bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

NMDAR-mediated transcriptional control of expression during the development of medial ganglionic eminence-derived interneurons

Vivek Mahadevana, Apratim Mitrab, Yajun Zhangc, Xiaoqing Yuana, Areg Peltekiana, Ramesh Chittajallua, Caroline Esnaultb, Dragan Maricd, Christopher Rhodesc, Kenneth A. Pelkeya, Ryan Daleb, Timothy J. Petrosc, Chris J. McBaina,∗ aSection on Cellular and Synaptic Physiology, Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), Bethesda, 20892, MD, USA bBioinformatics and Scientific Programming Core, NICHD, Bethesda, 20892, MD, USA cUnit on Cellular and Molecular Neurodevelopment, NICHD, Bethesda, 20892, MD, USA dFlow and Imaging Cytometry Core Facility, National Institute of Neurological Disorders and Stroke (NINDS), Bethesda, 20852, MD, USA

ABSTRACT Medial ganglionic eminence (MGE)-derived parvalbumin (PV)+, somatostatin (SST)+ and Neurogli- aform (NGFC)-type cortical and hippocampal interneurons, have distinct molecular, anatomical and physiological properties. However, the molecular mechanisms regulating their diversity remain poorly understood. Here, via single-cell transcriptomics, we show that the obligate NMDA-type glutamate (NMDAR) subunit gene Grin1 mediates subtype-specific transcriptional regulation of gene ex- pression in MGE-derived interneurons, leading to altered subtype abundances. Notably, MGE-specific conditional Grin1 loss results in a systemic downregulation of diverse transcriptional, synaptogenic and membrane excitability regulatory programs. These widespread abnormalities mirror aberrations that are typically associated with neurodevelopmental disorders, particularly schizophre- nia. Our study hence provides a road map for the systematic examination of NMDAR signaling in interneuron subtypes, revealing potential MGE-specific genetic targets that could instruct future therapies of psychiatric disorders. Keywords: Interneurons, Medial ganglionic eminence, PV, SST, Neurogliaform, NMDA receptor, Transcriptional regulation, Neurodevelopmental disorders, Schizophrenia, NMDA-hypofunction, , Frontal Cortex, Mouse model, scRNAseq

1 1. INTRODUCTION entire forebrain accounting for approximately 6 60% of all cortical interneurons [1, 2]. In ad- 7 2 Medial ganglionic eminence (MGE)-derived dition, approximately half of all hippocampal 8 3 forebrain GABAergic interneurons comprise the neurogliaform-type cells (NGFCs), the so called 9 4 parvalbumin-containing (PV) and somatostatin- Ivy cells, originate from the MGE [3, 4]. In- 10 5 containing (SST) subpopulations throughout the terestingly, though only rarely found in rodent 11

neocortex such MGE-derived NGFCs are sig- 12 ∗Corresponding author. nificantly more populous in primate neocortex, 13 Email addresses: [email protected] (Vivek Mahadevan), [email protected] (Ryan Dale), including humans [5]. While PV neurons exert ro- 14 [email protected] (Timothy J. Petros), bust somatic, and proximal dendritic inhibition, 15 [email protected] (Chris J. McBain)

Mahadevan et al., 2021 1/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

16 the SST and NGFCs mediate domain-specific resembles global Grin1 -mutants [32] in their con- 63

17 dendritic inhibition on their downstream pyrami- stellation of schizophrenia-like behavioral aber- 64

18 dal neuron targets [6]. Collectively these classes rations. This indicates that Grin1 dysfunction 65

19 of interneurons shape diverse aspects of corti- across multiple interneuron-subtypes precipitates 66

20 cal and hippocampal circuit maturation during schizophrenia-like abnormalities [33]. In addition, 67

21 development, and critically regulate information this interneuron-specific NMDAR-hypofunction 68

22 processing in mature circuits by maintaining ap- model is sensitive to developmental age, since 69

23 propriate excitation-inhibition (E-I) balance [7]. adult-onset Grin1 loss does not result in the same 70

24 Recent evidence indicates a critical role for activ- phenotypes [29]. Despite the importance of devel- 71

25 ity, particularly through ionotropic glutamate opmental NMDAR function in interneurons, and 72

26 receptors (iGluRs), in driving the morpho- its relevance to human neurodevelopmental disor- 73

27 physiological maturation of MGE-derived in- ders, a comprehensive interrogation of the impact 74

28 terneurons [8–12]. Unlike mature interneurons of developmental NMDAR ablation in interneu- 75

29 where iGluRs differentially contribute towards rons, particularly across MGE-derived interneu- 76

30 synaptic transmission, immature and migrating rons, is lacking. 77

31 interneurons express different glutamate receptor It is clear that during the developmental win- 78

32 subunits including the NMDA-type iGluR (NM- dow between embryonic day (ED) 13.5 and post- 79

33 DAR) and AMPA/Kainate-type iGluR (AM- natal day (PD) ~10 [34], a combination of innate 80

34 PAR/KAR) [13–15] prior to the expression of genetic programs, external environment, and neu- 81

35 any functional synapses. This becomes particu- ronal activity shapes interneuron subtype spec- 82

36 larly important as the developing brain contains ification leading to remarkable diversity [2, 21, 83

37 higher ambient glutamate levels than the adult 35, 36] The NMDAR signaling complex comprises 84

38 brain [16]. Collectively, higher ambient gluta- an essential node for regulating gene expression 85

39 mate, developmental expression of iGluRs and via excitation-transcription (E-T) coupling in ma- 86

40 recruitment of glutamatergic signaling is con- ture circuits [37–39]. Moreover, different NMDAR 87

41 sidered to be trophic [8, 17, 18] and thought to subunits are widely expressed in the developing 88

42 engage mechanisms to regulate various aspects of brain [40] where they provide a critical source of 89 2+ 43 interneuron development including morphological Ca -entry via trophic glutamate signaling prior 90

44 and electrical maturation to promote appropriate to synaptogenesis [15, 19, 41]. However, it is not 91 2+ 45 circuit integration [9, 11, 14, 16, 19–22]. clear whether the NMDAR-mediated Ca cas- 92

46 Interneuron-specific impairments are increas- cades in nascent and developing MGE-derived in- 93

47 ingly considered central to the etiology of multi- terneurons engage transcriptional programs nec- 94

48 ple neurodevelopmental and circuit disorders [23]. essary for MGE-derived interneuron diversity. To 95

49 The importance of interneuron-expressed iGluRs investigate this, we conditionally deleted Grin1 96

50 is most notable in psychiatric disorders exhibiting in MGE progenitors that give rise to cortical and 97

51 impaired NMDAR-associated systems [24, 25]. In hippocampal PV, SST, and NGFC subsets, us- 98

52 the adult brain, acute pharmacological NMDAR ing the Nkx2-1-Cre mouse line [3, 4, 42]. In 99

53 blockade results in circuit disinhibition and psy- this model, Nkx2-1 -driven Cre expression is re- 100

54 chotic symptoms [26], mediated in-part, by the ported in cycling/proliferating MGE cells, well 101

55 enhanced sensitivity of interneuronal NMDARs to before the cells become postmitotic, allowing for 102

56 their antagonists [27]. Indeed, direct blockade of assessment of the developmental impact of em- 103

57 interneuron activity also precipitates distinct be- bryonic loss of Grin1 activity across all subsets of 104

58 havioral deficits relevant to schizophrenia [28]. In MGE-derived interneurons. Applying a combina- 105

59 particular, ablation of the obligate NMDAR sub- tion of high-throughput single-cell RNA sequenc- 106

60 unit gene Grin1 in interneuron-specific early post- ing (scRNAseq), MGE-interneuron-specific ribo- 107

61 natal mouse [29], but not PV-specific [30], or glu- somal tagging, and quantitative immunostaining 108

62 tamatergic neuron-specific Grin1 ablation [31], and in situ RNAscope analyses we establish that 109

Mahadevan et al., 2021 2/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

110 NMDAR-mediated transcriptional cascades reg- markers Pvalb, Sst, and Lamp5, marking PV and 155 111 ulate MGE subtype abundances, by regulating SST, NGFC subsets respectively (Figure1B, 156 112 the expression of diverse transcriptional, synap- Figure1-Supplement3A). While we did not 157 113 togenic and membrane excitability genetic pro- recover cells expressing the CGE-markers Prox1, 158 114 grams. Notably, we identify numerous disease- Htr3a or Vip, we recovered a minor fraction of 159 115 relevant that are misexpressed in MGE- cells corresponding to glutamatergic neurons, 160 116 derived interneurons upon Grin1 -ablation, pro- and microglia. In addition, ~25% 161 + 117 viding a broad road map for examination of MGE- and 71% TdT MGE-sorts from cortex and 162 118 subtype specific regulation via NMDAR signaling. hippocampus respectively were enriched in oligo- 163 119 dendrocytes marked by Olig1 expression across 164

all replicates (Figure1-Supplement 2B,2C). 165

However, we focused our subsequent analyses on 166 120 2. RESULTS the 5656 and 3002 Gad1 / Gad2 positive cortical 167

and hippocampal MGE-derived interneurons. 168 121 scRNAseq recapitulates cardinal MGE 122 subtypes and a continuum of molecular Unbiased cell clustering by Seurat v3 identi- 169

123 profiles fied six subtypes of SST, two subtypes of PV, 170

124 To examine the molecular heterogeneity of two subtypes of NGFCs, and one subtype of Ty- 171

125 MGE-derived GABAergic interneurons by scR- rosine hydroxylse (TH) expressing interneurons, 172

126 NAseq, we microdissected frontal cortex (CX) expressing the markers Sst, Pvalb, Lamp5 and 173

127 and hippocampus (HPC) from fresh brain slices Th respectively, across the two brain regions ex- 174

128 obtained from PD18-20 Nkx2.1 -Cre:Ai14 mouse amined (Figure1C). Notably, all but two sub- 175

129 (Figure1A, Figure1-Supplement1A). Ai14- types (SST.5 and SST.6) expressed high levels 176 + 130 TdTomato (TdT ) single-cell suspensions were of Lhx6, and 2 clusters corresponding to PV.2 177

131 harvested by fluorescence-activated cell sort- and NGFC.1 expressed Nkx2.1 at this develop- 178

132 ing (FACS) using stringent gating constraints mental time. While the PV- SST- and NGFC- 179

133 including viability and doublet discrimination clusters clearly exhibited robust gene expression 180

134 (Figure1-Supplement1B) as previously de- differences among each other (Figure1D), the 181

135 scribed [43–45], and subsequently processed TH cluster appeared to express genes that cor- 182

136 through the 10X Genomics Chromium controller. respond to both PV: SST clusters, including Sst 183 + 137 9064 and 9964 TdT cells were recovered from and Pvalb expressions (Figure1C, Figure1- 184

138 cortex and hippocampus respectively across 3 Supplement3B). Particularly, at this develop- 185

139 biological replicates. To minimize the effect of mental window we could not observe robustly dif- 186

140 excitotoxicity and stress-related transcriptional ferent gene expression variances between the cor- 187

141 noise, the tissue processing, FACS, and sam- tical and hippocampal counterparts, barring a few 188

142 ple collection steps were performed in buffers marginal, but significant differences (Figure1- 189

143 supplemented with Tetrodotoxin (TTX), DL Supplement4B) . This gave us sufficient ratio- 190

144 -2-Amino-5-phosphonopentanoic acid (APV) nale to perform subsequent analyses using the 191

145 and Actinomycin-D (Act-D) [46]. Because we MGE-derived interneurons pooled from cortex 192

146 observed concordant cell clustering across the and hippocampus. 193

147 replicates during preliminary analysis by Seu- Among the SST sub clusters, SST.1-5 194

148 rat v3 [47, 48] (Figure1-Supplement2A), the uniquely expresses Chodl, Igf2bp3, Cdh7, Pld5 195

149 replicates were pooled for in-depth analysis. and Nfix respectively, while SST.6 expresses 196

150 Subsequent clustering and marker gene analy- only markers that are common with other 197 + 151 ses revealed that ~ 62% and 33% of the TdT SST clusters (Figure1-Supplement3B). With 198

152 MGE-sorts from cortex and hippocampus respec- the exception of SST.6 the remaining SST- 199

153 tively, express classical GABA markers including expressing subclusters are described in previous 200

154 Gad1 / Gad2, Lhx6 ; and the MGE-subclass scRNAseq assays (Figure1-Supplement5A). 201

Mahadevan et al., 2021 3/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure1 | Identification of MGE-derived interneuron subtypes in the cortex and hippocampus. A, Overview of the experimental workflow. Bi, Uniform Manifold Approximation and Projection (UMAP) dimensional reduction of 19,028 single-cell transcriptomes (9,064 from frontal cortex and 9,964 from hippocampus of 6 mouse brains, 3 biological replicates), showing the cardinal MGE populations. Cell clusters were color coded and annotated post hoc based on their transcriptional profile identities (Cell type abbreviations: PVALB, Parvalbumin; NGFC, Neurogliaform; TH, Tyrosine Hydroxylase; SST, Somatostatin; GLUT, Glutamatergic; CP, Choroid Plexus; MG, Microglia; ASTR, ; MUR, Mural; OLIGO, Oligodendrocyte). Bii, UMAP visualization of 11 MGE-derived interneuron subtypes from cortex (MGE.CX) and hippocampus (MGE.HPC), and the recovery of cell numbers from the subtypes. Biii, Table indicating the number of Gad1/Gad2+cells recovered in each MGE subtype from the cortex and hippocampus, and the defining genes enriched in each subtype. C, Violin plot showing the distribution of expression levels of well-known representative cell- type-enriched marker genes across the11 MGE subtypes. D, −log10 False Discovery Rate (FDR) versus log2 fold change (FC) between each of the MGE cardinal class, representing the top enriched markers at a fold change ≥0.5 and FDR <10e-25. E, UMAP representation of PVALB clusters highlighting the cortex-specific enrichment of Pthlh-expressing PVALB.2 subtype that is not observed in the hippocampus.

Mahadevan et al., 2021 4/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

202 For example, the Chodl-expressing SST.1 clus- comparison to their cortical counterparts, consis- 249

203 ter co-expresses high Nos1, Tacr1, Penk, and tent with preferential localization of MGE-derived 250

204 Npy, and it has been previously described as NGFCs to HPC over CX in rodents [1, 3, 4]. (ii) 251

205 putative GABAergic long-range projection neu- Next, the Pthlh-expressing PVALB.2 subclus- 252

206 rons [49, 50]. Clusters SST.2/3/4 express Elfn1, ter splits into two islands, only in the cortex 253

207 Reln and Grm1 characteristic of putative cor- and distinctly lacking from the hippocampus. 254

208 tical martinotti and their hippocampal coun- Only one of the PVALB.2 islands expresses 255

209 terpart, oriens-lacunosum/moleculare (O-LM) C1ql1, while the other cortex-enriched island 256

210 cells [43, 51, 52](Figure1-Supplement5B). expresses unique markers Etv1, Cnr1, Pcp4, 257

211 Lastly, Zbtb20 -expressing SST.5 is predicted to Crabp1, Necab2, Epha4 and Hapln1 (Fig- 258

212 be septal-projecting interneurons [43]. Among ure1E, Figure1-Supplement5D) . Whether 259

213 the PV sub clusters, while both PVALB.1&2 this represents a novel subclass of chandelier cells 260

214 coexpresses several common markers includ- remains to be determined. (iii) Lasty, we also 261

215 ing Pvalb, Kcnip2, Tcap and Kcnc1 there are observed a distinction in the hippocampal SST.3 262

216 several notable differences between the two clus- corresponding to a subset of O-LM interneurons 263 217 ters. PVALB.1 appears to contain continuous, (Figure1-Supplement3Aii). The overall MGE 264 218 but non-overlapping populations expressing Syt2 cell numbers indicate that the SST cells account 265

219 representing putative fast-spiking basket cells for the majority of MGE cell population recov- 266

220 or Rbp4/Sst containing putative bistratified ered in the scRNAseq assay from both brain 267 221 cells [1, 43, 53](Figure1-Supplement5D) . regions (Figure1-Supplement3Aii, Aiii). The 268 222 PVALB.2 contains cells that uniquely expresses PV and TH clusters accounted for a greater share 269

223 Pthlh, C1ql1, Fgf13 and Unc5b representing of MGE-derived interneurons in the CX than 270

224 putative axo-axonic chandelier cells [43, 49, 54]. in the HPC. While it is plausible these relative 271

225 We also observed a TH cluster, which, in addi- cell proportions may be skewed by differential 272

226 tion to expressing several genes common to the survivability of these subtypes during tissue dis- 273

227 SST: PV clusters, expresses several unique genes sociation, sorting and single-cell barcoding, these 274

228 including Rasgrp1, Bcl6, Myo1b that segregated relative percentages were similar across biological 275

229 into mutually exclusive cluster space express- replicates. 276

230 ing Crh or Nr4a2 (Figure1-Supplement3B, NMDAR signaling maintains MGE- 277

231 Figure1-Supplement5B) . This cluster is derived interneuron subtype abundance 278

232 also described previously as putative bistratified- Because neuronal activity and glutamatergic 279

233 like cells [43, 53] . Among the NGFC sub signaling are known to regulate multiple facets 280

234 clusters, while both NGFC.1&2 coexpress sev- of interneuronal development [2, 21] [11, 36, 55] 281

235 eral common markers including Lamp5, Hapln1, , we hypothesized that the key obligate sub- 282

236 Cacna2d1, Sema3c and Id2, the NGFC.1 cluster unit Grin1 and the NMDAR signaling complex 283

237 uniquely expresses several genes like Reln, Ngf, may play an instructive role in determining MGE 284

238 Egfr, Gabra5 that are not expressed by NGFC.2. subtype identities. To test whether NMDAR 285

239 (Figure1-Supplement3B). While the Reln+ signaling impacts the development and func- 286

240 represents MGE-derived neurogliaforms, the tion of MGE-derived interneurons, we ablated 287

241 Reln-population may represent putative ivy them in MGE progenitors by crossing floxed- 288 Cre 242 cells [43] (Figure1-Supplement5C). Grin1 mice with the Nkx2.1 mouse line [42] 289

243 While the majority of the UMAP space (Figure1-Supplement1A). The earliest expres- 290

244 aligns well between the cortical and hippocam- sions of Nkx2.1 and Grin1 in the developing ro- 291

245 pal MGE-derived interneurons, we observed dent brains is reported around ~embryonic day 292

246 some regional differences as well. (Figure1- (ED) 10.5 and ~ED14 respectively [56–58]. More- 293 2+ 247 Supplement3AiAiii). (i) First, we observed over, NMDAR-mediated Ca signaling in mi- 294 248 an increase in the HPC-expressed NGFC.1&2 in grating interneurons is reported by ~ED16 [15]. 295

Mahadevan et al., 2021 5/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure2 | Altered interneuron subtype proportions upon Grin1 -ablation A, Integrated UMAP visualization of 12 subtypes of MGE-derived interneurons obtained from cortex (CX) and hippocampus (HPC) of Grin1 wt/wt and Grin1 fl/flmice. UMAP represents the following numbers of MGE-derived interneurons from Grin1 wt/wt and Grin1 fl/flrespectively: 5624 (CX.WT) and 1387 (CX.NULL); 2998 (HPC.WT) and 2309 (HPC.NULL). 3 and 2 independent biological replicates of the scRNAseq assay from Grin1 wt/wt and Grin1 fl/ flrespectively. B, Violin plot showing the distribution of expression levels of well-known representative cell-type-enriched marker genes across the 12 interneuron subtypes. C, Violin plot from both genotypes indicating the expression of Grin1 in the cardinal of MGE-derived interneuron subtypes. D, UMAP representation colored by brain-region, highlighting the differential enrichments of cells (brown arrows) within interneuron subsets in Grin1 -WT and Grin1 -null from CX and HPC. E, Stacked-barplots representing the proportions of recovered cell numbers within Ei, pooled cardinal MGE subtypes, Eii, SST subtypes; Eiii, PVALB subtypes and Eiv, NGFC subtypes in Grin1 -WT and Grin1 -null from cortex or hippocampus. χ2, Chi-square test of proportions; ns, not significant.

Mahadevan et al., 2021 6/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

296 Because the expression and activity of Nkx2.1 tal PV/SST cell type counts at PD30 (Figure2- 343 297 precedes Grin1 expression, we rationalized that Supplement1Bii), we observed a modest reduc- 344 Cre 298 utilizing Nkx2.1 mouse will ablate Grin1 and tion in cortical PV cell type counts along with an 345

299 NMDAR signaling in MGE progenitors from increase in cortical SST cell type counts at the 346 300 the earliest developmental point. We sorted same age (Figure2-Supplement1Aii). This in- 347 + 301 TdT cells from the cortex and hippocampus of dicated differential impact of Grin1 -ablation on 348 Cre fl/fl 302 Nkx2.1 :Grin1 :Ai14 mice and performed cortical and hippocampal interneurons. 349

303 scRNAseq using the 10X platform. The scR- Despite observing no major changes in the 350

304 NAseq experiments were performed using juve- recoveries of cardinal MGE-interneuron subtypes 351

305 nile mice (PD18-20) of both sexes and from the by scRNAseq, we observed marked changes in the 352

306 same litters as the wildtypes (WT) to enable sub- recovery percentages of the subsets of SST, PV 353

307 sequent direct comparison. Similar to the WT- and NGFCs from both cortex and hippocampus 354 fl/fl 308 datasets, the MGE-Grin1 mutants also re- (Figure2Eii−iv, Supplement2Aii,C). Particu- 355 + 309 vealed an enrichment of TdT oligodendrocytes larly, we observed a robust increase in cortical 356

310 (Figure2-Supplement4B), however, we again Chodl-expressing cortical SST.1 population, hip- 357

311 focused our attention on the Gad1/2 positive in- pocampal Reln-expressing SST2-4 populations, 358

312 terneurons. and a decrease in hippocampal SST.6 population 359 fl/fl 2 313 We next performed integrated analyses of in MGE-Grin1 (CX, HPC: χ = 286, 209; 360 wt fl/fl 314 the MGE-Grin1 and MGE-Grin1 cortical p-value = 2.2e-16 for both regions). In addition, 361

315 and hippocampal scRNAseq datasets. Apply- we found a reduction in the cortical PVALB.1 362

316 ing similar unbiased clustering parameters used population, and a compensatory increase in 363 wt fl/fl 317 for the MGE-Grin1 analyses, we observed a PVALB.2/3 populations in MGE-Grin1 (CX, 364 2 318 total of twelve Gad1/2 positive clusters in the HPC: χ = 236, 8.4; p-value = 2.2e-16, 0.14). 365

319 integrated dataset (Figure2A, B). As a robust Finally, we observed an increase in the NGFC.1 366

320 control, Grin1 appeared to be absent or vastly along with a compensatory decrease in NGFC.2 367 2 321 reduced in all MGE subsets in both brain re- in both cortex and hippocampus (CX, HPC: χ = 368 fl/fl 322 gions from MGE-Grin1 (Figure2C), but 13, 232; p-value = 0.0003, 0.14). Among the dif- 369

323 not in the Slc17a7 expressing glutamatergic ferentially enriched subclusters, Pthlh-expressing 370

324 neurons (Figure2-Supplement4A). Overlay- PVALB.3 is quite notable (Figure2B, Figure2- 371

325 ing the WT and NULL datasets from the brain Supplement3A,B). This cortex-enriched clus- 372

326 regions revealed differential enrichments among ter lacking in the hippocampus was identified 373

327 the recovered cells between the genotypes (Fig- within the PVALB.2 putative-chandelier cells 374 wt/wt 328 ure2D). Intriguingly, Grin1 -ablation did not in the MGE-Grin1 (Figure1E, Figure1- 375

329 seem to alter the SST or PV recovery percent- Supplement5D), However, subsequent to inte- 376 fl/fl 330 ages, with the exception of a modest increase in gration of the MGE-Grin1 scRNAseq dataset, 377 2 331 the cortical NGFCs (χ = 11.6, p = 0.003), but it segregated as a unique cluster, far from other 378 2 332 not hippocampal NGFCs (χ = 4.07, p = 0.13) PVALB clusters in the UMAP space. We ob- 379 333 (Figure2Ei, Figure2-Supplement2Ai,B). served Prox1 expression in PVALB.3, which is 380

334 To independently examine whether Grin1 ab- uncharacteristic of MGE-derived interneurons, 381

335 lation impacts interneuron abundances, we con- additional to robust expressions of genes asso- 382

336 ducted immunostaining experiments to probe ciated with NGFCs such as Hapln1 and Reln 383

337 the PV and SST subtypes from postnatal days (Figure2-Supplement3A,B). Moreover, we 384

338 (PD)30 brains from both genontypes. First we observed an increase in recovery of the cortical 385 + 339 observed no change in the total TdT cell counts PVALB.3 cell numbers, including the emergence 386

340 from both cortex and hippocampus (Figure2- of these cells in the hippocampus subsequent to 387 341 Supplement1 Ai,Bi). Next, while we ob- Grin1-ablation (Figure2-Supplement2A,B). 388 342 served no change in hippocampal expressed to- To independently examine whether the pre- 389

Mahadevan et al., 2021 7/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure3 | Cell-autonomous transcriptional changes subsequent to MGE-specific developmental Grin1- ablation A, Combined heatmap representing the 802 differentially expressed (DEGs) in the cortical and hippocampal MGE-derived interneurons upon Grin1 -ablation, at a FDR<0.01 and FC>10%, as determined by MAST analysis (see details in Methods). Bi, iRegulon in silico analysis identifying high-confidence master upstream transcriptional reg- ulators(indicated in red) of the DEGs (indicated in lavender). Representative DNA-binding motifs are indicated next to the transcriptional regulators. Bii, Top five transcriptional regulators predicted by iRegulon, associated normalized enrichment score (NES) and number of predicted targets and motifs associated with each cluster, indicated for the three interneuron subtypes (CX and HPC pooled). C, Ingenuity Pathway Analysis of significantly overrepresented molecular pathways in each MGE-subtype. FDR determined by Fisher’s Exact Test.

Mahadevan et al., 2021 8/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

390 dicted differences in cell recovery percentages Among all DEGs only ~10% and 1% are upreg- 437

391 among the cardinal MGE subtypes are true, ulated in the cortex and hippocampus respec- 438

392 we conducted RNAscope in situ hybridization tively, while the remaining genes were all down- 439 393 assays from PD20-25 cortex and hippocampus regulated (Figure3-Supplement1Aii,B,C). 440 394 from both genotypes. We particularly focused (ii) While Grin1 ablation resulted in several 441

395 on the subtypes of SST interneurons, namely the unique DEGs between the MGE classes, ~10 442

396 high-Nos1 -expressing Chodl-SST.1 and the Reln- and 27% of the DEGs are common within cor- 443

397 expressing SST.2-4 subtypes. First, similar to our tex and hippocampus respectively (Figure3A, 444

398 scRNAseq prediction, we observed a significant Figure3-Supplement2). For instance, while 445 + 399 increase in cortical Sst cells that co-express Nos1 S100a10, Hapln1, Hcrt2 are uniquely upregu- 446 + 400 and a concomitant reduction in the Sst cells that lated in cortical SST, PV and NGFC respectively 447

401 lack Nos1 after Grin1 -ablation in MGE-derived (Figure3A), Apoe, Kcns3, Wnt5a were uniquely 448 402 interneurons (Figure2-Supplement1Ci, Cii). altered in hippocampal SST, PV and NGFC 449 403 Next, similar to our scRNAseq prediction, we ob- respectively. In contrast, Grin1 ablation induced 450 + 404 served an increase in hippocampal Sst cells that common changes in Penk1 and Erbb4 expres- 451 405 co-express Reln (Figure2-Supplement1Di, sion patterns across all MGE-derived interneuron 452 406 Dii), without changes in total Sst+ cell numbers classes in the cortex and hippocampus respec- 453 407 in both brain regions. It is unclear whether the tively. (iii) ~27-43% of all DEGs were shared 454

408 changes in subtype numbers reflect a change in by MGE classes across brain regions (Figure3- 455 409 cell identity or whether this is alterered subtype Supplement2Aii ). For example, Npas3, Cdh9, 456 410 proportions due to differential survival. Never- Grm1 are commonly downregulated in all SST 457

411 theless, these data demonstrate clear changes in subclasses; Bcl6, Epha7, Gabra4 common to PV 458

412 MGE-derived interneuron subtype abundances class; and Rgs12, Gabrad, Sema5a common to 459

413 following early embryonic loss of Grin1 function. NGFCs from both brain regions. (iv) Lastly, 460

414 NMDAR signaling shapes the transcrip- 28 genes are commonly differentially expressed 461

415 tional landscape in MGE-derived interneu- across both brain regions, across all MGE sub- 462

416 rons types. For example, Grin1, Neto1, Cdh2, Scn2b 463

417 It is now well-established that transcrip- are commonly downregulated across the board, 464

418 tional signatures defines the subtype identities of while Epha5, Olfm1 are commonly downregulated 465

419 GABAergic interneurons [59]. To examine the full across all, but cortical PV cells (Figure3A). 466

420 range of transcriptional impairments triggered by Gene expression co-regulation is intrinsic to cel- 467

421 Grin1 ablation in MGE-derived interneurons, lular diversity [60, 61]. Since the majority of 468

422 we next performed differential gene expression DEGs are downregulated across the MGE sub- 469

423 testing by pooling the SST / PVALB / NGFC types, we examined whether they correspond to 470

424 subtypes into their cardinal MGE classes to iden- clusters of coordinated co-regulation. We applied 471

425 tify the genes that are differentially expressed the iRegulon in silico framework [62], which iden- 472

426 between the genotypes. For instance SST1-6 and tifies transcription factor binding motifs that are 473

427 TH.1 are pooled together as SST; PVALB1-3 are enriched in genomic regions of the DEGs upon 474

428 pooled together as PVALB, and NGFC1-2 are Grin1 -ablation, and predicts the transcription 475

429 pooled together as NGFC cardinal classes for this factors that bind to the motifs. This in silico 476

430 assay . At a stringent false-discovery rate (FDR) analysis predicted 51 significantly enriched motifs 477

431 <0.01, 802 genes passed the 10%-foldchange (normalized enrichment score > 3) that clustered 478

432 (FC) threshold across the MGE subtypes from into 10 groups by similarity, 33 of which were 479

433 both brain regions (Figure3A, Supplemental associated with transcription factors (Figure3B, 480

434 Table1). Several interesting features were ob- Supplemental Table2). Put together, 10 tran- 481

435 served in the differentially expressed gene (DEG) scription factors were predicted to bind with the 482

436 pattern upon MGE-specific Grin1 -ablation. (i) motifs with high confidence, strongly support- 483

Mahadevan et al., 2021 9/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

484 ing targeted co-regulation of 617 among the 802 mediators of NMDAR signaling to the nucleus 531

485 DEG genes upon Grin1 -ablation. Notably, the for transcriptional regulation. Theoretically, an 532 2+ 486 RE1-silencing transcription factor (Rest) is a mas- early first wave impairment of Ca signaling in 533

487 ter transcriptional repressor that mediates the Grin1 -lacking MGE progenitors could result in 534 2+ 488 transcriptional accessibility for several synaptic transcriptional silencing of the mediators of Ca 535

489 genes [63], including NMDAR subunits them- signaling cascades and second messenger systems, 536

490 selves [64]. It is intriguing to observe that the which would sustain the transcriptional impair- 537

491 downregulation of the DEGs upon MGE-specific ments. Indeed, we also observed a downregulation 538 2+ 492 Grin1 -ablation are, in part, predicted to occur via of various Ca homeostasis-regulators, kinases 539

493 Rest-mediated transcriptional repression. / phosphatases and second messengers that 540

494 To examine the broad biological impact of the are activated downstream of Grin1 (Figure4- 541

495 DEGs, we performed (GO) anal- Supplement1A,B,C). Furthermore, we noted 542

496 yses. Broad GO analyses on all DEGs indicates that hippocampal MGE neurons had a greater 543

497 that these genes serve to regulate multiple molec- proportion of DE-TFs and kinase signaling cas- 544

498 ular functions in interneurons, including regula- cade effectors that were downregulated across all 545

499 tion of GABAergic and glutamatergic synapses, 3 subtypes compared to their cortical counter- 546

500 additional to biological pathways related to ad- parts. Together, this suggests that hippocampal 547

501 diction and circadian entrainment (Figure3- MGE-derived interneurons may be more vul- 548

502 Supplement2B). Further classification of DEGs nerable than cortical MGE-derived interneurons 549 2+ 503 based on their cellular functions within the MGE towards Grin1 -mediated Ca transcriptional 550

504 subtypes revealed genes critical for regulation of silencing at this age. 551

505 membrane excitability, gene expression, synaptic Interestingly, among the early TF cascades in 552

506 partnering and assembly, as well as major intra- the progenitors that sequentially determine and 553 2+ 507 cellular Ca signaling cascades and second mes- maintain MGE fate, several members appear to 554

508 sengers (Figure3C, Figure3-Supplement2C, be expressed at ~P20, and starkly downregulated 555

509 Supplemental Table3) . upon Grin1 -ablation. For instance, Lhx6, Maf, 556

510 Transcription factor expression is a key Arx, Myt1l, Dlx1 are among the genes broadly 557

511 component of NMDAR-mediated regula- downregulated across all hippocampal MGE sub- 558

512 tion of MGE-derived interneurons types and within specific class(es) in their cor- 559

513 Because transcriptional regulation underlies tical parallels (Figure4A ). Other MGE fate- 560

514 numerous fundamental processes including the determining TFs, Nkx2-1, Mafb, Satb1, Nr2f1 561

515 expression of other classes of genes, we next (CoupTf1), Sp9, also appear to be downregu- 562

516 examined the DE-transcriptional regulators in lated in discrete populations. This also includes 563

517 detail. We first examined the 67 genes that are a downregulation of Bcl11b (Ctip2) in both hip- 564

518 differentially expressed upon Grin1 -ablation and pocampal and cortical NGFCs, a gene recently 565

519 are known to mediate transcriptional regulation linked to regulation of NGFC morphology and 566

520 of gene expression. Of these, 35 genes are pre- function [65]. Among the few transcriptional reg- 567

521 viously established to be expressed in different ulators upregulated are Sirt2, Elmo1, Zcchc12, 568

522 GABAergic interneuron classes including some none of which have been characterized in the 569

523 notable MGE-expressed transcription factors context of MGE function (Figure4B ). Sirt2 is 570 524 (Figure4Ai). The remaining 32 are broadly ex- an established transcriptional repressor [66, 67] 571 525 pressed TFs (Figure4Aii ), that include a small that may regulate the repression of several target 572 526 subset of 15 genes that are regulated by neu- genes in an MGE-specific manner, and Elmo1 has 573

527 ronal activity. Barring a few genes, we observed been previously characterized during the activity- 574

528 the majority of TFs to be down regulated in dependent migration of CGE subtypes [21]. Fi- 575 2+ 529 both brain regions. Intracellular Ca signaling nally, a recent study has predicted that the ex- 576

530 cascades and second messenger systems are key pression of Zcchc12 correlates with slower in- 577

Mahadevan et al., 2021 10/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure 4 | Grin1-signaling in MGE-derived interneurons are highly dedicated to the transcriptional control of interneuron identity A,Heatmap of log2 FC of significant DEGs in cortical and hippocampal MGE cardinal subtypes, showing a subset of (Ai) Transcription factors (TFs) that are previously established to regulate MGE subtype identity and function; (Aii) broadly expressed TFs and transcriptional regulators (TRs) that are not currently known to regulate MGE function. Clusters of commonly differentially expressed genes in cortex and hippocampus are indicated in yellow or green boxes. * indicates a subset of neuronal activity-regulated TFs. (Bi) Overview of the experimental workflow for RNAseq validation, by ribosomal tagging / Ribotag strategy specific to MGE-derived interneurons. n = 4 biological replicates from each genotype for MGE-Ribotag-seq; Error bars reflect s.e.m; DeSeq2 for statistical analysis. (Bii) MGE-Ribotag-seq normalized reads for Grin1 expression in cortex and hippocampus. MGE-Ribotag-seq normalized reads cross-validating the differential expressions of (C) MGE-enriched TFs, and (D) neuronal activity-regulated TFs. (Ei) Representative immunostaining of P25-30 somatosensory cortex using anti-Satb1 (cyan), and endogenous Ai14 reporter expression. Ai14:Satb1 double- positive cells are represented as yellow surfaces using Imaris9.6. (Eii) Boxplots indicate the normalized density of cortical Ai14(+) and Ai14(+) :Satb1(+) double-positive cell counts. n = 3 brains from each genotype for immunostaining; Error bars reflect s.e.m from individual sections; two-tailed unpaired t-test, for statistical analysis.

Mahadevan et al., 2021 11/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

578 trinsic firing among hippocampal CA1 interneu- sion of Npas3 in both brain regions, after Grin1 625

579 rons [43]. This suggests that increased Zcchc12 ablation (Figure4C, top). 626

580 expression might regulate the expression of synap- Key differences exist between the two strate- 627

581 tic genes enabling reduced intrinsic excitability in gies scRNAseq and MGE-Ribotag-seq. The 628

582 the MGE subsets. Related to such putative de- scRNAseq assays only the mRNAs present in 629

583 creased excitability in the MGE-derived interneu- the cytoplasm of the neurons while providing 630

584 rons, among the activity-regulated TFs, we ob- a single-cell resolution across the subtypes of 631

585 serve broad downregulation of Jun, Egr1, Fos, MGE-derived interneurons. The MGE-Ribotag- 632

586 Fosb, Arc, Satb1, Arnt2 across all classes of MGE seq however lacks the single-cell resolution across 633

587 in both brain regions (Figure4A,B, indicated MGE-interneuron subtypes but includes the mR- 634

588 by *). While most of these are well-established NAs that are bound to the translational ma- 635

589 activity-regulated TFs, Arnt2 has been recently chinery and much closer to the cellular pro- 636

590 described to partner with Npas4, downstream teome. Moreover, MGE-Ribotag-seq includes 637 2+ 591 of Ca signaling in response to neuronal activ- the translating-mRNAs present in the neuronal 638

592 ity [68]. Unsurprisingly, Npas4 is also signifi- processes. Despite these technical differences, 639

593 cantly downregulated in discrete MGE subtypes. the congruity in the DEGs after Grin1 -ablation 640

594 MGE-Ribotag-seq cross-validates im- strongly favors our claims related to NMDAR- 641

595 paired transcription factor expressions due signaling dependent regulation of key transcrip- 642

596 to MGE-Grin1 -ablation tion factors. Having a gene not validated by 643

597 To independently validate the changes in gene MGE-Ribotag-seq does not imply a false-positive 644

598 expressions observed by scRNAseq, we employed scRNAseq result, and the above-described tech- 645

599 a Ribotag-based strategy where we generated nical differences could contribute to it. Also, the 646

600 a triple-transgenic mice by breeding the Nkx2- genes that are commonly differentially expressed 647 cre fl/fl HA/HA− 601 1 :Grin1 with the Rpl22 containing across the PV-SST-NGFC subtypes by scRNAseq 648

602 Ribotag mice [69], hereafter called MGE-Ribotag will have a greater possibility of being identified 649 603 (Figure4Bi, Figure4-Supplement1D). We re- as a DEG via MGE-Ribotag-seq. Accordingly, 650 604 cently established that this is a robust tool to genes such as Dlx6, Sp9, Sirt2 and Zcchc12 that 651

605 examine the mRNAs associated with the trans- showed alternating differences among subtypes 652

606 lational machinery, in a manner exclusive to in the scRNAseq, did not reveal significant dif- 653

607 MGE-derived GABAergic interneurons [70]. mR- ferences via MGE-Ribotag-seq (Figure4C, bot- 654

608 NAs copurified with anti-HA immunoprecipita- tom). 655

609 tion were utilized to prepare cDNA libraries that We also observed a significant decrease in 656

610 were subsequently sequenced in order to inves- the expressions of several activity-regulated 657

611 tigate changes in gene expression changes re- TFs by MGE-Ribotag-seq in both brain re- 658

612 sulting from MGE-specific Grin1 -ablation. By gions (Figure4D). While Fos, Junb, Btg2, Nfib 659

613 utilizing cortical and hippocampal tissues from expressions were significantly downregulated in 660

614 age-matched MGE-Ribotag from both genotypes, either cortex or hippocampal MGE-derived in- 661

615 we first established a robust reduction in Grin1 terneurons, the expressions of Nr4a1 and Npas4 662 616 expression in both brain regions (Figure4Bii). appear to be commonly downregulated in both 663 617 Similar to the observed changes in the scRNAseq brain regions after Grin1 -ablation. Satb1 is 664

618 assay we observed robust changes in several key a key activity-regulated TF that is expressed 665

619 TFs by MGE-Ribotag-seq, particularly in genes in MGE-derived interneurons, critically reg- 666

620 that have established functions in maintaining ulating the identity and survival of different 667

621 MGE-identities. Notably, we observed a signif- subtypes [71, 72]. Independent immunostain- 668

622 icant downregulation of hippocampus-expressed ing experiments revealed that Grin1 -ablation 669

623 Lhx6 and Id2, an upregulation in cortical Mafb resulted in a robust reduction in MGE-derived 670

624 and hippocampal Npas1, and an increased expres- interneurons co-expressing Satb1 in PD30 so- 671

Mahadevan et al., 2021 12/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

672 matosensory cortex (Figure4Ei,Eii), further ders [23]. Thus, we questioned whether the Grin1 719 673 substantiating our findings from scRNAseq. ablation induced DEGs presently identified corre- 720

674 Taken together, this establishes a framework for late with disease etiology. Disease-ontology based 721

675 simultaneous regulation of neuronal activity and Fisher’s Exact Test conducted on the DEGs 722

676 expression of distinct sets of TFs (including the showed significant over-representation of genes 723

677 activity-regulated TFs) by NMDAR-signaling in implicated in ‘Schizophrenia’, ‘Psychiatric dis- 724

678 MGE-derived interneurons. orders’ and ‘Movement disorders’, among other 725

679 Impaired NMDAR signaling alters cellular impairments involving aberrant mor- 726

680 region-specific MGE subtype marker ex- phology of neurons (Figure6A, Supplemental 727

681 pression Table4). To independently examine the DEGs 728

682 Several GABAergic/MGE markers were mis- for potential enrichment for neurodevelopmen- 729

683 regulated upon Grin1 -ablation (Figure5A). For tal disorders, we obtained the risk genes for 730

684 example, genes S100a0, Pthlh, Hcrtr2 that are schizophrenia (Sz) and autism spectrum (As) 731

685 normally expressed in SST, PV and NGFCs re- from the SZDB [73] and SFARI [74] databases 732

686 spectively, are upregulated in the same clusters of respectively. These databases curate and rank 733 fl/fl 687 MGE-Grin1 (Figure5Bi), indicating a mis- disease-relevant gene sets, based on multiple evi- 734 688 expression in a subtype-specific manner. Next, dence sources including genome-wide association 735

689 while certain genes such as Reln, Tenm1 are studies, copy-number variations, known human 736

690 broadly downregulated across MGE classes, some mutations and other integrative analyses. In 737

691 genes like Thsd7a show an upregulation in cer- particular, we mapped the DEGs with the top- 738

692 tain classes but a downregulation in the other ranked genes from these disease datasets (see 739 693 classes (Figure5Bii). Interestingly, a few genes methods for details). While 592 DEGs could 740 694 that are normally abundant in one MGE class, not be mapped with either disease genes, 25 741

695 appear to be misexpressed in another MGE class genes mapped exclusively with the SFARI-AS 742

696 where they are not abundant. For instance, Tcap, gene list, 164 genes mapped exclusively with 743

697 that is normally expressed in PV cells, in addi- the SZDB-Sz gene list and 21 genes mapped 744

698 tion to being decreased in PV cells, is upregu- with both datasets (Figure6Bi, Supplemen- 745

699 lated in NGFCs in both cortex and hippocam- tal Table5 ). It is now well-established that 746

700 pus. Similarly, Hapln1 expression which is typi- several neurodevelopmental disorders exhibit a 747

701 cally limited to NGFCs, is upregulated in PV sub- high degree of converging molecular pathways 748 702 sets (Figure5Biii). Indeed, MGE-Ribotag-seq employing that exist in physical com- 749 703 cross-validates a robust downregulation in Tcap, plexes [75–78]. Therefore, we examined whether 750

704 Tenm1 and Hapln1, and a robust upregulation these disease-associated DEGs are known to form 751

705 in S100a0, Tacr1 and Nos1 (Figure5C ). Lastly, complexes between each other, by map- 752

706 we observed an upregulation in the Gad1 and ping curated protein-protein interaction (PPI) 753

707 Slc32a1 (vesicular GABA transporter, vGAT) datasets for all 802 DEG products. Indeed, we 754 + 708 and a downregulation in Gad2 and Slc6a1 (Na - observed that >95% of disease-annotated DEG 755 ¯ 709 Cl dependent GABA transporter, GAT1), corre- products are known to exist with PPIs, while 756

710 sponding with GABA synthesis and reuptake ma- only ~75% of DEG products not annotated with 757

711 chineries respectively (Figure5-Supplement5A Sz/As are known to exist with PPIs (Supple- 758

712 ). These findings closely match the trending dif- mental Table5C). Interestingly, despite not 759

713 ferences observed via MGE-Ribotag-seq. mapping directly with the high-ranked disease 760

714 NMDAR signaling regulates MGE gene sets, the remaining 592 genes are observed 761

715 subtype-specific expression of neurode- to exist in tightly-knit PPIs with the disease 762

716 velopmental disorder risk genes annotated genes. However, the PPIs mapped 763

717 Interneuron-centric disease etiology is an with SZDB form the most interconnected clus- 764

718 emerging centrality in multiple psychiatric disor- ters in comparison to the SFARI-mapped PPI 765

Mahadevan et al., 2021 13/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure5 | Differential expression of interneuron marker genes across subtypes upon Grin1-ablation A, Heatmap of log2 FC of significant DEGs in cortical and hippocampal MGE cardinal subtypes, showing a subset notable MGE marker genes. Representative box-violin plots of top differentially expressed genes from the above that represent Bi, fl/fl fl/fl a misexpression unique to a single MGE-Grin1 subtype; Bii, a misexpression across all MGE-Grin1 subtypes; fl/fl Biii, a misexpression between MGE-Grin1 subtypes. (C) MGE-Ribotag-seq normalized reads, cross-validating the misexpressed interneuron marker genes in cortex and hippocampus.

766 network (Figure6Bii), as indicated by relatively Ntng1, Sema3e, Slit2, Cx3cl1, and some of their 787 767 higher clustering coefficient. Put together, both receptors, Unc5c, Nrp1, Neto1/2, Robo2 that are 788

768 unbiased disease ontology prediction, targeted decreased in a MGE-class-specific manner. We 789

769 analysis of high-confidence neuropsychiatric observed Fgf13 that was recently demonstrated 790

770 disorder genes, indicate that members of the to mediate MGE-subtype specific synapse assem- 791

771 DEGs share physical, and functional pathways in bly [79], to be upregulated in cortical PV cells, but 792

772 MGE-derived interneurons, contributing towards downregulated in cortical SST, while Apoe to be 793

773 disease etiology. upregulated in hippocampal SST cells. In addi- 794

tion to synaptic assembly molecules, we observed 795 774 Among the 210 DEGs mapped with to Sz and DE in a variety of synaptic adhesion molecules, 796 775 As, 45 genes are established regulators of axon corresponding to protocadherin, cadherin, ephrin 797 776 path-finding, synapse formation and assembly, and contactin families (Figure6Cii). Notably, we 798 777 while 38 members are established regulators of also observed a downregulation of Erbb4 across 799 778 membrane excitability and neuronal firing. Be- all hippocampal MGE-subtypes. Lastly, we ob- 800 779 cause both of these gene classes are intimately served increased expression of extracellular ma- 801 780 associated with interneuron function, we exam- trix components Mgp, Ndst4, Hapln1, Adamts5, 802 781 ined these classes in detail. We observed multiple Mxra7, Thsd7a and the matrix modifying en- 803 782 classes of secreted ligands and cognate receptor zymes Hs3st2/4 in cortical SST/PV subtypes 804 783 families corresponding to semaphorin, netrin, slit, (Figure6Ciii). In parallel, MGE-Ribotag-seq 805 784 chemokine and growth factors, and their intracel- cross-validates a robust down regulation in hip- 806 785 lular effectors that are downregulated upon MGE- pocampal Erbb4, and a robust differential expres- 807 786 Grin1 -ablation (Figure6Ci,iv). These include

Mahadevan et al., 2021 14/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure6 | Aberrant Grin1 -signaling result in misexpression of high-risk Sz genes A, Ingenuity Pathway Analysis of significantly overrepresented disease pathways in each MGE-subtype. FDR determined by Fisher’s Exact Test. Bi, Global protein-protein interaction (PPI)map among all differentially expressed genes (DEGs). Red circles indicate the DEGs annotated to be top-ranked Sz-risk genes; Blue circles indicate the DEGs annotated to be top-ranked As-risk genes; Yellow circles indicate the DEGs annotated with both Sz and As-risk genes. Black circles in the periphery indicate the DEGs not annotated with high-risk Sz/As genes. The PPIs between DEGs indicated in grey lines. Bii, PPIs between Sz / As / dually enriched clusters, and other genes. The PPIs between disease-annotated DEGs and other disease-annotated DEGs or with other non-annotated DEGs are indicated in black lines. ThePPI between non- annotated DEGs indicated in grey lines. C, Heatmap of log2 FC of significant DEGs in cortical and hippocampal MGE cardinal subtypes, showing a subset of Ci, secreted trophic factors and secreted ligands and guidance cues. Cii, membrane- bound synaptogenic receptors and cell adhesion molecules (CAMs) Ciii, Extracellular Matrix (ECM) components and matrix modifying enzymes. Civ, Intracellular effectors of guidance and synaptogenic cues. D, MGE-Ribotag-seq normalized reads, cross-validating the differentially expressed synaptic partnering molecules.

Mahadevan et al., 2021 15/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

808 sion of several synaptic partnering and adhesion Our unbiased transcriptional profiling approach 853

809 molecules including Cdh9, Ntng1, Ephb6, Pcdh1, indicates that developmental NMDAR signaling 854

810 Cdh12, Sema5a, and key members of CAM mod- participates in MGE-derived interneuron identi- 855

811 ifiers St3gal1 and St8sia2 (Figure6D ). ties by regulating the expression of transcrip- 856

812 Among the regulators of neuronal excitabil- tion factors (67 genes), synaptogenic (53 genes) 857

813 ity, we observed a downregulation of multiple and connectivity factors/adhesion molecules (61 858 814 members of postsynaptic glutamate receptor sub- genes), and regulators of membrane excitability 859 815 units, GABA receptors and their associated part- (78 genes), among the 802 DEGs in interneurons 860 816 ners (Figure6-Supplement1Bii). Interestingly, (Figure7). We employed bioinformatic analyses 861 817 while we noted a broad downregulation of several to examine whether system-wide downregulation 862 818 members of potassium and sodium channel sub- of target genes can be attributed to transcription- 863 819 units, a few discrete members of the Kcn-families repression elements. Indeed, we identify a set 864 820 were upregulated in cortical PV and NGFC sub- of 10 transcriptional regulators that commonly 865 821 types. Finally, we also observed multiple mem- recognize the DNA-motifs present in the identi- 866 822 bers of presynaptic GABA synthesis, release and fied DEGs, including the master-repressor Rest. 867 823 uptake machineries including Gad1, Syt2/10, and Future studies are needed to examine the role 868 824 Slc6a1 differentially expressed in discrete MGE of these putative repressors and and associated 869 825 subtypes (Figure6, Supplement1Bi ). Collec- cis-regulatory elements that have not been pre- 870 826 tively, these findings highlight the centrality of viously associated with MGE-specific transcrip- 871 827 MGE-expressed Grin1 -signaling during synapse tion. However, based on broad transcriptional 872 828 formation and connectivity, which when aber- downregulation of target genes, we can make sev- 873 829 rantly expressed, can lead to neurodevelopmental eral predictions that should guide future investi- 874

830 disorders. gations. 875

Shaping interneuron identity and granu- 876

larity amongst subtypes 877 831 3. DISCUSSION Interneuron development from MGE is replete 878

832 Centrality of MGE-derived interneuron- with combinatorial expressions of numerous tran- 879 833 expressed NMDARs from juvenile brain scription factors, leading to diversity [82, 83]. 880 834 NMDARs serve as critical activity dependent Several transcription factors that are impacted 881

835 signaling hubs for myriad neuronal functions due by NMDAR-signaling are established regulators 882

836 to their innate ability to directly link network of MGE fate, subtype abundances and identities 883

837 dynamics to cellular calcium events, and asso- (34 genes) including Nkx2-1, Lhx6, Satb1, Dlx1, 884

838 ciated transcriptional coupling. Such NMDAR- Dlx6, Maf, Mafb, Mef2c, Etv1, Npas1, Npas3 and 885

839 dependent excitation-transcription coupling is Sp9 [34, 56, 84–87] [88]. While the scRNAseq 886

840 widely established in glutamatergic neurons [80], landscapes of interneurons predict several tran- 887

841 and in specific interneurons using candidate ap- scriptomic features that would classify them as 888

842 proaches [81] within mature circuits. However, distinct ‘cell-types’ or cell-states’ [49, 89, 90] [50] 889

843 the detailed unbiased evaluation of the transcrip- the precise mechanisms responsible for such gran- 890

844 tional landscape of NMDAR signaling within ularity is still emerging [2]. 891

845 interneurons in developing circuits undergoing re- Three key possibilities that enable NMDAR- 892

846 finement is lacking. Our study provides the first dependent interneuron development exists, and 893

847 systematic “fingerprinting” of the transcriptional our dataset provides potential examples for all 894

848 coupling associated with NMDAR signaling, ex- three scenarios. (i) First, it is possible that 895 2+ 849 clusive to MGE-derived interneurons, providing NMDAR-dependent Ca signaling in the devel- 896

850 a road map for examining NMDAR regulation of oping interneuron progenitors provides a combi- 897

851 MGE-derived interneurons in a subtype specific natorial cue that will couple with innate genetic 898

852 manner. programs to generate the diversity in interneu- 899

Mahadevan et al., 2021 16/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure7 | Transcriptional control of MGE development, synaptic partnering and excitablity are mediated by NMDAR signaling A, Nkx2.1 expression appears at ~ED10.5, driving MGE subtype fate in interneuronal progen-itors [42] [56]. Subsequently, their sequential developmental milestones towards circuit refinement appears to be under a combination of innate genetic mechanisms and neuronal activity. While the earliest Grin1 expression is reported at ~ED14 in developing brain [58] [57] , MGE-specific Grin1 -mediated Ca2+ is recorded at ~ED16 [15]. However, the broad role played by interneuron-expressed NMDAR signaling during interneuron development until now is not well delineated. B, By driving Grin1 -ablation using Nkx2.1 -driven Cre-recombinase, we report the earliest developmental loss of NMDAR signaling, across MGE-derived interneuron subtypes. In particular, by performing scRNAseq assay in MGE-derived interneurons from the cortex and hippocampus of the mouse brain, we report a broad transcriptional aber-ration subsequent to loss of NMDAR-signaling. Notably, this expression abnormality involves numerous transcriptional factors, synaptogenic and regulators of interneuron excitability, that collectively estabish MGE subtype identities. ED, Embryonic day; PD, Postnatal day; PCD, Programmed cell death; LiGC, Ligand-gated channel; VGCC, Voltage-gated calcium channel

900 ron subtypes. We observe key MGE-specification the identity and survival of different subtypes of 914

901 genes Lhx6 and Id2 downregulated and other SST-interneurons [71, 72]. Additionally, we ob- 915

902 genes such as Mafb, Npas1 and Npas3 upregu- serve an increase in Etv1 expression in cortical 916

903 lated in Grin1 -lacking interneurons in both scR- PV cells, and Etv1 was previously demonstrated 917

904 NAseq and MGE-Ribotag-seq assays. (ii) It is to be an activity-dependent TF that inversely cor- 918 2+ 905 also possible that NMDAR signaling critically relates Ca influx, regulating the identity of a 919

906 regulates neuronal activity that further shapes the subset of PV-interneurons [84]. (iii) Lastly, it 920

907 expressions of MGE-subtype determining and dis- is possible that altered neuronal activity due to 921

908 tinct activity-dependent TFs. In line with this aberrant NMDAR -signaling result in differential 922

909 possibility, we establish a robust downregulation survival and/or cell-death in developing interneu- 923

910 in the activity-regulated TF, Satb1, and a ro- rons [2, 11, 21, 35]. Indicating that the Grin1 - 924 + 911 bust reduction in Satb1 cortical interneurons lacking MGE-derived interneurons have altered 925

912 that lack Grin1. It is notable that Satb1 is a interneuron survival and/or cell-death, we observe 926

913 key activity-regulated TF that critically regulates differential recoveries of the subtypes within SST, 927

Mahadevan et al., 2021 17/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

928 PV and NGFC in the scRNAseq assay. Par- candidates for examining subtype specific synapse 975

929 ticularly, we observe an abundance of cortical assembly, which are centrally regulated by NM- 976

930 Chodl-SST.1 in the Grin1 -ablated MGE-derived DAR signaling. Of particular interest are the fam- 977

931 interneurons, and a concomitant decrease in cor- ily of protocadherins that are reported recently to 978

932 tical SST.2-4 subtypes. Moreover, we report an be commonly downregulated in cortical interneu- 979

933 increase in hippocampal Reln-SST2-4 subtypes rons generated from Sz patient-derived induced 980

934 upon Grin1 -ablation by RNA in situ hybridiza- pluripotent stem cells [100]. 981

935 tion. Additionally, by independent immunostain- Subsequent to synapse formation nascent con- 982

936 ing experiments, we observe a robust decrease nections remain susceptible to strength modifi- 983

937 in cortical MGE-derived interneurons that coex- cations according to neuronal activity. Again, 984

938 press the TF Satb1, which is established to pro- NMDAR-signaling in MGE-derived interneurons 985

939 mote interneuron survival. However, detailed fu- seems to regulate this process by the transcrip- 986

940 ture studies are necessary to uncover which of tional regulation of the expressions of both presy- 987

941 the above indicated NMDAR-dependant mech- naptic and postsynaptic members, including ex- 988

942 anisms regulate the abundances and diversity citatory and inhibitory synaptic molecules and 989

943 within PV/SST/NGFC subclasses. their auxiliary subunits, as well as presynap- 990

944 Shaping interneuron subtype-specific tic GABA release machinery molecules such as 991

945 synaptic assembly and connectivity Cplx1/2, Stx1b, Rab3c. However, most dramatic 992

946 What is the biological context of differential is the massive down regulation of several mem- 993

947 expression of the TFs, in the juvenile forebrain bers of the potassium channel subunits and their 994

948 when MGE-derived interneuron fate is assumed auxiliary subunits across MGE subtypes, with the 995

949 to be already sealed, and subtype identities estab- exception of an upregulation of a few Kcn-genes 996

950 lished? It is emerging that some of these TFs are in cortical PVs and NGFCs. While the precise 997

951 continually required for the maintenance of MGE impact of the diverse changes in these genes on 998

952 fate, post development [91]. One of the ways the MGE firing are currently unclear, the pattern of 999

953 TFs maintain MGE subtype fate into adulthood, expression of the activity-dependent transcription 1000

954 is by controlling the expression of genes that are factors provides us an indication. 1001

955 essential for ongoing interneuron function. Ac- Notable activity-dependent TFs such as Jun, 1002

956 cordingly, we predict that NMDAR-dependent ex- Egr1 are downregulated across all MGE subtypes, 1003

957 pression of synaptogenic and synaptic partnering while Fosb, Fos, Arx are down regulated across all 1004

958 molecules regulate the assembly of synapses with hippocampal MGEs, and Satb1, Arnt2 are down- 1005

959 appropriate targets. Secreted semaphorin, ephrin, regulated across all cortical MGEs. In addition, 1006

960 slit, netrin and neurotrophin-based signaling sys- Npas4, an established early-response TF [101– 1007 2+ 961 tems have been investigated in GABAergic neu- 103] activated upon neuronal activity and Ca 1008

962 rons, during axonal pathfinding, and cell migra- influx in MGE-derived interneurons [104], was 1009

963 tion [92–99]. However, only recently have inroads downregulated in cortical NGFCs upon Grin1 - 1010

964 been made into delineating their expression, and ablation. Lastly, Ostn was recently established 1011

965 interaction with appropriate receptor systems in as an activity-regulated secreted factor [105], and 1012

966 target synapses during accurate synaptogenesis. we observed Ostn to be downregulated specifically 1013

967 In addition, the NMDAR-dependent expression of in cortical PV subtypes. Together, these changes 1014

968 synaptic adhesion molecules will further promote are consistent with reduced neuronal activity in 1015

969 stability of newly formed synapses. Here, the mis- MGE subtypes upon Grin1 -ablation, recapitu- 1016

970 expression of diverse secreted cues, their receptors lating previous reports indicating that NMDAR- 1017

971 and adhesion molecules by MGE subtypes during antagonists can directly reduce the activity of 1018

972 Grin1 -ablation, provides unique insight into the GABAergic interneurons in adult mice [27]. In- 1019

973 molecular diversity employed during synapse es- terpreting the differential expressions of activity- 1020

974 tablishment. Our findings also reveal numerous dependent genes during scRNAseq has been chal- 1021

Mahadevan et al., 2021 18/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1022 lenging, particularly, when these genes could get NMDAR signaling in neurons is thought to under- 1069

1023 activated by the very process involved in cell dis- lie a wide range of these neurological disorders, an 1070

1024 sociation and sorting [106, 107]. However, our use interneuron-centric developmental NMDAR aber- 1071

1025 of activity-blockers and actinomycin-D through- ration is emerging central to schizophrenia-related 1072 wt fl/fl 1026 out our MGE-Grin1 and MGE-Grin1 scR- syndromes. Indeed, in the present study, dis- 1073

1027 NAseq pipelines [46] and independent validations ease mapping of the DEGs using high-ranked 1074

1028 via MGE-Ribotag-seq gives confidence that the SZDB-Sz and SFARI-As datasets indicate that 1075

1029 differential expressions of activity-dependent TFs many more DEGs map with the Sz than the As 1076

1030 reflect biological relevance. database. Moreover, these disease-relevent DEGs 1077

1031 NMDAR signaling in NGFCs exist in physical and functional complexes with 1078

1032 Among the MGE subtypes, the PV and SST other DEGs that are not directly mapped to the 1079

1033 interneurons are traditionally widely studied in Sz database. We used only stringent, high-ranked 1080

1034 comparison to the dendrite-targeting NGFC sub- disease genes from the database that pass sev- 1081

1035 types (that include the Ivy cells). In the present eral disease-relevant criteria. However, there are 1082

1036 study we provide the first detailed molecular in- other DEGs that still map to lower-ranked Sz and 1083

1037 sight into the cortical and hippocampal NGFCs, As datasets that are ‘non-annotated’ in present 1084

1038 subsequent to NMDAR ablation. We anticipated study. While our study can be argued as an ‘ex- 1085

1039 that these cell types could be particularly sus- treme’ case of NMDAR hypofunction in MGE- 1086

1040 ceptible to loss of NMDARs, since we previously derived interneurons, it provides a starting point 1087

1041 reported that NGFCs exhibit the most robust highlighting the centrality and broad range of 1088

1042 synaptic NMDAR conductances among the MGE interneuronal NMDAR-transcriptional pathways 1089

1043 subtypes [12]. Intriguingly, while the cortical during development. 1090

1044 NGFCs had comparable numbers of both total A multitude of studies implicate NMDAR- 1091

1045 and unique DEGs with respect to other cortical hypofunction specific to PV cell types as a cen- 1092

1046 MGE-derived interneurons (Figure3B), we ob- tral underlying feature of schizophrenia etiol- 1093

1047 served far fewer total and NGFC-specific DEGs in ogy [116, 117]. However, the measurable NM- 1094

1048 the hippocampus, compared to other hippocam- DAR conductances within PV interneurons are 1095

1049 pal MGEs. However, based on the scRNAseq relatively small in comparison to other MGE sub- 1096

1050 cell type recoveries, we predict an elaboration of types [12] . Additionally, NMDA signaling in non- 1097

1051 NGFC.1, and a reduction in the NGFC.2 sub- PV interneuron subtypes drives robust dendritic 1098

1052 type upon Grin1 -ablation. Finally, NGFCs ex- inhibition in pyramidal neurons [118, 119]. More- 1099

1053 hibited dendritic arborization impairments sub- over, while NMDAR-ablation in Pvalb-Cre lines 1100

1054 sequent to impaired NMDAR signaling [9, 11]. produces other behavioral deficits unrelated to 1101

1055 Indeed, we observe 49 genes among the DEGs the Sz-like phenotypes [30, 33] , a developmental, 1102

1056 (Supplemental Table1 ) that have established but not adult-onset Grin1 -ablation in Ppp1r2 - 1103

1057 roles in regulating neuronal cytoskeleton and as- Cre line [29] that targets a subset of PV interneu- 1104

1058 sociated signaling, likely mediating the observed rons among other subtypes [120], recapitulates 1105

1059 dendritic impairments in NGFCs. core Sz-like phenotypes. Lastly, studies that map 1106

1060 Developmental NMDAR ablation in in- interneuron subtypes to Sz-like phenotypes indeed 1107

1061 terneurons and s chizophrenia support the role of different interneuron classes 1108

1062 Impaired NMDAR function observed during beyond PV cells towards disease etiology [28, 33]. 1109

1063 human NMDAR gene mutations [108], and anti- Integrating these ideas and based on findings 1110

1064 NMDAR-encephalitis [109] results in a wide range from the present study, we propose the follow- 1111

1065 of neuropsychiatric disorders including autism ing: (i) Despite a smaller NMDAR conductance 1112

1066 spectrum disorders [110, 111], intellectual disabil- in PV interneurons, we observe a robust transcrip- 1113

1067 ity [112], psychosis [113], epilepsy and associated tional coupling via NMDARs, as observed by sev- 1114

1068 comorbidities [114, 115]. While broadly aberrant eral distinct gene expression abnormalities in this 1115

Mahadevan et al., 2021 19/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1116 cell type relevant to human Sz. Therefore, PV- 1117 expressed NMDARs primarily serve to regulate 1118 transcriptional coupling, mediating PV-subtype 1119 abundances. (ii) The developmental window for 1120 NMDAR loss of function is particularly impor- 1121 tant because, its transcriptional regulation main- 1122 tains the correct synaptogenic and assembly cues, 1123 which when lost, lead to disease-causing impaired fl/fl 1124 connectivity. Perhaps, in the Grin1 : Pvalb- 1125 Cre mouse line, the Grin1 -ablation occurs only 1126 at a developmental window when synaptic con- 1127 nectivity is sufficiently complete, explaining why 1128 the animal model does not lead to profound Sz- 1129 like impairments. (iii) The dendrite targeting SST 1130 and NGFC interneurons also exhibit robust NM- 1131 DAR signaling and transcriptional coupling. Dur- 1132 ing aberrant NMDAR-transcriptional coupling, it 1133 is therefore likely that impaired dendritic connec- 1134 tivity and inhibition onto pyramidal neurons also 1135 contributes towards disease etiology. Therefore, 1136 our dataset provides credence to interneuronal 1137 subtype-specific granularilty, connectivity and ex- 1138 citability, all playing combinatorial and mutually- 1139 supporting roles during disease etiology. Taken 1140 together, our study presents a rich resource, lay- 1141 ing the road map for systematic examination of 1142 NMDAR signaling in interneuron subtypes, by 1143 providing multiple molecular targets for exami- 1144 nation in both normal and impaired circuits.

Mahadevan et al., 2021 20/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1145 4. STARFMETHODS

Table 1: Key Resource Table

Reagent/Resource Source Product catalog, Identifier 1. Transgenic mice Nkx2.1-Cre driver line Jackson Laboratory Stock No: 008661, RRID:IMSR_JAX:008661 Ai14, Floxed reporter Jackson Laboratory Stock No: 007914, RRID:IMSR_JAX:007914 Grin1, Floxed line Jackson Laboratory Stock No: 005246, RRID:IMSR_JAX:005246 Ribotag, Floxed line Jackson Laboratory Stock No: 011029, RRID:IMSR_JAX:011029 2. Antibodies Mouse anti-Parvalbumin Sigma-Aldrich P3088, RRID: AB_477329 Rat anti-Somatostatin Millipore MAB354, RRID: AB_2255365 Rabbit anti-Satb1 Abcam AB109122, RRID: AB_10862207 Rabbit Anti-hemagglutinin (HA) Abcam AB9110, RRID: AB_307019 3. Commercial Kits RNA 6000 Pico Kit Agilent 5067-1513 RNeasy Plus Micro Kit Qiagen 74034 SMARTer. Stranded Total RNA-Seq TaKaRa 634413 Kit v2—Pico Input Mammalian Chromium Single Cell 3” GEM, Li- 10X Genomics 1000075 brary & Gel Bead Kit v2, v3, 16 re- actions RNAscope Hiplex 12 Reagent Kit Advanced Cell Diagnostics 324102

Papain dissociation system Worthington LK003150 4. Chemicals/reagents Actinomycin D Sigma-Aldrich A1410

Cycloheximide Sigma-Aldrich C7698

Heparin Sigma-Aldrich H3393

Dynabeads Protein G ThermoFisher Scientific 10004D

DRAQ5 ThermoFisher Scientific 62251 DAPI ThermoFisher Scientific 62248 5. Oligonucleotides Mm-Reln-T9 Advanced Cell Diagnostics 405981-T9 Mm-Sst-T10 Advanced Cell Diagnostics 404631-T10 Continued on next page

Mahadevan et al., 2021 21/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Table 1 continued Mm-Nos1-T12 Advanced Cell Diagnostics 437651-T12 6. Softwares, algorithms R v3.5.1, The R Project for Statisti- https://cran.r-project.org/mirrors.html cal Computing [121] CellRanger v2.2.0, 10X Genomics https://support.10xgenomics.com/single-cell-gene-expressio n/software/pipelines/latest/installation Seurat, v3.1.5, v3.1.5.9000 [47, 122] https://github.com/satijalab/seurat/releases/tag/v3.0.0 MAST, within the R framework [123] https://github.com/RGLab/MAST/ DeSeq2 (1.22.1), within the R frame- https://bioconductor.org/packages/release/bioc/html/DES work [124] eq2.html Morpheus, within the R frame- https://software.broadinstitute.org/morpheus work [125] Enhanced Volcano, within the R https://Github.Com/Kevinblighe/EnhancedVolcano framework [126] Cytoscape v3.8.0 [127] https://cytoscape.org/ iRegulon, within the Cytoscape http://iregulon.aertslab.org/help.html framework [62] Bitplane Imaris 9.5.1, 9.6 http://www.bitplane.com Graphpad Prism 8 https://www.graphpad.com/scientific-software/prism/ 7. Data availability Gene Expression Omnibus Accession https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GS GSE156201 E156201

Mahadevan et al., 2021 22/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

fl/fl + 1146 5. DETAILED MATERIALS AND 1 -Cre: Grin1 : TdT mice were processed 1192 1147 METHODS on consecutive days for single-cell sequencing 1193

experiments. TdT negative animals were pro- 1194

1148 Contact for Reagent and Resource Shar- cessed in parallel for initially setting FACS gate 1195

1149 ing for the Tomato-channel. Across the replicates, 1196

1150 Further information and requests for resources 10XMGE-Grin1-WT and 6XMGE-Grin1-null 1197

1151 and reagents should be directed to and reason- animals were used for the scRNAseq. Coro- 1198

1152 able requests will be fulfilled by the Lead Contact, nal slices containing frontal cortex and hip- 1199

1153 Chris McBain ([email protected]). pocampus (350mM) were cut using VT-1000S 1200

1154 Animals vibratome (Leica Microsystems) in cold NMDG- 1201 2+ 1155 All experiments were conducted in accor- HEPES–based high-Mg cutting solution. Slices 1202 ◦ 1156 dance with animal protocols approved by the were recovered in the same solution at 20 C for 1203

1157 National Institutes of Health. MGE-derived 30 minutes during when, they were visually in- 1204 fl/fl 1158 interneuron-specific Grin1 line, and Ribotag spected under fluorescence microscope and micro 1205

1159 crosses were conducted as indicated in Figure1- dissected, all under constant carbogenation. The 1206

1160 Supplement1A, and Figure4-Supplement1D. recovery and microdissection were conducted in 1207 wt/wt 2+ 1161 Littermate MGE-Grin1 controls, and MGE- the NMDG-HEPES high-Mg solution supple- 1208 fl/fl 1162 Grin1 male and female mice were used during mented with 0.5µM tetrodotoxin (TTX), 50 µM 1209

1163 this study. Mice were housed and bred in conven- DL -2-Amino-5-phosphonopentanoic acid (APV) 1210

1164 tional vivarium with standard laboratory chow and 10 µM Actinomycin-D (Act-D). 1211

1165 and water in standard animal cages under a 12hr Cell dissociation was performed using the Wor- 1212

1166 circadian cycle. Genotyping of the mice were per- thington Papain Dissociation System according 1213

1167 formed as indicated in the appropriate Jackson to manufacturer instructions with minor modi- 1214

1168 Laboratory mice catalog. fications. Briefly, single-cell suspensions of the 1215

1169 Single-cell dissociation and FACS micro dissected frontal cortices or hippocampus 1216 wt/wt + 1170 P18-20 juvenile Nkx2 -1-Cre: Grin1 : TdT were prepared using sufficiently carbogenated dis- 1217 fl/fl + 1171 and Nkx2-1-Cre: Grin1 : TdT mice were sociation solution (containing Papain, DNAse in 1218

1172 used for single-cell sequencing experiments. All Earle’s Balanced Salt Solution, EBSS), supple- 1219

1173 mice were anesthetized with isoflurane and then mented with 1µM TTX, 100 µM APV and 20 1220

1174 decapitated. Brain dissection, slicing and FACS µM Act-D. After a 60 min enzymatic digestion at 1221 ◦ 1175 sorting were carried out as described [43, 44], with 37 C, followed by gentle manual trituration with 1222

1176 slight modifications. NMDG-HEPES–based solu- fire-polished Pasteur pipettes, the cell dissociates 1223 ◦ 1177 tion was used in all steps to enable better recovery were centrifuged at 300g for 5 minutes at 20 C, 1224

1178 of the cells [45] during FACS sorting and single- and the supernatants were discarded. The en- 1225

1179 cell bar coding. Briefly, the brain sectioning solu- zymatic digestion was quenched in the next step 1226 2+ 1180 tion contained NMDG-HEPES–based high-Mg by the addition of ovomucoid protease inhibitor. 1227

1181 cutting solution contained 93 mM NMDG, 2.5 Albumin density gradient was performed on the 1228 1182 mM KCl, 1.2 mM NaH2PO4, 30 mM NaHCO3, 20 pellets, using a sufficiently carbogenated debris 1229 1183 mM HEPES, 25 mM glucose, 5 mM sodium ascor- removal solution (containing albumin-ovomucoid 1230

1184 bate, 3 mM sodium pyruvate, 2mM Thiourea, 10 inhibitor, DNAse in EBSS). The resulting cell pel- 1231 1185 mM MgSO4*7H2O, and 0.5 mM CaCl2*2H2O; it lets were resuspended in 1ml FACS buffer contain- 1232 1186 was adjusted to pH 7.4 with 12.1N HCl, an osmo- ing 10% FBS, 10U/µl of DNAse, 1µM TTX, 100 1233

1187 larity of 300-310 mOsm, and carbogenated (mix µM APV and 20 µM Act-D in a 50:50 mix of car- 1234 1188 of 95% O2 and 5% CO2) before use. This solution bogenated EBSS: NMDG-HEPES–based cutting 1235 1189 was chilled and the process of sectioning were con- saline (with 1mM MgSO4*7H2O, it is important 1236 2+ 1190 ducted on a ice-chamber in the vibratome. to not use High-Mg in the FACS buffer, as it 1237 wt/wt + 1191 3-4, Nkx2-1 -Cre: Grin1 : TdT or Nkx2- interferes with the subsequent 10X scRNAseq re- 1238

Mahadevan et al., 2021 23/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1239 action). Cells were placed in polystyrene tubes reads per cell (raw), median genes per cell respec- 1286

1240 (Falcon 352235) on ice during the FACS. tively, are as follows Cortical WT: 3851, 22.8K, 1287 + 1241 For single cell sorting of TdT expressing cells 1536; Cortical NULL: 2898, 23.5K, 2759; Hip- 1288

1242 by FACS, resuspended cell dissociates were fil- pocampal WT: 4600, 23.6K, 850; Hippocampal 1289

1243 tered through 35mm cell strainer (Falcon 352235) NULL: 4436, 25.8K, 3143. Replicate 3 of the 1290

1244 to remove cell clumps. The single cell suspensions scRNAseq were performed using 10X v3 reac- 1291

1245 were then incubated with 1mg/ml DAPI (1:500) tion from which, cell estimates, mean reads per 1292 ◦ 1246 and 1mM DRAQ5 at 4 C for 5 minutes to label cell (raw), median genes per cell respectively, are 1293

1247 dead cells and live cells respectively. Samples were as follows Cortical WT: 3960, 24.8K, 2870; Hip- 1294

1248 analyzed for TdTomato expression and sorted us- pocampal WT: 3159, 26.9K, 2956. Representa- 1295

1249 ing a MoFlo Astrios EQ high speed cell sorter tive quality metrics from Replicate 2 are indi- 1296

1250 (Beckman Coulter). TdT-negative cells were used cated in Figure1-Supplement1B,C ,D,E. De- 1297

1251 as a control to set the thresholding FACS gate for multiplexed samples were aligned to the mouse 1298

1252 the detection and sorting of the Ai14-TdTomato- reference genome (mm10). The end definitions 1299

1253 expressing cells, and the same gate was then of genes were extended 4k bp downstream (or 1300

1254 applied for all subsequent experiments. Flow halfway to the next feature if closer), and con- 1301

1255 data analysis and setting of sorting gates on verted to mRNA counts using the Cell Ranger 1302

1256 live (DAPI-negative, DRAQ5-positive) and Ai14- Version 2.1.1, provided by the manufacturer. 1303

1257 TdTomato-expressing cells were carried out using MGE-Ribotag-seq 1304

1258 Summit software V6.3.016900 (Beckman Coul- This assay was performed as recently de- 1305 cre fl/wt 1259 ter). Per sample/session, 20,000 – 40,000 indi- scribed [70]. Notably, Nkx2.1 :Grin1 were 1306

1260 vidual cells were sorted into a FBS-precoated, bred with Ribotag mice for >10 generations 1307 HA/HA 1261 Eppendorf LoBind Microcentrifuge tubes contain- to obtain homozygosity in Rpl22 . Age- 1308 cre wt/wt HA/HA 1262 ing carbogenated 10ml FACS buffer, that served matched Nkx2.1 :Grin1 :Rpl22 and 1309 cre fl/fl HA/HA 1263 as the starting material for 10X Genomics bar- Nkx2.1 :Grin1 :Rpl22 littermates (2 1310

1264 coding. males and 2 females per genotype) were utilized 1311

1265 10X Genomics Chromium for the MGE-Ribotag-seq. RNAs bound with 1312

1266 The cells were inspected for viability, counted, anti-HA immunoprecipitates and RNA from bulk 1313

1267 and loaded on the 10X Genomics Chromium sys- tissue (input) were purified using RNeasy Plus 1314

1268 tem, aiming to recover ~5000 cells per condi- Micro Kit and the quality of RNA were measured 1315

1269 tion. 12 PCR cycles were conducted for cDNA using RNA 6000 Pico kit and 2100 Bioanalyzer 1316

1270 amplification, and the subsequent library prepa- system. cDNA libraries were constructed from 1317

1271 ration and sequencing were carried out in ac- 250 pg RNA using the SMARTer Stranded Total 1318

1272 cordance with the manufacturer recommendation RNA-Seq Kit v2 only from samples with RNA In- 1319 TM 1273 (Chromium Single Cell 3’ Library & Gel Bead tegrity Numbers > 6. Sequencing of the libraries 1320

1274 Kit v2 and v3, 16 reactions). Sequencing of the li- were performed on the Illumina HiSeq2500, at 50 1321

1275 braries were performed on the Illumina HiSeq2500 million 2 x 100bp paired-end reads per sample. 1322

1276 at the NICHD, Molecular Genomics Core facil- ~75% of reads were uniquely mapped to genomic 1323

1277 ity. Replicate 1 of the scRNAseq were performed features in the reference genome. Bioconductor 1324

1278 using 10X v2 reaction from which, the cell esti- package DESeq2(1.22.1) was used to identify 1325

1279 mates, mean reads per cell (raw), median genes differentially expressed genes (DEG). This pack- 1326

1280 per cell respectively, are as follows Cortical WT: age allows for statistical determination of DEGs 1327

1281 1277, 149K, 4615; Cortical NULL: 181, 159K, using a negative binomial distribution model. 1328

1282 4826; Hippocampal WT: 2221, 92K, 2578; Hip- The resulting values were then adjusted using the 1329

1283 pocampal NULL: 404, 154K, 4903. Replicate 2 Benjamini and Hochberg’s method for controlling 1330

1284 of the scRNAseq were performed using 10X v3 the false discovery rate (FDR). 1331

1285 reaction from which, the cell estimates, mean Data processing, analyses, visualization 1332

Mahadevan et al., 2021 24/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1333 and differential expression testing MGE subcluster using the FindAllMarkers() func- 1380

1334 Processing (load, align, merge, cluster, dif- tion. The genes thus identified for the integrated 1381

1335 ferential expression testing) and visualization of data is presented in Supplemental Table1b. De- 1382

1336 the scRNAseq datasets were performed with termination of MGE and non-MGE identities are 1383

1337 the R statistical programming environment [128] performed based on existing interneuron litera- 1384

1338 (v3.5.1) and Seurat package (v3.1.5, a develop- ture and other scRNAseq datasets [1, 43, 50, 53, 1385

1339 ment version of Seurat v3.1.5.9000 was used to 89, 91, 129–131]. The labels from Figures 1 and 2 1386

1340 generate violin plots in 2C and 5B) [47, 48] . Data are matched with the top gene markers identified 1387

1341 set preprocessing, comparison of WT- and NULL- by the FindAllMarkers() function and the simi- 1388

1342 Ai14 cells, canonical correlation analyses, and dif- larly named clusters in Figures 1 and 2 have the 1389 same identities. Lastly, for the integrated anal- 1390 1343 ferential expression of genes (padj < 0.01) within 1344 the same cluster between WT- and NULL-Ai14 yses and differential expression testing, we first 1391

1345 cells were performed according to default Seurat merged the identities of the subclusters SST.1- 1392

1346 parameters, unless otherwise mentioned. Qual- SST.6 and TH.1, and relabelled as SST sub- 1393

1347 ity control filtering was performed by only includ- set; PVALB.1-3 relabelled as PVALB subset; and 1394

1348 ing cells that had between 200-6000 unique genes, NGFC.1-2 relabelled as the NGFC subset during 1395

1349 and that had <30% of reads from mitochondrial subsequent analysis (Figure3). 1396

1350 genes. While the WT replicates had no cells above Differential gene expression testing were per- 1397

1351 30% mitochondrial genes, only NULL replicates formed using the MAST package within the 1398

1352 from both brain regions exhibited 7-12% of cells FindMarkers function to identify the differen- 1399

1353 above this threshold. Suggestive of inherent bi- tially expressed genes between two subclusters. 1400

1354 ological impact of Grin1 -ablation, we repeated MAST utilizes a hurdle model with normalized 1401

1355 the clustering and subsequent analyses without UMI as covariate to generate the differential fold 1402

1356 excluding any cells. These analyses did not alter changes [123] , and is known to result in un- 1403

1357 the clustering or skew the gene list. Clustering derestimation of the magnitude of fold change 1404

1358 was performed on the top 25 PCs using the func- (FC) [132]. Therefore, while applying a strin- 1405

1359 tion FindClusters() by applying the shared near- gent false-discovery rate <0.01, we determined 1406

1360 est neighbor modularity optimization with vary- the minimum FC based on the control gene 1407

1361 ing clustering resolution. A cluster resolution of Grin1, which is the target gene knocked out in 1408

1362 1.0 was determined to be biologically meaningful, MGE-derived interneuron celltypes. Notably for 1409

1363 that yielded all known MGE cardinal classes. Ini- Grin1, we had previously demonstrated that the 1410

1364 tial analyses were performed on the WT datasets NGFCs which carry maximum NMDAR compo- 1411

1365 separately (WT.alone), and similar set of analysis nent among MGEs, are devoid of NMDAR cur- 1412

1366 parameters were applied when the WT and NULL rent at this comparable age [9] . In the present 1413

1367 samples were merged (WT.NULL.integrated) for scRNAseq assay, we observe a logFC for Grin1 1414

1368 subsequent differential expression testing. Phy- ranging between -0.1 to -0.35 across both brain 1415

1369 logenetic tree relating the ’average’ cell from regions and all MGE subtypes. Therefore, we de- 1416

1370 each identity class based on a distance matrix termined a minimum logFC in our DEGs as ±0.1 1417

1371 constructed in gene expression space using the to be meaningful. Previous studies have demon- 1418

1372 BuildClusterTree() function. Overall, we iden- strated the MAST approach for DEG testing to be 1419

1373 tified 27, and 33 clusters using this approach powerful in determining subtle changes in highly 1420

1374 in the WT.alone, and WT.NULL.integrated as- transcribed genes, and among abundant popula- 1421

1375 says respectively. The WT.alone correspond to tions, additional to under representing changes 1422

1376 11 MGE.GAD1/2 clusters (Figure1&2), while among weakly transcribed genes [123, 132] . Vol- 1423

1377 the WT.NULL.integrated assay correspond to cano plots and Heat maps for the DEG were gen- 1424

1378 12 clusters (Figure5-Supplement.1). We first erated using EnhancedVolcano package and Mor- 1425

1379 searched for the top differential markers for each pheus package within the R framework. Associ- 1426

Mahadevan et al., 2021 25/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1427 ated scRNAseq source data, including sequencing paraformaldehyde. The brains were post-fixed in 1474 ◦ 1428 reads and single cell expression matrices, is avail- the same fixative for overnight at 4 C for the im- 1475

1429 able from the Gene Expression Omnibus (GEO) munostaining assays. Postfixed brains were seri- 1476

1430 under accession number GSE156201. ally dehydrated using 10%/20%/30% sucrose so- 1477 ◦ 1431 Pathway analyses PPI network mapping lutions at 4 C. Coronal sections (50 µm) were 1478

1432 and disease mapping cut on a freezing microtome. Immunostaining 1479

1433 Ingenuity Pathway Analyses were conducted was performed on free-floating sections. Tissue 1480

1434 on the differentially expressed genes to generate sections were permeabilized and blocked in 1 × 1481

1435 the molecular functional annotation and to iden- PBS + 1% bovine serum albumin + 10% normal 1482

1436 tify the biological pathways and disease path- goat serum + 0.5% Triton X-100 (Carrier PB) at 1483

1437 ways overrepresented. This tool was also used room temperature for 2 h, followed by incubation 1484

1438 to annotate genes with their known cellular func- in primary antibodies, listed below, diluted with 1485

1439 tional classes. Additional Gene Ontology map- 1 × PBS + 1% bovine serum albumin + 1% nor- 1486

1440 ping and KEGG analyses were conducted us- mal goat serum + 0.1% Triton X-100 overnight 1487 ◦ 1441 ing ShinyGO [133]. Protein-protein interaction at 4 C. Tissue sections were then incubated with 1488

1442 (PPI) mapping datasets from a variety of curated secondary antibodies, listed below, diluted in Car- 1489

1443 databases [134–136] were conducted as previously rier Solution (1:1000), and DAPI (1:2000) at room 1490

1444 described [76] [137] using the Cytoscape [127] temperature for 1–2 h and mounted on Super- 1491

1445 platform (v3.8.0). Schizophrenia risk genes in- frost glass slides, and coverslipped using Mowiol 1492

1446 tegrated from various sources including genome- mounting medium and 1.5 mm cover glasses. 1493

1447 wide association studies (GWAS), copy num- Antibodies 1494

1448 ber variation (CNV), association and linkage The following primary antibodies were used: 1495

1449 studies, post-mortem human brain gene expres- mouse anti-PV (1:1000; Sigma-Aldrich Cat# 1496

1450 sion, expression quantitative trait loci (eQTL) P3088, RRID: AB_477329), rat anti-SST 1497

1451 and encyclopedia of DNA elements (ENCODE), (1:1000; Millipore Cat# MAB354, RRID: 1498

1452 were downloaded from http://www.szdb.org/ [73] AB_2255365), rabbit anti-Satb1 (1:1000; Ab- 1499

1453 . Autism Spectrum Disorder risk genes inte- cam Cat# ab109122, RRID: AB_10862207), 1500

1454 grated from various sources were downloaded rabbit anti-HA (Abcam, Cat#ab9110, RRID 1501

1455 from Simons Foundation https://gene.sfari.org/ AB_307019). Secondary antibodies were conju- 1502

1456 [74]. SZDB genes that had a integrated total score gated with Alexa Fluor dyes 488 or 633 (1:1000; 1503

1457 of 3-6 (1419 genes, 22% out of 6387) were con- Thermo Fisher Scientific). 1504

1458 sidered ‘high-risk’ for DEG mapping (Supple- RNAscope HiPlex in situ hybridization 1505

1459 mental Table5a). SFARI genes scored 1-2 with All apparatus, dissection instruments for this 1506

1460 accounting for a high strength of evidence (392 assay were treated to maintain RNAse-free condi- 1507

1461 genes, 42% out of 943), were considered ‘high-risk’ tions. Brains from two sets of P20-25 littermates 1508

1462 for DEG mapping (Supplemental Table5b). from both genotypes were rapidly dissected and 1509

1463 Transcriptional factor motif enrichment search us- rinsed in fresh ice-cold RNAse-free 1XPBS. In- 1510

1464 ing the iRegulon was also conducted using Cy- dividual brains were placed into cryomolds and 1511

1465 toscape using default parameters. snap-freezed by dipping into isopentene:dry ice 1512 ◦ 1466 Immunostaining mix. Frozen brains were stored in -80 C. 15µm 1513

1467 All solutions were freshly prepared and filtered fresh frozen saggital sections were obtained us- 1514

1468 using 0.22µm syringe filters for parallel treat- ing a RNAse-free cryostat, and sections serially 1515

1469 ments of wildtype and MGE-Grin1-null groups. mounted on RNAse-free Superfrost glass slides. 1516

1470 Adult mice of postnatal day (PD) 30/60/210 were After a series of pretreatments with 4%fresh 1517

1471 Mice were deeply anesthetized with isoflurane and PFA and dehydration using 50%, 70% and 100% 1518

1472 perfused transcardially with 1X phosphate buffer EtOH, the slides were processed for HiPlex assay 1519

1473 saline (PBS) and followed by the fixative 4% according to ACDBio manufacturer’s instructions 1520

Mahadevan et al., 2021 26/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1521 within a week. A panel of HiPlex probes were mals (2-3 sections per animal) were used for each 1568 1522 utilized counterstained with DAPI (1:2000), but condition, and section depth were matched be- 1569 1523 only data from Reln, Sst, Nos1 are presented in tween the genotypes for parallel HiPlex assay. 1570 1524 this study. Fluorescent images were captured using the 40X- 1571 1525 Image acquisition and analysis oil objective of a Nikon Widefield Fluorescence, 1572

1526 Immunostaining: Mouse brains from 3–6 ani- Spinning Disk Confocal microscope. All sections 1573

1527 mals (2-3 sections per animal) were used for each were imaged using 10x/0.45 CFI PlanApo objec- 1574

1528 condition, and section depth were matched be- tive (imaging settings: Numerical Aperture 0.75, 1575

1529 tween the genotypes for parallel immunostaining. bit depth 16-bit, Exposure 100ms). Confocal 1576

1530 Fluorescent images were captured using the 20X stacks were stitched using NIS Elements (Nikon) 1577

1531 objective of a Nikon Widefield Fluorescence, Spin- before importing them into Imaris software (Bit- 1578

1532 ning Disk Confocal microscope. For all slices with plane, version 9.6). Sst+ Reln+ Nos1 + RNA 1579

1533 immunostained or genetically reported signal, 50 signals that contain DAPI were identified using 1580

1534 µm thin sections were imaged using 10x/0.45 CFI ’Spot’ function as described above, and the over- 1581

1535 PlanApo objective (imaging settings: Numerical lap of Sst+ spots with Reln+ or Nos1 + spots 1582

1536 Aperture 0.75, bit depth 16-bit, Exposure 100ms). were identified by thresholding the signals of Reln 1583

1537 Confocal stacks were stitched using NIS Elements or Nos1 respectively. Total numbers of cell counts 1584

1538 (Nikon) before importing them into Imaris soft- are normalized to the area of the thin section im- 1585

1539 ware (Bitplane, version 9.6). Cell bodies were aged. In Figure2-Supplement1C,D error bars 1586

1540 marked in Imaris software using the ‘Spots’ func- reflect standard error of mean (s.e.m); Two-tailed 1587 + 1541 tion. Nkx2-1 -Cre:TdT RFP+, PV+, SST+, unpaired t-test was performed using Prism8. 1588

1542 SATB1+ cell bodies were detected using the au-

1543 tomatic function, with a signal detection radius of 6. AUTHOR CONTRIBUTIONS 1589 1544 15 µm. The Imaris ‘Quality’ filter was set above 1545 an empirically determined threshold to maximize VM and CJM conceived the project. VM, TJP 1590 1546 the number of detected cells while minimizing and CJM designed the experiments, DM per- 1591 1547 observed false positives. SST+ cell bodies were formed FACS sorting and analysis. VM, YZ, TJP 1592 1548 marked manually using the Imaris ‘Spots’ func- performed 10X scRNAseq. VM performed Ribo- 1593 1549 tion. ROI 3D borders around hippocampus or tag assays. VM, AM, CJR, CE, RD performed 1594 1550 somatosensory cortex, drawn manually using the 10X scRNAseq bioinformatic analyses. VM, XY 1595 1551 Imaris function ‘Surfaces’. Spots were then split and AP conducted RNAscope, immunufloures- 1596 1552 within each ROI using the Imaris function ‘Split cent staining, imaging and analysis. CJM super- 1597 1553 Spots’. Overlap of RFP+ cells with other mark- vised the study. VM and CJM wrote the paper 1598 1554 ers (PV, SST, SATB1) was addressed by filtering and all authors edited the manuscript. 1599 1555 the RFP+ Spots above an empirically determined

1556 threshold intensity in the channel relative to the ACKNOWLEDGEMENT 1600 1557 marker of interest. Each image with an automatic

1558 analysis by Imaris was checked by an expert and This work was supported by Eunice Kennedy 1601

1559 incorrectly identified cell bodies where refined if Shriver NICHD Intramural Award to CJM. We 1602

1560 required. Total numbers of cell counts are normal- thank Steven L. Coon, Tianwei Li and James R. 1603

1561 ized to the volume of the section imaged, and nor- Iben at the Molecular Genomics Core, NICHD, 1604

1562 malized to the Ai14 cell numbers where applica- for RNA sequencing and bioinformatics support. 1605

1563 ble. In Figure4E, Figure2-Supplement1A,B We thank Vincent Schram and Lynne Holtzclaw 1606

1564 error bars reflect standard error of mean (s.e.m); of the NICHD Microscopy and Imaging Core for 1607

1565 Two-tailed unpaired t-test was performed using imaging support, and we thank Carolina Bengts- 1608

1566 Prism8. son Gonzales for assistance with improving cell 1609

1567 RNAscope-HiPlex: Mouse brains from 2 ani- viability during dissociation and FACS. We also 1610

Mahadevan et al., 2021 27/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure1, Supp.1 | Schematic overview and quality control for scRNAseq A, Breeding strat-egy, Bi,Ci,Di,Ei, Representative FACS gates to sequentially isolate live: dead cells using DAPI: DRAQ5 staining, singlet-gating and TdT+-reporter gating to obtain reporter-positive MGE-derived interneurons from frontal cortex and hippocampus. Bii,Cii,Dii,Eii, Barcode Rank Plots for cells from WT and NULL mice, demonstrating separation of cell-associated barcodes and those associated with empty partitions. UMI, unique molecular identifier; MR, Mean Reads; MG, Median Genes. Biii,Ciii,Diii,Eiii, Distributions of the total number of genes, percentage of mitochondrial genes and UMIs per cell in control mice Biv,Civ,Div,Eiv , Pearson correlation coefficient of the distributions of the total number of genes and the UMI

Mahadevan et al., 2021 28/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure1, Supp.2 | Biological replicates Representative UMAP plots of the 3 biological replicates from cortex and hippocampus indicates A, similar clustering, and B, similar expression profiles of Nkx2-1 -derived, Gad1 -expressing MGE-derived interneurons and Nkx2-1 derived, Olig1 - expressing oligodendrocytes.

Mahadevan et al., 2021 29/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure1, Supp.3 | Select marker gene expression across the subtypes of merged cortical and hippocampal MGE-Grin1wt/wt Ai, UMAP representation of cardinal MGE markers genes in the cortical and hippocampal merged dataset. Aii, UMAP representation colored by region, highlighting the region-specific enrichments of MGE subsets. Aiii, Pie chart indicating the percentages of cells recovered across the interneuron subtypes from cortex and the hippocampus. B, Violin plot showing the distribution of expression levels of well-known representative cell-type-enriched marker genes across the 11 MGE subtypes.

Mahadevan et al., 2021 30/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure1, Supp.4 | MGE subtype differences between cortex and hippocampus Volcano plot representing the −log10 False Discovery Rate (FDR) versus log2 fold change (FC) between A, TH-expressing MGE subsets and the remaining MGE subset SST, PVALB and NGFC; B, Differential expression of the cardinal MGE classes between cortex and hippocampus, at a fold change ≥0.5 and FDR <10e-5.

Mahadevan et al., 2021 31/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure1, Supp.5 | MGE subtype annotation based on marker expression A, Table indicating the subype-defining marker genes observed in the present study and their descriptions in the previous scRNAseq datasets (*indicates the genes expressed in the cortex-exclusive PVALB.2 subcluster). Representative UMAP plots of MGE subtype-enriched genes in B, SST subclusters, C, NGFC subclusters, and D, PVALB subclusters.

Mahadevan et al., 2021 32/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure2, Supp.1 | Validation of MGE subtype abundances by immunostaining and RNA in situ hybridization subsequent toGrin1 -ablation A, Boxplots indicating the cell counts of MGE-derived interneurons expressing (Ai) Ai14/tdTomato, (Aii) PV, SST immunostaining from P30 somatosensory cortex of Grin1 wt/wtand Grin1 fl/fl. B, Boxplots indicating the cell counts of hippocampal MGE-derived interneurons express-ing (Bi) Ai14/tdTomato, and (Bii) PV, SST immunostaining from P30 wt/wt fl/fl Grin1 and Grin1 . C, Exami-nation of cortical Nos1-expressing Chodl-SST.1 subtype abundances by (Ci) in situ hybridization using Sst and Nos1 RNAscope probes from P20-25 somatosensory cortex, counterstained with DAPI. (Cii) Boxplots indicate Sst(+), Nos1(+), SST(+):Nos1(+)or SST(+):Nos1(−) cell counts.D, Examination of hippocampal Reln- expressing SST.2-4 subtype abun-dances by (Di) in situ hybridization using Sst and Reln RNAscope probes from P20-25 (+) (+) (+) (+) (−) (+) hippocampus, counterstained with DAPI. (Dii) Boxplots indicate Sst , Reln , SST :Reln or SST :Reln cell counts. n = 4-6 brains from each genotype for immunostaining; n = 2 brains (4-6 sections) from each genotype for RNAscope. Error bars reflect s.e.m.; two-tailed unpaired t-test, for statistical analysis.

Mahadevan et al., 2021 33/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure2, Supp.2 | scRNAseq differential recoveries of MGE subtypes Ai,Number of cells recovered across cardinal subtypes SST, PV and NGFC. Aii, Number of cells recovered within the subtypes of PV / SST/ NGFC. B, UMAP representation colored by cardinal MGE-derived interneuron subtypes SST, PVALB and NGFC, highlighting the differential enrichments of cells C, Representative UMAP plots indicating the granularity among PV/SST/NGFC subtypes between both brain regions and both genotypes.

Mahadevan et al., 2021 34/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure2, Supp.3 | Select marker gene expression across the subtypes of merged MGE-derived interneurons from Grin1 wt/wt and Grin1 fl/fl A, Violin plot showing the distribution of expression levels of well-known representative cell-type-enriched marker genes across the MGE subtypes. Bi, −log10 False Discovery Rate (FDR) versus log2 fold change (FC) between Pthlh-PVALB.2 and Pthlh-PVALB.3 at a fold change ≥0.5 and FDR <10e-3. Bii, Dot plots representing the normalized expressions of NGFC marker genes mis expressed in Pthlh-PVALB.3 upon Grin1 -ablation.

1611 thank Steven Hunt, Daniel Abebe for assistance (2011) 10948–10970. doi:10.1523/jneurosci.0323- 1630 11.2011. 1631 1612 with animal colony maintenance. URL https://dx.doi.org/10.1523/jneurosci.0323-11. 1632 2011 1633 1613 References [4] L. Tricoire, K. A. Pelkey, M. I. Daw, V. H. Sousa, 1634 G. Miyoshi, B. Jeffries, B. Cauli, G. Fishell, C. J. 1635 1614 [1] K. A. Pelkey, R. Chittajallu, M. T. Craig, McBain, Common Origins of Hippocampal Ivy and 1636 1615 L. Tricoire, J. C. Wester, C. J. McBain, Hip- Nitric Oxide Synthase Expressing Neurogliaform 1637 1616 pocampal GABAergic Inhibitory Interneurons, Cells, Journal of Neuroscience 30 (6) (2010) 2165– 1638 1617 Physiological Reviews 97 (4) (2017) 1619–1747. 2176. doi:10.1523/jneurosci.5123-09.2010. 1639 1618 doi:10.1152/physrev.00007.2017. URL https://dx.doi.org/10.1523/jneurosci.5123-09. 1640 1619 URL https://dx.doi.org/10.1152/physrev.00007. 2010 1641 1620 2017 [5] F. M. Krienen, M. Goldman, Q. Zhang, R. D. 1642 1621 [2] B. Wamsley, G. Fishell, Genetic and activity- Rosario, M. Florio, R. Machold, A. Saunders, 1643 1622 dependent mechanisms underlying interneuron di- K. Levandowski, H. Zaniewski, B. Schuman (2019). 1644 1623 versity, Nature Reviews Neuroscience 18 (5) (2017) [6] L. Overstreet-Wadiche, C. J. McBain, Neu- 1645 1624 299–309. doi:10.1038/nrn.2017.30. rogliaform cells in cortical circuits, Nature 1646 1625 URL https://dx.doi.org/10.1038/nrn.2017.30 Reviews Neuroscience 16 (8) (2015) 458–468. 1647 1626 [3] L. Tricoire, K. A. Pelkey, B. E. Erkkila, B. W. Jef- doi:10.1038/nrn3969. 1648 1627 fries, X. Yuan, C. J. McBain, A Blueprint for the URL https://dx.doi.org/10.1038/nrn3969 1649 1628 Spatiotemporal Origins of Mouse Hippocampal In- [7] T. Klausberger, P. Somogyi (2008). 1650 1629 terneuron Diversity, Journal of Neuroscience 31 (30)

Mahadevan et al., 2021 35/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure2, Supp.4 | Expression of control genes in the MGE subtypes subsequent to Grin1 -ablation Ai, Split-violin plot from both genotypes indicating the expression of Grin1 in the MGE-derived interneurons, pyramidal neurons and non-neurons. B, Representative UMAP plots of cardinal MGE markers genes.

Mahadevan et al., 2021 36/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure3, Supp.1 | Differential gene expression in the MGE subtypes subsequent to Grin1-ablation Ai, Split-violin plot from both genotypes indicating the expression of Grin1 in the MGE-derived interneurons, pyramidal neurons and non-neurons. A ii , Bar plot denoting the number of genes up/downregulated in the cortical and hippocampal MGE clusters. Volcano plot representing the −log10 False Discovery Rate (FDR) versus log2 fold change (FC) between B, hippocampal and C,cortical MGE cardinal clusters upon Grin1 -loss, at a fold change ≥0.2 and FDR <10e-6.

Mahadevan et al., 2021 37/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure3, Supp.2 | Molecular pathways differentially expressed in MGE subtypes Ai, Venn-diagrams indicating the percentages of DEGs common within MGE subtypes from cortex or hippocampus. A i,Venn-diagrams indicating the percentages of DEGs common within MGE subtypes from cortex and hippocampus. B, Hier-archical clustering tree summarizing the correlation among significant pathways enriched among the DEGs. Pathways with many shared genes are clustered together. Bigger dots indicate more significant P-values. C, Bar plot showing the classification of molecular functions of the DEGs, across the MGE subtypes. Total number of DEGs in the particular molecular class indicated in parentheses.

Mahadevan et al., 2021 38/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure4, Supp.1 | Differential expression of intracellular signaling cascades across subtypes upon Grin1- ablation Heatmap of log2 FC of significant DEGs in cortical and hippocampal MGE cardinal subtypes, showing a subset of A, genes regulating intracellular Ca2+ homeostasis and Ca2+ binding proteins; B, notable second messengers; and, C, Ca2 +dependent / activated kinases and phosphatases. D, Breeding strategy to obtain MGE-derived interneuron-specific expression of Ribotag (Rpl22-HA) transgene in the background of Grin1 fl/fl alleles.

Mahadevan et al., 2021 39/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure5, Supp.1 | Differential expression of fundamental interneuron marker genes across sub-types upon Grin1-ablation Differential expression of cardinal interneuron marker genes that are broadly expressed across interneurons by Ai, scRNAseq approach, and cross-validated using Aii, the Ribotag-seq approach.

Mahadevan et al., 2021 40/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

Figure6, Supp.1 | Aberrant NMDAR-signaling result in misexpression of regulators of membrane excitability that are high-risk Sz genes A, Schema representing the field of hippocampal pyramidal cell innervated by the interneurons. Heatmap of log2 FC of significant DEGs in cortical and hippocampal MGE cardinal subtypes, showing a subset of Bi, Neurotransmitter release machinery. B ii, Postsynaptic GABA / Glutamate receptor complexes, Kcn-, Scn- ion channel complexes and other miscellaneous regulators of excitability.

1651 [8] G. Akgül, C. J. McBain, Diverse roles for y 1665 1652 ionotropic glutamate receptors on inhibitory in- [10] J. H. Cornford, M. S. Mercier, M. Leite, V. Ma- 1666 1653 terneurons in developing and adult brain, The gloire, M. Häusser, D. M. Kullmann, Dendritic 1667 1654 Journal of Physiology 594 (19) (2016) 5471–5490. NMDA receptors in parvalbumin neurons enable 1668 1655 doi:10.1113/jp271764. strong and stable neuronal assemblies (2019). 1669 1656 URL https://dx.doi.org/10.1113/jp271764 doi:10.7554/elife.49872. 1670 1657 [9] R. Chittajallu, J. C. Wester, M. T. Craig, E. Barks- URL https://dx.doi.org/10.7554/elife.49872 1671 1658 dale, X. Q. Yuan, G. Akgül, C. Fang, D. Collins, [11] N. V. D. M. García, R. Priya, S. N. Tuncdemir, 1672 1659 S. Hunt, K. A. Pelkey, C. J. McBain, Affer- G. Fishell, T. Karayannis, Sensory inputs control 1673 1660 ent specific role of NMDA receptors for the cir- the integration of neurogliaform interneurons into 1674 1661 cuit integration of hippocampal neurogliaform cells, cortical circuits, Nature Neuroscience 18 (3) (2015) 1675 1662 Nature Communications 8 (1) (2017) 152–152. 393–401. doi:10.1038/nn.3946. 1676 1663 doi:10.1038/s41467-017-00218-y. URL https://dx.doi.org/10.1038/nn.3946 1677 1664 URL https://dx.doi.org/10.1038/s41467-017-00218- [12] J. A. Matta, K. A. Pelkey, M. T. Craig, R. Chit- 1678

Mahadevan et al., 2021 41/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1679 tajallu, B. W. Jeffries, C. J. McBain, Devel- 034 1735 1680 opmental origin dictates interneuron AMPA and [20] C. L. Magueresse, H. Monyer, GABAergic In- 1736 1681 NMDA receptor subunit composition and plastic- terneurons Shape the Functional Maturation of 1737 1682 ity, Nature Neuroscience 16 (8) (2013) 1032–1041. the Cortex, Neuron 77 (3) (2013) 388–405. 1738 1683 doi:10.1038/nn.3459. doi:10.1016/j.neuron.2013.01.011. 1739 1684 URL https://dx.doi.org/10.1038/nn.3459 URL https://dx.doi.org/10.1016/j.neuron.2013.01. 1740 1685 [13] R. Luján, R. Shigemoto, G. López-Bendito, Glu- 011 1741 1686 tamate and GABA receptor signalling in the [21] N. V. D. M. García, T. Karayannis, G. Fishell, 1742 1687 developing brain, Neuroscience 130 (3) (2005) Neuronal activity is required for the devel- 1743 1688 567–580. doi:10.1016/j.neuroscience.2004.09.042. opment of specific cortical interneuron sub- 1744 1689 URL https://dx.doi.org/10.1016/j.neuroscience. types, Nature 472 (7343) (2011) 351–355. 1745 1690 2004.09.042 doi:10.1038/nature09865. 1746 1691 [14] J. B. Manent, Glutamate Acting on AMPA URL https://dx.doi.org/10.1038/nature09865 1747 1692 But Not NMDA Receptors Modulates the [22] M. Yozu, H. Tabata, N. König, K. Nakajima, Mi- 1748 1693 Migration of Hippocampal Interneurons, Jour- gratory Behavior of Presumptive Interneurons Is 1749 1694 nal of Neuroscience 26 (22) (2006) 5901–5909. Affected by AMPA Receptor Activation in Slice 1750 1695 doi:10.1523/jneurosci.1033-06.2006. Cultures of Embryonic Mouse Neocortex, Devel- 1751 1696 URL https://dx.doi.org/10.1523/jneurosci.1033-06. opmental Neuroscience 30 (1-3) (2008) 105–116. 1752 1697 2006 doi:10.1159/000109856. 1753 1698 [15] J. M. Soria, Receptor-activated Calcium URL https://dx.doi.org/10.1159/000109856 1754 1699 Signals in Tangentially Migrating Cortical [23] O. Marín, Interneuron dysfunction in psychiatric 1755 1700 Cells, Cerebral Cortex 12 (8) (2002) 831–839. disorders, Nat. Rev. Neurosci 13 (2012) 107–120. 1756 1701 doi:10.1093/cercor/12.8.831. [24] K. Nakazawa, K. Sapkota, The origin of NMDA 1757 1702 URL https://dx.doi.org/10.1093/cercor/12.8.831 receptor hypofunction in schizophrenia, Pharma- 1758 1703 [16] E. Hanson, M. Armbruster, L. A. Lau, M. E. Som- cology & Therapeutics 205 (2020) 107426–107426. 1759 1704 mer, Z.-J. Klaft, S. A. Swanger, S. F. Traynelis, S. J. doi:10.1016/j.pharmthera.2019.107426. 1760 1705 Moss, F. Noubary, J. Chadchankar, C. G. Dulla, URL https://dx.doi.org/10.1016/j.pharmthera. 1761 1706 Tonic Activation of GluN2C/GluN2D-Containing 2019.107426 1762 1707 NMDA Receptors by Ambient Glutamate Facili- [25] B. Rico, O. Marín, Neuregulin signaling, cortical 1763 1708 tates Cortical Interneuron Maturation, The Jour- circuitry development and schizophrenia, Current 1764 1709 nal of Neuroscience 39 (19) (2019) 3611–3626. Opinion in Genetics & Development 21 (3) (2011) 1765 1710 doi:10.1523/jneurosci.1392-18.2019. 262–270. doi:10.1016/j.gde.2010.12.010. 1766 1711 URL https://dx.doi.org/10.1523/jneurosci.1392-18. URL https://dx.doi.org/10.1016/j.gde.2010.12.010 1767 1712 2019 [26] A. K. Malhotra, D. A. Pinals, C. M. Adler, I. Elman, 1768 1713 [17] W. Kelsch, Z. Li, S. Wieland, O. Senkov, A. Herb, A. Clifton, D. Pickar, A. Breier, Ketamine-induced 1769 1714 C. Gongrich, H. Monyer, GluN2B-Containing exacerbation of psychotic symptoms and cognitive 1770 1715 NMDA Receptors Promote Glutamate Synapse impairment in neuroleptic-free schizophrenics, Neu- 1771 1716 Development in Hippocampal Interneurons, Jour- ropsychopharmacology 17 (1997) 141–150. 1772 1717 nal of Neuroscience 34 (48) (2014) 16022–16030. [27] H. Homayoun, B. Moghaddam, NMDA Receptor 1773 1718 doi:10.1523/jneurosci.1210-14.2014. Hypofunction Produces Opposite Effects on Pre- 1774 1719 URL https://dx.doi.org/10.1523/jneurosci.1210-14. frontal Cortex Interneurons and Pyramidal Neurons, 1775 1720 2014 Journal of Neuroscience 27 (43) (2007) 11496–11500. 1776 1721 [18] G. Akgül, C. J. McBain, AMPA receptor deletion doi:10.1523/jneurosci.2213-07.2007. 1777 1722 in developing MGE-derived hippocampal interneu- URL https://dx.doi.org/10.1523/jneurosci.2213-07. 1778 1723 rons causes a redistribution of excitatory synapses 2007 1779 1724 and attenuates postnatal network oscillatory ac- [28] R. Nguyen, M. D. Morrissey, V. Mahadevan, 1780 1725 tivity, Scientific Reports 10 (1) (2020) 1333–1333. J. D. Cajanding, M. A. Woodin, J. S. Yeo- 1781 1726 doi:10.1038/s41598-020-58068-6. mans, K. Takehara-Nishiuchi, J. C. Kim, Par- 1782 1727 URL https://dx.doi.org/10.1038/s41598-020-58068- valbumin and GAD65 Interneuron Inhibition in 1783 1728 6 the Ventral Hippocampus Induces Distinct Be- 1784 1729 [19] D. Bortone, F. Polleux, KCC2 Expression Pro- havioral Deficits Relevant to Schizophrenia, Jour- 1785 1730 motes the Termination of Cortical Interneu- nal of Neuroscience 34 (45) (2014) 14948–14960. 1786 1731 ron Migration in a Voltage-Sensitive Calcium- doi:10.1523/jneurosci.2204-14.2014. 1787 1732 Dependent Manner, Neuron 62 (1) (2009) 53–71. URL https://dx.doi.org/10.1523/jneurosci.2204-14. 1788 1733 doi:10.1016/j.neuron.2009.01.034. 2014 1789 1734 URL https://dx.doi.org/10.1016/j.neuron.2009.01. [29] J. E. Belforte, V. Zsiros, E. R. Sklar, Z. Jiang, 1790

Mahadevan et al., 2021 42/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1791 G. Yu, Y. Li, E. M. Quinlan, K. Nakazawa, talés, M. Fernández-Otero, O. Marín, Pyramidal 1847 1792 Postnatal NMDA receptor ablation in corticolim- cell regulation of interneuron survival sculpts cor- 1848 1793 bic interneurons confers schizophrenia-like pheno- tical networks, Nature 557 (7707) (2018) 668–673. 1849 1794 types, Nature Neuroscience 13 (1) (2010) 76–83. doi:10.1038/s41586-018-0139-6. 1850 1795 doi:10.1038/nn.2447. [37] H. Bading, D. Ginty, M. Greenberg, Regulation 1851 1796 URL https://dx.doi.org/10.1038/nn.2447 of gene expression in hippocampal neurons by dis- 1852 1797 [30] T. Korotkova, E. C. Fuchs, A. Ponomarenko, J. von tinct calcium signaling pathways, Science 260 (5105) 1853 1798 Engelhardt, H. Monyer, NMDA Receptor Ablation (1993) 181–186. doi:10.1126/science.8097060. 1854 1799 on Parvalbumin-Positive Interneurons Impairs Hip- URL https://dx.doi.org/10.1126/science.8097060 1855 1800 pocampal Synchrony, Spatial Representations, and [38] H. Bading, M. M. Segal, N. J. Sucher, . H. Dudek, 1856 1801 Working Memory, Neuron 68 (3) (2010) 557–569. ~. Lipton, S. A. Greenberg, M. E (1995). 1857 1802 doi:10.1016/j.neuron.2010.09.017. [39] E. L. Yap, M. E. Greenberg, Activity-Regulated 1858 1803 URL https://dx.doi.org/10.1016/j.neuron.2010.09. Transcription: Bridging the Gap between Neural 1859 1804 017 Activity and Behavior (2018). 1860 1805 [31] V. M. Tatard-Leitman, C. R. Jutzeler, J. Suh, J. A. URL https://doi.org/10.1016/j.neuron.2018.10.013 1861 1806 Saunders, E. N. Billingslea, S. Morita, R. White, [40] S. Cull-Candy, S. Brickley, M. Farrant, NMDA re- 1862 1807 R. E. Featherstone, R. Ray, P. I. Ortinski, A. Baner- ceptor subunits: diversity, development and disease, 1863 1808 jee, M. J. Gandal, R. Lin, A. Alexandrescu, Current Opinion in Neurobiology 11 (3) (2001) 1864 1809 Y. Liang, R. E. Gur, K. E. Borgmann-Winter, 327–335. doi:10.1016/s0959-4388(00)00215-4. 1865 1810 G. C. Carlson, C.-G. Hahn, S. J. Siegel, Pyramidal URL https://dx.doi.org/10.1016/s0959-4388(00) 1866 1811 Cell Selective Ablation of N-Methyl-D-Aspartate 00215-4 1867 1812 Receptor 1 Causes Increase in Cellular and Network [41] H. Komuro, P. Rakic (1993). 1868 1813 Excitability, Biological Psychiatry 77 (6) (2015) [42] Q. Xu, M. Tam, S. A. Anderson, Fate mapping 1869 1814 556–568. doi:10.1016/j.biopsych.2014.06.026. Nkx2.1-lineage cells in the mouse telencephalon, The 1870 1815 URL https://dx.doi.org/10.1016/j.biopsych.2014. Journal of Comparative Neurology 506 (1) (2008) 1871 1816 06.026 16–29. doi:10.1002/cne.21529. 1872 1817 [32] A. R. Mohn, R. R. Gainetdinov, M. G. Caron, B. H. URL https://dx.doi.org/10.1002/cne.21529 1873 1818 Koller, Mice with Reduced NMDA Receptor Expres- [43] K. D. Harris, H. Hochgerner, N. G. Skene, L. Magno, 1874 1819 sion Display Behaviors Related to Schizophrenia, L. Katona, C. B. Gonzales, P. Somogyi, N. Kessaris, 1875 1820 Cell 98 (4) (1999) 427–436. doi:10.1016/s0092- S. Linnarsson, J. Hjerling-Leffler, Classes and 1876 1821 8674(00)81972-8. continua of hippocampal CA1 inhibitory neu- 1877 1822 URL https://dx.doi.org/10.1016/s0092-8674(00) rons revealed by single-cell transcriptomics, 1878 1823 81972-8 PLOS Biology 16 (6) (2018) e2006387–e2006387. 1879 1824 [33] A. M. Bygrave, K. Kilonzo, D. M. Kull- doi:10.1371/journal.pbio.2006387. 1880 1825 mann, D. M. Bannerman, D. Kätzel, Can URL https://dx.doi.org/10.1371/journal.pbio. 1881 1826 N-Methyl-D-Aspartate Receptor Hypofunction in 2006387 1882 1827 Schizophrenia Be Localized to an Individual Cell [44] A. B. Muñoz-Manchado, C. B. Gonzales, A. Zeisel, 1883 1828 Type?, Frontiers in Psychiatry 10 (2019) 835–835. H. Munguba, B. Bekkouche, N. G. Skene, P. Lön- 1884 1829 doi:10.3389/fpsyt.2019.00835. nerberg, J. Ryge, K. D. Harris, S. Linnarsson, 1885 1830 URL https://dx.doi.org/10.3389/fpsyt.2019.00835 J. Hjerling-Leffler, Diversity of Interneurons in the 1886 1831 [34] C. Mayer, C. Hafemeister, R. C. Bandler, Dorsal Striatum Revealed by Single-Cell RNA Se- 1887 1832 R. Machold, R. B. Brito, X. Jaglin, K. All- quencing and PatchSeq, Cell Reports 24 (8) (2018) 1888 1833 away, A. Butler, G. Fishell, R. Satija, Devel- 2179–2190.e7. doi:10.1016/j.celrep.2018.07.053. 1889 1834 opmental diversification of cortical inhibitory in- URL https://dx.doi.org/10.1016/j.celrep.2018.07. 1890 1835 terneurons, Nature 555 (7697) (2018) 457–462. 053 1891 1836 doi:10.1038/nature25999. [45] Y. Tanaka, Y. Tanaka, T. Furuta, Y. Yana- 1892 1837 URL https://dx.doi.org/10.1038/nature25999 gawa, T. Kaneko, The effects of cutting solu- 1893 1838 [35] R. Priya, M. F. Paredes, T. Karayannis, N. Yusuf, tions on the viability of GABAergic interneurons 1894 1839 X. Liu, X. Jaglin, I. Graef, A. Alvarez-Buylla, in cerebral cortical slices of adult mice, Journal 1895 1840 G. Fishell, Activity Regulates Cell Death within of Neuroscience Methods 171 (1) (2008) 118–125. 1896 1841 Cortical Interneurons through a Calcineurin- doi:10.1016/j.jneumeth.2008.02.021. 1897 1842 Dependent Mechanism, Cell Reports 22 (7) (2018) URL https://dx.doi.org/10.1016/j.jneumeth.2008. 1898 1843 1695–1709. doi:10.1016/j.celrep.2018.01.007. 02.021 1899 1844 URL https://dx.doi.org/10.1016/j.celrep.2018.01. [46] Y. E. Wu, L. Pan, Y. Zuo, X. Li, W. Hong, Detecting 1900 1845 007 Activated Cell Populations Using Single-Cell RNA- 1901 1846 [36] F. K. Wong, K. Bercsenyi, V. Sreenivasan, A. Por- Seq, Neuron 96 (2017) 1–27. 1902

Mahadevan et al., 2021 43/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

1903 [47] A. Butler, P. Hoffman, P. Smibert, E. Papalexi, 1695–1709. doi:10.1016/j.celrep.2018.01.007. 1959 1904 R. Satija, Integrating single-cell transcriptomic [56] S. J. Butt, V. H. Sousa, M. V. Fuccillo, J. Hjerling- 1960 1905 data across different conditions, technologies, and Leffler, G. Miyoshi, S. Kimura, G. Fishell, The Re- 1961 1906 species, Nature Biotechnology 36 (5) (2018) 411– quirement of Nkx2-1 in the Temporal Specification 1962 1907 420. doi:10.1038/nbt.4096. of Cortical Interneuron Subtypes, Neuron 59 (5) 1963 1908 URL https://dx.doi.org/10.1038/nbt.4096 (2008) 722–732. doi:10.1016/j.neuron.2008.07.031. 1964 1909 [48] T. Stuart, A. Butler, P. Hoffman, C. Hafemeister, URL https://dx.doi.org/10.1016/j.neuron.2008.07. 1965 1910 E. Papalexi, W. M. Mauck, Y. Hao, M. Stoeckius, 031 1966 1911 P. Smibert, R. Satija, Comprehensive Integration of [57] D. J. Laurie, P. H. Seeburg, Regional and devel- 1967 1912 Single-Cell Data. Cell (2019). opmental heterogeneity in splicing of the rat brain 1968 1913 [49] A. Paul, M. Crow, R. Raudales, M. He, J. Gillis, NMDAR1 mRNA, The Journal of Neuroscience 1969 1914 Z. J. Huang, Transcriptional Architecture of Synap- 14 (5) (1994) 3180–3194. doi:10.1523/jneurosci.14- 1970 1915 tic Communication Delineates GABAergic Neu- 05-03180.1994. 1971 1916 ron Identity, Cell 171 (3) (2017) 522–539.e20. URL https://dx.doi.org/10.1523/jneurosci.14-05- 1972 1917 doi:10.1016/j.cell.2017.08.032. 03180.1994 1973 1918 URL https://dx.doi.org/10.1016/j.cell.2017.08.032 [58] H. Monyer, N. Burnashev, D. J. Laurie, B. Sak- 1974 1919 [50] B. Tasic, V. Menon, T. N. T. Nguyen, T. T. K. mann, P. H. Seeburg, Developmental and regional 1975 1920 Kim, T. Jarsky, Z. Yao, B. B. Levi, L. T. Gray, S. A. expression in the rat brain and functional properties 1976 1921 Sorensen, T. Dolbeare, Adult mouse cortical cell tax- of four NMDA receptors, Neuron 12 (3) (1994) 1977 1922 onomy revealed by single cell transcriptomics, Nat. 529–540. doi:10.1016/0896-6273(94)90210-0. 1978 1923 Neurosci. advance on (2016) 1–37. URL https://dx.doi.org/10.1016/0896-6273(94) 1979 1924 [51] B. Tasic, V. Menon, T. N. T. Nguyen, T. T. K. Kim, 90210-0 1980 1925 T. Jarsky, Z. Yao, B. B. Levi, L. T. Gray, S. A. [59] Z. J. Huang, A. Paul, The diversity of GABAergic 1981 1926 Sorensen, T. Dolbeare, D. Bertagnolli, J. Goldy, neurons and neural communication elements, Na- 1982 1927 N. Shapovalova, S. Parry, C. C. Lee, K. Smith, ture Reviews Neuroscience 20 (9) (2019) 563–572. 1983 1928 A. Bernard, L. Madisen, S. M. Sunkin, M. Hawry- doi:10.1038/s41583-019-0195-4. 1984 1929 lycz, C. Koch, H. Zeng, Z. Yao, C. C. Lee, URL https://dx.doi.org/10.1038/s41583-019-0195-4 1985 1930 N. Shapovalova, S. Parry, L. Madisen, S. M. Sunkin, [60] O. Hobert, I. Carrera, N. Stefanakis, The molec- 1986 1931 M. Hawrylycz, C. Koch, H. Zeng, Adult mouse cor- ular and gene regulatory signature of a neuron, 1987 1932 tical cell taxonomy revealed by single cell transcrip- Trends in Neurosciences 33 (10) (2010) 435–445. 1988 1933 tomics, Nature Neuroscience advance on (January) doi:10.1016/j.tins.2010.05.006. 1989 1934 (2016) 1–37. doi:10.1038/nn.4216. URL https://dx.doi.org/10.1016/j.tins.2010.05.006 1990 1935 [52] J. Winterer, D. Lukacsovich, L. Que, A. M. Sar- [61] K. Y. Kwan, N. Sestan, E. S. Anton, Transcrip- 1991 1936 tori, W. Luo, C. Földy, Single-cell RNA-Seq charac- tional co-regulation of neuronal migration and lam- 1992 1937 terization of anatomically identified OLM interneu- inar identity in the neocortex, Development 139 (9) 1993 1938 rons in different transgenic mouse lines, European (2012) 1535–1546. doi:10.1242/dev.069963. 1994 1939 Journal of Neuroscience 50 (11) (2019) 3750–3771. URL https://dx.doi.org/10.1242/dev.069963 1995 1940 doi:10.1111/ejn.14549. [62] R. Janky, A. Verfaillie, H. Imrichová, B. V. 1996 1941 URL https://dx.doi.org/10.1111/ejn.14549 de Sande, L. Standaert, V. Christiaens, G. Hulsel- 1997 1942 [53] B. Tasic, Z. Yao, L. T. Graybuck, K. A. Smith, T. N. mans, K. Herten, M. N. Sanchez, D. Potier, 1998 1943 Nguyen, D. Bertagnolli, J. Goldy, E. Garren, M. N. D. Svetlichnyy, Z. K. Atak, M. Fiers, J.-C. 1999 1944 Economo, S. Viswanathan, Shared and distinct tran- Marine, S. Aerts, iRegulon: From a Gene List 2000 1945 scriptomic cell types across neocortical areas, Nature to a Gene Regulatory Network Using Large 2001 1946 563 (2018) 72–78. Motif and Track Collections, PLoS Computa- 2002 1947 [54] E. Favuzzi, R. Deogracias, A. Marques-Smith, tional Biology 10 (7) (2014) e1003731–e1003731. 2003 1948 P. Maeso, J. Jezequel, D. Exposito-Alonso, doi:10.1371/journal.pcbi.1003731. 2004 1949 M. Balia, T. Kroon, A. J. Hinojosa, E. F. URL https://dx.doi.org/10.1371/journal.pcbi. 2005 1950 Maraver, B. Rico, Distinct molecular programs 1003731 2006 1951 regulate synapse specificity in cortical inhibitory [63] S. J. Otto, S. R. McCorkle, J. Hover, C. Conaco, 2007 1952 circuits, Science 363 (6425) (2019) 413–417. J. J. Han, S. Impey, G. S. Yochum, J. J. Dunn, 2008 1953 doi:10.1126/science.aau8977. R. H. Goodman, G. Mandel, A New Binding Motif 2009 1954 [55] R. Priya, M. F. Paredes, T. Karayannis, N. Yusuf, for the Transcriptional Repressor REST Uncovers 2010 1955 X. Liu, X. Jaglin, I. Graef, A. Alvarez- Large Gene Networks Devoted to Neuronal Func- 2011 1956 Buylla, G. Fishell, Activity Regulates Cell Death tions, Journal of Neuroscience 27 (25) (2007) 6729– 2012 1957 within Cortical Interneurons through a Calcineurin- 6739. doi:10.1523/jneurosci.0091-07.2007. 2013 1958 Dependent Mechanism., Cell reports 22 (7) (2018) URL https://dx.doi.org/10.1523/jneurosci.0091-07. 2014

Mahadevan et al., 2021 44/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

2015 2007 doi:https://doi.org/10.1523/JNEUROSCI.3583- 2071 2016 [64] A. Rodenas-Ruano, A. E. Chávez, M. J. Cossio, 12.2012. 2072 2017 P. E. Castillo, R. S. Zukin, REST-dependent epige- [72] V. Pachnis, A. Achimastou, R. Lasrado, A. Gare- 2073 2018 netic remodeling promotes the developmental switch falaki, T. Maes, M. Kalaitzidou, M. Denaxa, 2074 2019 in synaptic NMDA receptors, Nature Neuroscience Maturation-Promoting Activity of SATB1 in MGE- 2075 2020 15 (10) (2012) 1382–1390. doi:10.1038/nn.3214. Derived Cortical Interneurons, Cell Reports (2012). 2076 2021 URL https://dx.doi.org/10.1038/nn.3214 doi:10.1016/j.celrep.2012.10.003. 2077 2022 [65] K. Nikouei, A. B. Muñoz-Manchado, J. Hjerling- [73] Y. Wu, Y. G. Yao, X. J. Luo, SZDB: A Database 2078 2023 Leffler, BCL11B/CTIP2 is highly expressed in for Schizophrenia Genetic Research, Schizophr. Bull 2079 2024 GABAergic interneurons of the mouse somatosen- 43 (2017) 459–471. 2080 2025 sory cortex, Journal of Chemical Neuroanatomy 71 [74] S. Foundation (2018). 2081 2026 (2016) 1–5. doi:10.1016/j.jchemneu.2015.12.004. [75] A. S. Cristino, S. M. Williams, Z. Hawi, J. Y. An, 2082 2027 URL https://dx.doi.org/10.1016/j.jchemneu.2015. M. A. Bellgrove, C. E. Schwartz, L. da F Costa, 2083 2028 12.004 C. Claudianos, Neurodevelopmental and neuropsy- 2084 2029 [66] M. Erburu, I. Muñoz-Cobo, T. Diaz-Perdigon, chiatric disorders represent an interconnected molec- 2085 2030 P. Mellini, T. Suzuki, E. Puerta, R. M. ular system, Molecular Psychiatry 19 (3) (2014) 2086 2031 Tordera, SIRT2 inhibition modulate gluta- 294–301. doi:10.1038/mp.2013.16. 2087 2032 mate and serotonin systems in the prefrontal URL https://dx.doi.org/10.1038/mp.2013.16 2088 2033 cortex and induces antidepressant-like ac- [76] J. Liu, M. Li, X. J. Luo, B. Su, Systems-level anal- 2089 2034 tion, Neuropharmacology 117 (2017) 195–208. ysis of risk genes reveals the modular nature of 2090 2035 doi:10.1016/j.neuropharm.2017.01.033. schizophrenia, Schizophr. Res 201 (2018) 261–269. 2091 2036 URL https://dx.doi.org/10.1016/j.neuropharm. [77] K. J. Mitchell, The genetics of neurodevelopmental 2092 2037 2017.01.033 disease, Curr. Opin. Neurobiol 21 (2011) 197–203. 2093 2038 [67] M. Wątroba, D. Szukiewicz, The role of sir- [78] Y. Sakai, C. A. Shaw, B. C. Dawson, D. V. 2094 2039 tuins in aging and age-related diseases, Ad- Dugas, Z. Al-Mohtaseb, D. E. Hill, H. Y. Zoghbi, 2095 2040 vances in Medical Sciences 61 (1) (2016) 52–62. Protein Interactome Reveals Converging Molecular 2096 2041 doi:10.1016/j.advms.2015.09.003. Pathways Among Autism Disorders, Science Trans- 2097 2042 URL https://dx.doi.org/10.1016/j.advms.2015.09. lational Medicine 3 (86) (2011) 86ra49–86ra49. 2098 2043 003 doi:10.1126/scitranslmed.3002166. 2099 2044 [68] N. Sharma, E. A. Pollina, M. A. Nagy, E.- URL https://dx.doi.org/10.1126/scitranslmed. 2100 2045 L. Yap, F. A. DiBiase, S. Hrvatin, L. Hu, 3002166 2101 2046 C. Lin, M. E. Greenberg, ARNT2 Tunes Activity- [79] E. Favuzzi, R. Deogracias, A. Marques-Smith, 2102 2047 Dependent Gene Expression through NCoR2- P. Maeso, J. Jezequel, D. Exposito-Alonso, 2103 2048 Mediated Repression and NPAS4-Mediated Ac- M. Balia, T. Kroon, A. J. Hinojosa, E. F. 2104 2049 tivation, Neuron 102 (2) (2019) 390–406.e9. Maraver, B. Rico, Distinct molecular programs 2105 2050 doi:10.1016/j.neuron.2019.02.007. regulate synapse specificity in cortical inhibitory 2106 2051 URL https://dx.doi.org/10.1016/j.neuron.2019.02. circuits, Science 363 (6425) (2019) 413–417. 2107 2052 007 doi:10.1126/science.aau8977. 2108 2053 [69] E. Sanz, L. Yang, T. Su, D. R. Morris, URL https://dx.doi.org/10.1126/science.aau8977 2109 2054 G. S. McKnight, P. S. Amieux, Cell-type- [80] R. Dolmetsch, Excitation-Transcription Coupling: 2110 2055 specific isolation of ribosome-associated mRNA Signaling by Ion Channels to the Nucleus, 2111 2056 from complex tissues, Proceedings of the National Science Signaling 2003 (166) (2003) pe4–pe4. 2112 2057 Academy of Sciences 106 (33) (2009) 13939–13944. doi:10.1126/stke.2003.166.pe4. 2113 2058 doi:10.1073/pnas.0907143106. URL https://dx.doi.org/10.1126/stke.2003.166.pe4 2114 2059 URL https://dx.doi.org/10.1073/pnas.0907143106 [81] S. M. Cohen, H. Ma, K. V. Kuchibhotla, B. O. 2115 2060 [70] V. Mahadevan, A. Peltekian, C. J. McBain, Trans- Watson, G. Buzsáki, R. C. Froemke, R. W. Tsien, 2116 2061 latome Analyses Using Conditional Ribosomal Tag- Excitation-Transcription Coupling in Parvalbumin- 2117 2062 ging in GABAergic Interneurons and Other Sparse Positive Interneurons Employs a Novel CaM 2118 2063 Cell Types, Current Protocols in Neuroscience 92 (1) Kinase-Dependent Pathway Distinct from Exci- 2119 2064 (2020). doi:10.1002/cpns.93. tatory Neurons, Neuron 90 (2) (2016) 292–307. 2120 2065 URL https://dx.doi.org/10.1002/cpns.93 doi:10.1016/j.neuron.2016.03.001. 2121 2066 [71] Satb1 is an activity-modulated transcription URL https://dx.doi.org/10.1016/j.neuron.2016.03. 2122 2067 factor required for the terminal differentia- 001 2123 2068 tion and connectivity of medial ganglionic [82] N. Flames, R. Pla, D. M. Gelman, J. L. R. Ruben- 2124 2069 eminence-derived cortical interneurons, Journal stein, L. Puelles, O. Marin, Delineation of Multi- 2125 2070 of Neuroscience 32 (49) (2012) 17690–17705. ple Subpallial Progenitor Domains by the Combi- 2126

Mahadevan et al., 2021 45/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

2127 natorial Expression of Transcriptional Codes, Jour- (2020). 2183 2128 nal of Neuroscience 27 (36) (2007) 9682–9695. [91] G. Fishell, A. Kepecs, Interneuron Types as Attrac- 2184 2129 doi:10.1523/jneurosci.2750-07.2007. tors and Controllers (2020). doi:10.1146/annurev- 2185 2130 URL https://dx.doi.org/10.1523/jneurosci.2750-07. neuro-070918-050421. 2186 2131 2007 URL https://dx.doi.org/10.1146/annurev-neuro- 2187 2132 [83] L. Lim, D. Mi, A. Llorca, O. Marín, Devel- 070918-050421 2188 2133 opment and Functional Diversification of Corti- [92] W. Andrews, M. Barber, L. R. Hernadez-Miranda, 2189 2134 cal Interneurons, Neuron 100 (2) (2018) 294–313. J. Xian, S. Rakic, V. Sundaresan, T. H. Rabbitts, 2190 2135 doi:10.1016/j.neuron.2018.10.009. R. Pannell, P. Rabbitts, H. Thompson, L. Ersk- 2191 2136 URL https://dx.doi.org/10.1016/j.neuron.2018.10. ine, F. Murakami, J. G. Parnavelas, The role of 2192 2137 009 Slit-Robo signaling in the generation, migration 2193 2138 [84] N. Dehorter, G. Ciceri, G. Bartolini, L. Lim, I. del and morphological differentiation of cortical in- 2194 2139 Pino, O. Marín, Tuning of fast-spiking interneu- terneurons, Developmental Biology 313 (2) (2008) 2195 2140 ron properties by an activity-dependent transcrip- 648–658. doi:10.1016/j.ydbio.2007.10.052. 2196 2141 tional switch, Science 349 (6253) (2015) 1216–1220. URL https://dx.doi.org/10.1016/j.ydbio.2007.10. 2197 2142 doi:10.1126/science.aab3415. 052 2198 2143 URL https://dx.doi.org/10.1126/science.aab3415 [93] F. Polleux, K. L. Whitford, P. A. Dijkhuizen, T. Vi- 2199 2144 [85] P. Flandin, Y. Zhao, D. Vogt, J. Jeong, J. Long, talis, A. Ghosh, Control of cortical interneuron mi- 2200 2145 G. Potter, H. Westphal, J. L. R. Rubenstein, Lhx6 gration by neurotrophins and PI3-kinase signaling, 2201 2146 and Lhx8 Coordinately Induce Neuronal Expres- Development 129 (2002) 3147–3160. 2202 2147 sion of Shh that Controls the Generation of In- [94] O. Marín, A. Yaron, A. Bagri, M. Tessier-Lavigne, 2203 2148 terneuron Progenitors, Neuron 70 (5) (2011) 939– J. L. R. Rubenstein, Sorting of Striatal and Cortical 2204 2149 950. doi:10.1016/j.neuron.2011.04.020. Interneurons Regulated by Semaphorin-Neuropilin 2205 2150 URL https://dx.doi.org/10.1016/j.neuron.2011.04. Interactions, Science 293 (2001) 872–875. 2206 2151 020 [95] S. Nobrega-Pereira, O. Marin, Transcriptional Con- 2207 2152 [86] E. L.-L. Pai, D. Vogt, A. Clemente-Perez, G. L. trol of Neuronal Migration in the Developing Mouse 2208 2153 McKinsey, F. S. Cho, J. S. Hu, M. Wimer, A. Paul, Brain, Cerebral Cortex 19 (suppl 1) (2009) i107– 2209 2154 S. F. Darbandi, R. Pla, T. J. Nowakowski, L. V. i113. doi:10.1093/cercor/bhp044. 2210 2155 Goodrich, J. T. Paz, J. L. Rubenstein, Mafb and URL https://dx.doi.org/10.1093/cercor/bhp044 2211 2156 c-Maf Have Prenatal Compensatory and Postnatal [96] S. Nóbrega-Pereira, N. Kessaris, T. Du, S. Kimura, 2212 2157 Antagonistic Roles in Cortical Interneuron Fate and S. A. Anderson, O. Marín, Postmitotic Nkx2- 2213 2158 Function, Cell Reports 26 (5) (2019) 1157–1173.e5. 1 Controls the Migration of Telencephalic 2214 2159 doi:10.1016/j.celrep.2019.01.031. Interneurons by Direct Repression of Guid- 2215 2160 URL https://dx.doi.org/10.1016/j.celrep.2019.01. ance Receptors, Neuron 59 (5) (2008) 733–745. 2216 2161 031 doi:10.1016/j.neuron.2008.07.024. 2217 2162 [87] A. Stanco, R. Pla, D. Vogt, Y. Chen, S. Man- URL https://dx.doi.org/10.1016/j.neuron.2008.07. 2218 2163 dal, J. Walker, R. F. Hunt, S. Lindtner, C. A. 024 2219 2164 Erdman, A. A. Pieper, S. P. Hamilton, D. Xu, [97] A. Stanco, C. Szekeres, N. Patel, S. Rao, K. Camp- 2220 2165 S. C. Baraban, J. L. R. Rubenstein, NPAS1 Re- bell, J. A. Kreidberg, F. Polleux, E. S. An- 2221 2166 presses the Generation of Specific Subtypes of Cor- ton, Netrin-1- 3 1 integrin interactions regulate 2222 2167 tical Interneurons, Neuron 84 (5) (2014) 940–953. the migration of interneurons through the corti- 2223 2168 doi:10.1016/j.neuron.2014.10.040. cal marginal zone, Proceedings of the National 2224 2169 URL https://dx.doi.org/10.1016/j.neuron.2014.10. Academy of Sciences 106 (18) (2009) 7595–7600. 2225 2170 040 doi:10.1073/pnas.0811343106. 2226 2171 [88] Z. Liu, Z. Zhang, S. Lindtner, Z. Li, Z. Xu, S. Wei, URL https://dx.doi.org/10.1073/pnas.0811343106 2227 2172 Q. Liang, Y. Wen, G. Tao, Y. You, Sp9 Regulates [98] A. Steinecke, C. Gampe, G. Zimmer, J. Rudolph, 2228 2173 Medial Ganglionic Eminence-Derived Cortical In- J. Bolz, EphA/ephrin A reverse signaling promotes 2229 2174 terneuron Development, Cereb. Cortex (2018) 1–15. the migration of cortical interneurons from the 2230 2175 [89] R. D. Hodge, T. E. Bakken, J. A. Miller, K. A. medial ganglionic eminence, Development 141 (2) 2231 2176 Smith, E. R. Barkan, L. T. Graybuck, J. L. Close, (2014) 460–471. doi:10.1242/dev.101691. 2232 2177 B. Long, N. Johansen, O. Penn, Conserved cell types URL https://dx.doi.org/10.1242/dev.101691 2233 2178 with divergent features in human versus mouse cor- [99] G. Zimmer, P. Garcez, J. Rudolph, R. Niehage, 2234 2179 tex, Nature (2019). F. Weth, R. Lent, J. Bolz, Ephrin-A5 acts as a re- 2235 2180 [90] Z. Yao, T. N. Nguyen, C. T. J. Velthoven, Van, pulsive cue for migrating cortical interneurons, Eu- 2236 2181 J. Goldy, A. E. Sedeno-Cortes, F. Baftizadeh, ropean Journal of Neuroscience 28 (1) (2008) 62–73. 2237 2182 D. Bertagnolli, T. Casper, K. Crichton, S. L. Ding doi:10.1111/j.1460-9568.2008.06320.x. 2238

Mahadevan et al., 2021 46/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

2239 URL https://dx.doi.org/10.1111/j.1460-9568.2008. doi:10.1016/j.cophys.2017.12.013. 2295 2240 06320.x URL https://dx.doi.org/10.1016/j.cophys.2017.12. 2296 2241 [100] Z. Shao, H. Noh, W. B. Kim, P. Ni, C. Nguyen, 013 2297 2242 S. E. Cote, E. Noyes, J. Zhao, T. Parsons, J. M. [109] J. C. Masdeu, J. Dalmau, K. F. Berman, 2298 2243 Park, Dysregulated protocadherin-pathway activity NMDA Receptor Internalization by Autoantibod- 2299 2244 as an intrinsic defect in induced pluripotent stem ies: A Reversible Mechanism Underlying Psychosis?, 2300 2245 cell-derived cortical interneurons from subjects with Trends in Neurosciences 39 (5) (2016) 300–310. 2301 2246 schizophrenia, Nat. Neurosci 22 (2019) 229–242. doi:10.1016/j.tins.2016.02.006. 2302 2247 [101] B. L. Bloodgood, N. Sharma, H. A. Browne, URL https://dx.doi.org/10.1016/j.tins.2016.02.006 2303 2248 A. Z. Trepman, M. E. Greenberg, The activity- [110] J. Tarabeux, , O. Kebir, J. Gauthier, F. F. Ham- 2304 2249 dependent transcription factor NPAS4 regulates dan, L. Xiong, A. Piton, D. Spiegelman, E. Henrion, 2305 2250 domain-specific inhibition, Nature 503 (7474) (2013) B. Millet, F. Fathalli, R. Joober, J. L. Rapoport, 2306 2251 121–125. doi:10.1038/nature12743. L. E. DeLisi, E. Fombonne, L. Mottron, N. Forget- 2307 2252 URL https://dx.doi.org/10.1038/nature12743 Dubois, M. Boivin, J. L. Michaud, P. Drapeau, R. G. 2308 2253 [102] Y. Lin, B. L. Bloodgood, J. L. Hauser, A. D. Lafrenière, G. A. Rouleau, M. O. Krebs, Rare mu- 2309 2254 Lapan, A. C. Koon, T.-K. Kim, L. S. Hu, tations in N-methyl-D-aspartate glutamate recep- 2310 2255 A. N. Malik, M. E. Greenberg, Activity-dependent tors in autism spectrum disorders and schizophre- 2311 2256 regulation of inhibitory synapse development by nia, Translational Psychiatry 1 (11) (2011) e55–e55. 2312 2257 Npas4, Nature 455 (7217) (2008) 1198–1204. doi:10.1038/tp.2011.52. 2313 2258 doi:10.1038/nature07319. URL https://dx.doi.org/10.1038/tp.2011.52 2314 2259 URL https://dx.doi.org/10.1038/nature07319 [111] Y. Yu, Y. Lin, Y. Takasaki, C. Wang, H. Kimura, 2315 2260 [103] X. Sun, Lin, Y, Npas4: Linking Neuronal Activity J. Xing, K. Ishizuka, M. Toyama, I. Kushima, 2316 2261 to Memory, Trends Neurosci 39 (2016) 264–275. D. Mori, Rare loss of function mutations in N- 2317 2262 [104] I. Spiegel, A. R. Mardinly, H. W. Gabel, J. E. methyl-d-aspartate glutamate receptors and their 2318 2263 Bazinet, C. H. Couch, C. P. Tzeng, D. A. Harmin, contributions to schizophrenia susceptibility, Transl. 2319 2264 M. E. Greenberg, Npas4 Regulates Excitatory- Psychiatry 8 (2018) 12–12. 2320 2265 Inhibitory Balance within Neural Circuits through [112] W. Chen, C. Shieh, S. A. Swanger, A. Tankovic, 2321 2266 Cell-Type-Specific Gene Programs, Cell 157 (5) M. Au, M. McGuire, M. Tagliati, J. M. Graham, 2322 2267 (2014) 1216–1229. doi:10.1016/j.cell.2014.03.058. S. Madan-Khetarpal, S. F. Traynelis, H. Yuan, T. M. 2323 2268 URL https://dx.doi.org/10.1016/j.cell.2014.03.058 Pierson, GRIN1 mutation associated with intellec- 2324 2269 [105] B. Ataman, G. L. Boulting, D. A. Harmin, M. G. tual disability alters NMDA receptor trafficking and 2325 2270 Yang, M. Baker-Salisbury, E.-L. Yap, A. N. Ma- function, Journal of Human Genetics 62 (6) (2017) 2326 2271 lik, K. Mei, A. A. Rubin, I. Spiegel, E. Durresi, 589–597. doi:10.1038/jhg.2017.19. 2327 2272 N. Sharma, L. S. Hu, M. Pletikos, E. C. Griffith, URL https://dx.doi.org/10.1038/jhg.2017.19 2328 2273 J. N. Partlow, C. R. Stevens, M. Adli, M. Chahrour, [113] L. L. Gibson, T. A. Pollak, G. Blackman, 2329 2274 N. Sestan, C. A. Walsh, V. K. Berezovskii, M. S. M. Thornton, N. Moran, A. S. David, The 2330 2275 Livingstone, M. E. Greenberg, Evolution of Os- Psychiatric Phenotype of Anti-NMDA Receptor 2331 2276 teocrin as an activity-regulated factor in the pri- Encephalitis, The Journal of Neuropsychiatry 2332 2277 mate brain, Nature 539 (7628) (2016) 242–247. and Clinical Neurosciences 31 (1) (2019) 70–79. 2333 2278 doi:10.1038/nature20111. doi:10.1176/appi.neuropsych.17120343. 2334 2279 URL https://dx.doi.org/10.1038/nature20111 URL https://dx.doi.org/10.1176/appi.neuropsych. 2335 2280 [106] B. Lacar, S. B. Linker, B. N. Jaeger, S. R. Krish- 17120343 2336 2281 naswami, J. J. Barron, M. J. E. Kelder, S. L. Pary- [114] G. L. Carvill, B. M. Regan, S. C. Yendle, B. J. 2337 2282 lak, A. C. M. Paquola, P. Venepally, M. Novotny, O’Roak, N. Lozovaya, N. Bruneau, N. Burnashev, 2338 2283 Nuclear RNA-seq of single neurons reveals molecu- A. Khan, J. Cook, E. Geraghty, L. G. Sadleir, 2339 2284 lar signatures of activation, Nat. Commun 7 (2016) S. J. Turner, M.-H. Tsai, R. Webster, R. Ouvrier, 2340 2285 11022–11022. J. A. Damiano, S. F. Berkovic, J. Shendure, M. S. 2341 2286 [107] C. M. Müller, A. Vlachos, T. Deller, Calcium home- Hildebrand, P. Szepetowski, I. E. Scheffer, H. C. 2342 2287 ostasis of acutely denervated and lesioned den- Mefford, GRIN2A mutations cause epilepsy-aphasia 2343 2288 tate gyrus in organotypic entorhino-hippocampal spectrum disorders, Nature Genetics 45 (9) (2013) 2344 2289 co-cultures, Cell Calcium 47 (3) (2010) 242–252. 1073–1076. doi:10.1038/ng.2727. 2345 2290 doi:10.1016/j.ceca.2009.12.006. URL https://dx.doi.org/10.1038/ng.2727 2346 2291 URL https://dx.doi.org/10.1016/j.ceca.2009.12.006 [115] J. R. Lemke, D. Lal, E. M. Reinthaler, I. Steiner, 2347 2292 [108] W. XiangWei, Y. Jiang, H. Yuan, De novo muta- M. Nothnagel, M. Alber, K. Geider, B. Laube, 2348 2293 tions and rare variants occurring in NMDA recep- M. Schwake, K. Finsterwalder, Mutations in 2349 2294 tors, Current Opinion in Physiology 2 (2018) 27–35. GRIN2A cause idiopathic focal epilepsy with 2350

Mahadevan et al., 2021 47/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

2351 rolandic spikes, Nat. Genet 45 (2013) 1067–1072. URL https://dx.doi.org/10.1093/bioinformatics/ 2407 2352 [116] S. J. Dienel, D. A. Lewis, Alterations in cortical in- btq675 2408 2353 terneurons and cognitive function in schizophrenia, [128] R. C. Team (2013). 2409 2354 Neurobiology of Disease 131 (2019) 104208–104208. [129] A. Paul, M. Crow, R. Raudales, J. Gillis, Z. J. 2410 2355 doi:10.1016/j.nbd.2018.06.020. Huang, Transcriptional Architecture of Synap- 2411 2356 URL https://dx.doi.org/10.1016/j.nbd.2018.06.020 tic Communication Delineates Cortical GABAer- 2412 2357 [117] D. A. Lewis, T. Hashimoto, D. W. Volk, Cor- gic Neuron Identity, Cell 171 (3) (2017) 522–539. 2413 2358 tical inhibitory neurons and schizophrenia, Na- doi:10.1016/j.cell.2017.08.032. 2414 2359 ture Reviews Neuroscience 6 (4) (2005) 312–324. [130] A. Saunders, E. Z. Macosko, A. Wysoker, M. Gold- 2415 2360 doi:10.1038/nrn1648. man, F. M. Krienen, H. de Rivera, E. Bien, 2416 2361 URL https://dx.doi.org/10.1038/nrn1648 M. Baum, L. Bortolin, S. Wang, A. Go- 2417 2362 [118] F. Ali, D. M. Gerhard, K. Sweasy, S. Pothula, C. Pit- eva, J. Nemesh, N. Kamitaki, S. Brumbaugh, 2418 2363 tenger, R. S. Duman, A. C. Kwan, Ketamine dis- D. Kulp, S. A. McCarroll, Molecular Diversity 2419 2364 inhibits dendrites and enhances calcium signals in and Specializations among the Cells of the Adult 2420 2365 prefrontal dendritic spines, Nature Communications Mouse Brain, Cell 174 (4) (2018) 1015–1030.e16. 2421 2366 11 (1) (2020) 72–72. doi:10.1038/s41467-019-13809- doi:10.1016/j.cell.2018.07.028. 2422 2367 8. URL https://dx.doi.org/10.1016/j.cell.2018.07.028 2423 2368 URL https://dx.doi.org/10.1038/s41467-019-13809- [131] Z. Yao, T. N. Nguyen, C. T. J. van Velthoven, 2424 2369 8 J. Goldy, A. E. Sedeno-Cortes, F. Baftizadeh, 2425 2370 [119] R. Chittajallu, K. Auville, V. Mahadevan, M. Lai, D. Bertagnolli, T. Casper, K. Crichton, S.-L. Ding, 2426 2371 S. Hunt, D. Calvigioni, K. A. Pelkey, K. Zaghloul, O. Fong, E. Garren, A. Glandon, J. Gray, L. T. 2427 2372 C. J. Mcbain (2020). Graybuck, D. Hirschstein, M. Kroll, K. Lathia, 2428 2373 [120] K. Nakao, V. Jeevakumar, S. Z. Jiang, Y. Fujita, B. Levi, D. McMillen, S. Mok, T. Pham, Q. Ren, 2429 2374 N. B. Diaz, C. A. P. Annan, K. L. E. Jaunarajs, C. Rimorin, N. Shapovalova, J. Sulc, S. M. Sunkin, 2430 2375 K. Hashimoto, J. E. Belforte, K. Nakazawa (2018). M. Tieu, A. Torkelson, H. Tung, K. Ward, N. Dee, 2431 2376 [121] R-Core-Team, R: A Language and Environment for K. A. Smith, B. Tasic, H. Zeng, A taxonomy 2432 2377 Statistical Computing (2013). of transcriptomic cell types across the isocortex 2433 2378 [122] T. Stuart, A. Butler, P. Hoffman, C. Hafemeister, and hippocampal formation, bioRxiv (mar 2020). 2434 2379 E. Papalexi, W. M. Mauck, Y. Hao, M. Stoeck- doi:10.1101/2020.03.30.015214. 2435 2380 ius, P. Smibert, R. Satija, Comprehensive In- [132] M. Ximerakis, S. L. Lipnick, B. T. Innes, S. K. 2436 2381 tegration of Single-Cell Data, Cell (jun 2019). Simmons, X. Adiconis, D. Dionne, B. A. May- 2437 2382 doi:10.1016/j.cell.2019.05.031. weather, L. Nguyen, Z. Niziolek, C. Ozek, V. L. 2438 2383 [123] G. Finak, A. McDavid, M. Yajima, J. Deng, V. Ger- Butty, R. Isserlin, S. M. Buchanan, S. S. Levine, 2439 2384 suk, A. K. Shalek, C. K. Slichter, H. W. Miller, A. Regev, G. D. Bader, J. Z. Levin, L. L. Ru- 2440 2385 M. J. McElrath, M. Prlic, P. S. Linsley, R. Gottardo, bin, Single-cell transcriptomic profiling of the ag- 2441 2386 MAST: a flexible statistical framework for assessing ing mouse brain, Nature Neuroscience 22 (10) (2019) 2442 2387 transcriptional changes and characterizing hetero- 1696–1708. doi:10.1038/s41593-019-0491-3. 2443 2388 geneity in single-cell RNA sequencing data, Genome URL https://dx.doi.org/10.1038/s41593-019-0491-3 2444 2389 Biology 16 (1) (2015). doi:10.1186/s13059-015-0844- [133] S. X. Ge, D. Jung, R. Yao, ShinyGO: a graph- 2445 2390 5. ical gene-set enrichment tool for animals and 2446 2391 URL https://dx.doi.org/10.1186/s13059-015-0844-5 plants, Bioinformatics 36 (8) (2020) 2628–2629. 2447 2392 [124] M. I. Love, W. Huber, S. Anders, Moderated esti- doi:10.1093/bioinformatics/btz931. 2448 2393 mation of fold change and dispersion for RNA-seq URL https://dx.doi.org/10.1093/bioinformatics/ 2449 2394 data with DESeq2, Genome Biology 15 (12) (2014) btz931 2450 2395 550–550. doi:10.1186/s13059-014-0550-8. [134] K. R. Brown, I. Jurisica, Unequal evolutionary con- 2451 2396 URL https://dx.doi.org/10.1186/s13059-014-0550-8 servation of human protein interactions in interolo- 2452 2397 [125] B. Institute, Morpheus, https://software.broadinsti gous networks, Genome Biology 8 (5) (2007) R95– 2453 2398 tute.org/morpheus. R95. doi:10.1186/gb-2007-8-5-r95. 2454 2399 [126] B. Kevin, EnhancedVolcano: Publication-ready vol- URL https://dx.doi.org/10.1186/gb-2007-8-5-r95 2455 2400 cano plots with enhanced colouring and labeling. [135] M. Kotlyar, C. Pastrello, F. Pivetta, A. L. Sardo, 2456 2401 (2019). C. Cumbaa, H. Li, T. Naranian, Y. Niu, Z. Ding, 2457 2402 [127] M. E. Smoot, K. Ono, J. Ruscheinski, P. L. F. Vafaee, F. Broackes-Carter, J. Petschnigg, G. B. 2458 2403 Wang, T. Ideker, Cytoscape 2.8: new features Mills, A. Jurisicova, I. Stagljar, R. Maestro, I. Ju- 2459 2404 for data integration and network visualiza- risica, In silico prediction of physical protein in- 2460 2405 tion, Bioinformatics 27 (3) (2011) 431–432. teractions and characterization of interactome or- 2461 2406 doi:10.1093/bioinformatics/btq675. phans, Nature Methods 12 (1) (2015) 79–84. 2462

Mahadevan et al., 2021 48/49 NMDAR-mediated control of gene expression in MGE-derived interneurons bioRxiv preprint doi: https://doi.org/10.1101/2020.06.10.144295; this version posted January 29, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

2463 doi:10.1038/nmeth.3178. 2464 URL https://dx.doi.org/10.1038/nmeth.3178 2465 [136] C. Stark, BioGRID: a general repository for inter- 2466 action datasets, Nucleic Acids Research 34 (90001) 2467 (2006) D535–D539. doi:10.1093/nar/gkj109. 2468 URL https://dx.doi.org/10.1093/nar/gkj109 2469 [137] V. Mahadevan, C. S. Khademullah, Z. Dargaei, 2470 J. Chevrier, P. Uvarov, J. Kwan, R. D. Bagshaw, 2471 T. Pawson, A. Emili, Y. D. Koninck (2017).

Mahadevan et al., 2021 49/49 NMDAR-mediated control of gene expression in MGE-derived interneurons