and edit distance of 3’-most 50 bases ) of at least 6 .","The use of sequencing primers unique to each sub-library enabled us to submit samples for sequencing in multiplexed format and accurately assign reads to the correct sub-library computationally .","We further leveraged the edit distance margin built into the library to enable mapping of sequencing reads with a small number of errors .","Sequencing read alignment was performed using custom bash script built on top of existing tools and additional custom scripts .","It takes as input arguments the UMI length , the sub-library sequencing primer , the edit distance threshold for that sub-library , and raw FASTQ files .","For pull-down experiments , unique molecular identifiers ( UMI ) were extracted from reads and appended to the read names using umi_tools 1 . 0 . 0 ( Smith et al . , 2017 ) .","cutadapt 1 . 18 was used to discard reads without matching paired-end sub-library sequencing primers and trim the primers in reads with matching primers; the default error tolerance was used ( Martin , 2011 ) .","bwa-mem 0 . 7 . 17-r1188 was used to perform a first-pass alignment of reads to the DNA fragment library ( Li , 2013 ) .","Imperfectly mapped read pairs ( i . e . those without paired read SAM flags of 99 and 147 ) were re-mapped to the library sequence with minimal edit distance .","This was necessary because bwa-mem did not always correctly map paired reads as a pair , a problem most evident in the mutant library with many similar sequences .","Pairwise Levenshtein edit distance was computed using the Python package editdistance 0 . 5 . 3 ( https://github . com/roy-ht/editdistance , Tanaka , 2019 ) .","Paired reads exceeding the edit distance threshold were discarded using reformat . sh from BBTools 38 . 61 ( https://sourceforge . net/projects/bbmap/ ) .","Duplicate reads were identified and deduplicated using utmi_tools 1 . 0 . 0 .","Finally , reads mapped to each DNA library fragment sequence were counted .","Calculations of GFP signal and fold-activation required the following inputs: First , to make cell counts comparable between sorting bins 1–4 versus bins 5–8 , the number of cells sorted into each GFP bin were normalized based on the number of events assayed in each sort .","Second , to estimate the number of cells sorted into each of the 8 GFP bins that expressed a given protein fragment F , the total number of cells sorted into each bin was multiplied by the fraction of sequencing reads for that bin that mapped to F . Any fragment represented by fewer than 10 cells were excluded from further analysis .","Third , the average GFP signal of cells expressing F was calculated as a weighted average from the average GFP signal observed in each bin during sorting weighted by the distribution of the number of cells expressing F across the eight bins .","Because GFP distributions cells expressing single AD sequences were normally distributed in FACS data when GFP signal ( arbitrary units ) was plotted in log-scale , all averages of GFP distributions were calculated in log-scale ( equivalently , a geometric mean of GFP values ) .","The standard deviation of GFP signal of cells expressing F was similarly calculated in log-scale and used as a metric for measurement precision .","All averages involving activation reported in the paper are calculated in log-scale ( equivalently , a geometric mean ) .","Fourth , the fold-activation of F was calculated as the mean GFP signal of F divided by the mean GFP signal of negative control cells ( see below ) .","Z-scores were calculated by standardizing the GFP signal of negative control cells to mean 0 and variance one in log scale and p-values were calculated with one tailed Z-tests .","The mean and standard deviation GFP signal of negative control cells was defined in one of two ways .","In the experiment with sub-libraries A and B , a sample of estrogen-induced cells containing only the empty pRS414-TFchassis plasmid was assayed under identical conditions and the mean and standard deviation GFP signal was computed from FACS data .","In the experiment with sub-libraries D and E , these values were instead determined internally from the 50 random sequence negative controls in the pooled library .","Since a small fraction of random sequences activate , outlier sequences were iteratively excluded if they had GFP signal greater than three standard deviations away from the mean , resulting in three sequences excluded from each sub-library .","Then mean and standard deviation were computed from the sum of the negative control GFP distributions .","In activation assays with sub-libraries D and E , we noticed that many sequencing reads that were mapped with a small number of errors corresponded to fragments with a very large GFP standard deviation .","Upon closer examination of sequencing reads of example fragments , the same sequencing errors were consistently found , which likely corresponded to mutations that occurred in vivo that affected function of the protein fragment ( e . g . inactivation of a strong AD by a frameshift mutation ) .","We therefore used only reads that aligned perfectly for this experiment .","Furthermore , fragments with GFP standard deviation greater than 4 . 5 were excluded from further analysis .","The following three metrics were tracked: the number of total unique fragments , number of fragments observed in at least one of the eight GFP bins , and number of fragments that passed filtering .","For the activation assay with sub-libraries A and B , this was ( 7637 , 7604 , 7549 ) for sub-library A and ( 2900 , 2833 , 2813 ) for sub-library B . The same metrics for the activation assay with sub-libraries D and E was ( 7733 , 7716 , 7345 ) for sub-library D and ( 4261 , 4261 , 4252 ) for sub-library E . These numbers include additional fragments in the peptide libraries that are not reported in this paper .","To define ADs across all TFs , at each protein position a mean Z-score was calculated from the Z-scores of all fragments overlapping that position .","ADs were defined for all intervals of the protein that had a mean positional Z-score ( MPZ ) greater than 3 . 5 .","Starting with intervals that had the highest maximal MPZ score , the start and stop positions of the AD were defined by the full width half maximum; that is , the positions in the protein at which the MPZ fell below half of the maximal MPZ value in that interval .","If the AD defined in this manner includes a new region with an MPZ score higher than the original maximal value , this indicated that the MPZ peak was only small relative to nearby peaks , so the AD was then excluded .","We compared our ADs with the Induction Dynamics gene Expression Atlas ( IDEA ) ( Hackett et al . , 2020 ) , which induced expression of every TF individually and measured ‘v_inter’ parameters that describe the resulting fold-change in the expression of all target genes .","We said that a TF activated a gene if the v_inter > 1 , and for each TF that regulated at least five genes , we calculated the proportion of regulated genes that were activated .","We then compared these proportions for TFs that had a strong AD ( overlapping a fragment with Z-score at least 6 ) versus TFs without ADs .","Significance was calculated using a Kolmogorov–Smirnov test .","Hydrophobicity measurements were calculated using the Wimley-White interfacial scale , which measures whole-residue ∆G for transfer from water to a water-octanol interface ( Wimley and White , 1996 ) .","Because the goal was to quantify the total content of hydrophobic residues rather than the net hydrophobicity/hydrophilicity , hydrophobic content of a peptide is computed as the sum of the scores of only the hydrophobic residues ( i . e . those with negative ∆G ) .","Net charge was computed by adding the number of Arg and Lys residues and subtracting the number of Asp and Glu residues .","Motif finding using DREME was done through the web portal ( http://meme-suite . org/tools/dreme ) using default settings and shuffled input sequences as control ( Bailey , 2011 ) .","Searching for 9aaTAD sequences was done by matching to the ‘most stringent pattern’ [MDENQSTYG]{KRHCGP}[ILVFWM]{KRHCGP}{CGP}{KRHCGP}[ILVFWM][ILVFWMAY]{KRHC} listed at the prediction portal https://www . med . muni . cz/9aaTAD/index . php ( Piskacek et al . , 2007 ) .","ADpred was installed locally from https://github . com/FredHutch/adpred , Erijman , 2020a and predictions were computed for all 30-aa tiles on all yeast TFs ( Erijman et al . , 2020b ) .","ADs predicted by ADpred were defined using the authors’ criteria , namely sites with five or more contiguous residues with a score >= 0 . 8 .","An ADpred-predicted region was considered correct if it overlapped by at least five aa with a significantly activating tile ( p<0 . 0001 ) .","With this approach , ADpred achieved a precision and recall of 0 . 366 and 0 . 692 in predicting experimental ADs , respectively .","When ADpred-predicted regions were randomly shuffled 100 times to different proteins and positions , the average precision and recall of these shuffled predictions were 0 . 120 and 0 . 295 respectively .","Alternatively , to compare ADpred predictions directly to our experimental measurements of activation , a representative ADpred score for each 53-aa TF tile in our library was calculated by taking the 80th percentile of ADpred scores of 30-aa tiles contained within the 53-aa tile .","These scores were used for calculating tile-wise precision-recall curves for ADpred predictions ( Figure 3—figure supplement 1K ) .","The fractional pull-down of each fragment F was calculated as ( the fraction of the total input library that bound to beads , measured from qPCR ) x ( the fraction of UMI-deduplicated sequencing reads in the bound sample matching F ) / ( the fraction of UMI-deduplicated sequencing reads in the input sample matching F ) .","Fragments with fewer than 30 reads in the input sample or fewer than five reads in the bound sample were excluded from downstream analysis .","Baseline binding was defined using the 50 random protein control sequences in each library , except the sequences that were outliers in the activation assay were excluded .","Specifically , random control outliers were iteratively removed if their activation was more than three standard deviations away from the mean ( in log scale ) , resulting in the exclusion of random control #2 , 12 , 16 , 38 from sub-library A and #17 , 18 , 21 , 22 , 31 , 43 from sub-library B . Binding enrichment of each fragment was then computed as the fractional pull-down of the fragment divided by the mean fractional pull-down ( in log scale ) of the random controls .","Because fractional pull-down and enrichment were normally distributed in log-scale , all averages involving pull-down or enrichment reported in the paper are calculated in log-scale ( equivalently , a geometric mean ) .","Z-scores were computed by standardizing the binding enrichment of random controls to mean 0 and variance one in log-scale and P-values were calculated with one tailed Z-tests .","Triplicate measurements for Med15 pull-downs and duplicate measurements for TFIID subcomplex pull-downs were combined by averaging the binding enrichment values .","A small number of highly positively-charged fragments bound in Med15 pull-downs when the mRNA display library was purified on a sucrose gradient but did not bind when the sucrose gradient was omitted .","Suspecting that this binding was artifactual , potentially through non-specific binding to mRNA tags , we excluded from analysis 92 fragments that bound with Z-score between −2 . 5 and 1 without the sucrose gradient and had a Z-score at least 3 . 5 higher with the sucrose gradient than without .","Med15-binding domains were annotated in the same manner as ADs using a mean positional Z-score , but with a cutoff of 2 . 23 .","An AD was considered to bind Med15 or TFIID subcomplex if it overlapped a fragment that bound significantly ( p<0 . 001 ) by at least 26 aa .","A Med15-binding domain was considered to activate if it overlapped an AD .","Neural network models were trained to predict activation Z-scores .","Approximately 15% of sequences from sub-library A and B were randomly held-out as a test set , and the remaining library elements were split into 10 parts for training and cross-validation .","Because adjacent TF tiles have significant overlap in sequence , all tiles from each protein were placed into the same train or test split .","From sub-library B , all mutants and homologs of the Pdr1 AD and all scramble mutants were held-out in the test set and the remaining sequences were distributed evenly across the training splits .","Models were developed in TensorFlow 2 . 2 ( Abadi et al . , 2016 ) and trained using CPUs on Stanford’s Sherlock computing cluster .","All models used mean squared error as the loss function .","Sequence encodings and activation Z-scores were standardized to mean 0 variance one across the training dataset for more efficient learning .","For each architecture , 10 models were trained on each subset of nine training splits and validated on the 10th split .","The best model architecture was chosen by the mean validation score across the 10 splits on sub-library A sequences only .","For final evaluation on the test dataset , the mean prediction across all 10 models was used .","Protein fragment sequences were each encoded as a single 20-element vector giving the proportions of each of the 20 amino acids .","The best architecture , as determined by a grid search over hyperparameters , was a neural with three fully connected layers of width 40 with Swish activation ( Ramachandran et al . , 2017 ) , batch normalization , L2 weight penalty of 0 . 01 , and dropout rate of 0 . 4 .","The model was trained with an initial learning rate of 0 . 0001 using the Adam optimizer for up to 300 epochs with two scheduling callbacks: reduction of the learning rate by fivefold if training loss did not improve for 20 epochs , and early stopping if no improvement on the validation loss ( averaged in a 10-epoch window ) was observed for 75 epochs , upon which the best model from previous epochs was saved .","One-hot encodings were used to encode each 53-aa protein fragment sequence as a 53-by-20 matrix .","Secondary structure predictions were generated using PSIPRED v . 4 . 01 without PSI-BLAST ( Jones , 1999 ) , and a 53-by-2 matrix of the alpha helix and coil predictions were also used as input .","Disorder predictions were generated using IUPred2A ( Mészáros et al . , 2018 ) , and a 53-by-2 matrix of the disorder predictions in ‘short’ and ‘long’ mode were also used as input .","In total , each 53-aa protein fragment was encoded as a 53-by-24 matrix .","The best model architecture , as determined by a grid search over hyperparameters , was a convolutional neural network with nine convolutional layers with kernel size 10 and channel width 30 ( i . e . number of filters ) followed by max-pooling along the sequence-length dimension and two fully-connected layers of width 20 , with Swish activation .","For regularization , the L2 weight penalty was 0 . 001 , dropout rate was 0 . 1 , and batch normalization was used .","The model was trained with an initial learning rate of 0 . 001 in the same manner as the amino acid composition-based network , except for up to 500 epochs .","After finding this optimal architecture , the test dataset was distributed across the 10 training datasets and 10 new models were trained , together comprising PADDLE .","The mean value of these 10 predictors was used as the final prediction .","This model was used for all predictions on yeast TFs and nuclear proteins , human TFs , and virus transcriptional regulators .","Because secondary structure prediction was by far the slowest aspect of running PADDLE predictions , we trained a version of PADDLE called PADDLE-noSS which only used the one-hot sequence encoding for input , no secondary structure or disorder predictions .","PADDLE-noSS was nearly as accurate as PADDLE ( R2 scores of 0 . 76 , 0 . 52 , and 0 . 89 for TF tiles , scramble mutants , and Pdr1 mutants and homologs respectively; compare to Figure 3—figure supplement 1B ) and was used for predictions of core ADs and mutant sequences in Figure 3E–G , Figure 3—figure supplement 1E–I , and Figure 4—figure supplement 1F .","Human TFs were defined by Gene Ontology term GO:003700 ( DNA-binding transcription factor activity ) .","Virus transcriptional regulators ( vTRs ) were taken from Figure 1—source data 1 from Liu et al . , 2020 .","Within each human TF or vTR , predictions on 53-aa tiles at 1-aa resolution were generated using PADDLE and smoothed with a 9-aa moving average .","High-strength ADs were defined as the union of at least 5 consecutive 53-aa tiles with a predicted Z-score greater than 6 .","ADs defined in this way that overlapped by more than 26 aa were combined .","Medium-strength ADs were defined similarly with a Z-score cutoff of 4 and were required to not overlap with high-strength ADs by more than 26 aa .","All protein positions reported in the text are 1-indexed and inclusive; all annotations reported in supplemental tables are 0-indexed and inclusive of the start position but exclusive of the stop position .","For testing in a luciferase assay , 50 predicted high-strength ADs from human TFs were chosen at random and the tile with the strongest ( smoothed ) predicted Z-score from each AD was used .","Yeast nuclear proteins were defined by Gene Ontology term GO:0005634 ( Nucleus ) .","Within each protein , predictions on 53-aa tiles at 1-aa resolution were generated using PADDLE and smoothed with a 9-aa moving average .","All predicted Z-score local maxima higher than 3 . 09 ( p<0 . 001 ) were considered for experimental testing .","The 53-aa tiles corresponding to the local maxima were included , starting from those with the largest predicted Z-score , unless they overlapped a previously-included tile by at least 15 aa .","Analysis of experimentally-confirmed ADs focused on proteins that had experimental or high-throughput evidence for nuclear localization .","Proteins involved in transcriptional regulation were defined by GO:0006355 ( regulation of transcription , DNA-templated ) .","To calculate enrichment of activating tiles in coactivator proteins , we generated a null model in which ‘activating’ tiles , equal in number to actual activating tiles discovered in our screen , were randomly selected from among non-TF nuclear proteins .","We then counted how many of these randomly-selected tiles were in coactivator proteins , and then repeated this sampling 10 , 000 times to generate a null distribution .","The actual number of tiles , 42 , is 2 . 9-fold more than the mean of the null; p<10−12 by one tailed Z-test .","Predicted disorder for each tile was computed from D2P2 annotations as the fraction of the tile’s residues that were annotated as disordered among all 8 of the predictors listed in D2P2 .","To define core ADs , predictions of sequences shorter than 53 aa were done using PADDLE-noSS by embedding the shorter region in 100 different neutral sequences composed of AGSTNQV residues and averaging the predicted values .","Predictions of all tiles 5–30 aa in length within every AD were computed in this manner , and used to define core ADs for Figure 3F–G , Figure 7A , and Figure 7—figure supplement 1B .","Specifically , core ADs were defined as the shortest tiles with predicted Z-score greater than 3 . 09 ( p<0 . 001 ) that do not overlap another shorter core AD .","In Figure 3—figure supplement 1I , the 20-aa region within each AD with the highest predicted Z-score was included if it had Z-score greater than 3 . 09 , and the effects of single AA mutants in these 20-aa core ADs were predicted in the same manner using PADDLE-noSS .","Code for running PADDLE and PADDLE-noSS is available at https://github . com/asanborn/PADDLE ( Sanborn , 2021; copy archived at swh:1:rev:17aa36985e474d8593b99d593f7cc5e7c3e205a0 ) .","Pre-generated PADDLE predictions from multiple species are available for download at http://paddle . stanford . edu .","To generate structural models of binding , we took 28 core ADs peptides 13 residues in length and docked them to the ABDs using Rosetta3 ( Leaver-Fay et al . , 2011 ) , specifically its FlexPepDock ab initio pipeline ( Raveh et al . , 2011 ) .","Core ADs were used as the peptide , and activator-binding domain as the receptor .","Peptides were initialized in an extended conformation using Rosetta’s BuildPeptide application and manually placed near the binding pocket .","For the KIX domain , NMR chemical shift data was used to aid in placement ( Thakur et al . , 2008 ) .","For ABD1 , the AD of Gcn4 was used to aid in placement .","For all models of ABD1 , the first conformer from PDB structure 2LPB was used ( Brzovic et al . , 2011 ) .","For each AD-ABD pair , a set of 50 , 000 structural models were generated and ranked by FlexPepDock’s reweighted_sc score .","Binding poses were generated by clustering the top 500 ranked models using Rosetta’s cluster application with the default clustering distance threshold of 2 Å .","The best-scoring model from each cluster was designated as the cluster representative , and the cluster representatives that were also among the 10 best-scoring models were used for further analysis .","Additionally , larger runs with 500 , 000 structural models and clustering on the top 5000 models were run for six individual ADs for the KIX ABD .","The full list of ABDs and ADs used is available in Figure 6—source data","1 . PDB files of 10 best-scoring models from different binding poses and the score , cluster number , and RMSD to best model for the 500 best-scoring models are available in Figure 6—source data","2 . Due to the large number of peptides that needed to be docked , the FlexPepDock ab-initio pipeline was modified to run in a cluster setting at a large scale and used approximately 100 , 000 CPU hours on the Stanford Sherlock academic cluster .","Secondary structure was calculated using DSSP ( Kabsch and Sander , 1983; Touw et al . , 2015 ) and annotations of G , H , or I were considered alpha-helical .","Interaction types were computed with GetContacts ( Venkatakrishnan et al . , 2019 ) and per-AA accessible surface areas were computed with FreeSASA ( Mitternacht , 2016 ) .","For each final docked structural model , buried peptide surface area was taken by subtracting the accessible surface area of the peptide in complex with the receptor from the accessible surface area of the peptide on its own .","For Figure 6E , marked positions correspond to atom CG of Asp , atom CG of Leu , and the midpoint of the CG and CZ atoms in Phe ( the center of the aromatic ring ) .","Residues are shown from all the 10 best models from different clusters for all core AD structures if the marked position is within 3 Angstroms of the KIX domain or ABD1 structure .","Adapted FlexPepDock code , and wrapper code for GetContacts and FreeSASA are available at https://github . com/drorlab/med15 ."]],"string":"[\n [\n \"Transcription factors ( TFs ) perform the last step in signal transduction pathways .\",\n \"They thus serve key roles in central processes such as growth , stress response , and development , and their mutation or misregulation underlies many human diseases ( Spitz and Furlong , 2012 ) .\",\n \"A TF includes a sequence-specific DNA-binding domain ( DBD ) and an effector domain that regulates nearby gene transcription .\",\n \"Activation domains ( ADs ) – effector domains that increase transcription – have long been of particular interest due to their roles as oncogenic drivers and use as scientific tools ( Bradner et al . , 2017; Brückner et al . , 2009; Dominguez et al . , 2016 ) .\",\n \"ADs were discovered as regions that could independently stimulate transcription when ectopically recruited to a gene promoter ( Brent and Ptashne , 1985 ) .\",\n \"Early experiments showed that ADs were unlike structured domains because progressive truncations showed graded reductions in activity ( Hope and Struhl , 1986; Hope et al . , 1988 ) .\",\n \"Subsequent studies showed that ADs were disordered and had few similarities in their primary sequence ( Mitchell and Tjian , 1989 ) .\",\n \"Instead , ADs were classified based on their enrichment of certain residues , whether acidic , glutamine-rich , or proline-rich .\",\n \"Acidic ADs are the most common and best characterized .\",\n \"Acidic ADs retain activity when transferred between yeast and animals , pointing to a conserved eukaryotic mechanism ( Fischer et al . , 1988; Struhl , 1988 ) .\",\n \"While some have found that acidic residues are necessary for activation , others have found that they are dispensable ( Brzovic et al . , 2011; Pacheco et al . , 2018; Staller et al . , 2018; Warfield et al . , 2014 ) .\",\n \"Besides their negative charge , acidic ADs are rich in bulky hydrophobic residues .\",\n \"Mutating these hydrophobic residues reduces activation , often in proportion to the number mutated .\",\n \"Because AD sequences are highly diverse and poorly conserved , only a small fraction of all ADs in eukaryotic TFs have likely been annotated .\",\n \"Sequence motifs have been proposed based on analysis of select ADs but have not been used for large-scale prediction ( Piskacek et al . , 2007 ) .\",\n \"Screens of random sequences in yeast identified many activating sequences that represented as many as 1–4% of elements tested ( Hackett et al . , 2020; Ravarani et al . , 2018 ) .\",\n \"Actual protein sequences are , however , highly non-random .\",\n \"Direct screening of protein sequences has identified relatively few ADs at low resolution ( Arnold et al . , 2018; Tycko et al . , 2020 ) .\",\n \"There is a need for methods to experimentally detect or computationally predict all ADs .\",\n \"ADs stimulate transcription of genes by recruiting coactivator complexes , especially the Mediator complex , which interacts with RNA Polymerase II and regulates its transcription initiation ( Kornberg , 2005 ) .\",\n \"Mediator is necessary for TF-dependent activation in vitro and in vivo and is required for regulation by enhancers in all eukaryotes ( Allen and Taatjes , 2015 ) .\",\n \"Mediator and TFs are concentrated at strong enhancers , perhaps in a phase-separated state , which may play a role in gene activation ( Boija et al . , 2018; Chong et al . , 2018; Sabari et al . , 2018; Shrinivas et al . , 2019; Whyte et al . , 2013 ) .\",\n \"Beyond Mediator , TFs have been suggested to recruit other conserved multiprotein complexes , including TFIID , SAGA , and SWI/SNF , that play roles in activation of various genes ( Hahn and Young , 2011; Mitchell and Tjian , 1989 ) .\",\n \"While many such TF interactions have been described , their occurrence and roles remain to be determined .\",\n \"Biochemical studies of TF-coactivator interaction have given insight into the mechanism of activation .\",\n \"Nuclear magnetic resonance ( NMR ) studies revealed short alpha helices formed by acidic ADs of yeast Gcn4 protein , with hydrophobic faces contacting hydrophobic surfaces of the Med15 subunit of Mediator ( Brzovic et al . , 2011; Tuttle et al . , 2018 ) .\",\n \"NMR constraints were consistent with multiple possible AD binding poses , suggesting a dynamic ‘fuzzy complex’ ( Tompa and Fuxreiter , 2008 ) .\",\n \"Further consistent with these ideas , solvent exposure of aromatic residues was associated with activation in a screen of Gcn4 AD variants ( Staller et al . , 2018 ) .\",\n \"Here , we combine quantitative , high-throughput measurements of in vivo activation and in vitro interaction with computational modeling to characterize ADs .\",\n \"We identify 150 ADs in budding yeast and describe their shared sequence attributes .\",\n \"We extend the analysis by training and validating a neural network that predicts new ADs across eukaryotes .\",\n \"Guided by predictions , we design and measure activation of thousands of AD mutants to derive a deeper understanding of the principles underlying activation .\",\n \"Correlating activation with measurements of binding of ADs to Mediator in vitro identifies the key protein interactions driving activation , for which we predict atomic structures using a peptide-docking algorithm .\",\n \"We derive structural features common to AD-Mediator interactions and use them to explain measurements of Mediator recruitment kinetics in vitro .\",\n \"In total , we develop an integrated model for function , sequence determinants , molecular interactions , and kinetic features of TF ADs .\"\n ],\n [\n \"We identified ADs across all TFs in budding yeast with the use of a quantitative , uniform , and high-throughput activation assay .\",\n \"Variability due to protein expression , activation duration , and secondary genetic effects was minimized by fusing domains of interest to a three-part artificial TF ( aTF ) that ( 1 ) is tracked by an mCherry tag , ( 2 ) localizes to the nucleus only upon induction with estrogen , and ( 3 ) binds uniquely through its mouse DBD in the promoter of a chromosomally-integrated GFP reporter gene ( Figure 1A; McIsaac et al . , 2013; Staller et al . , 2018 ) .\",\n \"In this assay , three known ADs stimulated GFP expression greater than 100-fold upon estrogen treatment ( Figure 1—figure supplement 1A ) .\",\n \"We used a pooled screen to measure activation by 7460 protein segments , each 53 amino acids ( aa ) in length , that tiled all 164 TFs ( identified by Gene Ontology terms ) with a step size of 12–13 aa ( 3 . 8-fold average coverage; Figure 1B ) .\",\n \"Outgrowth of transformed yeast for 5 days ensured that each cell contained a unique aTF expression plasmid due to mechanisms maintaining it at single-copy levels ( Materials and methods ) ( Scanlon et al . , 2009 ) .\",\n \"After induction with estrogen , cells with a defined level of aTF expression were selected by mCherry signal and sorted into eight bins based on GFP expression ( Figure 1—figure supplement 1B ) .\",\n \"By next-generation sequencing , the distribution of each protein tile across the bins was determined and the mean value was used to calculate activation—namely , the fold increase in GFP relative to background ( Figure 1—source data 1 , Materials and methods ) .\",\n \"Activation was highly concordant between distinct DNA sequences encoding the same protein fragment ( Figure 1—figure supplement 1C ) .\",\n \"GFP distributions of three known ADs in the pooled screen exactly reproduced measurements from individual pure populations ( Figure 1—figure supplement 1D ) .\",\n \"Controlling by a defined aTF expression level was critical for precisely measuring activation ( Figure 1—figure supplement 1E–F ) .\",\n \"Activation measurements of fragments shared between two independently cloned and assayed libraries were reproducible ( Pearson's r = 0 . 953 , Figure 1—figure supplement 1G ) .\",\n \"Over 2 , 850 , 000 cells were assayed in total , giving excellent coverage of library elements ( approximately 99% of TF tiles were observed in at least 10 cells each ) .\",\n \"Our assay exhibited high signal-to-noise: tiles in previously known and newly identified ADs activated nearly 200-fold while 88% of fragments activated less than twofold ( Figure 1C ) .\",\n \"Across the library , 451 tiles showed significant activation ( p<0 . 0001 by Z-test ) .\",\n \"When plotted by protein position , activating tiles clustered into discrete , well-defined ADs ( Figure 1D ) .\",\n \"Using a positional activation score , we identified 150 ADs in 96 TFs ( Figure 1E and Figure 1—figure supplement 1H , Figure 1—source data 2 , Materials and methods ) .\",\n \"These ADs overlapped 75% of all previously-reported ADs in TFs ( Figure 1—source data 2 ) .\",\n \"Furthermore , the 53-aa tile length was not limiting , since our screen successfully identified ADs in over 85% of TFs that activated in a previous one-hybrid screen testing full-length proteins ( Figure 1—source data 2; Titz et al . , 2006 ) .\",\n \"These results show that our screen is both sensitive and comprehensive .\",\n \"Three-quarters ( 112 ) of ADs identified here were previously unknown ( Figure 1—source data 2 ) .\",\n \"While 63 TFs contained just a single AD , 33 TFs had multiple , including up to seven distinct ADs in Adr1 ( Figure 1D and F ) .\",\n \"C-terminal ADs were common—found in nearly half of all AD-containing TFs—and were stronger and shorter on average than other ADs ( Figure 1—figure supplement 1I ) .\",\n \"Consistent with AD function in their native context , TFs that contained ADs upregulated a higher proportion of downstream genes than TFs without ADs ( Figure 1—figure supplement 1J; Hackett et al . , 2020 ) .\",\n \"Consistent with the diverse functions of TFs beyond transcriptional activation ( e . g . repression ) , 68 TFs did not contain any ADs ( Figure 1F ) .\",\n \"ADs were enriched in negative charge and hydrophobic content , and activation was strongest for fragments that were high in both ( Figure 2A and Figure 2—figure supplement 1A–B; Wimley-White hydrophobicity scale ) ( Wimley and White , 1996 ) .\",\n \"Past experiments have clearly shown the importance of bulky hydrophobic residues but presented conflicting evidence for acidic residues in activation ( Brzovic et al . , 2011; Cress and Triezenberg , 1991; Drysdale et al . , 1995; Jackson et al . , 1996; Pacheco et al . , 2018; Staller et al . , 2018; Sullivan et al . , 1998; Warfield et al . , 2014 ) .\",\n \"To determine the role of negative charge , we took the strongest-activating fragment from each AD and measured its activation with all acidic residues mutated .\",\n \"Except for a single example that activated just as strongly at +7 net charge without its acidic residues ( the Gln3 C-terminal AD ) , activation was abolished in all ADs , showing definitively that acidic residues are necessary in budding yeast ADs ( Figure 2B and Figure 2—figure supplement 1C ) .\",\n \"We also noticed that activation of wild-type tiles was more strongly correlated with the number of Asp residues than the number of Glu residues , and hypothesized that Asp promotes activation more strongly than Glu ( Figure 2—figure supplement 1D ) .\",\n \"Indeed , ADs with all Glu residues mutated to Asp consistently activated as well as or better than ADs with all Asp residues mutated to Glu ( Figure 2—figure supplement 1E ) .\",\n \"Having identified the important residue types , we looked for sequence motifs that could predict activation .\",\n \"However , AD sequences appeared to share no common motifs upon inspection .\",\n \"The 9aaTAD motif , originally proposed based on yeast ADs ( Piskacek et al . , 2007 ) , predicted ADs poorly: only 10% of 9aaTAD-containing tiles activated in our assay , which was not significantly more predictive than scrambled versions of the motif ( Figure 2C ) .\",\n \"A de novo search for motifs using DREME also found none that were shared by more than two ADs ( Bailey , 2011 ) .\",\n \"Evidently , ADs are diverse in sequence and not predictable from simple motifs .\",\n \"The lack of motifs suggests that activation is determined not by positions of individual residues but by distributed or redundant features .\",\n \"We pursued this point by measuring activation of a panel of designed mutants of eight ADs .\",\n \"Indeed , 94% of scanning single-amino acid mutants affected activation by less than twofold , and all still activated more than 10-fold ( Figure 2—figure supplement 1F–G ) .\",\n \"To produce larger effects , we varied the acidic and hydrophobic content of the ADs by mutating successively larger subsets of acidic or aromatic residues ( Figure 2D , Materials and methods ) .\",\n \"Activation was gradually , and finally completely , eliminated as either characteristic was reduced .\",\n \"It is noteworthy that the Gcn4 AD , a focus of many past studies , uniquely tolerated mutation of up to six acidic residues .\",\n \"Most notably , activation by acidic and hydrophobic mutants were related to the number , rather than position , of mutated residues ( Figure 2D and Figure 2—figure supplement 1H ) .\",\n \"Thus , activation strength is determined in large part by acidic and hydrophobic content , not by motifs .\",\n \"Since motifs predicted ADs poorly , we turned to more sophisticated approaches using machine learning .\",\n \"We first trained a simple neural network to predict activation based solely on tiles’ amino acid ( aa ) composition ( Materials and methods ) .\",\n \"When tested on TF tiles withheld from training , this neural network performed remarkably well , explaining 66% of observed variation ( Figure 3—figure supplement 1A ) .\",\n \"However , this algorithm was limited by its lack of information about residue positions .\",\n \"To experimentally disentangle the contributions of aa positioning and aa composition , we measured activation of eight scrambled sequences from each of the eight ADs tested before .\",\n \"Despite their shared abundance of acidic and hydrophobic residues , these mutants spanned a wide range of activation strengths , and some mutants activated stronger than wild-type while others failed to activate entirely ( Figure 3A ) .\",\n \"Clearly , full sequence information beyond aa composition is needed to predict activation .\",\n \"We therefore trained a deep convolutional neural network ( CNN ) to predict activation based on protein sequence , predicted secondary structure , and predicted disorder ( Materials and methods ) .\",\n \"CNNs evaluate sequences by hierarchically integrating matches to a suite of learned patterns and have recently found great success in many genomic prediction tasks ( Ching et al . , 2018; Kelley et al . , 2016; Wang et al . , 2016 ) .\",\n \"Our Predictor of Activation Domains using Deep Learning in Eukaryotes , or ‘PADDLE’ , explained 81% of observed variation in TF tiles withheld from training ( alternatively , area under the precision-recall curve of 0 . 805; Figure 3—figure supplement 1B–C , Figure 3—source data 1 ) , markedly better than the aa composition-based predictor .\",\n \"De novo , PADDLE accurately predicted the activation strength of ( 1 ) new ADs within TFs omitted from training ( R2 score = 0 . 81 across tiles from 22 TFs; example predictions for Arg81 in Figure 3B ) ; ( 2 ) scrambled AD sequences , despite their identical amino acid composition ( R2 score = 0 . 61 ) ; and ( 3 ) 232 mutants and 178 orthologs of the Pdr1 AD ( R2 score = 0 . 90 ) ( Figure 3—figure supplement 1B ) .\",\n \"Altogether , the performance of PADDLE was validated across a wide range of both wild-type and mutant sequences in yeast .\",\n \"Classic experiments showed that many acidic ADs retained activity even when transferred between yeast and animals ( Fischer et al . , 1988; Struhl , 1988 ) , so we investigated whether PADDLE could identify ADs in human TFs that would activate in human cells .\",\n \"Using PADDLE , we predicted 236 high-strength and 366 moderate-strength ADs , together spanning 462 ( 27% ) human TFs ( Figure 3C , Figure 3—source data 1 ) .\",\n \"These predicted ADs overlapped many known ADs of TFs from diverse families , including p53 , NFkB , Myc , Klf4 , Fos , PPARA , SREBF1 , E2F proteins , and the glucocorticoid receptor ( Figure 3—figure supplement 1D ) .\",\n \"PADDLE also predicted 41 high-strength and 45 moderate-strength ADs from among 419 transcription-regulating viral proteins ( Figure 3—source data 1; Liu et al . , 2020 ) .\",\n \"We randomly selected 25 high-strength predicted ADs from human TFs and measured their activation individually using a luciferase reporter in HEK293T cells .\",\n \"Remarkably , 23 domains ( 92% ) activated luciferase expression ( Figure 3D ) .\",\n \"While other classes of human ADs , such as glutamine-rich or proline-rich ADs , are not predictable by PADDLE since they are not present in S . cerevisiae , acidic activation mechanisms are evidently conserved and the patterns learned by PADDLE are generalizable across eukaryotes .\",\n \"To uncover the principles of acidic activation learned by PADDLE , we analyzed predictions on designed mutant sequences and developed hypotheses , which we then experimentally tested in a second library .\",\n \"We noticed from predictions on sequences less than 30 aa that yeast ADs contained short , independently activating regions , which we term core ADs ( cADs ) ( Figure 3—figure supplement 1E ) .\",\n \"As an experimental test , we measured in vivo activation of 13-aa-long fragments tiling ten ADs in 1-aa steps .\",\n \"Predictions proved highly accurate ( R2 score = 0 . 84 , Figure 3—figure supplement 1F ) , and both predictions and experiments identified at least one significantly activating 13-aa cAD within each AD ( Figure 3E; p<0 . 001 ) .\",\n \"Across all yeast TFs , PADDLE predicted that the minimal region for activation could be localized to within a 20-aa cAD in 85% of ADs ( Figure 3F , Figure 3—source data 1 ) .\",\n \"These cADs still shared no motifs ( Figure 3—figure supplement 1G ) , but instead had an especially high density of acidic and hydrophobic residues , even more so than the full ADs ( Figure 3G ) .\",\n \"To quantify the contribution of each residue to activation , we predicted the impact of mutating each position in each cAD to alanine ( Figure 3—figure supplement 1H ) .\",\n \"Single-aa mutants in these short cADs had a large predicted impact on activation , unlike in the longer 53-aa ADs ( Figure 2—figure supplement 1F–G ) , and showed clear trends when grouped by residue identity .\",\n \"Mutating the bulkiest hydrophobic residues led to the greatest decreases in predicted activation , and the three most important were Trp , Phe , and Leu ( Figure 3—figure supplement 1I ) .\",\n \"Mutating acidic or basic residues had opposite but comparable effects on predicted activation .\",\n \"Together , these analyses identify the regions and residues across ADs that contribute most to activation .\",\n \"Having identified the core AD regions , we sought to understand the sequence properties that drive their activation .\",\n \"We focused on the 28 strongest 13-aa cADs ( Figure 3—source data 1 ) , and first experimentally quantified the importance of their aa composition by comparing each cAD with 33 random scrambles of its sequence .\",\n \"Remarkably , only nine cADs ( 32% ) showed activation by the wild-type sequence greater than threefold greater than the average activation by scrambled sequences ( Figure 4A ) .\",\n \"In these nine cADs , maximal activation evidently requires a specific positioning of residues .\",\n \"On the other hand , 18 cADs ( 64% ) showed activation by the wild-type sequence within twofold of the scrambled sequence average , indicating that their activation is primarily determined by aa composition and not by a unique positioning of residues .\",\n \"The wild-type-to-scramble ratio was not correlated with wild-type activation strength ( Figure 4—figure supplement 1A , Pearson’s r = 0 . 32 , p=0 . 09 ) .\",\n \"Instead , cADs with the greatest hydrophobic content had the lowest wild-type-to-scramble ratios ( Figure 4B , Pearson’s r = −0 . 51 , p=0 . 006 ) .\",\n \"Thus , the majority of cADs studied here activate by a composition-driven mechanism that exploits an excess of bulky hydrophobic residues .\",\n \"Alpha-helical folding is thought to be important for activation , which is at odds with the prevalence of composition-driven cADs .\",\n \"We directly determined whether the 28 strongest 13-aa cADs require helical folding by measuring the impact of inserting a helix-breaking proline .\",\n \"Within the seven central positions of each 13-aa cAD , we individually mutated each non-hydrophobic residue to alanine or proline and asked whether proline inhibited activation relative to alanine ( Figure 4C ) .\",\n \"Proline mutations inhibited activation more than threefold over alanine in nine ADs ( 32% ) , including up to an 11-fold drop in activation in the Tda9 cAD .\",\n \"However , proline mutations inhibited activation by less than twofold in 16 cADs ( 58% ) .\",\n \"These effects were consistent between different positions within the same cAD ( Figure 4—figure supplement 1B ) .\",\n \"Interestingly , all three cADs that contained a basic residue ( Pul4 , Tda9 , and Rsf2:586 ) were strongly inhibited by proline , suggesting that a helix is necessary to position their inhibitory positive charge away from the coactivator binding interface .\",\n \"Most notably , the magnitude of proline disruption was tightly correlated with the wild-type-to-scramble ratio ( Figure 4D , Pearson’s r = 0 . 842 ) , showing that constraints on structure and sequence , or a lack thereof , were closely linked .\",\n \"Thus , composition-driven cADs typically do not need to form a helix and likely sample disordered conformations , while cADs that require their wild-type sequence typically also require helical folding , employing a structure-driven mechanism .\",\n \"To understand these two contrasting mechanisms , we studied them in a simplified system of artificial cADs that used only the four most activation-promoting residues .\",\n \"We examined 9-aa sequences ( denoted ‘9mers’ ) consisting entirely of only two types of amino acids each—Asp and either Leu , Phe , or Trp—so that all possible such sequences ( 3 × 29 = 1536 ) could be systematically assayed ( Figure 4E ) .\",\n \"As expected , activation required a balance of hydrophobic and acidic residues ( Figure 4—figure supplement 1C ) , so we focused analysis on aa compositions with the strongest median activation: 9mers with five or six hydrophobic residues .\",\n \"Within these , nearly all Phe 9mers and Trp 9mers activated regardless of their sequence ( Figure 4F ) , characteristic of a composition-driven mechanism and consistent with the highly hydrophobic nature of Phe and Trp residues .\",\n \"However , Leu 9mers spanned a wide range of activity with only a minority that were strongly activating , characteristic of structure-driven activation requiring a specific ordering of residues .\",\n \"To see what allows only certain Leu 9mers to activate , we examined all possible short motifs containing two or three Leu residues .\",\n \"The only motif strongly associated with activation was LxxLL ( Figure 4G and Figure 4—figure supplement 1D; p<10−8 by Kolmogorov-Smirnov test ) .\",\n \"This motif was strictly required for activation in these 9mers , present in two copies in the three strongest 9mers , found in two wild-type structure-driven cADs ( Put3 and Crz1 ) , and previously described in ADs of many nuclear receptors and other TFs ( Plevin et al . , 2005 ) .\",\n \"This motif also suggests an explanation for why helical folding was necessary for cADs with less hydrophobicity: by placing Leu residues along one face of an alpha helix , the motif efficiently forms a hydrophobic binding surface .\",\n \"The analogous FxxFF and WxxWW motifs were also significantly but more weakly associated with activation of Phe 9mers and Trp 9mers ( Figure 4—figure supplement 1E ) .\",\n \"However , their strongest predictor of activation was the average position of acidic residues , with N-terminal skew highly favored ( Figure 4H ) .\",\n \"Most dramatically , DDDFFFFFF activated 20-fold while its reverse sequence FFFFFFDDD activated just 1 . 3-fold .\",\n \"One role of N-terminal negative charge could be to neutralize the macroscopic dipole that arises from alpha helix backbone hydrogen bonds , stabilizing helical conformations ( Rocklin et al . , 2017 ) .\",\n \"Even though a helix is not necessary for composition-driven activation , this effect could reduce the entropic cost of binding to coactivators by promoting transient folding .\",\n \"Despite the importance of hydrophobic residues for activation , PADDLE also predicted that some TFs contained cADs whose activation was inhibited by nearby clusters of hydrophobic residues .\",\n \"We confirmed this experimentally for Rds1 , in which residues 239–263 alone activated 6 . 8-fold but extending the sequence by eight aa , five of them hydrophobic , abolished activation ( Figure 4I ) .\",\n \"Inhibition by hydrophobic residues also explains the weak activation of some scrambled 53-aa ADs ( Figure 3A , Figure 4—figure supplement 1F ) .\",\n \"To systematically quantify inhibition , we measured the effect on activation when the Pdr1 cAD was placed next to a library of 2177 random 20-aa sequences chosen to span a wide range of net charge and hydrophobic content ( Figure 4J ) .\",\n \"As expected , negative charge with moderate hydrophobicity bolstered activation .\",\n \"Positive charge without hydrophobicity inhibited activation , but even variable regions with +8 charge ( +4 net charge ) still activated an average of 4 . 4-fold .\",\n \"Therefore , the local clustering of negative charges near hydrophobic residues renders the Pdr1 cAD resistant to global changes in net charge , suggesting that interactions between opposite charges within the peptide do not fully prevent binding to coactivators .\",\n \"In contrast , hydrophobic residues at high density inhibited activation completely .\",\n \"This suggests that inhibitory hydrophobic clusters interact with hydrophobic residues in the cAD , completely preventing their interaction with coactivators .\",\n \"Consistent with two independent mechanisms of inhibition , basic and hydrophobic residues together reduced activation additively .\",\n \"In total , these experiments mapped the biochemical and structural properties that drive both activation and inhibition .\",\n \"The surprising ability of simple biochemical features to explain a large and diverse set of ADs suggests that these features may also drive interactions with coactivator complexes .\",\n \"To gain a mechanistic understanding of the AD-coactivator contacts that underlie activation , we mapped the interactions of wild-type and mutant ADs with Mediator .\",\n \"We focused on Med15 , the primary subunit of Mediator targeted by many ADs in budding yeast , which contains four tandem activator-binding domains ( ABDs ) ( Herbig et al . , 2010; Jedidi et al . , 2010; Thakur et al . , 2008 ) .\",\n \"We isolated the N-terminal portion of Med15 consisting of its four ABDs ( hereafter just called 'Med15' ) , confirmed its in vitro interaction with Gcn4 , and used it for binding experiments ( Figure 5—figure supplement 1A ) .\",\n \"To measure binding in high-throughput , we used mRNA display ( Takahashi and Roberts , 2009 ) , expressing our library of TF tiles as a pool of protein fragments covalently tagged with their mRNA sequences , and using this pool in Med15 pull-down experiments ( Figure 5A ) .\",\n \"Direct counts of Med15-bound and input protein molecules were obtained by amplifying and sequencing their mRNA tags and using unique molecular identifiers ( UMIs ) to remove PCR duplicates .\",\n \"Finally , the fractional pull-down of each protein fragment was computed by its abundance in the Med15-bound sample relative to input , normalized to total library concentrations measured by qPCR ( Materials and methods ) .\",\n \"Pull-down measurements were reproducible between replicates ( Pearson’s r > 0 . 90 ) and consistent between identical protein fragments encoded by distinct but synonymous mRNA sequences ( Figure 5—figure supplement 1B–C ) .\",\n \"Fragments known to bind Med15 enriched up to 17-fold above random sequence controls and 1000-fold higher than in mock pull-downs with beads alone ( Figure 5—figure supplement 1D , Figure 5—source data 1 ) .\",\n \"Binding to Med15 was widespread among ADs and coincident with activation activity .\",\n \"In total , 324 tiles bound Med15 significantly more than negative controls ( p<0 . 001 by Z-test , Figure 5—figure supplement 1E , Materials and methods ) .\",\n \"When plotted by protein position , Med15-binding tiles clustered into 153 discrete domains , the vast majority of which were not previously known to bind Mediator ( Figure 5B , Figure 5—source data 1 ) .\",\n \"ADs and Med15-binding domains overlapped substantially: 73% of ADs bound Med15 , and 71% of Med15-binding domains functioned as ADs ( Figure 5C ) .\",\n \"Moreover , ADs that did not bind Med15 tended to activate weakly ( Figure 5—figure supplement 1F ) , so it is possible that many bind Med15 at levels below the detection limit of our pull-down assay , which was less sensitive than our activation assay ( compare Figure 1C and Figure 5—figure supplement 1E ) .\",\n \"Just as with activation , high acidic and hydrophobic content were associated with Med15 binding ( Figure 5—figure supplement 1G ) .\",\n \"The Gln3 AD that did not require acidic residues also did not bind Med15 substantially ( 2 . 1-fold enrichment ) , suggesting that it activates through other mechanisms .\",\n \"Most notably , activation strength and Med15 binding of tiles were directly proportional ( Figure 5D ) .\",\n \"To test this relationship further , we measured Med15 binding of the set of AD mutants in which aromatic residues were systematically removed ( Figure 2D ) .\",\n \"In every case , Med15 binding and activation were strongly correlated ( Figure 5E ) , showing that the hydrophobic features that drive activation similarly underlie Med15 binding .\",\n \"Moreover , the ability for binding of a single protein in vitro to quantitatively explain activation in vivo suggests that Mediator recruitment is a key driver for gene activation and determined largely by AD binding to its Med15 subunit .\",\n \"ADs can bind different coactivators , though the extent of such interaction is unknown .\",\n \"To explore the possibility , we also examined binding of TF tiles to TFIID , a key factor in the transcription of nearly all genes ( Warfield et al . , 2017 ) .\",\n \"Yeast TFIID contains three known AD-binding subunits , Taf4 , Taf5 , and Taf12 , which form a subcomplex with Taf6 and Taf9 ( Layer et al . , 2010; Wright et al . , 2006 ) .\",\n \"We isolated this TFIID subcomplex and confirmed its interaction with the VP16 AD ( Figure 5—figure supplement 1H ) .\",\n \"Pull-downs of the TF tiling library with the TFIID subcomplex enriched fragments up to 6 . 7-fold above random sequence controls ( Figure 5—figure supplement 1I–J , Figure 5—source data 1 ) .\",\n \"In total , 27 ADs ( 18% ) across 26 TFs bound the TFIID subcomplex ( p<0 . 001 ) , greatly expanding its repertoire of known interactions ( Figure 5B–C ) .\",\n \"However , all ADs that bound TFIID also bound Med15 , and despite the identical pull-down conditions , ADs bound more weakly to TFIID than to Med15 ( Figure 5—figure supplement 1K ) .\",\n \"While more TFIID interactions may be discoverable if assay sensitivity can be further optimized , these results suggest that TFIID has lower affinity for most ADs and provide no evidence for its specific targeting as a coactivator .\",\n \"Since TFIID binding by ADs is redundant and less frequent , its role in most TF-directed activation is secondary to that of Mediator .\",\n \"What structural features of Med15 enable its promiscuous yet functional interaction with diverse AD sequences ?\",\n \"In the best studied example , the Gcn4 AD interacts with hydrophobic patches and basic residues on multiple Med15 activator-binding domains ( ABDs ) in a large number of binding poses to form a ‘fuzzy’ complex ( Brzovic et al . , 2011; Tuttle et al . , 2018 ) .\",\n \"We addressed the question by using FlexPepDock , a peptide docking algorithm from the Rosetta suite ( Leaver-Fay et al . , 2011; Raveh et al . , 2011 ) , to systematically build structural models of ABD-AD interactions , made computationally tractable by our identification of short cADs .\",\n \"Our modeling focused on two ABDs with known structures: the KIX domain , which has homology to human Med15 and the p300/CBP coactivator family , and ABD1 ( Figure 6A; Brzovic et al . , 2011; Thakur et al . , 2008 ) .\",\n \"To sample diverse sequences , aa composition , and secondary structure , we modeled interactions of the 28 13-aa cADs described above ( Figure 4A–D ) with the KIX domain and ABD1 .\",\n \"For each interaction , 50 , 000 candidate structural models were sampled and ranked by the Rosetta energy score , and the 10 best-scoring models from each interaction were used in subsequent analyses ( Figure 6—source data 1 and Figure 6—source data 2 ) .\",\n \"We validated our structural modeling by comparison to experimental data in two ways .\",\n \"First , KIX domain residues previously shown by NMR to be important for interacting with the Pdr1 AD were recapitulated by the best-scoring model of this interaction ( Figure 6—figure supplement 1A; Thakur et al . , 2008 ) .\",\n \"Second , the degree of alpha helix formation across the 28 cADs was consistent with the effect of proline insertion on in vivo activation ( Figure 6—figure supplement 1B ) .\",\n \"As expected , cADs that were inhibited by proline predominantly bound both domains as an alpha helix , and conversely cADs that bound both domains in disordered conformations were minimally affected by proline insertion .\",\n \"Some cADs that were minimally affected by proline insertion were modeled as alpha helical , possibly because disordered regions are underrepresented in protein structure databases on which Rosetta was trained .\",\n \"A unifying feature of the interactions was the prominent role of hydrophobic contacts at the protein interface .\",\n \"Hydrophobicity-driven interactions between proteins often employ high shape complementarity to maximize interface area and minimize solvent-exposed area .\",\n \"However , the flat surface of the KIX domain was ineffective in burying hydrophobic cAD residues , which on average had as much surface area exposed to solvent as area contacting the KIX domain ( Figure 6B , top ) .\",\n \"For example , the best-scoring Pdr1 AD model used Ile and Leu residues along one helical face to contact the KIX domain , but this left Trp and Tyr residues on the opposite face exposed to solvent ( Figure 6—figure supplement 1C ) .\",\n \"The poor hydrophobic shape complementarity of the KIX domain was frequently mitigated by multiple salt bridges and cation-pi interactions ( Figure 6—figure supplement 1D ) .\",\n \"ABD1 , which formed fewer ionic interactions , engaged cADs through a larger surface formed by its contoured hydrophobic ridge ( Figure 6B , bottom ) .\",\n \"Nevertheless , hydrophobic residues of most cADs were incompletely buried , with those residues showing greater than 20% of area exposed to solvent in over half of all structures .\",\n \"Thus , despite the central role of hydrophobic AD residues in activation and Med15 binding , hydrophobic shape complementarity is not an important feature of ABD-cAD interactions .\",\n \"Consistent with weak constraints on the shape of the binding interface , all cADs bound in diverse poses .\",\n \"For example , the 10 best-scoring models of the War1 cAD bound the KIX domain with its helix axis at different orientations and with different helical faces presented toward the interaction surface ( Figure 6C ) .\",\n \"To quantify this diversity , we defined the unique binding poses of each ABD-cAD interaction by clustering similar structures based on cAD backbone root mean square distance ( 2 Å distance cutoff ) , and then counted the number of poses adopted by the 10 best-scoring models .\",\n \"The 10 best-scoring models of nearly every cAD , when interacting with either domain , adopted six or more distinct poses ( Figure 6D ) .\",\n \"To ensure that this did not result from insufficient sampling , we generated tenfold more candidate models ( 500 , 000 total ) for six cADs binding the KIX domain .\",\n \"For all six cADs , the 10 best-scoring models still adopted seven or more distinct poses , despite spanning a narrower range of Rosetta scores ( Figure 6—figure supplement 1E–F ) .\",\n \"Disordered cADs , lacking helical content , adopted especially many distinct poses; also , poses were frequently sampled only once , suggesting that disordered cADs inhabit an extremely large conformational space that may reduce the entropic cost of binding ( Figure 6—figure supplement 1G–H ) .\",\n \"Altogether , these structures point to a shape-agnostic nature of Med15 ABD interaction surfaces and support the idea of a multi-pose , fuzzy mode of interaction .\",\n \"To explore the consequence of fuzzy binding for amino acid sequence constraints , we took Leu , Phe , and Asp residues from models of all cADs and plotted their binding positions on the KIX and ABD1 interfaces ( Figure 6E ) .\",\n \"As expected , the hydrophobic Leu and Phe residues occupied regions distinct from the negatively charged Asp residues .\",\n \"However , the Leu and Phe distributions were similar to each other: there was no binding pocket that selectively preferred one residue over the other , despite large differences in their size and shape .\",\n \"The diversity of AD sequences may thus be understood in terms of the fuzzy nature of the AD-ABD interaction , which only loosely constrains the characteristics of hydrophobic AD residues .\",\n \"Pull-down experiments with individual ABDs showed no appreciable binding to any of nine ADs ( Figure 7—figure supplement 1A ) , consistent with the previous suggestion that multiple ABDs are required for strong binding to Med15 ( Brzovic et al . , 2011; Tuttle et al . , 2018 ) .\",\n \"Multiple interaction sites were also common within ADs: PADDLE identified 42 ADs ( 28% ) that contained two or more non-overlapping cADs ( Figure 7—figure supplement 1B , Figure 3—source data 1 , Materials and methods ) .\",\n \"We took 47 pairs of adjacent cADs and measured activation by the two cADs individually or in tandem .\",\n \"For 40 of the pairs , the tandem cADs activated more strongly than expected from the combined activation of the individual cADs ( 4 . 0-fold median increase , Figure 7A ) .\",\n \"Thus , an increased valence of interaction sites in both Med15 and ADs drives stronger binding and activation .\",\n \"At larger scales , the number of interaction sites is further multiplied because some TFs contain multiple ADs , many TFs bind DNA as dimers , and many genes have several TF binding sites ( Hahn and Young , 2011; Spitz and Furlong , 2012 ) .\",\n \"To understand the advantages conferred by the high valence of TF-Mediator interactions , we studied the kinetics in vitro of the initial step of activation: TF-driven recruitment of Mediator to DNA .\",\n \"We performed these experiments with Gcn4 , which forms homodimers and has an extended AD consisting of three cADs .\",\n \"We isolated budding yeast Mediator complex , coupled DNA containing a Gcn4 motif to a NeutrAvidin-coated surface , added Gcn4 , and measured real-time binding of Mediator by surface plasmon resonance ( Figure 7B , Figure 7—figure supplement 1C ) .\",\n \"Mediator bound to the Gcn4-DNA complex with an apparent dissociation constant of KD = 14 nM ± 2 nM and an interaction half-life of 2 . 7 ± 0 . 2 min .\",\n \"We repeated the experiment with DNA containing 2 , 4 , or 6 copies of the Gcn4 motif .\",\n \"Mediator bound the resulting Gcn4-DNA complexes with identical on-rates .\",\n \"Mediator was also recruited more efficiently , with more Mediator bound per mole of Gcn4 than on templates with only one motif ( Figure 7—figure supplement 1D ) .\",\n \"The interaction half-life , however , increased with additional copies of the Gcn4 motif—up to 16 min when six motifs were present ( Figure 7B–C ) .\",\n \"If all copies bound Mediator independently , then the half-life should be the same , regardless of the number of copies .\",\n \"The presence of additional , neighboring ADs retards dissociation by enabling recapture of Mediator released from one AD through binding to another .\",\n \"A slower dissociation rate corresponds to an increase in apparent affinity constant .\",\n \"Conversely , we found that the multiplicity of ABDs in Mediator facilitated binding of Gcn4 ADs .\",\n \"When the surface plasmon resonance experiment was repeated with large excess Gcn4 in solution , there was almost no change in the measured on-rate .\",\n \"Even 160-fold excess of Gcn4 had only minimal effect on the rate of Mediator binding and final level achieved ( Figure 7—figure supplement 1E ) .\",\n \"To maintain stable TF-DNA complexes without TFs in solution , we repeated these experiments using a fusion of the Gcn4 AD to nuclease deficient EcoRI ( E111Q ) , which resides on DNA for many hours and also forms homodimers ( Wright et al . , 1989 ) .\",\n \"Still , Mediator with no Gcn4 competitor or with 250-fold excess of Gcn4 competitor bound similarly ( Figure 7D ) .\",\n \"Excess competitor Gcn4-DNA complexes also failed to slow Mediator binding to the surface .\",\n \"Evidently , binding of Gcn4 to Mediator does not appreciably block binding of other Gcn4 molecules .\",\n \"This behavior may be explained by the occurrence of multiple Gcn4-binding sites on Mediator and weak interaction of a cAD with any one site .\",\n \"Rapid association-dissociation equilibrium of a cAD allows a second Gcn4 molecule to interact with Gcn4-bound Mediator , increasing the frequency of exchange of one Gcn4 molecule by another ( Figure 7E ) .\",\n \"Such facilitated exchange provides a rationale for both the fuzzy nature of the AD-ABD interaction and the multiplicity of ABDs .\",\n \"PADDLE predictions and previous experiments ( Liu and Myers , 2015 ) identified two strong ADs in the Med2 subunit of Mediator , which is adjacent to Med15 in the Mediator complex , suggesting a broader role for ADs beyond TF proteins ( Figure 7—figure supplement 1F ) .\",\n \"We used PADDLE to search for ADs across all nuclear-localized non-TF proteins in yeast and selected 1485 fragments that showed potential for activation ( Materials and methods ) .\",\n \"Upon pooled in vivo testing , we found 290 fragments across 229 proteins that activated significantly ( p<0 . 001 by Z-test , Figure 1—source data 1 ) .\",\n \"To see how often PADDLE fails to identify acidic ADs , we also included 291 control fragments that had extremely high acidic and hydrophobic content but were predicted not to activate .\",\n \"Zero control fragments activated ( Figure 7—figure supplement 1G ) , showing that PADDLE likely identified all acidic ADs that exist and again demonstrating that AD function cannot be predicted from amino acid composition alone .\",\n \"ADs were remarkably widespread and were found in proteins associated with diverse functions including transcriptional regulation , RNA splicing , DNA repair , ribosome biogenesis , and protein degradation .\",\n \"We focused on the ADs in transcription-associated proteins .\",\n \"Fourteen ADs were found in proteins that bind DNA and regulate pathway-specific genes; these proteins were not included in the original tiling library because they bind DNA only through a TF partner .\",\n \"The two predicted Med2 ADs were also confirmed to activate , 36-fold and 25-fold in our assay .\",\n \"Notably , ADs were identified in all major coactivator and chromatin modifying complexes , including Mediator , TFIID , cohesin , condensin , SAGA , RSC , SWI/SNF , ISWI , NuA4 , and FACT ( Figure 7F ) .\",\n \"These complexes contained 2 . 9-fold more activating fragments than expected by chance ( p<10−12 compared to randomly shuffled controls ) , suggesting a functional role .\",\n \"Domains that activate on their own could be inactive in their natural context if buried within a protein or protein complex .\",\n \"To investigate this possibility , we cross-referenced coactivator ADs with all available structures in the Protein Data Bank ( PDB ) and with the Database of Disordered Protein Predictions ( D2P2 ) ( Berman et al . , 2000; Oates et al . , 2013 ) .\",\n \"In fact , 23 ADs were predicted in D2P2 to be highly disordered ( more than 30% ) or were predominantly disordered or unresolved in PDB structures ( Figure 7F ) , and so may be exposed and active in vivo .\",\n \"For example , cohesin , which binds Mediator and forms enhancer-promoter contacts through loop extrusion ( Davidson et al . , 2019; Kagey et al . , 2010; Kim et al . , 2019; Sanborn et al . , 2015 ) , contained strong and likely-disordered ADs in its subunits Scc1 ( RAD21 homolog ) , Scc3 ( STAG1/2 homolog ) , and Smc3 , as well as multiple ADs on the meiotic-specific subunit Rec8 .\",\n \"These cohesin ADs , as well as ADs within other coactivators , may play a role in gene regulation by driving interactions with Mediator .\"\n ],\n [\n \"We obtained high-throughput measurements of activation strength that were both quantitative and precise .\",\n \"This enabled us to detect—and predict with PADDLE—how activation strength depends on amino acid sequence , amino acid composition , and valence .\",\n \"Quantitation was achieved by inducing artificial TF ( aTF ) binding for a brief , defined period so that GFP levels were representative of transcriptional activation ( McIsaac et al . , 2013 ) , and by partitioning the range of GFP signal into eight levels for sorting .\",\n \"Others have measured activation only as an on-off binary system or screened for non-quantitative measures of activation , such as cell survival or pull-down of a cell surface marker ( Arnold et al . , 2018; Ravarani et al . , 2018; Tycko et al . , 2020 ) .\",\n \"We achieved high precision by restricting measurements to cells with a defined level of aTF expression .\",\n \"Controlling for this frequently overlooked confounding variable was critical because activation was strongly dependent on aTF abundance , especially at low levels of expression ( Figure 1—figure supplement 1E ) .\",\n \"Another approach is to account for TF expression by sorting based on the GFP reporter signal divided by aTF expression ( Staller et al . , 2018; Staller et al . , 2021 ) ; this introduces substantial variability , because activation has a non-linear relationship with aTF expression ( Figure 1—figure supplement 1E ) , and because this ratio can be systematically biased by features that affect aTF expression ( Figure 1—figure supplement 1F ) .\",\n \"This may be why it has been concluded that acidic residues are not necessary for Gcn4 activation , whereas using a similar reporter system we found definitive evidence to the contrary .\",\n \"By this approach , we found that all ADs in S . cerevisiae employ a shared acidic and hydrophobic basis and trained a neural network termed PADDLE to predict sequences that activate on this basis .\",\n \"PADDLE predicted ADs in human TFs with 92% accuracy , indicative of broad conservation of the acidic activation mechanism , and identified hundreds of new ADs in human TFs and virus proteins .\",\n \"PADDLE could further be used to interpret the functional impact of cancer-associated and naturally occurring mutations in human ADs .\",\n \"Predicted ADs from multiple species and instructions for running PADDLE are available online at paddle . stanford . edu .\",\n \"During our development of PADDLE , a predictor of yeast activation named 'ADpred'’ was published ( Erijman et al . , 2020b ) .\",\n \"ADpred , a shallow neural network trained on random protein sequences , achieved a prediction accuracy of 93% on random sequences .\",\n \"However , its accuracy on wild-type sequences was much lower: only 33% of ADs predicted by ADpred activated in our experiments , and 29% of the ADs we identified were not predicted by ADpred , only 3 . 1-fold and 2 . 3-fold better than random predictions , respectively ( alternatively , tile-wise AUPRC of 0 . 294; Figure 3—figure supplement 1J–K , Materials and methods ) .\",\n \"This discrepancy may be because the random sequences on which ADpred was trained do not sufficiently represent the vastly larger space of actual protein sequences; for example , random sequences rarely form significant secondary structure .\",\n \"This failure to generalize highlights the importance of matching neural network training data to the prediction task and underscores the advantage of experiments on wild-type and related mutant sequences .\",\n \"PADDLE predictions were also crucial for generating , refining , and testing hypotheses for activation mechanisms .\",\n \"For example , we unexpectedly discovered that hydrophobic residues could inhibit activation by noticing that sub-domains of certain non-activating regions were predicted to activate on their own .\",\n \"More generally , predictions on AD subsequences identified the core domains responsible for activation at single-amino acid resolution , focusing subsequent experiments on these domains .\",\n \"By combining a high-throughput assay to measure and a machine learning algorithm to predict activity of protein sequences , we have demonstrated a design-build-test-learn cycle that could be applied to accelerate discovery of protein function in other areas of research .\",\n \"This approach yielded the most detailed view to date of the principles governing acidic AD sequence , which we summarize as follows .\",\n \"Activation arises from an abundance of acidic and bulky hydrophobic residues , especially Asp , Trp , Phe , and Leu .\",\n \"The most potent activators cluster these residues densely , forming short core ADs .\",\n \"In most ADs , and especially with high hydrophobic content or abundant Phe and Trp residues , this clustering is sufficient to activate in a composition-driven manner , regardless of sequence or secondary structure .\",\n \"ADs with lower hydrophobic content , particularly those with many Leu residues , instead must position their hydrophobic residues along one face of an alpha helix to activate .\",\n \"Overall , the loose constraints on sequence explain the remarkable diversity and lack of evolutionary conservation of ADs .\",\n \"The unusual plasticity in AD sequence has several advantages .\",\n \"It facilitates spontaneous evolution of new ADs , since an appreciable proportion of even random sequences can activate .\",\n \"This flexibility would allow organisms to develop new transcriptional circuits responsive to their unique needs .\",\n \"Furthermore , in contrast to classical interactions in which structure and function are dependent on a few key residues , activation strength can be adjusted gradually by mutation .\",\n \"Indeed , the observation that nearly all core ADs could increase their strength with simple mutations demonstrates that proper regulation requires precisely adjusted rather than maximized activation .\",\n \"Similarly , AD sequences can preserve their activity while evolving other features , such as in the Gal4 C-terminal AD , whose sequence is specifically bound and inhibited by Gal80 in the absence of galactose ( Johnston et al . , 1987; Ma and Ptashne , 1987 ) .\",\n \"With the use of mRNA display , we found that 73% of ADs bound the Med15 subunit of Mediator and that binding strength was strongly predictive of activation for thousands of wild-type and mutant protein sequences .\",\n \"In contrast , a five-protein TFIID subcomplex bound only 18% of ADs , all of which also bound Mediator .\",\n \"We conclude that Mediator recruitment by TFs is a key driver for gene activation , and Mediator recruitment is largely determined by affinity of ADs for the Med15 subunit .\",\n \"These findings are in keeping with reports that Mediator recruitment occurs independently of other coactivators and can stimulate the recruitment of other coactivators ( Ansari and Morse , 2013; Ansari et al . , 2014; Govind et al . , 2005 ) .\",\n \"How is Med15 , through its four activator-binding domains ( ABDs ) , able to interact with such a large diversity of AD sequences ?\",\n \"Despite the importance of hydrophobic interactions , our structural modeling showed that neither the Med15 KIX domain nor ABD1 provided enough shape complementarity to bury the hydrophobic residues of cADs .\",\n \"Consistent with this , most cADs did not need to form an alpha helix to activate .\",\n \"The lack of shape constraint has two important consequences .\",\n \"First , all modeled cADs bound each ABD in many distinct , equally favored poses .\",\n \"We therefore suggest that fuzzy binding to Med15 , previously demonstrated for the Gcn4 AD , is employed by all ADs .\",\n \"Second , neither ABD showed favored binding positions of Leu versus Phe residues of the cAD , suggesting that the shapes of those residues were unimportant .\",\n \"Thus , the loose constraints on binding shape explain the loose constraints on AD sequence .\",\n \"Instead , each ABD can be approximated as a hydrophobic surface , which engages hydrophobic AD residues through simple hydrophobic forces , flanked by basic residues , which form salt bridges and cation-pi interactions with acidic and aromatic AD residues .\",\n \"Clusters of hydrophobic residues adjacent to ADs inhibit activation , presumably because they compete for interaction with hydrophobic AD residues .\",\n \"Many experiments have shown that transcriptional initiation involves a cycle in which ( 1 ) TFs recruit Mediator to enhancers or upstream activating sites , ( 2 ) Mediator contacts the promoter and orchestrates a series of events starting with chromatin rearrangement , followed by pre-initiation complex formation and finally RNA polymerase II recruitment , and ( 3 ) Mediator facilitates phosphorylation of the polymerase C-terminal domain , which releases polymerase into transcriptional elongation and triggers dissociation of Mediator ( Jeronimo and Robert , 2014; Knoll et al . , 2018; Robinson et al . , 2016; Whyte et al . , 2013; Wong et al . , 2014 ) .\",\n \"Mediator release is apparently necessary because stably recruiting Mediator by fusing individual subunits to a DNA-binding domain failed to activate in nearly all subunits tested ( Wang et al . , 2010 ) .\",\n \"The cycling of Mediator complexes increases the responsiveness of transcriptional regulation in at least two ways .\",\n \"First , it requires certain steps to be repeated in order to continue initiating transcription , preventing the system from locking into an activated state .\",\n \"Second , it frees Mediator complexes to participate in regulation of other genes , which all depend upon and compete for a relatively limited supply of Mediator ( Flanagan et al . , 1991; Gill and Ptashne , 1988 ) .\",\n \"Thus , TFs must recruit Mediator in a manner that is specific and high-affinity but still dynamic .\",\n \"Our kinetic experiments showed that Mediator binds Gcn4-DNA complexes with high affinity but that excess Gcn4 does not impede this interaction , showing that interacting Mediator and Gcn4 molecules can exchange rapidly .\",\n \"If a bimolecular interaction occurs through a single site , a competing molecule cannot bind until the first molecule dissociates which , in a case of high affinity interaction , is a slow process .\",\n \"Our observations are indicative of facilitated exchange , arising from the multiple sites and fuzzy nature of the Gcn4-Mediator interaction ( Figure 7E ) .\",\n \"Fuzzy interactions allow for rapid association and dissociation , because a high proportion of random encounters lead to weak but productive binding ( Ferreira et al . , 2005; Sugase et al . , 2007; Tompa and Fuxreiter , 2008 ) .\",\n \"One among multiple binding sites will frequently be available for a second molecule , which can invade and facilitate release of the first .\",\n \"Similarly , accelerated dissociation or exchange of the E . coli sequence-flexible DNA-binding protein Fis in the presence of competitor proteins was proposed to occur by transitioning through a destabilizing Fis-DNA-Fis ternary complex ( Graham et al . , 2011; Kamar et al . , 2017 ) .\",\n \"These advantageous kinetics provide a rationale for the prevalence of multivalent fuzzy interactions among transcription proteins .\",\n \"ADs have primarily been characterized in TFs where they serve to recruit coactivators such as Mediator , TFIID , SAGA , SWI/SNF , and NuA4 ( Hahn and Young , 2011; Mitchell and Tjian , 1989 ) .\",\n \"Our finding that functional ADs are also present in all coactivator complexes suggests that ADs have broader roles .\",\n \"First , coactivator ADs could interact with ABDs within the same complex , limiting their weak or non-specific binding to TFs .\",\n \"These auto-inhibited coactivators could still bind to ADs clustered on DNA through facilitated exchange .\",\n \"For example , we found that two cADs activated more efficiently and two DNA-bound Gcn4 dimers recruited Mediator more efficiently in tandem than individually .\",\n \"Second , coactivator ADs could amplify activation by mediating interactions between coactivators ( Ansari and Morse , 2013; Liu and Myers , 2015 ) .\",\n \"For example , upon Mediator binding to TFs , the two ADs on Med2 would be liberated from auto-inhibition and could recruit other coactivators .\",\n \"Alternatively , Mediator and other coactivators could cluster in the nucleoplasm through AD-ABD cross interactions and bind together to promoters .\",\n \"Either mode of interaction would explain why recruitment of the Med2/Med3/Med15 subcomplex suffices to recruit SWI/SNF to the CHA1 promoter and suffices for activation of the ARG1 gene , while deletion of both Med2 ADs impairs transcription of galactose-inducible genes ( Ansari et al . , 2014; Liu and Myers , 2015; Zhang et al . , 2004 ) .\",\n \"Phase separation of Mediator and TFs has recently been demonstrated in vitro and is proposed to underlie enhancer-promoter clustering and transcriptional activation in vivo ( Boija et al . , 2018; Chong et al . , 2018; Sabari et al . , 2018; Shrinivas et al . , 2019 ) .\",\n \"Phase separation may occur when a protein makes dynamic self-interactions through two or more binding domains , ( Banani et al . , 2016 ) .\",\n \"Interactions between the four Med15 ABDs and the two Med2 ADs make possible phase separation of the Mediator tail subcomplex; other coactivators , with both ADs and ABDs , might phase separate as well .\",\n \"The functional role of such phase separation , if it occurs , is unclear .\",\n \"A parsimonious interpretation suggests that phase separation of coactivators and TFs is simply a macroscopic consequence of the high valence and dynamism of coactivator-TF interactions , features which may serve an altogether different purpose .\",\n \"ADs have long been enigmatic , due to the apparent mismatch of their structure and function: AD sequences are abundant among random polypeptides , and yet ADs bind their targets with high specificity; AD peptides are apparently disordered , and yet they bind with high affinity .\",\n \"We now understand that these structural features of ADs are ideally suited to their functions—they render AD-target interaction dynamic , through rapid yet weak binding to single sites and strong binding yet rapid displacement from multiple sites .\",\n \"Dynamic AD-target interaction may serve various purposes , such as a rapid response to changing conditions , and the recruitment of multiple AD-bearing proteins to a single target .\",\n \"For example , Mediator bound to RNA polymerase II at a promoter might interact transiently with TFIID , SAGA , SWI/SNF complex and other proteins during transcription .\",\n \"This mechanism may be employed with other unstructured sequences and high valence targets in other cellular processes .\"\n ],\n [\n \"Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact , Roger Kornberg ( kornberg@stanford . edu ) .\",\n \"Raw sequencing data for in vivo activation and in vitro binding assays have been deposited in the Gene Expression Omnibus ( GEO ) database , accession number GSE173156 .\",\n \"The S . cerevisiae strain BY4711 ( ATCC 200873 ) with genotype MATalpha trp1delta63 was used for activation experiments .\",\n \"Standard cell growth conditions were 30°C incubation in YPD medium ( 1% w/v yeast extract , 2% w/v peptone , 2% w/v dextrose ) or synthetic complete ( SC ) medium ( Dunham , 2015 ) .\",\n \"Growth in SC medium lacking tryptophan ( SC-Trp ) was used to select for transformants with the pRS414 plasmid backbone .\",\n \"Selection for stable transformants harboring the natMX6 marker gene was performed by supplementing media with 100 µg/mL nourseothricin ( NAT ) .\",\n \"Selection for stable transformants harboring the KanMX marker gene was performed by supplementing media with 200 µg/mL geneticin ( G418 ) .\",\n \"In activation assays , overnight cultures were diluted to OD600 = 0 . 4 in SC-Trp-NAT-G418 with 10 nM beta-estradiol and grown for 4 hr at 30°C before fluorescence-activated cell sorting ( FACS ) .\",\n \"HEK293T cells ( ATCC CRL-3216 ) were cultured in DMEM with high glucose and GlutaMAX ( Thermo 10566016 ) supplemented with 10% fetal bovine serum ( Millipore TMS-013-B ) at 37°C in 5% CO2 .\",\n \"Yeast and human cells were used directly from ATCC , which authenticates all cell lines using STR profiling and tests for mycoplasma contamination .\",\n \"The artificial TF ( aTF ) expression cassette , which contains a KanMX gene and an ACT1 promoter driving expression of an aTF ( comprising an mCherry tag , mouse Zif269 DBD , and estrogen binding domain ) with a multiple cloning site for insertion of protein fragments of interest , was derived from pMVS142_pACT1_mCherry_Zif268_EBD_MCS_KAN , which was a gift from Barak Cohen ( Addgene plasmid # 99049 ) ( Staller et al . , 2018 ) .\",\n \"This cassette was cloned into the pRS414 backbone , which contains a TRP1 marker ( Sikorski and Hieter , 1989 ) .\",\n \"pRS414 plasmid was digested with EcoRI-HF and SacI , the aTF cassette was PCR-amplified using Phusion polymerase ( NEB ) , and DNA fragments were assembled through Gibson assembly and transformed into XL10-Gold Ultracompetent Cells ( Agilent ) .\",\n \"The final construct , called pRS414-TFchassis , was confirmed by Sanger sequencing .\",\n \"Oligo pools encoding all the protein fragments to be tested in the screen were synthesized at 12K scale ( Twist Bioscience ) .\",\n \"Each oligo was 200 nt in length and contained a 20-nt 5’ constant region and a 21-nt 3’ constant region which were used for PCR amplification .\",\n \"Different sub-libraries had different 3’ constant regions , which allowed specific amplification of the sub-library from the oligo pool .\",\n \"To enable cloning of the oligo pool into pRS414-TFchassis , 20 bp Gibson homology arms matching pRS414-TFchassis were added using PCR .\",\n \"To amplify each sub-library , 2 . 5 ng of oligo template was PCR-amplified for nine cycles using Phusion polymerase , followed by column cleanup .\",\n \"pRS414-TFchassis was digested with NheI-HF and AscI , the amplified oligo pool was cloned in by Gibson assembly , and the product was transformed in two replicates , each into 100 µL of Agilent XL10-Gold cells with 5 µL of Gibson mix .\",\n \"Transformed cells were plated on lysogeny broth ( LB ) plates with Carbenicillin ( Carb ) to estimate efficiency and also grown in 50 mL LB Carb overnight at 37°C for next-day plasmid extraction using Qiagen Midiprep kit .\",\n \"To generate positive control plasmids for the activation screen , IDT gBlocks containing genes expressing 53-aa regions from VP16 , Gcn4 , and Pho4 ADs were also cloned in the same manner as above .\",\n \"The uncloned pRS414-TFchassis , which lacks an introduced insert , was used as a negative control .\",\n \"The GFP reporter construct pMVS102_P3-GFP-NATMX6 was a gift from Barak Cohen ( Addgene plasmid # 99048 ) ( Staller et al . , 2018 ) .\",\n \"To construct a yeast GFP reporter cell line , a cassette from pMVS102_P3-GFP-NATMX6 comprising a natMX6 marker and a P3 promoter ( a modified GAL1 promoter with six copies of the Zif268 binding site ) that drives expression of a fast maturing EGFP variant ( McIsaac et al . , 2013 ) , was PCR-amplified using Phusion polymerase .\",\n \"PCR primers were designed to add 40 bp overhangs with homology to the dubious ORF YBR032W .\",\n \"The PCR product was transformed into yeast using a high-efficiency method ( Gietz et al . , 1992 ) .\",\n \"After overnight outgrowth at 30°C , transformed cells were plated onto YPD-NAT , colonies were picked , and integration into YBR032W was confirmed by PCR amplification of the integrant and Sanger sequencing .\",\n \"The in vivo activation screen requires a yeast library in which each cell contains one pRS414-TFchassis plasmid with a unique insert .\",\n \"The plasmid is maintained at single-copy or low levels due to its CEN/ARS origin of replication , so if a cell receives multiple different vectors upon transformation then repeated cell divisions will eventually separate the distinct vectors ( Scanlon et al . , 2009 ) .\",\n \"To ensure generation of a cellular library in which a unique pRS414-TFchassis plasmid was present in each cell , an assay was developed to quantify the fraction of cells initially receiving multiple plasmids and the time required for cells to reach single-copy plasmid levels by repeated cell divisions .\",\n \"First , using a high efficiency transformation protocol ( Gietz et al . , 1992 ) , approximately 50 million yeast cells in a volume of 360 µL were transformed with 2 µg total plasmid DNA , consisting of an equal mixture of aTF plasmids expressing one AD from VP16 , Gcn4 , or Pho4 .\",\n \"Transformed cells were sparsely plated on SC-Trp-NAT plates either immediately or after a 24 hr or 48 hr outgrowth in SC-Trp-NAT liquid media .\",\n \"For each plating condition , colonies were picked for DNA extraction , followed by qPCR to measure the abundance of VP16 , Gcn4 , and Pho4 plasmids .\",\n \"After 0 hr , 24 hr , and 48 hr of growth , 2 ( of 15 ) , 2 ( of 15 ) , and 1 ( of 24 ) colonies had multiple vectors present , respectively .\",\n \"Because this method cannot detect multiple transformation events of the same plasmid sequence in the same cell ( e . g . double VP16 transformation ) , a correction was applied , yielding multiple vector rates of 20% , 20% , and 6% for each outgrowth condition , respectively .\",\n \"The cloned pRS414-TFchassis plasmid library was transformed into the GFP reporter yeast strain using the same method as in the multiple transformant outgrowth assay above but scaled up 2 . 5-fold .\",\n \"To generate clonal control strains , cells were also separately transformed with a positive control plasmid expressing a VP16 , Gcn4 , or Pho4 AD or with a negative control plasmid , the pRS414-TFchassis plasmid without a protein fragment insert .\",\n \"Transformed cells were grown in 200 mL SC-Trp-NAT-G418 for 6 days ( first library ) or 5 days ( second library ) ; cells were diluted down to OD600 = 0 . 05 every day upon reaching saturation .\",\n \"After this period , cells were diluted to OD600 = 0 . 4 in SC-Trp-NAT-G418 containing 10 nM beta-estradiol and grown at 30°C for 4 hr .\",\n \"Cells were then pelleted and washed once in FACS buffer ( PBS , 2 mM EDTA , 0 . 1% BSA; sterile filtered ) , and finally suspended in FACS buffer at roughly 40 million cells/mL .\",\n \"Cells were analyzed and sorted on a BD Influx Cell Sorter at the Stanford Shared FACS Facility .\",\n \"Cells from different sub-libraries were pooled just prior to sorting .\",\n \"The 10–15% largest cells were excluded from sorting , and then an additional filter was applied to limit mCherry signal to within a twofold range .\",\n \"For the remaining cells , the full range of GFP signal was divided into eight evenly spaced bins in log scale; cells were first sorted into the four low bins and then sorted into the four high bins .\",\n \"In total , 2 . 85 million cells and 2 . 65 million cells were sorted for the first and second libraries , respectively .\",\n \"To define the background GFP signal , cells containing the negative control plasmid were also induced with 10 nM beta-estradiol and their distribution across the eight GFP bins defined above was measured .\",\n \"We found that these cells had slightly increased levels of GFP signal relative to untransformed cells , potentially because binding of the aTF alone slightly opens up chromatin around the GFP promoter .\",\n \"Cells expressing just one plasmid containing the VP16 , Gcn4 , or Pho4 AD were also individually analyzed ( by similarly measuring their distribution across same eight GFP bins ) and compared to results obtained for identical sequences in the pooled activation screen .\",\n \"Cells sorted into each of the 8 GFP bins and a sample of total input cells ( nine samples total ) were grown overnight and then pelleted .\",\n \"Pellets were treated with lyticase and then plasmid DNA was extracted using a Qiagen Miniprep kit .\",\n \"The region of the plasmid encoding the variable protein fragment was PCR-amplified for 11–20 cycles ( based on the OD600 of the culture ) using Phusion polymerase with PCR primers that added flanking sequences corresponding to the standard Illumina sequencing primers .\",\n \"After a column cleanup , a second PCR was performed with primers that added sample-specific barcodes and the P5 and P7 sequences for Illumina sequencing .\",\n \"This second PCR product was SPRI size-selected using AMPure XP beads ( Beckman Coulter ) at 1x dilution , and the different samples were pooled for paired-end sequencing on the Illumina NextSeq 550 .\",\n \"Roughly 2 million reads of 2 × 76 bp length were sequenced per sample .\",\n \"The pFN26A BIND plasmid ( Promega ) , which contains a Renilla luciferase gene used for normalization , was used to express domains of interest fused to a GAL4 DBD .\",\n \"Genes expressing 50 randomly chosen putative human ADs predicted by PADDLE were synthesized in a single oPool ( IDT ) with Gibson homology arms , amplified using Phusion polymerase , and cloned into pFN26A BIND ( digested with AsiSI and PmeI ) using Gibson assembly .\",\n \"The product was transformed into Agilent XL10 Gold cells , and then plated onto LB Carb plates .\",\n \"Forty-three colonies were picked , which yielded clones of 25 correct unique ADs .\",\n \"In addition , plasmids expressing a VP16 AD-positive control or three random protein sequences for negative controls were individually cloned in the same manner .\",\n \"HEK293T cells were plated at 24 , 000 cells per well of a 96-well plate .\",\n \"The next day , cells were transfected using Lipofectamine 3000 with 50 ng of cloned pFN26A BIND plasmid and 50 ng of pGL4 . 35 ( Promega ) , which contains a Firefly luciferase reporter gene under the control of a minimal promoter driven by nine upstream GAL4 DNA-binding sites .\",\n \"Each plasmid was transfected into triplicate wells .\",\n \"Seventy-two hours after transfection , luminescence from renilla and firefly luciferases were measured using the Dual-Glo Luciferase assay system ( Promega ) on a BioTek Synergy two plate reader .\",\n \"Measured activity of each predicted AD or control sequence was computed by dividing the Firefly luciferase signal by the Renilla signal and then averaging the triplicate values .\",\n \"Fold-activation was computed by dividing the activity of each predicted AD by the mean activity of the three negative controls , Z-scores were computed based on log-scale fold-activation by standardizing the negative controls to mean of 0 and variance of 1 , and p-values were computed with one tailed Z-tests .\",\n \"An mRNA display library was made for each of sub-library A , sub-library B , and six pooled controls ( three libraries total ) , using the following method .\",\n \"Sub-libraries of wild-type and mutant TF fragments were expressed as mRNA-tagged peptides using mRNA display with in vitro translation as described in Takahashi and Roberts , 2009 .\",\n \"Specifically , sub-library pools were PCR-amplified using primers that added an upstream T7 promoter sequence and a downstream linker and HA-tag but no stop codon .\",\n \"Thus , all encoded peptides were of the form MSGT[variable protein sequence]TSVGGSGSYPYDVPDYA and fused to mRNA at the C-terminus .\",\n \"In addition to the two sub-libraries , similar DNA sequences encoding positive control ADs from VP16 , Gcn4 , and Pho4 and three negative control random protein sequences were synthesized as gBlocks ( IDT ) ( containing an upstream T7 promoter sequence and a downstream linker and HA-tag but no stop codon ) .\",\n \"All subsequent work was performed in RNase-free conditions .\",\n \"mRNA was in vitro transcribed for 4 hr at 37°C using the MEGAscript T7 transcription kit ( Thermo ) and purified with RNeasy mini columns ( Qiagen ) .\",\n \"The PF30P oligo containing a poly-A sequence , spacers , and puromycin was synthesized by IDT and ligated to the 3’ end of the mRNA using Moore-sharp splinted ligation by T4 DNA ligase .\",\n \"The ligation product was desalted by three rounds of concentrating in 0 . 5 mL Amicon Ultracel 30K MWCO columns and diluting in water .\",\n \"The desalted product was run on a denaturing 4% polyacrylamide gel with 7 M urea .\",\n \"The band corresponding to full-length ligated mRNA was excised from the gel and recovered using the Model 422 Electro-Eluter with 3500 MWCO caps ( Bio Rad ) .\",\n \"After concentrating and desalting in the same manner , the ligated and purified mRNA was used in an in vitro translation reaction using the Retic Lysate IVT kit ( Thermo ) for 60 min at 30°C .\",\n \"MgCl2 and KCl were added to final concentrations of 60 mM and 500 mM , respectively , and the sample was incubated for 15 min at room temperature to encourage puromycin fusion .\",\n \"The sample was then diluted in binding buffer ( 20 mM Tris pH 8 . 0 , 100 mM KCl , 1 mM MgCl2 , 0 . 01% Tween-20 , 0 . 1 mg/mL BSA , 0 . 1 mM DTT ) and bound to Anti-HA magnetic beads ( Pierce ) for 30 min at room temperature with mixing .\",\n \"After washing three times for 5 min with binding buffer without BSA , peptides were eluted by incubating in 1x SuperScript II buffer ( 50 mM Tris-HCl pH 8 . 3 , 75 mM KCl , 3 mM MgCl2; Thermo ) with 0 . 01% Tween-20 and 0 . 5 mg/mL synthetic HA peptide ( Thermo ) for 1 hr at room temperature with mixing .\",\n \"The mRNA tag on the fused peptides was reverse transcribed using SuperScript II ( Thermo ) and a primer that contained ( 1 ) a sequence that hybridizes just downstream of variable protein sequence coding region , ( 2 ) a 12-nt random sequence that serves as a unique molecular identifier ( UMI ) , and ( 3 ) the Illumina read two sequencing primer sequence .\",\n \"Enzyme was heat inactivated by adding EDTA to 10 mM and heating to 65°C for 10 min .\",\n \"To remove aggregates , the library was fractionated in 20 mM Tris pH 8 . 0 , 100 mM KCl , 1 mM MgCl2 , 1 mM DTT over a 6 mL 10–30% sucrose gradient by spinning at 60 , 000 rpm in a SW60 rotor for 10 hr .\",\n \"Fractions of 100–200 µL were collected and the abundance of peptide-mRNA-cDNA fusion in each fraction was assayed using qPCR .\",\n \"The 3–4 peak fractions encompassing 80–90% of total input was pooled and used in pull-down experiments or flash frozen in aliquots and stored at −80°C .\",\n \"Biotinylated bait proteins ( 12 µg of Med15 or 8 µg of TFIID subcomplex ) were bound to 50 µL Dynabeads MyOne Streptavidin T1 ( Thermo ) for 15 min at room temperature in binding buffer ( 20 mM Tris pH 8 . 0 , 100 mM KCl , 10 mM MgCl2 , 0 . 01% Tween-20 , 1 mM DTT ) and then washed three times for 5 min in binding buffer .\",\n \"In the third wash , beads were blocked by also including 100 ng/µL salmon sperm DNA ( Thermo ) and 10 ng/µL BSA ( NEB ) in the buffer .\",\n \"mRNA display library input was prepared by mixing sub-library A , sub-library B , and the three positive and three negative controls .\",\n \"This input was then incubated with the bait-coated beads in binding buffer with salmon sperm DNA and BSA at 100 µL total volume for 8 min with mixing .\",\n \"Afterward , the beads were separated on a magnet , briefly washed once with cold binding buffer , and then immediately separated on a magnet again .\",\n \"cDNA from bound fragments was eluted by suspending the beads in 10 mM Tris pH 8 . 0 with proteinase K ( NEB ) and incubated for 30 min at room temperature .\",\n \"Proteinase K was heat inactivated at 95°C for 10 min , beads were separated on a magnet , and the eluate was collected .\",\n \"Total abundance of the library cDNA in the input , bound , and eluted samples was measured by qPCR .\",\n \"To measure the total abundance of the whole library ( without regard to the identity of the library elements ) , qPCR primers were designed to hybridize to the constant sequences flanking the variable region of the library .\",\n \"Additionally , this design enabled the use of different primers to separately track sub-library A and sub-library B , as well as each of the six control sequences individually .\",\n \"The controls were used to provide an estimate of binding efficiency .\",\n \"Binding percentage was calculated by dividing the abundance of each library in the bound sample by the abundance in the input sample .\",\n \"cDNA of the input and bound samples were prepared for sequencing in the same manner as the purified plasmids in the activation assay .\",\n \"Biotinylated 104 bp double-stranded DNA templates were assembled by annealing a 104-nt oligo to a 20-nt oligo with a 5’ biotin modification ( IDT ) and elongating using Klenow Fragment ( 3’-->5’ exo- ) ( NEB ) .\",\n \"The oligos were designed such that the TF DNA binding sites were placed away from the biotinylated end .\",\n \"SPR experiments were performed on the ProteOn XPR36 Surface Plasmon Resonance System using NeutrAvidin-coated NLC Sensor Chips ( Bio Rad ) .\",\n \"Mass of biomolecules bound to the chip surface was measured in arbitrary units ( AU ) approximately every 1 s .\",\n \"All plotted data were normalized to the first ligand channel and the first analyte channel which were used as negative controls in experiments as described below .\",\n \"In SPR experiments with GCN4 , 104 bp DNA templates with 0 , 1 , 2 , 4 , or 6 Gcn4 motifs were bound to the chip along the ligand channels , except for the first ligand channel , which was left without DNA for normalization .\",\n \"Binding buffer for all conditions was 20 mM Tris-HCl ( pH 7 . 5 ) , 100 mM KCl , 10 mM MgCl2 , 2 mM EDTA , 10 mM AmSO4 , 0 . 05% Tween-20 , with 1 mg/mL BSA ( NEB ) .\",\n \"Bound DNA was washed twice with 1 M NaCl .\",\n \"Gcn4 at 7 . 5 nM was then pre-bound to DNA for 480 s along the analyte channels , except for the first analyte channel , which was left without Gcn4 for normalization .\",\n \"After Gcn4 binding , Gcn4 at 7 . 5 nM with Mediator at 0 , 1 . 25 , 2 . 5 , 5 , or 10 nM were flowed across analyte channels 2–6 respectively for 300 s , then dissociation was observed for 20 min by flowing in binding buffer .\",\n \"Association , dissociation , and affinity constants of Mediator binding to Gcn4 on DNA were estimated by fitting to a Langmuir kinetic model using the ProteOn Manager Software ( Bio Rad ) .\",\n \"Reported values are the mean and standard deviation across the four different Mediator concentrations .\",\n \"Note that Mediator dissociation rates in conditions with multiple Gcn4 binding sites may be overestimated because it also includes the dissociation of Gcn4 from DNA .\",\n \"For SPR experiments with EcoRI , biotinylated 60 bp DNA templates containing 0 or 1 EcoRI motif were annealed and elongated as described above , and then bound along the ligand channels and washed with 1 M NaCl .\",\n \"Gcn4AD-EcoRI ( E111Q ) was bound along analyte channels at 27 nM monomer in 20 mM Tris-HCl ( pH 7 . 5 ) , 10 mM MgCl2 , 300 mM NaCl , 10 mM bME , 0 . 05% Tween-20 , 1 mg/mL BSA ( higher salt was used to reduce non-specific binding to DNA ) .\",\n \"Minimal dissociation was observed after binding .\",\n \"Then , Mediator at 5 nM with competitor Gcn4 protein ( in molar excess as specified in Figure 7D ) was flowed across the chip for 180 s in 20 mM Tris-HCl ( pH 7 . 5 ) , 10 mM MgCl2 , 100 mM NaCl , 10 mM AmSO4 , 10 mM bME , 0 . 05% Tween-20 , 1 mg/mL BSA .\",\n \"All protein positions reported in the text are 1-indexed and inclusive .\",\n \"The first library of protein sequences was composed of sub-library A , which contained wild-type tiles spanning all S . cerevisiae transcription factors , and sub-library B , which contains mutants of 8 known S . cerevisiae ADs .\",\n \"The second library was composed of sub-library D and sub-library E , which contained multiple collections of wild-type , mutant , and designed sequences .\",\n \"Sequences in sub-libraries A , B , and D were 53 amino acids in length and sequences in sub-library E were 30 amino acids in length .\",\n \"Protein and DNA sequences , as well as activation and pull-down data where available , are listed in Tables S1 and S4 .\",\n \"Libraries contained additional fragments which were included in experiments but are not reported here .\",\n \"In the reverse translation design process , we aimed to optimize our library DNA fragments for compatibility and consistency with our in vitro and in vivo assays , as well as with standard RNA-seq protocols .\",\n \"We also sought to build in redundancy for error-correcting reads .\",\n \"In particular , we used the Python package dnachisel 1 . 4 . 1 ( Zulkower and Rosser , 2020 ) to optimize the following objectives: To allow paired-end sequencing to uniquely identify each element , we additionally enforced an edit distance of 6 among the first 48 bases and last 48 bases of any two sequences in the same sub-library .\",\n \"This was performed in a randomized , brute-force , iterative approach , with each iteration consisting of the following steps: Finally , we verified that all sequences sharing the same sub-library primer ( e . g . all sequences in sub-library A , or all sequences in sub-library B ) had a paired-end edit distance ( sum of edit distance of 5’-most 50 bases and edit distance of 3’-most 50 bases ) of at least 6 .\",\n \"The use of sequencing primers unique to each sub-library enabled us to submit samples for sequencing in multiplexed format and accurately assign reads to the correct sub-library computationally .\",\n \"We further leveraged the edit distance margin built into the library to enable mapping of sequencing reads with a small number of errors .\",\n \"Sequencing read alignment was performed using custom bash script built on top of existing tools and additional custom scripts .\",\n \"It takes as input arguments the UMI length , the sub-library sequencing primer , the edit distance threshold for that sub-library , and raw FASTQ files .\",\n \"For pull-down experiments , unique molecular identifiers ( UMI ) were extracted from reads and appended to the read names using umi_tools 1 . 0 . 0 ( Smith et al . , 2017 ) .\",\n \"cutadapt 1 . 18 was used to discard reads without matching paired-end sub-library sequencing primers and trim the primers in reads with matching primers; the default error tolerance was used ( Martin , 2011 ) .\",\n \"bwa-mem 0 . 7 . 17-r1188 was used to perform a first-pass alignment of reads to the DNA fragment library ( Li , 2013 ) .\",\n \"Imperfectly mapped read pairs ( i . e . those without paired read SAM flags of 99 and 147 ) were re-mapped to the library sequence with minimal edit distance .\",\n \"This was necessary because bwa-mem did not always correctly map paired reads as a pair , a problem most evident in the mutant library with many similar sequences .\",\n \"Pairwise Levenshtein edit distance was computed using the Python package editdistance 0 . 5 . 3 ( https://github . com/roy-ht/editdistance , Tanaka , 2019 ) .\",\n \"Paired reads exceeding the edit distance threshold were discarded using reformat . sh from BBTools 38 . 61 ( https://sourceforge . net/projects/bbmap/ ) .\",\n \"Duplicate reads were identified and deduplicated using utmi_tools 1 . 0 . 0 .\",\n \"Finally , reads mapped to each DNA library fragment sequence were counted .\",\n \"Calculations of GFP signal and fold-activation required the following inputs: First , to make cell counts comparable between sorting bins 1–4 versus bins 5–8 , the number of cells sorted into each GFP bin were normalized based on the number of events assayed in each sort .\",\n \"Second , to estimate the number of cells sorted into each of the 8 GFP bins that expressed a given protein fragment F , the total number of cells sorted into each bin was multiplied by the fraction of sequencing reads for that bin that mapped to F . Any fragment represented by fewer than 10 cells were excluded from further analysis .\",\n \"Third , the average GFP signal of cells expressing F was calculated as a weighted average from the average GFP signal observed in each bin during sorting weighted by the distribution of the number of cells expressing F across the eight bins .\",\n \"Because GFP distributions cells expressing single AD sequences were normally distributed in FACS data when GFP signal ( arbitrary units ) was plotted in log-scale , all averages of GFP distributions were calculated in log-scale ( equivalently , a geometric mean of GFP values ) .\",\n \"The standard deviation of GFP signal of cells expressing F was similarly calculated in log-scale and used as a metric for measurement precision .\",\n \"All averages involving activation reported in the paper are calculated in log-scale ( equivalently , a geometric mean ) .\",\n \"Fourth , the fold-activation of F was calculated as the mean GFP signal of F divided by the mean GFP signal of negative control cells ( see below ) .\",\n \"Z-scores were calculated by standardizing the GFP signal of negative control cells to mean 0 and variance one in log scale and p-values were calculated with one tailed Z-tests .\",\n \"The mean and standard deviation GFP signal of negative control cells was defined in one of two ways .\",\n \"In the experiment with sub-libraries A and B , a sample of estrogen-induced cells containing only the empty pRS414-TFchassis plasmid was assayed under identical conditions and the mean and standard deviation GFP signal was computed from FACS data .\",\n \"In the experiment with sub-libraries D and E , these values were instead determined internally from the 50 random sequence negative controls in the pooled library .\",\n \"Since a small fraction of random sequences activate , outlier sequences were iteratively excluded if they had GFP signal greater than three standard deviations away from the mean , resulting in three sequences excluded from each sub-library .\",\n \"Then mean and standard deviation were computed from the sum of the negative control GFP distributions .\",\n \"In activation assays with sub-libraries D and E , we noticed that many sequencing reads that were mapped with a small number of errors corresponded to fragments with a very large GFP standard deviation .\",\n \"Upon closer examination of sequencing reads of example fragments , the same sequencing errors were consistently found , which likely corresponded to mutations that occurred in vivo that affected function of the protein fragment ( e . g . inactivation of a strong AD by a frameshift mutation ) .\",\n \"We therefore used only reads that aligned perfectly for this experiment .\",\n \"Furthermore , fragments with GFP standard deviation greater than 4 . 5 were excluded from further analysis .\",\n \"The following three metrics were tracked: the number of total unique fragments , number of fragments observed in at least one of the eight GFP bins , and number of fragments that passed filtering .\",\n \"For the activation assay with sub-libraries A and B , this was ( 7637 , 7604 , 7549 ) for sub-library A and ( 2900 , 2833 , 2813 ) for sub-library B . The same metrics for the activation assay with sub-libraries D and E was ( 7733 , 7716 , 7345 ) for sub-library D and ( 4261 , 4261 , 4252 ) for sub-library E . These numbers include additional fragments in the peptide libraries that are not reported in this paper .\",\n \"To define ADs across all TFs , at each protein position a mean Z-score was calculated from the Z-scores of all fragments overlapping that position .\",\n \"ADs were defined for all intervals of the protein that had a mean positional Z-score ( MPZ ) greater than 3 . 5 .\",\n \"Starting with intervals that had the highest maximal MPZ score , the start and stop positions of the AD were defined by the full width half maximum; that is , the positions in the protein at which the MPZ fell below half of the maximal MPZ value in that interval .\",\n \"If the AD defined in this manner includes a new region with an MPZ score higher than the original maximal value , this indicated that the MPZ peak was only small relative to nearby peaks , so the AD was then excluded .\",\n \"We compared our ADs with the Induction Dynamics gene Expression Atlas ( IDEA ) ( Hackett et al . , 2020 ) , which induced expression of every TF individually and measured ‘v_inter’ parameters that describe the resulting fold-change in the expression of all target genes .\",\n \"We said that a TF activated a gene if the v_inter > 1 , and for each TF that regulated at least five genes , we calculated the proportion of regulated genes that were activated .\",\n \"We then compared these proportions for TFs that had a strong AD ( overlapping a fragment with Z-score at least 6 ) versus TFs without ADs .\",\n \"Significance was calculated using a Kolmogorov–Smirnov test .\",\n \"Hydrophobicity measurements were calculated using the Wimley-White interfacial scale , which measures whole-residue ∆G for transfer from water to a water-octanol interface ( Wimley and White , 1996 ) .\",\n \"Because the goal was to quantify the total content of hydrophobic residues rather than the net hydrophobicity/hydrophilicity , hydrophobic content of a peptide is computed as the sum of the scores of only the hydrophobic residues ( i . e . those with negative ∆G ) .\",\n \"Net charge was computed by adding the number of Arg and Lys residues and subtracting the number of Asp and Glu residues .\",\n \"Motif finding using DREME was done through the web portal ( http://meme-suite . org/tools/dreme ) using default settings and shuffled input sequences as control ( Bailey , 2011 ) .\",\n \"Searching for 9aaTAD sequences was done by matching to the ‘most stringent pattern’ [MDENQSTYG]{KRHCGP}[ILVFWM]{KRHCGP}{CGP}{KRHCGP}[ILVFWM][ILVFWMAY]{KRHC} listed at the prediction portal https://www . med . muni . cz/9aaTAD/index . php ( Piskacek et al . , 2007 ) .\",\n \"ADpred was installed locally from https://github . com/FredHutch/adpred , Erijman , 2020a and predictions were computed for all 30-aa tiles on all yeast TFs ( Erijman et al . , 2020b ) .\",\n \"ADs predicted by ADpred were defined using the authors’ criteria , namely sites with five or more contiguous residues with a score >= 0 . 8 .\",\n \"An ADpred-predicted region was considered correct if it overlapped by at least five aa with a significantly activating tile ( p<0 . 0001 ) .\",\n \"With this approach , ADpred achieved a precision and recall of 0 . 366 and 0 . 692 in predicting experimental ADs , respectively .\",\n \"When ADpred-predicted regions were randomly shuffled 100 times to different proteins and positions , the average precision and recall of these shuffled predictions were 0 . 120 and 0 . 295 respectively .\",\n \"Alternatively , to compare ADpred predictions directly to our experimental measurements of activation , a representative ADpred score for each 53-aa TF tile in our library was calculated by taking the 80th percentile of ADpred scores of 30-aa tiles contained within the 53-aa tile .\",\n \"These scores were used for calculating tile-wise precision-recall curves for ADpred predictions ( Figure 3—figure supplement 1K ) .\",\n \"The fractional pull-down of each fragment F was calculated as ( the fraction of the total input library that bound to beads , measured from qPCR ) x ( the fraction of UMI-deduplicated sequencing reads in the bound sample matching F ) / ( the fraction of UMI-deduplicated sequencing reads in the input sample matching F ) .\",\n \"Fragments with fewer than 30 reads in the input sample or fewer than five reads in the bound sample were excluded from downstream analysis .\",\n \"Baseline binding was defined using the 50 random protein control sequences in each library , except the sequences that were outliers in the activation assay were excluded .\",\n \"Specifically , random control outliers were iteratively removed if their activation was more than three standard deviations away from the mean ( in log scale ) , resulting in the exclusion of random control #2 , 12 , 16 , 38 from sub-library A and #17 , 18 , 21 , 22 , 31 , 43 from sub-library B . Binding enrichment of each fragment was then computed as the fractional pull-down of the fragment divided by the mean fractional pull-down ( in log scale ) of the random controls .\",\n \"Because fractional pull-down and enrichment were normally distributed in log-scale , all averages involving pull-down or enrichment reported in the paper are calculated in log-scale ( equivalently , a geometric mean ) .\",\n \"Z-scores were computed by standardizing the binding enrichment of random controls to mean 0 and variance one in log-scale and P-values were calculated with one tailed Z-tests .\",\n \"Triplicate measurements for Med15 pull-downs and duplicate measurements for TFIID subcomplex pull-downs were combined by averaging the binding enrichment values .\",\n \"A small number of highly positively-charged fragments bound in Med15 pull-downs when the mRNA display library was purified on a sucrose gradient but did not bind when the sucrose gradient was omitted .\",\n \"Suspecting that this binding was artifactual , potentially through non-specific binding to mRNA tags , we excluded from analysis 92 fragments that bound with Z-score between −2 . 5 and 1 without the sucrose gradient and had a Z-score at least 3 . 5 higher with the sucrose gradient than without .\",\n \"Med15-binding domains were annotated in the same manner as ADs using a mean positional Z-score , but with a cutoff of 2 . 23 .\",\n \"An AD was considered to bind Med15 or TFIID subcomplex if it overlapped a fragment that bound significantly ( p<0 . 001 ) by at least 26 aa .\",\n \"A Med15-binding domain was considered to activate if it overlapped an AD .\",\n \"Neural network models were trained to predict activation Z-scores .\",\n \"Approximately 15% of sequences from sub-library A and B were randomly held-out as a test set , and the remaining library elements were split into 10 parts for training and cross-validation .\",\n \"Because adjacent TF tiles have significant overlap in sequence , all tiles from each protein were placed into the same train or test split .\",\n \"From sub-library B , all mutants and homologs of the Pdr1 AD and all scramble mutants were held-out in the test set and the remaining sequences were distributed evenly across the training splits .\",\n \"Models were developed in TensorFlow 2 . 2 ( Abadi et al . , 2016 ) and trained using CPUs on Stanford’s Sherlock computing cluster .\",\n \"All models used mean squared error as the loss function .\",\n \"Sequence encodings and activation Z-scores were standardized to mean 0 variance one across the training dataset for more efficient learning .\",\n \"For each architecture , 10 models were trained on each subset of nine training splits and validated on the 10th split .\",\n \"The best model architecture was chosen by the mean validation score across the 10 splits on sub-library A sequences only .\",\n \"For final evaluation on the test dataset , the mean prediction across all 10 models was used .\",\n \"Protein fragment sequences were each encoded as a single 20-element vector giving the proportions of each of the 20 amino acids .\",\n \"The best architecture , as determined by a grid search over hyperparameters , was a neural with three fully connected layers of width 40 with Swish activation ( Ramachandran et al . , 2017 ) , batch normalization , L2 weight penalty of 0 . 01 , and dropout rate of 0 . 4 .\",\n \"The model was trained with an initial learning rate of 0 . 0001 using the Adam optimizer for up to 300 epochs with two scheduling callbacks: reduction of the learning rate by fivefold if training loss did not improve for 20 epochs , and early stopping if no improvement on the validation loss ( averaged in a 10-epoch window ) was observed for 75 epochs , upon which the best model from previous epochs was saved .\",\n \"One-hot encodings were used to encode each 53-aa protein fragment sequence as a 53-by-20 matrix .\",\n \"Secondary structure predictions were generated using PSIPRED v . 4 . 01 without PSI-BLAST ( Jones , 1999 ) , and a 53-by-2 matrix of the alpha helix and coil predictions were also used as input .\",\n \"Disorder predictions were generated using IUPred2A ( Mészáros et al . , 2018 ) , and a 53-by-2 matrix of the disorder predictions in ‘short’ and ‘long’ mode were also used as input .\",\n \"In total , each 53-aa protein fragment was encoded as a 53-by-24 matrix .\",\n \"The best model architecture , as determined by a grid search over hyperparameters , was a convolutional neural network with nine convolutional layers with kernel size 10 and channel width 30 ( i . e . number of filters ) followed by max-pooling along the sequence-length dimension and two fully-connected layers of width 20 , with Swish activation .\",\n \"For regularization , the L2 weight penalty was 0 . 001 , dropout rate was 0 . 1 , and batch normalization was used .\",\n \"The model was trained with an initial learning rate of 0 . 001 in the same manner as the amino acid composition-based network , except for up to 500 epochs .\",\n \"After finding this optimal architecture , the test dataset was distributed across the 10 training datasets and 10 new models were trained , together comprising PADDLE .\",\n \"The mean value of these 10 predictors was used as the final prediction .\",\n \"This model was used for all predictions on yeast TFs and nuclear proteins , human TFs , and virus transcriptional regulators .\",\n \"Because secondary structure prediction was by far the slowest aspect of running PADDLE predictions , we trained a version of PADDLE called PADDLE-noSS which only used the one-hot sequence encoding for input , no secondary structure or disorder predictions .\",\n \"PADDLE-noSS was nearly as accurate as PADDLE ( R2 scores of 0 . 76 , 0 . 52 , and 0 . 89 for TF tiles , scramble mutants , and Pdr1 mutants and homologs respectively; compare to Figure 3—figure supplement 1B ) and was used for predictions of core ADs and mutant sequences in Figure 3E–G , Figure 3—figure supplement 1E–I , and Figure 4—figure supplement 1F .\",\n \"Human TFs were defined by Gene Ontology term GO:003700 ( DNA-binding transcription factor activity ) .\",\n \"Virus transcriptional regulators ( vTRs ) were taken from Figure 1—source data 1 from Liu et al . , 2020 .\",\n \"Within each human TF or vTR , predictions on 53-aa tiles at 1-aa resolution were generated using PADDLE and smoothed with a 9-aa moving average .\",\n \"High-strength ADs were defined as the union of at least 5 consecutive 53-aa tiles with a predicted Z-score greater than 6 .\",\n \"ADs defined in this way that overlapped by more than 26 aa were combined .\",\n \"Medium-strength ADs were defined similarly with a Z-score cutoff of 4 and were required to not overlap with high-strength ADs by more than 26 aa .\",\n \"All protein positions reported in the text are 1-indexed and inclusive; all annotations reported in supplemental tables are 0-indexed and inclusive of the start position but exclusive of the stop position .\",\n \"For testing in a luciferase assay , 50 predicted high-strength ADs from human TFs were chosen at random and the tile with the strongest ( smoothed ) predicted Z-score from each AD was used .\",\n \"Yeast nuclear proteins were defined by Gene Ontology term GO:0005634 ( Nucleus ) .\",\n \"Within each protein , predictions on 53-aa tiles at 1-aa resolution were generated using PADDLE and smoothed with a 9-aa moving average .\",\n \"All predicted Z-score local maxima higher than 3 . 09 ( p<0 . 001 ) were considered for experimental testing .\",\n \"The 53-aa tiles corresponding to the local maxima were included , starting from those with the largest predicted Z-score , unless they overlapped a previously-included tile by at least 15 aa .\",\n \"Analysis of experimentally-confirmed ADs focused on proteins that had experimental or high-throughput evidence for nuclear localization .\",\n \"Proteins involved in transcriptional regulation were defined by GO:0006355 ( regulation of transcription , DNA-templated ) .\",\n \"To calculate enrichment of activating tiles in coactivator proteins , we generated a null model in which ‘activating’ tiles , equal in number to actual activating tiles discovered in our screen , were randomly selected from among non-TF nuclear proteins .\",\n \"We then counted how many of these randomly-selected tiles were in coactivator proteins , and then repeated this sampling 10 , 000 times to generate a null distribution .\",\n \"The actual number of tiles , 42 , is 2 . 9-fold more than the mean of the null; p<10−12 by one tailed Z-test .\",\n \"Predicted disorder for each tile was computed from D2P2 annotations as the fraction of the tile’s residues that were annotated as disordered among all 8 of the predictors listed in D2P2 .\",\n \"To define core ADs , predictions of sequences shorter than 53 aa were done using PADDLE-noSS by embedding the shorter region in 100 different neutral sequences composed of AGSTNQV residues and averaging the predicted values .\",\n \"Predictions of all tiles 5–30 aa in length within every AD were computed in this manner , and used to define core ADs for Figure 3F–G , Figure 7A , and Figure 7—figure supplement 1B .\",\n \"Specifically , core ADs were defined as the shortest tiles with predicted Z-score greater than 3 . 09 ( p<0 . 001 ) that do not overlap another shorter core AD .\",\n \"In Figure 3—figure supplement 1I , the 20-aa region within each AD with the highest predicted Z-score was included if it had Z-score greater than 3 . 09 , and the effects of single AA mutants in these 20-aa core ADs were predicted in the same manner using PADDLE-noSS .\",\n \"Code for running PADDLE and PADDLE-noSS is available at https://github . com/asanborn/PADDLE ( Sanborn , 2021; copy archived at swh:1:rev:17aa36985e474d8593b99d593f7cc5e7c3e205a0 ) .\",\n \"Pre-generated PADDLE predictions from multiple species are available for download at http://paddle . stanford . edu .\",\n \"To generate structural models of binding , we took 28 core ADs peptides 13 residues in length and docked them to the ABDs using Rosetta3 ( Leaver-Fay et al . , 2011 ) , specifically its FlexPepDock ab initio pipeline ( Raveh et al . , 2011 ) .\",\n \"Core ADs were used as the peptide , and activator-binding domain as the receptor .\",\n \"Peptides were initialized in an extended conformation using Rosetta’s BuildPeptide application and manually placed near the binding pocket .\",\n \"For the KIX domain , NMR chemical shift data was used to aid in placement ( Thakur et al . , 2008 ) .\",\n \"For ABD1 , the AD of Gcn4 was used to aid in placement .\",\n \"For all models of ABD1 , the first conformer from PDB structure 2LPB was used ( Brzovic et al . , 2011 ) .\",\n \"For each AD-ABD pair , a set of 50 , 000 structural models were generated and ranked by FlexPepDock’s reweighted_sc score .\",\n \"Binding poses were generated by clustering the top 500 ranked models using Rosetta’s cluster application with the default clustering distance threshold of 2 Å .\",\n \"The best-scoring model from each cluster was designated as the cluster representative , and the cluster representatives that were also among the 10 best-scoring models were used for further analysis .\",\n \"Additionally , larger runs with 500 , 000 structural models and clustering on the top 5000 models were run for six individual ADs for the KIX ABD .\",\n \"The full list of ABDs and ADs used is available in Figure 6—source data\",\n \"1 . PDB files of 10 best-scoring models from different binding poses and the score , cluster number , and RMSD to best model for the 500 best-scoring models are available in Figure 6—source data\",\n \"2 . Due to the large number of peptides that needed to be docked , the FlexPepDock ab-initio pipeline was modified to run in a cluster setting at a large scale and used approximately 100 , 000 CPU hours on the Stanford Sherlock academic cluster .\",\n \"Secondary structure was calculated using DSSP ( Kabsch and Sander , 1983; Touw et al . , 2015 ) and annotations of G , H , or I were considered alpha-helical .\",\n \"Interaction types were computed with GetContacts ( Venkatakrishnan et al . , 2019 ) and per-AA accessible surface areas were computed with FreeSASA ( Mitternacht , 2016 ) .\",\n \"For each final docked structural model , buried peptide surface area was taken by subtracting the accessible surface area of the peptide in complex with the receptor from the accessible surface area of the peptide on its own .\",\n \"For Figure 6E , marked positions correspond to atom CG of Asp , atom CG of Leu , and the midpoint of the CG and CZ atoms in Phe ( the center of the aromatic ring ) .\",\n \"Residues are shown from all the 10 best models from different clusters for all core AD structures if the marked position is within 3 Angstroms of the KIX domain or ABD1 structure .\",\n \"Adapted FlexPepDock code , and wrapper code for GetContacts and FreeSASA are available at https://github . com/drorlab/med15 .\"\n ]\n]","isConversational":false},"abstract":{"kind":"list like","value":["Gene activator proteins comprise distinct DNA-binding and transcriptional activation domains ( ADs ) .","Because few ADs have been described , we tested domains tiling all yeast transcription factors for activation in vivo and identified 150 ADs .","By mRNA display , we showed that 73% of ADs bound the Med15 subunit of Mediator , and that binding strength was correlated with activation .","AD-Mediator interaction in vitro was unaffected by a large excess of free activator protein , pointing to a dynamic mechanism of interaction .","Structural modeling showed that ADs interact with Med15 without shape complementarity ( ‘fuzzy’ binding ) .","ADs shared no sequence motifs , but mutagenesis revealed biochemical and structural constraints .","Finally , a neural network trained on AD sequences accurately predicted ADs in human proteins and in other yeast proteins , including chromosomal proteins and chromatin remodeling complexes .","These findings solve the longstanding enigma of AD structure and function and provide a rationale for their role in biology ."],"string":"[\n \"Gene activator proteins comprise distinct DNA-binding and transcriptional activation domains ( ADs ) .\",\n \"Because few ADs have been described , we tested domains tiling all yeast transcription factors for activation in vivo and identified 150 ADs .\",\n \"By mRNA display , we showed that 73% of ADs bound the Med15 subunit of Mediator , and that binding strength was correlated with activation .\",\n \"AD-Mediator interaction in vitro was unaffected by a large excess of free activator protein , pointing to a dynamic mechanism of interaction .\",\n \"Structural modeling showed that ADs interact with Med15 without shape complementarity ( ‘fuzzy’ binding ) .\",\n \"ADs shared no sequence motifs , but mutagenesis revealed biochemical and structural constraints .\",\n \"Finally , a neural network trained on AD sequences accurately predicted ADs in human proteins and in other yeast proteins , including chromosomal proteins and chromatin remodeling complexes .\",\n \"These findings solve the longstanding enigma of AD structure and function and provide a rationale for their role in biology .\"\n]","isConversational":false},"summary":{"kind":"list like","value":["Cells adapt and respond to changes by regulating the activity of their genes .","To turn genes on or off , they use a family of proteins called transcription factors .","Transcription factors influence specific but overlapping groups of genes , so that each gene is controlled by several transcription factors that act together like a dimmer switch to regulate gene activity .","The presence of transcription factors attracts proteins such as the Mediator complex , which activates genes by gathering the protein machines that read the genes .","The more transcription factors are found near a specific gene , the more strongly they attract Mediator and the more active the gene is .","A specific region on the transcription factor called the activation domain is necessary for this process .","The biochemical sequences of these domains vary greatly between species , yet activation domains from , for example , yeast and human proteins are often interchangeable .","To understand why this is the case , Sanborn et al . analyzed the genome of baker’s yeast and identified 150 activation domains , each very different in sequence .","Three-quarters of them bound to a subunit of the Mediator complex called Med15 .","Sanborn et al . then developed a machine learning algorithm to predict activation domains in both yeast and humans .","This algorithm also showed that negatively charged and greasy regions on the activation domains were essential to be activated by the Mediator complex .","Further analyses revealed that activation domains used different poses to bind multiple sites on Med15 , a behavior known as ‘fuzzy’ binding .","This creates a high overall affinity even though the binding strength at each individual site is low , enabling the protein complexes to remain dynamic .","These weak interactions together permit fine control over the activity of several genes , allowing cells to respond quickly and precisely to many changes .","The computer algorithm used here provides a new way to identify activation domains across species and could improve our understanding of how living things grow , adapt and evolve .","It could also give new insights into mechanisms of disease , particularly cancer , where transcription factors are often faulty ."],"string":"[\n \"Cells adapt and respond to changes by regulating the activity of their genes .\",\n \"To turn genes on or off , they use a family of proteins called transcription factors .\",\n \"Transcription factors influence specific but overlapping groups of genes , so that each gene is controlled by several transcription factors that act together like a dimmer switch to regulate gene activity .\",\n \"The presence of transcription factors attracts proteins such as the Mediator complex , which activates genes by gathering the protein machines that read the genes .\",\n \"The more transcription factors are found near a specific gene , the more strongly they attract Mediator and the more active the gene is .\",\n \"A specific region on the transcription factor called the activation domain is necessary for this process .\",\n \"The biochemical sequences of these domains vary greatly between species , yet activation domains from , for example , yeast and human proteins are often interchangeable .\",\n \"To understand why this is the case , Sanborn et al . analyzed the genome of baker’s yeast and identified 150 activation domains , each very different in sequence .\",\n \"Three-quarters of them bound to a subunit of the Mediator complex called Med15 .\",\n \"Sanborn et al . then developed a machine learning algorithm to predict activation domains in both yeast and humans .\",\n \"This algorithm also showed that negatively charged and greasy regions on the activation domains were essential to be activated by the Mediator complex .\",\n \"Further analyses revealed that activation domains used different poses to bind multiple sites on Med15 , a behavior known as ‘fuzzy’ binding .\",\n \"This creates a high overall affinity even though the binding strength at each individual site is low , enabling the protein complexes to remain dynamic .\",\n \"These weak interactions together permit fine control over the activity of several genes , allowing cells to respond quickly and precisely to many changes .\",\n \"The computer algorithm used here provides a new way to identify activation domains across species and could improve our understanding of how living things grow , adapt and evolve .\",\n \"It could also give new insights into mechanisms of disease , particularly cancer , where transcription factors are often faulty .\"\n]","isConversational":false},"year":{"kind":"string","value":"2021"}}}],"truncated":false,"partial":false},"paginationData":{"pageIndex":0,"numItemsPerPage":100,"numTotalItems":4346,"offset":0,"length":100}},"jwt":"eyJhbGciOiJFZERTQSIsImtpZCI6IjVHZDBvd0g5MTM2eDZjc1FvbE1zcktNWUZoRVFoUm5PVVVybEpjOUhaUEEifQ.eyJyZWFkIjp0cnVlLCJwZXJtaXNzaW9ucyI6eyJyZXBvLmNvbnRlbnQucmVhZCI6dHJ1ZX0sImlhdCI6MTc4NDMyMTg4OCwianRpIjoiOGMxOWE2Y2EtNGY0YS00ODJmLWFmNTgtYTU1NDAxMzBkZDY4Iiwic3ViIjoiL2RhdGFzZXRzL3NhbmR5MzcvZWxpZmUiLCJleHAiOjE3ODQzMjU0ODgsImlzcyI6Imh0dHBzOi8vaHVnZ2luZ2ZhY2UuY28ifQ.Cf6PozBiyd3EMRfNTGd1YfaA5wXcI-58zgzFLEpK_IISpynf2R3JmIKJGVHnxOMWMST5v55AoUiQFYfnWiM-Cw","displayUrls":true,"splitSizeSummaries":[{"config":"default","split":"train","numRows":4346,"numBytesParquet":131728263},{"config":"default","split":"test","numRows":241,"numBytesParquet":7238778},{"config":"default","split":"validation","numRows":241,"numBytesParquet":7230553}]},"discussionsStats":{"closed":0,"open":1,"total":1},"fullWidth":true,"hasGatedAccess":true,"hasFullAccess":true,"isEmbedded":false,"savedQueries":{"community":[],"user":[]}}">
headings list | keywords list | title stringlengths 30 189 | id stringlengths 14 14 | sections list | abstract list | summary list | year stringclasses 11
values |
|---|---|---|---|---|---|---|---|
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"epidemiology and global health"
] | National and regional seasonal dynamics of all-cause and cause-specific mortality in the USA from 1980 to 2016 | elife-35500-v1 | [
[
"It is well-established that death rates vary throughout the year , and in temperate climates there tend to be more deaths in winter than in summer ( Campbell , 2017; Fowler et al . , 2015; Healy , 2003; McKee , 1989 ) .",
"It has therefore been hypothesized that a warmer world may lower winter mortality ... | [
"In temperate climates , winter deaths exceed summer ones .",
"However , there is limited information on the timing and the relative magnitudes of maximum and minimum mortality , by local climate , age group , sex and medical cause of death .",
"We used geo-coded mortality data and wavelets to analyse the seaso... | [
"In the USA , more deaths happen in the winter than the summer .",
"But when deaths occur varies greatly by sex , age , cause of death , and possibly region .",
"Seasonal differences in death rates can change over time due to changes in factors that cause disease or affect treatment .",
"Analyzing the seasona... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"microbiology and infectious disease",
"immunology and inflammation"
] | Complement and CD4+ T cells drive context-specific corneal sensory neuropathy | elife-48378-v2 | [
[
"Dysregulated complement activation is increasingly recognized as a significant pathological event in a variety of neurodegenerative and neuroinflammatory diseases of the central nervous system ( Tenner et al . , 2018 ) .",
"These include conditions from Alzheimer’s disease and age-related macular degener... | [
"Whether complement dysregulation directly contributes to the pathogenesis of peripheral nervous system diseases , including sensory neuropathies , is unclear .",
"We addressed this important question in a mouse model of ocular HSV-1 infection , where sensory nerve damage is a common clinical problem .",
"Throu... | [
"Most people have likely experienced the discomfort of an eyelash falling onto the surface of their eye .",
"Or that gritty sensation when dust blows into the eye and irritates the surface .",
"These sensations are warnings from sensory nerves in the cornea , the transparent tissue that covers the iris and pupi... | 2019 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"microbiology and infectious disease",
"immunology and inflammation"
] | Phenotypic complementation of genetic immunodeficiency by chronic herpesvirus infection | elife-04494-v1 | [
[
"HOIL-1 ( encoded by the RBCK1 gene ) , HOIP ( RNF31 ) and SHARPIN form the linear ubiquitin chain assembly complex ( LUBAC ) , which linearly ubiquitinates receptor signaling complex components such as NEMO to enhance NF-κB activation after engagement of immune receptors including TNF-R1 , IL-1R , CD40 , TLR... | [
"Variation in the presentation of hereditary immunodeficiencies may be explained by genetic or environmental factors .",
"Patients with mutations in HOIL1 ( RBCK1 ) present with amylopectinosis-associated myopathy with or without hyper-inflammation and immunodeficiency .",
"We report that barrier-raised HOIL-1-... | [
"The immune system protects an individual from invading bacteria , viruses and parasites , as well as malfunctioning or cancerous host cells .",
"However , some people inherit genetic defects that cause part of the immune system to be missing or to not work properly .",
"This is called a genetic immunodeficienc... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Cascade of neural processing orchestrates cognitive control in human frontal cortex | elife-12352-v2 | [
[
"Flexible control of cognitive processes is fundamental to daily activities , including the execution of goal-directed tasks according to stimulus inputs and context dependencies .",
"An important case of cognitive control arises when input stimuli elicit conflicting responses and subjects must select the... | [
"Rapid and flexible interpretation of conflicting sensory inputs in the context of current goals is a critical component of cognitive control that is orchestrated by frontal cortex .",
"The relative roles of distinct subregions within frontal cortex are poorly understood .",
"To examine the dynamics underlying ... | [
"The brain adapts to control our behavior in different ways depending on the specific situation , which is particularly useful when deciding how to interpret conflicting sets of information .",
"The 'Stroop task' is a classic demonstration of this process .",
"In this task , individuals are shown words where th... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"structural biology and molecular biophysics"
] | Structural dynamics of myosin 5 during processive motion revealed by interferometric scattering microscopy | elife-05413-v2 | [
[
"Myosin 5a moves in a hand-over-hand fashion with individual heads moving by 74 nm and the center of mass by 37 nm for each step ( Vale , 2003; Sellers and Veigel , 2006; Hammer and Sellers , 2012 ) .",
"A finely tuned kinetic sequence of ATP binding , hydrolysis , phosphate release and eventually ADP rel... | [
"Myosin 5a is a dual-headed molecular motor that transports cargo along actin filaments .",
"By following the motion of individual heads with interferometric scattering microscopy at nm spatial and ms temporal precision we found that the detached head occupies a loosely fixed position to one side of actin from wh... | [
"Cells use motor proteins that to move organelles and other cargos from one place to another .",
"The myosins are a family of motor proteins that pull cargo along filaments made of another protein called actin .",
"The ‘head’ end of myosin attaches to the actin filament and the ‘tail’ end binds to the cargo .",... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"microbiology and infectious disease"
] | Phenotypic analysis of the unstimulated in vivo HIV CD4 T cell reservoir | elife-60933-v1 | [
[
"Combination antiretroviral therapy ( ART ) suppresses HIV replication but does not eliminate the latent reservoir , which persists for decades in CD4+ T cells and forms a major barrier to HIV cure ( Chun et al . , 1997; Finzi et al . , 1997; Wong et al . , 1997 ) .",
"Our understanding of the phenotypic ... | [
"The latent reservoir is a major barrier to HIV cure .",
"As latently infected cells cannot be phenotyped directly , the features of the in vivo reservoir have remained elusive .",
"Here , we describe a method that leverages high-dimensional phenotyping using CyTOF to trace latently infected cells reactivated e... | [
"There is no cure for the human immunodeficiency virus infection ( HIV ) , but anti-retroviral drugs allow infected people to keep the virus at bay and lead a normal life .",
"These drugs suppress the growth of HIV , but they do not eliminate the virus .",
"If the treatment is interrupted , the virus bounces ba... | 2020 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"ecology",
"computational and systems biology"
] | Tree crickets optimize the acoustics of baffles to exaggerate their mate-attraction signal | elife-32763-v2 | [
[
"Animals produce conspicuous intraspecific communication signals whose intensity is maximised through sexual selection .",
"Several insect species , including tree crickets , use sound to advertise for mates .",
"The louder the call , the farther it travels and the more detectable and attractive it is... | [
"Object manufacture in insects is typically inherited , and believed to be highly stereotyped .",
"Optimization , the ability to select the functionally best material and modify it appropriately for a specific function , implies flexibility and is usually thought to be incompatible with inherited behaviour .",
... | [
"Male tree crickets produce sounds at a specific pitch to attract females .",
"The louder the call , the further the sound travels and the more females he can attract .",
"But making loud sounds is difficult for small animals like insects .",
"To produce sounds , tree crickets rub their wings together and set... | 2017 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Valosin-containing protein (VCP/p97) inhibitors relieve Mitofusin-dependent mitochondrial defects due to VCP disease mutants | elife-17834-v2 | [
[
"IBMPFD is an autosomal dominant disease that afflicts multiple body systems ( Kimonis et al . , 2008a ) .",
"90% of IBMPFD patients display skeletal muscle weakness ( myopathy ) , the primary and earliest symptom of IBMPFD ( Weihl et al . , 2009 ) .",
"With progression of the disease , 50% of patient... | [
"Missense mutations of valosin-containing protein ( VCP ) cause an autosomal dominant disease known as inclusion body myopathy , Paget disease with frontotemporal dementia ( IBMPFD ) and other neurodegenerative disorders .",
"The pathological mechanism of IBMPFD is not clear and there is no treatment .",
"We sh... | [
"A disease called “inclusion body myopathy , Paget disease and frontotemporal dementia ( IBMPFD ) ” is an inherited disorder that can affect the muscles , brain and bones .",
"People affected by the disease find that their muscles become progressively weaker , and may go on to develop a bone disorder and a form o... | 2017 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"structural biology and molecular biophysics",
"microbiology and infectious disease"
] | Structural basis of malodour precursor transport in the human axilla | elife-34995-v2 | [
[
"In Homo sapiens the skin of the underarm ( axilla ) provides a unique niche for bacteria .",
"Through the secretions of various glands , that open onto the skin or into hair follicles , this environment is nutrient-rich and hosts a unique microbial community ( Taylor et al . , 2003 ) .",
"Esterified ... | [
"Mammals produce volatile odours that convey different types of societal information .",
"In Homo sapiens , this is now recognised as body odour , a key chemical component of which is the sulphurous thioalcohol , 3-methyl-3-sulfanylhexan-1-ol ( 3M3SH ) .",
"Volatile 3M3SH is produced in the underarm as a result... | [
"Human body odour contains a number of chemicals , but the most pungent and recognisable are thioalcohols .",
"These molecules are created through a series of chemical reactions that start with an odourless precursor , a compound produced in glands located in our armpits .",
"Then , a type of bacteria called St... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"cell biology",
"immunology and inflammation"
] | Immune surveillance of the lung by migrating tissue monocytes | elife-07847-v2 | [
[
"The mononuclear phagocytic system ( MPS ) consists of a family of cells—including monocytes , macrophages , and dendritic cells ( DC ) —that are derived from common committed bone marrow progenitors and perform related functions ( Hume , 2006; Hashimoto et al . , 2011 ) .",
"The lung is a mucosal surface... | [
"Monocytes are phagocytic effector cells in the blood and precursors of resident and inflammatory tissue macrophages .",
"The aim of the current study was to analyse and compare their contribution to innate immune surveillance of the lung in the steady state with macrophage and dendritic cells ( DC ) .",
"ECFP ... | [
"White blood cells form part of the immune system , which protects the body against infectious diseases and other harmful agents .",
"Some of these cells , including ‘mononuclear phagocytes’ , can reside within different tissues of the body , such as the lungs .",
"Other less specialized cells , called monocyte... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Material and methods"
] | [
"biochemistry and chemical biology",
"genetics and genomics"
] | Genome-wide dynamics of Pol II elongation and its interplay with promoter proximal pausing, chromatin, and exons | elife-02407-v1 | [
[
"Many steps throughout the transcription cycle of RNA polymerase II ( Pol II ) can be regulated , and modulation of any step has the potential to alter the timing and output of mRNA production .",
"After initiation of Pol II , transcription regulation is mediated mostly by the dynamics of Pol II elongatio... | [
"Production of mRNA depends critically on the rate of RNA polymerase II ( Pol II ) elongation .",
"To dissect Pol II dynamics in mouse ES cells , we inhibited Pol II transcription at either initiation or promoter-proximal pause escape with Triptolide or Flavopiridol , and tracked Pol II kinetically using GRO-seq ... | [
"Many different factors determine how quickly the DNA in genes is transcribed to produce molecules of messenger RNA .",
"The start of the transcription process features two milestones: first , an enzyme called RNA Polymerase II starts the process; shortly afterwards , however , the process pauses and only starts ... | 2014 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Ciliary and rhabdomeric photoreceptor-cell circuits form a spectral depth gauge in marine zooplankton | elife-36440-v2 | [
[
"Bilaterian animals have two major photoreceptor cell-types , the rhabdomeric- and the ciliary-type photoreceptor cells ( rPRC and cPRC , respectively ) ( Arendt , 2003; Arendt et al . , 2004; Eakin , 1979; Erclik et al . , 2009 ) .",
"These cells have distinct morphologies and express different classes o... | [
"Ciliary and rhabdomeric photoreceptor cells represent two main lines of photoreceptor-cell evolution in animals .",
"The two cell types coexist in some animals , however how these cells functionally integrate is unknown .",
"We used connectomics to map synaptic paths between ciliary and rhabdomeric photorecept... | [
"The animal kingdom contains many different types of eyes , but all share certain features in common .",
"All detect light using specialized cells called photoreceptors , of which there are two main kinds: ciliary and rhabdomeric .",
"Crustaceans and their relatives , including insects , have rhabdomeric photor... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"medicine",
"microbiology and infectious disease"
] | HIV-1 DNA predicts disease progression and post-treatment virological control | elife-03821-v2 | [
[
"Renewed interest in exploring avenues for curing Human Immunodeficiency Virus Type 1 ( HIV-1 ) infection ( Henrich et al . , 2013; Persaud et al . , 2013; Saez-Cirion et al . , 2013; Denton et al . , 2014; Tebas et al . , 2014 ) has resulted in the investigation of interventions to eradicate cells in which H... | [
"In HIV-1 infection , a population of latently infected cells facilitates viral persistence despite antiretroviral therapy ( ART ) .",
"With the aim of identifying individuals in whom ART might induce a period of viraemic control on stopping therapy , we hypothesised that quantification of the pool of latently in... | [
"HIV is a virus that can hide in , and hijack , the cells of the immune system and force them to make new copies of the virus .",
"This eventually destroys the infected cells and weakens the ability of a person with HIV to fight off infections and disease .",
"If diagnosed early and treated , most people with H... | 2014 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"physics of living systems",
"neuroscience"
] | Analysis of stochastic fluctuations in responsiveness is a critical step toward personalized anesthesia | elife-50143-v1 | [
[
"One of the great promises of personalized medicine is the delivery of a maximally efficacious and minimally harmful dose of appropriate medication to every patient ( Fitzgerald et al . , 2006 ) .",
"Traditionally , dosing decisions are based on the relationship between drug concentration and the magnitud... | [
"Traditionally , drug dosing is based on a concentration-response relationship estimated in a population .",
"Yet , in specific individuals , decisions based on the population-level effects frequently result in over or under-dosing .",
"Here , we interrogate the relationship between population-based and individ... | [
"Every year , millions of patients undergo general anesthesia for complex or life-saving surgeries .",
"In the vast majority of cases , the drugs work as intended .",
"But a minority of patients take longer than expected to regain consciousness after anesthetic , and a few wake up during the surgery itself .",
... | 2019 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Foggy perception slows us down | elife-00031-v1 | [
[
"Visual contrast is usually referred to as the difference in brightness between an object and the background ( Hofstetter et al . , 2000 ) .",
"Classical vision research experiments systematically investigated how visual contrast affects objects motion perception ( Thompson , 1982; Stone and Thompson , 19... | [
"Visual speed is believed to be underestimated at low contrast , which has been proposed as an explanation of excessive driving speed in fog .",
"Combining psychophysics measurements and driving simulation , we confirm that speed is underestimated when contrast is reduced uniformly for all objects of the visual s... | [
"The ways people respond to conditions of reduced visibility is a central topic in vision research .",
"Notably , it has been shown that people tend to underestimate speeds when visibility is reduced equally at all distances , as for example , when driving with a fogged up windshield .",
"But what happens when ... | 2012 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"stem cells and regenerative medicine",
"neuroscience"
] | A functional genomics screen in planarians reveals regulators of whole-brain regeneration | elife-17002-v2 | [
[
"An improved understanding of brain regeneration holds the potential to stimulate new therapies geared toward the treatment of stroke , neurological disease , and brain injury .",
"Brain regeneration , as a process , must occur through several interrelated steps: wound signaling and healing; production of... | [
"Planarians regenerate all body parts after injury , including the central nervous system ( CNS ) .",
"We capitalized on this distinctive trait and completed a gene expression-guided functional screen to identify factors that regulate diverse aspects of neural regeneration in Schmidtea mediterranea .",
"Our scr... | [
"Animals differ in the extent to which they can regenerate missing body parts after injury .",
"Humans regenerate poorly after many injuries , especially when the brain becomes damaged after stroke , disease or trauma .",
"On the other hand , planarians – small worms that live in fresh water – regenerate except... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"biochemistry and chemical biology",
"neuroscience"
] | Real-time in vivo imaging of extracellular ATP in the brain with a hybrid-type fluorescent sensor | elife-57544-v2 | [
[
"Adenosine 5’ triphosphate ( ATP ) is the essential energy currency of intracellular metabolism of all living systems , but also plays important roles as an extracellular signaling molecule in a wide variety of physiological and pathological processes , including neurotransmission ( Burnstock , 2007 ) , neuro... | [
"Adenosine 5’ triphosphate ( ATP ) is a ubiquitous extracellular signaling messenger .",
"Here , we describe a method for in-vivo imaging of extracellular ATP with high spatiotemporal resolution .",
"We prepared a comprehensive set of cysteine-substitution mutants of ATP-binding protein , Bacillus FoF1-ATP synt... | [
"Biologists often refer to a small molecule called adenosine triphosphate – or ATP for short – as ‘the currency of life’ .",
"This molecule carries energy all through the body , and most cells and proteins require ATP to perform their various roles .",
"Nerve cells ( also known as neurons ) in the brain release... | 2020 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"immunology and inflammation"
] | The RNA binding protein IMP3 facilitates tumor immune escape by downregulating the stress-induced ligands ULPB2 and MICB | elife-13426-v1 | [
[
"The regulation of protein expression is essential for development , differentiation , maintenance , division and death of cells as well as homeostasis of tissues ( Buszczak et al . , 2014 ) .",
"Regulation of protein biogenesis is controlled on transcriptional , translational and post-translational level... | [
"Expression of the stress-induced ligands MICA , MICB and ULBP 1–6 are up-regulated as a cellular response to DNA damage , excessive proliferation or viral infection; thereby , they enable recognition and annihilation by immune cells that express the powerful activating receptor NKG2D .",
"This receptor is presen... | [
"Tumor cells differ from healthy cells in many aspects .",
"Importantly , tumor cells have the ability to divide and grow much faster than normal cells .",
"To protect ourselves from full-grown cancers , our bodies have developed a surveillance system: when a tumor cell starts to divide without restraint , “str... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"ecology",
"computational and systems biology"
] | Nomadic-colonial life strategies enable paradoxical survival and growth despite habitat destruction | elife-21673-v4 | [
[
"Two sub-populations comprise our model: the nomadic organisms , and the colonial ones .",
"In a similar vein to habitat-patch models , organisms that exist in multiple sub-populations can be modelled as follows: ( 1 ) dnidt=gi ( ni ) +∑jsijnj-∑jsjiniwhere ni is the size of sub-population i , gi is... | [
"Organisms often exhibit behavioral or phenotypic diversity to improve population fitness in the face of environmental variability .",
"When each behavior or phenotype is individually maladaptive , alternating between these losing strategies can counter-intuitively result in population persistence–an outcome simi... | [
"Many organisms , from slime molds to jellyfish , alternate between life as free-moving “nomadic” individuals and communal life in a more stationary colony .",
"So what evolutionary reasons lie behind such stark behavioral diversity in a single species ?",
"What benefits are obtained by switching from one behav... | 2017 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | A time-stamp mechanism may provide temporal information necessary for egocentric to allocentric spatial transformations | elife-36769-v2 | [
[
"Learning to navigate within the spatial organization of different habitats is essential for animals’ survival ( Geva-Sagiv et al . , 2015 ) .",
"Electric fish , for example , occupied a lucrative ecological niche by evolving the ability to navigate and localize food sources in complete darkness using sho... | [
"Learning the spatial organization of the environment is essential for most animals’ survival .",
"This requires the animal to derive allocentric spatial information from egocentric sensory and motor experience .",
"The neural mechanisms underlying this transformation are mostly unknown .",
"We addressed this... | [
"Finding their way around is an essential part of survival for many animals and helps them to locate food , mates and shelter .",
"Animals have evolved the ability to form a 'map' or representation of their surroundings .",
"For example , the electric fish Apteronotus leptorhynchus , is able to precisely learn ... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Phase-amplitude coupling supports phase coding in human ECoG | elife-07886-v2 | [
[
"Perceptual representations of the environment are critical to an animal's survival and are believed to occur through coactivated neuronal groups known as cell assemblies .",
"Human neuronal firing ( Ekstrom et al . , 2007; Kraskov et al . , 2007; Chan et al . , 2011; Rey et al . , 2014 ) and increases in... | [
"Prior studies have shown that high-frequency activity ( HFA ) is modulated by the phase of low-frequency activity .",
"This phenomenon of phase-amplitude coupling ( PAC ) is often interpreted as reflecting phase coding of neural representations , although evidence for this link is still lacking in humans .",
"... | [
"Electrocorticography , or ECoG , is a technique that is used to record the electrical activity of the brain via electrodes placed inside the skull .",
"This electrical activity repeatedly rises and falls , and can therefore be represented as a series of waves .",
"All waves have three basic properties: amplitu... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"short report",
"structural biology and molecular biophysics",
"microbiology and infectious disease"
] | Crystal structure of PfRh5, an essential P. falciparum ligand for invasion of human erythrocytes | elife-04187-v2 | [
[
"Plasmodium falciparum is the causative agent of the most severe form of malaria with over 700 , 000 deaths each year , mostly in sub-Saharan Africa .",
"The asexual blood cycle of this parasite begins with the invasion of human erythrocytes by the merozoite form of P . falciparum in a complex multistep p... | [
"Plasmodium falciparum causes the most severe form of malaria in humans and is responsible for over 700 , 000 deaths annually .",
"It is an obligate intracellular parasite and invades erythrocytes where it grows in a relatively protected niche .",
"Invasion of erythrocytes is essential for parasite survival and... | [
"Malaria is a disease caused by a single-celled parasite called Plasmodium , which is transmitted between humans by mosquitoes .",
"It is estimated that 3 . 4 billion people worldwide live in regions where they are at risk of malaria , and malaria infections cause hundreds of thousands of deaths each year .",
"... | 2014 |
[
"Introduction",
"Materials and methods",
"Results",
"Discussion"
] | [
"ecology",
"epidemiology and global health"
] | Both consumptive and non-consumptive effects of predators impact mosquito populations and have implications for disease transmission | elife-71503-v1 | [
[
"While it is well known that predation reduces vector populations through consumptive effects , non-consumptive effects of predators can also greatly impact prey demographics ( Preisser et al . , 2005 ) .",
"Mosquitoes are vectors of a variety of debilitating and deadly diseases , including malaria , lymp... | [
"Predator-prey interactions influence prey traits through both consumptive and non-consumptive effects , and variation in these traits can shape vector-borne disease dynamics .",
"Meta-analysis methods were employed to generate predation effect sizes by different categories of predators and mosquito prey .",
"T... | [
"Mosquitoes are often referred to as the deadliest animals on earth because some species spread malaria , West Nile virus or other dangerous diseases when they bite humans and other animals .",
"Adult mosquitoes fly to streams , ponds and other freshwater environments to lay their eggs .",
"When the eggs hatch ... | 2022 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"evolutionary biology",
"biochemistry and chemical biology"
] | Host proteostasis modulates influenza evolution | elife-28652-v1 | [
[
"Minimalist pathogens like RNA viruses survive dynamic host environments by virtue of their extreme adaptability .",
"This adaptability is driven by a high rate of genetic variation , mediated by error-prone genome replication ( Sanjuán et al . , 2010 ) .",
"Most missense mutations have deleterious co... | [
"Predicting and constraining RNA virus evolution require understanding the molecular factors that define the mutational landscape accessible to these pathogens .",
"RNA viruses typically have high mutation rates , resulting in frequent production of protein variants with compromised biophysical properties .",
"... | [
"Influenza viruses , commonly called flu , can evade our immune system and develop resistance to treatments by changing frequently .",
"Specifically , mutations in their genome cause influenza proteins to change in ways that can help the virus evade our defences .",
"However , these mutations come at a cost and... | 2017 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"short report",
"neuroscience"
] | Peroxisomal dysfunctions cause lysosomal storage and axonal Kv1 channel redistribution in peripheral neuropathy | elife-23332-v1 | [
[
"Schwann cells , the myelinating cells of the peripheral nervous system ( PNS ) , contain peroxisomes in distant cytoplasmic compartments , including myelin channels and cytosolic loop regions close to nodes of Ranvier ( Kassmann et al . , 2011 ) .",
"Peroxisomes degrade fatty acids derived from myelin li... | [
"Impairment of peripheral nerve function is frequent in neurometabolic diseases , but mechanistically not well understood .",
"Here , we report a novel disease mechanism and the finding that glial lipid metabolism is critical for axon function , independent of myelin itself .",
"Surprisingly , nerves of Schwann... | [
"Nerve cells transmit messages along their length in the form of electrical signals .",
"Much like an electrical wire , the nerve fiber or axon is coated by a multiple-layered insulation , called the myelin sheath .",
"However , unlike electrical insulation , the myelin sheath is regularly interrupted to expose... | 2017 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"stem cells and regenerative medicine"
] | Stem cells repurpose proliferation to contain a breach in their niche barrier | elife-41661-v2 | [
[
"Adult stem cells reside in most if not all tissues .",
"They are endowed with the ability to self-renew long-term and differentiate into specialized cell lineages in order to maintain , renew and repair tissues .",
"In turn , stem cells are heavily influenced by the tissue microenvironment in which t... | [
"Adult stem cells are responsible for life-long tissue maintenance .",
"They reside in and interact with specialized tissue microenvironments ( niches ) .",
"Using murine hair follicle as a model , we show that when junctional perturbations in the niche disrupt barrier function , adjacent stem cells dramaticall... | [
"Most , if not all , tissues of an adult animal contain stem cells .",
"These stem cells regenerate and repair damaged tissues and organs for the entire lifetime of an animal , contributing to a healthy life .",
"They divide to make daughter cells that become either new stem cells or specialized cells of that o... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"developmental biology"
] | Extracellular interactions and ligand degradation shape the nodal morphogen gradient | elife-13879-v2 | [
[
"In many animals , development from a single cell to a complex multicellular organism requires the graded distribution and activity of diffusible proteins , which are known as morphogens .",
"How morphogen gradients are formed is not fully understood .",
"Studies in many organisms have suggested three... | [
"The correct distribution and activity of secreted signaling proteins called morphogens is required for many developmental processes .",
"Nodal morphogens play critical roles in embryonic axis formation in many organisms .",
"Models proposed to generate the Nodal gradient include diffusivity , ligand processing... | [
"Animals develop from a single fertilized egg cell into multicellular organisms .",
"This process requires chemical signals called “morphogens” that instruct the cells how to behave during development .",
"The morphogens move across cells and tissues to form gradients of the signal .",
"Cells then respond in ... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"evolutionary biology",
"cell biology"
] | Genetic diversity of CHC22 clathrin impacts its function in glucose metabolism | elife-41517-v1 | [
[
"Clathrin-coated vesicles ( CCVs ) are key players in eukaryotic intracellular membrane traffic ( Brodsky , 2012 ) .",
"Their characteristic lattice-like coat is self-assembled from cytoplasmic clathrin proteins , captures membrane-embedded protein cargo and deforms the membrane into a vesicle .",
"Th... | [
"CHC22 clathrin plays a key role in intracellular membrane traffic of the insulin-responsive glucose transporter GLUT4 in humans .",
"We performed population genetic and phylogenetic analyses of the CHC22-encoding CLTCL1 gene , revealing independent gene loss in at least two vertebrate lineages , after arising fr... | [
"When we eat carbohydrates , they are digested into sugars that circulate in the blood to provide energy for the brain and other parts of the body .",
"But too much blood sugar can be poisonous .",
"The body regulates blood sugar balance using the hormone insulin , which triggers the removal of sugar from the b... | 2019 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"chromosomes and gene expression",
"tools and resources"
] | A promoter interaction map for cardiovascular disease genetics | elife-35788-v2 | [
[
"A major goal in human genetics research is to understand genetic contributions to complex diseases , specifically the molecular mechanisms by which common DNA variants impact disease etiology .",
"Most genome-wide association studies ( GWAS ) implicate non-coding variants that are far from genes , compli... | [
"Over 500 genetic loci have been associated with risk of cardiovascular diseases ( CVDs ) ; however , most loci are located in gene-distal non-coding regions and their target genes are not known .",
"Here , we generated high-resolution promoter capture Hi-C ( PCHi-C ) maps in human induced pluripotent stem cells ... | [
"Our genomes contain around 20 , 000 different genes that code for instructions to create proteins and other important molecules .",
"When changes , or mutations , occur within these genes , malfunctioning proteins that are damaging to the cell may be produced .",
"Researchers of human genetics have tried to sp... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"ecology"
] | The golden mimicry complex uses a wide spectrum of defence to deter a community of predators | elife-22089-v1 | [
[
"Mimicry complexes ( i . e . a composite of several Müllerian and Batesian mimics ) have long fascinated biologists for the theoretical and empirical challenges they pose .",
"Classic Müllerian mimicry was thought to only include equally defended species ( Müller , 1878 ) , while Batesian mimicry ( Bates ... | [
"Mimicry complexes typically consist of multiple species that deter predators using similar anti-predatory signals .",
"Mimics in these complexes are assumed to vary in their level of defence from highly defended through to moderately defended , or not defended at all .",
"Here , we report a new multi-order mim... | [
"Many animals use bright colours to warn a potential predator that they can defend themselves .",
"Wasps , for instance , are armed with a harmful sting and advertise this fact via their distinctive yellow and black stripes .",
"Predators often learn to heed such warnings and avoid these unpalatable animals in ... | 2017 |
[
"Results",
"Discussion"
] | [
"computational and systems biology",
"neuroscience"
] | Continuous attractors for dynamic memories | elife-69499-v1 | [
[
"The spontaneous dynamics produced by the network is constrained to the low dimensional manifold codified in the connectivity matrix and spanned by the parameter x→ .",
"The short-range interactions and the uniform inhibition enforced by the firing threshold h0 produce a localized 'bump' of activity in th... | [
"Episodic memory has a dynamic nature: when we recall past episodes , we retrieve not only their content , but also their temporal structure .",
"The phenomenon of replay , in the hippocampus of mammals , offers a remarkable example of this temporal dynamics .",
"However , most quantitative models of memory tre... | [
"When we recall a past experience , accessing what is known as an ‘episodic memory’ , it usually does not appear as a still image or a snapshot of what occurred .",
"Instead , our memories tend to be dynamic: we remember how a sequence of events unfolded , and when we do this , we often re-experience at least par... | 2021 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"developmental biology",
"physics of living systems"
] | Self-organized patterning of cell morphology via mechanosensitive feedback | elife-57964-v3 | [
[
"During morphogenesis , tissues with complex morphologies are formed from the collective interplay of many cells .",
"This process involves spatial patterns of signaling activity , and previous work has discovered mechanisms for generating tissue-scale patterns of activity in signaling pathways such as He... | [
"Tissue organization is often characterized by specific patterns of cell morphology .",
"How such patterns emerge in developing tissues is a fundamental open question .",
"Here , we investigate the emergence of tissue-scale patterns of cell shape and mechanical tissue stress in the Drosophila wing imaginal disc... | [
"During development , carefully choreographed cell movements ensure the creation of a healthy organism .",
"To determine their identity and place across a tissue , cells can read gradients of far-reaching signaling molecules called morphogens; in addition , physical forces can play a part in helping cells acquire... | 2021 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"microbiology and infectious disease",
"immunology and inflammation"
] | Mannose receptor is an HIV restriction factor counteracted by Vpr in macrophages | elife-51035-v1 | [
[
"Vpr is a highly conserved HIV accessory protein that is necessary for optimal replication in macrophages ( Balliet et al . , 1994 ) but its mechanism of action is poorly understood .",
"Studies using human lymphoid tissue ( HLT ) , which are rich in both T cells and macrophages , have found that loss of ... | [
"HIV-1 Vpr is necessary for maximal HIV infection and spread in macrophages .",
"Evolutionary conservation of Vpr suggests an important yet poorly understood role for macrophages in HIV pathogenesis .",
"Vpr counteracts a previously unknown macrophage-specific restriction factor that targets and reduces the exp... | [
"Human cells have defense mechanisms against viral infection known as restriction factors .",
"These are proteins that break down parts of a virus including its DNA or proteins .",
"To evade these defenses , viruses in turn make proteins that block or break down restriction factors .",
"This battle between hu... | 2020 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"cell biology",
"neuroscience"
] | Sam68 promotes self-renewal and glycolytic metabolism in mouse neural progenitor cells by modulating Aldh1a3 pre-mRNA 3'-end processing | elife-20750-v2 | [
[
"Neurogenesis is the process that leads to a regionalized brain from an embryonic neuroepithelium .",
"In the mouse developing brain , most neurons are generated between the 10th day of embryonal life ( E10 ) and birth ( Martynoga et al . , 2012 ) , even though adult neurogenesis persists in restricted ar... | [
"The balance between self-renewal and differentiation of neural progenitor cells ( NPCs ) dictates neurogenesis and proper brain development .",
"We found that the RNA- binding protein Sam68 ( Khdrbs1 ) is strongly expressed in neurogenic areas of the neocortex and supports the self-renewing potential of mouse NP... | [
"Neurons develop from cells called neural progenitors .",
"These cells can either divide to produce more progenitor cells or develop into specific types of neurons .",
"These two activities – known as self-renewal and differentiation – must be balanced to produce the right number of specialized neurons , withou... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and method"
] | [
"short report",
"neuroscience"
] | Hippocampal place cells construct reward related sequences through unexplored space | elife-06063-v1 | [
[
"We investigated whether the presence of an inaccessible goal in an unvisited portion of an environment was sufficient to elicit pre-activation ( ‘preplay’ ) of hippocampal place cell sequences that will subsequently represent runs through the unvisited environment .",
"To this end , we recorded from ense... | [
"Dominant theories of hippocampal function propose that place cell representations are formed during an animal's first encounter with a novel environment and are subsequently replayed during off-line states to support consolidation and future behaviour .",
"Here we report that viewing the delivery of food to an u... | [
"As an animal explores an area , part of the brain called the hippocampus creates a mental map of the space .",
"When the animal is in one location , a few neurons called ‘place cells’ will fire .",
"If the animal moves to a new spot , other place cells fire instead .",
"Each time the animal returns to that s... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"biochemistry and chemical biology",
"structural biology and molecular biophysics"
] | Cooperative base pair melting by helicase and polymerase positioned one nucleotide from each other | elife-06562-v2 | [
[
"Replicative helicases and DNA polymerases ( DNAPs ) are not efficient at unwinding the duplex DNA when they are working independently .",
"The unwinding rates are slower than their translocation rates on single-stranded ( ss ) DNA and slower than the rates of DNA replication ( Kim et al . , 1996; Delagou... | [
"Leading strand DNA synthesis requires functional coupling between replicative helicase and DNA polymerase ( DNAP ) enzymes , but the structural and mechanistic basis of coupling is poorly understood .",
"This study defines the precise positions of T7 helicase and T7 DNAP at the replication fork junction with sin... | [
"DNA replication is the process whereby a molecule of DNA is copied to form two identical molecules .",
"First , an enzyme called a DNA helicase separates the two strands of the DNA double helix .",
"This forms a structure called a replication fork that has two exposed single strands .",
"Other enzymes called... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Genetic dissection of the different roles of hypothalamic kisspeptin neurons in regulating female reproduction | elife-43999-v2 | [
[
"Infertility is a common clinical problem affecting 15% of couples; ovulatory disorders account for 25% of this total ( Macaluso et al . , 2010 ) .",
"The hypothalamic-pituitary-gonadal axis controls reproduction and malfunction of this axis can cause ovulatory dysfunction and/or other disturbances of the... | [
"The brain regulates fertility through gonadotropin-releasing hormone ( GnRH ) neurons .",
"Estradiol induces negative feedback on pulsatile GnRH/luteinizing hormone ( LH ) release and positive feedback generating preovulatory GnRH/LH surges .",
"Negative and positive feedbacks are postulated to be mediated by ... | [
"Female reproduction relies on a complex balance of hormones that drive the reproductive cycle ( menstrual cycle in humans ) and influence fertility .",
"A hormone called GnRH , which stands for gonadotropin-releasing hormone , plays a major role in regulating this balance .",
"GnRH is transmitted from the brai... | 2019 |
[
"Introduction",
"Results",
"Discussion"
] | [
"structural biology and molecular biophysics"
] | Conserved conformational selection mechanism of Hsp70 chaperone-substrate interactions | elife-32764-v1 | [
[
"An elaborate network of chaperones is present in organisms across all domains of life to oversee the health of the cellular proteome ( Balchin et al . , 2016 ) .",
"The Hsp70 family of molecular chaperones occupies a central node in this network , steering proteins synthesized on the ribosome to their na... | [
"Molecular recognition is integral to biological function and frequently involves preferred binding of a molecule to one of several exchanging ligand conformations in solution .",
"In such a process the bound structure can be selected from the ensemble of interconverting ligands a priori ( conformational selectio... | [
"Proteins are the workhorses of a cell and are involved in almost all biological processes .",
"Newly made proteins need to ‘fold’ into precise three-dimensional shapes in order to carry out their roles .",
"However , proteins sometimes fold incorrectly or unfold .",
"These protein forms are not able to work ... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"structural biology and molecular biophysics"
] | Direct single molecule measurement of TCR triggering by agonist pMHC in living primary T cells | elife-00778-v1 | [
[
"An essential aspect of adaptive immunity is the ability of T cells to discriminate between structurally similar agonist and non-stimulatory self peptides bound to major histocompatibility complex ( pMHC ) molecules presented on the surface of antigen-presenting cells ( APCs ) ( Weiss and Littman , 1994 ) .",... | [
"T cells discriminate between self and foreign antigenic peptides , displayed on antigen presenting cell surfaces , via the TCR .",
"While the molecular interactions between TCR and its ligands are well characterized in vitro , quantitative measurements of these interactions in living cells are required to accura... | [
"The immune system identifies and combats foreign objects , including pathogens , in the body .",
"T cells are key components of the immune system , and each has a unique variant of a signalling complex known as the T cell receptor on its surface .",
"T cells scan the surfaces of other cells in search of antige... | 2013 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"physics of living systems",
"microbiology and infectious disease"
] | Cyanobacteria use micro-optics to sense light
direction | elife-12620-v1 | [
[
"Many prokaryotes move directionally in response to a chemical or physical stimulus .",
"However , it is generally assumed that bacteria are too small for direct sensing of a",
"concentration gradient across the cell: instead they probe changes in stimulus",
"concentration over time , as in the cl... | [
"Bacterial phototaxis was first recognized over a century ago , but the method by",
"which such small cells can sense the direction of illumination has remained",
"puzzling .",
"The unicellular cyanobacterium Synechocystis sp .",
"PCC",
"6803 moves with Type IV pili and measures light intensity and color ... | [
"Cyanobacteria are blue-green bacteria that are abundant in the environment .",
"Cyanobacteria in the oceans are among the world’s most important oxygen",
"producers and carbon dioxide consumers .",
"Synechocystis is a",
"spherical single-celled cyanobacterium that measures about three thousandths of",
"a... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"cell biology"
] | Actin filaments target the oligomeric maturation of the dynamin GTPase Drp1 to mitochondrial fission sites | elife-11553-v3 | [
[
"Mitochondrial fission and fusion create a dynamic network that is necessary for mitochondrial distribution and homeostasis .",
"During mitosis , fission is necessary to distribute mitochondria between daughter cells .",
"In neurons , fission and transport maintains mitochondrial distribution througho... | [
"While the dynamin GTPase Drp1 plays a critical role during mitochondrial fission , mechanisms controlling its recruitment to fission sites are unclear .",
"A current assumption is that cytosolic Drp1 is recruited directly to fission sites immediately prior to fission .",
"Using live-cell microscopy , we find e... | [
"Inside cells , structures called mitochondria supply the energy needed to carry out the processes that sustain life .",
"Mitochondria constantly divide ( a process known as fission ) or fuse together , which helps to keep them in good working condition and well distributed around the cell .",
"Several neurolog... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"developmental biology"
] | Genetic dissection of the Transcription Factor code controlling serial specification of muscle identities in Drosophila | elife-14979-v2 | [
[
"The morphological diversity of body wall muscles is necessary for precision , strength and coordination of body movements specific to each animal species .",
"The development of the complex architecture of the body wall musculature of the Drosophila larva – 30 different muscles in each hemi-segment ( Bat... | [
"Each Drosophila muscle is seeded by one Founder Cell issued from terminal division of a Progenitor Cell ( PC ) .",
"Muscle identity reflects the expression by each PC of a specific combination of identity Transcription Factors ( iTFs ) .",
"Sequential emergence of several PCs at the same position raised the qu... | [
"Animals have many different muscles of various shapes and sizes that are suited to specific tasks and behaviors .",
"The fruit fly known as Drosophila has a fairly simple musculature , which makes it an ideal model animal to investigate how different muscles form .",
"In fruit fly embryos , cells called progen... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"chromosomes and gene expression"
] | Mutations in L-type amino acid transporter-2 support SLC7A8 as a novel gene involved in age-related hearing loss | elife-31511-v4 | [
[
"Age-related hearing loss ( ARHL ) or presbycusis is one of the most prevalent chronic medical conditions associated with aging .",
"Indeed , more than 30% of people aged over 65 years suffer ARHL ( Gates and Mills , 2005; Huang and Tang , 2010; Van Eyken et al . , 2007 ) .",
"Clinically , ARHL is def... | [
"Age-related hearing loss ( ARHL ) is the most common sensory deficit in the elderly .",
"The disease has a multifactorial etiology with both environmental and genetic factors involved being largely unknown .",
"SLC7A8/SLC3A2 heterodimer is a neutral amino acid exchanger .",
"Here , we demonstrated that SLC7A... | [
"Age-related hearing loss affects about one in three individuals between the ages of 65 and 74 .",
"The first symptom is difficulty hearing high-pitched sounds like children’s voices .",
"The disease starts gradually and worsens over time .",
"Changes in the ear , the nerve that connects it to the brain , or ... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"biochemistry and chemical biology",
"structural biology and molecular biophysics"
] | Structure and mechanism of a phage-encoded SAM lyase revises catalytic function of enzyme family | elife-61818-v1 | [
[
"S-adenosyl methionine ( SAM ) plays many important roles in biology .",
"It is an essential methyl donor for methyltransferases that act on nucleic acids , proteins , lipids , and small molecules , but is also involved in many other reactions , for example , as a substrate used in biosynthesis of polyami... | [
"The first S-adenosyl methionine ( SAM ) degrading enzyme ( SAMase ) was discovered in bacteriophage T3 , as a counter-defense against the bacterial restriction-modification system , and annotated as a SAM hydrolase forming 5’-methyl-thioadenosine ( MTA ) and L-homoserine .",
"From environmental phages , we recen... | [
"Bacteria can be infected by viruses known as bacteriophages .",
"These viruses inject their genetic material into bacterial cells and use the bacteria’s own machinery to build the proteins they need to survive and infect other cells .",
"To protect themselves , bacteria produce a molecule called S-adenosyl met... | 2021 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"evolutionary biology",
"microbiology and infectious disease"
] | Mutational resilience of antiviral restriction favors primate TRIM5α in host-virus evolutionary arms races | elife-59988-v1 | [
[
"Mammalian genomes combat the persistent threat of viruses by encoding a battery of cell-intrinsic antiviral proteins , termed restriction factors , that recognize and inhibit viral replication within host cells .",
"The potency of restriction factors places selective pressure on viruses to evade recognit... | [
"Host antiviral proteins engage in evolutionary arms races with viruses , in which both sides rapidly evolve at interaction interfaces to gain or evade immune defense .",
"For example , primate TRIM5α uses its rapidly evolving ‘v1’ loop to bind retroviral capsids , and single mutations in this loop can dramatical... | [
"The evolutionary battle between viruses and the immune system is essentially a high-stakes arms race .",
"The immune system makes antiviral proteins , called restriction factors , which can stop the virus from replicating .",
"In response , viruses evolve to evade the effects of restriction factors .",
"To c... | 2020 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"biochemistry and chemical biology",
"cell biology"
] | Tim29 is a novel subunit of the human TIM22 translocase and is involved in complex assembly and stability | elife-17463-v3 | [
[
"Mitochondria cannot be created de novo and pre-existing mitochondria are used as templates for mitochondrial biogenesis .",
"This genesis requires the ~1500 different mitochondrial proteins to be imported via dynamic translocation machines to one of four subcompartments of the organelle – outer and inner... | [
"The TIM22 complex mediates the import of hydrophobic carrier proteins into the mitochondrial inner membrane .",
"While the TIM22 machinery has been well characterised in yeast , the human complex remains poorly characterised .",
"Here , we identify Tim29 ( C19orf52 ) as a novel , metazoan-specific subunit of t... | [
"Mitochondria are like tiny bean-shaped “power stations” that provide our cells with the vast majority of the energy that they need .",
"These structures , however , are not self-sufficient and instead rely on proteins and chemicals that are imported from elsewhere in the cell .",
"Two layers of membrane enclos... | 2016 |
[
"Introduction",
"Results",
"Discussion"
] | [
"biochemistry and chemical biology",
"physics of living systems",
"tools and resources"
] | Optical estimation of absolute membrane potential using fluorescence lifetime imaging | elife-44522-v2 | [
[
"Membrane potential ( Vmem ) is an essential facet of cellular physiology .",
"In electrically excitable cells , such as neurons and cardiomyocytes , voltage-gated ion channels enable rapid changes in membrane potential .",
"These fast membrane potential changes , on the order of milliseconds to secon... | [
"All cells maintain ionic gradients across their plasma membranes , producing transmembrane potentials ( Vmem ) .",
"Mounting evidence suggests a relationship between resting Vmem and the physiology of non-excitable cells with implications in diverse areas , including cancer , cellular differentiation , and body ... | [
"All living cells are like tiny batteries .",
"As long as a cell is alive , it actively maintains a difference in electrical charge between its interior and exterior .",
"This charge difference , or voltage , is called the membrane potential , and it is vital for our bodies to work properly .",
"For example ,... | 2019 |
[
"Introduction",
"Results and discussion",
"Materials and methods"
] | [
"short report",
"neuroscience"
] | A molecular mechanism underlying gustatory memory trace for an association in the insular cortex | elife-07582-v2 | [
[
"Association between events in time and space is a major mechanism for all animals , including humans , to learn about the world , which can potentially change their behavior in future circumstances .",
"The most influential books concerning learning theory published early in the 20th century were Animal ... | [
"Events separated in time are associatively learned in trace conditioning , recruiting more neuronal circuits and molecular mechanisms than in delay conditioning .",
"However , it remains unknown whether a given sensory memory trace is being maintained as a unitary item to associate .",
"Here , we used conditio... | [
"The survival of animals , including us humans , depends on the ability to discriminate good food from bad .",
"We would prefer eating a given taste if it did not cause any negative feelings after eating it for the first time; however , we would avoid eating that specific taste if it caused any digestive discomfo... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"microbiology and infectious disease",
"genetics and genomics"
] | Virophages and retrotransposons colonize the genomes of a heterotrophic flagellate | elife-72674-v1 | [
[
"Many eukaryotic genomes harbor endogenous viral elements ( EVEs ) ( Feschotte and Gilbert , 2012 ) .",
"For retroviruses , integration as a provirus is an essential part of their replication cycles , but other viruses also occasionally endogenize , for instance with the help of cellular retroelements ( H... | [
"Virophages can parasitize giant DNA viruses and may provide adaptive anti-giant virus defense in unicellular eukaryotes .",
"Under laboratory conditions , the virophage mavirus integrates into the nuclear genome of the marine flagellate Cafeteria burkhardae and reactivates upon superinfection with the giant viru... | [
"Viruses exist in all ecosystems in vast numbers and infect many organisms .",
"Some of them are harmful but others can protect the organisms they infect .",
"For example , a group of viruses called virophages protect microscopic sea creatures called plankton from deadly infections by so-called giant viruses ."... | 2021 |
[
"Main text",
"Materials and methods"
] | [
"short report",
"genetics and genomics"
] | A feedback loop between nonsense-mediated decay and the retrogene DUX4 in facioscapulohumeral muscular dystrophy | elife-04996-v3 | [
[
"Facioscapulohumeral muscular dystrophy ( FSHD ) is typically an adult-onset muscular dystrophy characterized by muscle weakness initially affecting the face ( facio ) , shoulders ( scapulo ) , and upper arms ( humeral ) .",
"FSHD is caused by decreased epigenetic repression of the D4Z4 macrosatellite arr... | [
"Facioscapulohumeral muscular dystrophy ( FSHD ) is a muscular dystrophy caused by inefficient epigenetic repression of the D4Z4 macrosatellite array and somatic expression of the DUX4 retrogene .",
"DUX4 is a double homeobox transcription factor that is normally expressed in the testis and causes apoptosis and F... | [
"Genes are sequences of DNA that contain instructions for the cell that must be carefully controlled because it is not always appropriate or safe for these instructions to be followed .",
"When genes are active , copies of the DNA are made using molecules of ribonucleic acid ( RNA ) and these can then be used as ... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"plant biology"
] | A regulatory module controlling stress-induced cell cycle arrest in Arabidopsis | elife-43944-v1 | [
[
"To survive under fluctuating environmental conditions , all organisms have the ability to deal with various kinds of internal and external stresses .",
"In animals , failure to adapt against disadvantageous stresses leads to various diseases such as cancer , and , in severe cases , cells perish ( Twayana... | [
"Cell cycle arrest is an active response to stresses that enables organisms to survive under fluctuating environmental conditions .",
"While signalling pathways that inhibit cell cycle progression have been elucidated , the putative core module orchestrating cell cycle arrest in response to various stresses is st... | [
"During environmental stresses , such as high light or a drought , plants do not have the opportunity to up and leave .",
"Instead , they need to buy time and energy by pausing their growth , which means stopping their cells from dividing .",
"In this case , the cell cycle , a series of stages during which a ce... | 2019 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"biochemistry and chemical biology",
"cell biology"
] | ULK3 regulates cytokinetic abscission by phosphorylating ESCRT-III proteins | elife-06547-v2 | [
[
"Cytokinesis is the final stage of cell division when two daughter cells are physically separated .",
"The process comprises several steps , including cleavage furrow ingression and abscission of the midbody—the thin intercellular bridge connecting the nascent daughter cells ( Caballe and Martin-Serrano ,... | [
"The endosomal sorting complexes required for transport ( ESCRT ) machinery mediates the physical separation between daughter cells during cytokinetic abscission .",
"This process is regulated by the abscission checkpoint , a genome protection mechanism that relies on Aurora B and the ESCRT-III subunit CHMP4C to ... | [
"Our cells multiply by dividing into two .",
"Many proteins are involved in this process , including a group called the ESCRT-III complex .",
"This group is required to complete the final stage of cell division when the single membrane that surrounds the two new daughter cells separates .",
"Before the cell d... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"ecology",
"microbiology and infectious disease"
] | Recurring patterns in bacterioplankton dynamics during coastal spring algae blooms | elife-11888-v2 | [
[
"Pelagic zones of the world's oceans seemingly constitute rather homogenous habitats , however , they feature enough spatial and temporal variation to support a large number of species with distinct niches .",
"This phenomenon has been termed 'paradox of the plankton' by G . Evelyn Hutchinson ( Hutchinson... | [
"A process of global importance in carbon cycling is the remineralization of algae biomass by heterotrophic bacteria , most notably during massive marine algae blooms .",
"Such blooms can trigger secondary blooms of planktonic bacteria that consist of swift successions of distinct bacterial clades , most prominen... | [
"Small algae in the world's oceans remove about as much carbon dioxide from the atmosphere as land plants .",
"These algae do not grow continuously , but often surge in numbers during temporary blooms .",
"Such blooms can be large enough to be seen from space by satellites .",
"The lifespan of algae within su... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Risk of punishment influences discrete and coordinated encoding of reward-guided actions by prefrontal cortex and VTA neurons | elife-30056-v2 | [
[
"Goal-directed actions aimed at obtaining a reward often involve exposure to an aversive event or punishment .",
"For example , foraging for food in the wild may result in encountering a predator .",
"In a causally and socially complex world , appropriate representation of punishment that lurks around... | [
"Actions motivated by rewards are often associated with risk of punishment .",
"Little is known about the neural representation of punishment risk during reward-seeking behavior .",
"We modeled this circumstance in rats by designing a task where actions were consistently rewarded but probabilistically punished ... | [
"When deciding what to do , we usually try to predict the likely outcomes of our actions .",
"This helps us choose behaviors that will lead to positive outcomes , or rewards , and avoid those that will lead to negative outcomes , or punishments .",
"But in practice , actions that offer the possibility of reward... | 2017 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"evolutionary biology",
"biochemistry and chemical biology"
] | Non-classical amine recognition evolved in a large clade of olfactory receptors | elife-10441-v2 | [
[
"Biogenic amines are key intercellular signaling molecules that regulate physiology and behavior .",
"In vertebrates , biogenic amines include adrenaline , serotonin , acetylcholine , histamine , and dopamine , and these biogenic amines can be detected by G Protein-Coupled Receptors ( GPCRs ) expressed on... | [
"Biogenic amines are important signaling molecules , and the structural basis for their recognition by G Protein-Coupled Receptors ( GPCRs ) is well understood .",
"Amines are also potent odors , with some activating olfactory trace amine-associated receptors ( TAARs ) .",
"Here , we report that teleost TAARs e... | [
"Many organisms make molecules called biogenic amines .",
"These molecules , which include the human hormones adrenaline and histamine , have important roles in regulating the biology and behaviour of many animals .",
"Some biogenic amines bind to receptor proteins called GPCRs on the surface of cells .",
"Ma... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Maximally informative foraging by Caenorhabditis elegans | elife-04220-v1 | [
[
"In considering animal behavior and decision-making , it is exciting to consider the proposal ( Polani , 2009; Tishby and Polani , 2011 ) that animals may be guided by fundamental statistical quantities , such as the maximization of Shannon mutual information ( Cover and Thomas , 1991 ) .",
"The advantage... | [
"Animals have evolved intricate search strategies to find new sources of food .",
"Here , we analyze a complex food seeking behavior in the nematode Caenorhabditis elegans ( C . elegans ) to derive a general theory describing different searches .",
"We show that C . elegans , like many other animals , uses a mu... | [
"How an animal forages for food can make the difference between life and death , and there are several different searching strategies that may be adopted .",
"Foraging could be more productive if animals could take into account any of the patterns with which food is distributed in their environment , but how much... | 2014 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"tools and resources",
"neuroscience"
] | Sensitive red protein calcium indicators for imaging neural activity | elife-12727-v2 | [
[
"Genetically-encoded calcium indicators ( GECIs ) enable non-invasive measurement of neuronal activity in vivo .",
"Activity can be tracked across multiple spatial scales , from synapses ( Chen et al . , 2013b ) to populations of thousands of neurons ( Ahrens et al . , 2013; Peron et al . , 2015 ) .",
... | [
"Genetically encoded calcium indicators ( GECIs ) allow measurement of activity in large populations of neurons and in small neuronal compartments , over times of milliseconds to months .",
"Although GFP-based GECIs are widely used for in vivo neurophysiology , GECIs with red-shifted excitation and emission spect... | [
"Neurons encode information with brief electrical pulses called spikes .",
"Monitoring spikes in large populations of neurons is a powerful method for studying how networks of neurons process information and produce behavior .",
"This activity can be detected using fluorescent protein indicators , or “probes” ,... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"chromosomes and gene expression",
"evolutionary biology"
] | Intrinsic cooperativity potentiates parallel cis-regulatory evolution | elife-37563-v2 | [
[
"Contingency is widespread in evolution , with chance historical changes dictating the repertoire of future possibilities ( Bershtein et al . , 2006; Bloom et al . , 2006; Blount et al . , 2012; Ortlund et al . , 2007; Sorrells et al . , 2015; Starr et al . , 2017 ) .",
"Nevertheless , repeated evolutiona... | [
"Convergent evolutionary events in independent lineages provide an opportunity to understand why evolution favors certain outcomes over others .",
"We studied such a case where a large set of genes—those coding for the ribosomal proteins—gained cis-regulatory sequences for a particular transcription regulator ( M... | [
"Sometimes evolution repeats itself .",
"For example , independent butterfly species can evolve the same warning pattern to ward off predators .",
"In many cases , the reason that a certain trait crops up again and again in parallel evolution is unknown .",
"One example is from the evolution of fungi , where ... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"biochemistry and chemical biology"
] | Mitochondrial fatty acid synthesis coordinates oxidative metabolism in mammalian mitochondria | elife-58041-v2 | [
[
"Fatty acids play diverse cellular roles , including providing the hydrophobic tails of membrane phospholipids , energy storage in the form of triglycerides , and as cell signaling molecules .",
"Aside from the import of exogenous fatty acids , mammalian cells use the well-known and well-studied cytoplasm... | [
"Cells harbor two systems for fatty acid synthesis , one in the cytoplasm ( catalyzed by fatty acid synthase , FASN ) and one in the mitochondria ( mtFAS ) .",
"In contrast to FASN , mtFAS is poorly characterized , especially in higher eukaryotes , with the major product ( s ) , metabolic roles , and cellular fun... | [
"In human , plant and other eukaryotic cells , fats are an important source of energy and also play many other roles including waterproofing , thermal insulation and energy storage .",
"Eukaryotic cells have two systems that make the building blocks of fats ( known as fatty acids ) and one of these systems , call... | 2020 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | A Cambrian origin for vertebrate rods | elife-07166-v2 | [
[
"The fossil record shows that by the middle Cambrian , camera-type eyes were already present in stem vertebrates ( Morris and Caron , 2014 ) , supporting the emerging concept that spatially resolved vision provided a major competitive advantage in those biota ( Paterson et al . , 2011 ) .",
"Lampreys , th... | [
"Vertebrates acquired dim-light vision when an ancestral cone evolved into the rod photoreceptor at an unknown stage preceding the last common ancestor of extant jawed vertebrates ( ∼420 million years ago Ma ) .",
"The jawless lampreys provide a unique opportunity to constrain the timing of this advance , as thei... | [
"The eyes of humans and many other animals with backbones contain two different types of cells that can detect light , which are known as rod and cone cells .",
"Rod cells are much more sensitive to light than cone cells .",
"The rods allow us to see in dim light by amplifying weak light signals and transmittin... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"chromosomes and gene expression"
] | Synthetic CpG islands reveal DNA sequence determinants of chromatin structure | elife-03397-v2 | [
[
"CpG islands ( CGIs ) are stretches of atypical genomic DNA sequences that mark most gene promoters ( Bird , 1986; Deaton and Bird , 2011 ) .",
"Though individually unique in nucleotide sequence , mammalian CGIs share two features that often correlate but can vary independently: a G + C-rich base composit... | [
"The mammalian genome is punctuated by CpG islands ( CGIs ) , which differ sharply from the bulk genome by being rich in G + C and the dinucleotide CpG .",
"CGIs often include transcription initiation sites and display ‘active’ histone marks , notably histone H3 lysine 4 methylation .",
"In embryonic stem cells... | [
"The building blocks of DNA are four molecules commonly named ‘A’ , ‘T’ , ‘C’ and ‘G’ .",
"The order of these DNA letters in a gene contains the instructions to make specific proteins or other molecules .",
"Other stretches of DNA contain codes that direct the cell's machinery to genes that need to be switched ... | 2014 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"developmental biology",
"immunology and inflammation"
] | The NFκB-inducing kinase is essential for the developmental programming of skin-resident and IL-17-producing γδ T cells | elife-10087-v2 | [
[
"γδ T cells , together with αβ T cells and B cells , are the only cells in mammals capable of generating diverse antigenic receptors by somatic gene rearrangement , which enables them to specifically recognize and respond to a vast array of antigens .",
"However , while the essential role of αβ T cells an... | [
"γδ T cells contribute to first line immune defense , particularly through their ability for rapid production of proinflammatory cytokines .",
"The cytokine profile of γδ T cells is hard-wired already during thymic development .",
"Yet , the molecular pathways underlying this phenomenon are incompletely underst... | [
"Our bodies are protected from infection and disease by several different types of immune cells .",
"Gamma delta T cells are unusual in that they only make up a small proportion of the immune cells of the body , yet are present in many different animal species .",
"These peculiar T cells are primarily found in ... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience",
"genetics and genomics"
] | Neuropeptide ACP facilitates lipid oxidation and utilization during long-term flight in locusts | elife-65279-v3 | [
[
"Flight is an extraordinary biological trait that has only evolved in several kinds of animals ( e . g . insects , bats , and birds ) and is notably effective in searching for food and mates , finding habitats , defending against predators , and adapting to seasonal changes in environment ( Krause and Godin ,... | [
"Long-term flight depends heavily on intensive energy metabolism in animals; however , the neuroendocrine mechanisms underlying efficient substrate utilization remain elusive .",
"Here , we report that the adipokinetic hormone/corazonin-related peptide ( ACP ) can facilitate muscle lipid utilization in a famous l... | [
"Flight allows insects to find food or seek a better environment .",
"Some insects have developed the ability of ‘long-term flight’ , which allows them to make continuous journeys over large distances .",
"For example , one locust species regularly crosses the Red Sea which is up to 300 km wide – a spectacular ... | 2021 |
[
"Introduction",
"Results and discussion",
"Materials and methods"
] | [
"evolutionary biology",
"microbiology and infectious disease"
] | A new family of cell surface located purine transporters in Microsporidia and related fungal endoparasites | elife-47037-v2 | [
[
"Microsporidia are a highly successful group of strict intracellular eukaryotic parasites that infect a broad range of animal hosts , including humans and economically important species of fish , honeybees , and silkworms ( Stentiford et al . , 2016; Vávra and Lukeš , 2013 ) .",
"Microsporidia can only co... | [
"Plasma membrane-located transport proteins are key adaptations for obligate intracellular Microsporidia parasites , because they can use them to steal host metabolites the parasites need to grow and replicate .",
"However , despite their importance , the functions and substrate specificities of most Microsporidi... | [
"Microsporidia are a group of microscopic parasites that spend part of their lives inside the cells of a broad range of animal hosts , including humans .",
"These parasites are considered to be related to fungi , some of which also live within the cells of other species and are known as fungal endoparasites .",
... | 2019 |
[
"Introduction",
"Results",
"Discussion",
"Material and methods"
] | [
"plant biology",
"developmental biology"
] | A stochastic multicellular model identifies biological watermarks from disorders in self-organized patterns of phyllotaxis | elife-14093-v1 | [
[
"Developmental systems strikingly produce regular patterns and analysis of eukaryote development has classically been focused on regularities as the main source of information to understand these complex systems .",
"However , it is becoming increasingly evident that intrinsic molecular noise is an inhere... | [
"Exploration of developmental mechanisms classically relies on analysis of pattern regularities .",
"Whether disorders induced by biological noise may carry information on building principles of developmental systems is an important debated question .",
"Here , we addressed theoretically this question using phy... | [
"Plants grow throughout their lifetime , forming new flowers and leaves at the tips of their stems through a patterning process called phyllotaxis , which occurs in spirals for a vast number of plant species .",
"The classical view suggests that the positioning of each new leaf or flower bud at the tip of a growi... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"cell biology",
"immunology and inflammation"
] | Epidermal barrier defects link atopic dermatitis with altered skin cancer susceptibility | elife-01888-v1 | [
[
"There is an ongoing debate as to whether allergic disease is a risk factor for cancer or whether it is protective , with recent studies indicating that the relationship is complex and site specific ( Arana et al . , 2010; Wedgeworth et al . , 2011; Hwang et al . , 2012 ) .",
"Several epidemiological stud... | [
"Atopic dermatitis can result from loss of structural proteins in the outermost epidermal layers , leading to a defective epidermal barrier .",
"To test whether this influences tumour formation , we chemically induced tumours in EPI−/− mice , which lack three barrier proteins—Envoplakin , Periplakin , and Involuc... | [
"Skin cancer is a common and growing problem—according to the World Health Organization , skin cancers account for one in every three cancers diagnosed world wide .",
"There is some evidence from epidemiological studies that patients with certain allergies might be protected against cancer and , in particular , t... | 2014 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"biochemistry and chemical biology",
"cell biology"
] | PPP1R15A-mediated dephosphorylation of eIF2α is unaffected by Sephin1 or Guanabenz | elife-26109-v2 | [
[
"Protein folding homeostasis ( proteostasis ) is achieved by balancing the rate of production , folding and protein degradation .",
"Proteostasis is strongly influenced by the phosphorylation state of serine 51 of the α subunit of eukaryotic translation initiation factor 2 ( eIF2α ) ( Sonenberg and Hinneb... | [
"Dephosphorylation of translation initiation factor 2 ( eIF2α ) terminates signalling in the mammalian integrated stress response ( ISR ) and has emerged as a promising target for modifying the course of protein misfolding diseases .",
"The [ ( o-chlorobenzylidene ) amino]guanidines ( Guanabenz and Sephin1 ) have... | [
"Most drugs work by tweaking the way that cells are regulated .",
"Adding or removing a phosphate group from proteins regulates many cellular decisions .",
"There are known drugs that bind to and inhibit the enzymes that add phosphate to proteins , thereby controlling various aspects of cell behaviour .",
"Ho... | 2017 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"tools and resources",
"neuroscience"
] | Scanned optogenetic control of mammalian somatosensory input to map input-specific behavioral outputs | elife-62026-v2 | [
[
"The survival of an organism depends on its ability to detect and respond appropriately to its environment .",
"Afferent neurons innervating the skin provide sensory information to guide and refine behavior ( Seymour , 2019; Zimmerman et al . , 2014 ) .",
"Cutaneous stimuli are used to study a wide ra... | [
"Somatosensory stimuli guide and shape behavior , from immediate protective reflexes to longer-term learning and higher-order processes related to pain and touch .",
"However , somatosensory inputs are challenging to control in awake mammals due to the diversity and nature of contact stimuli .",
"Application of... | [
"To safely navigate their world , animals need to be able to tell apart a gentle touch from an eye-watering pinch , detect cold water or sense the throbbing pain stemming from an infected cut .",
"These ‘somatic’ sensations are relayed through thousands of nerve endings embedded in the skin and other tissues .",
... | 2021 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"microbiology and infectious disease"
] | Regulatory roles of Escherichia coli 5' UTR and ORF-internal RNAs detected by 3' end mapping | elife-62438-v2 | [
[
"The expression of many bacterial genes is controlled by elements in the 5´ untranslated regions ( UTRs ) of mRNAs .",
"Changes in the secondary structures of these cis-acting RNA elements lead to altered expression of the associated gene ( s ) by modulating accessibility of ribosomes to sites of translat... | [
"Many bacterial genes are regulated by RNA elements in their 5´ untranslated regions ( UTRs ) .",
"However , the full complement of these elements is not known even in the model bacterium Escherichia coli .",
"Using complementary RNA-sequencing approaches , we detected large numbers of 3´ ends in 5´ UTRs and op... | [
"In most organisms , specific segments of a cell’s genetic information are copied to form single-stranded molecules of various sizes and purposes .",
"Each of these RNA molecules , as they are known , is constructed as a chain that starts at the 5´ end and terminates at the 3´ end .",
"Certain RNAs carry the in... | 2021 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"cell biology",
"immunology and inflammation"
] | Fibroblastic reticular cell-derived lysophosphatidic acid regulates confined intranodal T-cell motility | elife-10561-v1 | [
[
"Blood-borne naïve lymphocytes migrate along the fibroblastic reticular cell ( FRC ) network in lymph nodes ( LNs ) ( von Andrian and Mempel , 2003; Miyasaka and Tanaka , 2004; Girard et al . , 2012 ) .",
"B cells then migrate into LN follicles , whereas T cells remain in the paracortex and migrate contin... | [
"Lymph nodes ( LNs ) are highly confined environments with a cell-dense three-dimensional meshwork , in which lymphocyte migration is regulated by intracellular contractile proteins .",
"However , the molecular cues directing intranodal cell migration remain poorly characterized .",
"Here we demonstrate that ly... | [
"Small organs called lymph nodes are found throughout the body and help to filter out harmful particles and cells .",
"Lymph nodes are packed with different types of immune cells , such as the T-cells that play a number of roles in detecting and destroying bacteria , viruses and other disease-causing microbes .",... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"chromosomes and gene expression"
] | Histone H3G34R mutation causes replication stress, homologous recombination defects and genomic instability in S. pombe | elife-27406-v1 | [
[
"The genomic DNA of eukaryotes is packaged into chromatin , which regulates all DNA transactions including transcription , replication and repair .",
"The basic subunit of chromatin is the nucleosome , which is made up of 147 bp of DNA wrapped around a core octamer of histone proteins .",
"Proteins th... | [
"Recurrent somatic mutations of H3F3A in aggressive pediatric high-grade gliomas generate K27M or G34R/V mutant histone H3 . 3 .",
"H3 . 3-G34R/V mutants are common in tumors with mutations in p53 and ATRX , an H3 . 3-specific chromatin remodeler .",
"To gain insight into the role of H3-G34R , we generated fiss... | [
"Children suffering from a brain cancer called high-grade glioma rarely recover because there are no therapies that can effectively target this disease .",
"Recently , mutations in a gene that encodes a protein called histone H3 were found in children’s glioma cells .",
"Histone proteins bind to DNA to help pac... | 2017 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"developmental biology"
] | Functional genome-wide siRNA screen identifies KIAA0586 as mutated in Joubert syndrome | elife-06602-v2 | [
[
"A range of disorders from isolated organ defects like blindness or nephronophthisis to multi-system disorders like Joubert ( JS ) , Bardet–Biedl , or Meckel–Gruber syndromes are correlated with mutations in genes involved in formation or stability of the primary cilium ( Goetz and Anderson , 2010; Waters and... | [
"Defective primary ciliogenesis or cilium stability forms the basis of human ciliopathies , including Joubert syndrome ( JS ) , with defective cerebellar vermis development .",
"We performed a high-content genome-wide small interfering RNA ( siRNA ) screen to identify genes regulating ciliogenesis as candidates f... | [
"Joubert syndrome is a rare disorder that affects the brain and causes physical , mental , and sometimes visual impairments .",
"In individuals with this condition , two parts of the brain called the cerebellar vermis and the brainstem do not develop properly .",
"This is thought to be due to defects in the dev... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"immunology and inflammation",
"cancer biology"
] | Tumor initiating cells induce Cxcr4-mediated infiltration of pro-tumoral macrophages into the brain | elife-31918-v1 | [
[
"Microglia are the resident macrophages of the brain ( for review see [Kettenmann et al . , 2011; Casano and Peri , 2015] ) .",
"Microglia are derived from primitive macrophages that invade the brain during development ( Ginhoux et al . , 2010; Schulz et al . , 2012; Herbomel et al . , 2001 ) .",
"Due... | [
"It is now clear that microglia and macrophages are present in brain tumors , but whether or how they affect initiation and development of tumors is not known .",
"Exploiting the advantages of the zebrafish ( Danio rerio ) model , we showed that macrophages and microglia respond immediately upon oncogene activati... | [
"Brain tumors can be aggressive , difficult to treat and are often incurable .",
"Removing brain tumors by surgery can be challenging because the tumor cells infiltrate into the healthy tissue .",
"Brain tumors grow in close physical contact with other cells , such as cells of the immune system .",
"This incl... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"cell biology",
"cancer biology"
] | Negative regulation of urokinase receptor activity by a GPI-specific phospholipase C in breast cancer cells | elife-23649-v1 | [
[
"The urokinase-type plasminogen activator receptor ( uPAR ) is a central player in a complex signaling network implicated in a variety of remodeling processes , both physiological and pathological , ranging from embryo implantation to wound healing and tumor progression ( Boonstra et al . , 2011; Ferraris and... | [
"The urokinase receptor ( uPAR ) is a glycosylphosphatidylinositol ( GPI ) -anchored protein that promotes tissue remodeling , tumor cell adhesion , migration and invasion .",
"uPAR mediates degradation of the extracellular matrix through protease recruitment and enhances cell adhesion , migration and signaling t... | [
"Every process in the body , from how cells divide to how they move around , is tightly regulated .",
"For example , cells only migrate when they receive the correct signals from their environment .",
"These signals are recognised by receptor proteins that sit on the cell surface and connect the outside signal ... | 2017 |
[
"Introduction",
"Results and discussion",
"Materials and methods"
] | [
"ecology",
"tools and resources",
"genetics and genomics"
] | Viral dark matter and virus–host interactions resolved from publicly available microbial genomes | elife-08490-v3 | [
[
"Over the past two decades , our collective understanding of microbial diversity has been profoundly expanded by cultivation-independent molecular methods ( Pace , 1997; Whitman et al . , 1998; Rappé and Giovannoni , 2003; DeLong , 2009; Hanson et al . , 2012 ) .",
"It is now widely recognized that interc... | [
"The ecological importance of viruses is now widely recognized , yet our limited knowledge of viral sequence space and virus–host interactions precludes accurate prediction of their roles and impacts .",
"In this study , we mined publicly available bacterial and archaeal genomic data sets to identify 12 , 498 hig... | [
"Viruses are infectious particles that can only multiply inside the cells of microbes and other organisms .",
"Little is known about the genetic differences between virus particles ( so-called ‘genetic diversity’ ) , especially compared to what we know about the diversity of bacteria , archaea , and other single-... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"cell biology",
"neuroscience"
] | Unconventional secretory processing diversifies neuronal ion channel properties | elife-20609-v3 | [
[
"Most membrane and secreted proteins are N-glycosylated during their synthesis and processing in the secretory pathway ( Moremen et al . , 2012 ) .",
"During this process , as nascent proteins emerge in the lumen of the endoplasmic reticulum ( ER ) , a mannose-rich precursor is first transferred en bloc t... | [
"N-glycosylation – the sequential addition of complex sugars to adhesion proteins , neurotransmitter receptors , ion channels and secreted trophic factors as they progress through the endoplasmic reticulum and the Golgi apparatus – is one of the most frequent protein modifications .",
"In mammals , most organ-spe... | [
"Information is carried around the nervous system by cells called neurons .",
"The ability of neurons to communicate with each other relies on many proteins that are found on the surfaces of the cells .",
"Like in all animal cells , surface proteins are made inside the cell in a compartment called the endoplasm... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"ecology",
"microbiology and infectious disease"
] | The diversity and function of sourdough starter microbiomes | elife-61644-v1 | [
[
"Sourdough bread is a globally distributed fermented food that is made using a microbial community of yeasts and bacteria .",
"The sourdough microbiome is maintained in a starter that is used to inoculate dough for bread production ( Figure 1A ) .",
"Yeasts , lactic acid bacteria ( LAB ) , and acetic ... | [
"Humans have relied on sourdough starter microbial communities to make leavened bread for thousands of years , but only a small fraction of global sourdough biodiversity has been characterized .",
"Working with a community-scientist network of bread bakers , we determined the microbial diversity of 500 sourdough ... | [
"Sourdough bread is an ancient fermented food that has sustained humans around the world for thousands of years .",
"It is made from a sourdough ‘starter culture’ which is maintained , portioned , and shared among bread bakers around the world .",
"The starter culture contains a community of microbes made up of... | 2021 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"structural biology and molecular biophysics"
] | Multivalency regulates activity in an intrinsically disordered transcription factor | elife-36258-v3 | [
[
"Regulation of transcriptional activity is essential for every biological process .",
"Common mechanisms for regulation include post-translational modifications and/or cooperativity among multiple activators and repressors ( Banerjee and Zhang , 2003 ) .",
"Some transcription factors contain multiple ... | [
"The transcription factor ASCIZ ( ATMIN , ZNF822 ) has an unusually high number of recognition motifs for the product of its main target gene , the hub protein LC8 ( DYNLL1 ) .",
"Using a combination of biophysical methods , structural analysis by NMR and electron microscopy , and cellular transcription assays , ... | [
"Proteins help to regulate almost every process in the body , and come in various forms , sizes and purposes .",
"Cells contain thousands of different proteins , but not every protein is needed at all times .",
"To create new proteins , the information on a gene first needs to be transcribed into RNA ( template... | 2018 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | The neural basis for a persistent internal state in Drosophila females | elife-59502-v7 | [
[
"Social behaviors are known to be affected by persistent internal states ( Anderson , 2016; Berridge , 2004; Lorenz and Leyhausen , 1973 ) .",
"These states correspond with levels of arousal or drive , and can impact whether and how individuals interact , with consequences for mating decisions and reprodu... | [
"Sustained changes in mood or action require persistent changes in neural activity , but it has been difficult to identify the neural circuit mechanisms that underlie persistent activity and contribute to long-lasting changes in behavior .",
"Here , we show that a subset of Doublesex+ pC1 neurons in the Drosophil... | [
"Long-term mental states such as arousal and mood variations rely on persistent changes in the activity of certain neural circuits which have been difficult to identify .",
"For instance , in male fruit flies , the activation of a particular circuit containing ‘P1 neurons’ can escalate aggressive and mating behav... | 2020 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"research communication",
"neuroscience"
] | Oxytocin promotes coordinated out-group attack during intergroup conflict in humans | elife-40698-v3 | [
[
"Conflict between groups of people can profoundly change their demographic , economic and cultural outlook .",
"Across human history , intergroup conflict functioned as a critical change agent and selection pressure that may have shaped the biological preparedness for in-group oriented cooperation and sel... | [
"Intergroup conflict contributes to human discrimination and violence , but persists because individuals make costly contributions to their group’s fighting capacity .",
"Yet how group members effectively coordinate their contributions during intergroup conflict remains poorly understood .",
"Here we examine th... | [
"Conflict between groups is a recurring theme in human history .",
"We tend to form social bonds with others who share the same characteristics as ourselves , whether that is nationality , ethnicity , or supporting the same football team .",
"Individuals that belong to the same group as us comprise our ‘in-grou... | 2019 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"plant biology",
"evolutionary biology"
] | Independent evolution of ancestral and novel defenses in a genus of toxic plants (Erysimum, Brassicaceae) | elife-51712-v2 | [
[
"Plant chemical defenses play a central role in the coevolutionary arms race with herbivorous insects .",
"In response to diverse environmental challenges , plants have evolved a plethora of structurally diverse organic compounds with repellent , antinutritive , or toxic properties ( Fraenkel , 1959; Mith... | [
"Phytochemical diversity is thought to result from coevolutionary cycles as specialization in herbivores imposes diversifying selection on plant chemical defenses .",
"Plants in the speciose genus Erysimum ( Brassicaceae ) produce both ancestral glucosinolates and evolutionarily novel cardenolides as defenses .",... | [
"Plants are often attacked by insects and other herbivores .",
"As a result , they have evolved to defend themselves by producing many different chemicals that are toxic to these pests .",
"As producing each chemical costs energy , individual plants often only produce one type of chemical that is targeted towar... | 2020 |
[
"Introduction",
"Results and discussion",
"Methods"
] | [
"evolutionary biology",
"short report",
"computational and systems biology"
] | Further support for aneuploidy tolerance in wild yeast and effects of dosage compensation on gene copy-number evolution | elife-14409-v2 | [
[
"In our previous work ( Hose et al . , 2015 ) , we reported from a genome-sequencing survey that aneuploidy was frequently observed in wild strains of S . cerevisiae , and that 10–30% of amplified genes , depending on the strain and affected chromosome , show lower-than-expected expression in naturally aneupl... | [
"In our prior work by Hose et al . , we performed a genome-sequencing survey and reported that aneuploidy was frequently observed in wild strains of S . cerevisiae .",
"We also profiled transcriptome abundance in naturally aneuploid isolates compared to isogenic euploid controls and found that 10–30% of amplified... | [
"Cells package their DNA into structures called chromosomes .",
"Sometimes when a cell divides , it fails to allocate the right number of chromosomes to each new cell and so they end up with too many or too few chromosomes .",
"The extra copies of the genes on an additional chromosome can be harmful to the cell... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"biochemistry and chemical biology",
"cell biology",
"tools and resources"
] | AirID, a novel proximity biotinylation enzyme, for analysis of protein–protein interactions | elife-54983-v2 | [
[
"Many cellular proteins function under the control of biological regulatory systems .",
"Protein–protein interactions ( PPIs ) comprise part of the biological regulation system for proteins .",
"Besides PPIs , biological protein function is post-translationally promoted by multiple modifications such ... | [
"Proximity biotinylation based on Escherichia coli BirA enzymes such as BioID ( BirA* ) and TurboID is a key technology for identifying proteins that interact with a target protein in a cell or organism .",
"However , there have been some improvements in the enzymes that are used for that purpose .",
"Here , we... | [
"Proteins in a cell need to interact with each other to perform the many tasks required for organisms to thrive .",
"A technique called proximity biotinylation helps scientists to pinpoint the identity of the proteins that partner together .",
"It relies on attaching an enzyme ( either BioID or TurboID ) to a p... | 2020 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"chromosomes and gene expression"
] | The molecular basis of coupling between poly(A)-tail length and translational efficiency | elife-66493-v1 | [
[
"Most eukaryotic mRNAs are polyadenylated at their 3′ ends in a process associated with transcriptional termination .",
"In the nucleus , these poly ( A ) tails can facilitate mRNA nucleocytoplasmic export ( Kühn and Wahle , 2004 ) , whereas in the cytoplasm , they serve as molecular timers for mRNA decay... | [
"In animal oocytes and early embryos , mRNA poly ( A ) -tail length strongly influences translational efficiency ( TE ) , but later in development this coupling between tail length and TE disappears .",
"Here , we elucidate how this coupling is first established and why it disappears .",
"Overexpressing cytopla... | [
"Cells are microscopic biological factories that are constantly creating new proteins .",
"To do so , a cell must first convert its master genetic blueprint , the DNA , into strands of messenger RNA or mRNA .",
"These strands are subsequently translated to make proteins .",
"Cells have two ways to adjust the ... | 2021 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Synaptic organization of the Drosophila antennal lobe and its regulation by the Teneurins | elife-03726-v1 | [
[
"Understanding how neural circuits process information requires knowing not only the neuroanatomical connectivity and electrophysiological properties of individual neurons , but also the organization of synapses between specific neuronal types .",
"Knowledge of these basic organizational principles , as w... | [
"Understanding information flow through neuronal circuits requires knowledge of their synaptic organization .",
"In this study , we utilized fluorescent pre- and postsynaptic markers to map synaptic organization in the Drosophila antennal lobe , the first olfactory processing center .",
"Olfactory receptor neur... | [
"Just as progress in science relies on researchers communicating their findings to other people working in their field , our bodies rely on neurons being able to communicate with other neurons .",
"This is where structures called synapses come in: synapses allow signals to be passed from one neuron to another .",... | 2014 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Decoupling sensory from decisional choice biases in perceptual decision making | elife-43994-v2 | [
[
"You ask a friend about the tilt of a canvas that is perfectly horizontal and she says that the top right corner is up .",
"What causes her inaccuracy ?",
"One possibility is that her sensory representation is biased and she perceives the canvas tilted .",
"Another possibility , however , is that ... | [
"The contribution of sensory and decisional processes to perceptual decision making is still unclear , even in simple perceptual tasks .",
"When decision makers need to select an action from a set of balanced alternatives , any tendency to choose one alternative more often—choice bias—is consistent with a bias in... | [
"Imagine that every day , you split a chocolate bar into two and offer one half to your friend .",
"Even though you take care to divide the bar into equal pieces , your friend nearly always chooses the left half .",
"Why is that ?",
"One possibility is that sensory bias in her visual system makes her perceive... | 2019 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"cell biology"
] | Cardiac mitochondrial function depends on BUD23 mediated ribosome programming | elife-50705-v2 | [
[
"Regulation of protein abundance within a cell is of fundamental importance to homeostasis , cell identity and responsiveness to changes in external environment .",
"Many layers of regulatory control have evolved to fine-tune gene expression to meet cellular needs , including transcriptional regulation , ... | [
"Efficient mitochondrial function is required in tissues with high energy demand such as the heart , and mitochondrial dysfunction is associated with cardiovascular disease .",
"Expression of mitochondrial proteins is tightly regulated in response to internal and external stimuli .",
"Here we identify a novel m... | [
"Cells need to make proteins to survive , so they have protein-making machines called ribosomes .",
"Ribosomes are themselves made out of proteins and RNA ( a molecule similar to DNA ) , and they are assembled by other proteins that bring ribosomal components together and modify them until the ribosomes are funct... | 2020 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"ecology",
"neuroscience"
] | Burst muscle performance predicts the speed, acceleration, and turning performance of Anna’s hummingbirds | elife-11159-v2 | [
[
"The ability of an animal to change the speed and direction of movement , defined as maneuverability ( Dudley , 2002 ) , can determine its success at avoiding predators , obtaining food , and performing other behaviors that determine the margin between life and death ( Webb , 1976; Hedenström and Rosén , 2001... | [
"Despite recent advances in the study of animal flight , the biomechanical determinants of maneuverability are poorly understood .",
"It is thought that maneuverability may be influenced by intrinsic body mass and wing morphology , and by physiological muscle capacity , but this hypothesis has not yet been evalua... | [
"The ability of an animal to maneuver can determine its success at avoiding predators , catching prey , and outperforming its competitors .",
"However , little is known about the characteristics that determine maneuverability .",
"Why are some individuals more maneuverable than others ?",
"To investigate this... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"structural biology and molecular biophysics"
] | Cryo-EM structures of the autoinhibited E. coli ATP synthase in three rotational states | elife-21598-v4 | [
[
"In most cells , the bulk of ATP , the principal source of cellular energy , is synthesized by ATP synthase .",
"This molecular generator couples ion flow across membranes with the addition of inorganic phosphate ( Pi ) to ADP thereby generating ATP ( Iino and Noji , 2013; Stewart et al . , 2014 ) .",
... | [
"A molecular model that provides a framework for interpreting the wealth of functional information obtained on the E . coli F-ATP synthase has been generated using cryo-electron microscopy .",
"Three different states that relate to rotation of the enzyme were observed , with the central stalk’s ε subunit in an ex... | [
"ATP synthase is a biological motor that produces a molecule called adenosine tri-phosphate ( ATP for short ) , which acts as the major store of chemical energy in cells .",
"A single molecule of ATP contains three phosphate groups: the cell can remove one of these phosphates to make a molecule called adenosine d... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"cell biology",
"neuroscience"
] | Modeling of axonal endoplasmic reticulum network by spastic paraplegia proteins | elife-23882-v2 | [
[
"Axons allow long-range bidirectional communication in neurons .",
"They carry action potentials along their plasma membrane from dendrites and cell body to presynaptic terminals; and they transport cell components and signaling complexes using motor proteins , anterogradely and retrogradely .",
"Anot... | [
"Axons contain a smooth tubular endoplasmic reticulum ( ER ) network that is thought to be continuous with ER throughout the neuron; the mechanisms that form this axonal network are unknown .",
"Mutations affecting reticulon or REEP proteins , with intramembrane hairpin domains that model ER membranes , cause an ... | [
"The way we move – from simple motions like reaching out to grab something , to playing the piano or dancing – is coordinated in our brain .",
"These processes involve many regions and steps , in which nerve cells transport signals along projections known as axons .",
"Axons rely on sophisticated ‘engineering’ ... | 2017 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"developmental biology",
"neuroscience"
] | Large-scale phenotypic drug screen identifies neuroprotectants in zebrafish and mouse models of retinitis pigmentosa | elife-57245-v4 | [
[
"Inherited retinal diseases ( IRDs ) encompass a group of genetically linked retinopathies characterized by progressive photoreceptor death ( Duncan et al . , 2018 ) .",
"IRDs lead to irreversible vision loss , for which treatment strategies are limited .",
"Retinitis pigmentosa ( RP ) , the most comm... | [
"Retinitis pigmentosa ( RP ) and associated inherited retinal diseases ( IRDs ) are caused by rod photoreceptor degeneration , necessitating therapeutics promoting rod photoreceptor survival .",
"To address this , we tested compounds for neuroprotective effects in multiple zebrafish and mouse RP models , reasonin... | [
"Photoreceptors are the cells responsible for vision .",
"They are part of the retina: the light-sensing tissue at the back of the eye .",
"They come in two types: rods and cones .",
"Rods specialise in night vision , while cones specialise in daytime colour vision .",
"The death of these cells can cause a ... | 2021 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"developmental biology",
"genetics and genomics"
] | Deployment of a retinal determination gene network drives directed cell migration in the sea urchin embryo | elife-08827-v2 | [
[
"Much research has been done to understand the cell biology underlying the events of directed cell migration; however , how the events are encoded in the genome and how gene regulatory networks ( GRNs ) control this process are works in progress .",
"Cell migration is the directed movement of a single cel... | [
"Gene regulatory networks ( GRNs ) provide a systems-level orchestration of an organism's genome encoded anatomy .",
"As biological networks are revealed , they continue to answer many questions including knowledge of how GRNs control morphogenetic movements and how GRNs evolve .",
"The migration of the small m... | [
"Within an animal embryo , groups of cells tend to move , or migrate , between different areas before they form into tissues and organs .",
"These cell migrations are regulated by hundreds of genes , which must be expressed at the right time and in the right place .",
"Cells use proteins called transcription fa... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"biochemistry and chemical biology"
] | A chemical probe of CARM1 alters epigenetic plasticity against breast cancer cell invasion | elife-47110-v3 | [
[
"Numerous biological events are orchestrated epigenetically upon defining cellular fates ( Atlasi and Stunnenberg , 2017; Berdasco and Esteller , 2019 ) .",
"Among the key epigenetic regulators are protein methyltransferases ( PMTs ) , which can render downstream signals by modifying specific Arg or Lys r... | [
"CARM1 is a cancer-relevant protein arginine methyltransferase that regulates many aspects of transcription .",
"Its pharmacological inhibition is a promising anti-cancer strategy .",
"Here SKI-73 ( 6a in this work ) is presented as a CARM1 chemical probe with pro-drug properties .",
"SKI-73 ( 6a ) can rapidl... | [
"Drugs that are small molecules have the potential to block the individual proteins that drive the spread of cancer , but their design is a challenge .",
"This is because they need to get inside the cell and find their target without binding to other proteins on the way .",
"However , small molecule drugs often... | 2019 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"structural biology and molecular biophysics",
"microbiology and infectious disease"
] | Massive antibody discovery used to probe structure–function relationships of the essential outer membrane protein LptD | elife-46258-v1 | [
[
"The outer membrane ( OM ) of Gram-negative bacteria is a permeability barrier to antibiotics and other cytotoxic agents , such as detergents ( Nikaido , 2003 ) .",
"A key feature of the OM is its distinctive asymmetrical bilayer populated with lipopolysaccharide ( LPS ) in the outer leaflet and phospholi... | [
"Outer membrane proteins ( OMPs ) in Gram-negative bacteria dictate permeability of metabolites , antibiotics , and toxins .",
"Elucidating the structure-function relationships governing OMPs within native membrane environments remains challenging .",
"We constructed a diverse library of >3000 monoclonal antibo... | [
"The overuse and misuse of antibiotics has led to the rise of multi-drug resistant bacteria which threaten global public health .",
"Antibiotics interfere with essential processes in bacteria so they are unable to divide or survive , but over time , the microbes have found ways to become immune to the drugs .",
... | 2019 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"developmental biology",
"neuroscience"
] | The Drosophila Sp8 transcription factor Buttonhead prevents premature differentiation of intermediate neural progenitors | elife-03596-v2 | [
[
"Intermediate neural progenitor cells ( INPs ) play a critical role in increasing the brain size and complexity .",
"Transient amplification of INPs dramatically boosts the neural output from neural stem cells ( NSCs ) ( Kriegstein and Alvarez-Buylla , 2009; Florio and Huttner , 2014 ) .",
"Recent stu... | [
"Intermediate neural progenitor cells ( INPs ) need to avoid differentiation and cell cycle exit while maintaining restricted developmental potential , but mechanisms preventing differentiation and cell cycle exit of INPs are not well understood .",
"In this study , we report that the Drosophila homolog of mammal... | [
"Whereas the majority of cells in the brain are unable to divide to produce new cells , neural stem cells can divide numerous times and have the potential to become many different types of brain cells .",
"However , in between these two extremes there is another group of cells called neural progenitors .",
"The... | 2014 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"structural biology and molecular biophysics"
] | Molecular architecture of the yeast Mediator complex | elife-08719-v2 | [
[
"Mediator is a complex of at least 21 subunits , with a mass of greater than 1 million Daltons , conserved from yeast to humans ( reviewed in Kornberg , 2005; Conaway and Conaway , 2011; Ansari and Morse , 2013 ) .",
"Genetic studies have shown that Mediator is essential for RNA polymerase II ( pol II ) t... | [
"The 21-subunit Mediator complex transduces regulatory information from enhancers to promoters , and performs an essential role in the initiation of transcription in all eukaryotes .",
"Structural information on two-thirds of the complex has been limited to coarse subunit mapping onto 2-D images from electron mic... | [
"Inside a cell , proteins are made from instructions encoded by DNA .",
"To produce a particular protein , a section of DNA within a gene is copied into a molecule of messenger ribonucleic acid ( or mRNA ) .",
"This process is called transcription and is carried out by an enzyme known as RNA polymerase .",
"T... | 2015 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"neuroscience"
] | Working memory capacity of crows and monkeys arises from similar neuronal computations | elife-72783-v2 | [
[
"Working memory ( WM ) can hold information for a short period of time to allow further processing in the absence of sensory input ( Cowan , 2017; Oberauer et al . , 2018 ) .",
"By bridging this gap between the immediate sensory environment and behavior , WM is a cornerstone for complex cognition .",
... | [
"Complex cognition relies on flexible working memory , which is severely limited in its capacity .",
"The neuronal computations underlying these capacity limits have been extensively studied in humans and in monkeys , resulting in competing theoretical models .",
"We probed the working memory capacity of crows ... | [
"Working memory is the brain’s ability to temporarily hold and manipulate information .",
"It is essential for carrying out complex cognitive tasks , such as reasoning , planning , following instructions or solving problems .",
"Unlike long-term memory , information is not stored and recalled , but held in an a... | 2021 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"epidemiology and global health",
"immunology and inflammation"
] | The complex relationship of exposure to new Plasmodium infections and incidence of clinical malaria in Papua New Guinea | elife-23708-v2 | [
[
"Renewed emphasis on malaria control has resulted in substantial reductions in overall malaria prevalence and incidence in many endemic countries ( World Health Organization , 2015 ) .",
"However , where transmission persists , it is highly heterogeneous even on small spatial scales ( Bousema et al . , 20... | [
"The molecular force of blood-stage infection ( molFOB ) is a quantitative surrogate metric for malaria transmission at population level and for exposure at individual level .",
"Relationships between molFOB , parasite prevalence and clinical incidence were assessed in a treatment-to-reinfection cohort , where P ... | [
"Malaria is caused by five different species of parasites that are transmitted to humans by bites from parasite-carrying mosquitos .",
"Once in human blood , the parasites rapidly multiply .",
"People who live in countries where malaria is common may become infected and never show any symptoms because their imm... | 2017 |
[
"Introduction",
"Results and discussion",
"Materials and methods"
] | [
"cell biology",
"physics of living systems"
] | Cortical flow aligns actin filaments to form a furrow | elife-17807-v3 | [
[
"Cytokinesis begins in late anaphase , triggered by regulatory pathways with feedback from the mitotic spindle that ensure appropriate positioning of the molecular machinery ( reviewed in Fededa and Gerlich , 2012; Green et al . , 2012 ) .",
"As a consequence , the small regulatory GTPase RhoA ( Piekny et... | [
"Cytokinesis in eukaryotic cells is often accompanied by actomyosin cortical flow .",
"Over 30 years ago , Borisy and White proposed that cortical flow converging upon the cell equator compresses the actomyosin network to mechanically align actin filaments .",
"However , actin filaments also align via search-an... | [
"Just under the surface of every animal cell , a thin and dynamic network of filaments called the cell cortex acts as a scaffold and determines the cell’s shape .",
"When the cell divides , this material re-organizes to make a ring of filaments – known as the cytokinetic ring – across the middle of the cell .",
... | 2016 |
[
"Introduction",
"Results",
"Discussion",
"Materials and methods"
] | [
"chromosomes and gene expression",
"computational and systems biology"
] | Simple biochemical features underlie transcriptional activation domain diversity and dynamic, fuzzy binding to Mediator | elife-68068-v3 | [
[
"Transcription factors ( TFs ) perform the last step in signal transduction pathways .",
"They thus serve key roles in central processes such as growth , stress response , and development , and their mutation or misregulation underlies many human diseases ( Spitz and Furlong , 2012 ) .",
"A TF include... | [
"Gene activator proteins comprise distinct DNA-binding and transcriptional activation domains ( ADs ) .",
"Because few ADs have been described , we tested domains tiling all yeast transcription factors for activation in vivo and identified 150 ADs .",
"By mRNA display , we showed that 73% of ADs bound the Med15... | [
"Cells adapt and respond to changes by regulating the activity of their genes .",
"To turn genes on or off , they use a family of proteins called transcription factors .",
"Transcription factors influence specific but overlapping groups of genes , so that each gene is controlled by several transcription factors... | 2021 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.