Amelioration of Brain Histone Methylopathies by Balancing a Writer-Eraser Duo

bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 Amelioration of Brain Histone Methylopathies by Balancing a Writer-Eraser Duo

2 KMT2A-KDM5C

3 Christina N. Vallianatos1, 2, Brynne Raines3, Robert S. Porter1,2, Katherine M.

4 Bonefas1,4, Michael C. Wu5, Patricia M. Garay1,4, Katie M. Collette3, Young Ah Seo6,

5 Yali Dou7, Catherine E. Keegan1,8, Natalie C. Tronson3,*, and Shigeki Iwase1,*

7 1Department of Human Genetics, Michigan Medicine, University of Michigan, Ann Arbor,

8 MI 48109

9 2Genetics and Genomics Graduate Program, University of Michigan, Ann Arbor, MI

10 48109

11 3Department of Psychology, College of LS&A, University of Michigan, Ann Arbor, MI

12 48109

13 4The University of Michigan Neuroscience Graduate Program

14 5Neurodigitech, LLC, San Diego, CA

15 6Department of Nutritional Sciences, School of Public Health, University of Michigan,

16 Ann Arbor, MI 48109

17 7Department of Pathology, Michigan Medicine, University of Michigan, Ann Arbor, MI

18 48109

19 8Department of Pediatrics, Michigan Medicine, University of Michigan, Ann Arbor, MI

20 48109

21 *Co-corresponding authors

23 Please address correspondence to: [email protected] and [email protected]

24 Abstract

25 Histone H3 lysine 4 methylation (H3K4me) is extensively regulated by seven writer and

26 six eraser enzymes in mammals. Nine H3K4me enzymes are associated with

27 neurodevelopmental disorders to date, indicating their important roles in the brain.

28 Opposing activities of writer-eraser enzymes highlight activity modulation as a

29 therapeutic strategy. However, interplay among H3K4me enzymes in the brain remains

30 largely unknown. Here, we show functional interactions of a writer-eraser duo, KMT2A

31 and KDM5C, which are responsible for Wiedemann-Steiner Syndrome (WDSTS), and

32 mental retardation X-linked syndromic Claes-Jensen type (MRXSCJ), respectively.

33 Despite opposite enzymatic activities, the WDSTS and MRXSCJ mouse models,

34 deficient for either Kmt2a or Kdm5c, shared reduced dendritic spines and increased

35 aggression. Double mutation of Kmt2a and Kdm5c clearly reversed dendritic

36 morphology deficits and key behavioral traits including aggression, and partially

37 corrected altered transcriptomes and H3K4me landscapes. Thus, our study uncovers

38 common yet mutually suppressive aspects of the WDSTS and MRXSCJ models and

39 provides a proof of principle for balancing a single writer-eraser pair to ameliorate their

40 associated disorders.

41 Introduction

42 Dysregulation of histone methylation has emerged as a major contributor to

43 neurodevelopmental disorders (NDDs) such as autism spectrum disorders and

44 intellectual disabilities (1-3). Histone methylation can be placed on a subset of lysines

45 and arginines by histone methyltransferases (writer enzymes) and serves as a signaling

46 platform for a variety of nuclear events including transcription (4). Reader proteins

47 specifically recognize methylated histones, thereby converting methylation signals into

48 higher-order chromatin structures (5). Histone methylation can be removed by a set of

49 histone demethylases (eraser enzymes) (6). Pathogenic variants in all three classes of

50 methyl-histone regulators cause NDDs, indicating critical, yet poorly understood roles of

51 histone methylation dynamics in brain development and function (7-9).

53 Histone H3 lysine 4 methylation (H3K4me) is one of the most well-characterized histone

54 modifications. H3K4me is primarily found at transcriptionally active areas of the

55 genome. The three states, mono-, di-, and tri-methylation (H3K4me1-3), uniquely mark

56 gene regulatory elements and play pivotal roles in distinct steps of transcription. While

57 H3K4me3/2 are enriched at transcriptionally engaged promoters, H3K4me1 is a

58 hallmark of transcriptional enhancers (10, 11). At promoters, H3K4me3 contributes to

59 recruitment of general transcription machinery TFIID and RNA polymerase II (12, 13).

60 H3K4me1 has been shown to tether BAF, an ATP-dependent chromatin remodeling

61 complex, to enhancers (14).

63 H3K4me is extensively regulated by seven methyltransferases and six demethylases in

64 mammals (6). Consistent with the important roles of H3K4me in transcriptional

65 regulation, genomic distribution of H3K4me appears highly dynamic during brain

66 development (15), in which widespread gene expression changes take place.

67 Developmental H3K4me dynamics appear to be altered in the prefrontal cortices of

68 individuals with autism (16, 17). However, the contributions of each of the 13 enzymes

69 in the dynamic H3K4me landscapes of the developing brain remain largely unknown.

70 Strikingly, genetic alterations in nine of the 13 H3K4me enzymes and at least two

71 H3K4me readers have been associated with human NDDs to date, indicating the critical

72 roles of H3K4me balance (18) (Figure 1A). These human conditions can be collectively

73 referred to as brain H3K4 methylopathies and point to non-redundant yet poorly

74 understood roles of these genes controlling this single post-translational modification for

75 faithful brain development.

77 As histone modifications are reversible, one can, in theory, correct an imbalance by

78 modulating the writers or erasers. Chemical inhibitors of histone deacetylases (HDACs)

79 have been successfully used to rescue phenotypes in mouse models of NDDs. HDAC

80 inhibitors were able to ameliorate learning disabilities in mouse models of Rubinstein-

81 Taybi and Kabuki syndromes, which are deficient for CREBBP or KMT2D, writer

82 enzymes for histone acetylation or H3K4me, respectively (19, 20). However, none of

83 these chemical compounds has yet been proven applicable to human NDDs. Moreover,

84 the HDAC inhibitors, such as SAHA and AR-42, used in these studies interfere with

85 multiple HDACs (21), which could potentially result in widespread side effects. Given

86 the non-redundant roles of the H3K4me enzymes, a more specific perturbation is

87 desirable.

89 In order to achieve specific modulation of H3K4me rather than inhibiting multiple histone

90 modifiers in a given NDD, an important first step is to delineate functional relationships

91 between the H3K4 enzymes. In the present work, we focus on a pair of NDD-associated

92 writer/eraser enzymes: KMT2A and KDM5C. Haploinsufficiency of KMT2A underlies

93 Weidemann-Steiner Syndrome (WDSTS), characterized by developmental delay,

94 intellectual disability, characteristic facial features, short stature, and hypotonia (22, 23).

95 Loss of KDM5C function defines Mental Retardation, X-linked, syndromic, Claes Jensen

96 type (MRXSCJ), in which individuals display an intellectual disability syndrome with

97 aggression, short stature, and occasional autism comorbidity (24-27). Mouse models

98 have provided experimental support for causative impacts of KMT2A and KDM5C

99 deficiencies in impaired cognitive development. Heterozygous Kmt2a knockout (KO)

100 mice show compromised fear learning (28), and excitatory-neuron-specific Kmt2a

101 deletion similarly led to impaired learning, memory, and anxiety behaviors, as well as

102 altered H3K4me3 distributions and transcriptomes (29, 30). In these studies, however,

103 the social behavior and relevant cellular consequences of Kmt2a loss were not

104 examined. Kdm5c-knockout (KO) mice mimic many MRXSCJ features, including small

105 body size, aggressive behavior, and reduced social activity and learning (31, 32).

106 Kdm5c-KO neurons in the basolateral amygdala exhibit malformation of dendritic arbors

107 and spines along with misregulation of neurodevelopmental genes. However, the

108 functional relationships between KMT2A and KDM5C, e.g. how deficiencies in opposing

109 enzymatic activities can lead to same learning deficits, remain elusive. Additionally,

110 despite the strong association with NDDs, successful modulation of histone methylation

111 to restore normal brain physiology has not been reported in these or any other animal

112 models of histone methylopathies.

113

114 In the present work, we tested whether modulating either single H3K4me writer or

115 eraser can ameliorate the neurodevelopmental symptoms observed in the WDSTS and

116 MRXSCJ mouse models. We generated Kmt2a-, Kdm5c-double mutant (DM) mice, and

117 performed systematic comparisons between wild-type (WT), single mutants, and DM

118 mice.

119 Results

120

121 KMT2A and KDM5C co-exist broadly in the brain

122 We first examined expression patterns of KMT2A and KDM5C using publicly available

123 resources, and found the two genes are broadly expressed throughout brain regions of

124 adult mice and humans (Figure S1) (33-37) (38). Kmt2a and Kdm5c are expressed at

125 comparable levels in all major excitatory and inhibitory neuron subtypes as well as glia

126 cells in mouse visual cortices (Figure S1A), and also throughout mouse brains (Figure

127 S1B). Consistently, developing and aging human brains express KMT2A and KDM5C at

128 high, steady levels (Figure S1C). Thus, both writer and eraser are co-expressed across

129 brain cell types, regions, and developmental stages in both humans and mice.

130

131 Generation of Kmt2a-Kdm5c double-mutant (DM) mice

132 To test genetic interaction of Kmt2a and Kdm5c, we generated Kmt2a-Kdm5c double-

133 mutant (DM) mice. Experimental mice were F1 hybrids of the two fully congenic

134 laboratory mouse strains: 129S1/SvImJ Kmt2a+/- males (39) and C57BL/6J Kdm5c+/-

135 females (31) (Figure 1B). This cross resulted in the following genotypes of male mice in

136 an identical genetic background: wildtype (WT); Kmt2a heterozygote (Kmt2a-HET:

137 Kmt2a+/-), Kdm5c hemizygous knock-out (Kdm5c-KO: Kdm5c-/y), and Kmt2a-Kdm5c

138 double-mutant (DM: Kmt2a+/-, Kdm5c-/y), thereby allowing us to perform a systematic

139 comparison between the WDSTS model (Kmt2a-HET), the MRXSCJ model (Kdm5c-

140 KO), and their composite (DM). We focus on males, because MRXSCJ predominantly

141 affects males and Kdm5c-heterozygous female mice exhibit only minor cognitive deficits

142 (32).

143

144 These mice were born at expected Mendelian ratios of 25% per genotype,

145 demonstrating the DM mice were not synthetic lethal (Figure 1C). Genotypes were

146 confirmed at RNA and DNA levels (Figure S2A-C), and protein level for KDM5C (Figure

147 S2D). Brain anatomy showed no gross deformities in any of the genotypes (Figure

148 S2E). Body weight, during the course of development and in adult, however, was

149 reduced in the two single mutants as well as in DM mice (Figure 1D, Figure S2F). Note

150 that for all the four-way comparisons in this study, we first represent the p-values from

151 one-way ANOVA tests of genotype-phenotype interaction with histograms of measured

152 values. We then report p-values from post-hoc tests of all six genotype comparisons

153 with 95% confidence intervals of group mean differences. Both Kmt2a-Het and Kdm5c-

154 KO mice showed significant body weight reduction compared to WT (Figure 1D, One

155 Way ANOVA: F(3, 55) = 10.28, p < 1.0 x 10-4, Tukey's multiple comparison test: WT vs

156 2A: p = 0.008, WT vs 5C: p = 0.008). DM body weight was significantly smaller

157 compared to WT (DM vs WT: p < 1.0 x 10-4). Thus, loss of Kdm5c and Kmt2a

158 heterozygosity both led to growth retardation, which was not corrected but rather slightly

159 exacerbated in DM mice.

160

161 Memory impairments in Kdm5c-KO were ameliorated in DM

162 We first sought to determine the effect of loss of Kmt2a and/or Kdm5c on mouse

163 behavior through a battery of behavioral tests. Learning and memory measured by two

164 independent tests, contextual fear conditioning (CFC) and novel object recognition

165 (NOR), significantly interacted with genotypes (CFC: F (3,64) = 2.83, p = 0.046). NOR:

166 F(3,64) = 3.20, p = 0.030). In accordance with previous findings (31, 32), Kdm5c-KO

167 showed significant deficits in associative fear memory in the CFC (Figure 2A, WT vs 5C:

168 p = 0.018, WT vs 2A: p = 0.789). In the NOR, where WT mice showed preference for

169 the new object, Kdm5c-KO mice tended to avoid the novel object (Figure 2B) (WT vs

170 5C: p = 0.007). Previous work reported that homozygous deletion of Kmt2a in excitatory

171 hippocampal neurons leads to impaired fear memory in the CFC (30). In our tests,

172 Kmt2a-HET mice showed no deficits in either CFC or NOR (Figure 2A-B) (CFC: p =

173 0.789; NOR; p = 0.888), indicating that Kmt2a-heterozygosity does not lead to learning

174 impairment measured in these assays. Importantly, DM mice also showed no

175 differences from WT mice (Figure 2A-B) (CFC: p = 0.246; NOR: p = 0.756), suggesting

176 that Kmt2a heterozygosity rescues memory deficits of Kdm5c-KO mice. These

177 differences in memory tasks were not attributable to differences in locomotor activity or

178 shock responsiveness, as none of these parameters showed significant differences

179 among the genotypes (Figure 2C-D).

180

181 Social behavior in the single and double mutants

182 We next examined social behavior using the three independent behavioral paradigms.

183 First, social interaction was tested with the three-chambered preference test. Kmt2a-

184 HET mice showed no differences from WT (Figure 3A, WT vs 2A: p = 0.082) in

185 accordance with previous tests in conditional Kmt2a-KO mice (30). In contrast, Kdm5c-

186 KO mice, as previously shown (31), exhibited significantly less preference for the

187 stranger mouse compared with WT animals (WT vs 5C: p= 0.002). Similar to Kdm5c-

188 KO, DM mice spent less time with stranger mice compared to WT (WT vs DM: p =

189 0.011). No difference was detected between Kdm5c-KO and DM (5C vs DM: p= 0.709).

190 These data indicate that Kmt2a heterozygosity does not alter social preference or

191 rescue the deficit of Kdm5c-KO.

192

193 In tests of social dominance (Figure 3B), both Kmt2a-HET and Kdm5c-KO mice won

194 more frequently against WTs (2A vs WT: 60.9%, p = 0.091, 5C vs WT: 68.4%, p =

195 0.008). In contrast, DM animals lost more than 80% of their bouts against WT (DM vs

196 WT: p = 1.47 x 10-5). Although DM mice were slightly smaller compared to single

197 mutants (Figure 1D), this is unlikely to drive submissive behaviors, as body mass has

198 been shown to have minimal impact on social hierarchy unless excess difference (>

199 30%) is present between animals (40-42), which is not the case in our study (Figure 1D)

200 (2A vs WT:11%, 5C vs WT: 17%, DM vs WT: 25%). These results indicate that the two

201 single mutants share heightened social dominance and the double mutation reverses

202 the social dominance.

203

204 Lastly, in the resident-intruder test, similar to other behavioral paradigms, overall effects

205 of genotype were evident in both aggressive and submissive behavior (Figure 3C,

206 aggressive behavior: F(3,61) = 4.071, p = 0.011, submissive behavior: (F(3,61) = 4.071,

207 p = 0.011). The genotype effect on aggressive behaviors inversely correlated with that

208 of submissive behavior, reinforcing the changes in specific behaviors rather than

209 general locomotor activity. Both Kmt2a-HET and Kdm5c-KO show a trend of increased

210 aggression and decreased submission when we combine frequency of all aggressive or

211 submissive behavior types compared to WT (Figure 3C). With regard to specific

212 aggressive behavior, we found significantly increased darting (Figure S3A) in both

213 Kdm5c-KO (p = 0.006) and Kmt2a-HET (p = 0.032) mice.

214

215 DM mice showed significantly reduced overall aggression compared to the two single

216 mutants (Fig. 3C, DM vs 5C: p = 0.015, DM vs 2A: p = 0.011). Reciprocally, DM mice

217 were more submissive compared to single mutants (DM vs 5C: p = 0.001, DM vs 2A: p

218 = 0.011). Comparison between DM and WT mice did not yield any significant

219 differences. However, DM mice showed a strong trend of greater submission compared

220 to WT (aggression: DM vs WT: p = 0.189; submission: DM vs WT: p = 0.054); these

221 findings are consistent with the results of the social dominance test described above.

222 The decreased aggression and increased submission of DM relative to the single

223 mutants were also observed in multiple behavior types, including mounting, chasing for

224 aggression (Figure S3A) and cowering and running away for submission (Figures S3B).

225 Thus, these results suggest that double mutations alleviate aggressive behavior of both

226 Kmt2a-HET and Kdm5c-KO mice.

227

228 Together, the behavioral studies revealed more pronounced deficits in Kdm5c-KO

229 animals compared to Kmt2a-HET mice in terms of memory and social interaction, while

230 Kmt2a-HET and Kdm5c-KO mice shared increased social dominance and aggression.

231 The consequences of double mutations varied between the tests, with clear rescue

232 effects on cognitive tasks, dominant behavior, and aggression, and no effect on social

233 interactions. No behavioral traits were exacerbated in DM. Interestingly, DM mice

234 showed opposite phenotypes in social dominance and aggression compared to single

235 mutants, which is reminiscent of the increased spine density in the DM basolateral

236 amygdala but not in the hippocampal CA1 (Figure 4). These results establish common

237 and unique behavioral deficits of Kdm5c-KO and Kmt2a-HET mice and mutual

238 suppression between the two genes in some of the behavioral traits.

239

240 Roles of KMT2A, KDM5C, and their interplay in neuronal morphology

241 Altered dendrite morphology is a hallmark of many human NDDs, as well as animal

242 models of NDDs (31, 43-46). We previously found reduced dendritic length and spine

243 density in basolateral amygdala (BLA) neurons of Kdm5c-KO adult male mice (31).

244 Assessment of dendritic morphology in Kmt2a-HET has not been reported. We first

245 performed comparative dendrite morphometry of pyramidal neurons in the BLA using

246 Golgi staining for the four genotypes (Figure 4). For Kdm5c-KO neurons, we

247 recapitulated our previous findings of reduced dendrite lengths (Figure 4A and B, one-

248 way ANOVA followed by Tukey’s multiple comparison test, WT vs 5C: p = 0.034) and

249 lower spine density (Figure 4A and C, WT vs 5C: p = 0.008). Similar to Kdm5c-KO

250 neurons, Kmt2a-HET cells exhibited a reduction in spine density (Figure 4B and C, WT

251 vs 2A: p = 0.005) but not in dendritic length (WT vs 2A: p = 0.177).

252

253 Dendrite lengths of DM did not differ significantly from WT (Figure 4B, WT vs. DM: p =

254 0.550); however, these lengths were also not different from Kdm5c-KO (Kdm5c-KO vs.

255 DM: p = 0.534), representing a weak restorative effect. In contrast, dendritic spine

256 density of DM showed a significant increase that surpassed a rescue effect: DM spine

257 density was significantly higher than those of the other three genotypes (Figure 4C). As

258 morphology of dendritic spines progressively changes during synaptogenesis and

259 development (47) (Figure S4), we also asked whether developmental subtypes of

260 dendritic spines were altered in any genotype. We did not find dramatic changes in

261 spine morphology among the four genotypes (Figure 4D), indicating selective

262 requirement of Kdm5c and Kmt2a for regulation of spine numbers.

263

264 Next we tested the impact of single and double mutations on morphology of

265 hippocampal neurons (Fig. 4E and 4F). Since the ventral hippocampus CA1 (vCA1)

266 receives inputs from BLA (48), we performed morphometry analyses of pyramidal

267 neurons in this region. Genotype had significant impact on dendritic spine density (one-

268 way ANOVA, F (3, 92) 11.51, p < 1.0 x 10-4) but not on dendritic length (F (3, 92) =

269 0.564, p=0.639). While dendritic length did not show any difference in Kdm5c-KO and

270 Kmt2a-HET neurons, spine density was concomitantly decreased in the two single

271 mutants compared to WT (WT vs 2A: p = 2.0 x 10-4, WT vs 5C: p < 1.0 x 10-4).

272 In DM vCA1, spine density showed a trend of degrease compared to WT, yet this did

273 not reach statistical significance (WT vs DM: p = 0.066). DM spine density was

274 significantly higher than Kdm5c-KO (DM vs 5C: p = 0.021) but not compared to Kmt2a-

275 HET (DM vs 2A: p=0.223). These analyses indicate reduced spine density in both

276 Kdm5c-KO and Kmt2a-HET vCA1 neurons and its partial correction in DM.

277

278 Overall, we conclude that Kmt2a-HET and Kdm5c-KO share a reduced spine density in

279 both BLA and vCA1. Double mutations led to reversal of spine phenotype in BLA and

280 partial restoration of normal spine density in vCA1, supporting mutually suppressive

281 roles of KMT2A and KDM5C in dendritic spine development.

282

283 Roles of KMT2A and KDM5C in mRNA expression

284 We previously showed Kdm5c-KO mice exhibit aberrant gene expression patterns in the

285 amygdala and frontal cortex (31), and hippocampus (32). Excitatory-neuron specific

286 conditional Kmt2a-KO mice were also characterized with altered transcriptomes in the

287 hippocampus and cortex (29, 30). However, the global gene expression of Kmt2a-HET,

288 which is akin to the WDSTS syndrome genotypes, has not been determined. To

289 compare the impact of Kmt2a-heterozygosity and Kdm5c-KO on the transcriptomes, we

290 performed mRNA-seq (49) using the amygdala and the hippocampus of adult mice with

291 four animals per genotype. We confirmed the lack or reduction of reads from Kdm5c

292 exons 11 and 12 and Kmt2a exons 8 and 9 in the corresponding mutants (Figure S2A-

293 B). The accurate microdissection of brain regions was confirmed by stronger co-

294 clustering of our data with another published mRNA-seq dataset from amygdala and

295 hippocampus tissues (50) (Figure S5A). Principal component analysis (PCA) indicated

296 that transcriptomic divergence was much larger between brain regions compared to the

297 four genotypes (Figure S6A)

298

299 In the amygdala, we identified 132 differentially expressed genes (DEGs, padj < 0.1) in

300 Kdm5c-KO (5C-DEGs) and no 2A-DEG (Figure 5A and S6). The hippocampus yielded a

301 consistently higher number of DEGs across the genotypes (344 5C-DEGs and 4 2A-

302 DEGs). The small number of 2A-DEGs is likely due to the one remaining copy of Kmt2a.

303 Given the increased social dominance (Figure 3B) and clear reduction of dendritic

304 spines (Figure 4) in Kmt2a-HET mice, we reasoned that such phenotypes might be

305 associated with mild gene misregulation, which was not detected by DEseq2. To be

306 able to analyze the gene expression in Kmt2a-HET, we set a relaxed cut off (p < 0.005)

307 and retrieved 78 and 139 genes as 2A-DEG from the amygdala and hippocampus,

308 respectively (Figure 5B and S6). Table S1 lists all DEGs found in the study.

309

310 Between the amygdala and hippocampus, 5C-DEGs overlap substantially (72 DEGs,

311 Figure S7A). The heat map analysis of DEG points to a similarity in dysregulation of not

312 only 5C-DEGs but also 2A-DEGs between the two brain regions, which was not

313 detected as identity of genes (Figure S7B). These data suggest that the two brain

314 regions share gene expression programs that require KMT2A or KDM5C for their

315 normal expression.

316

317 We first sought to compare the gene expression profiles between the single mutants.

318 The majority of 5C-DEGs were upregulated and 2A-DEGs were downregulated (Figure

319 5A and 5B, S6). This result agrees with KMT2A’s primary role as a transcriptional

320 coactivator by placing H3K4me (39, 51, 52) and KDM5C’s suppressive activity on

321 transcription by removing this mark (53), yet KDM5C can also act as a positive regulator

322 of transcription (54, 55). If KMT2A and KDM5C counteract, DEGs would be oppositely

323 misregulated between the single mutants. Indeed, we found that 28-29% of 5C-DEGs

324 show signs of upregulation in Kmt2a-HET, while 8-17% of 2A-DEGs tend to be

325 downregulated in Kdm5c-KO brain tissues (Figure S8A and S8B). While these minor

326 fractions of DEGs show opposite changes between the single mutants, substantial

327 fractions of DEGs were misregulated in the same directions (Figure S8A and S8B),

328 which might be due to indirect consequences of gene manipulations. The same-

329 direction changes reflect on the similar phenotypes observed in some of behavioral

330 tests, including social dominance, and reduced spine density in the two single mutants

331 (Figure 3 and 4).

332

333 Impact of double mutations on mRNA expression

334 If the double mutations fully restore the normal transcriptome, the DM transcriptome

335 should resemble to that of WT over single mutants. However, the DM brain tissues had

336 a similar number of DEGs as Kdm5c-KO (Figure 5A). Many DM-DEGs overlap with 5C-

337 DEGs (84 in the amygdala and 116 in the hippocampus), and 30 DM-DEGs overlap with

338 2A-DEGs in the hippocampus (Figure S8C). Thus, the DM transcriptome still retains

339 some mRNA misregulations of single mutants and does not fully return to the normal

340 state, which may not be surprising given that only one of six writer or eraser enzymes

341 was manipulated (Figure 1).

342

343 Next, we sought to test more quantitatively whether the transcriptome alterations in

344 single mutants were restored in DM to some extent. We first compared fold changes of

345 single-mutant DEGs as a group in DM. Expression of 5C-up-regulated genes was

346 significantly lower in the DM amygdala and hippocampus (Paired Wilcoxon signed-rank

347 test, p < 1.41 x 10-8, Figure 5C). Reciprocally, expression of 2A-down-regulated genes

348 was significantly higher in DM brain tissues (p < 1.26 x 10-8, Figure 5C). We then

349 examined how individual DEGs behaved in DM. The fold changes of 5C-DEGs and 2A-

350 DEGs in single mutants clearly correlated with those in DM, indicating that single-

351 mutant DEGs are similarly dysregulated in DM in general. However, the slopes of linear

352 regression were lower than 1 with statistical significance, suggesting the partial rescue

353 effect (Figure 5D). The mildest rescue effect was observed in 5C-DEGs in the DM

354 amygdala (linear regression slope=0.91±0.03), while 5C-DEGs in the hippocampus and

355 2A-DEGs-R exhibit more pronounced rescue effects (slope: 0.51~0.73). These results

356 indicate that misregulation of genes in both Kdm5c-KO or Kmt2a-HET was partially

357 corrected in the DM brain tissues.

358

359 We then sought to isolate genes that might have contributed to the rescue effects at

360 cellular and behavioral levels. As shown by the red circles in Figure 5D, some DEGs

361 exhibit stronger rescue effects than others. We selected these genes as potential

362 drivers of rescue effects (see Materials & Methods). While 10 genes in the amygdala

363 and 81 genes in the hippocampus drove rescue effects of 5C-DEGs, 36 genes in the

364 amygdala and 46 genes in the hippocampus contributed to the partial restoration of

365 normal 2A-DEGs expression in DM (Table S2). We then performed pathway analysis on

366 these rescue-driving genes using the Enrichr program (56, 57). While 2A-rescue genes

367 did not yield any statistically significant enrichment (padj < 0.05), 5C-rescued genes

368 showed strong enrichment of a mouse KEGG pathway, “neuroactive ligand-receptor

369 interaction” (padj = 5.60 x 10-3, Odds ratio: 6.01). The following 10 genes contributed to

370 the enrichment: Pomc, Adcy2, Agt, Adora2a, Mc3r, Gabre, P2ry14, Tacr1, Trh, and Lhb.

371 Except for P2ry14, all these genes have known roles in learning and memory or

372 aggression (Table 1).

373

374 Interestingly, most of these genes encode peptide hormones, their receptors, or

375 downstream signaling molecules. For example, Pro-opiomelanocortin (POMC) is a

376 peptide hormone primarily known for its roles in regulating hypothalamic functions, such

377 as feeding (58). In addition to its high expression in the hypothalamus, Pomc is also

378 expressed in the amygdala and the hippocampus (58, 59). POMC signaling in these

379 regions has been implicated in the cognitive decline of Alzheimer’s disease as well as in

380 drug-induced stress response (60, 61). Another gene in the list, Mc3r, encodes a POMC

381 receptor. Mc3r is expressed in the hypothalamus and limbic structures such as the

382 amygdala and hippocampus (62). Similar to Pomc, Mc3r’s roles have been primarily

383 studied in the hypothalamus (63), yet this gene has also been implicated in

384 hippocampal memory consolidation (64). Collectively, these results raise the possibility

385 that aberrant peptide hormone signaling and its normalization underlie phenotypic

386 outcomes of Kdm5c-KO and DM mice.

387

388 H3K4me3 landscapes in WDSTS and MRXSCJ models

389 H3K4me3 is a reaction product of KMT2A-mediated methylation (51), while a substrate

390 for KDM5C-mediated demethylation (53, 65). We sought to determine the impact of

391 KMT2A- and KDM5C-deficiencies and double mutation on the H3K4me3 landscape

392 within the amygdala. In Western blot analyses, global H3K4me1-3 levels were not

393 altered dramatically in any mutant (Figure 6A). We thus performed H3K4me3 chromatin

394 immunoprecipitation coupled with deep sequencing (ChIP-seq) to probe local changes

395 genome-wide. To assess the IP specificity, we spiked-in an array of recombinant

396 nucleosomes carrying 15 common methylations along with DNA barcodes appended to

397 the Widom601 nucleosome positioning sequence (66) (see Methods). The two

398 H3K4me3 nucleosomes dominated the Widom601-containing DNA in all IP reactions

399 with negligible signals from other methylation states such as H3K4me1 or H3K4me2

400 (Figure S9A), demonstrating a superb specificity of the ChIP.

401

402 PCA analyses of H3K4me3 distribution indicated that WT and Kmt2a-HET are close

403 each other, while Kdm5c-KO and DM cluster separately from the WT-2a-HET cluster.

404 Reminiscent of the transcriptome data (Figure 5), we did not recover any differentially

405 methylated regions (DMRs) in Kmt2a-HET, while Kdm5c-KO had 1,147

406 hypermethylated regions (Figure 6C, S9, padj < 0.05 in DEseq2, see method). In

407 Kdm5c-KO, 583 loci are hypomethylated compared to WT, and these regions are highly

408 methylated regions at baseline (Figure 6C and S9C). By relaxing the threshold, we

409 obtained weak 2A-HET DMRs (Figure 6C, 357 hypermethylated and 240

410 hypomethylated). We then examined how these single-mutant DMRs behaved in the

411 other single mutant. Compared to the transcriptome data (Figure S8A and S8B), we

412 were able to recover more DMRs that are misregulated in opposite directions between

413 the single mutants (Figure 6D). These results suggest that, as a primary substrate of the

414 two enzymes, H3K4me3 changes represent the action of KMT2A and KDM5C more

415 directly compared to mRNA expression. Comparable numbers of DMRs still show

416 same-direction H3K4me3 changes, which may again reflect indirect consequences of

417 Kmt2a heterozygosity and Kdm5c deficiency.

418

419 Rescue effect of H3K4me3 landscapes in DM

420 We next examined the impact of double mutations on H3K4me3 distributions. Notably,

421 in the PCA plot, the second replicate of DM was distant from the other two replicates

422 and clustered closer with the WT-2A group, indicating that expressivity of rescue effect

423 varies among individual animals (Figure 6B). The number of DMRs in DM are smaller

424 than those of single mutants (Figure 6C). We then assessed how single-mutant DMRs

425 behave in DM. Note that, we focus on hypomethylated loci in Kmt2a-HET and

426 hypermethylated loci in Kdm5c-KO (colored loci in Figure 6D) based on the enzymatic

427 activity of KMT2A and KDM5C. The opposite-direction 2A-DMRs showed complete

428 normalization in DM, while the same-direction 2A-DMRs were even more

429 hypomethylated in DM (Figure 6E). Likewise, the opposite-direction 5C-DMRs showed

430 more pronounced rescue effect in DM compared to same-direction 5C-DMRs (Figure

431 6E). This direction-dependent rescue effect persisted even when we omitted DM rep2

432 from the analysis (Figure S10). These results strongly support the idea that KMT2A and

433 KDM5C counteract to normalize H3K4me3 levels in specific genomic loci.

434

435 Peak annotation revealed that these genomic loci, which are misregulated in single

436 mutants and rescued in DM, tend to be outside the promoters, such as distal intergenic

437 regions, introns, and internal exons (Figure 6F). Examples of in such changes are

438 represented in Figure S11. Finally, we tested whether the rescue effects in gene

439 expression involve normalization of H3K4me3. We took all rescued 2A-DEGs

440 regardless of direction of changes in single mutants and plotted DM vs 2A fold changes

441 of H3K4me3 as a function of DM vs 2A mRNA changes (Figure 6G and see method).

442 Rescued 2A-DEGs mostly showed an increase in mRNA levels in DM (Q2 and Q4).

443 When mRNA expression increases from 2A-HET to DM, H3K4me3 levels also tend to

444 increase, which results in the larger number of genes within the upper right quadrant of

445 the plot (Q2, 47 genes, p < 1.0 x 10-4, χ2 test) compared to lower-right quadrant (Q4, 30

446 genes). Similarly, when mRNA expression decreases from 5C-KO to DM, the majority of

447 genes are accompanied by decreased H3K4me3 nearby (Q3: 45 genes vs Q1: 20

448 genes, p < 1.0 x 10-4, χ2 test). Furthermore, the magnitude of mRNA changes and

449 H3K4me3 changes show positively correlation to some extent (Pearson r = 0.16, p =

450 0.110, and r = 0.27, p = 0.020 for 2A- and 5C-DEGs, respectively). Thus, misregulation

451 of mRNA expression and its partial correction involve corresponding changes in the

452 H3K4me3 levels.

453

454

455 Discussion

456

457 The present work, to our knowledge, represents the first genetic interactions between

458 chromatin modification writer and eraser enzymes in vivo. Discovery of chromatin-

459 modifying enzymes in the past decades made it clear that virtually no DNA- or histone-

460 modifications are irreversible, and instead are subjected to dynamic regulation by writer

461 and eraser enzymes. The genes encoding these enzymes appear to have undergone

462 duplication events during evolution. Complex organisms carry a greater number of

463 genes that encode enzymes for single chromatin modification. For example, only one

464 enzyme, Set1, is responsible for H3K4 methylation in fission yeast. In flies three genes,

465 Trx, Trr, and Set1 mediate H3K4me installation, and all three genes were duplicated in

466 the mammalian genomes, which resulted in six SET-family H3K4 writers (67). A

467 plethora of work has demonstrated specialized as well as redundant roles of individual

468 histone-modifying enzymes within a family, in broad biological processes such as

469 cancer and embryonic development (68). A fundamental question remained — is there

470 any specific writer-eraser pairing in such highly duplicated gene families for a single

471 chromatin modification? Mishra et al. showed that KDM5A antagonizes KMT2A,

472 KMT2B, and KMT2D to modulate the transient site-specific DNA copy number gains in

473 immortalized human cells (69). Cao et al. found that failed differentiation of mouse

474 embryonic stem cells due to Kmt2d deletion can be rescued by Kdm1a knockdown (70).

475 These pioneering efforts identified functional interplay between the opposing enzymes

476 in vitro; however, no in vivo study has been reported. Thus, the present study

477 substantially advances our understanding of how methyl-histone writer and eraser

478 enzymes functionally interact during brain development and function.

479

480 Brain development is particularly relevant to the H3K4me dynamics, because a cohort

481 of neurodevelopmental disorders have been genetically associated with impaired

482 functions of these enzymes, as discussed earlier. Unlike previous studies using

483 chemical approaches that block multiple chromatin regulators (19-21), we demonstrated

484 that manipulation of a single enzyme, KMT2A or KDM5C, is sufficient to reverse many

485 neurological traits in both of the two single-mutant models. It may not be surprising that

486 not all traits were reversed in DM mice, because compensatory actions by remaining

487 H3K4me-regulators may potentially mediate these un-rescued traits. Our work opens a

488 new avenue for future studies to delineate the full interplay between the 13 H3K4me-

489 regulatory enzymes throughout brain development and function.

490

491 Several challenges remain, especially in linking molecular functions of KMT2A and

492 KDM5C to cellular and behavioral outcomes. First, gene expression profiles did not

493 perfectly correlate with cellular outcomes. While restorative effects on spine density

494 were stronger in DM amygdala compared to hippocampus (Figure 4), the misregulation

495 and rescue effects in the gene expression profile appear to be more pronounced in the

496 hippocampus (Figure 5). One possible explanation is the difference in cell-type

497 composition between the two brain regions, which is masked in our bulk tissue mRNA-

498 seq. In addition, “causative” gene misregulation underlying cellular and behavioral traits

499 may occur during a critical developmental time window and therefore not be sensitively

500 captured in our study using adult brains. Second, histone modifying enzymes, including

501 KDM5C, have been increasingly shown to exert their non-enzymatic function (27, 71-

502 74). Third, although it has not been reported, there might be a non-histone substrate for

503 KMT2A and KDM5C, as shown in other histone modifying enzymes (75). Finally, in a fly

504 model of KMD5C-intellctual disability (76), impaired intestinal barrier and the altered

505 microbiota appears to contribute to abnormal social behavior of the kdm5-mutant flies

506 (77). One technical aspect is that while our RNA-seq was done primarily using 4–5-

507 month-old animals, ChIP-seq was carried out using 8–14-month-old mice (Table S3).

508 Nonetheless, we found that the rescue effects in mRNA levels do involve H3K4me3

509 normalization (Figure 6G). Future studies are warranted to isolate the causal events

510 between cell autonomous and non-cell autonomous mechanisms, enzymatic and non-

511 enzymatic functions, and histone and non-histone substrates. The first step would be to

512 test the functional contribution of candidate genes, many of which encode peptide-

513 hormone signaling (Table 1), to the phenotypes in the single mutants and its rescue

514 effect in DM mice.

515

516 Increased social dominance is a novel behavioral trait we observed in both WDSTS and

517 MRXSCJ mouse models. The amygdala is well known to mediate social behaviors (78,

518 79). For example, lesions of BLA result in decreased aggression-like behavior and

519 increased social interactions (80, 81), and changes in transcriptional regulation in BLA

520 are observed after social interactions (79). The dorsal hippocampus CA1 is a direct

521 recipient region of BLA inputs (48). Decreased BLA and CA1 spine densities in Kmt2a-

522 HET and Kdm5c-KO mice inversely correlate with increased social dominance and

523 aggression (Figures 4 and 6). Together, these observations imply that decreased spine

524 density is not a cause of increased social dominance, but rather reflects compensatory

525 reduction of net synaptic strength due to increased excitation and/or a loss of inhibitory

526 control in this circuit. Thus, determining the connectivity of the amygdala with other

527 regions, including prelimbic, infralimbic, and orbitofrontal cortices (82) as well as the

528 ventral hippocampus, (78) will be critical for understanding the changes in social

529 behaviors in both WDSTS and MRXSCJ models.

530

531 It is important to note that the double mutations introduced in our mice were constitutive,

532 and therefore a lifetime of adaptation to loss of these two major chromatin regulators

533 may occur from early developmental stages. A more realistic therapeutic strategy may

534 be acute inhibition of KDM5C and KMT2A in the juvenile or mature brain. Previous work

535 characterizing mouse models with excitatory-neuron specific ablation of Kdm5c or

536 Kmt2a via CamKII-Cre found that conditional Kmt2a deletion led to clear learning

537 deficits (30), while cognitive impairments in the conditional Kdm5c-KO mice were much

538 milder than those of constitutive Kdm5c-KO mice (32). These results suggest a

539 developmental origin of phenotypes in Kdm5c-KO. Future investigations are needed to

540 address whether the effects of acute inhibition of opposing enzymes in these mouse

541 models can restore such neurodevelopmental deficits.

542

543 Materials & methods

544

545 Mouse models

546 The Kdm5c-KO allele was previously described and maintained in C57BL/6J congenic

547 background (31). Kmt2a-HET mice were generated by crossing previously described

548 Kmt2a-flox (exons 8 and 9) mice with B6.129-Gt(ROSA)26Sortm1(cre/ERT2)Tyj/J-Cre mice

549 (83). Our strategy was to use F1 hybrid for all studies as previously recommended as a

550 standard practice to eliminate deleterious homozygous mutations, which can result in

551 abnormalities in the congenic lines, e.g. deafness in C57/BL6 (84). We backcrossed

552 Kmt2a+/- mice onto the desired 129S1/SvImJ strain, by marker-assisted accelerated

553 backcrossing through Charles River Labs. Kmt2a+/- mice were bred to the N4 generation

554 at minimum, where mice were >99% congenic for 129S1/SvImJ. All experimental mice

555 were generated as F1 generation hybrids from mating between 129S1/SvImJ Kmt2a+/-

556 males and C57Bl/6 Kdm5c+/- females: WT males (Kmt2a+/+, Kdm5c+/y); Kdm5c-KO

557 males (Kmt2a+/+, Kdm5c-/y); Kmt2a-HET males (Kmt2a+/-, Kdm5c+/y); and Kdm5c-Kmt2a-

558 DM males (Kmt2a+/-, Kdm5c-/y). Genotypes were confirmed using the following primers:

559 for Kmt2a, 5’-GCCAGTCAGTCCGAAAGTAC, 5’-AGGATGTTCAAAGTGCCTGC, 5’-

560 GCTCTAGAACTAGTGGATCCC; for Kdm5c, 5’-

561 CAGGTGGCTTACTGTGACATTGATG, 5’-TGGGTTTGAGGGATACTTTAGG, 5’-

562 GGTTCTCAACACTCACATAGTG.

563

564 Western blot analysis

565 Total proteins from adult brain tissues were subjected to Western blot analysis using in-

566 house anti-KDM5C (31) and anti-GAPDH antibodies (G-9, Santa Cruz). For histone

567 proteins, nuclei were enriched from the Dounce-homogenized brain tissues using Nuclei

568 EZ prep Kit (Sigma, NUC-101). DNA were digested with micrococcal nuclease (MNase,

569 NEB) for 10 minutes at room temperature, and total nuclear proteins were extracted by

570 boiling the samples with the SDS-PAGE sample buffer. The following antibodies were

571 used for Western blot analyses: anti-H3K4me3 (Abcam, ab8580), anti-H3K4me2

572 (Thermo, #710796), anti-H3K4me1 (Abcam, ab8895), and anti-H3 C-terminus (Millipore,

573 CS204377).

574

575 Brain histology

576 Mice were subjected to transcardial perfusion according to standard procedures. Fixed

577 brains were sliced on a freeze microtome, yielding 30 µm sections that were then fixed,

578 permeabilized, blocked, and stained with DAPI. Slides were imaged on an Olympus

579 SZX16 microscope, with an Olympus U-HGLGPS fluorescence source and Q Imaging

580 Retiga 6000 camera. Images were captured using Q-Capture Pro 7 software. Data were

581 collected in a blind fashion, where samples were coded and genotypes only revealed

582 after data collection was complete.

583

584 Behavioral paradigms

585 Prior to behavioral testing, mice were acclimated to the animal colony room for one

586 week single-housing in standard cages provided with a lab diet and water ad libitum. A

587 12-hour light-dark cycle (7:00AM-7:00PM) was utilized with temperature and humidity

588 maintained at 20 ±2 ºC and >30%, respectively. The University of Michigan Committee

589 on the Use and Care of Animals approved all tests performed in this research. Five

590 tests, listed in order of testing, were performed: Novel Object Recognition (five days),

591 Context Fear Conditioning (two days), Three-Chambered Social Interaction (two days),

592 Social Dominance Tube Test (three to four days), and Resident-intruder (two to three

593 days). All testing was conducted in the morning by experimenters blind to genotype.

594 The cleaning agent used in every test between each trial was 70% ethanol. Data were

595 collected in a blind fashion, where mice were coded and genotypes were only revealed

596 after testing was complete.

597

598 Novel Object Recognition: Mice were first habituated to testing arenas (40 x 30 x 32.5

599 cm3) in three, 10-minute sessions over six consecutive days (85, 86). Twenty-four hours

600 later, mice were allowed to explore two identical objects (a jar or egg, counterbalanced

601 across animals) for two, 10-minute trials spaced three hours apart. All animals were

602 returned to the arena, tested 24 hours after the first training session, and presented with

603 one training object (“familiar” object: jar or egg) and one “novel” object (egg or jar).

604 Exploration of the objects was defined as nose-point (sniffing) within 2 cm of the object.

605 Behavior was automatically measured by Ethovision XT9 software using a Euresys

606 Picolo U4H.264No/0 camera (Noldus, Cincinnati, OH). Preference was calculated as

607 the time spent exploring novel object/total time exploring both objects. One-sample t-

608 tests against 50% (no preference) were used to establish whether animals remembered

609 the original objects.

610

611 Contextual Fear Conditioning: Context fear conditioning was assessed as previously

612 described (87). Mice were placed into a distinct context with white walls (9 ¾ × 12 ¾ × 9

613 ¾ in) and a 36 steel rod grid floor (1/8 in diameter; ¼ spaced apart) (Med-Associates,

614 St. Albans, VT) and allowed to explore for three minutes, followed by a two-second 0.8

615 mA shock, after which mice were immediately returned to their home cages in the

616 colony room. Twenty-four hours later, mice were returned to the context and freezing

617 behavior was assessed with NIR camera (VID-CAM-MONO-2A) and VideoFreeze

618 (MedAssociates, St Albans, VT). Freezing levels were compared between genotypes

619 using a between-groups analysis (one-way ANOVA) with genotype as the between-

620 subjects factor.

621

622 Three-Chambered Social Interaction: Mice were placed into a three-chambered

623 apparatus consisting of one central chamber (24 x 20 x 30 cm3) and two identical side

624 chambers (24.5 x 20 x 30 cm3) each with a containment enclosure (8 cm diameter; 18

625 cm height; grey stainless steel grid 3 mm diameter spaced 7.4 mm apart) and allowed

626 to habituate for 10 minutes. Twenty-four hours later, mice were returned to the

627 apparatus that now included a 2-3 month-old stranger male mouse (C57BL/6N) on one

628 side of the box (“stranger”), and a toy mouse approximately the same size and color as

629 the stranger mouse on the other (“toy”). Exploration of either the stranger or toy was

630 defined as nose-point (sniffing) within 2 cm of the enclosure and used as a measure of

631 social interaction (88). Behavior was automatically scored by Ethovision XT9 software

632 as described above, and social preference was defined as time exploring stranger/total

633 exploration time. Social preference was analyzed using one-sample t-tests for each

634 genotype. A repeated measures analysis was used for each aggressive (genotype x

635 aggression measures ANOVA) and submissive behavior (genotype x submissive) to

636 analyze aggressive behaviors.

637

638 Social Dominance Tube Test: Twenty-four hours prior to testing, mice were habituated

639 to a plastic clear cylindrical tube (1.5 in diameter; 50 cm length) for 10 minutes. During

640 the test, two mice of different genotypes were placed at opposite ends of the tube and

641 allowed to walk to the middle. The match concluded when the one mouse (the dominant

642 mouse) forced the other mouse (the submissive mouse) to retreat with all four paws

643 outside of the tube (a “win” for the dominant mouse) (89-91). Each mouse underwent a

644 total of three matches against three different opponents for counterbalancing. Videos

645 were recorded by Ethovision XT9 software as described above, and videos were

646 manually scored by trained experimenters blind to genotype. The number of “wins” was

647 reported as a percentage of total number of matches. Data were analyzed using an

648 Exact Binomial Test with 0.5 as the probability of success (win or loss).

649

650 Resident-intruder aggression: Resident-intruder tests were used to assess aggression.

651 Tests were performed on consecutive days, where the resident mouse was exposed to

652 an unfamiliar intruder mouse for 15 minutes (92, 93). A trial was terminated prematurely

653 if blood was drawn, if an attack lasted continuously for 30 seconds, or if an intruder

654 showed visible signs of injury after an attack. Resident mice were assessed for active

655 aggression (darting, mounting, chasing/following, tail rattling, and boxing/parrying), as

656 well as submissive behaviors (cowering, upright, running away). Intruder mice were

657 assessed for passive defense (freezing, cowering, and digging). Behavior was recorded

658 and videos scored manually by experimenters blind to genotype. Data were analyzed

659 using a between-groups analysis (one-way ANOVA) with genotype as the between-

660 subjects factor.

661

662 Neuronal Golgi staining and morphological analyses

663 Brains from adult (2-8 months) mice were dissected and incubated in a modified Golgi-

664 Cox solution for two weeks at room temperature. The remaining procedure of Golgi

665 immersion, cryosectioning, staining and coverslipping was performed as described

666 previously (94).

667

668 Morphological analyses of dendrites were carried out as described previously (94). Four

669 animals were used for each genotype, and pyramidal neurons in the basolateral

670 amygdala and dorsal hippocampus CA1 per animal were quantified: N=24 neurons for

671 WT, Kmt2a-HET and Kdm5c-KO, and N=27 neurons for DM. Quantification was done

672 using commercially available software, NeuroLucida (v10, Microbrightfield, VT), installed

673 on a Dell PC workstation that controlled a Zeiss Axioplan microscope with a CCD

674 camera (1600 x 1200 pixels) and with a motorized X, Y, and Z focus for high-resolution

675 image acquisition (100X oil immersion) and quantifications. The morphological analyses

676 included: dendritic lengths, spine counts, and spine subtype morphology. All sample

677 genotypes were blinded to the analysts throughout the course of the analysis.

678

679 The criteria for selecting candidate neurons for analysis were based on: (1) visualization

680 of a completely filled soma with no overlap of neighboring soma and completely filled

681 dendrites, (2) the tapering of most distal dendrites; and (3) the visualization of the

682 complete 3-D profile of dendritic trees using the 3-D display of the imaging software.

683

684 For quantitative analysis of spine subtypes (thin, stubby, mushroom, filopodia, and

685 branched spines), only spines orthogonal to the dendritic shaft were included in this

686 analysis, whereas spines protruding above or beneath the dendritic shaft were not

687 sampled. This principle remained consistent throughout the course of analysis.

688

689 After completion, the digital profile of neuron morphology was extrapolated and

690 transported to a multi-panel computer workstation, then quantitated using

691 NeuroExplorer program (Microbrightfield, VT), followed by statistical analysis (one- and

692 two-way ANOVAs, p < 0.05).

693

694 RNA-seq

695 Brains from adult (4.5 to 8 months) male mice were microdissected to obtain the

696 amygdala and hippocampus from Bregma ~ 4.80 mm regions. N=4 animals were used

697 per genotype. The ages of mice used for genomics study, RNA-seq and ChIP-seq, are

698 summarized in Table S3. Tissue was homogenized in Tri Reagent (Sigma). Samples

699 were subjected to total RNA isolation, and RNA was purified using RNEasy Mini Kit

700 (Qiagen). ERCC spike-in RNA was added at this stage, according to manufacturer’s

701 instructions (Life Technologies). Libraries were prepared using the NEBNext® Ultra™ II

702 Directional RNA Library Prep Kit with oligo-dT priming. Multiplexed libraries were pooled

703 in approximately equimolar ratios and purified using Agencourt RNAClean XP beads

704 (Beckman Coulter).

705

706 Libraries were sequenced on the Illumina Novaseq 6000 platform, with paired-end 150

707 base pair reads (24-35 million reads/library), according to standard procedures. Reads

708 were mapped to the mm10 mouse genome (Gencode) using STAR (v2.5.3a) (95),

709 where only uniquely mapped reads were used for downstream analyses. Duplicates

710 were removed by STAR, and a counts file was generated using FeatureCounts

711 (Subread v1.5.0) (96). BAM files were converted to bigwigs using deeptools (v3.1.3)

712 (97, 98). Differentially expressed (DE) genes were called using DESeq2 (v1.14.1) (99,

713 100). Data analyses were performed with RStudio (v1.0.136). Fold change heatmaps

714 was created using gplots heatmap.2 function.

715

716 To validate the microdissection of hippocampus and amygdala, we compared our RNA-

717 seq datasets with similar RNA-seq data from Arbeitman (50) datasets, which involved

718 the cerebellum, cortex, hippocampus, and amygdala. Briefly, count data underwent

719 variance stabilizing transformation via DEseq2 vst function and Euclidean distances of

720 the transformed values were calculated by dist command and the heatmap was

721 generated by the pheatmap function. Linear regression analysis was performed using

722 the smtr v3.4-8 slope.test function with intercept = FALSE and robust=FALSE options.

723 Rescue-driving genes were chosen as genes that satisfy two conditions using R, 1) abs

724 (log2FC (single mutant vs DM) > 2, 2) abs (log2FC (DM vs WT) < 0.7.

725

726 The RNA-seq data have been deposited in NCBI’s Gene Expression Omnibus (101).

727 Data are accessible through GEO series accession numbers, SuperSeries GSE127818

728 and SubSeries GSE127722, and can be accessed with secure token ovuhykuazryxrad.

729

730 ChIP-seq

731 Amygdala tissues were microdissected from adult (8-14.5 months) male mice. N=2

732 animals were used for WT, and N=3 animals were used for Kmt2a-HET, Kdm5c-KO,

733 and DM as biological replicates. Nuclei were isolated using a Nuclei EZ prep Kit (Sigma,

734 NUC-101), and counted after Trypan blue staining. Approximately 20,000 nuclei for

735 each replicate were subjected to MNase digestion as previously described (102). We

736 essentially followed the native ChIP-seq protocol (102) with two modifications. One was

737 to use a kit to generate sequencing libraries in one-tube reactions (NEB, E7103S).

738 Another modification was to spike-in the panel of synthetic nucleosomes carrying major

739 histone methylations (EpiCypher, SKU: 19-1001) (66). For ChIP, we used the rabbit

740 monoclonal H3K4me3 antibody (Thermo, clone #RM340).

741

742 Libraries were sequenced on the Illumina NextSeq 500 platform, with single-end 75

743 base-pair sequencing, according to standard procedures. We obtained 20-59 million

744 reads per sample. Reads were aligned to the mm10 mouse genome (Gencode) and a

745 custom genome containing the sequences from our standardized, synthetic

746 nucleosomes (EpiCypher) for normalization (103), using Bowtie allowing up to two

747 mismatches. Only uniquely mapped reads were used for analysis. The range of

748 uniquely mapped reads for input samples was 38-44 million reads. All IP replicates had

749 a mean of 9.1 million uniquely mapped reads (range: 7.4-13.9 million). The enrichment

750 of mapped synthetic spike-in nucleosomes compared to input was calculated and used

751 as a normalization coefficient for read depth of each ChIP-seq replicate (103).

752

753 Peaks were called using MACS2 software (v 2.1.0.20140616) (104) using input BAM

754 files for normalization, with filters for a q-value < 0.1 and a fold enrichment greater than

755 1. Common peak sets were obtained via DiffBind (105), and count tables for the

756 common peaks were generated with the Bedtools multicov command. We removed

757 “black-list” regions that often give aberrant signals (106). Resulting count tables were

758 piped into DEseq2 to identify DMRs incorporating the synthetic nucleosome

759 normalization into the read depth factor. We used ChIPseeker (107) to annotate

760 H3K4me3 peaks with genomic features including mm10 promoters (defined here as ±1

761 kb from annotated transcription start site [TSS]). Normalized bam files were converted

762 to bigwigs for visualization in the UCSC genome browser. Genes near peaks were

763 identified by ChIPseeker. RNA-seq and ChIP-seq data were integrated using standard

764 R commands and rescued amygdala DEGs were chosen as genes that meet two

765 criteria, 1) abs (log2FC (single mutant vs DM) > 1, 2) abs (log2FC (DM vs WT) < 0.7. All

766 scripts used in this study are available upon request.

767

768 The ChIP-seq data have been deposited in NCBI’s Gene Expression Omnibus (101).

769 Data are accessible through GEO series accession numbers, SuperSeries GSE127818

770 and SubSeries GSE127817, and can be accessed with secure token ovuhykuazryxrad.

771

772 Acknowledgements

773 We thank Dr. Ken Kwan, Mandy Lam, and Own Funk for their assistance with the RNA-

774 seq library preparation protocol and use of their microscope; Chris Gates for his

775 assistance with RNA-seq analyses; and Clara Farrehi, Jordan Rich, and Demetri

776 Tsirukis for their assistance with experiments for transcriptome analyses, global histone

777 methylation Western blots, and brain histology, respectively. We also thank Drs. Sally

778 Camper, Stephen Parker, Stephanie Bielas, and Michael-Christopher Keogh, as well as

779 the members of the Iwase and Bielas labs, for helpful discussions and critical review of

780 the data.

781

782 Conflict of interest

783 MCW is CEO of Neurodigitech, LLC. The other authors declare no conflict of interest.

784

785 Author contributions

786 CNV, NCT, and SI conceived the study and designed the experiments. BR and KMC

787 performed the mouse behavioral tests under the guidance of NCT. MCW oversaw

788 dendritic morphometry analyses. CNV, KMB, PMG, and SI analyzed RNAseq data. YAS

789 performed global H3K4me3 analyses. RSP and SI performed H3K4me3 ChIP-seq

790 analyses. YD and CEK provided key experimental recourse and made important

791 intellectual contributions. All authors contributed to the writing and editing of the

792 manuscript.

793

794 Funding

795 This work was funded by an NIH National Research Service Award T32-GM07544

796 (University of Michigan Predoctoral Genetics Training Program) from the National

797 Institute of General Medicine Sciences (to CNV), an NIH National Research Service

798 Award T32-HD079342 (University of Michigan Predoctoral Career Training in the

799 Reproductive Sciences Program) from the National Institute of Child Health and Human

800 Development (NICHD) (to CNV), University of Michigan Rackham Predoctoral Research

801 Grants (to CNV), a Michigan Institute for Clinical and Health Research fellowship

802 (Translational Research Education Certificate, supported by UL1TR000433 and

803 UL1TR002240) (to CNV), a University of Michigan Rackham Predoctoral Fellowship

804 award (to CNV), an Autism Science Foundation Predoctoral Fellowship award (to CNV),

805 an NIH National Research Service Award F31NS103377 from the National Institute of

806 Neurological Disease & Stroke (NINDS) (to RSP), NIH NINDS Awards (R01NS089896

807 and R21NS104774) (to SI), Basil O'Connor Starter Scholar Research Awards from

808 March of Dimes Foundation (to SI), and a Farrehi Family Foundation Grant (to SI).

809

810

811 Figures and Legends

812 Figure 1. The H3K4 methylopathies and generation of the Kmt2a-Kdm5c double-

813 mutant (DM) mouse. (A) Histone H3 lysine 4 (H3K4me) methyltransferases (writers)

814 and demethylases (erasers) depicted by their ability to place or remove H3K4me.

815 Reader proteins recognizing specific H3K4me substrates (arrows) are depicted below.

816 Genes are listed next to their associated neurodevelopmental disorder. KMT2A and

817 KDM5C are highlighted in purple and green, respectively. WDSTS: Weideman-Steiner

818 Syndrome; ID: intellectual disability; ASD: autism spectrum disorder, CPRF: cleft palate,

819 psychomotor retardation, and distinctive facial features; ARID: autosomal recessive ID;

820 MRXSCJ: mental retardation, X-linked, syndromic, Claes-Jensen type. (B) Mouse

821 breeding scheme crossing congenic 129S1/SvlmJ Kmt2a-heterozygous males with

822 congenic C57/Bl6 Kdm5c-heterozygous females, resulting in F1 generation mice. Only

823 males were used in this study. (C) Numbers of male offspring across 30 litters, showing

824 Mendelian ratios of expected genotypes. (D) Left panel: Body weight of adult mice > 2

825 months of age (mean ± SEM, ****p < 0.0001 in one-way ANOVA). Right panel:

826 Difference between group means of weight (mean ± 95% confidence intervals, *p<0.05,

827 **p<0.01, ***p<0.001, ****p<0.0001 in Tukey’s multiple comparison test).

828

829 Figure 2. Deficit of memory-related behavior in Kdm5c-KO and its rescue in DM.

830 (A) Contextual fear conditioning test. Left panel: Freezing levels after shock on test day

831 (mean ± SEM, *p < 0.05 in one-way ANOVA). Right panel: Difference between group

832 means of freezing (mean ± 95% confidence intervals, *p<0.05 in Least Significant

833 Difference (LSD) test). (B) Novel object recognition test. Left panel: Preference for novel

834 versus familiar object (mean ± SEM, *p < 0.05 in one-way ANOVA). Right panel:

835 Difference between group means of freeze response (mean ± 95% confidence intervals,

836 *p<0.05 in Least Significant Difference (LSD) test). (C) Response to mild foot-shock

837 (mean ± 95% confidence intervals, no statistical significance [n.s.], one-way ANOVA).

838 (D) Locomotor activity (mean ± 95% confidence intervals, no statistical significance

839 [n.s.], one-way ANOVA). N=21 WT, N=16 Kmt2a-HET, N=16 Kdm5c-KO, and N=12 DM

840 animals were used for all studies.

841

842 Figure 3. Differential impacts of double mutation in social behavior. (A) Three

843 chamber test for social interaction. Left panel: preference for stranger versus toy mouse

844 (mean ± SEM, **p < 0.01 in one-way ANOVA). Right panel: Difference between group

845 means of preference (mean ± 95% confidence intervals, *p<0.05, **p<0.01 in Least

846 Significant Difference (LSD) test). (B) Tube test for social dominance. Proportion of wins

847 in matches of each mutant versus WT. Numbers on colored bars represent total number

848 of wins for WT (grey, above) or each mutant (below) in every matchup. **p<0.01,

849 ***p<0.001, Exact binomial test. (C) Resident intruder test. Left panel: average number

850 of all aggressive behaviors (mean ± SEM, *p < 0.05 in one-way ANOVA). Right panel:

851 Difference between group means of aggressive behaviors (mean ± 95% confidence

852 intervals, *p<0.05 in Least Significant Difference (LSD) test). N=21 WT, N=16 Kmt2a-

853 HET, N=16 Kdm5c-KO, and N=12 DM animals were used for all studies.

854

855 Figure 4. Altered dendrite morphology of Kdm5c-KO and Kmt2a-HET was

856 reversed in DM animals. (A) Representative images of basolateral amygdala (BLA)

857 pyramidal neurons across all genotypes, depicting overall neuron morphology including

858 dendrite lengths and dendritic spines. Scale bars represent: 100µm (above, whole

859 neuron image), 10µm (below, spine image). (B and C) Left panel: Total dendrite lengths

860 (B) or spine density (C) (mean ± 95% *p<0.05, ****p<0.0001, one-way ANOVA). Right

861 panel: Difference between group means (mean ± 95% confidence intervals, *p<0.05,

862 **p<0.01, ***p<0.001, ****p<0.0001 in Tukey’s multiple comparison test). (D)

863 Quantification of spine morphology subtypes represented as percentage of total spines

864 counted. (E and F) Morphometry of pyramidal neurons within the dorsal hippocampus

865 CA1. At least 20 neurons from four animals per genotype were quantified for all panels.

866

867 Figure 5. The transcriptomes in the amygdala and hippocampus. (A) Number of

868 differentially expressed (DE) genes across genotypes were determined using a

869 threshold of padj-value < 0.1. or relaxed cut-off of p < 0.005 for Kmt2a-HET in (B) (see

870 also Table S1). We analyzed amygdala and hippocampal tissues from four animals for

871 each genotype. (C) Behavior of single-mutant DEGs in DM. Log2 fold change of DEGs

872 relative to WT were plotted across the three mutants. Boxplot features: box, interquartile

873 range (IQR); bold line, median; gray dots, individual genes. Associated p values result

874 from Wilcoxon signed-rank tests. (D) Identification of rescue-driving genes and

875 regression analysis. Blue fitting lines and slopes result from linear regression of log2

876 fold changes between the two genotypes. Gray shade: 95% confidence interval. P-

877 values indicate probability of the null hypothesis that the fitting line does not differ from 1

878 (108). Red circles: rescue-driving genes (see method).

879

880

881 Figure 6. Altered H3K4me3 landscapes in the amygdala and rescue effect in DM.

882 (A) Western blot of whole brain lysates showing unchanged global H3K4 methylation

883 across genotypes. Total histone H3 was detected using an antibody recognizing the C-

884 terminus of H3, and used as a control for equal loading. (B) PCA analysis of H3K4me3

885 ChIPseq replicates. We analyzed amygdala tissue from 2 to 3 animals (rep1-3) for each

886 genotype (see Methods). (C) Number of H3K4me3 DMRs in 2a-HET, 5c-KO, or DM

887 compared to WT across the genome. 2A-HET DMRs are retrieved with a relaxed

888 threshold (p < 0.05). (D) Direction of H3K4me3 changes between single mutants.

889 Genes are colored based on the direction of misregulation between the single mutants.

890 (E) Behavior of single-mutant DMRs in DM. Log2 fold change of DMRs relative to WT

891 were plotted across the three mutants. Boxplot features: box, interquartile range (IQR);

892 bold line, median; gray dots, individual genes. Associated p values result from Wilcoxon

893 signed-rank tests. (F) Genomic features of H3K4me3 peaks. (G) Involvement of

894 H3K4me3 restoration near rescued DEGs. H3K4me3 log2 fold changes (FC) are plotted

895 as function of mRNA log2 FC between DM and single mutants. Red line: linear

896 regression fitting line with 95% confidence interval (gray shade). The gene numbers fall

897 into each of the four quadrants as indicated (Q1-Q4). Results (p) of χ2 test and

898 Pearson’s correlation coefficient (r) and p-value (p) are noted.

899

900 Supplementary Figure 1. Expression of KMT2A and KDM5C. (A) Expression of

901 Kmt2a and Kdm5c, from FACS-sorted single cells of mouse visual cortex, shown in

902 reads per kilobase of transcript per million mapped reads (RPKM). Neuronal cells:

903 GABAergic (GABA), Glutamatergic (Glu). Non-neuronal cells: astrocytes (A); endothelial

904 cells (E); microglia (M), oligodendrocyte precursor cells (OPC); oligodendrocytes (O);

905 smooth muscle cells (SMC). Image credit: Broad Institute “Single Cell Portal”

906 transcriptome of adult mouse visual cortex (38). (B) Expression of Kmt2a and Kdm5c

907 mRNA from adult mouse brain, shown in log2 of raw expression value from in situ

908 hybridization. Brain regions: Isocortex, olfactory areas (OLF), hippocampal formation

909 (HPF), cortical subplate (CTXsp), striatum (STR), pallidum (PAL), thalamus (TH),

910 hypothalamus (HY), midbrain (MB), pons (P), medulla (MY), cerebellum (CB). Image

911 credit: Allen Institute, Allen Mouse Brain Atlas (2004) (34). (C) Expression of KMT2A

912 and KDM5C transcripts, from developing and adult human brains, shown in RPKM.

913 Human development and adulthood were split into the following Periods: 1-7 fetal

914 development; 8-9 birth and infancy; 10-11 childhood; 12 adolescence; and 13-15

915 adulthood. Image credit: Human Brain Transcriptome Atlas (36, 109)

916

917 Supplementary Figure 2. Basic features of mutant mice. (A-B) RNA-seq read

918 coverage of Kmt2a (A) and Kdm5c (B) genes and targeted exons (highlight) confirmed

919 the intended gene manipulations. (C) Genotyping using genomic DNA, confirming

920 presence of Kmt2a and/or Kdm5c deleted alleles (“del”) only in appropriate genotypes.

921 (D) Western blot for KDM5C protein. Stars indicate non-specific bands present in all

922 samples. GAPDH shown for equal loading. (E) Serial brain sections 30 µm thick stained

923 with DAPI to mark nuclei. Sections shown at Bregma regions 1.41, 0.49, -2.15, and -

924 2.91 mm (top to bottom). Regions highlighted: anterior forceps of the corpus callosum

925 (fmi), caudate putamen (CPu), corpus callosum (cc), lateral ventricle (LV), piriform

926 cortex (Pir), olfactory tubercle (Tu), hippocampal fields CA1 and CA2, dentate gyrus

927 (DG), anteromedial nucleus (AM), third ventricle (3V), substantia nigra pars reticularis

928 (SNR). Scale bar: 1mm. (F) Body weight tracked from birth, postnatal day 1 (P1).

929

930 Supplementary Figure 3. Individual behavior types in the resident intruder test.

931 (A) Individual aggressive behaviors (mean ± SEM, **p < 0.01 in one-way ANOVA). N.S.

932 depicts no statistical difference. Left panel: average number of all submissive behaviors

933 (mean ± SEM, *p < 0.05 in one-way ANOVA). Right panel: Difference between group

934 means of submissive behaviors (mean ± 95% confidence intervals, **p<0.01,

935 ***p<0.001 in Least Significant Difference (LSD) test). (B) Individual submissive

936 behaviors (mean ± SEM, **p < 0.01 in one-way ANOVA). N.S. depicts no statistical

937 difference. N=21 WT, N=16 Kmt2a-HET, N=16 Kdm5c-KO, and N=12 DM animals were

938 used for all studies. Differences between group means all aggressive (A) and

939 submissive (B) behaviors (mean ± 95% confidence intervals, *p<0.05, **p<0.01 in Least

940 Significant Difference (LSD) test).

941

942 Supplementary Figure 4. Schematic of dendrite spine subtype analysis. (A)

943 Projection image of a dendritic segment from a series of Z stack images derived from

944 BLA pyramidal cells. Numerical marks adjacent to corresponding spine subtypes,

945 represented as: 1. Filopodia, 2. Thin, 3, Stubby, 4. Mushroom, and 5. Branched.

946

947 Supplementary Figure 5. Validation of brain microdissection. (A) To validate

948 accuracy of our brain microdissection, we compared our RNAseq data of the

949 hippocampus (HIP) and amygdala (AMY) with those of the Arbeitman paper (50), which

950 involved the hippocampus (Hippo), amygdala (Amy), cerebellum (Cer), cerebral cortex

951 (Ctx). Euclidian distance between all combination of RNA-seq data are plotted (see

952 method). Samples from this study and the Arbeitman study are clustered separately,

953 likely due to difference in experimental procedures and/or sex of mice. Our study used

954 adult male mice, while the Arbeitman study used adult females. Nonetheless, our

955 amygdala samples showed the shortest distance to the Arbeitman amygdala data

956 compared to any other brain tissues (red rectangles). Likewise, our hippocampus data

957 are closest to the Arbeitman hippocampus data among the four brain regions. The data

958 demonstrate that the microdissection of brain regions in the two studies are consistent

959 with each other.

960

961 Supplementary Figure 6. Basic analysis of RNA-seq data. (A) PCA analysis of RNA-

962 seq libraries. Tissue types are a stronger segregating factor than genotypes. (B)

963 Volcano plot representation of differential expressed genes (DEGs) in mutant vs WT

964 comparisons. Red dots: DEGs with padj < 0.1 (see Table S1). (C) Mildly dysregulated

965 genes in Kmt2a-HET were recovered with a relaxed threshold.

966

967 Supplementary Figure 7. Similarity of gene misregulation between the amygdala

968 and hippocampus. (A) Overlap of single mutant DEGs between the amygdala (AMY)

969 and hippocampus (HIP). 5C-DEGs overlap between the two brain regions. (B) Heatmap

970 representation of log2 fold changes of 2A-DEGs (left) and 5C-DEGs (right). Overall the

971 patterns of gene misregulation are similar in the two brain regions.

972

973 Supplementary Figure 8. Direction of gene misregulation between Kmt2a-HET and

974 Kdm5c-KO and their overlap in DM brain tissues. (A) Behavior of single-mutant

975 DEGs in the other single mutant. Log2 fold changes of DEGs found in a single mutant

976 were plotted as a function of log2 fold changes in the other single mutant. Red shade

977 covers genes that are regulated in the opposite direction between Kmt2a-HET and

978 Kdm5c-KO brain tissues. Blue shade covers deregulated genes in the same direction.

979 (B) The larger number of genes is dysregulated in the same direction between the two

980 single mutants. (C) Overlap of DEGs between mutants. 5C- and DM-DEGs overlap

981 substantially.

982

983 Supplementary Figure 9. Basic characterization of H3K4me3 ChIP-seq data. (A)

984 Validation of H3K4me3 ChIP-seq specificity. Barcode reads originating from spike-in

985 nucleosomes were counted. The two synthetic nucleosomes with the H3K4me3

986 barcodes dominated all ChIP samples, with H3K4me1/2 nucleosomes rarely detected.

987 (B) Volcano plot represents the statistical significance of differentially methylated

988 regions (DMRs). (C) MA plots of DMRs revealed signal intensity-dependent

989 misregulation of H3K4me3 in Kdm5c-KO, where hypomethylated regions in the mutant

990 were highly methylated in WT.

991

992 Supplementary Figure 10. Rescue effect without DM rep2. (A) DM rep2 showed

993 strong rescue effect (Figure 6B). To test if H3K4me3 misregulations were alleviated in

994 other DM replicates, we removed DM rep2 and then examined the behavior of single-

995 mutant DMRs in DM. Log2 fold change of DMRs relative to WT were plotted across the

996 three mutants. Boxplot features: box, interquartile range (IQR); bold line, median; gray

997 dots, individual genes. Associated p values result from Wilcoxon signed-rank tests.

998 Rescue effects were still evident in DM and also dependent on the direction of

999 misregulation between the single mutants.

1000

1001 Supplementary Figure 11. Representative loci found in the H3K4me3 ChIP-seq

1002 analysis. (A) Representative genome browser view of two representative loci for each

1003 of the major genome areas: rescued intergenic DMRs (A), rescued promoter DMRs (B),

1004 un-rescued DMRs (C).

Gene Encoded protein Chromosome Description symbol Tachykinin receptor Tachykinin receptor (aka neurokinin-1 or Substance-P receptor). G-protein coupled receptors found across the brain (110). Elevated Tacr1 chr6 1 TACR1 mRNA levels in ASD patients with heightened aggression (111). Decreased aggression in Tacr1-KO mice (112). Beta subunit of Major roles of LH in testosterone production (113). Additional roles of LH in learning and memory functions, e.g. elevated LH impairs Lhb luteinizing hormone chr7 spatial memory of rodents (114) and is associated with poor memory recall in men (115). (LH) Thyrotropin- Primary TRH production in hypothalamic neurons. Some production and activity within the hippocampus (116). TRH stimulates GABA Trh releasing chr6 release within CA1 of the hippocampus (117). TRH increases during spatial learning tasks and is correlated with performance of rats hormone (TRH) (118). The epsilon subunit of GABA-A Receptor, an inhibitory chloride channel. Agonist-independent activity produces a tonic inhibitory tone The epsilon subunit Gabre chrX onto neurons (119). Highest in the hypothalamus (120). of the GABA-A Expressed at low levels throughout the brain and is stimulated by G-protein signaling and PKC (121). High expression in the soma and Adcy2 Adenylyl cyclase 2 chr13 dendrites of hippocampal neurons and a proposed role in coincidence detection (122).

A G-protein coupled receptor of extracellular adenosine (123). A regulator of neurogenesis and hippocampal volume. Contribution of Adenosine A2a ADORA2A SNPs in Alzheimer’s disease progression (124). Modulatory roles of BDNF expression and GABA and glutamate signaling in the Adora2a chr10 receptor hippocampus (125). Increased body mass, anxiety, and heightened aggression in Adora2a-KO mice (123).

A peptide hormone regulating the hypothalamic function such as feeding (58). High expression in the hypothalamus, and lower Pro- expression in the amygdala and hippocampus (58, 59). There, proposed roles in responding to stress (60). Impaired POMC expression in Pomc chr12 opiomelanocortin CA3 of the hippocampus is associated with cognitive decline in Alzheimer’s disease (61).

MC3R is receptor for cleavage products of POMC (64). Expression in the hypothalamus and limbic structures such as the amygdala and Melanocortin 3 Mc3r chr2 hippocampus (62). The hypothalamic function in weight and feeding (63). Roles in hippocampal memory consolidation (64). receptor

A peptide hormone primarily produced by astrocytes and converted into Angiotensin, neuroactive peptides (126). Well-known roles in Agt Angiotensinogen chr8 vascular regulation, such as blood pressure (127). Angiotensin II inhibits LTP in the dentate gyrus in the hippocampus (128). Roles in mood and cognition (129). Purinergic receptor G-protein coupled receptor that binds UDP-glucose (130). Highly expressed in the placenta but moderately expressed in the brain (131). P2ry14 P2Y, G-protein chr3 Upregulated in response to immune challenge and typically found within glia (130). coupled 005

006 Table 1. Rescued 5C-DEGs that drove enrichment of the KEGG pathway “Neuroactive ligand-receptor interaction”

1007 References

1008 1. Faundes V, Newman WG, Bernardini L, Canham N, Clayton-Smith J, Dallapiccola B, et 1009 al. Histone Lysine Methylases and Demethylases in the Landscape of Human 1010 Developmental Disorders. Am J Hum Genet. 2018;102(1):175-87. 1011 2. De Rubeis S, He X, Goldberg AP, Poultney CS, Samocha K, Ercument Cicek A, et al. 1012 Synaptic, transcriptional and chromatin genes disrupted in autism. Nature. 1013 2014;515(7526):209-15. 1014 3. Iossifov I, O'Roak BJ, Sanders SJ, Ronemus M, Krumm N, Levy D, et al. The 1015 contribution of de novo coding mutations to autism spectrum disorder. Nature. 1016 2014;515(7526):216-21. 1017 4. Kouzarides T. Chromatin modifications and their function. Cell. 2007;128(4):693- 1018 705. 1019 5. Taverna SD, Li H, Ruthenburg AJ, Allis CD, and Patel DJ. How chromatin-binding 1020 modules interpret histone modifications: lessons from professional pocket pickers. 1021 Nat Struct Mol Biol. 2007;14(11):1025-40. 1022 6. Allis CD, Berger SL, Cote J, Dent S, Jenuwien T, Kouzarides T, et al. New 1023 nomenclature for chromatin-modifying enzymes. Cell. 2007;131(4):633-6. 1024 7. Iwase S, Berube NG, Zhou Z, Kasri NN, Battaglioli E, Scandaglia M, et al. Epigenetic 1025 Etiology of Intellectual Disability. The Journal of neuroscience : the official journal of 1026 the Society for Neuroscience. 2017;37(45):10773-82. 1027 8. Ronan JL, Wu W, and Crabtree GR. From neural development to cognition: 1028 unexpected roles for chromatin. Nat Rev Genet. 2013;14(5):347-59. 1029 9. Iwase S, and Shi Y. Histone and DNA modifications in mental retardation. Progress in 1030 drug research Fortschritte der Arzneimittelforschung Progres des recherches 1031 pharmaceutiques. 2011;67:147-73. 1032 10. Barski A, Cuddapah S, Cui K, Roh TY, Schones DE, Wang Z, et al. High-resolution 1033 profiling of histone methylations in the human genome. Cell. 2007;129(4):823-37. 1034 11. Heintzman ND, Stuart RK, Hon G, Fu Y, Ching CW, Hawkins RD, et al. Distinct and 1035 predictive chromatin signatures of transcriptional promoters and enhancers in the 1036 human genome. Nat Genet. 2007;39(3):311-8. 1037 12. Vermeulen M, Mulder KW, Denissov S, Pijnappel WW, van Schaik FM, Varier RA, et 1038 al. Selective anchoring of TFIID to nucleosomes by trimethylation of histone H3 1039 lysine 4. Cell. 2007;131(1):58-69. 1040 13. Lauberth SM, Nakayama T, Wu X, Ferris AL, Tang Z, Hughes SH, et al. H3K4me3 1041 Interactions with TAF3 Regulate Preinitiation Complex Assembly and Selective Gene 1042 Activation. Cell. 2013;152(5):1021-36. 1043 14. Local A, Huang H, Albuquerque CP, Singh N, Lee AY, Wang W, et al. Identification of 1044 H3K4me1-associated proteins at mammalian enhancers. Nat Genet. 2018;50(1):73- 1045 82. 1046 15. Cheung I, Shulha HP, Jiang Y, Matevossian A, Wang J, Weng Z, et al. Developmental 1047 regulation and individual differences of neuronal H3K4me3 epigenomes in the 1048 prefrontal cortex. Proceedings of the National Academy of Sciences of the United 1049 States of America. 2010;107(19):8824-9.

1050 16. Shulha HP, Cheung I, Whittle C, Wang J, Virgil D, Lin CL, et al. Epigenetic signatures 1051 of autism: trimethylated H3K4 landscapes in prefrontal neurons. Arch Gen 1052 Psychiatry. 2012;69(3):314-24. 1053 17. Shulha HP, Crisci JL, Reshetov D, Tushir JS, Cheung I, Bharadwaj R, et al. Human- 1054 specific histone methylation signatures at transcription start sites in prefrontal 1055 neurons. PLoS biology. 2012;10(11):e1001427. 1056 18. Vallianatos CN, and Iwase S. Disrupted intricacy of histone H3K4 methylation in 1057 neurodevelopmental disorders. Epigenomics. 2015;7(3):503-19. 1058 19. Alarcon JM, Malleret G, Touzani K, Vronskaya S, Ishii S, Kandel ER, et al. Chromatin 1059 acetylation, memory, and LTP are impaired in CBP+/- mice: a model for the 1060 cognitive deficit in Rubinstein-Taybi syndrome and its amelioration. Neuron. 1061 2004;42(6):947-59. 1062 20. Bjornsson HT, Benjamin JS, Zhang L, Weissman J, Gerber EE, Chen YC, et al. Histone 1063 deacetylase inhibition rescues structural and functional brain deficits in a mouse 1064 model of Kabuki syndrome. Science translational medicine. 2014;6(256):256ra135. 1065 21. Park J, Thomas S, and Munster PN. Epigenetic modulation with histone deacetylase 1066 inhibitors in combination with immunotherapy. Epigenomics. 2015;7(4):641-52. 1067 22. Strom SP, Lozano R, Lee H, Dorrani N, Mann J, O'Lague PF, et al. De Novo variants in 1068 the KMT2A (MLL) gene causing atypical Wiedemann-Steiner syndrome in two 1069 unrelated individuals identified by clinical exome sequencing. BMC Med Genet. 1070 2014;15:49. 1071 23. Jones WD, Dafou D, McEntagart M, Woollard WJ, Elmslie FV, Holder-Espinasse M, et 1072 al. De Novo Mutations in MLL Cause Wiedemann-Steiner Syndrome. American 1073 Journal of Human Genetics. 2012;91(2):358-64. 1074 24. Claes S, Devriendt K, Van Goethem G, Roelen L, Meireleire J, Raeymaekers P, et al. 1075 Novel syndromic form of X-linked complicated spastic paraplegia. Am J Med Genet. 1076 2000;94(1):1-4. 1077 25. Jensen LR, Amende M, Gurok U, Moser B, Gimmel V, Tzschach A, et al. Mutations in 1078 the JARID1C gene, which is involved in transcriptional regulation and chromatin 1079 remodeling, cause X-linked mental retardation. Am J Hum Genet. 2005;76(2):227-36. 1080 26. Adegbola A, Gao H, Sommer S, and Browning M. A novel mutation in JARID1C/SMCX 1081 in a patient with autism spectrum disorder (ASD). American journal of medical 1082 genetics Part A. 2008;146A(4):505-11. 1083 27. Vallianatos CN, Farrehi C, Friez MJ, Burmeister M, Keegan CE, and Iwase S. Altered 1084 Gene-Regulatory Function of KDM5C by a Novel Mutation Associated With Autism 1085 and Intellectual Disability. Front Mol Neurosci. 2018;11:104. 1086 28. Gupta S, Kim SY, Artis S, Molfese DL, Schumacher A, Sweatt JD, et al. Histone 1087 methylation regulates memory formation. The Journal of neuroscience : the official 1088 journal of the Society for Neuroscience. 2010;30(10):3589-99. 1089 29. Jakovcevski M, Ruan H, Shen EY, Dincer A, Javidfar B, Ma Q, et al. Neuronal 1090 Kmt2a/Mll1 histone methyltransferase is essential for prefrontal synaptic plasticity 1091 and working memory. The Journal of neuroscience : the official journal of the Society 1092 for Neuroscience. 2015;35(13):5097-108. 1093 30. Kerimoglu C, Sakib MS, Jain G, Benito E, Burkhardt S, Capece V, et al. KMT2A and 1094 KMT2B Mediate Memory Function by Affecting Distinct Genomic Regions. Cell 1095 reports. 2017;20(3):538-48. 49

1096 31. Iwase S, Brookes E, Agarwal S, Badeaux AI, Ito H, Vallianatos CN, et al. A Mouse 1097 Model of X-linked Intellectual Disability Associated with Impaired Removal of 1098 Histone Methylation. Cell reports. 2016. 1099 32. Scandaglia M, Lopez-Atalaya JP, Medrano-Fernandez A, Lopez-Cascales MT, Del 1100 Blanco B, Lipinski M, et al. Loss of Kdm5c Causes Spurious Transcription and 1101 Prevents the Fine-Tuning of Activity-Regulated Enhancers in Neurons. Cell reports. 1102 2017;21(1):47-59. 1103 33. Hawrylycz MJ, Lein ES, Guillozet-Bongaarts AL, Shen EH, Ng L, Miller JA, et al. An 1104 anatomically comprehensive atlas of the adult human brain transcriptome. Nature. 1105 2012;489(7416):391-9. 1106 34. Lein ES, Hawrylycz MJ, Ao N, Ayres M, Bensinger A, Bernard A, et al. Genome-wide 1107 atlas of gene expression in the adult mouse brain. Nature. 2007;445(7124):168-76. 1108 35. Zhang XQ, Wang ZL, Poon MW, and Yang JH. Spatial-temporal transcriptional 1109 dynamics of long non-coding RNAs in human brain. Hum Mol Genet. 1110 2017;26(16):3202-11. 1111 36. Pletikos M, Sousa AM, Sedmak G, Meyer KA, Zhu Y, Cheng F, et al. Temporal 1112 specification and bilaterality of human neocortical topographic gene expression. 1113 Neuron. 2014;81(2):321-32. 1114 37. Johnson MB, Kawasawa YI, Mason CE, Krsnik Z, Coppola G, Bogdanovic D, et al. 1115 Functional and evolutionary insights into human brain development through global 1116 transcriptome analysis. Neuron. 2009;62(4):494-509. 1117 38. Tasic B, Menon V, Nguyen TN, Kim TK, Jarsky T, Yao Z, et al. Adult mouse cortical cell 1118 taxonomy revealed by single cell transcriptomics. Nature neuroscience. 1119 2016;19(2):335-46. 1120 39. Cao F, Townsend EC, Karatas H, Xu J, Li L, Lee S, et al. Targeting MLL1 H3K4 1121 methyltransferase activity in mixed-lineage leukemia. Mol Cell. 2014;53(2):247-61. 1122 40. Wang F, Kessels HW, and Hu H. The mouse that roared: neural mechanisms of social 1123 hierarchy. Trends Neurosci. 2014;37(11):674-82. 1124 41. Kim B, Colon E, Chawla S, Vandenberg LN, and Suvorov A. Endocrine disruptors alter 1125 social behaviors and indirectly influence social hierarchies via changes in body 1126 weight. Environ Health. 2015;14:64. 1127 42. Varholick JA, Bailoo JD, Palme R, and Würbel H. Phenotypic variability between 1128 Social Dominance Ranks in laboratory mice. Sci Rep. 2018;8(1):6593. 1129 43. Penzes P, Cahill ME, Jones KA, VanLeeuwen JE, and Woolfrey KM. Dendritic spine 1130 pathology in neuropsychiatric disorders. Nat Neurosci. 2011;14(3):285-93. 1131 44. Dierssen M, and Ramakers GJ. Dendritic pathology in mental retardation: from 1132 molecular genetics to neurobiology. Genes Brain Behav. 2006;5 Suppl 2:48-60. 1133 45. Xu X, Miller EC, and Pozzo-Miller L. Dendritic spine dysgenesis in Rett syndrome. 1134 Front Neuroanat. 2014;8:97. 1135 46. Irwin SA, Galvez R, and Greenough WT. Dendritic spine structural anomalies in 1136 fragile-X mental retardation syndrome. Cereb Cortex. 2000;10(10):1038-44. 1137 47. Arikkath J. Molecular mechanisms of dendrite morphogenesis. Front Cell Neurosci. 1138 2012;6:61. 1139 48. Yang Y, and Wang JZ. From Structure to Behavior in Basolateral Amygdala- 1140 Hippocampus Circuits. Front Neural Circuits. 2017;11:86.

1141 49. Jaworski E, and Routh A. ClickSeq: Replacing Fragmentation and Enzymatic Ligation 1142 with Click-Chemistry to Prevent Sequence Chimeras. Methods Mol Biol. 1143 2018;1712:71-85. 1144 50. Arbeitman MN. Maternal Experience Leads to Lasting Gene Expression Changes in 1145 Some Regions of the Mouse Brain. G3 (Bethesda). 2019;9(8):2623-8. 1146 51. Dou YL, Milne TA, Ruthenburg AJ, Lee S, Lee JW, Verdine GL, et al. Regulation of 1147 MLL1 H3K4 methyltransferase activity by its core components. Nature Structural & 1148 Molecular Biology. 2006;13(8):713-9. 1149 52. Zhang H, Gayen S, Xiong J, Zhou B, Shanmugam AK, Sun Y, et al. MLL1 Inhibition 1150 Reprograms Epiblast Stem Cells to Naive Pluripotency. Cell stem cell. 1151 2016;18(4):481-94. 1152 53. Iwase S, Lan F, Bayliss P, de la Torre-Ubieta L, Huarte M, Qi HH, et al. The X-linked 1153 mental retardation gene SMCX/JARID1C defines a family of histone H3 lysine 4 1154 demethylases. Cell. 2007;128(6):1077-88. 1155 54. Outchkourov NS, Muino JM, Kaufmann K, van Ijcken WF, Groot Koerkamp MJ, van 1156 Leenen D, et al. Balancing of histone H3K4 methylation states by the Kdm5c/SMCX 1157 histone demethylase modulates promoter and enhancer function. Cell reports. 1158 2013;3(4):1071-9. 1159 55. S G, E M, Y M-N, CN V, RS P, PM G, et al. biorxiv: Cold Spring Harbor Laboratory; 1160 2017. 1161 56. Chen EY, Tan CM, Kou Y, Duan Q, Wang Z, Meirelles GV, et al. Enrichr: interactive and 1162 collaborative HTML5 gene list enrichment analysis tool. BMC bioinformatics. 1163 2013;14:128. 1164 57. Kuleshov MV, Jones MR, Rouillard AD, Fernandez NF, Duan Q, Wang Z, et al. Enrichr: 1165 a comprehensive gene set enrichment analysis web server 2016 update. Nucleic 1166 acids research. 2016;44(W1):W90-7. 1167 58. Civelli O, Birnberg N, and Herbert E. Detection and quantitation of pro- 1168 opiomelanocortin mRNA in pituitary and brain tissues from different species. The 1169 Journal of biological chemistry. 1982;257(12):6783-7. 1170 59. 1171 olanzapine administration on POMC and neuropeptide Y mRNA levels in the male 1172 ratPałasz amygdala A, Bandyszewska and hippocampus. M, Rojczyk Pharmacol E, and Wiaderkiewicz Rep. 2016;68(2):292 R. Effect-6. of extended 1173 60. Niikura K, Zhou Y, Ho A, and Kreek MJ. Proopiomelanocortin (POMC) expression and 1174 conditioned place aversion during protracted withdrawal from chronic intermittent 1175 escalating-dose heroin in POMC-EGFP promoter transgenic mice. Neuroscience. 1176 2013;236:220-32. 1177 61. Shen Y, Tian M, Zheng Y, Gong F, Fu AKY, and Ip NY. Stimulation of the Hippocampal 1178 POMC/MC4R Circuit Alleviates Synaptic Plasticity Impairment in an Alzheimer's 1179 Disease Model. Cell reports. 2016;17(7):1819-31. 1180 62. Roselli-Rehfuss L, Mountjoy KG, Robbins LS, Mortrud MT, Low MJ, Tatro JB, et al. 1181 Identification of a receptor for gamma melanotropin and other 1182 proopiomelanocortin peptides in the hypothalamus and limbic system. Proceedings 1183 of the National Academy of Sciences of the United States of America. 1184 1993;90(19):8856-60.

1185 63. Pei H, Patterson CM, Sutton AK, Burnett KH, Myers MG, and Olson DP. Lateral 1186 Hypothalamic Mc3R-Expressing Neurons Modulate Locomotor Activity, Energy 1187 Expenditure, and Adiposity in Male Mice. Endocrinology. 2019;160(2):343-58. 1188 64. -Melanocyte- 1189 -MSH) reverses impairment of memory reconsolidation 1190 inducedMachado by I, Gonzálezinterleukin P,- Schiöth1 beta (IL HB,-1 Lasagabeta) hippocampal M, and Scimonelli infusions. TN. αPeptides. 1191 2010;31(11):2141stimulating hormone-4. (α 1192 65. Tahiliani M, Mei PC, Fang R, Leonor T, Rutenberg M, Shimizu F, et al. The histone 1193 H3K4 demethylase SMCX links REST target genes to X-linked mental retardation. 1194 Nature. 2007;447(7144):601-+. 1195 66. Shah RN, Grzybowski AT, Cornett EM, Johnstone AL, Dickson BM, Boone BA, et al. 1196 Examining the Roles of H3K4 Methylation States with Systematically Characterized 1197 Antibodies. Molecular cell. 2018;72(1):162-77 e7. 1198 67. Eissenberg JC, and Shilatifard A. Histone H3 lysine 4 (H3K4) methylation in 1199 development and differentiation. Dev Biol. 2010;339(2):240-9. 1200 68. Shilatifard A. The COMPASS family of histone H3K4 methylases: mechanisms of 1201 regulation in development and disease pathogenesis. Annual review of biochemistry. 1202 2012;81:65-95. 1203 69. Mishra S, Van Rechem C, Pal S, Clarke TL, Chakraborty D, Mahan SD, et al. Cross-talk 1204 between Lysine-Modifying Enzymes Controls Site-Specific DNA Amplifications. Cell. 1205 2018;175(6):1716. 1206 70. Cao K, Collings CK, Morgan MA, Marshall SA, Rendleman EJ, Ozark PA, et al. An 1207 Mll4/COMPASS-Lsd1 epigenetic axis governs enhancer function and pluripotency 1208 transition in embryonic stem cells. Sci Adv. 2018;4(1):eaap8747. 1209 71. Liu X, and Secombe J. The Histone Demethylase KDM5 Activates Gene Expression by 1210 Recognizing Chromatin Context through Its PHD Reader Motif. Cell reports. 1211 2015;13(10):2219-31. 1212 72. Dimitrova E, Turberfield AH, and Klose RJ. Histone demethylases in chromatin 1213 biology and beyond. EMBO Rep. 2015;16(12):1620-39. 1214 73. Dorighi KM, Swigut T, Henriques T, Bhanu NV, Scruggs BS, Nady N, et al. Mll3 and 1215 Mll4 Facilitate Enhancer RNA Synthesis and Transcription from Promoters 1216 Independently of H3K4 Monomethylation. Molecular cell. 2017;66(4):568-76 e4. 1217 74. Aubert Y, Egolf S, and Capell BC. The Unexpected Noncatalytic Roles of Histone 1218 Modifiers in Development and Disease. Trends Genet. 2019;35(9):645-57. 1219 75. Zhang X, Wen H, and Shi X. Lysine methylation: beyond histones. Acta Biochim 1220 Biophys Sin (Shanghai). 2012;44(1):14-27. 1221 76. Zamurrad S, Hatch HAM, Drelon C, Belalcazar HM, and Secombe J. A Drosophila 1222 Model of Intellectual Disability Caused by Mutations in the Histone Demethylase 1223 KDM5. Cell reports. 2018;22(9):2359-69. 1224 77. Chen K, Luan X, Liu Q, Wang J, Chang X, Snijders AM, et al. Drosophila Histone 1225 Demethylase KDM5 Regulates Social Behavior through Immune Control and Gut 1226 Microbiota Maintenance. Cell Host Microbe. 2019;25(4):537-52 e8. 1227 78. Felix-Ortiz AC, and Tye KM. Amygdala inputs to the ventral hippocampus 1228 bidirectionally modulate social behavior. J Neurosci. 2014;34(2):586-95. 1229 79. McCann KE, Sinkiewicz DM, Rosenhauer AM, Beach LQ, and Huhman KL. 1230 Transcriptomic Analysis Reveals Sex-Dependent Expression Patterns in the

1231 Basolateral Amygdala of Dominant and Subordinate Animals After Acute Social 1232 Conflict. Mol Neurobiol. 2018. 1233 80. Levinson DM, Reeves DL, and Buchanan DR. Reductions in aggression and 1234 dominance status in guinea pigs following bilateral lesions in the basolateral 1235 amygdala or lateral septum. Physiol Behav. 1980;25(6):963-71. 1236 81. McGregor A, and Herbert J. Differential effects of excitotoxic basolateral and 1237 corticomedial lesions of the amygdala on the behavioural and endocrine responses 1238 to either sexual or aggression-promoting stimuli in the male rat. Brain Res. 1239 1992;574(1-2):9-20. 1240 82. Felix-Ortiz AC, Burgos-Robles A, Bhagat ND, Leppla CA, and Tye KM. Bidirectional 1241 modulation of anxiety-related and social behaviors by amygdala projections to the 1242 medial prefrontal cortex. Neuroscience. 2016;321:197-209. 1243 83. McMahon KA, Hiew SY, Hadjur S, Veiga-Fernandes H, Menzel U, Price AJ, et al. Mll 1244 has a critical role in fetal and adult hematopoietic stem cell self-renewal. Cell stem 1245 cell. 2007;1(3):338-45. 1246 84. Silva AJ, Simpson EM, Takahashi JS, Lipp HP, Nakanishi S, Wehner JM, et al. Mutant 1247 mice and neuroscience: recommendations concerning genetic background. Banbury 1248 Conference on genetic background in mice. Neuron. 1997;19(4):755-9. 1249 85. Vogel-Ciernia A, and Wood MA. Examining object location and object recognition 1250 memory in mice. Curr Protoc Neurosci. 2014;69:8.31.1-17. 1251 86. Tchessalova D, and Tronson NC. Memory deficits in males and females long after 1252 subchronic immune challenge. Neurobiol Learn Mem. 2019. 1253 87. Keiser AA, Turnbull LM, Darian MA, Feldman DE, Song I, and Tronson NC. Sex 1254 Differences in Context Fear Generalization and Recruitment of Hippocampus and 1255 Amygdala during Retrieval. Neuropsychopharmacology : official publication of the 1256 American College of Neuropsychopharmacology. 2017;42(2):397-407. 1257 88. Crawley JN. Mouse behavioral assays relevant to the symptoms of autism. Brain 1258 Pathol. 2007;17(4):448-59. 1259 89. Zhou T, Zhu H, Fan Z, Wang F, Chen Y, Liang H, et al. History of winning remodels 1260 thalamo-PFC circuit to reinforce social dominance. Science. 2017;357(6347):162-8. 1261 90. Larrieu T, Cherix A, Duque A, Rodrigues J, Lei H, Gruetter R, et al. Hierarchical Status 1262 Predicts Behavioral Vulnerability and Nucleus Accumbens Metabolic Profile 1263 Following Chronic Social Defeat Stress. Curr Biol. 2017;27(14):2202-10.e4. 1264 91. Moretti P, Bouwknecht JA, Teague R, Paylor R, and Zoghbi HY. Abnormalities of 1265 social interactions and home-cage behavior in a mouse model of Rett syndrome. 1266 Hum Mol Genet. 2005;14(2):205-20. 1267 92. Guzmán YF, Tronson NC, Jovasevic V, Sato K, Guedea AL, Mizukami H, et al. Fear- 1268 enhancing effects of septal oxytocin receptors. Nat Neurosci. 2013;16(9):1185-7. 1269 93. Miczek KA, Maxson SC, Fish EW, and Faccidomo S. Aggressive behavioral 1270 phenotypes in mice. Behav Brain Res. 2001;125(1-2):167-81. 1271 94. Shmelkov SV, Hormigo A, Jing D, Proenca CC, Bath KG, Milde T, et al. Slitrk5 1272 deficiency impairs corticostriatal circuitry and leads to obsessive-compulsive-like 1273 behaviors in mice. Nature medicine. 2010;16(5):598-602, 1p following 1274 95. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast 1275 universal RNA-seq aligner. Bioinformatics. 2013;29(1):15-21.

1276 96. Liao Y, Smyth GK, and Shi W. featureCounts: an efficient general purpose program 1277 for assigning sequence reads to genomic features. Bioinformatics. 2014;30(7):923- 1278 30. 1279 97. Ramírez F, Dündar F, Diehl S, Grüning BA, and Manke T. deepTools: a flexible 1280 platform for exploring deep-sequencing data. Nucleic Acids Res. 2014;42(Web Server 1281 issue):W187-91. 1282 98. Ramírez F, Ryan DP, Grüning B, Bhardwaj V, Kilpert F, Richter AS, et al. deepTools2: 1283 a next generation web server for deep-sequencing data analysis. Nucleic Acids Res. 1284 2016;44(W1):W160-5. 1285 99. Anders S, and Huber W. Differential expression analysis for sequence count data. 1286 Genome biology. 2010;11(10):R106. 1287 100. Love MI, Huber W, and Anders S. Moderated estimation of fold change and 1288 dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. 1289 101. Edgar R, Domrachev M, and Lash AE. Gene Expression Omnibus: NCBI gene 1290 expression and hybridization array data repository. Nucleic Acids Res. 1291 2002;30(1):207-10. 1292 102. Brind'Amour J, Liu S, Hudson M, Chen C, Karimi MM, and Lorincz MC. An ultra-low- 1293 input native ChIP-seq protocol for genome-wide profiling of rare cell populations. 1294 Nat Commun. 2015;6:6033. 1295 103. Grzybowski AT, Chen Z, and Ruthenburg AJ. Calibrating ChIP-Seq with Nucleosomal 1296 Internal Standards to Measure Histone Modification Density Genome Wide. 1297 Molecular cell. 2015;58(5):886-99. 1298 104. Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, et al. Model-based 1299 analysis of ChIP-Seq (MACS). Genome biology. 2008;9(9):R137. 1300 105. Ross-Innes CS, Stark R, Teschendorff AE, Holmes KA, Ali HR, Dunning MJ, et al. 1301 Differential oestrogen receptor binding is associated with clinical outcome in breast 1302 cancer. Nature. 2012;481(7381):389-93. 1303 106. Amemiya HM, Kundaje A, and Boyle AP. The ENCODE Blacklist: Identification of 1304 Problematic Regions of the Genome. Sci Rep. 2019;9(1):9354. 1305 107. Yu G, Wang LG, and He QY. ChIPseeker: an R/Bioconductor package for ChIP peak 1306 annotation, comparison and visualization. Bioinformatics. 2015;31(14):2382-3. 1307 108. Warton DI, Wright IJ, Falster DS, and Westoby M. Bivariate line-fitting methods for 1308 allometry. Biol Rev Camb Philos Soc. 2006;81(2):259-91. 1309 109. Kang HJ, Kawasawa YI, Cheng F, Zhu Y, Xu X, Li M, et al. Spatio-temporal 1310 transcriptome of the human brain. Nature. 2011;478(7370):483-9. 1311 110. Rigby M, O'Donnell R, and Rupniak NM. Species differences in tachykinin receptor 1312 distribution: further evidence that the substance P (NK1) receptor predominates in 1313 human brain. The Journal of comparative neurology. 2005;490(4):335-53. 1314 111. Sener EF, Taheri S, Sahin MC, Bayramov KK, Marasli MK, Zararsiz G, et al. Altered 1315 Global mRNA Expressions of Pain and Aggression Related Genes in the Blood of 1316 Children with Autism Spectrum Disorders. J Mol Neurosci. 2019;67(1):89-96. 1317 112. De Felipe C, Herrero JF, O'Brien JA, Palmer JA, Doyle CA, Smith AJ, et al. Altered 1318 nociception, analgesia and aggression in mice lacking the receptor for substance P. 1319 Nature. 1998;392(6674):394-7.

1320 113. Huhtaniemi I, Zhang FP, Kero J, Hamalainen T, and Poutanen M. Transgenic and 1321 knockout mouse models for the study of luteinizing hormone and luteinizing 1322 hormone receptor function. Mol Cell Endocrinol. 2002;187(1-2):49-56. 1323 114. Burnham V, Sundby C, Laman-Maharg A, and Thornton J. Luteinizing hormone acts 1324 at the hippocampus to dampen spatial memory. Horm Behav. 2017;89:55-63. 1325 115. Hyde Z, Flicker L, Almeida OP, McCaul KA, Jamrozik K, Hankey GJ, et al. Higher 1326 luteinizing hormone is associated with poor memory recall: the health in men study. 1327 J Alzheimers Dis. 2010;19(3):943-51. 1328 116. Low WC, Roepke J, Farber SD, Hill TG, Sattin A, and Kubek MJ. Distribution of 1329 thyrotropin-releasing hormone (TRH) in the hippocampal formation as determined 1330 by radioimmunoassay. Neuroscience letters. 1989;103(3):314-9. 1331 117. Deng PY, Porter JE, Shin HS, and Lei S. Thyrotropin-releasing hormone increases 1332 GABA release in rat hippocampus. The Journal of physiology. 2006;577(Pt 2):497- 1333 511. 1334 118. Aguilar-Valles A, Sánchez E, de Gortari P, García-Vazquez AI, Ramírez-Amaya V, 1335 Bermúdez-Rattoni F, et al. The expression of TRH, its receptors and degrading 1336 enzyme is differentially modulated in the rat limbic system during training in the 1337 Morris water maze. Neurochemistry international. 2007;50(2):404-17. 1338 119. Neelands TR, Fisher JL, Bianchi M, and Macdonald RL. Spontaneous and gamma- 1339 aminobutyric acid (GABA)-activated GABA(A) receptor channels formed by epsilon 1340 subunit-containing isoforms. Molecular pharmacology. 1999;55(1):168-78. 1341 120. Mehta AK, and Ticku MK. An update on GABAA receptors. Brain Res Brain Res Rev. 1342 1999;29(2-3):196-217. 1343 121. Mons N, Harry A, Dubourg P, Premont RT, Iyengar R, and Cooper DM. 1344 Immunohistochemical localization of adenylyl cyclase in rat brain indicates a highly 1345 selective concentration at synapses. Proceedings of the National Academy of Sciences 1346 of the United States of America. 1995;92(18):8473-7. 1347 122. Baker LP, Nielsen MD, Impey S, Hacker BM, Poser SW, Chan MY, et al. Regulation and 1348 immunohistochemical localization of betagamma-stimulated adenylyl cyclases in 1349 mouse hippocampus. The Journal of neuroscience : the official journal of the Society 1350 for Neuroscience. 1999;19(1):180-92. 1351 123. Ledent C, Vaugeois JM, Schiffmann SN, Pedrazzini T, El Yacoubi M, Vanderhaeghen JJ, 1352 et al. Aggressiveness, hypoalgesia and high blood pressure in mice lacking the 1353 adenosine A2a receptor. Nature. 1997;388(6643):674-8. 1354 124. Horgusluoglu-Moloch E, Nho K, Risacher SL, Kim S, Foroud T, Shaw LM, et al. 1355 Targeted neurogenesis pathway-based gene analysis identifies ADORA2A associated 1356 with hippocampal volume in mild cognitive impairment and Alzheimer's disease. 1357 Neurobiology of aging. 2017;60:92-103. 1358 125. Vaz SH, Lérias SR, Parreira S, Diógenes MJ, and Sebastião AM. Adenosine A2A 1359 receptor activation is determinant for BDNF actions upon GABA and glutamate 1360 release from rat hippocampal synaptosomes. Purinergic Signal. 2015;11(4):607-12. 1361 126. O'Caoimh R, Kehoe PG, and Molloy DW. Renin Angiotensin aldosterone system 1362 inhibition in controlling dementia-related cognitive decline. J Alzheimers Dis. 1363 2014;42 Suppl 4:S575-86. 1364 127. Jackson L, Eldahshan W, Fagan SC, and Ergul A. Within the Brain: The Renin 1365 Angiotensin System. Int J Mol Sci. 2018;19(3). 55

1366 128. Denny JB, Polan-Curtain J, Wayner MJ, and Armstrong DL. Angiotensin II blocks 1367 hippocampal long-term potentiation. Brain research. 1991;567(2):321-4. 1368 129. Wayner MJ, Armstrong DL, Polan-Curtain JL, and Denny JB. Role of angiotensin II 1369 and AT1 receptors in hippocampal LTP. Pharmacol Biochem Behav. 1993;45(2):455- 1370 64. 1371 130. Abbracchio MP, Boeynaems JM, Barnard EA, Boyer JL, Kennedy C, Miras-Portugal 1372 MT, et al. Characterization of the UDP-glucose receptor (re-named here the P2Y14 1373 receptor) adds diversity to the P2Y receptor family. Trends Pharmacol Sci. 1374 2003;24(2):52-5. 1375 131. Chambers JK, Macdonald LE, Sarau HM, Ames RS, Freeman K, Foley JJ, et al. A G 1376 protein-coupled receptor for UDP-glucose. The Journal of biological chemistry. 1377 2000;275(15):10767-71. 1378

A H3K4me Writers me0 me1 me2 me3 WDSTS KMT2A (MLL1) Dystonia/ID KMT2B (MLL2/4) Kleefstra syndrome/ASD KMT2C (MLL3) Kabuki syndrome KMT2D (MLL4/2) Schizophrenia KMT2F (SET1A) KMT2G (SET1B) PRDM16 Erasers KDM1A CPRF KDM1B KDM5A ARID KDM5B ID/ASD KDM5CMRXSCJ KDM5D

Readers PHF21A PHF8 Potocki-Shaffer syndrome Siderius X-linked ID

B 129S1/SvlmJ x C57/Bl6 Kmt2a-HET Kdm5c-HET Kmt2a +/- Kdm5c +/-

F1:

WT Kmt2a-HET DM Kdm5c-KO -/y Kmt2a +/+ , Kdm5c +/y Kmt2a +/- , Kdm5c +/y Kmt2a +/+ , Kdm5c Kmt2a +/- , Kdm5c -/y C Number of offspring across 30 litters (% total males)

WT 2a-HET 5c-KO DM 17 16 21 14 (25%) (23%) (31%) (21%)

Adult body weight D 50 **** 5c-KO - DM 40 2a-HET - DM 30 * 2a-HET - 5c-KO 20 WT - DM **** WT - 5c-KO Body weight (g ) 10 *** WT - 2a-HET * 0 -5 0 5 10 15 WT 2a-HET5c-KO DM Difference between group means

Figure 1 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Contextual fear conditioning (CFC) Shock reactivity A * C 100 2500 n.s. 5c-KO - DM 80 2a-HET - DM 2000 60 2a-HET - 5c-KO * 1500 40 WT - DM 1000 Freeze (% ) WT - 5c-KO 20 * 500

WT - 2a-HET Shock reactivit y 0 0 WT 2a-HET 5c-KO DM -40 -20 0 20 40 60 WT 2a-HET 5c-KO DM Difference between group means B Novel object recognition (NOR) D Locomotor activity 100 * 300 n.s. 5c-KO - DM 80 2a-HET - DM 200 60 2a-HET - 5c-KO 40 WT - DM 100 20 WT - 5c-KO * Time spent with novel object (% ) WT - 2a-HET 0 verage motion inde x 0 WT 2a-HET 5c-KO DM -40 -20 0 20 40 A WT 2a-HET 5c-KO DM Difference between group means

Figure 2 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A Three chamber test for social interaction ** 100 5c-KO - DM 2a-HET - DM 2a-HET - 5c-KO 50 WT - DM *

stranger (% ) WT - 5c-KO ** Time spent with WT - 2a-HET 0 WT 2a-HET 5c-KO DM -40 -20 0 20 40 60 Difference between group means

B Tube test for social dominance ******* 100 22 21 46 90 80 70 WT 60 2a-HET 50 5c-KO Wins (%) 40 DM 30 20 n=3010 n=13 n=18 41 39 11 300

C Resident intruder: Aggressive behaviors 8 * 5c-KO - DM * 6 2a-HET - DM * 2a-HET - 5c-KO 4 WT - DM 2 WT - 5c-KO of behaviors

Average number WT - 2a-HET 0 WT 2a-HET 5c-KO DM -4 -2 0 2 4 6 Difference between group means Submissive behaviors 10 * 5c-KO - DM 8 *** 2a-HET - DM ** 6 2a-HET - 5c-KO 4 WT - DM WT - 5c-KO 2 of behaviors WT - 2a-HET Average number 0 WT 2a-HET 5c-KO DM -10 -5 0 5 10 Difference between group means

Figure 3 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not Basolateral amygdalacertified (BLA) by peer review)B is the author/funder. All rights reserved. No reuse allowed without permission. A WT 5c-KO Total dendritic lengths (BLA) 5 * 5c-KO - DM 4 2a-HET - DM 3 2a-HET - 5c-KO

2 WT - DM WT - 5c-KO

Lengt h (mm) 1 * WT - 2a-HET 0 -1000 -500 0 500 1000 1500 WT 2a-HET5c-KO DM Difference between group means

C Dendritic spine density (BLA) 4 **** 5c-KO - DM m) **** 3 2a-HET - DM **** 2a-HET - 5c-KO DM 2a-HET 2 WT - DM **** 1 WT - 5c-KO Density (#/ m ** WT - 2a-HET 0 ** WT 2a-HET 5c-KO DM -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 Difference between group means D Spine morphology, percent of total WT Thin 2a-HET Stubby Mushroom 5c-KO Filopodia DM Branched E Dorsal hippocampus CA1 (CA1) 0% 20% 40% 60% 80% 100% WT 5c-KO F

Total dendritic lengths (CA1) 5 n.s. 5c-KO - DM 4 2a-HET - DM 3 2a-HET - 5c-KO 2 WT - DM WT - 5c-KO Lengt h (mm) 1 WT - 2a-HET 0 -1000 -500 0 500 1000 WT 2a-HET 5c-KO DM Difference between group means DM 2a-HET Dendritic spine density (CA1) 2.0 ****

1.5 5c-KO - DM * 2a-HET - DM 1.0 2a-HET - 5c-KO WT - DM 0.5 Density (#/mm ) WT - 5c-KO **** WT - 2a-HET 0 *** -0.4 -0.2 0.0 0.2 0.4 WT 2a-HET5c-KO DM Difference between group means

Figure 4 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. A mRNA-seq (DEseq2: Padj < 0.1) B 2A-DEGs (P < 0.005) 300 300 271 250 250

200 167 200 150 127 125 150 100 100 41 50 50 0 1 14 0 0 0 3 5 14

Number of DEGs 50 Number of DEGs 50 100 73 100 64 Down Up Down Up 98 150 112 150 AMY HIP AMY HIP AMY HIP AMY HIP 2A-HET vs WT 5C-KO vs WT DM vs WT 2A-HET vs WT

C Rescue effect of mRNA expression in DM Amygdala Hippocampus Amygdala Hippocampus 5C-up genes 5C-up genes 2A-down genes 2A-down genes –8 –16 p = 1.41 x 10 p < 2.20 x 10 p = 1.26 x 10 –8 0 3 3 1 2 2 −0.5 0 1 1 −1.0 −1 0 Log2FC to WT Log2FC to WT Log2FC to WT Log2FC to WT 0 −1 −1.5 p = 1.61 x 10 –12 −2 −1 2A 5C DM 2A 5C DM 2A 5C DM 2A 5C DM

D Identifiction of rescue-driving genes Amygdala Hippocampus 5C-DEGs 5C-DEGs

2 2

0 0 Log2FC DM /WT Log2FC DM /WT −2 −2 slope = 0.91 ± 0.03 slope = 0.73 ± 0.03 p =1.43 x 10 –6 p < 6.02 x 10 –22 −2 0 2 −2 0 2 Log2FC_5C-KO/WT Log2FC_5C-KO/WT

2A-DEGs 2A-DEGs

2 2

0 0 Log2FC DM /WT Log2FC DM /WT −2 −2 slope = 0.51 ± 0.06 slope = 0.61 ± 0.07 p < 6.02 x 10 –22 p < 6.02 x 10 –22 −2 0 2 −2 0 2 Log2FC_2A-HET/WT Log2FC_2A-HET/WT Figure 5 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Global H3K4me level (AMY) B Inter-replicate distance C Number of DMRs in AMY A 1400 1200 1,147 2.5 1000 µl: 5 20 10 800 600 520 400 357 0.0 WT 200 0 WTWT WT 2A HET5C KODM 2A-HET 0 H3K4me1 200 0 95 5C-KO 240 H3K4me2 −2.5 DM-rep2 DM 400

Down600 Up

H3K4me3 PC2: 9% variance 800 563 Histone H3 −5.0 2A-HET 5C-KO DM 2A-HET −4 0 4 8 vs vs vs vs WT WT WT WT PC1: 77% variance padj < 0.05 p < 0.05 D Direction of H3K4me3 changes between single mutants E Behavior of 2A- and 5C-DMRs in DM KMT2A-DMRs in 5C-KO KDM5C-DMRs in 2A-HET 2A-DMRs-opposite 2A-DMRs-same 5C-DMRs-5C-DMRs-oppositeopposite 5C-DMRs-same 2A-DMRs-opposite 5C-DMRs-same (62) (155) p = 1.9 x 10 –6

Log2FC 5C vs WT –11

2A-DMRs-same Log2FC 2A vs WT Log2FC 2A vs WT 5C-DMRs-opposite p < 3.4 x 10 (107) (283) –12 p = 8.1 x 10 p < 2.2 x 10 –16 H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT Log2FC 2A vs WT Log2FC 5C vs WT 2A 5C DM 2A 5C DM 2A 5C DM 2A 5C DM Same direction Opposite direction Correlation of mRNA vs H3K4me3 levels of rescued genes F Genomic feature distrubution of H3K4me3 peaks G Rescued 2A-DEGs Rescued 5C-DEGs Feature Q1: 12 genes Q2: 47 Q1: 20 genes Q2: 1 All peaks (25,867) Promoter 5' UTR KMT2A-down 3' UTR 1st Exon KDM5C-up Other Exon 1st Intron 2 2 2A-DMR-opposite Other Intron p (χ ) < 0.0001 p (χ ) < 0.0001 r = 0.16 r = 0.27 Downstream Pearson Pearson p Pearson = 0.11 p Pearson = 0.02 5C-DMR-opposite5C-DMR-opposite Distal Intergenic H3K4me3 Log2FC DM vs 2A Q3: 8 Q4: 30 Q3: 45 Q4: 1 H3K4me3 Log2FC DM vs 2A H3K4me3 Log2FC DM vs 2A 0 25 50 75 100 Percentage (%) mRNA Log2FC DM vs 2A mRNA Log2FC DM vs 5C Figure 6 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. A Mouse adult visual cortex 60 - Kmt2a

40 -

20 -

0 -

60 - Kdm5c

40 - Expression (RPKM) 20 -

0 - L2 L4 L5 A E M O Igtp Sst Vip L5a L5b L6a L6b NdnfPvalb Sncg L2_3 OPC SMC Smad3 GABA Glu Non-neuronal

B Mouse adult brain 4 -Kmt2a 3 - 2 - 1 - 0 -

4 - Kdm5c 3 - Expression 2 - 1 - 0 - P TH HY MB MY CB OLF HPF STR PAL CTXsp Isocortex Human brain development C Brain regions Neocortical regions KMT2A KMT2A

7 7

6 6

5 5

4 4 Expression

3 3

50 100 200 500 2000 10000 30000 50 100 200 500 2000 10000 30000

KDM5C KDM5C 10 10

9 9 8 8 7 7

6 6 Expression 5 5 4 4 3 3

50 100 200 500 2000 10000 30000 50 100 200 500 2000 10000 30000 Age (days) Supplementary Figure 1 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A Chr. X B Chr. 9

49 kb

WT WT 2a-HET 2a-HET AMY AMY 5c-KO 5c-KO DM DM WT WT 2a-HET 2a-HET HIP 5c-KO HIP 5c-KO DM DM

Kdm5c Exon 11 and 12 Kmt2a Exon 9 and 10

C D E WT 2a-HET 5c-KO DM DM DM2A HET2A HETWT2A HET5C KO WT 2a-HET5c-KODM bp Kmt2a fmi KDM5C * CPu 770 - - WT * 300 - - del GAPDH 1mm

bp Kdm5c cc LV CPu 500 - del- - - - Pir 400 - - WT Tu

cc CA1 CA2 DG

Body weight timeline 3V F Am 40 WT 30 2a-HET DG 5c-KO CA1 20 DM CA2

SNR 10 Body weight (g ) 0 0 50 100 Postnatal day

Supplementary Figure 2 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A Resident intruder: aggressive behaviors

Darting Mounting Chasing/following Tail rattling Boxing/parrying 20 ** 20 ** 30 ** 25 n.s. 25 n.s.

15 15 20 20 20 15 15 10 10 10 10 10 5 5 5 5 0 0 0 0 0 Number of behaviors WT DM WT DM WT DM WT DM WT DM 5c-KO 5c-KO 2a-HET5c-KO 2a-HET 2a-HET 2a-HET5c-KO 2a-HET5c-KO Darting Mounting Chasing/following Tail rattling Boxing/parrying

5c-KO - DM * ** 2a-HET - DM ** ** 2a-HET - 5c-KO WT - DM * WT - 5c-KO ** WT - 2a-HET * -10 -5 0 5 10 -10 -5 0 5 10 -10 -5 0 5 10 15 20 -5 0 5 10 -6 -4 -2 0 2 4 6 Difference between group means

B Resident intruder: submissive behaviors Cowering Running away Upright 20 ** 30 ** 15 n.s.

15 20 10 10 10 5 5

0 0

Number of behaviors 0

WT DM WT DM WT DM 5c-KO 2a-HET5c-KO 2a-HET5c-KO 2a-HET Cowering Running away Upright

5c-KO - DM ** ** 2a-HET - DM ** 2a-HET - 5c-KO WT - DM * WT - 5c-KO WT - 2a-HET -10 -5 0 5 -15 -10 -5 0 5 10 15 -4 -2 0 2 4 Difference between group means

Supplementary Figure 3 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A Dendritic spine subtypes 1. Filopodia 2. Thin 3. Stubby 4. Mushroom 5. Branched

Supplementary Figure 4 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A Euclidean distances between RNA-seq samples from various brain regions Arbeitman (2019) This study

Arbeitman (2019) This study

Supplementary Figure 5 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A Principal component analysis 10

0 WT Amygdala DM 5C-KO −10 2A-HET PC2: 2% variance Hippocampus

−20

−25 0 25

B RNA-seq differentially-expressed genes (DEGs, padj < 0.1) C Kmt2a-HET DEGs, p < 0.005) Amygdala 2A vs WT 5C vs WT DM vs WT 2A vs WT

Hippocampus 2A vs WT 5C vs WT DM vs WT 2A vs WT

Supplementary Figure 6 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A Overlap DEGs between tissues

HIP_2A HIP_5C 1 gene AMY_5C

AMY_2A 72 genes

B Similarity in DEG expression between the two brain regions 2A-DEGs 5C-DEGs Amygala Hippocampus Amygala Hippocampus 2A 5C DM 2A 5C DM 2A 5C DM 2A 5C DM Gfap Gpr62 Ston1 Mxra8 Cd38 Ceacam1 Smim1 Fcgr2b Lims2 Serpinb9 Ctsz Slc38a2 B3gnt9 Ccdc173 Dglucy 2810414N06Rik Pnma3 Abca9 Rab34 Mpeg1 Hspa1a Pfkfb3 Celf6 Daam2 Zfp532 Klf3 Marveld1 Msantd1 Zmynd8 St8sia3os Parvb Ccar1 Prdm16 5031439G07Rik Psg16 Hcfc2 Ldah Cdh15 Vps36 Togaram2 Ccnl1 ENSMUSG00000097433 Mb Map10 Col6a6 Ctsc Tnfrsf12a Smoc1 Rasal3 Fzd7 Ctsw Bok Gm14308 Col27a1 ENSMUSG00000099767 Plin3 Trim63 Fam161a Rax Ttll9 4833423E24Rik Gm14164 Ccl6 Ampd3 Haao Smim20 ENSMUSG00000067389 Ppm1f Dnah3 ENSMUSG00000115650 ENSMUSG00000097451 ENSMUSG00000104388 Drc3 ENSMUSG00000097593 Rtl8a Smok4a Fmo1 ENSMUSG00000085242 Rhbdl1 ENSMUSG00000086524 Zc3h11a ENSMUSG00000115921 Fam136a Slc26a5 P2rx7 ENSMUSG00000105514 Rab29 ENSMUSG00000087179 Peg10 ENSMUSG00000111264 Cmbl Ankrd1 Six3 ENSMUSG00000103373 Adora2a Cnn1 Tmie ENSMUSG00000116358 Gaa Art1 Pvrig Gabre Trp63 Agt Hmx2 Iqub ENSMUSG00000115311 Frmpd2 Proca1 Tekt1 Sfrp1 Cox6b2 Evpl Scube3 Nkpd1 Rsph4a H2−Q2 Cfap61 ENSMUSG00000093898 Six3os1 Tcfl5 Frmd7 Neurog2 Dnaic2 Gm36251 Ttc16 Gm4779 Ccdc27 Atp6ap1l C4b Usp43 Ddo Tgm3 ENSMUSG00000113404 Castor1 Emilin3 Dnaic1 Met Neat1 ENSMUSG00000102179 Gamt ENSMUSG00000109648 Dlk1 ENSMUSG00000109505 Magel2 Cryga Ecel1 Tomt Mc3r Cc2d2b ENSMUSG00000111685 ENSMUSG00000110400 Bricd5 ENSMUSG00000104299 ENSMUSG00000116358 ENSMUSG00000106855 Pdzd3 Gm16867 Ptger4 Gins1 Vwa3b 9530062K07Rik Spink8 ENSMUSG00000097057 ENSMUSG00000097410 Prcd Asphd1 Gucy2e Rin1 Stk33 Espn Gck Gm3696 Pcbd1 Cep112 Rnase1 Eda Cd59a ENSMUSG00000109887 Ccdc153 Cenpe Ccdc187 ENSMUSG00000091050 ENSMUSG00000051606 Pde2a Xrcc3 Zfp593 Rxra ENSMUSG00000106019 Idua Kcnh3 Hyi Zbtb8a Fancc Slc40a1 Scnn1a Itpka Icam5 Arhgap27 Me3 Zfp783 Tnfaip8l1 Samd11 Akap5 Cabcoco1 Sult2b1 C230014O12Rik Ninl Amn ENSMUSG00000110249 Ltb4r2 Jpx BC049352 Cmtr1 Tcfl5 Rmnd1 Krt8 Nkain2 Drc7 Dtnb Loxl1 Speg 6430584L05Rik Nkapd1 Tcte1 Syngap1 Pyroxd2 Ipo11 Agap2 Rhcg Rasgef1a Sln Chrm1 Map3k15 Chn1 Col22a1 Lzts1 ENSMUSG00000110020 Kcnj4 Acta1 Synpr 4632428C04Rik ENSMUSG00000114019 Tnxb Jakmip1 Slc12a8 Cbx6 Cercam Parp6 4930447C04Rik Neurl1a Pgpep1l ENSMUSG00000116081 Stra6 Tmem121b Rnf17 Riox1 ENSMUSG00000111236 Adam33 Trim71 Gdf5 ENSMUSG00000101795 Ttc21a ENSMUSG00000084898 Cdhr4 Samd3 Ccl9 ENSMUSG00000087094 Ccdc194 ENSMUSG00000112466 Cfap43 Duxbl2 Serpinb1b Gm35394 Meig1 Vill Cdhr3 Mei1 T2 Otof ENSMUSG00000104674 Myl4 Slc14a2 ENSMUSG00000112441 Opn4 E130201H02Rik C230072F16Rik Gm5415 Ak9 B230319C09Rik Dnah3 ENSMUSG00000096968 Hkdc1 Sgms2 Tex9 Nhlrc4 ENSMUSG00000086108 ENSMUSG00000108778 Gpr88 Eppk1 Tpbg BC024139 Mak Rps4l Il17ra Gm11681 Dleu7 Tspo2 Sh2b2 Fkbp6 Doc2a Stk31 Cpq Lipo2 Cyb561 Glra4 Camk2d Zfp648 Filip1 ENSMUSG00000056418 Arhgap10 Dusp2 Efcab6 1700003F12Rik Pcdha4 Ppp1r16a Sptbn5 Mark1 ENSMUSG00000111681 Ankrd35 Lum Golga7b Rbm46 2610008E11Rik Col10a1 Plekhb2 ENSMUSG00000103378 Nkiras2 Tshz3 Ywhaz P2ry14 Stk11ip Nova1 Pitpnm1 Cxcl12 Arhgap1 Esyt1 Atp13a2 Stxbp6 Zfp777 Brsk1 Ctnnd2 Pcif1 Ddah1 Zfpm1 ENSMUSG00000111681 Zfp607b ENSMUSG00000095690 Isg20 Prmt3 ENSMUSG00000090907 Wee1 Hs3st2 Prickle1 Cux2 Ubqln4 B230216N24Rik Rhobtb2 Pamr1 Akap7 A930011G23Rik Dis3l Whrn Cdk16 Rab3b Camk2g Sptbn5 Csrnp2 Cntn6 Kazald1 Hrh1 ENSMUSG00000107728 Hs3st5 Ndnf Igsf21 Gucy2g Dpp10 Pla2g4e Atp2b4 Cdh13 Mcm6 Mgat5b Cntn4 Tmem132a ENSMUSG00000097604 Osbpl10 Gcnt4 Slc35f1 Pcdhga9 Kiss1r Cd34 Rnf144b Mgat5b Dact2 ENSMUSG00000087268 Cdh22 Hsf2bp Plxnd1 Ccnb1ip1 ENSMUSG00000104279 D1Pas1 Col5a1 Klk10 Garnl3 Cacng6 Fnbp1l Wdr95

−2 0 2 Log2 FC mut/WT

Supplementary Figure 7 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A Behavior of DEGs btween single mutants B Amygdala (AMY) Hippocampus (HIP) 5C-DEGs 5C-DEGs

Number of reciprocal- and similarly-misregulated genes 400 0 1 2 3 0 1 2 3 350 300

Log2FC 2A /WT Log2FC 2A /WT 250

−3 −2 −1 −3 −2 −1 71% 200 −3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3 150 Log2FC 5C-KO/WT Log2FC 5C-KO/WT 100

Number of DEGs 72% 83% 50 29% 92% 28% 2A-DEGs 2A-DEGs 0 8% 17% AMY HIP AMY HIP 5C-DEGs 2A-DEGs 5C /WT 0 1 2 3 0 1 2 3 Opposite change

Log2FC 5C /WT Log2FC_ Change in the same direction −3 −2 −1 −3 −2 −1

−3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3 Log2FC 2A-HET/WT Log2FC 2A-HET/WT

C Overlap of DEGs between mutants Amygdala Hippocampus

DM-DEGs DM-DEGs (139) (279)

Overlap: 84 34 116 2A-DEGs 7 (78) 2A-DEGs 5C-DEGs 5C-DEGs (139) (132) (344)

Supplementary Figure 8 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Input H3K4me3 ChIP libraries WT A WT B H3K4me3 B 100 H3K4me3 A A H3K9me3 A H3K9me3 B 90 H3K27me3 A H3K27me3 B 80 H3K36me3 A H3K36me3 B 70 H4K20me3 B H3K4me3 B H4K20me3 A 60 H3K4me1 A H3K4me1 B 50 H3K9me1 A H3K9me1 B 40 H3K27me1 A H3K27me1 B 30 H3K36me1 A H3K36me1 A 20 H4K20me1 A H4K20me1 B H3K4me3 A 10 H3K4me2 A H3K4me2 B 0 H3K9me2 A H3K9me2 B H3K27me2 A H3K27me2 B WT DMrep1rep2rep1rep2rep3rep1rep2rep3rep1rep2rep3 H3K36me2 A H3K36me2 B 5c-KO H4K20me2 A H4K20me2 B 2a-HET WT 2a-HET 5c-KO DM

B DMRs in volcano plots 2a-HET vs WT 5c-KO vs WT DM vs WT

C DMRs in MA plots 2a-HET vs WT 5c-KO vs WT DM vs WT

Normalized H3K4me3 levels

Supplementary Figure 9 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A Behavior of 2A- and 5C-DMRs in DM without DM-rep2

2A-DMRs-opposite 2A-DMRs-same 5C-DMRs-opposite5C-DMRs-opposite 5C-DMRs-same (200) (175) (972) (483)

p =2.2 x 10 –16 p =3.4 x 10 –9 p < 1.2 x 10 –6 p < 2.2 x 10 –16 H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT H3K4me3 Log2FC vs. WT 2A 5C DM 2A 5C DM 2A 5C DM 2A 5C DM

Supplementary Figure 10 bioRxiv preprint doi: https://doi.org/10.1101/567917; this version posted January 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Rescued DMRs: intergenic chr13: 5 kb 2 kb A chr18: 35,815,000 35,820,000 72,636,000 72,637,000 72,638,000 72,639,000 72,640,000 72,641,000 72,642,000 72,643,000 72,644,000 10 _ 10 _ chr16: 5 kb WT 20,236,000 20,237,000 20,238,000 20,239,000 20,240,000 20,241,000 20,242,000 20,243,000 20,244,000 Map6d10 _ 0 _ 1020 __ 10 _ 2a-HETWT 00 __ 0 _ 1020 __ 10 _ 2a-HET5c-KO 00 __ 0 _ 1020 __ 10 _ H3K4me3 5c-KODM 0 _ 0 _ 200 _ DM 0 _ Rescued DMRs: promoter chr16: 20 kb chrX: 5 kb 65,500,000 65,510,000 65,515,000 65,520,000 65,525,000 65,530,000 65,535,000 65,540,000 65,545,000 B 133,977,000 133,978,000 133,979,000 133,980,000 133,981,000 133,982,000 133,983,000 133,985,000 Chmp2b

10 _ 10 _ Sytl4 10 _ Pou1f1 WT 0 _ 0 _ 10 _ 10 _ 0 _ 10 _ 2a-HET

0 _ 0 _ 10 _ 10 _ 0 _ 5c-KO 10 _

0 _ 0 _ 10 _ 10 _ 0 _ H3K4me3 10 _ DM 0 _ 0 _

Unrescued DMRs chr16: 5 kb chr12: 20 kb C 20,236,000 20,237,000 20,238,000 20,239,000 20,240,000 20,241,000 20,242,000 20,243,000 20,244,000 28,590,000 28,600,000 28,610,000 28,620,000 28,630,000 28,640,000 20 _ Map6d1 10 _ Allc Colec11 Rps7

WT 0 _ 0 _ 20 _ 10 _ 2a-HET 0 _ 0 _ 20 _ 10 _ 5c-KO 0 _ 0 _ 20 _ 10 _ H3K4me3 DM 0 _ 0 _

chrX: 5 kb 133,977,000 133,978,000 133,979,000 133,980,000 133,981,000 133,982,000 133,983,000 133,985,000

Sytl4 10 _ WT 0 _ 10 _ 2a-HET

0 _ 10 _ 5c-KO

0 _ 10 _ DM 0 _

Supplementary Figure 11