UBC Faculty Research and Publications

Coordinated increase in inhibitory and excitatory synapses onto retinal ganglion cells during development Soto, Florentina; Bleckert, Adam; Lewis, Renate; Kang, Yunhee; Kerschensteiner, Daniel; Craig, Ann M; Wong, Rachel O Aug 24, 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13064_2011_Article_176.pdf [ 1.81MB ]
JSON: 52383-1.0223919.json
JSON-LD: 52383-1.0223919-ld.json
RDF/XML (Pretty): 52383-1.0223919-rdf.xml
RDF/JSON: 52383-1.0223919-rdf.json
Turtle: 52383-1.0223919-turtle.txt
N-Triples: 52383-1.0223919-rdf-ntriples.txt
Original Record: 52383-1.0223919-source.json
Full Text

Full Text

Coordinated increase in inhibitory and excitatorysynapses onto retinal ganglion cells duringdevelopmentSoto et al.Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31 (24 August 2011)RESEARCH ARTICLE Open AccessCoordinated increase in inhibitory and excitatorysynapses onto retinal ganglion cells duringdevelopmentFlorentina Soto1,2, Adam Bleckert1, Renate Lewis3, Yunhee Kang4, Daniel Kerschensteiner1,2, Ann Marie Craig4 andRachel OL Wong1*AbstractBackground: Neuronal output is shaped by a balance of excitation and inhibition. How this balance is attained inthe central nervous system during development is not well understood, and is complicated by the fact that, invivo, GABAergic and glycinergic synaptogenesis precedes that of glutamatergic synapses. Here, we determined thedistributions of inhibitory postsynaptic sites on the dendritic arbors of individual neurons, and compared theirdevelopmental patterns with that of excitatory postsynaptic sites. We focused on retinal ganglion cells (RGCs), theoutput neurons of the retina, which receive excitatory input from bipolar cells and inhibitory input from amacrinecells. To visualize and map inhibitory postsynaptic sites, we generated transgenic mice in which RGCs expressfluorescently tagged Neuroligin 2 (YFP-NL2) under the control of the Thy1 promoter. By labeling RGC dendritesbiolistically in YFP-NL2-expressing retinas, we were able to map the spatial distribution and thus densities ofinhibitory postsynaptic sites on the dendritic arbors of individual large-field RGCs across ages.Results: We demonstrate that YFP-NL2 is present at inhibitory synapses in the inner plexiform layer by its co-localization with gephyrin, the g2 subunit of the GABAA receptor and glycine receptors. YFP-NL2 puncta wereapposed to the vesicular inhibitory transmitter transporter VGAT but not to CtBP2, a marker of presynaptic ribbonsfound at bipolar cell terminals. Similar patterns of co-localization with synaptic markers were observed forendogenous NL2. We also verified that expression of YFP-NL2 in the transgenic line did not significantly alterspontaneous inhibitory synaptic transmission onto RGCs. Using these mice, we found that, on average, the densityof inhibitory synapses on individual arbors increased gradually until eye opening (postnatal day 15). A small centro-peripheral gradient in density found in mature arbors was apparent at the earliest age we examined (postnatal day8). Unexpectedly, the adult ratio of inhibitory/excitatory postsynaptic sites was rapidly attained, shortly afterglutamatergic synaptogenesis commenced (postnatal day 7).Conclusion: Our observations suggest that bipolar and amacrine cell synaptogenesis onto RGCs appearcoordinated to rapidly attain a balanced ratio of excitatory and inhibitory synapse densities prior to the onset ofvisual experience.BackgroundThe normal functioning of the nervous system requiresbalanced excitatory and inhibitory neurotransmission. Ifexcitation or inhibition is perturbed, neurons undergoalterations in their intrinsic excitability and synaptictransmission in order to restore a balance, and preventtheir circuits from undergoing epileptiform activity [1,2].While such homeostatic plasticity in mature neuronalnetworks is well studied [3], much less is known con-cerning how balanced excitation and inhibition is nor-mally achieved during development. In many parts ofthe central nervous system (CNS), interneurons contain-ing the classical inhibitory transmitters g-aminobutyricacid (GABA) or glycine form functional synaptic con-nections well before glutamatergic synapses emerge[4,5]. GABAergic or glycinergic synaptogenesis may thus* Correspondence: wongr2@u.washington.edu1Department of Biological Structure, University of Washington, 1950 PacificAve, Seattle, WA 98195, USAFull list of author information is available at the end of the articleSoto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31© 2011 Soto et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.outpace glutamatergic synaptogenesis, requiringmechanisms to adjust excitation and inhibition toachieve a balance throughout development. In contrast,it is possible that inhibitory and excitatory synapsesonto a given neuron develop largely in parallel, main-taining a constant ratio of synapse densities at all stagesshortly after glutamatergic synaptogenesis begins. Here,we distinguished between these two possibilities bycomparing the densities of inhibitory and excitatorysynapses on the dendritic arbors of cells of the sametype during the period of synaptogenesis.Many studies have tracked the distribution of gluta-matergic synapses on a given neuron by imaging spines[6-9] or by visualizing fluorescently labeled postsynapticdensities or receptors on the dendrites [10-15]. How-ever, spatial maps of inhibitory synapses across the den-dritic arbor of individual neurons during in vivodevelopment have not been charted. Because matureinhibitory synapse distributions vary across cell types[16] and even across the arbor of individual neurons[17], it is important to obtain and compare inhibitoryand excitatory synapse distributions on the dendrites ofthe same cell type. Here, we focused on retinal ganglioncells (RGCs) of the vertebrate retina because the phy-siology and morphology of these neurons are generallywell studied. We mapped the spatial distributions ofinhibitory postsynaptic sites on mouse RGCs and com-pared their developmental distributions with those ofglutamatergic postsynaptic sites [14].Inhibition in the inner retina is provided by amacrineinterneurons that primarily use either GABA or glycineas their neurotransmitter [18,19]. Serial electron micro-scopy (EM) of dendritic trees of RGCs in adult maca-que, cat and rabbit revealed that amacrine cellscontribute a significant fraction of the total number ofsynapses onto a RGC [20-24]. This fraction, however,differs greatly across species, and even amongst RGCsubtypes within a species. For example, in cat, amacrinecells make about 30 to 60% of all synapses onto beta-RGCs [25,26] but about 80 to 86% of synapses ontoalpha-RGCs [22,26]. However, the ratio of amacrine tobipolar cell synapses appears to be consistent within aRGC subtype [25]. Combined immunolabeling for iono-tropic glutamate receptors and gephyrin recentlydemonstrated relatively similar densities of excitatoryand inhibitory synapses on individually labeled smallbistratified [27] and sparse or thorny arbored [28] RGCsof the primate retina.The ratio of amacrine to bipolar cell synapses is notknown for developing RGCs. EM of the mouse retinasuggested that conventional (presumed amacrine)synapses are present shortly after birth, and ribbon(bipolar) synapses appear only after postnatal day (P)10[29]. Conventional synapse density increases rapidly asribbon synapses form. This increase in conventionalsynapses may represent increased amacrine synaptogen-esis onto RGC dendrites and/or synaptogenesis largelyonto the newly formed axonal terminals of differentiat-ing bipolar cells [30]. However, it is also possible thatearly in development, conventional synapses includebipolar synapses that have not yet localized ribbons totheir presynaptic sites [31,32]. Thus, in order to distin-guish amacrine and bipolar cell synapses onto develop-ing RGCs, excitatory and inhibitory postsynapticmarkers are necessary.We previously expressed PSD95 fluorescently taggedwith yellow fluorescent protein (PSD95-YFP) in RGCsto label their glutamatergic postsynaptic densities [14].Here, we visualized inhibitory postsynaptic sites onRGCs in transgenic mice in which Neuroligin 2 (NL2)was tagged with YFP (YFP-NL2). Neuroligins are postsy-naptic cell adhesion molecules that interact with presy-naptic neurexins and are essential for normal synapticmaturation and function [33]. NL2 is selectivelyexpressed at inhibitory synapses in the CNS [34,35].Because NL2 protein is present at early stages of differ-entiating inhibitory synapses [35,36], we used YFP-NL2-expressing transgenic mice as a means of identifyinginhibitory postsynaptic sites on the RGCs. We labeledthe dendrites of large-field RGCs and mapped the spa-tial distribution of YFP-NL2 puncta on their dendriticarbors across different ages. We found that the densityof inhibitory postsynaptic sites on large-field RGCsincreases gradually with maturation, following a timecourse similar to that for their glutamatergic postsynap-tic sites. Interestingly, the ratio of YFP-NL2 and PSD95-YFP puncta density per cell remained constant shortlyafter bipolar cell synapses form, suggesting that excita-tory and inhibitory synaptogenesis onto RGCs may becoordinated, perhaps to ensure that neuronal excitabilityis suitably regulated and stable throughout development.ResultsNeuroligin 2 localizes to inhibitory postsynaptic sites inthe retinaIn order to compare the distribution of fluorescentlytagged NL2 in our transgenic mouse lines with theendogenous expression of NL2, we first carried outimmunostaining of endogenous NL2 and inhibitorypostsynaptic markers in wild-type retina. In the rodentretina, gephyrin appears to be present mostly at glyci-nergic synapses and at a subset of GABAergic synapsesmainly containing the a2 subunit of the GABAA recep-tor [37]. Figure 1A shows immunolabeling for endogen-ous NL2 and gephryin in the inner plexiform layer (IPL)of a P21 retina, and illustrates the co-localization ofthese two postsynaptic proteins. To ascertain whetherNL2 is present at both glycinergic and GABAergicSoto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 2 of 14synapses in the IPL, we combined immunolabeling forNL2 with immunostaining using an antibody against theg2 subunit of the GABAA receptor (gift of JM Fristchy[38]), or an antibody that recognizes all subunits of theglycine receptor (mAb4a) [39]. Similar to previousobservations [40], we found that in the IPL, NL2 coloca-lized with the g2 subunit of the GABAA receptor, andwith glycine receptors to some extent (Figure 1B).We performed two-dimensional cross-correlation ana-lysis to determine whether colocalization between twolabels was real or random (see Materials and methods).Our results indicate that there is a positive correlationof endogenous NL2 signal and immunolabeled gephyrin,GABAA and glycine receptors (Figure 1C). We con-firmed that such correlations were not an artifact of theanalysis by repeating the calculation with one channelrotated by 180°. Hoon et al. [40] demonstrated thatendogenous NL2 does not colocalize with the scaffold-ing protein PSD95, found at excitatory postsynaptic sitesin the retinal IPL. We similarly found that carboxy-terminal binding protein 2 (CtBP2), a marker of bipolarcell ribbons at glutamatergic presynaptic release sites,was not apposed to endogenous NL2 clusters (Figure1B, C). Taken together, our observations concur withpast observations [40] indicating that NL2 is a suitablemarker of inhibitory postsynaptic sites in the retinal IPL.YFP-NL2 expression in the transgenic retina matches thedistribution of the endogenous proteinIn order to determine the distribution of inhibitory post-synaptic sites on individual RGC dendritic arbors, weperformed our study on a transgenic line in which theThy1 promoter drives expression of NL2 fused to YFP.Retinas from Tg(Thy1-YFP-NL2) mice showed punctateexpression of YFP-NL2 in a large population of cells inthe ganglion cell layer. Bright YFP puncta were presentalong dendrites of the RGCs, as well as on their somata(Figure 2B). Although many cells in the ganglion cellFigure 1 NL2 colocalizes with markers of inhibitory synapses in wild-type (WT) mice. (A, B) Single plane confocal images of P21 to P25vertical sections of WT retina labeled with antibodies against NL2 and the inhibitory or excitatory synapse markers gephyrin, g2 subunit ofGABAA receptors or glycine receptors (GlyR) and ribbons (carboxy-terminal binding protein 2 (CtBP2)). IPL, inner plexiform layer. Highermagnification views (smaller panels) are shown for the boxed regions in (A, B). (C) Two-dimensional cross-correlation coefficients of pixelintensities in the red and green channels are plotted to demonstrate whether the signals show non-random colocalization (peak at 0,0), orrandom colocalization (flat distribution). See Materials and methods for analysis. As a control for random colocalization of signals, correlationplots were obtained after the red channel was rotated 180° relative to the green (NL2) channel.Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 3 of 14layer brightly expressed YFP-tagged NL2 in our trans-genic line, expression in neurons whose somata locatedin the inner nuclear layer, including amacrine and bipo-lar cells, was scarce and very dim (Figure 2C). Thus, notall inner retinal cells containing native NL2 show trans-genic expression of YFP-NLG2 in this transgenic line.YFP-NL2 puncta were apparent in the IPL at P8 andlocalized to this synaptic layer throughout neonataldevelopment and at maturity (Figure 2C). Diffuse intra-cellular staining was observed in RGCs during early neo-natal development but this expression became lessapparent by P21. Faint expression was also observed inhorizontal cells in the outer retina at P8 but this expres-sion disappeared by P21.To determine whether the subcellular localization ofYFP-NL2 matched the distribution of endogenous NL2in the mature retina, we first performed immunolabelingfor pre- and postsynaptic markers of inhibitory synapseson vertical sections of the transgenic retina at P21 toP24. As observed for endogenous NL2, YFP-NL2 wasfound to colocalize with gephyrin, the g2 subunit ofGABAA receptors and glycine receptors (Figure 3A). Inaddition, YFP-NL2 clusters were apposed to the inhibi-tory presynaptic marker, vesicular GABA and glycinetransporter (VGAT), but not the presynaptic marker ofglutamatergic synapses CtBP2 (Figure 3B). Similarly,immunolabeling for the same set of markers at P10demonstrates that YFP-NL2 is already appropriatelyFigure 2 Punctate expression is observed in the cell bodies in the ganglion cell layer (GCL) and in the inner plexiform layer (IPL) ofThy1-YFP-NL2 retinas. (A) En face (top) view of a P21 Thy1-YFP-NL2 retina at the level of the RGC layer or at the IPL. (B) High magnificationview of YFP-NL2 puncta in the GCL and IPL. Arrows indicate puncta on a RGC soma. (A, B) Maximal intensity projection encompassing 4.5 μmthickness. (C) Vertical slices through Thy1-YFP-NL2 retinas at various ages, showing punctate staining in the IPL. Faint expression in horizontalcells (HCs) was observed transiently during development. INL, inner nuclear layer. Maximal intensity projections of optical planes encompassing 5to 8 μm thickness.Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 4 of 14Figure 3 YFP-NL2 colocalizes with postsynaptic and presynaptic markers of inhibitory synapses in P21 to 24 retinas. (A) Single planeconfocal images of vertical slices from Thy1-YFP-NL2 retinas labeled with antibodies against the postsynaptic markers gephyrin, the g2 subunit ofGABAA receptors and glycine receptors (GlyR). (B) YFP-NL2 colocalizes with presynaptic markers of inhibitory but not excitatory synapses.Vesicular GABA and glycine transporter (VGAT) was used as a marker of inhibitory presynaptic sites, and CtBP2 as a marker of excitatorypresynaptic sites. Higher magnification views (smaller panels on right) are shown for the boxed regions. For (A, B), two-dimensional cross-correlation coefficient plots of the pixel intensities in the green (YFP-NL2) and red (immunolabeling) channels, and between pixels in the greenand rotated red channels, are shown in the right columns.Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 5 of 14localized to inhibitory synapses at the early stages ofbipolar cell synaptogenesis (Figure 4).Expression of YFP-NL2 in Thy1-YFP-NL2 retina does notalter spontaneous inhibitory postsynaptic currents inRGCsOver-expression of NL2 can cause increased localizationof NL2 to glutamatergic postsynaptic sites [41]. But, wedid not find colocalization of CtBP2 with YFP-NL2 inour transgenic line. Over-expression of NL2 alsoincreases the number of inhibitory synapses in culturedhippocampal neurons [42]. Thus, we performed wholecell patch-clamp experiments and compared the fre-quency and amplitude of spontaneous inhibitory postsy-naptic currents (sIPSCs) of P21 RGCs in YFP-NL2transgenic retinas (n = 3 retinas) and compared thefindings with recordings from wild-type retinas (n = 4retinas). Our results indicate that there is no significantdifference in either the median amplitude or frequencyof sIPSCs of YFP-NL2-expressing RGCs when comparedto RGCs in wild-type littermates (Figure 5).Developmental increases in YFP-NL2 and PSD95-YFPpuncta densities on RGC dendrites occur in parallelTo simultaneously visualize RGC dendrites and inhibi-tory postsynaptic sites, we biolistically transfected RGCsin YFP-NL2-expressing retinas with a plasmid encodingthe red fluorescent protein tdTomato (Figure 6). Werestricted our analysis to large-field RGCs whose den-drites stratify in sublamina ‘b’ of the IPL [43] becausethese cells were frequently labeled both by the ballisticmethod and in the transgenic line (Figure 6A). We gen-erated three-dimensional binary masks (Amira, VisageImaging) of RGC dendrites based on their tdTomatosignal. These masks were then used to isolate YFP-NL2puncta belonging to the tdTomato-labeled cell (Figure6A, right panels; see Materials and methods).Using this approach, we examined the distribution ofYFP-NL2 puncta for RGCs at several developmentalages: P7 to P8, during early phase of synapse formationbetween amacrine cells, bipolar cells and RGCs; P11 toP12, just before eye-opening and when RGCs begin todevelop light responses; P15 to P16, around the time ofeye-opening; P21, when retinal connections appearestablished, and P33, when inhibitory and excitatoryspontaneous currents attain maturity [44]. Figure 6Bshows examples of en face and orthogonal views ofrepresentative RGCs in wholemount preparations. Wethen used custom written Matlab routines to obtain thenumber and distribution of synaptic puncta, as well astotal dendritic length and dendritic area [14]. Fromthese parameters, we generated spatial maps of dendriticdensity (dendritic length divided by dendritic area) andlinear density (the number of puncta per unit dendriticlength) of YFP-NL2 (Figure 7), similar to what we hadcharted for PSD95-YFP [14].As we showed previously [14], the complexity of RGCarbors, represented by dendritic density, decreased withdevelopment. Comparison of the linear density mapsacross ages suggested that YFP-NL2 puncta densityincreased with maturation (Figure 7). Quantificationacross cells confirmed this impression, showing that, onaverage, YFP-NL2 linear density increased until aroundP15, whereupon it remained relatively unchanged (Fig-ure 8A). When we compared YFP-NL2 linear densitywithin the inner and outer halves of the dendritic arboras a function of age, we found that YFP-NL2 densitieswere higher in central compared to the peripheral partsof the arbor, and that this gradient was already apparentat the earliest ages studied, P7 to P8 (Figure 8B).Finally, we compared the developmental patterning ofYFP-NL2 with the distribution of excitatory postsynapticsites, marked by PSD95-YFP expression, on large-fieldON-type RGCs that stratify in the inner part of the IPL.It is evident in Figure 8A, B that both PSD95-YFP andYFP-NL2 puncta densities increased in parallel with age.The average ratio of YFP-NL2/PSD95-YFP puncta den-sities per RGC was relatively unchanged from P7 untilmaturity (Figure 8C). Thus, the mature ratio (1.16 ±0.14; P33) of inhibitory to excitatory synapse numberonto large-field ON RGCs is attained shortly after of theonset of glutamatergic synaptogenesis, and several weeksbefore inner retinal circuits are mature.DiscussionYFP-tagged NL2 expression resembles endogenous NL2expressionNL2 is one of the earliest components of the postsynap-tic specialization of inhibitory synapses, and helpsrecruit gephyrin [34,45] and GABAA receptors to thesesites [36,40]. In the retina of Thy1-YFP-NL2 transgenicmice, punctate distribution of YFP-NL2 in the IPLresembles clustering of endogenous NL2 [40,46]. Likeendogenous NL2 [40], YFP-NL2 colocalized with the g2subunit of the GABAA receptor, a key component ofsynaptic GABAA receptors [47]. In the retinal IPL,gephyrin is mainly associated with a subset of GABAA(a2-containing) receptors [48] and is present in all glyci-nergic synapses, with the exception of connections ontorod bipolar cell terminals [37]. We observed co-localiza-tion of NL2 with gephyrin that corroborates previousreports of a direct interaction between these two pro-teins [45]. This, together with the fact that we alsofound co-localization of NL2 puncta with glycine recep-tor subunits, indicates that YFP-NL2 is also present atglycinergic postsynaptic specializations. In addition, weobserved appositions with a presynaptic marker of inhi-bitory (VGAT) but not excitatory (CtBP2) synapses.Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 6 of 14Figure 4 YFP-NL2 colocalizes with postsynaptic and presynaptic markers of inhibitory synapses at an early developmental age (P10).(A) Postsynaptic markers. (B) Presynaptic marlers. Antibody labeling and cross-correlation analysis were carried out as described for experimentsat P21 to P24 (Figure 3). Shown here are single confocal planes of the immunolabeled vibratome sections. Higher magnification views areshown for the boxed regions.Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 7 of 14YFP-NL2 therefore appears to be a suitable marker ofGABAergic and glycinergic inhibitory postsynapses inthe mouse retinal IPL.One concern that we had was that in our transgenicline, expression of YFP-NL2 may alter synapse numberbecause NL2 over-expression increases the number ofboth inhibitory and excitatory synaptic connections [41].However, we found that RGCs in Thy1-YFP-NL2 miceshowed no significant change in the median amplitudeand frequency of spontaneous inhibitory synaptic cur-rents. Moreover, transgenic YFP-NL2 animals showedgrossly normal behavioral traits, life spans and offspringproduction. This is in contrast to what was describedfor transgenic lines with high levels of NL2 over-expres-sion (50% or more above endogenous levels), which dis-played a variety of phenotypes, including alteredsynaptic function, early postnatal death, and behavioralchanges such as limb clasping, repetitive behavior andanxiety [49]. Thus, we believe that the distribution ofNL2 is not grossly over-expressed in our transgenic line,and YFP-NL2 puncta across the dendritic arbors of theRGCs likely reflect endogenous distributions.Spatial maps of YFP-NL2 puncta for mature large-field ONRGCsTo date, the densities and distribution patterns of ama-crine synapses onto the dendrites of RGCs in themature retina have been obtained largely by serial EM.As such, only a handful of cells can be fully recon-structed. More recently, combining immunostaining fortransmitter receptors with single cell labeling hasenabled comparisons of the distributions of excitatoryand inhibitory synapses on adult primate RGCs [27,28].Here, we used the expression of YFP-NL2 to mark thelocations of inhibitory postsynaptic sites and revealedthe spatial distributions of amacrine synapses on thedendritic arbors of large field ON RGCs. One monthpostnatally, we found that, on average, the linear densityof YFP-NL2 is 0.38 ± 0.03 puncta/μm. This density issimilar to the density of amacrine synapses on the den-drites of a large-field ON alpha-RGC in the adult catretina (0.36 synapse/μm), previously reconstructed byEM [22].Serial EM and partial three-dimensional reconstruc-tions of RGC dendritic trees in the cat, marmoset andrabbit retina have also revealed a considerable variationin the ratio of inhibitory and excitatory synapses ontoindividual RGCs [20-24,50]. Amacrine synapses com-prise 50 to 80% of the total number of synapses on aRGC, depending on the species and the type of RGCanalyzed. Confocal reconstructions of rabbit direction-selective RGCs in which presynaptic ribbon-associatedterminals were labeled and inhibitory postsynaptic sitesrevealed by immunostaining for the a1 subunit of theGABAA receptor suggest that inhibitory synapses consti-tute at least one-third of all synapses onto these RGCs[21,50]. In our study, comparison between NL2 andPSD95 puncta densities suggests that amacrine andbipolar cells form almost equal numbers of synapses onmature large-field mouse ON RGCs. Additionally, wefound that there is a shallow gradient (about 1.5) ofNL2 puncta density from the center to peripheral partsof the dendritic arbor of the large-field RGCs, similar tothe gradient of glutamatergic postsynaptic sites on theseRGCs [14]. Thus, the densities of inhibitory and excita-tory postsynaptic sites appear matched across the den-dritic arbor of these mouse RGCs.Figure 5 Expression of YFP-NLG2 does not significantly alterthe spontaneous inhibitory drive onto retinal ganglion cells.(A) Example traces showing sIPSCs from whole-cell recordings ofwild-type retinas (WT; cells 1 and 2) and YFP-NLG2-expressingretinas (cells 3 and 4) at P21 (inset magnifications showingindividual events). (B) Mean frequencies and median amplitudes ofsIPSCS from WT RGCs (open-triangles, n = 6 cells), and YFP-NLG2-expressing RGCs (open-circles, n = 7 cells). The differences in meansbetween each group (dark grey symols) were not statisticallydifferent (frequency P = 0.23, amplitude P = 0.23, Wilcoxon Rank-sum test). Values for the example traces shown in (A) are indicatedby the numbers 1 to 4 (light grey symbols). Error bars are standarderror of the means.Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 8 of 14Coordinated development of inhibitory and excitatorypostsynaptic sites onto RGCsIt is evident from morphological and physiological stu-dies that, in vivo, GABAergic and glycinergic synapticconnections are formed before glutamatergic connec-tions. For example, in the hippocampus, the apical den-drite of CA1 pyramidal cells is first contacted byGABAergic inteneurons prior to extension of these den-drites into the stratum lacunosum-moleculare, wherethey later receive input from glutamatergic afferents ofthe perforant pathway [4,5]. Likewise, RGCs also obtainGABAergic and glycinergic drive from amacrine cellsprior to the formation of glutamatergic synapses frombipolar cells [51] (but see [52]). Unlike hippocampal pyr-amidal cells where dendrites need to reach available glu-tamatergic afferents, the delay in glutamatergicsynaptogenesis onto RGCs is because bipolar cells aregenerated much later than amacrine cells. As apparentin other model systems, GABAergic and glycinergictransmission onto RGCs is also initially depolarizing,but switches to hyperpolarizing when glutamatergicsynapses are established [53-55]. Thus, circuitry of theinner retina of vertebrates is a good representativemodel for comparing the development of inhibitory ver-sus excitatory synapses onto the same postsynaptic celltype. By mapping the distribution and density of YFP-NL2 on the dendrites of individually labeled large-fieldRGCs, we found that inhibitory synaptogenesis ontothese cells peaks at the end of the second postnatalweek, and stays relatively constant thereafter. This tem-poral profile in the generation of amacrine synapsesmatches that previously suggested by following changesin total inhibitory synapse density across the mouse IPL[29]. Rapid amacrine synaptogenesis during the first andsecond postnatal weeks also appears to occur in otherspecies, as revealed by immunostaining for GABAAreceptors, glycine receptor subunits or gephyrin onRGCs, for example in rat and rabbit retina [37,56].Our observations here show that the time course inthe increase in YFP-NL2 puncta density with age paral-lels that of PSD95-YFP on the dendrites of large-fieldRGCs [14], resulting in a surprisingly constant inhibi-tory/excitatory synapse ratio across the RGC arborshortly after bipolar cell synaptogenesis commences.Figure 6 RGC dendrites and putative inhibitory postsynaptic sites visualized upon expression of tdTomato in RGCs of Thy1-YFP-NL2retina. (A) Whole-mount or en face view of a P21 YFP-NL2-expressing (green) ON RGC, co-expressing tdTomato (red). The dendritic labeling wasused to perform a ‘masking’ function that digitally removes YFP-NL2 outside the labeled cell, as shown at high magnification in the right panels.The YFP-NL2 puncta identified using a semi-automated dotfinder program (magenta dots) are superimposed onto the image of the YFPfluorescence. (B) Whole-mount view (top) and orthogonal (bottom) projections of large-field ON RGCs at several postnatal ages.Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 9 of 14This inhibitory to excitatory synapse ratio appears main-tained throughout neonatal development despite signifi-cant dendritic remodeling [14] and large-scalesynaptogenesis during this period [29]. The increase inYFP-NL2 and PSD95-YFP puncta densities may also becoordinated locally on the dendritic arbor as well asacross ages because the centro-peripheral gradients ofboth YFP-NL2 and PSD95-YFP appeared by P12 [14]and persist into adulthood. Indeed, in hippocampal neu-rons in culture, the ratio of excitatory and inhibitorysynapses is matched at the level of individual dendriticbranches, producing a local balance of excitation andinhibition [57]. However, unlike our current findings forRGCs, this ratio for hippocampal neurons increasesbetween the second and third week in vitro [57,58]. Itshould be noted, however, that although the ratio ofinhibitory to excitatory synapses appears constant acrossages for the large-field ON RGCs, functional drive frombipolar cells and amacrine cells onto RGCs increaseswith maturation. Whole-cell recordings from neonatalmouse retina show that both spontaneous inhibitoryFigure 7 Spatial maps of dendritic and YFP-NL2 punctadensities of RGCs across development. Left panels:skeletonization of RGC dendrites (red) and identified YFP-NL2puncta (blue). Middle panels: dendritic territories determined byconvolving a 10 μm diameter disk centered at each pixel with thedendritic skeleton [14]; dendritic density (dendritic length within 100μm2 area) is mapped across the arbor. Right panels: spatialdistributions of YFP-NL2 puncta (linear density; number of punctaper micrometer of dendrite).Figure 8 Inhibitory and excitatory synapse densities on RGCdendrites as a function of age. (A) Linear densities of YFP-NL2puncta (shaded bars, n = number of cells) compared with thedistribution of PSD95 (open bars) during development (datareplotted from [14,64]). (B) Centro-peripheral gradient of inhibitoryand excitatory puncta obtained as described in Morgan et al. [14].The gradients were quantified by determining the ratio of punctadensity in the inner and outer halves of a circle encompassing 98%of the dendritic arbor. The analysis excludes a 10-μm region aroundthe cell soma. The inner and outer densities were significantlydifferent at all ages except for P12 (paired t-test; P < 0.05). (C) Ratioof average linear densities of YFP-NL2/PSD95-YFP across ages. Errorbars are standard error of the means.Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 10 of 14and excitatory postsynaptic current frequencies increaseconcurrently over the first postnatal month [59].What mechanisms regulate the ratio of inhibitory toexcitatory synapses on dendrites? It is likely that apartfrom glutamatergic transmission [60,61], GABAergictransmission itself plays a role in regulating the balance ofinhibitory and excitatory inputs in developing CNS circuits[62,63]. Glutamatergic transmission certainly regulates thenumber of synapses formed between bipolar cells and thelarge-field RGCs [64], but as yet, we do not know howamacrine synapse numbers or distributions on these RGCsmight be affected. Conversely, the influence of GABAergictransmission on amacrine or bipolar cell synaptogenesison to RGC dendrites has yet to be explored. It is evident,however, that whatever the role neurotransmission playsin setting up amacrine-bipolar cell synaptic ratios, the finalsynapse numbers from each input type are likely to beshaped by the relative addition and elimination of synapses[64]. Currently, we do not know whether, like bipolar cellcontacts, amacrine connections also undergo remodelingduring development, although it is possible given thatamacrine neurites show structural rearrangements duringthe period of synaptogenesis [65]. A further considerationis that glycinergic amacrine cells are born later thanGABAergic amacrine cells [66,67]. Because NL2 does notnecessarily distinguish GABAergic from glycinergic post-synaptic sites in the retinal IPL, future studies specificallymarking GABAergic or glycinergic synapses will help dis-tinguish the contributions of these two major amacrinecell populations to the observed increase in conventionalsynapses during postnatal development. Such knowledgewill help provide further insight into how the developmentof amacrine and bipolar cell connectivity onto RGCs iscoordinated.ConclusionsBased on YFP-NL2 expression in a Thy1-YFP-NL2 trans-genic mouse line we generated, we found that the densityof inhibitory amacrine synapses on the arbors of large-field ON RGCs increased gradually, until around eyeopening (P15). By comparing the spatial densities of YFP-NL2 with PSD95-YFP across the dendritic arbors of theseRGCs, we discovered that their adult ratio of inhibitory/excitatory postsynaptic sites was rapidly attained, shortlyafter bipolar cells form synapses in the IPL (P7). Ourobservations suggest that bipolar and amacrine cellsynaptogenesis onto RGCs are coordinated, and abalanced ratio of excitatory and inhibitory synapse densi-ties is established prior to the onset of visual experience.Materials and methodsGeneration of Thy1-YFP-NL2 transgenic miceYFP-NL2 expressed here was modified from Graf et al.[34] and consists of the signal sequence of mouseNeuroligin-1, hexahistidine (HHHHHH), Flag(GGDYKDDDDK), and EYFP tags followed by themature coding sequence of mouse NL2. The Thy1 pro-moter [68] was used to drive expression of YFP-NL2.The transgene was generated by cloning the Thy1 pro-moter fragment into shuttle vector LNL, which con-tained the needed restriction sites for transgene release.YFP-NL2 was digested with HindIII and Afl II andblunted into XhoI cut and blunted thy1-LNL. TheThy1-YFP-NL2 transgene was released from the vectorbackbone sequence by restriction digestion with AscIand PmeI and injected into B6/CBA F1 hybrid pronucleito generate founder mice.ImmunohistochemistryC57BL/6 (P21) and Thy1-YFP-NL2 (P7, P12, P15, P21)mice were deeply anesthetized with 5% isofluorane anddecapitated. Eyes were removed and placed in ice coldmouse artificial cerebrospinal fluid (mACSF; 119 mMNaCl, 2.5 mM KCl, 2.5 mM CaCl2, 1.3 mM MgCl2, 1mM NaH2PO4, 11 mM glucose (20), 20 mM HEPES,pH = 7.4). After removing the lens and vitreous, the eyecup was fixed in 4% paraformaldehyde for 15 to 30 min-utes. After fixation, the eye cups were rinsed in 0.1 MPBS. The retina was removed from the eye cup,embedded in 4% low-melting point agarose and cut into60 μm thick sections using a vibratome. Sections weremounted and used for imaging or processed for immu-nostaining as follows: blocked in 10% NGS in PBS for 1hour followed by overnight incubation in 5% NGS, 0.5%Triton-X100, with the corresponding primary antibodies.The primary antibodies used were: rabbit anti-NL2 anti-body (1:8,000; generous gift of F Varoqueaux and NBrose) [35,40], guinea-pig anti-g2 antibody (1:1,000; gen-erous gift of JM Fritschy), a monoclonal mouse antibody(mAb4a (P21) against all glycine receptor subunits,1:400;or mAb2b (P10), against the a1 subunit of the glycinereceptor, 1:400; Synaptic Systems, Goettingen, Ger-many), anti-gephyrin mAb7a (1:500; Synaptic Systems),anti-VGAT antibody (1:1,000; Millipore, Temecula, CA,USA), anti-CtBP2 antibody (1:1000; BD Transduction,Franklin Lakes, NJ, USA). Sections were then washedand incubated for 1 hour with the corresponding sec-ondary antibody conjugated to either Alexa-488 orAlexa-568 (1:1,000; Invitrogen, Carlsbad, CA, USA).Cell transfectionThy1-YFP-NL2 mice were deeply anesthetized with 5%isofluorane and decapitated. Eyes were removed andplaced in ice cold mouse mACSF. Retinas were removedfrom the eye cup and mounted RGC side up on blacknitrocellulose filter paper (HABP013, Millipore, Bedford,MA, USA). Gold particles were coated with 20 μg CMV:tdtomato DNA (gift of R Tsien) and delivered using aSoto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 11 of 14Bio-Rad Helios gene-gun as previously described [14].Retinas were incubated for 18 to 24 hours at 33°C, fixedfor 30 minutes in 4% parafolmaldehyde in mACSF,washed in PBS and mounted in Vectashield (VectorLabs, Burlingame, CA, USA). The data presented in thisstudy were obtained from six P8, five P12, five P15, tenP21 and five P30 RGCs.Imaging and image analysisImages were obtained using a 1.35 NA 60× oil objective(Olympus). Images were acquired at 0.069 × 0.069 × 0.3μm for double labeling immunohistochemistry (Figures1, 3 and 4), and 0.103 × 0.103 × 0.3 μm voxel sizes forvertical slices of YFP-NL2 retinas (Figure 2) and retinalwhole-mounts (Figures 2 and 6). Images were processedusing Metamorph (Molecular Devices, Sunnyvale, CA,USA), Image J (NCBI), Amira (Mercury Computer Sys-tems Inc., Chelmsford, MA, USA) and Matlab (MathWorks, Natick, MA, USA). Images were median-filteredto reduce noise. The contrast and gamma of the imageswere adjusted to increase visualization of dim objects.Using the ‘label-field’ function of the AMIRA softwareprogram, a threshold was applied, plane by plane, tocapture pixels representing the dendrites of the RGCs[14]. This procedure generated a binary mask of thedendrites that was then used to isolate puncta from theYFP-NL2 channel that resided within the mask (YFP-NL2-labeled voxels outside the mask were then dis-carded). In the same process, cell somas were removedbefore further analysis. Custom Matlab programs wereused to generate dendritic skeletons and to identifypuncta, dendritic lengths and dendritic areas as pre-viously described [14]. Dendritic density represents thetotal dendritic length divided by the total area of thedendritic territory.In order to assess whether NL2 signal significantlyoverlaps with other synapse markers, a custom Matlabprogram was used to calculate the two-dimensionalcross-correlation coefficients of the signals from bothchannels (Josh Morgan and Daniel Kerschensteiner).This approach does not require identification of puncta,yet allows us to determine whether the fluorescent sig-nals in the two channels are spatially correlated, or arerandomly associated (random association determined byrotating one image 180° relative to the other image).Within an optical plane, the signal intensity of a pixelin the green channel was compared with the signalintensity of the corresponding pixel in the red channel(0,0 location). To obtain the two-dimensional correla-tion plot, the intensity of a green pixel was comparedwith the intensity of red pixels (or vice versa) to theright, left, top and bottom, displaced from the referencepixel up to 2.8 μm. A pixel intensity threshold of 20 to35, defining the background, was used while analyzingimages obtained from wild-type mouse retina. Thresh-olding was not used when analyzing images obtainedfrom the YFP-NL2 transgenic mouse line due to thelower immunohistochemistry background level in thetransgenic retina.Representative examples shown in the figures arecross-correlation coefficients calculated based on 20 to45 image planes (z-step size = 0.3 μm) of an imagedfield of view. The equation used to calculate the correla-tion coefficient is:R(r, g)=C(r, g)√C(r, r), C(g, g)where R(r, g) is the correlation coefficient of the redand green channel and C is the covariance of the corre-sponding channels.For two identical images, the correlogram peak is 1.The value of the positive peak for two identical imagesis only weakly influenced by the absolute intensities ofthe pixels within an image but strongly corresponds tohow spatially ‘similar’ the two images are. For overlap-ping green and red puncta, the peak of the two-dimen-sional correlation plot falls off symmetrically on allsides. The correlograms with positive peaks have a full-width at half maximum of less than 1 μm, suggestingthat there exist structures in the two images within lessthan 1 μm overlap in space.ElectrophysiologyRetinal flat mounts were prepared as described abovefor cell transfection, but mounted on white filter paper(Anodisc 13, Whatman Inc., Piscataway, NJ, USA) forbetter visualization. Recordings were performed at roomtemperature and retinas were maintained in bicarbo-nate-buffered mACSF containing 125 mM NaCl, 2.5mM KCl, 2 mM CaCl2, 1 mM MgCl2, 1.25 mMNaH2PO4, 11 mM glucose and 26 mM NaHCO3 (equili-brated with 95% O2 and 5% CO2). Whole-cell recordingswere performed with electrodes (4 to 8 MΩ) filled with120 mM Cs-gluconate, 1 mM CaCl2, 1 mM MgCl2, 10mM Na-HEPES, 11 mM EGTA, and 10 mM TEA-Cl(pH 7.2 adjusted with CsOH). For some experiments 2mM QX314 was also included in the patch pipette. Aliquid junction potential of 15 mV was corrected beforethe cell was attached, and series resistance was not com-pensated. Data were acquired using an Axopatch 200 Bamplifier (Molecular Devices), low-pass filtered at 2 kHzand digitized at 5 kHz. sIPSCs were recorded at -0 mV,the reversal potential of cation currents in our recordingconditions. Area and amplitude thresholds (Mini Analy-sis, Synaptosoft, Decatur, GA, USA) were optimized todetect > 90% of the events identified by eye for theentirety of recordings analyzed. For overlapping events,Soto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 12 of 14the baseline for amplitude measurement of each eventwas estimated from exponential decay extrapolation ofthe previous event.AbbreviationsCNS: central nervous system; CtBP2: carboxy-terminal binding protein 2; EM:electron microscopy; GABA: γ-aminobutyric acid; IPL: inner plexiform layer;mACSF: mouse artificial cerebrospinal fluid; NL2: Neuroligin 2; P: postnatalday; PBS: phosphate-buffered saline; RGC: retinal ganglion cell; sIPSC:spontaneous inhibitory postsynaptic current; VGAT: vesicular inhibitorytransmitter transporter; YFP: yellow fluorescent protein.AcknowledgementsSupported by NIH grant EY10699 to R.O.W and a Canadian Institutes ofHealth Research grant MOP-69096 to A.M.C. A.B. is supported aDevelopmental Biology Predoctoral Training Grant T32HD007183 from theNational Institute of Child Health and Human Development. We thank Dr.Mrinalini Hoon for many helpful comments on the manuscript.Author details1Department of Biological Structure, University of Washington, 1950 PacificAve, Seattle, WA 98195, USA. 2Department of Ophthalmology and VisualSciences, Washington University in St Louis, 660 S. Euclid Ave, St Louis, MO63110, USA. 3Department of Anatomy and Neurobiology, WashingtonUniversity in St Louis, 660 S. Euclid Ave, St Louis, MO 63110, USA. 4BrainResearch Centre and Department of Psychiatry, University of BritishColumbia, 2211 Wesbrook Mall, Vancouver, BC, Canada V6T 2B5.Authors’ contributionsFS and AB carried out the experiments. FS, AB and DK performed theanalysis. RL, YK, and AMC generated the transgenic mice. FS and ROLWdesigned and coordinated the work. All authors read and approved the finalmanuscript.Competing interestsThe authors declare that they have no competing interests.Received: 22 March 2011 Accepted: 24 August 2011Published: 24 August 2011References1. Salinas E, Sejnowski TJ: Impact of correlated synaptic input on outputfiring rate and variability in simple neuronal models. J Neurosci 2000,20:6193-6209.2. Turrigiano GG: The self-tuning neuron: synaptic scaling of excitatorysynapses. Cell 2008, 135:422-435.3. Feldman DE: Synaptic mechanisms for plasticity in neocortex. Annu RevNeurosci 2009, 32:33-55.4. Hennou S, Khalilov I, Diabira D, Ben-Ari Y, Gozlan H: Early sequentialformation of functional GABA(A) and glutamatergic synapses on CA1interneurons of the rat foetal hippocampus. Eur J Neurosci 2002,16:197-208.5. Tyzio R, Represa A, Jorquera I, Ben-Ari Y, Gozlan H, Aniksztejn L: Theestablishment of GABAergic and glutamatergic synapses on CA1pyramidal neurons is sequential and correlates with the development ofthe apical dendrite. J Neurosci 1999, 19:10372-10382.6. Alvarez VA, Sabatini BL: Anatomical and physiological plasticity ofdendritic spines. Annu Rev Neurosci 2007, 30:79-97.7. Bourne JN, Harris KM: Balancing structure and function at hippocampaldendritic spines. Annu Rev Neurosci 2008, 31:47-67.8. Holtmaat A, Svoboda K: Experience-dependent structural synapticplasticity in the mammalian brain. Nat Rev Neurosci 2009, 10:647-658.9. Pan F, Gan WB: Two-photon imaging of dendritic spine development inthe mouse cortex. Dev Neurobiol 2008, 68:771-778.10. Erikoz B, Jusuf PR, Percival KA, Grunert U: Distribution of bipolar input tomidget and parasol ganglion cells in marmoset retina. Vis Neurosci 2008,25:67-76.11. Jakobs TC, Koizumi A, Masland RH: The spatial distribution ofglutamatergic inputs to dendrites of retinal ganglion cells. J Comp Neurol2008, 510:221-236.12. Koizumi A, Jakobs TC, Masland RH: Regular mosaic of synaptic contactsamong three retinal neurons. J Comp Neurol 2011, 519:341-357.13. Lin B, Martin PR, Grunert U: Expression and distribution of ionotropicglutamate receptor subunits on parasol ganglion cells in the primateretina. Vis Neurosci 2002, 19:453-465.14. Morgan JL, Schubert T, Wong RO: Developmental patterning ofglutamatergic synapses onto retinal ganglion cells. Neural Dev 2008, 3:8.15. Niell CM, Meyer MP, Smith SJ: In vivo imaging of synapse formation on agrowing dendritic arbor. Nat Neurosci 2004, 7:254-260.16. Huang ZJ: Activity-dependent development of inhibitory synapses andinnervation pattern: role of GABA signalling and beyond. J Physiol 2009,587:1881-1888.17. Megias M, Emri Z, Freund TF, Gulyas AI: Total number and distribution ofinhibitory and excitatory synapses on hippocampal CA1 pyramidal cells.Neuroscience 2001, 102:527-540.18. MacNeil MA, Masland RH: Extreme diversity among amacrine cells:implications for function. Neuron 1998, 20:971-982.19. Wassle H, Koulen P, Brandstatter JH, Fletcher EL, Becker CM: Glycine andGABA receptors in the mammalian retina. Vision Res 1998, 38:1411-1430.20. Dacheux RF, Chimento MF, Amthor FR: Synaptic input to the on-offdirectionally selective ganglion cell in the rabbit retina. J Comp Neurol2003, 456:267-278.21. Famiglietti EV: Synaptic organization of complex ganglion cells in rabbitretina: type and arrangement of inputs to directionally selective andlocal-edge-detector cells. J Comp Neurol 2005, 484:357-391.22. Freed MA, Sterling P: The ON-alpha ganglion cell of the cat retina and itspresynaptic cell types. J Neurosci 1988, 8:2303-2320.23. Ghosh KK, Grunert U: Synaptic input to small bistratified (blue-ON)ganglion cells in the retina of a new world monkey, the marmosetCallithrix jacchus. J Comp Neurol 1999, 413:417-428.24. Haverkamp S, Eldred WD, Ottersen OP, Pow D, Ammermuller J: Synapticinputs to identified color-coded amacrine and ganglion cells in theturtle retina. J Comp Neurol 1997, 389:235-248.25. Cohen E, Sterling P: Parallel circuits from cones to the on-beta ganglioncell. Eur J Neurosci 1992, 4:506-520.26. Kolb H, Nelson R: OFF-alpha and OFF-beta ganglion cells in cat retina: II.Neural circuitry as revealed by electron microscopy of HRP stains. JComp Neurol 1993, 329:85-110.27. Percival KA, Jusuf PR, Martin PR, Grunert U: Synaptic inputs onto smallbistratified (blue-ON/yellow-OFF) ganglion cells in marmoset retina. JComp Neurol 2009, 517:655-669.28. Percival KA, Martin PR, Grunert U: Synaptic inputs to two types ofkoniocellular pathway ganglion cells in marmoset retina. J Comp Neurol2011, 519:2135-2153.29. Fisher LJ, Easter SS Jr: Retinal synaptic arrays: continuing development inthe adult goldfish. J Comp Neurol 1979, 185:373-379.30. Schubert T, Kerschensteiner D, Eggers ED, Misgeld T, Kerschensteiner M,Lichtman JW, Lukasiewicz PD, Wong RO: Development of presynapticinhibition onto retinal bipolar cell axon terminals is subclass-specific. JNeurophysiol 2008, 100:304-316.31. Nishimura Y, Rakic P: Development of the rhesus monkey retina: II. Athree-dimensional analysis of the sequences of synaptic combinations inthe inner plexiform layer. J Comp Neurol 1987, 262:290-313.32. Crooks J, Okada M, Hendrickson AE: Quantitative analysis ofsynaptogenesis in the inner plexiform layer of macaque monkey fovea. JComp Neurol 1995, 360:349-362.33. Varoqueaux F, Aramuni G, Rawson RL, Mohrmann R, Missler M, Gottmann K,Zhang W, Sudhof TC, Brose N: Neuroligins determine synapse maturationand function. Neuron 2006, 51:741-754.34. Graf ER, Zhang X, Jin SX, Linhoff MW, Craig AM: Neurexins inducedifferentiation of GABA and glutamate postsynaptic specializations vianeuroligins. Cell 2004, 119:1013-1026.35. Varoqueaux F, Jamain S, Brose N: Neuroligin 2 is exclusively localized toinhibitory synapses. Eur J Cell Biol 2004, 83:449-456.36. Patrizi A, Scelfo B, Viltono L, Briatore F, Fukaya M, Watanabe M, Strata P,Varoqueaux F, Brose N, Fritschy JM, Sassoe-Pognetto M: Synapse formationSoto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 13 of 14and clustering of neuroligin-2 in the absence of GABAA receptors. ProcNatl Acad Sci USA 2008, 105:13151-13156.37. Sassoe-Pognetto M, Wassle H: Synaptogenesis in the rat retina:subcellular localization of glycine receptors, GABA(A) receptors, and theanchoring protein gephyrin. J Comp Neurol 1997, 381:158-174.38. Fritschy JM, Mohler H: GABAA-receptor heterogeneity in the adult ratbrain: differential regional and cellular distribution of seven majorsubunits. J Comp Neurol 1995, 359:154-194.39. Bechade C, Colin I, Kirsch J, Betz H, Triller A: Expression of glycine receptoralpha subunits and gephyrin in cultured spinal neurons. Eur J Neurosci1996, 8:429-435.40. Hoon M, Bauer G, Fritschy JM, Moser T, Falkenburger BH, Varoqueaux F:Neuroligin 2 controls the maturation of GABAergic synapses andinformation processing in the retina. J Neurosci 2009, 29:8039-8050.41. Levinson JN, El-Husseini A: Building excitatory and inhibitory synapses:balancing neuroligin partnerships. Neuron 2005, 48:171-174.42. Chubykin AA, Atasoy D, Etherton MR, Brose N, Kavalali ET, Gibson JR,Sudhof TC: Activity-dependent validation of excitatory versus inhibitorysynapses by neuroligin-1 versus neuroligin-2. Neuron 2007, 54:919-931.43. Volgyi B, Chheda S, Bloomfield SA: Tracer coupling patterns of theganglion cell subtypes in the mouse retina. J Comp Neurol 2009,512:664-687.44. Johnson J, Tian N, Caywood MS, Reimer RJ, Edwards RH, Copenhagen DR:Vesicular neurotransmitter transporter expression in developingpostnatal rodent retina: GABA and glycine precede glutamate. J Neurosci2003, 23:518-529.45. Poulopoulos A, Aramuni G, Meyer G, Soykan T, Hoon M, Papadopoulos T,Zhang M, Paarmann I, Fuchs C, Harvey K, Jedlicka P, Schwarzacher SW,Betz H, Harvey RJ, Brose N, Zhang W, Varoqueaux F: Neuroligin 2 drivespostsynaptic assembly at perisomatic inhibitory synapses throughgephyrin and collybistin. Neuron 2009, 63:628-642.46. Koulen P, Sassoe-Pognetto M, Grunert U, Wassle H: Selective clustering ofGABA(A) and glycine receptors in the mammalian retina. J Neurosci 1996,16:2127-2140.47. Essrich C, Lorez M, Benson JA, Fritschy JM, Luscher B: Postsynapticclustering of major GABAA receptor subtypes requires the gamma 2subunit and gephyrin. Nat Neurosci 1998, 1:563-571.48. Sassoe-Pognetto M, Kirsch J, Grunert U, Greferath U, Fritschy JM, Mohler H,Betz H, Wassle H: Colocalization of gephyrin and GABAA-receptorsubunits in the rat retina. J Comp Neurol 1995, 357:1-14.49. Hines RM, Wu L, Hines DJ, Steenland H, Mansour S, Dahlhaus R,Singaraja RR, Cao X, Sammler E, Hormuzdi SG, Zhuo M, El-Husseini A:Synaptic imbalance, stereotypies, and impaired social interactions inmice with altered neuroligin 2 expression. J Neurosci 2008, 28:6055-6067.50. Jeon CJ, Kong JH, Strettoi E, Rockhill R, Stasheff SF, Masland RH: Pattern ofsynaptic excitation and inhibition upon direction-selective retinalganglion cells. J Comp Neurol 2002, 449:195-205.51. Unsoeld T, Stradomska AM, Wang R, Rathjen FG, Juttner R: Earlymaturation of GABAergic synapses in mouse retinal ganglion cells. Int JDev Neurosci 2008, 26:233-238.52. Rorig B, Grantyn R: Rat retinal ganglion cells express Ca(2+)-permeablenon-NMDA glutamate receptors during the period of histogenetic celldeath. Neurosci Lett 1993, 153:32-36.53. Leitch E, Coaker J, Young C, Mehta V, Sernagor E: GABA type-A activitycontrols its own developmental polarity switch in the maturing retina. JNeurosci 2005, 25:4801-4805.54. Vu TQ, Payne JA, Copenhagen DR: Localization and developmentalexpression patterns of the neuronal K-Cl cotransporter (KCC2) in the ratretina. J Neurosci 2000, 20:1414-1423.55. Zhang LL, Pathak HR, Coulter DA, Freed MA, Vardi N: Shift of intracellularchloride concentration in ganglion and amacrine cells of developingmouse retina. J Neurophysiol 2006, 95:2404-2416.56. Koulen P: Postnatal development of GABAA receptor beta1, beta2/3, andgamma2 immunoreactivity in the rat retina. J Neurosci Res 1999,57:185-194.57. Liu G: Local structural balance and functional interaction of excitatoryand inhibitory synapses in hippocampal dendrites. Nat Neurosci 2004,7:373-379.58. Benson DL, Cohen PA: Activity-independent segregation of excitatoryand inhibitory synaptic terminals in cultured hippocampal neurons. JNeurosci 1996, 16:6424-6432.59. Tian N, Copenhagen DR: Visual deprivation alters development ofsynaptic function in inner retina after eye opening. Neuron 2001,32:439-449.60. Colin-Le Brun I, Ferrand N, Caillard O, Tosetti P, Ben-Ari Y, Gaiarsa JL:Spontaneous synaptic activity is required for the formation of functionalGABAergic synapses in the developing rat hippocampus. J Physiol 2004,559:129-139.61. Liu Y, Zhang LI, Tao HW: Heterosynaptic scaling of developing GABAergicsynapses: dependence on glutamatergic input and developmentalstage. J Neurosci 2007, 27:5301-5312.62. Akerman CJ, Cline HT: Depolarizing GABAergic conductances regulate thebalance of excitation to inhibition in the developing retinotectal circuitin vivo. J Neurosci 2006, 26:5117-5130.63. Hartman KN, Pal SK, Burrone J, Murthy VN: Activity-dependent regulationof inhibitory synaptic transmission in hippocampal neurons. Nat Neurosci2006, 9:642-649.64. Kerschensteiner D, Morgan JL, Parker ED, Lewis RM, Wong RO:Neurotransmission selectively regulates synapse formation in parallelcircuits in vivo. Nature 2009, 460:1016-1020.65. Wong RO, Collin SP: Dendritic maturation of displaced putativecholinergic amacrine cells in the rabbit retina. J Comp Neurol 1989,287:164-178.66. Cherry TJ, Trimarchi JM, Stadler MB, Cepko CL: Development anddiversification of retinal amacrine interneurons at single cell resolution.Proc Natl Acad Sci USA 2009, 106:9495-9500.67. Voinescu PE, Kay JN, Sanes JR: Birthdays of retinal amacrine cell subtypesare systematically related to their molecular identity and soma position.J Comp Neurol 2009, 517:737-750.68. Feng G, Mellor RH, Bernstein M, Keller-Peck C, Nguyen QT, Wallace M,Nerbonne JM, Lichtman JW, Sanes JR: Imaging neuronal subsets intransgenic mice expressing multiple spectral variants of GFP. Neuron2000, 28:41-51.doi:10.1186/1749-8104-6-31Cite this article as: Soto et al.: Coordinated increase in inhibitory andexcitatory synapses onto retinal ganglion cells during development.Neural Development 2011 6:31.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitSoto et al. Neural Development 2011, 6:31http://www.neuraldevelopment.com/content/6/1/31Page 14 of 14


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items