Molecular Codes for Cell Type Specification in Brn3 Retinal Ganglion Cells

Home , CPNE4, IRX1, PLXNA3, RAB6B, REEP5, STMN3

Molecular codes for cell type specification in Brn3 retinal ganglion cells

Szilard Sajgoa,1, Miruna Georgiana Ghiniaa,2, Matthew Brooksb, Friedrich Kretschmera,3, Katherine Chuanga,4, Suja Hiriyannac, Zhijian Wuc, Octavian Popescud,e, and Tudor Constantin Badeaa,5

aRetinal Circuits Development and Genetics Unit, Neurobiology–Neurodegeneration and Repair Laboratory, National Eye Institute, Bethesda, MD 20892; bGenomics Core, Neurobiology–Neurodegeneration and Repair Laboratory, National Eye Institute, Bethesda, MD 20892; cOcular Gene Therapy Core, National Eye Institute, Bethesda, MD 20892; dInstitute of Biology, Romanian Academy, Bucharest 060031, Romania; and eMolecular Biology Center, Interdisciplinary Research Institute on Bio-Nano-Science, Babes-Bolyai University, Cluj-Napoca 400084, Romania

Edited by Jeremy Nathans, Johns Hopkins University, Baltimore, MD, and approved April 12, 2017 (received for review November 8, 2016) Visual information is conveyed from the eye to the brain by Onecut2 (6, 13, 26–33). Together with Isl1 and Brn3b, these distinct types of retinal ganglion cells (RGCs). It is largely unknown downstream factors are expressed in partially overlapping patterns how RGCs acquire their defining morphological and physiological in RGC types, and some were shown to be required for survival features and connect to upstream and downstream synaptic and/or dendrite and axon formation in various RGC types. partners. The three Brn3/Pou4f transcription factors (TFs) participate However, many other TFs may be involved in generating the in a combinatorial code for RGC type specification, but their exact diversity of RGC types (34–38). We have previously used re- molecular roles are still unclear. We use deep sequencing to define porter knock-in alleles expressing alkaline phosphatase (AP; a i ii ( ) transcriptomes of Brn3a- and/or Brn3b-positive RGCs, ( )Brn3a- glycosylphosphatidylinositol (GPI)-linked, extracellular molecule) iii and/or Brn3b-dependent RGC transcripts, and ( ) transcriptomes of at the loci of Brn3a, Brn3b, and Brn3c (Brn3CKOAP)todescribe retinorecipient areas of the brain at developmental stages relevant their cell type distribution among RGCs and other sensory pro- for axon guidance, dendrite formation, and synaptogenesis. We re- jection neurons. We also identified axonal and dendrite arbor veal a combinatorial code of TFs, cell surface molecules, and deter- defects in RGCs missing Brn3a, Brn3b, or Brn3c either alone or in minants of neuronal morphology that is differentially expressed in specific RGC populations and selectively regulated by Brn3a and/or combination (6, 13, 31, 39, 40). We now describe an immunoaffinity purification strategy using anti-AP antibodies to isolate Brn3b. This comprehensive molecular code provides a basis for un- AP derstanding neuronal cell type specification in RGCs. RGCs from Brn3 RGCs that are either WT or KO for Brn3a or Brn3b. Using our knowledge of partially overlapping RGC pop- retinal ganglion cells | transcription factors | neuronal cell types | Pou4f1 | ulations expressing Brn3s, we can identify molecules selectively Pou4f2 enriched in RGCs, selectively expressed in distinct Brn3 RGC subpopulations, and/or regulated by Brn3a or Brn3b in these RGC he molecular analysis of neuronal circuits benefits signifi- Tcantly from modern approaches to gene expression profiling Significance and genetic manipulation. The mechanisms of cell type specification are still poorly understood, but experiments in model organisms We report here transcriptome analysis by RNA sequencing suggest a combination of transcriptional regulation, extracellular (RNASeq) of genetically labeled and affinity-purified mouse signals, and cell–cell interactions (1–4). Retinal ganglion cells retinal ganglion cell (RGC) populations. Using a previously (RGCs) are a particularly powerful system for illustrating the established conditional knock-in reporter strategy, we label molecular and activity-dependent mechanisms of cell type speci- RGCs from which specific transcription factors have been re- fication. Based on molecular markers, dendritic arbor morphol- moved and determine the consequences on transcriptional ogies, axonal projections to retinorecipient areas of the brain, programs at different stages critical to RGC development. We – synaptic partners, physiological properties, and roles within the find that Brn3b and Brn3a control only small subsets of Brn3 – visual circuit, mouse RGCs can be cataloged in 20–30 different RGC specific transcripts. We identify extensive combinatorial types (5–10). Some of the developmental mechanisms by which sets of RGC transcription factors and cell surface molecules and RGC features are combined to determine RGC types are begin- show that several RGC-specific genes can induce neurite-like ning to be uncovered. Mouse RGCs become postmitotic and start processes cell autonomously in a heterologous system. exhibiting specific molecular markers and morphological features Author contributions: S.S., M.G.G., F.K., K.C., O.P., and T.C.B. designed research; S.S., M.G.G., around embryonic day 11 (E11). As soon as E12, RGC axons cross M.B., F.K., K.C., and T.C.B. performed research; S.H. and Z.W. contributed new reagents/ the midline at the optic chiasm, and by E15, the first axons have analytic tools; S.S., M.G.G., M.B., and T.C.B. analyzed data; and S.S., M.G.G., O.P., and T.C.B. reached the superior colliculus (SC), the most remote retinor- wrote the paper. ecipient area of the brain (11, 12). RGC axons invade their target The authors declare no conflict of interest. nuclei only around birth, and the first 10 postnatal days are the This article is a PNAS Direct Submission. most active period for synapse formation. RGC dendritic arbors Data deposition: The next generation sequencing data reported in this paper have been develop mostly postnatally, with lamination within the inner deposited in the Gene Expression Omnibus (GEO) database, https://www.ncbi.nlm.nih. gov/geo (accession no. GSE87647). plexyform layer clearly visible at postnatal days 3–4(P3–P4) and – 1Present address: Yonehara Laboratory, Danish Research Institute of Translational Neu- reaching a nearly mature distribution by P7 (13 16). Combinato- roscience, Aarhus University, 8000 Aarhus, Denmark. rial transcriptional regulation may play a major role in RGC type 2Present address: Emerson Laboratory, Biology Department, The City College of New specification. Previous work suggests the following transcriptional York, New York, NY 10031. – – cascade: the basic helix loop helix (bHLH) transcription factor 3Present address: Scientific Computing Core, Max Planck Institute for Brain Research, (TF) Atoh7 is expressed in RGC precursors and controls the ex- Frankfurt am Main 60438, Germany. pression of the POU4 family TF Brn3b and the Lim domain TF 4Present address: School of Medicine, Yale University, New Haven, CT 06510. Isl1, which are both required for the initiation of the RGC tran- 5To whom correspondence should be addressed. Email: [email protected]. – scriptional program (17 25). Further downstream TFs include This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. Brn3a, Brn3c, Eomesodesmin (Tbr2), Ebf1, Ebf3, Onecut1, and 1073/pnas.1618551114/-/DCSupplemental.

E3974–E3983 | PNAS | Published online May 2, 2017 www.pnas.org/cgi/doi/10.1073/pnas.1618551114 Downloaded by guest on September 30, 2021 KO). When paired with a WT allele, recombination results in AP- PNAS PLUS A C i) Genes enriched in RGCs: tagged heterozygote cells that are phenotypically WT (Brn3aAP/WT or Brn3bAP/WT RGCs; WT). Although recombination happens E15 Brn3AP/WT RGCs vs. Other (supernatant) throughout the retina, other retinal cell types do not express Axon Guidance either Brn3a or Brn3b and therefore, appear AP-negative (6, 13, ii) Brn3-dependent RGC Genes: 18, 21, 25, 31, 41). Using this genetic labeling strategy, we can B compare several cell populations (Fig. 1C). (i) Comparing the Brn3AP/WT RGCs vs. Brn3AP/KO RGCs expression profiles of Brn3AP/WT RGCs and retinal supernatants, we can identify genes specific for or enriched in RGCs. (ii) RGC P3 iii) Genes Specific for distinct RGC populations genes regulated by a Brn3 TF should be differentially expressed Dendrite Formation in Brn3AP/WT vs. Brn3AP/KO RGCs. (iii) Genes specific for Axon - target interaction + − − + + + Synapse formation Brn3aAP/WT RGCsvs. Brn3bAP/WT RGCs Brn3a Brn3b , Brn3a Brn3b , or Brn3a Brn3b RGC populations can be identified by comparing expression profiles of D F iii AP/WT AP/WT Supernat Brn3a with Brn3b RGCs. We dissociated retinas + from Pax6α:Cre;Brn3aCKOAP/WT,Pax6α:Cre;Brn3aCKOAP/KO, RGC Pax6α:Cre;Brn3bCKOAP/WT, and Pax6α:Cre;Brn3bCKOAP/KO mice Brn3AP αAP AP Other cells and isolated the AP-expressing Brn3 RGCs using magnetic RGCs Beads iii iv beads coupled to anti-AP mouse mAbs (Materials and Methods E ChTB P 3.5 and Fig. 1 D and F). We also have labeled the lateral geniculate P 0.5 nucleus (LGN), SC, and pretectal area (PTA) of P3 WT mice by anterograde tracing and dissected and processed them for deep sequencing (Materials and Methods and Fig. 1E). In the following, we will present gene expression data that are Fig. 1. Experimental goal and design. (A) E15 retina containing heteroge- restricted to the RefSeq (https://www.ncbi.nlm.nih.gov/refseq/) neous undifferentiated cells (gray) and RGCs (purple), which are mostly post- subset of mouse transcripts given its highest level of quality and mitotic and extend axons. (B) P3 retina with RGCs extending dendrites. RGC annotation confidence (Material and Methods). We use two al- axons are involved in synapse formation. (C) Comparison strategies. (i)RGC- ternative strategies to identify differentially expressed transcripts BIOLOGY enriched genes were identified by comparing E15 or P3 RGCs (Brn3aAP or

in our datasets. The first, differential expression analysis algo- DEVELOPMENTAL Brn3bAP) with AP-negative cells. (ii) Genes regulated by Brn3s were inferred by comparing Brn3bAP/WT or Brn3aAP/WT (heterozygote) RGCs with Brn3bAP/KO or rithm (DESeq), is the differential expression analysis as de- Brn3aAP/KO (KO) RGCs, respectively. (iii) To identify genes expressed selectively scribed by Anders and Huber (42), with the false discovery rate in RGC subpopulations, we compare Brn3bAP/WT with Brn3aAP/WT RGCs. set at 0.1 and the fold change at two. The second, referred to in (D) Immunomagnetic purification of RGCs. E15 or P3 retinas are dissociated, and the text as the “Twofold” criterion, identifies a transcript as Brn3aAP or Brn3bAP RGCs are separated using anti-AP–coated magnetic beads. differentially expressed in a given RGC population if it has RNA was extracted from either supernatant or Brn3AP RGCs-coupled beads expression levels in RGCs more than two fragments per kilo- and processed for RNASeq. (E) Retinorecipient brain tissue isolation from base per million reads (FPKM) and is enriched in RGCs P3.5 pups. Cholera toxin B (ChTB-AF488) was injected in the eyes of P0.5 mouse compared with the corresponding retina samples (FPKM levels pups. At P3.5, brains were isolated and vibratome-sectioned, and green fluo- in Brn3AP/WT RGCs ≥ 2 × FPKM levels in Brn3AP/WT retina rescent regions [LGN, shown in Right, olivary pretectal nucleus (OPN), medial supernatant). If the transcript also has a more than twofold change terminal nucleus (MTN), and SC] were dissected. All residual brain regions were AP/WT AP/KO pooled and used as controls. (F)The10-μL samples from each RGC immuno- between Brn3 and Brn3 RGCs, it is also considered purification were spread on slides and stained for AP and DAPI: (i–iii)three “regulated” by the respective Brn3. examples of Brn3AP RGCs coupled to magnetic beads (top), DAPI nuclear counterstain (middle), and merged images (bottom). Note occasional DAPI- Deep Sequencing and Sample Quality Control. Because RGCs are a positive AP-negative cells (bead cluster on the right in i). (iv) Retina superna- rare retinal cell population (around 0.5%) (43), bead-coupled tant after immunopurification showing a Brn3AP RGC (arrowhead) and a RGCs from several mice (typically six to eight retinas) of identi- retinal pigment epithelium cell (arrow). Estimation of yield and purity is in cal genotype were pooled to constitute one RGC sample. For each μ μ Materials and Methods and Dataset S1.(Scalebars:E,250 m; F,40 m.) genotype described, two RGC samples and one retina supernatant sample were submitted to RNA extraction, reverse transcription, populations. We focused this study on E15 RGCs to identify axon amplification, and sequencing using an Illumina Sequencing Materials and Methods guidance molecules (Fig. 1A) and P3 RGCs to detect potential platform ( ). For WT brain regions, three replicates for LGN and SC (derived each from one mouse) and determinants for dendrite formation and synaptogenesis (Fig. two replicates for whole-brain homogenates (pooled whole-brain 1B). Furthermore, we isolated RNA from retinorecipient areas nonvisual regions; each from one mouse) were submitted to se- at P3 to identify differential markers for the cellular targets that quencing, and one PTA sample was generated by pooling tissues might interact with incoming RGC axons (Fig. 1E). We report a from three individual mice. Correlation coefficients between bi- combinatorial code of TFs and cell surface molecules (CSMs) ological replicates were very high (> 0.95) and between samples expressed in different RGC subpopulations and retinorecipient (e.g., RGC vs. retina) were lower, ranging from 0.55 to 0.95 (Fig. areas of the brain. A subset of these genes can intrinsically in- 2A, Dataset S2,andFig. S1A). Hierarchical clustering across all duce arbor-like processes in epithelial cells, suggesting cell- transcripts expressed in our samples reveals a good correlation of autonomous neuronal arbor formation mechanisms. replicates by RGC genotype, retina, age, or brain region (Fig. 2B and Dataset S3), confirmed also by principal component analysis Results (Fig. S1 B–D). We determined the success of our cell purification Screen Rationale and Sample Collection. We have induced retina- and deep sequencing strategy in several ways. Visualization of the specific Cre recombination of our conditional knock-in reporter reads mapping to the endogenous Brn3a and Brn3b loci (Fig. 2 C–E) CKOAP CKOAP alleles at the Pou4f1 (Brn3a )andPou4f2 (Brn3b )loci showsdramaticallyhigherlevelsinBrn3aAP/WT over Brn3aAP/KO using the Pax6α:Cre driver (13). When the sister chromosome [adjusted means = 85.68 WT, 2.81 KO; DESeq P = 3.59 e-27, ad- carries a conventional KO allele, recombination of the conditional justed p value (padj) = 5.02e-23; t test P = 8.67e-04] and Brn3bAP/WT allele results in cells lacking both copies of the endogenous gene over Brn3bAP/KO RGCs at both P3 (adjusted means = 78.67 WT, labeled by the AP marker (Brn3aAP/KO or Brn3bAP/KO RGCs; 6.43 KO; DESeq P = 3.13 e-23, padj = 4.4e-19; t test P = 0.045) and

Sajgo et al. PNAS | Published online May 2, 2017 | E3975 Downloaded by guest on September 30, 2021 AP/WT AP AP/WT A P3 Brn3a RGC P3 Brn3a RGC P3 Brn3a C Brn3a Brn3b S2 WT RGC E 3 AP P3 (Pou4f1) (Pou4f2) 2 Brain 1

0 Reads

log FPKM -1 AP/KO -2 RGC-Brn3a -P3 36 S1 KO retina -2 -1 3210 -2 -10- 321 2-13210 log FPKM

7 Reads B 1 2 3 4 5 6 AP/WT RGC-Brn3a -P3 887 P3 RGC Retina 0.5 Kb

E15 E15 P3 FPKM FPKM D 0 204060800 20 40 60 80 100

AP AP F G 60 Reads 50 RGC-Brn3bAP/KO-P3 74 40

30 Reads P3E15E15 P3 20

RGC-Brn3bAP/WT-P3 RGC Retina WT-1 WT-2 KO-1 KO-2 KO-1 KO-2 WT-1 WT-2 KO WT KO-1 KO-2 WT-1 WT-2 611 10 Brn3a RGC ret Brn3b SC-2 SC-3 SC-1 PTA Brain-2 Brain-1 LGN-1 LGN-2 LGN-3

Brn3a-KO Brn3a-WT Brn3b-KO Brn3b-WT FPKM P3-RGC Brn3b-E15 P3-RGC 0.5 Kb P3-ret 0 20406080100120 Isl1 Isl2 Nefl Trhr Tbr1 Fstl4 Thy1 Sncg Snca Nefm Cdh6 Jam2 −3 −2 −1 0 1 2 3 Opn4 Cartpt Eomes Rbpms Rbfox3 Pou4f1 Pou4f2 Pou4f3

Fig. 2. Sample characterization and validation. (A) Log-scale scatter plots comparing FPKM levels. (Left) Comparison of two samples (S1 vs. S2) derived from P3 Brn3aAP/WT RGC; R = 0.9915. (Center) Means of two P3 Brn3aAP/KO RGC samples (KO) vs. means of two P3 Brn3aAP/WT RGC samples (WT); R = 0.9922. (Right) Comparison of a P3 Brn3aAP/WT retinal supernatant (retina) with the mean of two P3 Brn3aAP/WT RGC samples (RGC); R = 0.7761. Red diagonals separate the twofold comparison lines, and the red corners enclose genes with less than two FPKM for both samples in the plot. (B) Clustergram across 18,185 transcripts that were expressed at greater than or equal to one FPKM in at least one of the samples. Clustering was performed on standardized sample values, first along the sample dimension (columns) and then along the transcript dimension (rows) (Materials and Methods). Branches are color-coded and labeled 1–7asfollows:branch1, Brn3aAP P3 RGCs; branch 2, E15 retina and Brn3bAP RGCs; branch 3, P3 retinas; branch 4, Brn3bAP P3 RGCs; branch 5, SC and PTA; branch 6, whole-brain controls; branch 7, LGN. For each sample (along the bottom), numbers indicate biological replicates. Color scale represents units in SDs of the distribution across all observations for each given row (gene). [Dataset S2 shows cross-correlation matrix of all samples, and Fig. S1 shows additional scatter plots and principal component analysis (PCA).] (C and D) Visualization of mapped reads. (C) Reads from (Upper)Brn3aAP/KO and (Lower)Brn3aAP/WT P3 RGCs mapping to the Brn3a locus. (D)Reads from either (Upper)Brn3bAP/KO or (Lower)Brn3bAP/WT P3 RGCs mapping to the Brn3b locus. The x axis is in kilobases (notches every 0.5 kb). The y axis is scaled to the highest read stack (indicated in the bottom right corner). The AP cDNA inserted in the recombined alleles is indicated. Gray bars flanked by black notches represent reads. Thin blue lines represent spliced reads reaching across two exons. Exons (rectangles) and introns (lines) are shown for Brn3a (Pou4f1, three exons) and Brn3b (Pou4f2, two exons) in C, Lower and D, Lower. Coding regions within exons are blue. (E) Expression levels (FPKM) for Brn3a and Brn3b genes. Mouse WT P3 brain samples are (from top to bottom) whole brain (white; median of two samples), PTA (black; one sample derived from three mice), and SC and LGN (dotted and gray bars, respectively; each medians of three samples). For retina and RGCs, samples are (from top to bottom) P3 Brn3bAP/WT (dark red), Brn3bAP/KO (light red), Brn3aAP/WT (dark green), Brn3aAP/KO (light green), E15 Brn3bAP/WT (dark blue), and Brn3bAP/KO (light blue). Retina values represent individual retinal samples, and RGC samples represent medians of two samples. (F) Expression (FPKM) of the knocked in AP cDNA color-coded as in E.(G)Heatmapforknown general and subtype-specific RGC markers. Expression levels are normalized to the maximum level for each gene and displayed on a 64-level scale (red is high).

E15 (adjusted means = 86.93 WT, 20.41 KO; DESeq P = 6.45e-13, selectively expressed in these partially overlapping cell populations. padj = 8.67e-10; t test P = 0.0405). The residual reads in Brn3aAP/KO We find that, at P3, a large number of transcripts are enriched in and Brn3bAP/KO RGCs are mapping to the 5′ and 3′ UTRs, RGCs, with most being selective for Brn3aAP/WT RGCs (1,423 consistent with the replacement of the endogenous coding exons by DESeq and 1,667 by Twofold), some being common to Brn3aAP/WT with the AP ORF and preservation of the 5′ and 3′ Brn3a and and Brn3bAP/WT RGCs (994 by DESeq and 1,285 by Twofold), Brn3b UTRs (Fig. 2 C, Upper and D, Upper). Compared with the and only a small number being selective for Brn3bAP/WT RGCs (71 by retina or the brain samples, RGC samples have higher expression DESeq and 407 by Twofold) (Fig. 3A, Dataset S4,andFig. S1E). levels for Brn3a (adjusted means = 86.63 for WT RGCs, 5.34 for Only a small fraction of these RGC-enriched transcripts seem to be P3 retinas; DESeq P = 9.21 e-52, padj = 1.61e-49; t test P = 6.57e-05) regulated by Brn3b or Brn3a. Strikingly, most of these are selec- and Brn3b transcripts at both P3 (adjusted means = 78.85 for WT tively affected by Brn3b ablation, and only very few are Brn3a- RGCs, 3.19 for P3 retinas; DESeq P = 4.37 e-51, padj = 2.18 e-48; dependent (Fig. 3 C and D, Dataset S4,andFig. S1F) in keeping t test P = 0.0015) and E15 (adjusted means = 86.95 for WT RGCs, with the more dramatic effect of Brn3b ablation on RGC devel- 12.21forE15retinas;DESeqP = 4.53 e-15, padj = 2.79 e-12; t test opment. Interestingly, the repertoire of transcripts enriched in P = 0.036) (Fig. 2E). Reads aligned to the AP cDNA [using a Brn3bAP/WT RGCs changes significantly between E15 and P3 (Fig. “minigenome” generatedwithBowtie(44)](Materials and Methods) 3B, Dataset S4,andFig. S1G). This fact may reflect the distinct display the expected distribution, with essentially no reads in retina RGC-specific programs required at the two ages. The set of Brn3b- samples and high levels of the transcripts in the RGC samples (Fig. dependent genes expressed by Brn3bAP RGCs is also dramatically 2F). Of note, in E15 but not P3 RGCs, the AP reporter is expressed distinct at the two ages (Fig. 3 C and D, Dataset S4,andFig. S1H). at much higher levels in the Brn3bAP/KO than in the Brn3bAP/WT The differences between Brn3a- and Brn3b-positive RGC tran- RGCs. Thus, the higher levels of Brn3b reads in the Brn3b KO at scriptomes may reflect differences in cell type distribution, E15 could be explained by the AP transcript carrying partial Brn3b whereas the temporal differences may define the distinct func- 5′ and 3′ UTRs along. Finally, previously described RGC markers tionality required in the early (E15; axon guidance) vs. late (P3; are visibly enriched in Brn3AP RGCs (Fig. 2G). Thus, we are con- dendrite formation, synaptogenesis, and myelination) stages of fident that our cell sorting, RNA isolation, and deep sequencing RGC maturation. An analysis of Gene Ontology terms among our approach is successful. candidate genes shows significant enrichment for known neuronal- associated processes and pathways, such as neurotransmitter re- Outcome of RGC Purification. We compared expression data from ceptors and release, voltage-gated channels, signaling cascades, Brn3aAP/WT and Brn3bAP/WT RGCs to identify genes that are synaptic transmission, neuronal projection, etc. Among these pathways,

E3976 | www.pnas.org/cgi/doi/10.1073/pnas.1618551114 Sajgo et al. Downloaded by guest on September 30, 2021 PNAS PLUS Transcripts enriched in Brn3AP RGCs vs. retina Brn3a and/or Brn3b dependent RGC transcripts G

AP P3 Brn3a Downregulated Upregulated Brain ABCD Pax6 P3 AP P3 Myt1l Rgs4 Slc18a2 Myt1 Tsc22d3 71 Brn3b 7 32 49 E15 E15 4 P3

36 RGC Retina 1423 994 202 863 4 186 158 E15 10 20 30 40 50 150 250 20 60 100 5 10 15 20 40 60 80 5 10 15 20 70 2 P3 33 47 H P3 Brn3bAP Brn3a - WTvsKO - P3 Brn3bAP Brn3b - WTvsKO - P3 E Brn3b - WTvsKO - E15 P3 Brain Myt1l Rgs4 Slc18a2 Myt1 Pax6 Tsc22d3 P3 Cpne4 Igsf6 Rab6b Nrxn1 Itga6 Lrfn3 E15 P3 In Situ Hybridization Screen E15 In Situ Hybridization (Allen Brain Data) P3 IJ RGCE15 Retina Brn3-RGC-enriched Brn3-RGC-regulated Brn3b-RGC-enriched Brn3b-RGC-regulated 10 30 50 70 2 4 6 8 50 150 250 10 30 50 70 2 4 6 8 5 10 15 1 123 58 30 7 37 16 19 Cpne4 Igsf6 Rab6b Nrxn1 Itga6 Lrfn3 F 5

22 4 5 20 RGC specific RGC enriched Whole retina Negative

Fig. 3. Only a small fraction of genes enriched in Brn3+ RGCs depend on Brn3a or Brn3b. (A–D) Venn diagrams representing transcripts enriched in Brn3AP RGCs. Transcripts reported here are derived from DESeq (42), with a 0.1 false discovery rate and a twofold change between compared conditions. (A and B) Transcripts enriched in Brn3AP RGCs over retina supernatants. (A) Partially overlapping gene expression profiles in P3 Brn3aAP/WT RGCs and Brn3bAP/WT RGCs. (B) Partially overlapping gene expression programs in E15 and P3 Brn3bAP/WT RGCs. (C and D) Transcripts (C) down- or (D) up-regulated in Brn3AP RGCs as a result of either Brn3a or Brn3b loss. (E–H) RGC-enriched transcripts validation by ISH. (E and G) FPKM values (indicated on the X axis) for RGC-enriched candidate genes. Plots are scaled to individual maximum levels. Sample color coding as in Fig. 2E.(F) P3 ISHs with probes directed against the 3′ UTR. Note the spectrum of outcomes from RGC-specific (Rab6b and Cpne4) to RGC-enriched (Igsf6 and Nrxn1) and from full retina expression (Itga6) to lack of expression (Lrfn3). (H) In situ analysis of RGC-specific genes at E15.5. Expression images reproduced with permission from the Allen Brain Institute atlas. (I) Pie chart summaries of 223 genes analyzed by ISH at P3 broken down by genes predicted by RNASeq to be RGC-enriched (Brn3–RGC-enriched from A)orBrn3- regulated (Brn3–RGC-regulated from C and D). (J) Pie chart summaries of 58 genes found in E15 ISHs. Fig. S1 K–N shows control experiments for the P3 and E15 in situ screens. Dataset S4 lists all of the transcripts in the Venn diagrams and pie charts. (Scale bar: F, 100 μm; H,400μm.) BIOLOGY DEVELOPMENTAL adhesion molecules and TFs represented each about 5–10% of the seems that DESeq is a more stringent selection method; how- identified genes, regardless of the selection criteria (Dataset S4). ever, it may miss many useful candidates compared with our Twofold criterion. Overall, more than one-half of the targets In Situ Validation. Because the starting RGC population from identified by deep sequencing show RGC specificity in ISH re- which RNA was extracted is relatively small and the degree of sults, but expression levels detected with the two methods amplification is quite significant, we sought to validate a subset differ significantly. of about 10% of the identified RGC-enriched molecules with an independent technique. We designed in situ hybridization (ISH) Retinorecipient Nuclei Transcriptomes and ISH Validation. To iden- probes for the 3′ UTR of 233 genes from our candidate pool and tify molecules that may be conferring specificity for the distinct tested them against WT P3 retinas (Fig. 3 E, F, and I, Dataset S4, retinorecipient brain areas, we have mined our RNASeq data for and Fig. S1 I, K, and L). Genes were selected based on literature transcripts enriched in SC, LGN, or PTA compared with the searches for potentially interesting yet less well-studied pathways whole-brain homogenate. Fig. 4 A and B show that essentially all related to neuronal development and exhibit a broad range of possible combinations can be found, with transcripts common to expression levels. In addition, we searched the Allen Brain In- all three nuclei, common to only two of them, or selective for stitute mouse brain development ISH atlas (developingmouse. each nucleus individually (complete lists are in Dataset S5). Of brain-map.org/) with our Brn3bAP/WT RGC E15 candidate gene these candidate genes, 122 (LGN), 116 (PTA), and 134 (SC) were tested by ISH by the Allen Brain Institute (examples are in list and identified 265 genes for which E15 eye sections were – available (Fig. 3 G, H, and J, Dataset S4, and Fig. S1 J, M, and Fig. 4 C H). For the LGN and SC, between 55 and 75% of N). There did not seem to be a strong correlation between RNA RNASeq-predicted genes were specific or regionally expressed sequencing (RNASeq) FPKM expression levels and intensity of in the expected nucleus, with others being broadly expressed or I J in situ signal in the ganglion cell layer (GCL) at P3. As an ex- negative (Fig. 4 and ). The number of specific transcripts was much lower for the PTA. Interestingly, several transcripts ample, in P3 Brn3AP/WT RGCs, Ig superfamily 6 (Igsf6) is showed lamination within the SC (Gpc3, Barhl1, and Foxb1), expressed at only 4–7 FPKM, whereas Rab6b has ranges between consistent with the possibility that these markers are selective for 150 and 300 FPKM (Fig. 3E). Whereas both genes are enriched SC functional laminae (Fig. 4 E and F). Of note, applying DESeq in the GCL, the signal to noise ratio for Igsf6 is much higher than modestly increased the ratio of positive hits by ISH compared that in Rab6b, although probes are comparable in length and with our Twofold criterion, while reducing the total number of melting temperature. In a similar fashion, Rgs4 and Myt1l have correct hits, similar to what we observed for the RGC data. We, comparable ISH GCL-specific signals at E15 (Fig. 3H), but AP/WT therefore, present throughout the text results based on both Rgs4 has almost six times higher expression in E15 Brn3b selection criteria. RGCs compared with Myt1l (75 vs. 13 FPKM). Nevertheless, about 60–75% of candidate genes predicted by RNASeq to be TF Program of Brn3AP RGCs and Retinorecipient Areas. TFs play a enriched in RGCs over the retina were confirmed by our in situ significant role in neuronal cell type diversification. We, there- screens at either P3 or E15. The enrichment with positive hits fore, compared our data with a merged list of 2,437 TFs and was not dramatically affected by using either the DESeq or transcriptional regulation-associated genes compiled by com- Twofold selection criteria (compare Fig. 3 I and J with Fig. S1 I bining recently published surveys (45, 46). Of these genes, almost and J; Dataset S4). Confirmation by ISH was consistently better one-third (1,647) had expression levels of more than one FPKM for Brn3-regulated genes identified by DESeq (Fig. 3 I, Right and in our RGC samples, but a more restricted subset was enriched J, Right) compared with the Twofold criterion (Fig. S1 I and J), in RGCs compared with the retina (DESeq = 153, Twofold = but the total number of hits was smaller (Dataset S4). Thus, it 322) (Fig. 5 A, C, and E and Dataset S6). An even smaller set of

Sajgo et al. PNAS | Published online May 2, 2017 | E3977 Downloaded by guest on September 30, 2021 P3 A C Cck E Gpc3 Brain LGN SC P3 E15 SC 99 23 183 LGN P3 17 40 45 RGCE15 Retina H Cck Esrrb Gpc3 Barhl1 Foxb1 143 0 100 200 300 0 5 10 15 0 5 15 25 0 4 8 12 0 5 10 15 PTA P4 ISH (Allen) F Barhl1 Nucleus Enriched Genes Nucleus Selective Genes SC LGN PTA SC LGN PTA SC 31 14 12 18 30 10 25 5 4 7 19 5 BLGN I SC 4 D 3 74 Esrrb 324 221 6 DESeq 5 117 9 5 107 85 Foxb1 48 27 73 58 60 17 11 5 8 18 5 PTA G SC J 386 PTA 2

Twofold 16 PTA 63 11 3 3 16 Specific/Regional Broad Negative

Fig. 4. Combinatorial gene expression in P3 retinorecipient brain nuclei. (A and B) Venn diagram comparisons of enriched unique or shared genes in LGN-, SC-, and PTA-derived samples. (A) Significantly differentially expressed transcripts identified by DESeq. (B) Transcripts passing the Twofold criterion. (C–G)ISH patterns from Allen Brain Institute atlas at P4 for genes predicted to be nucleus-specific. (C) Cck expression in LGN. (D) Esrrb expression in PTA. (E–G) Gpc3, Barhl1, and Foxb1 expression in three distinct layers of SC. Foxb1 is also expressed in the PTA. Insets show complete sagital brain sections for the genes, documenting expression in additional brain regions. (Scale bars: C–G,200μm.) (H) FPKM values across brain regions, retina supernatants, and RGCs for genes presented in C–G. Sample color coding as in Fig. 2E.(I and J) Allen Brain Institute atlas validation outcomes for genes identified in our screen by criteria used in A or B broken down by genes predicted to be expressed only in one retinorecipient nucleus (selective) or expressed at higher levels in two or more retinorecipient nuclei (intersection sets, enriched). From our candidate lists, 122 (LGN), 116 (PTA), and 134 (SC) were present in the Allen Brain Institute atlas, and of those, more than one-half were nucleus-specific for the LGN and SC, but only about one-quarter were nucleus-specific for the PTA. Transcripts in all Venn diagrams and pie charts are listed in Dataset S5.

genes was Brn3a- and/or Brn3b-dependent (DESeq = 43, genes, about one-quarter (DESeq = 156, Twofold = 237) were Twofold = 95) (Fig. 5 B, D, and E and Dataset S6). Fig. 5H differentially expressed in RGCs (Fig. 6 A and C and Dataset S7), shows an unsupervised clustering of 38 TFs previously implicated andanevensmallernumber(DESeq= 13, Twofold = 93) were in RGC development. Interestingly, samples are first separated by regulated by Brn3 TFs in RGCs (Fig. 6 B and D and Dataset S7). age (P3 vs. E15) regardless of retina or RGC origin (Fig. 5H, Only very few (DESeq = 20, Twofold = 69) were differentially branches a + b vs. c). In addition, at both P3 and E15, Brn3bAP/KO expressed in retinorecipient areas of the brain (Fig. 6E and Dataset (Brn3b KO) RGCs are grouped closer to the respective retina S7). Partially overlapping sets of CSMs were selective for Brn3a supernatants (e.g., Fig. 5H, branch b) than to the other RGC and Brn3b (Fig. 6A) and dynamically expressed during embryonic samples(e.g.,Fig.5H, branch a) (containing Brn3aAP/WT, and early postnatal development in RGCs (Fig. 6C). The number Brn3aAP/KO,andBrn3bAP/WT RGCs at P3). This similarity sug- of differentially regulated transcripts in Brn3bAP RGCs increases gests that loss of Brn3b turns Brn3bAP/KO RGCs into a more betweenE15andP3(Fig.6D). A clustergram example of 37 ad- undifferentiated “whole-retina” state. When contrasting the hor- hesion molecules and/or guidance receptors implicated in RGC izontal branches of the hierarchical tree, we can distinguish a development (Fig. 6F) reveals several interesting patterns. P3 “general RGC” group (branch 2, Pou4f1, Pou4f2, Irx2, Irx3, Brn3bAP RGCs are more closely grouped with the retina samples Irx4 and Irx6, Ebf1, and Ebf3), a “P3 RGC” group (branch 1, Isl2, (Fig. 6F, vertical branch d and horizontal branch 1) and share in- Pou4f3, Klf2, Klf8, Myt1l, and Tbr1), and a “Brn3b-only RGC” creased expression of several molecules implicated in marking and/or group (branch 4, Nhlh1, Eomes, Irx1, and Tbx20). A subset of TFs determining lamination in the retina (Jam2, Plxna2, Robo3, in branch 3 shows up-regulation in Brn3bAP/KO vs. Brn3bAP/WT Epha8, Sema5a, and Sema5b) (47–50). A second group is visibly RGCs at E15, consistent with their suppression by Brn3b in RGCs enriched in E15 Brn3bAP RGCs (Fig. 6F, branch 2). Among (Atoh7, Dlx1, Dlx2, Onecut2, Barhl2, Eya2, and Zic2). Many of them, many receptors and/or ligands implicated in axon guid- these TFs are known to be required upstream or in parallel with ance are enriched in RGCs on Brn3b loss (Sema3a, Plxna3, Brn3b in the RGC class specification cascade. Similar patterns are Plxna1, Epha6, Efnb2, Efna4, Efna5, Cntn2, Nfasc, Slit1, seen over the full datasets containing RGC-enriched or Brn3- and Tenm3) (Fig. 6F, column 1: compare Brn3bAP/WT with dependent TFs (Dataset S6 and Fig. S2). A similar-sized set of Brn3bAP/KO RGCs) (51–53). Finally, clusters 3 and 4 seem to be TFs is enriched in the retinorecipient areas, with partially over- highly expressed in most brain regions as well as RGCs at P3. lapping expression between all three nuclei (Fig. 5 F and G, Fig. S4 (Dataset S7) shows the clustergram for all Brn3a- and Dataset S6,andFig. S3). The identified sets of TFs and the Brn3b-dependent RGC-enriched transcripts identified in this powerful combinatorics provided by our approach thus yield a study. It highlights largely three groups. Branch 1 is P3 RGC- large set of potential transcriptional regulators of RGC and/or specific and mostly Brn3b-dependent. Branch 2 is specific to E15 retinorecipient nuclei cell types. Brn3b RGCs and either negatively or positively regulated by Brn3b. Finally, branch 3, by far the largest group, is common CSMs Involved in Cell–Cell Interactions, Axon Guidance, and Neurite to brain regions and P3 RGCs. Most of the molecules in this Formation. A variety of transmembrane proteins involved in cell– subgroup are down-regulated in Brn3bAP/KO compared with cell, cell–matrix, and receptor–ligand interactions (e.g., Integrins, Brn3bAP/WT RGCs. The extended set of 237 RGC-enriched ad- Cadherins, Igs, Leucine Rich Repeats, Ephrins, Semaphorins, hesion molecules (Dataset S7 and Fig. S4) follows similar trends. Plexins, Robo, and Tenneurins) are required for guiding axon and Thus, our screen defines combinatorial expression of many sur- dendrite formation, and/or establishing specific synaptic interac- face ligands and receptors implicated in cell–cell adhesion, tions between neuronal cell types. We, therefore, queried the Gene neurite formation, and synapse specificity. Brn3b seems to play a database (National Center for Biotechnology Information) and role in regulating many of these neuronal identity determinants, established a comprehensive list containing 822 genes that include thus explaining its important role in axon guidance and dendrite one or more of these protein domains in their structure. Of these formation, whereas only a few are regulated by Brn3a in keeping

E3978 | www.pnas.org/cgi/doi/10.1073/pnas.1618551114 Sajgo et al. Downloaded by guest on September 30, 2021 AC H ments. This observation prompted us to ask whether some of our PNAS PLUS P3 RGCs Brn3bAP RGCs RGC-specific genes could cell-autonomously support process Brn3aAP vs Brn3bAP E15 vs. P3 RGC Retina RGC

AP (neurite) extension and branching as proposed in other contexts Brn3a Brn3bAP E15 P3 P3 E15 P3 Brn3a Brn3b Brn3a Brn3b Brn3b Brn3b (56–58). To test this possibility, we overexpressed in HEK293 cells 123 84 42 167 54 72 a set of 10 candidates identified in our RNASeq data and analyzed

(92) (55) (8) (9) (18) (44) WT KO WT KO KO WT KO WT WT KO KO WT the effects on cellular morphology (Fig. 7). The 10 genes had been

Isl2 Pou4f3 confirmed by in situ and belong to molecular families known to af- BD 1 Klf8 17 7 45 39 14 38 Klf2 fect cellular morphology, but they were relatively unstudied in this (1) (1) (26) (16) (6) (21) Myt1l Tbr1 Brn3a Brn3b E15 P3 Pou4f1 context. They range from TFs (Irx4) to cytoskeletal adaptors WTvsKO WTvsKO WTvsKO WTvsKO Irx4 – 2 Irx6 (Ablim3 and Stmn3), molecules mediating cytoskeleton membrane Pou4f2 Overall Transcription Factor Expression Ebf1 Irx2 interactions [S100a10, copines (Cpne), and Epb4.1l3], membrane Irx3 2437 E Regulated by Brn3a or Brn3b Ebf3 folding or trafficking [Reep5 and reticulon receptor 4rl1 (Rtn4rl1)], Pou6f2 Enriched in RGCs Isl1 or putative cell–cell adhesion molecules (Rtn4rl1, PcdhA1, and 1647 Klf7 Expressed in RGCs Barhl2 Dlx2 Igsf) (Fig. 7 and Dataset S8). Target genes were overexpressed 322 Zic2 (153) All TF Dlx1 95 Onecut2 using transfection into HEK293-Cre cells using a Cre-dependent (43) G WT P3 Atoh7

Brain Eya2 PTA LGN FLEX approach (Fig. 7A) (59). The intracellular localization of SC 3 Sox11 Onecut1 FLGN SC Onecut3 the expressed gene can be tracked using its HA tag, whereas a 19 41 Sox2 45 (3) (23) Tbx18 Tle1 membrane-attached EGFP (meGFP) reveals the effects on cell (26) Zfhx4 Sox4 16 18 Nhlh2 16 Fbxl19 Nhlh1 (0) (10) (5) Sox14 Eomes Pax3 Irx1 50 4 Tbx20 Bhlhe23 (11) Irx5 60 Otx2 5 Tfap2a PTA Tfap2b 50 Tfap2b Pou4f1 A F 40 Pou4f2 ab c P3 RGCs RGC Brain RGC Retina AP AP 30 Sp5 Brn3a vs Brn3b Smyd1 E15 P3 P3 P3 P3 E15 Brn3b Brn3a Brn3b Brn3a Brn3b Brn3b

20 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 Brain LGN Sp9 AP PTA

Brn3a SC Brn3bAP 10 Zic1 Zic4 72 105 21 KO WT KO WT WT KO KO WT WT KO WT WT KO (69) (80) (3) BIOLOGY Fig. 5. TF repertoire of Brn3aAP and Brn3bAP RGCs and retinorecipient brain – AP DEVELOPMENTAL areas. (A D) Venn diagrams representing TF genes enriched in Brn3 RGCs. Jam2 4 Plxna2 Numbers represent transcripts identified using the Twofold and DESeq B (0) 1 Robo3 (parentheses) criteria. Note that the DESeq criteria find a much smaller 8 63 Epha8 (0) (10) Sema5a number of significantly differentially expressed TFs. (E) Venn diagram rep- Brn3a Brn3b Sema5b resenting global TF expression in Brn3AP RGCs. Of a list of 2,437 TFs and WTvsKO WTvsKO Efnb2 Efna4 transcriptional activity molecules (45, 46), 1,647 are expressed at more than Efna5 AP AP AP Brn3bAP RGCs Cntn2 one FPKM in Brn3 RGCs. TF genes enriched in Brn3a and/or Brn3b RGCs E15 vs. P3 Plxna3 at E15 and/or P3 number 322 by the Twofold criteria (more than two FPKM Epha6 Nfasc P3 in RGCs and more than twofold in RGCs compared with retina) and C Sema3a Slit1 153 by the DESeq protocol. TF genes differentially expressed in RGCs in a 59 64 62 2 Tenm3 Brn3a- and/or Brn3b-dependent manner are either 95 (Twofold criteria) or (4) (13) (73) Plxna1 Nrp1 43 (DESeq). (F) Venn diagram with TFs selectively expressed in LGN, SC, and E15 Sdk1 Fstl4 PTA. Numbers represent transcripts identified using the Twofold criteria and F11r 11 D (0) Sema3e DESeq (parentheses). (G) Examples for TFs expressed either selectively or Sema6a jointly in three retinorecipient nuclei. Expression levels normalized to the 19 56 Robo2 (3) (10) Dscam maximum level of each gene and displayed on a 64-level heat map (red is E15 P3 Cdh12 WTvsKO WTvsKO Cdh6 high). (H) Clustergram of TFs believed to be important for RGC development. Cntn5 Sample names are labeled at the top, and hierarchical cluster major branches 3 Efna3 Retinorecipient areas Cntn1 are color-coded and labeled a–c. TFs (highest expressed transcript) are an- Nrcam SC Cdh4 notated to the right, and hierarchical tree branches of interest are color- ELGN 3 21 Cdh13 coded and labeled 1–5. Color scale bar is at the bottom, and units are in SDs (0) (8) Alcam Chl1 16 (clustergram details are in Materials and Methods and Fig. 2B). Clustergrams 12 4 Cdh8 (7) 4 (2) Cdh5 of the complete sets of TFs regulated by Brn3s, enriched in RGCs, or selective 1 (1) (2) 12 b c for particular retinorecipient areas are provided in Figs. S2 and S3. Dataset a (0) d S6 lists all transcripts in the Venn diagrams. PTA

−3 −2 −1 0 1 2 3

AP AP with its more specific role in RGC type specification. The clus- Fig. 6. CSM repertoire of Brn3a and Brn3b RGCs and retinorecipient – tergram of transcripts enriched in retinorecipient areas (Fig. S5) brain areas. (A D) Venn diagrams representing adhesion molecule genes from several molecular families believed to be important in neurite forma- reveals clusters specific for three investigated nuclei (LGN, PTA, tion that were enriched in Brn3AP RGCs. Expression criteria and comparison and SC) that seem to have little expression in the retina or RGCs. sets are identical to those in Fig. 5 A–D. The survey includes 822 genes, and However, about 30 transcripts are expressed in RGCs at levels the extended analysis is provided in Dataset S7.(E) Venn diagram for ad- equal to or higher than those in retinorecipient areas. Collectively, hesion molecules enriched or selective for retinorecipient areas. Expression the identified CSMs and guidance receptors could represent homo- criteria and comparison are the same as in Fig. 5F (Dataset S7). For A–E, or heterotypic interacting cell–cell adhesion or ligand–receptor numbers represent transcripts identified using the Twofold and DESeq (pa- pairs involved in targeting the axons of specific RGC types to these rentheses) criteria. (F) Clustergram of a subset of adhesion molecules be- nuclei or establishing cognate synaptic connections. lieved to be important for RGC development. Clustering algorithm and annotations are same as in Fig. 4. Sample names are labeled along the top, and hierarchical cluster major branches are color-coded and labeled a–d. Molecular Determinants of Neuronal Morphology. Dendritic arbor Genes (highest expressed transcript) are annotated on the right, and hier- branching patterns of neurons as diverse as RGCs and somato- archical tree branches of interest are color-coded and labeled 1–4. The color sensory neurons of the peripheral nervous system [dorsal root scale bar is at the bottom, and units are in SDs. Clustergrams covering the ganglia (DRG)] bear striking resemblances (6, 40, 54, 55), despite complete sets of differentially regulated or expressed adhesion molecules developing in distinct cellular and extracellular matrix environ- and guidance receptors are provided in Dataset S7 and Figs. S4 and S5.

Sajgo et al. PNAS | Published online May 2, 2017 | E3979 Downloaded by guest on September 30, 2021

Subcellular Localization of Candidate Genes in RGCs in Vivo. Con-

P2A HA

meGFP GeneX A ITR ITR CMV Promoter SV40 polyA sistent with the extensive genetic diversification of molecular families in vertebrates, many of the genes identified in our screen B pAAV_FLEX_GeneHA_meGFP or or are members of large molecular families. Thus, loss of function Cre Cre Cre Cre screens are not expected to yield dramatic results. However, we studied the subcellular distribution and gain of function effect ab c deb’ C for some of our candidate molecules by overexpressing them in c’ vivo in RGCs. Cre-dependent adeno-associated virus 1 (AAV1) viral vectors for S100a10, Rtn4rl1, Cpne4, and Igsf6 were injected in the d’ PTPY Irx4 Ablim3 Reep5 PcdhA1 retinas of Brn3bCre/WT and Brn3bCre/Cre mice (60) at P0, and sub- f g cellular distribution of the expressed genes together with dendritic e’ arbor morphologies of infected neurons were revealed by immu- f’ nostaining in adult mice (Fig. 8, Dataset S8,andFig. S6). Although g’ many RGC bodies were meGFP-positive, accurate imaging was S100a10 Stmn3 possible mostly for the large, sharply laminated dendritic arbor types hijk h’ (characteristic of ON and OFF alpha RGCs). None of the overex- i’ pressed molecules induced dramatic changes in dendritic arbor morphologies in either Brn3bCre/WT (WT) or Brn3bCre/Cre (KO) j’ neurons. As an example, dendritic arbor areas of labeled RGCs did not show any changes on either Brn3bCre/Cre (KO) or Brn3bWT/ Igsf6 Epb4.1l3 Rtn4rl1 Cpne4 k’ Cre(WT)RGCs(Fig.8,Dataset S8,andFig. S6). The subcellular

D 104 localization of the four molecules varied widely. S100a10, a non- ) 2

m calcium-binding member of the S100 family involved in membrane 103 processes, including signaling and membrane fusion (61, 62), was Area ( μ 102 restricted to RGC bodies (including the nucleus) and axon (Fig. 8A and Fig. S6A). Rtn4rl1 (Nogo receptor 1), implicated in axon Irx4 Igsf6 PTPY

Stmn3 navigation across myelinated substrates and endoplasmic re- Cpne4 Reep5 Ablim3 Rtn4rl1 PcdhA1 Epb4.1l3 S100a10 ticulum function (63, 64), is restricted to the cell body and some 2+ Fig. 7. Heterologous overexpression of several RGC-derived genes can affect proximal dendrites (Fig. 8B and Fig. S6B). Cpne4, a Ca binding HEK293 morphology. (A and B) Adeno-associated vector and the overexpression strategy used. A minimized CMV promoter drives an expression cassette flanked by tandem inverted lox sites (black, loxP; white, lox2272) and followed by a Cre/WT minimized SV40 polyadenylation signal. The expression cassette is in reverse Brn3b + P0 : AAV1.CMV.FLEX.GeneX-HA.T2A.mGFP.SV40polyA orientation and will be activated by inversion/excision induced by the tandem S100a10 Rtn4rl1 Cpne4 Igsf6 lox sites (FLEX strategy). It consists sequentially of the cDNA for the gene to be ACB D expressed (GeneX), a triple HA tag (HA), a self-cleaving T2A peptide, and EGFP coupled to a GAP43 membrane localization signal (meGFP). ITRs are viral inverted terminal repeats. After Cre-mediated inversion–excision, the two peptides are transcribed and separated on translation; the meGFP reveals the plasma membrane, whereas GeneX distributes in its expected subcellular compartment (e.g., nucleus, intracellular compartment, or plasma membrane) and can be traced by immunostaining with αHA. (C) Examples of HEK293-Cre

cells transfected with the pAAV_FLEX_GeneX_meGFP vectors. PTPY is a control Dend vector containing teal fluorescent protein tagged with a PSD95 domain and

membrane attached yellow fluorescent protein (meYFP). The overexpressed Soma genes are indicated (b–k′). Note extensive process formation or cell area en-

largements in HEK293-Cre cells expressing S100a10, Stmn3, Igsf6, Epb4.1l3, Axon Rtn4rl1, and Cpne4. For b–k, highlighted marquees are enlarged, and green E

′– ′ ) 2.5 and red channels are shown separately in b k .(D) All overexpressed vectors 2 exhibited highly significant cell area enlargements compared with the PTPY 2 mm control (P < 0.005) (Dataset S8). The y axis is in log10 scale. Box and whiskers 5 1.5 plots: the tops and bottoms of each box are the 25th and 75th percentiles of 1 the samples, respectively (interquartile ranges). The lines in the middle of 0.5 Area (x10 0 boxes are the sample median. Whiskers are drawn from the ends of the Brn3bCre WT KO WT KO WT KO Het KO interquartile ranges to the farthest observations within the whisker length S100a10 Rtn4rl1 Cpne4 Igsf6 (the adjacent values). Short red lines represent outliers. (Scale bar: C,20μm.) Fig. 8. Subcellular localization of candidate genes in RGCs in vivo. (A–D) Examples of adult RGCs from retinas of Brn3bCre/WT mice infected at membrane shape (Fig. 7B). As evident from Fig. 7C, large cel- P0 with AAV1 viral vectors for S100a10, Rtn4rl1, Cpne4, and Igsf6 (constructs lular processes reminiscent of lamellipodia are evident when are in Fig. 7). Retinas were whole mount-fixed and immunostained with αHA Stmn3 and Igsf6 are overexpressed, whereas more branched, (red) and αGFP (green). (Upper) Merged image of RGC in flat mount per- spective; white marquee squares outline soma and segments of dendritic arbor dendritic-like processes are evident in cells expressing S100a10, and axon. Lower shows red, green, and merge channels for the highlighted Epb4.1l3, Rtn4rl1, or Cpne4, often resulting in a nominal in- areas. White arrowheads point to meGFP marking dendrite and axon, and crease of cell area as defined by the bounding polygon (Fig. 7D white arrows point to HA-tagged gene localization in dendrite and/or axon. and Dataset S8). More modest but still significant changes are Note that, in D, axons and dendritic arbors of two RGCs are visible, and Igsf6 is visible in the case of Irx4, Ablim3, Reep5, or PcdhA1. Thus, localized to the axon arbor of only one of them. (E) Dendritic arbor area measurements for Brn3bCre/WT (WT) or Brn3bCre/Cre (KO) RGCs infected with isolated overexpression of several molecules identified in our expression vectors for four genes (Datasets S1 and S2). Box and whiskers plots screen may induce morphological changes consistent with neu- conventions are the same as in Fig. 7D. Examples for the Brn3bCre/Cre (KO) RGC rite formation in cultured epithelial cells. infections are shown in Fig. S6.(Scalebars:A–D,100μm; Insets,35μm.)

E3980 | www.pnas.org/cgi/doi/10.1073/pnas.1618551114 Sajgo et al. Downloaded by guest on September 30, 2021 protein with potential synaptic functions (65, 66) was distributed contain many well-characterized RGC markers and Brn3b target PNAS PLUS largely in cell bodies (including the nucleus) and dendritic arbors genes. The relative lack of overlap could be explained by (punctate pattern) and to a lesser extent, axons (Fig. 8C and Fig. (i) dilution of the Brn3b-dependent, RGC-specific genes by the S6C). Igsf6, an Ig gene of undefined function, was detectable in whole retinal tissue or (ii) secondary effects of RGC loss on dendrite and cell body; however, its axonal distribution seemed to other retinal precursors. vary from cell to cell (Fig. 8D and Fig. S6D). These distinct sub- In our hands, only a small fraction of RGC-enriched transcripts cellular distributions were confirmed for S100a10, Cpne4, and is regulated by Brn3b or Brn3a in RGCs; hence, other TFs may Igsf6 by staining WT uninfected retinal sections with antibodies participate in achieving the full RGC phenotype. Of 226 retinal against the endogenous proteins (Fig. S6 E–G). Thus, tissue cul- genes affected by Atoh7 loss (76), only 52 where enriched in E15 ture overexpression, subcellular distribution, molecular structure, Brn3bAP/WT RGCs (Dataset S9 and Fig. S7E). However, from + and previously described functions in other systems collectively 165 genes enriched in Atoh7 cells (77), 120 were also enriched in point to a role for these molecules in the establishment of neu- Brn3bAP/WT RGCs (Dataset S9 and Fig. S7F). The RGC program ronal arbor morphology. Because S100a, Cpne, Rtn4r, and Igsf may not, therefore, be completely defined by either Atoh7 are large molecular families, it is possible that combinatorial ex- (Math5) or Pou4f2 (Brn3b), and Atoh7 may be required but may pression of one or a few members of these families contributes to not be sufficient for Brn3b–RGC specification in cell-autonomous + the diversification of neuronal arbors. Transcriptional regulation and -nonautonomous fashions. Several TFs enriched in Atoh7 − by Brn3 TFs could, therefore, contribute to RGC type specifica- Brn3b precursors (Dlx1, Dlx2, Onecut2, and Onecut3) (Dataset tion by driving the selective expression of these downstream tar- S9) are negatively regulated by Brn3b at E15 (Fig. 5H) together gets and hence, inducing distinct features in RGCs. with Atoh7 itself. Loss of function phenotypes of these genes include RGC defects linked to defects in Horizontal cells (Onecut), Discussion and Amacrines (Dlx) (27, 30, 78, 79). Among the TFs enriched in + + We use RNASeq of purified RGC populations and retina con- Atoh7 Brn3b early RGCs and regulated by Atoh7 and Brn3b, trols to identify combinatorial expression codes for TFs and some may control specific RGC subtypes (Eomes), whereas others CSMs proposed to be important for the development and have broader roles in the retina (Onecut1 and Isl1) combined with specification of RGC types. functions in RGCs (22, 30, 31). Because negative feedback loops Previously, RGCs have been purified and profiled using seem to exist between Brn3b, and Atoh7, Dlx, and Onecut (Fig.

immunopanning onto substrates coated with anti–Thy-1 mAb, 5H), it is possible that Brn3b cooperates with more broadly BIOLOGY

laser capture microdissection of the GCL, or FACS sorting of expressed TFs to define RGCs, while suppressing others (22, 28, DEVELOPMENTAL genetically labeled GFP-positive neurons (67–71). Although 79). Our previous work suggests that RGC types might be de- immunopanning yields a nearly pure population, it relies on termined by a combination of TF profiles, each encoding distinct maintaining RGCs in culture for prolonged periods of time, features of the RGC type definition. Similar combinatorial codes potentially altering expression profile. GCL microdissection (or have been described in the spinal cord of vertebrates, the visual laser capture) yields fresh cells; however, the sample will contain and olfactory systems of flies, and the nervous system of Caeno- many amacrine neurons. None of the GFP-labeled lines are rhabditis elegans (1–3). Our approach enables us to look at the cell- limited to RGCs at this point, and relatively small amounts of autonomous effects of Brn3a and Brn3b loss from RGCs. At least RGCs are typically collected. Compared with these methodolo- 43 TF genes were differentially expressed in Brn3 KO RGCs gies, our immunomagnetic purification approach is relatively fast compared with the WT controls, a majority in Brn3b. Because (about 90 min from mouse to freezer or lysis buffer) and has a these transcripts could indirectly depend on Brn3, say by the relatively high recovery rate (about 25% of potential target modulation of other TFs, ChipSeq analysis would help establish + cells), but it results in less pure samples. RNA profiling involving the direct transcriptional relationship. By contrasting Brn3a , + + + microarray expression analysis only covers a subset of genes and Brn3b ,andBrn3a Brn3b populations, one can gain insights into + − provides relative gene expression level based on probe hybridization. more narrow RGC subgroups. As an example, Brn3b Brn3a + RNASeq represents the entire collection of transcripts obtained RGCs make up Opn4 intrinsically photosensitive RGCs through reverse transcription in a relatively unbiased manner, and (ipRGCs)-positive cells (6, 13, 80). Here, we show that this sub- each transcript is typically covered by multiple hits. For both ap- set of RGCs expresses Eomes, Tbox20, and Irx1, consistent with a proaches, a limiting amount of starting RNA could be an issue, possible role of these TFs in ipRGC specification (29, 32). Fur- because no matter the depth of sequencing, saturation in number of thermore, the few genes regulated by Brn3a may be directly rel- recovered genes is achieved as recently shown by single-cell RNASeq evant to cell type specification of the few cell types missing from from retinal populations (72, 73). In addition, amplification biases Brn3aKO RGCs (5, 13, 31). The degree of similarity between RGC could occur early in the amplification process, skewing transcript transcriptional profiles and retina depends on both the de- representation. We, therefore, advocate for validation of RNASeq velopmental age and Brn3 expression status. In several of our experiments with ISH or protein detection techniques. clustergrams, Brn3bAP/KO RGCs are segregating together with the Our dataset confirms between 25 and 50% of the RGC- retina samples and separated from other RGC populations. specific genes previously identified by FACS sorting, immuno- We report a large set of CSMs overexpressed in RGCs com- + panning, or GCL laser capture (67, 68, 71), with the intersection pared with the retina. CSMs affected by Brn3 ablation in Brn3 consisting mostly of well-established RGC markers (Dataset S9 RGCs are limited in numbers, especially for Brn3aAP RGCs. and Fig. S7). We, however, expand the potential number of Thus, very specific candidate genes were identified for the subtle targets by an order of magnitude (Dataset S9 and Fig. S7). This dendritic arbor distinctions in Brn3aAP/KO RGCs. In contrast to increase could be caused by the particular comparisons per- the TFs clusters, CSMs seem to more clearly distinguish RGC formed in the different screens (e.g., developmental time points samples from retina controls (Figs. S4 and S5). About 30% of the of immunopanned RGCs), the purity of the samples (e.g., RGCs identified adhesion molecules have already been implicated in vs. Amacrines in the GCL laser capture experiment), the depth various aspects of RGC development, but the separation by of profiling in microarrays vs. RNASeq, or other experimental Brn3 assignment and the differential expression in Brn3b WT vs. differences. Expression profiling experiments using microarrays KO RGCs will no doubt help to better target the assignment of were previously performed on Brn3bKO/KO retinas (74, 75). Of individual CSMs to RGC subpopulations. Of particular interest the combined 234 identified target genes, only 49 are also pre- is the negative regulation by Brn3b of a subset of adhesion sent in our combined Brn3b regulated dataset comprising molecules known to mediate axon guidance decisions (Efna4, 1,008 genes (Dataset S9 and Fig. S7D). These 49 common genes Efna5, Efnb2, Epha6, Cntn2, Nfasc, Tenm3, Plxna1, Plxna3,

Sajgo et al. PNAS | Published online May 2, 2017 | E3981 Downloaded by guest on September 30, 2021 Sema3a, and Slit1) at E15 but not P3. Intriguingly, Brn3b seems could be influenced by RNA processing and trafficking mecha- to also suppress Zic2 expression in E15 RGCs, a TF required for nisms, tagging, and/or overexpression and should be confirmed correct ipsilateral vs. contralateral segregation at the optic chi- using antibodies against the endogenous protein. asm (Fig. 5H) (81). Many of the disregulated guidance cues The follow-up and individual investigation of the identified function as repellants, potentially explaining the abnormal target genes are beyond the scope of this study, but our dataset branching into the inner plexyform layer or nonvisual nuclei of could benefit a large group of researchers interested in cell type AP/KO the thalamus observed in Brn3b RGC axons. specification, cell surface molecular codes, and pathogenetic Whole-tissue profiling of retinorecipient nuclei yielded many mechanisms of glaucoma. The strategy and reagents outlined interesting markers, including some that appeared to be lami- here could be used to isolate and profile many subpopulations of nated in the SC. Many of the identified TFs confirm previous neurons in the DRG, vestibular and auditory ganglia, and tri- work on brain patterning; however, the relatively reduced num- geminal ganglia and some sympathetic and parasympathetic af- ber of differentially expressed adhesion and guidance molecules ferent neurons, all expressing Brn3s. is somewhat surprising. It could be that RGC axons are triaged during the axon guidance process in the various intermediate Materials and Methods points (optic nerve, chiasm, tract, and brachium) and thus, need Mouse lines and genetic recombination strategy were previously described (13). fewer cues after they have reached the target areas. Alterna- Retina dissociation was previously described (83), and immunoaffinity purifica- tively, RGCs may make connections in several nuclei and hence, tion was developed by us using magnetic beads coupled to anti-AP antibodies. use similar guidance/recognition codes. Undoubtedly, cell puri- RGC recovery was tested by AP staining of bead-coupled RGCs and supernatants fication, sorting, and profiling approaches using specific markers and estimated to about 25% of all RGCs in a retina, whereas fold enrichment of will distinguish more defined molecular cues. For now, our RGC samples over retinal supernatants was about 65× (Fig. 1F and Dataset S1). data provide a useful entry point into the molecular diversity of Retinorecipient nuclei from WT P3 mice were visualized by anterograde tracing, retinorecipient neurons. dissected under a fluorescent microscope, and processed for RNA isolation. The Cell-autonomous mechanisms could determine the shape and entire remaining brain tissue was homogenized and used as control (Fig. 1E). size of neuronal arbors. Dissociated cultures of cortical and hip- RNA extraction was done using the RNEasy Kit (Qiagen), and RNASeq library pocampal neurons, starburst amacrine cells, and cerebellar Pur- preparation sequencing on Illumina platforms and Read mapping were pre- kinje neurons can adopt morphologies similar to those found in viously described (44, 84, 85). Alignments were visualized with Igviewer (86). vivo (82). Our overexpression screen identifies several molecules DESeq, hierarchical clustering, and principal components analysis (PCA) were performed using dedicated components of the Matlab Bioinformatics toolbox capable of inducing neurite-like extensions in HEK293 cells. The (42, 87, 88). Gene Ontology analysis was done using the Panther suite. ISH val- molecular nature of these genes includes cytoskeleton interactors, idation was performed using the protocol described by Schaeren-Wiemers and transmembrane proteins with roles in intracellular membrane Gerfin-Moser (89) and queries in the Allen Brain Institute brain atlas. traffic, and membrane-associated molecules with somewhat HEK293 overexpression studies, transfections, viral expression vectors, and Brn3- complex roles in membrane remodeling. Most of the studied genes Cre mice AAV infections were previously described. Extensive description of are members of large molecular families. It is possible that neu- materials and methods is provided in SI Materials and Methods. Mouse handling ronal arbor morphologies could initially be shaped by the com- procedures were approved by the National Eye Institute Animal User Committee binatorial expression of these genes and then, adjusted and under protocol NEI640. Next generation sequencing data reported here are sculpted by negative or positive cues or activity. In some instances, available under Gene Expression Omnibus accession number GSE87647. RGCs expressing particularly large amounts of Rtn4rl1, Cpne4, and S100a10 did exhibit changes of the dendritic arbors; however, ACKNOWLEDGMENTS. We thank Norimoto Gotoh for retina dissociation these defects were rare and inconsistent enough not to be included protocols; Harsha K. Rajasimha for assistance with Bowtie analysis; Peter Colosi, Scott Sternson, and Rachel O. Wong for AAV constructs; Beverly Wu in this work. However, our expression system allowed us to study for help with gene expression cloning; and Nadia Parmhans for genotyping. the subcellular localization of our candidates in RGCs. It should Funding was provided by the Intramural Research Program of the National be noted that subcellular localization of the various candidates Eye Institute (Z.W. and T.C.B.).

1. Hobert O (2011) Regulation of terminal differentiation programs in the nervous 15. Edwards MA, Caviness VS, Jr, Schneider GE (1986) Development of cell and fiber system. Annu Rev Cell Dev Biol 27:681–696. lamination in the mouse superior colliculus. J Comp Neurol 248:395–409. 2. Jessell TM (2000) Neuronal specification in the spinal cord: Inductive signals and 16. Chen C, Regehr WG (2000) Developmental remodeling of the retinogeniculate syn- transcriptional codes. Nat Rev Genet 1:20–29. apse. Neuron 28:955–966. 3. Komiyama T, Luo L (2006) Development of wiring specificity in the olfactory system. 17. Brown NL, Patel S, Brzezinski J, Glaser T (2001) Math5 is required for retinal ganglion Curr Opin Neurobiol 16:67–73. cell and optic nerve formation. Development 128:2497–2508. 4. Bassett EA, Wallace VA (2012) Cell fate determination in the vertebrate retina. Trends 18. Gan L, et al. (1996) POU domain factor Brn-3b is required for the development of a Neurosci 35:565–573. large set of retinal ganglion cells. Proc Natl Acad Sci USA 93:3920–3925. 5. Badea TC, Nathans J (2004) Quantitative analysis of neuronal morphologies in the mouse 19. Wang SW, et al. (2001) Requirement for math5 in the development of retinal gan- retina visualized by using a genetically directed reporter. JCompNeurol480:331–351. glion cells. Genes Dev 15:24–29. 6. Badea TC, Nathans J (2011) Morphologies of mouse retinal ganglion cells expressing 20. Xiang M, et al. (1997) Essential role of POU-domain factor Brn-3c in auditory and transcription factors Brn3a, Brn3b, and Brn3c: Analysis of wild type and mutant cells vestibular hair cell development. Proc Natl Acad Sci USA 94:9445–9450. using genetically-directed sparse labeling. Vision Res 51:269–279. 21. Xiang M, Gan L, Zhou L, Klein WH, Nathans J (1996) Targeted deletion of the mouse 7. Baden T, et al. (2016) The functional diversity of retinal ganglion cells in the mouse. POU domain gene Brn-3a causes selective loss of neurons in the brainstem and tri- Nature 529:345–350. geminal ganglion, uncoordinated limb movement, and impaired suckling. Proc Natl 8. Coombs J, van der List D, Wang GY, Chalupa LM (2006) Morphological properties of Acad Sci USA 93:11950–11955. mouse retinal ganglion cells. Neuroscience 140:123–136. 22. Mu X, Fu X, Beremand PD, Thomas TL, Klein WH (2008) Gene regulation logic in 9. Helmstaedter M, et al. (2013) Connectomic reconstruction of the inner plexiform layer retinal ganglion cell development: Isl1 defines a critical branch distinct from but in the mouse retina. Nature 500:168–174. overlapping with Pou4f2. Proc Natl Acad Sci USA 105:6942–6947. 10. Sümbül U, et al. (2014) A genetic and computational approach to structurally classify 23. Sun Y, et al. (2008) A central role for Islet1 in sensory neuron development linking neuronal types. Nat Commun 5:3512. sensory and spinal gene regulatory programs. Nat Neurosci 11:1283–1293. 11. Dräger UC (1985) Birth dates of retinal ganglion cells giving rise to the crossed and 24. Wu F, et al. (2015) Two transcription factors, Pou4f2 and Isl1, are sufficient to specify uncrossed optic projections in the mouse. Proc R Soc Lond B Biol Sci 224:57–77. the retinal ganglion cell fate. Proc Natl Acad Sci USA 112:E1559–E1568. 12. Godement P, Salaün J, Imbert M (1984) Prenatal and postnatal development of reti- 25. Erkman L, et al. (1996) Role of transcription factors Brn-3.1 and Brn-3.2 in auditory nogeniculate and retinocollicular projections in the mouse. JCompNeurol230:552–575. and visual system development. Nature 381:603–606. 13. Badea TC, Cahill H, Ecker J, Hattar S, Nathans J (2009) Distinct roles of transcription 26. Jin K, Jiang H, Mo Z, Xiang M (2010) Early B-cell factors are required for specifying multiple factors brn3a and brn3b in controlling the development, morphology, and function retinal cell types and subtypes from postmitotic precursors. J Neurosci 30:11902–11916. of retinal ganglion cells. Neuron 61:852–864. 27. Klimova L, Antosova B, Kuzelova A, Strnad H, Kozmik Z (2015) Onecut1 and 14. Coombs JL, Van Der List D, Chalupa LM (2007) Morphological properties of mouse retinal Onecut2 transcription factors operate downstream of Pax6 to regulate horizontal cell ganglion cells during postnatal development. J Comp Neurol 503:803–814. development. Dev Biol 402:48–60.

E3982 | www.pnas.org/cgi/doi/10.1073/pnas.1618551114 Sajgo et al. Downloaded by guest on September 30, 2021 28. Mao C-A, et al. (2008) Eomesodermin, a target gene of Pou4f2, is required for retinal 61. Rescher U, Gerke V (2008) S100A10/p11: Family, friends and functions. Pflugers Arch PNAS PLUS ganglion cell and optic nerve development in the mouse. Development 135:271–280. 455:575–582. 29. Mao CA, et al. (2014) T-box transcription regulator Tbr2 is essential for the formation 62. Dempsey BR, et al. (2012) Structure of an asymmetric ternary protein complex pro- and maintenance of Opn4/melanopsin-expressing intrinsically photosensitive retinal vides insight for membrane interaction. Structure 20:1737–1745. ganglion cells. J Neurosci 34:13083–13095. 63. Wang X, et al. (2015) Intravitreal delivery of human NgR-Fc decoy protein regenerates 30. Sapkota D, et al. (2014) Onecut1 and Onecut2 redundantly regulate early retinal cell axons after optic nerve crush and protects ganglion cells in glaucoma models. Invest fates during development. Proc Natl Acad Sci USA 111:E4086–E4095. Ophthalmol Vis Sci 56:1357–1366. 31. Shi M, et al. (2013) Genetic interactions between Brn3 transcription factors in retinal 64. Dickendesher TL, et al. (2012) NgR1 and NgR3 are receptors for chondroitin sulfate ganglion cell type specification. PLoS One 8:e76347. proteoglycans. Nat Neurosci 15:703–712. 32. Sweeney NT, Tierney H, Feldheim DA (2014) Tbr2 is required to generate a neural 65. Creutz CE, et al. (1998) The copines, a novel class of C2 domain-containing, calcium- – circuit mediating the pupillary light reflex. J Neurosci 34:5447 5453. dependent, phospholipid-binding proteins conserved from Paramecium to humans. 33. Wang SW, et al. (2002) Brn3b/Brn3c double knockout mice reveal an unsuspected role J Biol Chem 273:1393–1402. – for Brn3c in retinal ganglion cell axon outgrowth. Development 129:467 477. 66. Reinhard JR, et al. (2016) The calcium sensor Copine-6 regulates spine structural 34. Li X, et al. (2016) Loss of AP-2delta reduces retinal ganglion cell numbers and axonal plasticity and learning and memory. Nat Commun 7:11613. projections to the superior colliculus. Mol Brain 9:62. 67. Kim CY, Kuehn MH, Clark AF, Kwon YH (2006) Gene expression profile of the adult 35. Rousso DL, et al. (2016) Two pairs of ON and OFF retinal ganglion cells are defined by human retinal ganglion cell layer. Mol Vis 12:1640–1648. intersectional patterns of transcription factor expression. Cell Reports 15:1930–1944. 68. Wang JT, et al. (2007) Disease gene candidates revealed by expression profiling of 36. Triplett JW, et al. (2014) Dendritic and axonal targeting patterns of a genetically- retinal ganglion cell development. J Neurosci 27:8593–8603. specified class of retinal ganglion cells that participate in image-forming circuits. 69. Winzeler A, Wang JT (2013) Purification and culture of retinal ganglion cells from Neural Dev 9:2. rodents. Cold Spring Harb Protoc 2013:643–652. 37. Jiang Y, et al. (2013) Transcription factors SOX4 and SOX11 function redundantly to 70. Gao F, et al. (2016) Comparative analysis of three purification protocols for retinal regulate the development of mouse retinal ganglion cells. JBiolChem288:18429–18438. ganglion cells from rat. Mol Vis 22:387–400. 38. Moore DL, et al. (2009) KLF family members regulate intrinsic axon regeneration 71. Siegert S, et al. (2012) Transcriptional code and disease map for adult retinal cell ability. Science 326:298–301. – 39. Sajgo S, Ali S, Popescu O, Badea TC (2016) Dynamic expression of transcription factor types. Nat Neurosci 15:487 495. Brn3b during mouse cranial nerve development. J Comp Neurol 524:1033–1061. 72. Shekhar K, et al. (2016) Comprehensive classification of retinal bipolar neurons by – 40. Badea TC, et al. (2012) Combinatorial expression of Brn3 transcription factors in so- single-cell transcriptomics. Cell 166:1308 1323.e30. matosensory neurons: Genetic and morphologic analysis. J Neurosci 32:995–1007. 73. Macosko EZ, et al. (2015) Highly parallel genome-wide expression profiling of indi- – 41. Xiang M, et al. (1993) Brn-3b: A POU domain gene expressed in a subset of retinal vidual cells using nanoliter droplets. Cell 161:1202 1214. ganglion cells. Neuron 11:689–701. 74. Mu X, et al. (2004) Discrete gene sets depend on POU domain transcription factor 42. Anders S, Huber W (2010) Differential expression analysis for sequence count data. Brn3b/Brn-3.2/POU4f2 for their expression in the mouse embryonic retina. Development Genome Biol 11:R106. 131:1197–1210. 43. Jeon CJ, Strettoi E, Masland RH (1998) The major cell populations of the mouse retina. 75. Qiu F, Jiang H, Xiang M (2008) A comprehensive negative regulatory program con-

J Neurosci 18:8936–8946. trolled by Brn3b to ensure ganglion cell specification from multipotential retinal BIOLOGY 44. Langmead B, Salzberg SL (2012) Fast gapped-read alignment with Bowtie 2. Nat precursors. J Neurosci 28:3392–3403. DEVELOPMENTAL Methods 9:357–359. 76. Mu X, et al. (2005) A gene network downstream of transcription factor 45. Ravasi T, et al. (2010) An atlas of combinatorial transcriptional regulation in mouse Math5 regulates retinal progenitor cell competence and ganglion cell fate. Dev Biol and man. Cell 140:744–752. 280:467–481. 46. Zhang HM, et al. (2015) AnimalTFDB 2.0: A resource for expression, prediction and 77. Gao Z, Mao CA, Pan P, Mu X, Klein WH (2014) Transcriptome of Atoh7 retinal pro- functional study of animal transcription factors. Nucleic Acids Res 43:D76–D81. genitor cells identifies new Atoh7-dependent regulatory genes for retinal ganglion 47. Kim I-J, Zhang Y, Yamagata M, Meister M, Sanes JR (2008) Molecular identification of cell formation. Dev Neurobiol 74:1123–1140. a retinal cell type that responds to upward motion. Nature 452:478–482. 78. de Melo J, et al. (2005) Dlx1 and Dlx2 function is necessary for terminal differentiation 48. Matsuoka RL, et al. (2011) Class 5 transmembrane semaphorins control selective and survival of late-born retinal ganglion cells in the developing mouse retina. – Mammalian retinal lamination and function. Neuron 71:460 473. Development 132:311–322. 49. Matsuoka RL, et al. (2011) Transmembrane semaphorin signalling controls laminar 79. Feng L, et al. (2011) Brn-3b inhibits generation of amacrine cells by binding to and – stratification in the mammalian retina. Nature 470:259 263. negatively regulating DLX1/2 in developing retina. Neuroscience 195:9–20. 50. Wong GK, Baudet ML, Norden C, Leung L, Harris WA (2012) Slit1b-Robo3 signaling 80. Chen SK, Badea TC, Hattar S (2011) Photoentrainment and pupillary light reflex are and N-cadherin regulate apical process retraction in developing retinal ganglion cells. mediated by distinct populations of ipRGCs. Nature 476:92–95. J Neurosci 32:223–228. 81. Herrera E, et al. (2003) Zic2 patterns binocular vision by specifying the uncrossed 51. Feldheim DA, O’Leary DD (2010) Visual map development: Bidirectional signaling, retinal projection. Cell 114:545–557. bifunctional guidance molecules, and competition. Cold Spring Harb Perspect Biol 2: 82. Lefebvre JL, Sanes JR, Kay JN (2015) Development of dendritic form and function. a001768. Annu Rev Cell Dev Biol 31:741–777. 52. Mosca TJ (2015) On the Teneurin track: A new synaptic organization molecule 83. Tabata T, et al. (2000) A reliable method for culture of dissociated mouse cerebellar emerges. Front Cell Neurosci 9:204. cells enriched for Purkinje neurons. J Neurosci Methods 104:45–53. 53. Tessier-Lavigne M, Goodman CS (1996) The molecular biology of axon guidance. 84. Brooks MJ, Rajasimha HK, Swaroop A (2012) Retinal transcriptome profiling by di- Science 274:1123–1133. rectional next-generation sequencing using 100 ng of total RNA. Methods Mol Biol 54. Li L, et al. (2011) The functional organization of cutaneous low-threshold mechano- – sensory neurons. Cell 147:1615–1627. 884:319 334. 55. Wu H, Williams J, Nathans J (2012) Morphologic diversity of cutaneous sensory af- 85. Kaewkhaw R, et al. (2015) Transcriptome dynamics of developing photoreceptors in ferents revealed by genetically directed sparse labeling. eLife 1:e00181. three-dimensional retina cultures recapitulates temporal sequence of human cone 56. Jan YN, Jan LY (2010) Branching out: Mechanisms of dendritic arborization. Nat Rev and rod differentiation revealing cell surface markers and gene networks. Stem Cells Neurosci 11:316–328. 33:3504–3518. 57. Zipursky SL, Grueber WB (2013) The molecular basis of self-avoidance. Annu Rev 86. Robinson JT, et al. (2011) Integrative genomics viewer. Nat Biotechnol 29:24–26. Neurosci 36:547–568. 87. Bar-Joseph Z, Gifford DK, Jaakkola TS (2001) Fast optimal leaf ordering for hierar- 58. Corty MM, Matthews BJ, Grueber WB (2009) Molecules and mechanisms of dendrite chical clustering. Bioinformatics 17:S22–S29. development in Drosophila. Development 136:1049–1061. 88. Eisen MB, Spellman PT, Brown PO, Botstein D (1998) Cluster analysis and display of 59. Chuang K, Nguyen E, Sergeev Y, Badea TC (2015) Novel heterotypic Rox sites for genome-wide expression patterns. Proc Natl Acad Sci USA 95:14863–14868. combinatorial Dre recombination strategies. G3 (Bethesda) 6:559–571. 89. Schaeren-Wiemers N, Gerfin-Moser A (1993) A single protocol to detect transcripts of 60. Sajgo S, et al. (2014) Dre - Cre sequential recombination provides new tools for retinal various types and expression levels in neural tissue and cultured cells: In situ hybrid- ganglion cell labeling and manipulation in mice. PLoS One 9:e91435. ization using digoxigenin-labelled cRNA probes. Histochemistry 100:431–440.

Sajgo et al. PNAS | Published online May 2, 2017 | E3983 Downloaded by guest on September 30, 2021