bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder.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 ADNP promotes neural differentiation by modulating Wnt/β-catenin

2 signaling

3 Xiaoyun Sun1, Xixia Peng1, Yuqing Cao1, Yan Zhou2 & Yuhua Sun1* 4 5 Abstract

6 ADNP (Activity Dependent Neuroprotective Protein) is proposed as a neuroprotective

7 protein whose aberrant expression has been frequently linked to neural developmental

8 disorders, including the Helsmoortel-Van der Aa syndrome. However, its role in

9 neural development and pathology remains unclear. Using mESC (mouse embryonic

10 stem cell) directional neural differentiation as a model, we show that ADNP is

11 required for ESC neural induction and neuronal differentiation by maintaining Wnt

12 signaling. Mechanistically, ADNP functions to maintain the proper protein levels of

13 β-Catenin through binding to its armadillo domain which prevents its association with

14 key components of the degradation complex: Axin and APC. Loss of ADNP promotes

15 the formation of the degradation complex and hyperphosphorylation of β-Catenin by

16 GSK3β and subsequent degradation via ubiquitin-proteasome pathway, resulting in

17 down-regulation of key neuroectoderm developmental genes. We further show that

18 ADNP plays key role in cerebellar neuron formation. Finally, adnp gene disruption in

19 zebrafish embryos recapitulates key features of the mouse phenotype, including the

20 reduced Wnt signaling, defective embryonic cerebral neuron formation and the

21 massive neuron death. Thus, our work provides important insights into the role of

22 ADNP in neural development and the pathology of the Helsmoortel-Van der Aa

23 syndrome caused by ADNP gene mutation.

24 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder.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.

25 Introduction

26 ADNP was first described as a neural protective protein and has been implicated in

27 various neural developmental disorders and cancers1. This protein contains nine zinc

28 fingers, a homeobox domain and a PxLxP motif, suggesting that it functions as a

29 transcription factor. The exact roles of ADNP remain unclear, but an increasing

30 number of studies have shown that it functions as a key chromatin regulator by

31 interacting with chromatin remodelers or regulators. For instances, ADNP interacts

32 with core sub-units of SWI/SNF chromatin remodeling complex such as BRG1 and

33 BAF2502. By association with the chromatin regulator HP1, ADNP localizes to

34 pericentromeric heterochromatin regions where it silences major satellite repeat

35 elements3. Recently, ADNP was shown to form a triplex with CHD4 and HP1 to

36 control the expression of lineage specifying genes in embryonic stem cells4.

37 Mouse Adnp mRNA is abundantly expressed during early gestation and reaches

38 its maximum expression on E9.5. Adnp-/- mice are early embryonic lethal, displaying

39 severe defects in neural tube closure and brain formation5. Of note, E8.5 Adnp mutant

40 embryos exhibited ectopic expression of pluripotency gene Pou5f1 and reduced

41 expression of neural developmental gene Pax6 in the anterior neural plate. These

42 observations strongly suggested that ADNP plays essential roles during neurogenesis

43 of mouse embryos.

44 De novo mutations in ADNP gene have recently been linked to neural

45 developmental disorders, including the Helsmoortel-Van der Aa syndrome6. Patients

46 display multiple symptoms, usually manifesting in early childhood with features such

47 as intellectual disability, facial dysmorphism, motor dysfunction and developmental bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

48 delay, which share many features with another neural developmental disorder: autism

49 spectrum disorder (ASD). In fact, ADNP has been proposed as one of the most

50 frequent ASD-associated genes as it is mutated in at least 0.17% of ASD cases7. In a

51 mouse model, haploinsufficiency of Adnp leads to defective neuronal/glial

52 formation and compromised cognitive function8. Due to the significant relevance of

53 ADNP to neural developmental disorders, it is of great interests and importance in the

54 field to elucidate the pathogenic mechanism by ADNP gene mutation6.

55 In this work, we hypothesized that loss of ADNP leads to neural developmental

56 defects. We took use of ES cell directional neural differentiation as a model system to

57 investigate the role of ADNP in embryonic neural development. We showed that

58 ADNP is required for proper neural induction by modulating Wnt/β-catenin signaling.

59 Mechanistically, ADNP functions to maintain protein levels of β-Catenin through

60 binding to its armadillo domain which prevents its interaction with key components of

61 degradation complex: Axin and APC. Loss of ADNP leads to hyperphosphorylation of

62 β-Catenin by GSK3β and subsequent degradation via ubiquitin-proteasome pathway,

63 resulting in down-regulation of neuroectoderm developmental genes. Small

64 molecule-mediated activation of Wnt signaling can rescue the defects by loss of

65 ADNP. This work provides important insights into the role of ADNP in neural

66 development which would be useful for understanding the pathology of the

67 Helsmoortel-Van der Aa syndrome caused by ADNP mutation.

68 .

69 70 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

71 Results

72 Generation and characterization of Adnp-/- ESCs. To understand the molecular

73 function of ADNP, we generated Adnp mutant ESCs by using CRISPR/Cas9

74 technology. gRNAs were designed to target the 3' end of exon 4 of mouse Adnp gene

75 (Fig. 1a). We have successfully generated 2 Adnp mutant alleles (-4 and -5 bp),

76 revealed by DNA genotyping around the CRISPR targeting site (Fig. 1b). ADNP

77 protein was hardly detectable in Adnp-/- ESCs by Western blot using ADNP

78 antibodies from different resources (Fig. 1c), which strongly indicated that the mutant

79 alleles are functional nulls.

80 In the traditional self-renewal medium containing LIF-KSR plus FBS, the newly

81 established Adnp-/- ESC colonies overall exhibited typical ESC-like morphology (Fig.

82 1d). To understand how loss of Adnp affected the global gene expression, we

83 performed RNA-sequencing (RNA-seq) experiments for control and early passaged

84 Adnp-/- ESCs. Our RNA-seq analysis showed that the primitive endoderm genes such

85 as Gata6, Gata4, Sox17, Sox7 and Sparc were slightly up-regulated in the absence of

86 ADNP. However, loss of ADNP effects on pluripotency-related, mesodermal and

87 neuroectodermal genes were subtle. Our qRT-PCR and immunofluorescence assay

88 confirmed the RNA-seq results (Fig. 1e; Supplementary Fig. 1a). Consistently,

89 Adnp-/- ESCs can maintain self-renewal capacity for many generations before

90 eventually adopting a flatten morphology and exhibiting reduced alkaline phosphotase

91 activity (Supplementary Fig. 1b). Thus, our results indicated that acute ADNP

92 depletion in ESCs does not result in sudden and complete loss of self-renewal, while bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

93 prolonged ADNP depletion causes loss of ESC phenotype in the LIF/KSR medium.

94 To confirm that the observed phenotypes were not due to the possible off-target

95 effect of the CRISPR/Cas9 technology, we also reduced Adnp transcript levels in

96 ESCs by shRNA knockdown approach. The infection of lentivirus made from each of

97 the two individual shRNAs led to about 70-80% reduction of Adnp mRNA levels (Fig.

98 1c; Supplementary Fig. 1c). Adnp knockdown ESCs in the LIF-KSR medium

99 exhibited similar morphological characteristics and gene expression profile to Adnp-/-

100 ESCs, and could be passaged for many generations without obvious morphological

101 change (Supplementary Fig. 1d,e).

102

103 ADNP promotes ES cell directional neural differentiation. Next, we examined the

104 pluripotency of Adnp-/- ESCs by performing the classical embryoid bodies (EBs)

105 formation assay. Control ESCs were capable of forming large smooth spheroid EBs,

106 whereas Adnp-/- ESCs formed smaller rough disorganized structures (Fig. 1f). To

107 understand how loss of Adnp affected the global gene expression, we performed RNA

108 sequencing (RNA-seq) using day 6 EBs derived from control and Adnp-/- ESCs. A

109 total of 2,004 differentially expressed genes (DEGs) (log2 fold change>1.5 and p≤

110 0.05) were identified based on two RNA-seq replicates. Of which, an average of

111 1,088 genes were significantly up-regulated and 916 genes were down-regulated.

112 Extra-embryonic endodermal genes such as Gata6, Gata4, Sox17, Sox7 and Foxa2

113 were substantially up-regulated, while pluripotency genes such as Nanog and Sox2,

114 and neural-ectodermal genes such as Pax6, Olig2, Sox1, Nestin and Wnt1 were bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

115 significantly down-regulated (Fig. 1g). Our qRT-PCR assay of selected genes

116 confirmed the RNA-seq results (Fig. 1h). Thus, based on the ESC-derived embryoid

117 body (EB) model, ADNP is required for the neuroectodermal differentiation of

118 mESCs, which was in line with the recent report4.

119 To better understand the role of ADNP in ESC neural differentiation, we directly

120 differentiated control and Adnp-/- ESCs towards a neural cell fate. We initially used a

121 published serum-free adherent monolayer culture protocol using N2B27 medium9.

122 Control ESCs could be efficiently differentiated into neuro-ectodermal precursors and

123 further induced into mature neurons. Unfortunately, Adnp-/- ESCs failed to do so as

124 the majority of cells died in about 3-4 days (data not shown). The massive cell death

125 in this culture system prevented us from further analyzing the role of ADNP in neural

126 induction. To overcome this issue, we adapted a modified three-dimensional (3D)

127 floating neurosphere protocol that allows fast yet efficient differentiation of ESCs into

128 neural progenitors (Stage 1) and mature neurons (Stage 2) (Fig. 2a). The key of this

129 protocol is the addition of FGF and EGF growth factors in Stage 1. It is known that

130 FGF signaling plays a pivotal role in neural lineage specification through a

131 “autocrine” induction mechanism9,10. And the growth factor EGF and bFGF

132 synergistically promote neural progenitor cell survival, proliferation and neuronal cell

133 commitment11,12.

134 At the end of Stage 1 of ESC neural differentiation, neurospheres from control

135 ESCs were large, round and even (Fig. 2b). In contrast, neurospheres from Adnp-/-

136 ESCs were smaller, rough and disorganized. The qRT-PCR assay showed that day 6 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

137 control ESC-derived neurospheres abundantly expressed Olig2, Pax2, Pax6 and

138 Nestin, indicating of efficient formation of neural progenitor cells (NPCs) (Fig. 2c;

139 Supplementary Fig. 2a). Flow cytometric assay revealed that more than 60% of NPCs

140 were PAX6+, indicating that our ESC neural induction was of high efficiency (Fig. 1d).

141 When day 6 Adnp-/- ESC-derived neurospheres were analyzed, the expression of

142 neural progenitor markers such as Olig2, Pax2, Pax6 and Nestin was substantially

143 reduced (Fig. 2c-e). Flow cytometric assay showed that only 29% of Adnp-/-

144 ESC-derived NPCs were PAX6+. Thus, the ability of ESCs to differentiate into NPCs

145 was severely compromised in the absence of ADNP.

146 When day 9 neurospheres from control ESCs were plated onto gelatin-coated

147 plates to allow further differentiation for an additional 7-14 days (Stage 2), obvious

148 neuronal fiber-like structures characterized by the branched and elongated

149 morphology, were observed (Fig. 2b). qRT-PCR analysis showed that day 19

150 ESC-derived cells abundantly expressed neuronal and glial marker genes Tubb3 and

151 Gfap, indicating that more mature neural cell types such as neurons and astrocytic cell

152 types were formed (Fig. 2f). IF (immunofluoresence) staining showed that TuJ1 and

153 GFAP were abundantly expressed in these cells (Fig. 2g, h). In contrast, when

154 neurospheres from Adnp-/- ESCs were plated onto gelatin-coated plates for further

155 differentiation, few neuronal fibers were observed at day 19 (Fig. 2b). qRT-PCR, IF

156 and WB analysis showed that the expression of GFAP and TuJ1 was substantially

157 reduced in day 19 Adnp-/- ESC-derived neural cell cultures compared to control

158 counterparts (Fig. 2f-h; Supplementary Fig. 2b, c). bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

159 Taken together, we concluded that ADNP is required to promote ESC neural

160 differentiation, and its loss led to defective formation of neural progenitor cells and

161 mature neuronal cell types.

162

163 Loss of ADNP blocks neural induction by inhibiting the expression of early

164 neuroectoderm developmental genes. During ES cell neural differentiation, ADNP

165 levels were gradually induced, reaching its maximum level of expression at around

166 day 4 (Supplementary Fig. 3a). The dynamic expression pattern of Adnp was closely

167 correlated to that of the key neuroectoderm developmental genes such as Olig2, Pax6

168 and Nestin (Supplementary Fig. 3b). Importantly, loss of ADNP caused a significant

169 down-regulation of these genes (Fig. 2c). These observations implied that ADNP may

170 function at early stage of neural differentiation program by controlling the expression

171 of early neuroectoderm developmental genes.

172 To gain insights into the ADNP-dependent transcriptional program during ES

173 cell differentiation toward a neural cell fate, we monitored the dynamic changes of

174 gene expression at different time points of neural induction, by performing RNA-seq

175 experiments at day 0, day 3, and day 6 of neural differentiation from control and

176 Adnp-/- ESCs (Fig. 3a). In day 3 ESC-derived neurospheres, ~1,200 DEGs (log2 fold

177 change>1.5 and p≤ 0.05) were identified, of which 627 genes were significantly

178 up-regulated, and 578 genes were significantly down-regulated in the absence of

179 ADNP. Among the down-regulated genes were enriched for neuroectodermal markers

180 such as Sox3, Fgf5, Foxd3, Nestin, Sall2 and Nptx2, while the up-regulated genes bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

181 were enriched for pluripotency-related markers such as Dppa4, Esrrb, Klf4, Sox2 and

182 Zfp57, as well as primitive endodermal markers such as Sox7, Gata6, Gata4, ApoE,

183 Fabp3, Cubn and Cited. In day 6 ESC-derived neural progenitors, ~2,430 DEGs (log2

184 fold change>1.5 and p≤ 0.05) were identified, of which 1,344 genes were significantly

185 up-regulated, and 1,086 genes were significantly down-regulated (Fig. 3b). Of note,

186 genes implicated in neuroectoderm development such as Fgf5, Sox1, Sox3, Pax2/6,

187 Pax3/7, Foxd3, Nestin, Irx2, Cdx2, Sall2 and Nptx2 were significantly down-regulated.

188 Heat map analysis of day 3 and day 6 DEGs showed that there was a significantly

189 reduced expression of neuroectoderm-related genes, and a significantly increased

190 expression of pluripotency-related genes in the absence of ADNP (Fig. 3c;

191 Supplementary Fig. 3c). The qRT-PCR assay confirmed the RNA-seq results (Fig. 2c).

192 These data strongly suggested that ADNP performs an important role at the early

193 stage of neural differentiation by promoting the expression of neuroectoderm

194 developmental genes and inhibiting pluripotency-related genes.

195 If ADNP functions upstream of the key neuroectoderm developmental genes to

196 control neural induction, restoring ADNP should rescue the neural differentiation

197 defects from Adnp-/- ESCs. To this end, we took use of a Tet-Express inducible

198 transgenic Adnp-/- ES cell line in which 3×FLAG-ADNP can be induced by the

199 addition of Tet-Express protein (Supplementary Fig. 3d). We added Tet-Express at

200 different time points during mutant ES cell neural induction: day 0, day 2, and day 4,

201 and collected day 6-7 ESC-derived neurospheres for further analysis. We found that

202 the addition of Tet-Express at early time points (day 0 or 2) could partially rescue the bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

203 morphological and gene expression defects, while addition of Tet-Express after day 4

204 had minimal effects (Fig. 3d-f). When ESC-derived neurospheres were allowed for

205 further differentiation, day 19 cell cultures treated with Tet-Express at early time

206 points showed no significant differences from control cultures. By contrast, cultures

207 treated with Tet-Express at or after day 4 exhibited few neuronal fiber-like structures

208 (Supplementary Fig. 3e).

209 Based on the above data, we concluded that ADNP plays important roles at the

210 early stages of ESC neural differentiation by promoting the expression of key

211 developmental genes of neuroectodermal lineage.

212

213 Wnt/β-catenin signaling is reduced in the absence of ADNP. Next, we sought to

214 explore the underlying mechanisms by which ADNP controls ES cell neural induction

215 and neuronal differentiation. To this end, we performed KEGG pathway enrichment

216 analysis of differentially expressed genes (DEGs) identified in our RNA-sequencing

217 results. This analysis revealed a strong enrichment among DEGs for annotations

218 associated with signal transduction as well as signaling molecules and interaction

219 (Supplementary Fig. 4a). Detailed analysis showed that ADNP-dependent

220 down-regulated genes were highly enriched for gene networks regulating Wnt

221 signaling (Fig. 4a). Consistently, heat map analysis of DEGs revealed that there was a

222 significant reduction in transcript levels of Wnt pathway-related genes (Fig. 4b;

223 Supplementary Fig. 4b). This observation strongly suggested that loss of ADNP

224 blocks Wnt signaling pathway. To confirm the RNA-seq results, qRT-PCR for day 3 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

225 and day 9 ESC-derived neurospheres was performed to analyze the expression of Wnt

226 target genes such as Ccnd1, Axin2, Lef1 and Tcf712. The expression of these genes at

227 the indicated time points was greatly reduced in the absence of ADNP (Fig. 4c). Note

228 that the expression of Ctnnb1 gene, which encodes for β-Catenin, was barely changed

229 in the absence of ADNP. To further confirm that Wnt signaling was compromised in

230 the absence of ADNP, we performed TopFlash luciferase reporter assay for day 3

231 control and mutant ESC-derived neural progenitor cells. As shown in Fig. 4d, loss of

232 ADNP substantially decreased the TopFlash reporter activities, in the absence and

233 presence of Wnt3a.

234 On the other hand, we investigated whether overexpressing ADNP could

235 facilitate Wnt signaling. To this end, we performed the TopFlash reporter assay in

236 HEK293T cells transfected with an increasing amount of plasmids encoding ADNP.

237 As shown in Supplementary Fig. 4c, the TopFlash reporter activities were gradually

238 enhanced in a ADNP dose-dependent manner, in the presence and absence of Wnt3a.

239 Based on these data, we concluded that ADNP positively regulates Wnt/β-catenin

240 signaling.

241 Wnt signaling is well-known for its critical role in neural cell induction,

242 proliferation and differentiation, and dysfunction of this signaling pathway is

243 frequently associated with neural developmental disorders13. The above data

244 suggested that loss of ADNP function may affect neural induction by inhibiting Wnt

245 signaling. If this was the case, enhancing Wnt signaling should rescue the neural

246 induction defects from Adnp-/- ESCs. We therefore performed rescue experiments by bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

247 the addition of CHIR99021 or WNT3a from early time points (from day 0 or 1) of

248 mutant ESC neural differentiation. CHIR99021 (CHIR), a potent and highly selective

249 inhibitor of glycogen synthase kinase 3β (GSK3β), is a well-characterized canonical

250 Wnt signaling pathway activator. We have tested different concentration of

251 CHIR99021, and found that the addition of 3 μM CHIR99021 partially rescued the

252 gene and protein expression as well as morphological defects of day 6 Adnp-/-

253 ESC-derived neurospheres (Fig. 4e, f; Supplementary Fig. 4d, e). However, addition

254 of different doses of WNT3a proteins had little effect on the gene expression profiles.

255 This data suggested that ADNP may regulate Wnt signaling downstream of the

256 activity of the Wnt ligands but upstream or parallel to GSK3β which is known to

257 phosphorylate and induce proteasomal degradation of β-Catenin14.

258 Taken together, we concluded that ADNP is critical for ES cell neural

259 differentiation by promoting the Wnt/β-catenin signaling pathway. In the absence of

260 ADNP, Wnt signaling was impaired and the ability of ESC differentiation towards

261 neuroectodermal lineage cells was compromised.

262

263 Identifying β-Catenin as an ADNP interacting protein. In our immunoprecipitation

264 combined with mass spectrometry assay, β-Catenin was detected as one of the top hits

265 of ADNP immunoprecipitates (Fig. 5a). As β-Catenin is the key player in the

266 canonical Wnt signaling pathway, we hypothesized that ADNP may regulate Wnt

267 signaling by targeting β-Catenin during ESC differentiation towards a neural fate. We

268 first investigated whether ADNP interacts with β-Catenin by performing bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

269 co-immunoprecipitation (Co-IP) experiments. The Co-IP results showed that

270 endogenous ADNP can interact with β-Catenin in both ESCs and day 3 ESC-derived

271 neurospheres, exhibiting much stronger interaction in the neurospheres (Fig. 5b, c).

272 We also performed the Co-IP experiments using the 3×FLAG-Adnp overexpressing

273 Adnp-/- ESC line. As shown from Figure 5d, FLAG-ADNP readily pulled down

274 endogenous β-Catenin in day 3 ESC-derived neural progenitors (Fig. 5d).

275 Next, we mapped the interacting domains for ADNP and β-Catenin by

276 co-transfecting plasmids encoding the truncated form of ADNPs and β-Catenin into

277 293T cells (Fig. 5e). We found that it was the N-terminal fragment (1-685) but not the

278 C-terminal fragment of ADNP (686-1108) that is responsible to interact with

279 β-Catenin (Fig. 5f; Supplementary Fig. 5a). To investigate whether they interact

280 directly, we used a reticulate system to synthesize FLAG-ADNP (1-685) and

281 HA-β-Catenin. When they were mixed together, FLAG-ADNP (1-685) could readily

282 pull down HA-β-Catenin (Fig. 5g). Besides the Zinc fingers, ADNP-Nter (1-685)

283 contains NAP (which is also called davunetide or CP201), an 8-amino acid

284 neuroprotective peptide (NAPVSIPQ) (354-361 within the ADNP-Nter ) derived from

285 ADNP protein15. It was shown that NAP could enhance microtubule assembly and

286 neurite outgrowth by interacting with tubulin, and protect neurons and glia exposed to

287 toxins in mouse and rat models15,16. Recently, it was shown that the major symptoms

288 of the Helsmoortel-Van der Aa syndrome, such as developmental delays, vocalization

289 impairments, motor and pace dysfunctions, and social and object memory

290 impediments, could be partially ameliorated by daily NAP administration17,18. We bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

291 wondered whether NAP alone could be sufficient to interact with β-Catenin. When we

292 transfected plasmids encoding FLAG-ADNP-NAP and HA-β-Catenin into 293T cells,

293 FLAG antibodies could pull down HA-β-Catenin (Fig. 5h). Of note, ADNP-Nter

294 without NAP could still interact with β-Catenin but this interaction was much weaker

295 than that of ADNP-Nter and β-Catenin (Supplementary Fig. 5b).

296 Next, we determined the fragments of β-Catenin that are responsible for

297 interaction with ADNP. β-Catenin has an N-terminal domain (residues 1-150) that

298 harbors the binding site for GSK3β and CK1 phosphorylation, a central armadillo

299 domain (residues 151-666) composed of 12 armadillo repeats, and a C-terminal

300 domain (residues 667-781) (Fig. 5i). Our mapping experiments showed that the

301 central armadillo domain of β-Catenin strongly interacts with ADNP, while the

302 N-terminal (1-150) and the C-terminal fragment (667-781) barely interact with ADNP

303 (Fig. 5j; Supplementary Fig. 5c). These data indicated that the armadillo domain of

304 β-Catenin primarily mediates the interaction with ADNP.

305 Finally, we performed IF experiments to investigate the cellular co-localization

306 of ADNP and β-Catenin. During ES cell neuronal differentiation, there was a marked

307 change in the intra-cellular localization of ADNP proteins: from predominantly

308 nuclear localization in undifferentiated ESCs to predominantly cytoplasmic

309 localization in ESC-derived NPCs and mature neurons (Supplementary Fig. 5d). In

310 ESC-derived NPCs and mature neurons, ADNP was extensively co-localized with

311 β-Catenin in the cytosol (Fig. 5k).

312 Taken together, we identified β-Catenin as an ADNP interacting protein, which bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

313 suggested that ADNP had important roles in the regulation of Wnt/β-catenin signaling

314 by physical association with cytosolic β-Catenin.

315

316 ADNP stabilizes β-Catenin during ESC neural differentiation. Given the physical

317 association of ADNP and β-Catenin, we asked whether ADNP is involved in the

318 turnover of β-Catenin protein. First, we examined the expression levels of β-Catenin

319 in day 3 control and Adnp-/- ESC-derived neurospheres. Loss of ADNP led to a

320 significant reduction of total β-Catenin levels (Fig. 6a). This reduction was primarily

321 contributed by the alteration of the cytoplasmic fraction of β-Catenin. In fact, the

322 levels of the nuclear fraction of β-Catenin were barely changed in the absence of

323 ADNP (Fig. 6b). Throughout ESC neural differentiation, total β-Catenin levels

324 remained lower in the absence of ADNP than its presence (Fig. 6c, d; Supplementary

325 Fig. 6a, b). These observations strongly suggested that ADNP regulates the turnover

326 of β-Catenin which primarily occurred in the cytosol. When ESC-derived neural

327 progenitors were treated with the proteasome inhibitor MG132, loss of ADNP failed

328 to induce β-Catenin degradation, reflecting that ADNP-regulated β-Catenin

329 degradation was dependent on the ubiquitin-proteasome pathway (Fig. 6e).

330 Consistently, loss of ADNP strongly enhanced the ubiquitylation levels of β-Catenin,

331 which could be reversed by restoring 3×FLAG-ADNP in Adnp-/- ESC-derived

332 neurospheres (Fig. 6f ; Supplementary Fig. 6c).

333 On the other hand, we investigated the effect of overexpressing ADNP on total

334 β-Catenin levels in HEK293T cells. Overexpressing ADNP led to higher expression bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

335 of β-Catenin (Fig. 6g). In addition, overexpressing ADNP increased β-Catenin levels

336 in a dose dependent fashion. Next, we asked whether overexpression of ADNP

337 proteins affects the stability of coexpressed β-Catenin by co-transfecting plasmids

338 encoding FLAG-ADNP and HA-β-Catenin into 293T cells. We found that

339 FLAG-ADNP could stabilize HA-β-Catenin in a dose-dependent fashion

340 (Supplementary Fig. 6d). Combining both loss of function and gain of function data,

341 we concluded that ADNP is required for the stability of β-Catenin in ESCs and during

342 ESC differentiation towards a neural cell fate.

343 It is well-known that β-Catenin is subjected to phosphorylation by GSK3β who

344 destabilizes β-Catenin by phosphorylating it at Ser33, Ser37 and Thr41. A previous

345 study reported that there might be an increase in the ratio of active/inactive GSK3β in

346 Adnp+/- mice, suggesting that ADNP may negatively regulate the phosphorylation

347 activity of GSK3β8. To investigate whether ADNP regulates the stability of β-Catenin

348 through modulating GSK-3β activities, WB analysis for phospho-GSK3β (Ser9) and

349 total GSK3β were performed for day 3 control and Adnp-/- ESC-derived neurospheres.

350 We found that both the phospho-GSK3β and total GSK3β levels were barely altered

351 in the absence of ADNP (Supplementary Fig. 6e). In addition, overexpression of

352 ADNP had little effect on phospho-GSK3β and total GSK3β levels (Fig. 6g). These

353 results suggested that ADNP does not modulate the levels or activities of GSK3β.

354 Next, we asked whether ADNP is involved in the regulation of β-Catenin

355 phosphorylation by GSK3β. We found that this was the case. As shown in Fig. 6a-g,

356 phospho-β-Catenin levels were elevated in the absence of ADNP and were bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

357 significantly reduced in ADNP overexpressing cells. To understand the molecular

358 mechanism by which ADNP may regulate β-Catenin phosphorylation levels, we

359 further mapped the armadillo repeat domain that is responsible for interacting with

360 ADNP. The armadillo repeat domain of β-Catenin has 12 armadillo repeats,

361 containing overlapping binding sites for many binding partners, including E-cadherin,

362 TCF, Axin and APC14. It is well-known that Axin, APC and β-Catenin form the

363 degradation complex that promotes the GSK3β-mediated phosphorylation and

364 subsequent degradation of β-Catenin. We divided the armadillo repeat domain into

365 two parts: one contains repeats 1-4 (Arm1) which are known to mediate β-Catenin for

366 the interaction with the scaffolding protein Axin, and another one composed of

367 repeats 5-12 (Arm2) which are known to mediate β-Catenin for the interaction with

368 APC (Fig. 7a)19. By performing IP experiments, we found that both the Arm1 and

369 Arm2 were able to interact with ADNP although Arm2 displayed a stronger capacity

370 to associate with ADNP (Fig. 7b, c). Based on the the result, we hypothesized that

371 ADNP may protect β-Catenin from hyperphosphorylation by GSK3β through

372 competing with both components of the degradation complex: Axin and APC. To

373 directly investigate whether ADNP competes with Axin or APC for β-Catenin, we

374 examined the relationship between ADNP – β-Catenin and β-Catenin – Axin

375 interactions by performing competitive protein-binding assay. The amount of Axin

376 co-immunoprecipitated with β-Catenin became reduced by the increased addition of

377 ADNP (Fig. 7d). Similar result was obtained for the relationship between ADNP–

378 β-Catenin and β-Catenin – APC interactions. The amount of APC bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

379 co-immunoprecipitated with β-Catenin became reduced by the increased addition of

380 ADNP (Fig. 7e).

381 To further confirm that ADNP can protect β-Catenin from hyperphosphorylation

382 by GSK3β, we transfected an increasing dose of plasmids encoding wild type ADNP

383 or ADNP-Cter (686-1108) mutants into 293T cells, and examined the effects on total

384 β-Catenin and p-β-Catenin levels. The ADNP-Cter here served as a negative control

385 as it was unable to interact with β-Catenin. Overexpressing wild type ADNP

386 decreased the p-β-Catenin levels and increased the total β-Catenin levels, in a dose

387 dependent fashion (Fig. 6g). However, ADNP-Cter overexpression had no effects on

388 either β-Catenin or p-β-Catenin levels (data not shown). This data suggested that the

389 interaction between ADNP and β-Catenin was required for the ADNP mediated

390 stabilization of β-Catenin through modulating its phosphorylation levels.

391 Finally, we asked whether increasing β-Catenin protein levels could rescue the

392 neural induction defects caused by loss of ADNP. To this end, we made a Tet-Express

393 inducible HA-β-catenin transgenic Adnp-/- ESC line and used it for neural induction.

394 Before performing rescue experiments, we confirmed that addition of Tet-Express

395 proteins could induce the expression of total β-Catenin at similar levels to that of

396 control ESCs (Supplementary Fig. 7f). When the Tet-Express transactivitor was added

397 every day at the early stage of neural induction, the transgenic Adnp-/- ESCs could

398 form neural progenitors and neuronal cells with gene expression profiles similar to

399 that of wild type ESCs (Fig. 6h, 7g). Thus, restoring β-Catenin levels had similar

400 rescue effects as CHIR99021 treatment or restoring ADNP on Adnp-/- ESC neural bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

401 induction and differentiation. Taken together, we concluded that ADNP controls ESC

402 neural induction primarily by stabilizing β-Catenin.

403

404 Loss of ADNP leads to defective cerebellar neuron formation. By biochemical

405 analysis, it has been shown that ADNP is abundantly expressed in the hippocampus,

406 cerebral cortex and cerebellum of human and mouse brain20, 21. However, the

407 spatiotemporal expression patterns of ANDP in mammalian embryonic brain have not

408 been reported in detail. To explore this, vector slices from ten-week-old mouse

409 embryonic brains were subjected to immunohistochemistry assay using ADNP

410 antibodies. ADNP was expressed broadly in the embryonic brain, including olfactory

411 bulb, cerebellum, hippocampus and cortex, with cerebellar region displaying very

412 strong signal (Supplementary Fig. 7a). This observation suggested that ADNP is

413 cell-autonomously required for mouse embryonic brain development and maturation,

414 including the cerebellum. This was in line with the observation that the cerebellum

415 and areas related to it are one of the most consistently abnormal brain regions in

416 various neural developmental disorders22, 23. In fact, children with ASD or the

417 Helsmoortel-Van der Aa syndrome often show different extent of intellectual

418 disability and cognitive deficits, reflecting the dysfunction of the cerebellum.

419 We hypothesized that loss of Adnp may cause developmental and functional

420 defects of embryonic cerebellum. To test this hypothesis, we took use of ESC-based in

421 vitro cerebellar neuron differentiation as a model system, and adapted the published

422 SFEB protocols that have been shown to efficiently induce ESCs into GABA receptor bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

423 alpha 6+ granule cells and L7+ Purkinje neurons via Math1 and Ncam expressing

424 precursors24-27(Fig. 8a). And control and mutant ESCs were differentiated into

425 cerebellar neurons and comparison assay was performed to gain insights into the role

426 of ADNP in this process.

427 Round and large floating aggregates were formed after 5-10 days of neural

428 differentiation from control ESCs. In contrast, the aggregates derived from Adnp-/-

429 ESCs were much smaller and uneven, suggesting of defective formation of neural

430 precursors (Fig. 8b). It has been shown that a significant proportion of

431 SFEB/BMP4/Wnt3a/FGF8-treated ES cells at day 6 will become the external granule

432 layer (EGL) precursors expressing En2, Pax6, Six3 and Zic127. We therefore examined

433 the expression of these markers for day 6 control and Adnp-/- ESC-derived SFEB

434 cultures. In control ESC-derived aggregates, En2, Six3, Pax2, Zic1, Wnt1 and Pax6

435 were abundantly expressed, suggesting that EGL precursors were successfully

436 induced. In Adnp-/- ESC-derived aggregates, however, the expression of these genes

437 was greatly reduced (Fig. 8c). After 10 days of ESC neural differentiation, aggregates

438 derived from control ESCs highly expressed Math1, Zic1 and Ncam, while the

439 counterparts from mutant ESCs exhibited much lower levels of these markers (Fig. 8d,

440 e; Supplementary Fig. 7b-d). Importantly, the reduced expression of the EGL markers

441 in the absence of ADNP was accompanied by the reduction of β-Catenin levels (Fig.

442 8e). This suggested that Wnt/β-catenin signaling under the control of ADNP is

443 required for proper generation of cerebellar neuron precursors.

444 The Math1 or Zic1 positive precursors could be further induced into GABA bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

445 receptor alpha 6 positive granule cells and L7 positive Purkinje neurons on day 20 of

446 ESC in vitro differentiation24, 27. The expression of Purkinje cell marker L7 was

447 detected by immunofluorescence staining of day 20 neuronal cell cultures. As shown

448 in Fig. 8f, day 20 aggregates from control ESCs highly expressed Purkinje cell marker

449 L7, while the counterparts from Adnp-/- ESCs expressed this marker at much lower

450 levels.

451 Taken together, we concluded that ADNP is required for proper cerebellar neuron

452 formation likely by modulating Wnt/β-catenin signaling. Based on this result, we

453 tentatively propose that cerebellar developmental defects may be important pathology

454 of the Helsmoortel-Van der Aa syndrome.

455

456 ADNP regulates ESC-derived neural progenitor proliferation by maintaining

457 Wnt signaling. Adnp-/- embryos die in uteri displaying a striking disintegration

458 phenotype, likely due to the marked growth inhibition commencing at about E8.05. In

459 addition, Adnp+/− mice show a significant neurodegeneration phenotype with

460 reduced neuronal survival 8. These observations suggested that ADNP may regulate

461 neural progenitor cell proliferation or apoptosis in a mouse model. We hypothesize

462 that defective Wnt signaling caused by loss of function of ADNP contributes to this

463 phenotype.

464 During ESC differentiation into neurospheres, a marked reduction of total neural

465 cells was observed in the absence of ADNP (Fig. 1b, 3d). This could be confirmed by

466 comparing the cell pellet from day 6 control and Adnp-/- ESC-derived aggregates (Fig. bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

467 9a). We immunostained day 6 neurospheres with a phospho-histone H3 (PH3)

468 antibody, which is an M-phase marker, and quantitatively scored the number of

469 proliferating cells. We found that PH3 signal in Adnp-/- ESC-derived aggregates was

470 dramatically decreased compared to control ESC-derived neurospheres, suggesting

471 that loss of ADNP led to cell proliferation defects (Fig. 9b, c). To confirm this, WB

472 analysis was performed during neural induction of control and Adnp-/- ESCs using

473 PH3 antibodies. PH3 levels were significantly higher in the presence of ADNP than

474 its absence (Fig. 9d). Importantly, the PH3 levels could be partially rescued by

475 restoring ADNP in day 6 Adnp-/- ESC-derived neurospheres (Fig. 9e).

476 To understand whether ADNP was involved in apoptosis during ES cell

477 differentiation towards to a neural fate, WB analysis was performed using Caspase-3

478 antibodies. Caspase-3 is a well-known key protease that is activated during the early

479 stages of apoptosis. As shown in Fig. 9f, there was an increased cleaved Caspase-3

480 levels in day 6 Adnp-/- ESC-derived neurospheres compared with control

481 ESC-derived neurospheres. This data indicated that loss of ADNP led to the increased

482 apoptosis during ESC neural induction.

483 Importantly, addition of Wnt signaling agonist CHIR or restoring FLAG-ADNP

484 could partially rescue the cell proliferation defects by loss of ADNP, which supported

485 that ADNP controls cell proliferation or apoptosis by modulating Wnt/β-catenin

486 signaling pathway (Fig. 9b-g). It is well-known that Wnt signaling plays key roles in

487 cell fate determination, differentiation, proliferation and apoptosis18.

488 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

489 Loss of adnp leads to the reduced Wnt signaling and defective brain development

490 in a zebrafish model. Zebrafish has been proven to be an excellent in vivo model to

491 study vertebrate neurogenesis and pathology of ASD-related genes28, 29. We decided to

492 use zebrafish as an in vivo model to further study the function of ADNP in neural

493 induction and early neurogenesis. Zebrafish has two adnp homologues: adnpa and

494 adnpb gene. Both genes were maternally deposited into the zygote and expressed

495 ubiquitously until the early gastrulation stage when the expression of adnpa persisted

496 while the expression of adnpb began to disappear. During early somitogenesis, the

497 expression of both genes re-appeared along the dorsal midline, and by 24 hpf was

498 gradually refined to head and gut regions (Supplementary Fig. 8a)30. After 72h, adnpa

499 expression restricted to the forebrain, the midbrain, the mid-hindbrain boundary and

500 eye area (Supplementary Fig. 8b).

501 To understand the function of Adnp in zebrafish, we generated adnpa and adnpb

502 mutants using the CRISPR/Cas9 technology (Fig. 10a, b). Both adnpa and adnpb

503 mutant embryos can grow up to adults without apparent morphological defects

504 (Supplementary Fig. 8c). We went on to obtain adnpa adnpb double mutant zebrafish

505 lines by crossing adnpa-/- and adnpb-/- fish. We observed that a small portion of

506 adnpa adnpb double mutant embryos (about 1/7) at 1 dpf showed abnormal

507 morphology, ranging from slightly ventralization to severe ventralized phenotypes

508 (Supplementary Fig. 8d). The majority of adnpa adnpb double mutant embryos

509 appeared normal at 1 dpf (Supplementary Fig. 8d). Strikingly, about 16% (30/190) of

510 these embryos at 48 hpf displayed abnormal morphology with massive neuronal death bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

511 in brain (Fig. 10c). To further study the neuronal death phenotype, we performed

512 immunofluorescence (IF) staining of HuC/D antibodies for 36 hpf control embryos

513 and mutant embryos that began to show neuronal death in the embryonic brain. As

514 shown in Fig. 10d, the HuC/D signal in adnpa adnpb double mutant embryos was

515 greatly decreased compared to that of controls. To further investigate the effects of

516 Adnp deficiency on embryonic brain neuron development, we crossed adnpa adnpb

517 double mutant females to the transgenic Tg(elavl3: GFP) male fish. We found that the

518 GFP signal was significantly reduced in 3 dpf Adnp deficient embryos compared to

519 the controls (Fig. 10e). Finally, Whole mount in situ hybridization was performed for

520 2 dpf control and adnpa adnpb double mutant embryos to examine the expression of

521 neural developmental genes. We found that the expression of dlx5a, neurod1 and

522 phox2a was greatly reduced in the double mutant embryos compared with sibling

523 controls (Fig. 10f).

524 Next, we asked whether loss of Adnp in zebrafish embryos led to the reduction

525 of Wnt/β-catenin signaling. We performed TopFlash luciferase assay for lysates from

526 8 hpf wild type and adnpa adnpb double mutant embryos. As shown in Fig. 10g, the

527 TopFlash luciferase activities were greatly reduced in the absence of Adnp. To

528 confirm that Wnt signaling was reduced in adnpa adnpb double mutant zebrafish

529 embryos, we examined the expression of known Wnt target genes such as lef1, ccnd1

530 and axin2. qRT-PCR assay showed that the expression of these genes was

531 significantly reduced in adnpa adnpb double mutant embryos, which could be

532 partially rescued by the micro-injection of adnpa mRNAs (Fig. 10h). bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

533 On the other hand, we overexpressed Adnpa by injecting adnpa mRNAs into

534 1-cell stage wild type zebrafish embryos. Overexpression of Adnpa leads to

535 dorsalization of embryos (Supplementary Fig. 10e) and activation of the

536 Wnt/β-catenin signaling pathway as revealed by the TopFlash luciferase reporter

537 assay (Fig. 10g).

538 Taken together, we concluded that loss of Adnp led to the reduction of Wnt

539 signaling and defective brain development in zebrafish embryos.

540

541 Discussion

542 De novo mutations in ADNP have recently linked to the Helsmoortel-Van der Aa

543 syndrome, a complex neural developmental disorder displaying multiple symptoms

544 usually manifest in early childhood. In fact, ADNP has been proposed as one of the

545 most frequent ASD-associated genes as it is mutated in at least 0.17% of ASD cases.

546 It is therefore important to determine the molecular function of this protein, especially

547 during embryonic neurogenesis.

548 We found that loss of ADNP had little effect on the mRNA levels of Ctnnb1 gene

549 which encodes β-Catenin in ESCs. This ruled out that ADNP controls β-Catenin levels

550 by transcriptionally regulating Ctnnb1. In fact, Ctnnb1 gene was barely bound by

551 ADNP as revealed by the ChIP-seq assay (our unpublished observation)4. During ESC

552 differentiation towards a neural fate, ADNP translocates from predominantly in the

553 nuclei in ESCs to predominantly in the cytoplasm in neural cell types. These

554 observations suggested that the intra-cellular localization of ADNP is tightly regulated bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

555 during neural induction, or that ADNP plays an important role in the cytoplasm for the

556 induction, maintenance and mature of neural cell types. By IP in combination with

557 mass spectrometry assay, β-Catenin was highly presented as a top ADNP interacting

558 protein. Considering that Wnt/catenin signaling pathway is critically required for

559 neural induction and differentiation, we reasoned that ADNP may control neural

560 induction or embryonic neurogeneis by regulating Wnt signaling via targeting its key

561 transducer β-Catenin. We speculated that the most likely mechanism was that ADNP

562 controls the stability of β-Catenin in the cytoplasm. This idea was in line with the

563 observation that although ADNP interacts with β-Catenin in both the nuclei and the

564 cytoplasm, loss of ADNP had a significant reduction of the cytoplasmic but not the

565 nuclear faction of β-Catenin. In this work, we provided solid evidences that ADNP

566 directly associates with β-Catenin and regulates its phosphorylation by GSK3β.

567 Mechanistically, ADNP functions to maintain the proper levels of β-Catenin through

568 binding to its armadillo domain which prevents its association with key components

569 of the degradation complex: Axin and APC.

570 NAP, an 8 aa peptide derived from ADNP, was shown to interact with tubulin to

571 enhance microtubule assembly and neurite outgrowth, and protect neurons and glia

572 exposed to toxins in mouse and rat models1, 15, 16. Recently, it was shown that major

573 symptoms of the Helsmoortel-vanderAa syndrome, such as developmental delays,

574 vocalization impediments, motor and pace dysfunctions, and social and object

575 memory impairments, could be efficiently ameliorated by daily NAP administration17.

576 Towards the clinical development of NAP (CP201) for the treatment of the bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

577 Helsmoortel-Van der Aa syndrome, it is important to comprehensively understand the

578 mechanisms underlying the neuroprotective role of NAP. In this work, we showed that

579 ADNP-NAP alone can interact with β-Catenin, suggesting that NAP can execute

580 similar function as ADNP in the regulation of Wnt/β-Catenin signaling pathway. This

581 result provides another possible mechanistic role of NAP in the improvement of

582 symptoms of the Helsmoortel-Van der Aa syndrome: through maintaining Wnt

583 signaling via stabilizing β-Catenin. In the future, it will be interesting to look into this

584 possibility.

585 Adnp-/- mice are early embryonic lethal, displaying severe defects in neural tube

586 closure and brain formation5. And E8.5 Adnp mutant embryos exhibited reduced

587 expression of Pax6. These observations strongly suggested that ADNP plays a key

588 role in embryonic neurogenesis. In this work, using a ESC in vitro neural

589 differentiation system, we showed that ADNP performs an important role at the early

590 stage of neural differentiation by promoting the expression of key neuroectoderm

591 developmental genes. Restoring ADNP levels or activation of Wnt signaling by the

592 small molecule CHIR at the early stage of ESC neural induction can partially rescue

593 the gene expression and morphological defects of day 6 Adnp-/- ESC-derived

594 neurospheres, suggesting that ADNP controls the expression of early neural

595 developmental genes through Wnt/β-catenin signaling pathway. In addition, we show

596 that ADNP is abundantly expressed in the whole brain of 10-week-old mice, and is

597 required for proper cerebellar neuron formation. Together, our work suggested that

598 ADNP is cell autonomously required for not only embryonic neurogenesis but also bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

599 the formation and maturation of vertebrate brain.

600 The majority of 2dpf adnpa-/- adnpb-/- zebrafish embryos exhibited mild

601 phenotype in the brain development. Strikingly, a small portion of the double mutant

602 embryos exhibited massive neuronal cell death in the head region, which recapitulates

603 the key phenotype of Adnp deficient mouse embryos. We also showed that ADNP

604 plays important role in cerebellar neuron formation using ESC-based in vitro

605 cerebellar neuron differentiation model. Thus, our work in both ESC and zebrafish

606 models strongly supported the hypothesis that ADNP is required for embryonic brain

607 formation and its loss causes severe neural developmental defects. In the future, more

608 work is needed to test this hypothesis in vivo. Because Adnp mutant mice are

609 embryonic lethal, it will be necessary to establish a conditional knockout mouse

610 model to dissect out the function of ADNP in brain formation.

611 While no single gene is mutated in more than 1% of patients, autism-related

612 genes may functionally converge to commonly affected cellular pathways and

613 protein-protein interaction networks7. It is proposed that the most commonly affected

614 networks are the Wnt signaling pathway, and the pathways involving neurogenesis,

615 synaptic function and the chromatin remodeling. An increasing number of studies

616 have shown that ADNP functions as a key chromatin regulator by interacting with

617 chromatin remodelers or regulators. For instances, ADNP interacts with core sub-units

618 of SWI/SNF chromatin remodeling complex such as BRG1 and BAF250 as well as

619 CHD4, a key component of the NuRD complex4,7. Because ADNP is closely involved

620 in the regulation of both the chromatin remodeling and the Wnt signaling pathways, bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

621 we propose that ADNP is a master regulator implicated in ASD and the related

622 disorders. This may imply why ADNP is one of the most frequently mutated genes

623 associated with neural developmental disorders, including the Helsmoortel-Van der

624 Aa syndrome.

625

626 Methods

627 ES Cell Culture. Mouse embryonic stem cells (mESCs) R1 were maintained in

628 Dulbecco's Modified Eagle Medium (DMEM, BI, 01-052-1ACS) high glucose media

629 containing 10% fetal bovine serum (FBS, Gibco, 10099141), 10% knockout serum

630 replacement (KSR, Gibco, 10828028), 1 mM sodium pyruvate (Sigma, S8636), 2 mM

631 L-Glutamine (Sigma, G7513), 1,000 U/ml leukemia inhibitory factor (LIF, Millipore,

632 ESG1107, ) and penicillin/streptomycin (Gibco, 15140-122) at 37°C with 5% CO2.

633 The 2i culture condition was used as described previously31. The commercial

634 ESGRO-2i Medium (Merck-Millipore, SF-016-200) was also used when necessary.

635 We found that in 2i medium, Adnp-/- ESCs adopted morphology indistinguishable to

636 that of control ESCs, and maintain self-renewal capacity for more than 20 passages

637 that we tested.

638

639 Embryoid body formation. ESCs differentiation into embryoid body was performed

640 in attachment or suspension culture in medium lacking LIF or knockout serum

641 replacement (KSR), as described previously32.

642 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

643 ESC differentiation towards a neural fate. To induce ESC directional neural

644 differentiation in suspension, a neurosphere suspension culture system was used. For

645 Stage 1 of neural induction, cells were dissociated and suspended at a density of

646 1.5×106 cells in 10 cm2 dish coated with 5 ml 1% agarose. The dissociated cells were

647 culture with 45% Neurobasal medium (Life Technologies, 21103-049) and 45%

648 DMEM/F12 (BI, 01-172-1ACS) containing 1×B27 (Gibco, 17504044), 1× N2 (Gibco,

649 17502048) and 20 ng/ml bFGF (Peprotech, AF-100-18B) and 10 ng/ml EGF

650 (Peprotech, 315-09-1000). Cells were allowed to be cultured on a shaker with low

651 speed at 37 °C in 5% CO2 incubator. Half of the culture medium was changed every

652 two days. For Stage 2 of neuronal differentiation, the neurospheres were plated onto

653 gelatin-coated 6-well plates at day 9, and were cultured for additional 8-10 days in

654 DMEM/F12 containing 2.5% FBS and 1 μM retinoic acid (Sigma, R2625) .

655 ESCs were differentiated into cerebellar neuron cells using an approach adapted

656 from the previously described protocols24-27. The differentiation medium (SFEM) was

657 prepared as follows: DMEM supplemented with 5% KSR, 0.1 mM nonessential

658 amino acids (Gibco, 11140050), 1 mM sodium pyruvate, 2 mM L-Glutamine and 0.1

659 mM 2-mercaptoethanol. For generation of granule and Purkinje progenitor cells,

660 mouse ESCs were dissociated into single cells and suspended at a density of 5.0×104

661 cells per ml with SFEM. The cells were allowed to aggregate into neurospheres for 5

662 days. To induce MATH1+ cerebellar precursors, final concentration of 6.5 ng/ml

663 BMP4 (Peprotech, 120-05ET-10), 20 ng/ml WNT3a (Peprotech, 315-20-10) and 50

664 ng/ml Fgf8b (Peprotech, 100-25-25) were added to the medium from day 5-10. For bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

665 the cerebellar neuron induction, the medium was changed to the Neurobasal medium

666 supplemented with 1×B27 and cells were cultured from day 10 to day 20.

667 For the inhibition of the proteasome degradation, mESCs or neurospheres were

668 treated with 10 μM MG132 (MCE, HY-13259) for 4 hours before harvesting cells for

669 the next step. For rescue experiments by activating Wnt signaling, mESCs or

670 neurospheres were treated with 25 ng/ml Wnt3a (Peprotech, 315-20-10) or 3 μM

671 CHIR99021 (Selleck, S1263) from the indicated time points during ESC neural

672 induction.

673

674 Lentiviral transduction for shRNA knockdown. The shRNA plasmids for Adnp

675 (TRCN0000081670; TRCN0000081671), and the GFP control (RHS4459) were

676 purchased from Dharmacon (USA). To make lentivirus, shRNA plasmids and

677 Trans-lenti shRNA packaging plasmids were co-transfected into H293T cells

678 according to the kit manual (Open Biosystems, TLP4615). After determining the virus

679 titer, mESCs were transduced at a multiplicity of infection of 5:1. Puromycin

680 selection (1 μg/ml) for 4 days was applied to select cells with stable viral integration.

681 Quantitative PCR (qPCR) and Western blot were used to assess the knockdown of

682 Adnp.

683

684 Generation of Adnp-/- ESCs. Adnp-/- mESCs were generated by CRISPR/Cas9

685 technology. Briefly, we designed two gRNAs on exon 4 of Adnp gene by using the

686 online website http://crispr.mit.edu/. The gRNAs sequences are: gRNA 1: bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

687 5'-CCCTTCTCTTACGAAAAATCAGG-3'; gRNA 2:

688 5'-CTACTTGGTGCGCTGGAGTTTGG-3'. gRNAs were cloned into the pUC57-U6

689 expression vector with G418 resistance. The plasmids encoding gRNA and hCas9

690 were co-transfected into mESCs using Lipofectamine 2000 (Gibco, 11668019). After

691 48 hours, mESCs were selected with 500 μg/ml G418 for 7 days. Then the cells were

692 re-seeded on a 10-cm dish coated with 0.1% gelatin to form colonies. The single

693 colony was picked up, trypsinized and passaged at low density. DNA from single

694 colonies from the passaged cells was extracted and used for genotyping.

695

696 Generation of 3×Flag Tagged Adnp-/- mESC Cell Lines. The full-length Adnp

697 cDNA (NM_009628.3) was amplified by PCR and then cloned into pCMV-3×Flag

698 vector. The full-length Adnp cDNA sequence containing N-terminal 3×Flag sequence

699 was subcloned into pCAG-IRES-Puro vector. To make stable transgenic cells, Adnp-/-

700 mESCs were transfected with pCAG-IRES-Puro-3×Flag-Adnp vector using

701 Lipofectamine 2000 (Gibco, 11668019). 48 hours later, cells were selected by 1 μg/m

702 l puromycin. After 4-5 days drug selection, cells were expanded and passaged.

703 Western Blot assay was performed to confirm the transgenic cell line using a FLAG

704 antibody.

705 To make Tet-Express inducible transgenic cells, 3×Flag-Adnp or HA-Ctnnb1

706 were subcloned into pTRE3G or pLVX-tight-puro expression vector (Clontech,

707 PT5167-1 and PT3996-5) using In-fusion HD cloning system (Clontech, 638910).

708 Stable transgenic cell lines were established according to the manual of the bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

709 Tet-Express inducible expression systems (Clontech, 631169). Briefly, Adnp-/- ESCs

710 were transfected with 2 μg pTRE3G-3×Flag-Adnp or pTRE3G-HA-Ctnnb1 with

711 linear 100 ng puromycin marker using Lipofectamine 2000 transfection reagent. 96h

712 later, 1 μg/ml puromycin was added and drug selection was performed for two weeks

713 to establish the stable transgenic cell line. To induce target gene expression, 3×106

714 transgenic cells were plated in 6-well plates. The next day, the Tet-Express

715 transactivitor (Clontech, 631178) was added (3 μl Tet-Express to a final 100 μl total

716 volume according to the kit manual) for 1 h in serum-free medium to induce the target

717 gene expression. Then cells were allowed to grow in complete medium for an

718 additional 12–24 h before assay for the target protein induction. Western blot were

719 used to assess the target protein expression levels using a FLAG or HA antibodies. In

720 the absence of Tet-Express transactivitor, pTRE3G provides very low background

721 expression, whereas addition of Tet-Express proteins strongly transactivates target

722 genes.

723 FACS experiments. Neurospheres and SFEB cell aggregates were washed with

724 DPBS. The aggregates were dissociated into single cells with 0.25% trypsin (BI,

725 03-050-1A). After fixing with 4% paraformaldehyde at room temperature for 10

726 minutes, the cells were washed with 0.5% PBSA (0.5% BSA in PBS) and treated with

727 90% cold methanol for 15 minutes. After extensive washes, the cells were suspended

728 with 200 μl PBSA and filtered with the flow tube. Then the cells were incubated with

729 anti-NESTIN (Abcam, ab7659), anti-PAX6 (Proteintech, 12323-1-AP) antibody at

730 room temperature for 15 minutes. After three-times wash with 0.5% PBSA, the cells bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

731 were incubated with secondary antibodies (1: 500 dilution in PBSA, AlexaFluor 488)

732 at room temperature for 15 minutes in the dark. The cells were washed with PBSA for

733 two-times and resuspended with 400 μl PBSA, and were analyzed by the BD

734 AccuriC6 flow cytometer.

735

736 RNA preparation, RT-qPCR and RNA-seq. Total RNA from mESCs, neurospheres

737 and neuronal cells was extracted with a Total RNA kit (Omega, R6834-01). A total of

738 1 μg RNA was reverse transcribed into cDNA using the TransScript All-in-One

739 First-Strand cDNA synthesis Supermix (Transgen Biotech, China, AT341).

740 Quantitative real-time PCR (RT-qPCR) was performed using the TransStart® Tip

741 Green qPCR SuperMix (Transgen Biotech, China, AQ-141). The primers used for

742 RT-qPCR were listed in Table 1. All experiments were repeated for at least three times.

743 The relative gene expression levels were calculated based on the 2-∆∆Ct method.

744 Data were shown as means ± S.D. The Student’s t test was used for the statistical

745 analysis. The significance is indicated as follows: *, p < 0.05; **, p < 0.01; ***, p <

746 0.001.

747 For RNA-seq, ESCs and day 3-, day 4- and day 6-control and mutant

748 ESC-derived neurospheres were collected and treated with Trizol (Invitrogen). RNAs

749 were quantified by a Nanodrop instrument, and sent to BGI Shenzhen (Wuhan, China)

750 for making RNA-seq libraries and deep sequencing. For each time points, at least two

751 biological repeats were sequenced. DEGs were defined by FDR < 0.05 and a Log2

752 fold change >1.5 fold was deemed to be differentially expressed genes (DEGs). bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

753

754 Protein extraction and Western blot analysis. For protein extraction, ESCs cells

755 and neurospheres were harvested and lysed in TEN buffer (50 mM Tris-HCl, 150 mM

756 NaCl, 5 mM EDTA, 1% Triton X-100, 0.5% Na-Deoxycholate, supplement with

757 Roche cOmplete ™ Protease Inhibitor). The lysates were quantified by the Bradford

758 method and equal amount of proteins were loaded for Western blot assay. Antibodies

759 used for WB were ADNP (AF5919, R&D Systems), non-phospho (Active) β-Catenin

760 (Cell Signaling Technology CST, #8814), Phospho-β-Catenin (Ser33/37/Thr41) (CST,

761 #9561), total β-Catenin (CST, #9562), anti-NESTIN (Abcam, ab7659), anti-PAX6

762 (Proteintech, 12323-1-AP), anti-PCNA (Cusabio, CSB-PA01567A0Rb), anti-GSK3β

763 (Proteintech, 2204-1-AP), anti-pGSK3β (CST, 9336S), anti-PH3 (Santa Cruz,

764 sc-8656-R), anti-Caspas3 (BD, 559565), anti-L7 (Takara, M202), anti-PAX2 (Abcam,

765 ab79389), anti-TuJ1 (Abcam, ab78078), anti-GFAP (Abcam, ab53554), anti-HuC/D

766 (Life Technologies, A21271), anti-β-Catenin (Sigma, C7207), Anti-FLAG (F3165,

767 Sigma, 1:1000), anti-MYC antibody (Transgen Biotech, HT101) and anti-HA

768 (Abbkine, A02040, 1:1000). Briefly, the proteins were separated by 10% SDS-PAGE

769 and transferred to a PVDF membrane. After blocking with 5% (w/v) non-fat milk for

770 1 hour at room temperature, the membrane was incubated overnight at 4ºC with the

771 primary antibodies. Then the membranes were incubated with a HRP-conjugated goat

772 anti-rabbit IgG (GtxRb-003-DHRPX, ImmunoReagents, 1:5000), a HRP-linked

773 anti-mouse IgG (7076S, Cell Signaling Technology, 1:5000) for 1 hour at room

774 temperature. The GE ImageQuant LAS4000 mini luminescent image analyzer was bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

775 used for photographing.

776 Co-immunoprecipitation assay (Co-IP). Co-immunoprecipitations were performed

777 with the Dynabeads Protein G (Life Technologies, 10004D) according to the

778 manufacturer's instructions33. Briefly, 1.5 mg Dynabeads was conjugated with

779 antibodies or IgG overnight at 4 ºC. Antibodies were used are: 10 μg IgG (Proteintech,

780 B900610), or 10 μg anti-ADNP antibody (R&D Systems, A5919), or 10 μg

781 anti-FLAG antibody (Proteintech, 20543-1-AP), or 10 μg anti-HA antibody (Abbkine,

782 A02040) or 10 μg anti-MYC antibody (Transgen Biotech, HT101). The next day, total

783 cell lysates and antibody-conjugated Dynabeads were incubated overnight at 4 ºC

784 with shaking. After three-times wash with PBS containing 0.1% Tween, the beads

785 were boiled at 95 ºC for 5 minutes with the 6×Protein loading buffer and the

786 supernatant was collected for future WB analysis.

787 Luciferase reporter assays. TopFlash-luc reporter was kindly provided by Prof.

788 Zongbin Cui from Institute of Hydrobiology. HEK293T cells (about 1×105 cells) were

789 seeded in 24-well plates in DMEM medium containing 10% FBS (Transgen Biotech,

790 FS101-02). After 24 hour, the cells were transiently transfected with the indicated

791 luciferase reporters using Liposomal Transfection Reagent (Yeasen, 40802ES03). For

792 mESCs transfection, the Neon transfection system was used according to the manual.

793 Briefly, the trypsinized mESCs were resuspended in the electroporation buffer at a

794 density of 1×107cells/ml. The TopFlash-luc reporter and pTK-Renilla plasmids were

795 added into cells. Then electroporation was performed by the Neon transfection system

796 according to the manufacture’s instructions. After transfection, the cells were treated bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

797 with or without 10 ng/ml Wnt3a for 16 hours. The luciferase activity was measured

798 with Dual-luciferase Reporter Assay System (Promega, Madison, E1910). Renilla was

799 used as an internal control.

800 For luciferase reporter assays in zebrafish embryos, 1-cell stage wild type and

801 adnpa adnpb double mutant embryos were injected with 200 pg TopFlash and 20 pg

802 pTK-Renilla plasmids, and the late gastrula stage embryos were collected. Three

803 pools of 10 embryos each were lysed with passive lysis buffer and assayed for

804 luciferase activity according to the Dual luciferase system (Promega). The Topflash

805 luciferase activities were measured with the Dual-luciferase Reporter Assay System.

806 All luciferase reporter assays represent the mean ± standard error from 3 independent

807 measurements of pools. At least three independent experiments were performed in

808 different batches of embryos.

809 Immunofluorescence assay. Neurospheres and SFEB aggregates were collected and

810 fixed with 4% paraformaldehyde for half an hour at room temperature. Then the cells

811 were washed with PBST (phosphate-buffered saline, 0.1% Triton X100) for three

812 times, each for 15 minutes. Following the incubation with blocking buffer (5% normal

813 horse serum, 0.1% Triton X-100, in PBS) for 2 hours at room temperature, the cells

814 were incubated with primary antibodies at 4 ºC overnight. After three-times wash with

815 PBST, the cells were incubated with secondary antibodies (1: 500 dilution in blocking

816 buffer, Alexa Fluor 488, Life Technologies) at room temperature for 1 hour in the dark.

817 The nuclei were counter-stained with DAPI (Sigma, D9542, 1:1000). After washing

818 with PBS for twice, the slides were mounted with 100% glycerol on histological bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

819 slides. Images were taken by a Leica SP8 laser scanning confocal microscope

820 (Wetzlar, Germany).

821

822 Immunohistochemistry. Ten-week-old mice brain frozen sections were pretreated

823 with an antigen unmasking solution (10mM Sodium Citrate, pH6.0) for 15min at 95

824 ºC in oven. After that, sections were naturally cooled to room temperature and

825 blocked with 3% normal donkey serum adding 0.1% Tween-20 at room temperature

826 for 2h. The tissue was then incubated overnight at 4 ºC with the anti-ADNP antibody

827 diluted in blocking buffer, followed by avidin-biotin-peroxidase complex (1:50; ABC

828 kit, VECTASTAIN Elite ABC system, Vector Labs, Burlingame, CA, USA).

829 Peroxidase was reacted in 0.03% 3,3'-diaminobenzidine (DAB) and 0.003% H2O2 in

830 PBS. Sections were dehydrated, cleared in xylene, and mounted in neutral balsam.

831 Measuring and quantification of IHC images were performed using the Image-pro

832 Plus 6.0 software (Media Cybernetics).

833 Immunoprecipitation in combination with mass spectrometry. For Mass

834 Spectrometry, the IP samples (immunoprecipitated by IgG or ADNP antibody) were

835 run on SDS-PAGE gels and stained with Coomassie Blue. Then the entire lanes for

836 each IP samples were cut off and transferred into a 15 ml tube containing deionized

837 water. The treatment of the samples and the Mass Spectrometry analysis were done by

838 GeneCreate Biological Engineering Company (Wuhan, China).

839 Protein-protein interaction assay using a rabbit reticulocyte lysate system. bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

840 Protein-protein interaction assay using a rabbit reticulocyte lysate system has been

841 described previously33. Tagged-ADNP, Tagged-ADNP mutants, Tagged-β-Catenin and

842 Tagged-β-Catenin mutants were synthesized using the TNT coupled reticulocyte

843 lysate system according to the manual (Promega, L5020, USA). Briefly, 1 μg of

844 circular PCS2- version of plasmids were added directly to the TNT lysates and

845 incubated for 1.5 hours at 30 ºC. 1 μl of the reaction products were subjected to WB

846 assay to evaluate the synthesized protein. For protein-protein interaction assay, 5-10

847 μl of the synthesized HA or FLAG tagged proteins were mixed in a 1.5 ml tube

848 loaded with the 300 μl TEN buffer, and the mixture was shaken for 30 minutes at

849 room temperature. Next, IP or pull-down assay was performed using Dynabeads

850 protein G coupled with anti-FLAG or anti-HA antibodies as described above.

851

852 Zebrafifish Maintenance. Zebrafish (Danio rerio) were maintained at 28.5 ºC on a

853 12 h light/12 h dark cycle. All procedures were performed with the approval of the

854 Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China.

855 Generation zebrafish adnp mutants by CRISPR/Cas9. Zebrafish adnp mutants

856 were generated by CRISPR/Cas9 technology. The gRNA targeting sequences for

857 adnpa and adnpb gene were 5’- GGACTCTGGAAACCCACGTC -3’ and 5’-

858 GGAGGACTTCATGGGGCAG - 3’, respectively. Guide RNAs (gRNAs) were

859 generated using the MEGAshortscript T7 kit (Thermo Fisher). The Cas9 mRNA was

860 synthesized using the mMESSAGE mMACHINE® SP6 Kit (AM1340, Thermo

861 Fisher). The mixture containing 200 ng/µl Cas9 mRNA and 80 ng/µl guide RNA was bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

862 co-injected into 1-cell stage zebrafish embryos. The genomic DNA of 20 injected

863 embryos at 24 hpf was extracted and subjected to PCR amplification. The DNA

864 fragments containing the gRNA targeting sequences were amplified by PCR using

865 primers 5’- TGCTCAGATTGCCCGTTT -3’ and 5’-

866 GATAGGTGCACTTCTTGCAGTA -3’ for adnpa gene,

867 5’-AATGTGCACAGCGAGGACTT-3’ and 5’- GACAAGGACTGTGTAGCCCC -3’

868 for adnpb gene, respectively. The genotype was confirmed by DNA sequencing.

869 Adults raised up from injected embryos were screened for mosaic founders by the

870 amplicon sequencing. The mosaic founders were outcrossed to wild type to obtain the

871 F1 offsprings with stable germline transmission. The F1 heterozygous carrying 7 bp

872 deletion (adnpa) and 5 bp deletion (adnpb) were outcrossed to wild type to generate

873 F2 heterozygous, respectively. The F2 heterozygous were inter-crossed to generate

874 homozygous adnpa-/- and adnpb-/-, respectively. The adnpa-/- and adnpb-/- adults were

875 outcrossed to each other to generate adnpa and adnpb double mutant. The subsequent

876 screening was performed by high resolution melt analysis (HRMA) using primers

877 5’-TTCCGCAATGTTCACAGGGA-3’ and 5’- GGCATATGGAAGAGCCTGACG-3’

878 for adnpa, 5’- CACATCAGGTTGTTTCATATGCCT -3’ and 5’-

879 TCACGTGTCTTTTCCAGACGA -3’ for adnpb, respectively.

880 Microinjections into zebrafish embryos. For overexpression study, full-length

881 cDNA encoding Adnpa (NC_007122.7) was amplified and cloned into the PCS2 (+)

882 vector. The primers we used are as followed

883 GATGTTTCAGCTTCCAGTGAATAACC (Forward), and bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

884 CACAAGCCATCATCTACCAAGC (Reverse). The Adnpa mRNA was transcribed

885 with mMESSAGE mMACHINE SP6 Kit (Invitrogen, AM1340) and was dissolved in

886 nuclease-free water to the final concentration of 150 ng/μl. A total of 1-5 nl mRNAs

887 were microinjected into one-cell stage wild-type zebrafish embryo using the WPI

888 microinjector (WPI, USA). The 1-2 dpf embryos were anesthetized in 0.016% tricaine

889 and photographed under a Leica stereomicroscope.

890 Whole-mount immunofluorescence. The whole-mount immunofluorescence

891 experiments were performed as described33. Briefly, 48 hpf zebrafish embryos were

892 fixed in 4% paraformaldehyde overnight at 4 °C and then underwent proteinase K

893 treatment for 8 minutes at RT. The primary antibody we used was anti-HuC/HuD

894 monoclonal antibody (Life Technologies, USA, 1: 350 dilution in PBST). After

895 blocking for 2 h at room temperature with solution (0.1% Triton X-100, 1% BSA, and

896 1% DMSO in PBS), the embryos were incubated overnight with the primary

897 antibodies at 4 °C with shaking. After three-times wash with PBST, embryos were

898 incubated with the secondary antibodies (1:1000, Goat anti-Mouse IgG secondary

899 antibody, Alexa Fluor 488, Life technologies, USA) for 2 h at RT. Finally, the

900 embryos were counter-stained with DAPI. Embryos were imaged in 4% methyl-

901 cellulose using a Leica SP8 Confocal microscope.

902 Whole-mount in situ hybridization (WISH). WISH was performed as described

903 previously33. For making the RNA probes, total RNA was extracted from zebrafish

904 embryos, and cDNAs were made by RT-PCR. Templates were amplified by PCR bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

905 using primers with SP6 or T7 promoters. Primers we used are listed in Table 3. The

906 DIG-labeled probes for adnpa, adnpb, dlx5a, neurod1 and phox2a were made using

907 the DIG RNA Labeling Kit (SP6/T7) (Roche). Zebrafish embryos at different

908 developmental stages were collected and fixed with 4% PFA overnight at 4 °C.

909 Following the WISH, the embryos were transferred to 6-well plates and submerged in

910 100% glycerol for imaging.

911 Quantification and statistical analysis. Data are presented as mean values ±SD

912 unless otherwise stated. Data were analyzed using Student’s t-test analysis. Error bars

913 represent s.e.m. Differences in means were statistically significance when p< 0.05.

914 Significance levels are: *p< 0.05; **P<0.01.

915

916 Data availability. All data will be available upon request.

917

918 References

919 1. Gozes, I., Yeheskel, A, and Pasmanik-Chor, M. Activity-Dependent

920 Neuroprotective Protein (ADNP): A Case Study for Highly Conserved

921 Chordata-Specific Genes Shaping the Brain and Mutated in Cancer. J. of

922 Alzheimer’s Disease. 45, 57-73 (2015)

923 2. Mandel, S. and Gozes, I. Activity-dependent neuroprotective protein constitutes a

924 novel element in the SWI/SNF chromatin remodeling complex. J. Biol. Chem.

925 282, 34448-34456 (2007)

926 3. Mosch, K. et al. HP1 Recruits Activity-Dependent Neuroprotective Protein to

927 H3K9me3 Marked Pericentromeric Heterochromatin for Silencing of Major bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

928 Satellite Repeats. PLoS ONE. 6(1): e15894. doi:10.1371/journal.pone.0015894

929 (2011)

930 4. Ostapcuk, V. et al. Activity-dependent neuroprotective protein recruits HP1 and

931 CHD4 to control lineage-specifying genes. Nature. 557, 739-743 (2018)

932 5. Pinhasov, A. et al. Activity-dependent neuroprotective protein: A novel gene

933 essential for brain formation. Brain Res. Dev. 144, 83-90 (2003)

934 6. Helsmoortel, C. et al. A SWI/SNF-related autism syndrome caused by de novo

935 mutations in ADNP. Nat. Genet. 46(4), 380-384 (2014)

936 7. Vandeweyer, G. et al. The transcriptional regulator ADNP links the BAF

937 (SWI/SNF) complexes with Autism. Am. J. Med. Genet. C. Semin. Med. Genet.

938 166 (3), 315–326 (2014)

939 8. Vulih-Shultzman, I. et al. Activity-dependent neuroprotective protein snippet

940 NAP reduces tau hyperphosphorylation and enhances learning in a novel

941 transgenic mouse model. J Pharmacol Exp Ther. 323(2), 438–449 (2007)

942 9. Ying, Q. L. et al. Conversion of embryonic stem cells into neuroectodermal

943 precursors in adherent monoculture. Nat. Biotechnol. 21, 183-186 (2003)

944 10. Stavridis, M. P., Lunn, J. S., Collins, B. J. and Storey, K. G. A discrete period of

945 FGF-induced Erk1/2 signalling is required for vertebrate neural specification.

946 Development 134, 2889 -2894 (2007).

947 11. Bressan, R. B. et al. EGF–FGF2 stimulates the proliferation and improves the

948 neuronal commitment of mouse epidermal neural crest stem cells (EPI-NCSCs).

949 Exp. Cell Res. 327, 37-4 7 (2014) bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

950 12. Conti, L. et al. Niche-independent symmetrical self-renewal of a mammalian

951 tissue stem cell. PLoS. Biol. 3(9), e283 (2005)

952 13. Chen, et al. The emerging picture of autism spectrum disorder: genetics and

953 pathology. Annu. Rev. Pathol. Mech. Dis. 10, 111-44 (2015).

954 14. Xu, W. Q. and Kimelman, D. Mechanistic insights from structural studies of

955 beta-catenin and its binding partners. J. of Cell Sci. 120, 3337-3344 (2007).

956 15. Divinski, I., Mittelman, L., Gozes, I. A femtomolar acting octapeptide interacts

957 with tubulin and protects astrocytes against zinc intoxication. J. of Biol. Chem.

958 279, 28531–28538 (2004)

959 16. Gozes, I, et al. NAP: research and development of a peptide derived from

960 activity-dependent neuroprotective protein (ADNP). CNS Drug Rev. 11(4),

961 353–368 (2005)

962 17. Hacohen-Kleiman, G. et al. Activity-dependent neuroprotective protein

963 deficiency models synaptic and developmental phenotypes of autism-like

964 syndrome. J Clin Invest 128, 4956-4969 (201 8)

965 18. Bocchi, R. et al. Perturbed Wnt signaling leads to neuronal migration delay,

966 altered interhemispheric connections and impaired social behavior. Nat. Commun.

967 8, 1158-1162 (2018)

968 19. Xing, Y. et al. Crystal Structure of a Catenin/APC Complex Reveals a Critical

969 Role for APC Phosphorylation in APC Function. Mol. Cell 15, 523–533 (2004) bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

970 20. Bassan, M., Zamostiano, R., Davidson, A. et al. Complete sequence of a novel

971 protein containing a femtomolaractivity-dependent neuroprotective peptide. J.

972 Neurochem. 72, 1283–1293 (1999).

973 21. Zamostiano, R., Pinhasov, A., Gelber, E. et al. Cloning and characterization of the

974 human activity dependent neuroprotective protein. J. Biol. Chem. 276, 708–714

975 (2001).

976 22. Fatemi, H. S. et al. Consensus Paper: Pathological Role of the Cerebellum in

977 Autism. Cerebellum 11, 777–807 (2012)

978 23. Hampson, D. R. and Blatt G. J. Autism spectrum disorders and neuropathology of

979 the cerebellum. Front. Neurosci. 9, 420-436 (2015)

980 24. Higuera, G. A. et al. An expandable embryonic stem cell-derived Purkinje neuron

981 progenitor population that exhibits in vivo maturation in the adult mouse

982 cerebellum. Scientific Reports. 7, 8863-8882 (2017)

983 25. Su, H. L. et al. Generation of cerebellar neuron precursors from embryonic stem

984 cells. Dev. Biol. 290, 287 – 296 (2006)

985 26. Tao, O. et al. Effificient Generation of Mature Cerebellar Purkinje Cells From

986 Mouse Embryonic Stem Cells. J. of Neuroscience Res. 88, 234–247 (2010)

987 27. Wang, S.Y. et al. Differentiation of human induced pluripotent stem cells to

988 mature functional Purkinje neurons. Scientific Reports 5 : 9232 (2015)

989 28. Sugathana, A. et al. CHD8 regulates neurodevelopmental pathwaysassociated

990 with autism spectrum disorder in neural progenitors. PNAS 111, 4468-4477

991 (2014) bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

992 29. Bernier, R. et al. Disruptive CHD8 Mutations Defifine a Subtype of Autism Early

993 in Development. Cell 158, 263–276 (2014)

994 30. Dresner, E. et al. Novel Evolutionary-conserved Role for the Activity-dependent

995 Neuroprotective Protein (ADNP) Family That Is Important for Erythropoiesis. J.

996 of Biol. Chem. 287, 40173–40185 (2012)

997 31. Ying, Q. L. et al. The ground state of embryonic stem cell self-renewal. Nature

998 453: 519–523 (2008)

999 32. Chappell,J., Sun, Y.H., Singh, A., Dalton, S. MYC/MAX control ERK signaling

1000 and pluripotency by regulation of dual-specificity phosphatases 2 and 7. Genes &

1001 Dev. 27, 725–733 (2013)

1002 33. Sun, X.Y. et al. Mga Modulates Bmpr1a Activity by Antagonizing Bs69 in

1003 Zebrafish. Front. Cell Dev. Biol. 6, 126-131 (2018)

1004 Acknowledgements. This work was supported by National Key Research and

1005 Development Program (2016YFA0101100), National Natural Science Foundation of

1006 China (31671526), and Hundred-Talent Program (CAS and IHB) (Y623041501) to

1007 YH Sun.

1008 Author contributions. XY Sun performed the stem cell part of work; XX Peng and

1009 YQ Cao performed the zebrafish part of work; Y Zhou and Yuhua Sun designed the

1010 work; Yuhua Sun provided the final support of the work, and wrote the paper.

1011 Competing interests. The authors declared no competing interests.

1012 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1013 Figure legends

1014 Fig. 1 ADNP depletion leads to loss of ESC phenotype. a Cartoon depicting the

1015 gRNA target sites at exon 4 of mouse Adnp gene. b Genotyping showing the mutant

1016 alleles. c Western blot analysis of ADNP levels in control, shRNA knockdown and

1017 Adnp-/- ESCs. WB has been repeated for three times. d Representative image

1018 showing morphology of control and early passaged Adnp-/- ESCs. e The mRNA

1019 expression levels of representative pluripotency-related, mesodermal,

1020 neuroectodermal, endodermal genes in control and early passaged Adnp-/- ESCs.

1021 qPCR has been repeated for at least three times. f Representative image showing

1022 morphology of embryoid bodies (EBs) at indicated time points. g Heat-map of DEGs

1023 of indicated lineage-specific genes for control and Adnp-/- ESC-derived day 6 EBs,

1024 based on three RNA-seq replicates. h qPCR analysis showing the dynamic expression

1025 of lineage-specifying genes during EB formation of control and Adnp-/- ESCs. qPCR

1026 experiments have been done for three times.

1027 Supplementary Fig. 1 Related to Figure 1. a IF staining of OCT4 for control and

1028 Adnp-/- ESCs. b Alkaline phosphotase staining for long-term passaged control and

1029 Adnp-/- ESCs (shown is successfully passaged for 10 times in LIF/KSR medium). c

1030 Relative Adnp mRNA levels for Gfp shRNA and Adnp shRNA knockown ESCs. d

1031 Representative image showing morphology of control and early passaged Adnp

1032 shRNA ESCs. e The expression of representative pluripotency-related, mesodermal,

1033 neuroectodermal, endodermal genes in control and early passaged Adnp shRNA ESCs.

1034 qPCR has been repeated for at least three times and represented as mean values± s.d. bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1035 Fig. 2 ADNP is required for proper ESC neural differentiation. a Cartoon showing the

1036 2-Stage ESC neural differentiation protocol. At Stage 1, ESCs were cultured in

1037 suspension to form neurospheres with indicated growth factors; at Stage 2, the

1038 neurospheres were plated onto 0.1% gelatin-coated plates to allow further

1039 differentiation for an additional 7-14 days. b Representative image showing

1040 morphology of day 6 and day 20 control and Adnp-/- ESC-derived neurospheres

1041 neuronal cultures. The white arrow showing the fibre-like neuronal structures. c The

1042 dynamic expression profile of representative neuroectoderm genes during control and

1043 Adnp-/- ESC neural induction. Experiments were repeated for three times. d Flow

1044 cytometry analysis for quantification of PAX6+ cells. Red: isotype control; purple:

1045 experimental group using PAX6 antibody. e IF staining of neural progenitor markers

1046 NESTIN and ADNP for control, Adnp-/- ESC and FLAG-ADNP restoring Adnp-/-

1047 ESC-derived day 6 neurospheres. f qRT-PCR analysis for the expression of Tubb3

1048 (encoding TuJ1) and Gfap for day 19 control and Adnp-/- ESC-derived neuronal cell

1049 cultures. g IF staining of neuronal marker TuJ1 and glial marker GFAP for day 19

1050 control and Adnp-/- ESC-derived neuronal cell cultures. The white arrows showing

1051 the neuronal fibre structures. h WB analysis of TuJ1 and GFAP levels in day 19

1052 control and Adnp-/- ESC-derived neuronal cell types. All experiments were repeated

1053 for three times, and shown are the representative data.

1054 Supplementary Fig. 2 Related to Figure 2. a The expression of representative neural

1055 genes at indicated time points. b IF staining of PAX6 for day 6 control, Adnp-/- and bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1056 FLAG-ADNP restoring Adnp-/- ESC-derived neurospheres. c IF staining of TuJ1 and

1057 GFAP for day 16 control and Adnp-/- ESC-derived neuronal cell cultures.

1058 Fig. 3 ADNP promotes the expression of neuroectoderm developmental genes. a

1059 Schematic representation of design of RNA-seq experiments; b Plot showing DEGs

1060 from day 3 and day 6 control and Adnp-/- ESC-derived neurospheres; RNA-seq for

1061 each time points were repeated for at least two times. The numbers of up- and

1062 down-regulated genes were shown as average. c Heatmap illustrating the expression

1063 of selected neuroectoderm and pluripotency-related genes that were shown as log2

1064 FPKM in day 6 control and Adnp-/- ESC-derived neurospheres. Each lane

1065 corresponds to an independent biological RNA-seq sample. d Representative

1066 morphology of day 6 Adnp-/- ESC-derived neurospheres after addition of Tet-Express

1067 proteins from the indicated time points. Tet stands for Tet-Express. Experiments were

1068 repeated for two times. e IF staining of PAX6 for day 6 Adnp-/- ESC-derived

1069 neurospheres after addition of Tet-Express proteins at the indicated time points. f

1070 Quantitative analysis of PAX6+ cells based on panel (e). All IF staining experiments

1071 were repeated for at least two times, and shown are the representative data.

1072 Supplementary Fig. 3 Related to Figure 3. a Dynamic expression profile of ADNP

1073 during first 6 days of neural induction of ESCs. Experiments were repeated for two

1074 times. b Dynamic expression profile of Adnp, Pax6, Nestin and Olig2 mRNAs during

1075 the first 6 days of neural induction of wild type ESCs. Experiments were repeated for

1076 three times. c Heatmap illustrating the expression of selected neuroectoderm genes

1077 that were shown as log2 FPKM in day 3 control and Adnp-/- ESC-derived bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1078 neurospheres. Each lane corresponds to an independent biological RNA-seq sample. d

1079 WB showing that addition of Tet-Express transactivated 3×FLAG-ADNP in Adnp-/-

1080 ESCs. e Representative morphology showing the neuronal fibre structure in Adnp-/-

1081 ESC-derived day 19 neuronal cell types that added with Tet-Express from day 0, day

1082 2 and day 4 of Adnp-/- ESC neural induction. Experiments were repeated for three

1083 times, and shown are the representative images.

1084 Fig.4 Wnt signaling is impaired in the absence of ADNP. a Dissection of KEGG data

1085 of DEGs from day 3 and day 6 control and Adnp-/- ESC-derived neurospheres,

1086 showing enrichment of the Wnt signaling pathway. b Heatmap illustrating the

1087 expression of selected Wnt-related genes that were shown as log2 FPKM in day 6

1088 control and Adnp-/- ESC-derived neurospheres. Each lane corresponds to an

1089 independent biological RNA-seq sample. c qRT-PCR analysis for the indicated Wnt

1090 target genes for day3 and day9 control and Adnp-/- ESC-derived neurospheres. The

1091 experiments were repeated for three times. d Top-Flash luciferase activity assay for

1092 lysates from day 3 control and Adnp-/- ESC-derived neurospheres, in the absence or

1093 presence of Wnt3a. Experiments were repeated for two times. e Rescue of the

1094 expression of the indicated neural developmental genes and putative Wnt target genes

1095 by addition of CHIR and Wnt3a, or by restoring 3×FLAG-ADNP. Genes in red box

1096 are representative neural developmental genes, and genes in blue box are

1097 representative Wnt target genes. The experiments were repeated for three times. f

1098 Rescue of NESTIN expression by addition of CHIR. Representative IF staining of

1099 NESTIN for day 6 control and Adnp-/- ESC-derived neurospheres. bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1100 Supplementary Fig. 4 Related to Figure 4. a KEGG analysis of of DEGs from day 3

1101 and day 6 control and Adnp-/- ESC-derived neurospheres, showing enrichment of

1102 pathway related to signaling transduction. b Heatmap illustrating the expression of

1103 selected Wnt-related genes that were shown as log2 FPKM in day 3 control and

1104 Adnp-/- ESC-derived neurospheres. Each lane corresponds to an independent

1105 biological RNA-seq sample. c TopFlash luciferase activity analysis of 293T cells

1106 transfected with increasing dose of plasmids encoding ADNP in the absence or

1107 presence of Wnt3a. Experiments were repeated for two times. d Rescue of PAX6

1108 expression by addition of CHIR. Representative IF staining showing PAX6 signal for

1109 day 6 control and Adnp-/- ESC-derived neurospheres. e Representative morphology of

1110 day 6 Adnp-/- ESC-derived neurospheres after addition of CHIR from the indicated

1111 time points. Experiments were repeated for two times.

1112 Fig. 5 Identifying β-Catenin as a ADNP interacting protein. a Schematic

1113 representation showing the experimental design of IP in combination mass

1114 spectrometry assay (left) and a list of representative ADNP interacting proteins (right).

1115 b Co-IP data for endogenous ADNP and β-Catenin in ESCs. Up: IP ADNP followed

1116 by WB β-Catenin; bottom: IP β-Catenin followed by WB ADNP. c Co-IP data for

1117 endogenous ADNP and β-Catenin of day 3 ESC-derived neurospheres. d Co-IP data

1118 for FLAG-ADNP and β-Catenin in 3×FLAG-ADNP overexpressing Adnp-/- ESCs. e

1119 Schematic representation of the full-length and the truncated ADNP mutants. f IP of

1120 FLAG-ADNP-Nter (1-685) and MYC-β-Catenin in 293T cells. g IP of in vitro

1121 synthesized HA-β-Catenin and FLAG-ADNP. h Co-IP of FLAG-ADNP-NAP and bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1122 HA-β-Catenin in 293T cells. i Schematic representation of the full-length and the

1123 truncated β-Catenin mutants. j Co-IP of FLAG-β-Catenin (151-666) and ADNP in

1124 293T cells. k Colocalization of ADNP and β-Catenin in day 19 ESC-derived neuronal

1125 cell cultures. All WB experiments have been repeated for three times, and shown are

1126 the representative images.

1127 Supplementary Fig. 5 Related to Figure 5. a WB showing that the C-terminal of

1128 ADNP (686-1108) does not interact with β-Catenin in 293T cells. b Dissection of

1129 FLAG-ADNP-Nter (1-685) that is responsible to interact with β-Catenin in 293T cells.

1130 ADNP fragment 1-224, 446-685 both can interact with β-Catenin. And ADNP-Nter

1131 (1-685) without NAP can still interact with β-Catenin. c WB showing that β-Catenin

1132 (1-150) barely interacts with ADNP in 293T cells. d IF immunostaining of ADNP

1133 showing that during ESC neural differentiation ADNP translocates from nuclei to

1134 cytoplasm, and that ADNP was colocalized with NESTIN and TuJ1 in day 6

1135 ESC-derived neurospheres and day 19 neuronal cell types, respectively. All WB and

1136 IF staining experiments have been repeated for at least two times.

1137 Fig. 6 ADNP stabilizes β-Catenin during ESC neural differentiation. a WB showing

1138 total β-Catenin levels in day 3 control and Adnp-/- ESC-derived neurospheres. b WB

1139 showing β-Catenin levels in the cytoplasmic and the nuclear fraction of lysates from

1140 day 3 control and Adnp-/- ESC-derived neurospheres. Note that β-Catenin levels were

1141 reduced in the cytoplasmic but not in the nuclear fraction of lysates. c Time course

1142 WB assay showing total and phosphorylated β-Catenin levels during control and

1143 mutant ESC differentiation towards a neural fate. d Representative IF staining of bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1144 β-Catenin in day 19 ESC-derived neuronal cultures; e Representative WB showing

1145 β-Catenin levels in the combination treatment of CHIR and MG132. f Representative

1146 WB showing the ubiquitylation levels of β-Catenin in control and Adnp-/- ESCs. g

1147 Representative WB showing the expression of the indicated proteins in 293T cells that

1148 transfected with an increasing dose of ADNP. h Representative IF staining of

1149 NESTIN showing the rescue of NESTIN in day 6 Tet-Express inducible HA-β-catenin

1150 Adnp-/- ESC derived neurospheres by the addition of Tet-Express every day at the

1151 early stage of neural induction .

1152 Supplementary Fig. 6 Related to Figure 6. a Representative WB showing the total

1153 β-Catenin levels in control and Adnp-/- ESCs. b WB showing the total β-Catenin

1154 levels in day 3 control and Adnp shRNA knockdown ESC-derived neurospheres. c

1155 Representative WB showing the ubiquitylation levels of β-Catenin in day 3 control,

1156 Adnp-/- and FLAG-ADNP overexpressing ESC-derived neurospheres. d

1157 Representative WB showing the effects of an increasing dose of ADNP on the

1158 co-transfected HA-β-Catenin and total β-Catenin levels. e Representative WB

1159 showing the effect of ADNP depletion on GSK3β and phosphorylated GSK3β levels.

1160 All WB experiments were repeated for two times.

1161 Fig. 7 ADNP stabilizes β-Catenin by preventing the formation of degradation

1162 complex. a Cartoon showing the armadillo domain of β-Catenin. Arm1 contains

1163 armadillo repeat 1-4 and Arm2 contains repeat 5-9. Known core binding armadillo

1164 repeats of β-Catenin for Axin and APC were indicated with red and green bars,

1165 respectively. b IP experiment showing that the Arm1 fragment can interact with bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1166 ADNP in 293T cells. c IP experiment showing that the Arm2 fragment can interact

1167 with ADNP in 293T cells. d The relationship between ADNP – β-Catenin and

1168 β-Catenin – Axin interactions were examined by performing competitive

1169 protein-binding assay. An increasing dose of plasmids encoding FLAG-ADNP and

1170 plasmids encoding MYC-AXIN1 and HA-β-Catenin were co-transfected into 293T

1171 cells. Lysates were extracted from the transfected cells and pull-downed by HA

1172 antibody followed by WB using MYC and HA antibodies. WB showing that the

1173 amount of Axin co-immunoprecipitated with β-Catenin became reduced by the

1174 increased addition of ADNP. e The competitive protein-binding assay showing that

1175 the amount of APC co-immunoprecipitated with β-Catenin became reduced by the

1176 increased addition of ADNP. f WB showing the expression of HA-β-Catenin by the

1177 addition of the Tet-Express in the inducible HA-β-Catenin Adnp-/- ESCs. g IF

1178 staining showing the rescue of TuJ1 signal in day 19 Adnp-/- ESC-derived neuronal

1179 cell cultures by adding Tet-Express transactivitor every day at early stage of neural

1180 induction. All WB experiments were repeated for at least two times.

1181 Fig. 8. ADNP plays key role in cerebellar neuron development. a Schematic method

1182 for ESC differentiation into cerebellar neurons. b Representative morphology of day 7

1183 control and Adnp-/- ESC-derived neurospheres. c qPCR showing the expression of

1184 neural developmental genes of day 6 control and Adnp-/- ESC-derived neurospheres.

1185 d qPCR showing the expression of EGL precursor markers of day 10 control and

1186 Adnp-/- ESC-derived EGL precursors. e IF double-staining of MATH1 and β-Catenin

1187 for day 10 control and Adnp-/- ESC-derived EGL precursors. f IF double-staining of bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1188 L7 and HuC/D for day 20 control and Adnp-/- ESC-derived cerebellar neurons. All

1189 experiments were repeated for 3 times.

1190 Supplementary Fig. 7. Related to Figure 8. a IHC image showing the expression of

1191 ADNP in 10-week old mouse brain. a, olfactory bulb, b, cerebellum, c, hippocampus, d,

1192 cortex. b IF staining of NESTIN for day 6 control and Adnp-/- ESC-derived SFEB

1193 aggregates. c The expression of the indicated neural genes in day 10 control and

1194 Adnp-/- ESC-derived EGL neural progenitors. d IF double-staining of L7 and HuC/D

1195 for day 18 control and Adnp-/- ESC-derived cerebellar cell types.

1196 Fig. 9 ADNP regulates the proliferation of ESC-derived neural progenitors. a Cell

1197 mass comparison of day 6 control and Adnp-/- ESC-derived neurospheres; b

1198 Immunostaining of proliferation marker phosphohistone H3 (PH3) for day 6 control

1199 and Adnp-/- ESC derived neurospheres; c Quantification of PH3 positive cells based

1200 on panel (b). d Representative WB data showing PH3 levels during early stage of

1201 control and Adnp-/- ESC neural differentiation. e Representative WB data showing

1202 PH3 levels for day 6 control and Adnp-/- ESC-derived neurospheres; f WB assay of

1203 Caspase-3 for day 6 control and Adnp-/- ESCs-derived neurospheres. g The

1204 morphology of day 6 neurospheres under the indicated conditions (rescue conditions).

1205 Fig. 10. Loss of adnp leads to reduced Wnt signaling and defective neural

1206 development in zebrafish embryos. a CRISPR gRNA design and genotyping of adnpa

1207 mutant; b gRNA design and genotyping of adnpb mutant; c Representative

1208 morphology of 2 dpf control and adnpa-/- adnpa-/- double mutants. Black arrow

1209 showing the obvious cell death (black area) in head region of mutant embryos. d 3D bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

1210 Z-stack image showing the reduced expression of HuC/D in 2 dpf control and double

1211 mutant zebrafish brain. Arrow show the distance between the convex tip of the eye

1212 cups. e GFP fluorescence signal showing the reduced elavl3-GFP signal in adnpa

1213 adnpb deficient embryos. f In situ hybridization images for the indicated neural

1214 markers for 2 dpf control and double mutant embryos. g TopFlash luciferase activity

1215 assay for total lysates made from 8 hpf control, double mutant and adnpa mRNA

1216 overexpressing embryos. h The expression of Wnt target genes in 8 hpf control,

1217 double mutant and adnpa mRNA injected mutant embryos.

1218 Supplementary Fig. 8. Related to Figure 10. a In situ hybridization showing the

1219 expression of adnpa and adnpb in embryos at different developmental stages. b In situ

1220 hybridization showing adnpa expression in head region of embryos at different

1221 developmental stages. Dorsal view of head region. c Representative morphology of 1

1222 dpf control and adnpa-/- adnpb-/- zebrafish embryos. d Representative morphology of

1223 1 dpf adnpa-/- adnpb-/- embryos. The ventralized phenotypes (V1-V3) were

1224 according to the DV patterning index33. e Representative morphology of 1 dpf adnpa

1225 overexpressing embryos. The dorsalized phenotypes (C1-C4) were according to the

1226 DV patterning index33

1227

1228 Table 1 The primers for RT-qPCR analysis Mouse genes Forword (5'-3') Reverse (5'-3') β-actin AGAGGGAAATCGTGCGTGAC CAATAGTGATGACCTGGCCGT Adnp ACGAAAAATCAGGACTATCGG GGACATTCCGGAAAGACTTT Pax6 AGTGAATGGGCGGAGTTATG ACTTGGACGGGAACTGACAC Nestin CCCTGAAGTCGAGGAGCTG CCCTGAAGTCGAGGAGCTG Pax2 AAGCCCGGAGTGATTGGTG CAGGCGAACATAGTCGGGTT

Pax3 CCGGGGCAGAATTACCCAC GCCGTTGATAAATACTCCT CCG bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

TuJ1 TAGACCCCAGCGGCAACTAT GTTCCAGGTTCCAAGTCCACC Gfap GAAAACCGCATCACCATTC TTGTGACTTTTTGGCCTTCC Ccnd1 CAGAGGCGGATGAGAACAAGT GCGGTAGCAGGAGAGGAAG Axin2 CTCCCCACCTTGAATGAAGA TGGCTGGTGCAAAGACATAG Myc GTTGGAAGAGCCGTGTGTG CGCTGATGTTGGGTCAGTC En2 ATGGGACATTGGACACTTCTTC CCCACAGACCAAATAGGAGCTA Math1 GAGTGGGCTGAGGTAAAAGAG GGTCGGTGCTATCCAGGAG T Olig2 TCCCCAGAACCCGATGATCTT CGTGGACGAGGACACAGTC Wnt1 GGTTTCTACTACGTTGCTACTG GGAATCCGTCAACAGGTTCGT G Zic1 CAGTATCCCGCGATTGGTGT GCGAACTGGGGTTGAGCTT 1229 1230 Table 2 The primers for cDNA cloning Mouse genes Forword (5'-3') Reverse (5'-3') Adnp TL ATGTTCCAACTTCCTGTCAACA GATGCAACACGGCCCATATGC ATC Adnp-Nter ATGTTCCAACTTCCTGTCAACA TCAGTGGACTAGATGCAGAGTGAT (1-685) ATC Adnp-Nter ATGTTCCAACTTCCTGTCAACA TCATTCATGGTCCTCAATGACATG (1-244) ATC CT Adnp ATGGAACGGATAGGCTATCAG TCAGAGGCATTTGCTAGTAAAATT (245-491) GTC GTG Adnp ATGCACAATTTTACTAGCAAAT TCAGTGGACTAGATGCAGAGTGAT (484-685) GCCTC Adnp-Cter ATGGTCCGCGACTGTGAAAAGT GATGCAACACGGCCCATATGC AC β-catenin-TL ATGGCTACTCAAGCTGACCTG TTAAACCTTATCGTCGTCATCCTT β-catenin ATGGCTACTCAAGCTGACCTG TCATTCTTCCTCAGGGTTGCCCTT (1-55) β-catenin ATGCGTGCAATTCCTGAGCTGA TCACTTGTCCTCAGACATTCGGAA (151-666) CA β-catenin ATGCGTGCAATTCCTGAGCTGA TCATCCACCACTGGCCAGAATGAT (151-320) CA β-catenin ATCATTCTGGCCAGTGGTGGA TCACTTGTCCTCAGACATTCGGAA (320-666) TGA 1231 1232 Table 3 The primers for making In situ hybridization probes Zebrafish genes Forword (5'-3') Reverse (5'-3') adnpa CTTTGGAGTCCATGGTGCTG TAATACGACTCACTATAGGGCCTTT GGGAGTCTGGCAAAC adnpb ATGTACCCATACGATGTTCCA TCCAATCGACAAGGACTGTGTAG GATTACGCTGAATGTTTCAA bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

CTCCCAGTGAACAAC phox2a CCGGACATCTACACGAGAGA TAATACGACTCACTATAGGGGGGAA G TTTGATGACCACGCTG neurod1 CGAGCAGAGCCAGGAGAT TAATACGACTCACTATAGGGAGGGT GGTGTCAAAGAACG dlx5a ATGACTGGAGTATTCGACAG TCAGTACAACGTTCCTGATC A 1233 1234 1235 1236

1237 Figure 1 ADNP depletion leads to the disruption of ESC pluripotency

a b c

WT control Adnp KD Adnp-/-

Allele1 ADNP 150 KD

bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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. Allele2 GAPDH 37 KD

WT: 5'-CCCTTCTCTTACGAAAAATCAGGTAAGTCTCACAGAAATGCCC-3' A1: 5'-CCCTTCTCTTACT------TCAGGTAAGTCTCACAGAAATGCCC-3‘ -5 A2: 5'-CCCTTCTCTTACGAAAA------GGTAAGTCTCACAGAAATGCCC-3‘ -4 d e f

control Adnp-/- 4 Control Day 2 Day 4 Day 9 Day 11 Adnp-/- 3

2 control 1

Relative mRNA levelsRelative mRNA 0 250μm250μm Adnp Sox7 Sox17 Foxa2 Gata4 Gata6

Adnp-/- 250μm Endoderm Day 6 EBs 1.5 g control Adnp-/- h 1.0 Olig2 Pax6 Pax6 60 T Sox17 Map1b 10 40 800 Olig2 0.5 8 30 600 Nes 6 20 400 Otx2 4 Ectoderm 2 10 200

Mesp1 levelsRelative mRNA 0 0 0 0 Mesp2 Adnp Pax6 Nes Gsc T Tbx6 10 Nes Mesdc2 Gsc Gata6 8 Foxc1 Ectoderm Mesoderm 80 800 1.5 6 Cdx2 60 600 4 Gsc 40 400 Mesoderm 2 Lamc1 20 200 1.0 0 Gata6 0 0 Sparc Lama1 Adnp 3 Pou5f1 0 3 6 9 Sox7 0.5 2.5

Col4a1 levelsRelative mRNA 2.0 2 Col4a2 1.5 Lamb1 0 1.0 1 Sox17 levelsRelative mRNA 0.5 Endoderm Nanog Klf4 Sox2 Pou5f1 Gata4 0.0 0 Amn Log2 FPKM Foxa2 Pluripotency Day 0 3 6 9 0 3 6 9 Dab2 -9 -5 -1 0 -1 1 -5 -9 5 9 Figure S1

a b

control Adnp-/- control Adnp-/- + FLAG-ADNP bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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.

250μm

c e

Gfp shRNA Adnp shRNA Adnp 4

1.0 3 3

2 2 0.5

1 1 Relative mRNA levelsRelative mRNA Relative mRNA levelsRelative mRNA

0 0 0 Gfp sgRNA Adnp shRNA Adnp Nes Pax6 Gsc T Gata4 Sox7 Nanog Pou5f1 Sox2 KLF4 Ectoderm Mesoderm Endoderm Pluripotency d

GFP shRNA Adnp shRNA

250μm Figure 2 ADNP is required for proper ESC neural differentiation

a b Day 6 Day 20 d

300 300 control Adnp-/-

control 200 200 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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. count 100 100

0

Adnp-/- 0 Stage 1 Stage 2 250μm PE-A PE-A c e DAPI NESTIN ADNP Merge

control 20 Nestin 40 Pax6 20 Olig2 Adnp-/-

15 30 15 control

10 20 10

5 10 5 Adnp-/- 0 0 0 Relative mRNA levelsRelative mRNA

10μm +FLAG-ADNP f g DAPI TuJ1 GFAP merge h control Adnp-/- Control Adnp-/- ADNP 150 KD Tubb3 Gfap

15 150 control TUJ1 55 KD 10 100

50 5 GFAP 50 KD 0 0 Relative mRNA levelsRelative mRNA D0 D19 D0 D19 37 KD

Adnp-/- GAPDH 25μm Figure S2

a b DAPI PAX6 Merge Control Adnp-/- bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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. 40 40 Pax2 Sox1 30 30 control 20 20 10 10 0 0 /-

15 10 Foxd3 Nptx2 Adnp-

10 Relative mRNA levelsRelative mRNA 5

5

0 0 Day 0 Day 6 Day 0 Day 6

+ 3XFLAG-ADNP 25μm

c

DAPI TuJ1 GFAP Merge control Adnp-/- 25μm Figure 3 ADNP promotes the expression of neuroectoderm developmental genes

a b

neural induction Up Day 3 DEGs Day 6 DEGs Unchanged Down 10 10 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder.Day 0 All rights reserved.Day No reuse 3 allowed without permission.Day 6 627 1,344 control ESCs 5 5

0 0

578 1,086 Adnp-/- ESCs -5 -5

RNA-seq -10 -10 -5 0 5 10 -5 0 5 10

c d Day 6 Adnp-/- Day 6 control e Adnp-/- + Adnp-/- + Adnp-/- + Rep #1 #2 #1 #2 control Adnp-/- DAPI PAX6 Merge Phox2b Tet-E day 0 Tet-E day 2 Tet-E day 4 Six3 Pax3 Edn3 /- En1 e Olig2 250μm Pax2 Adnp- Nr2f1 Otx1 Nkx6-1 Hes5 Sox9 Pax6 f 100 Nsg2

Meis1 Day 0 En2 ** treatment Ncam1 Prdm1 cells Sox1 + Nptx2 50 Irx2 * Gli3

Zic5 Day 2 Neuroectoderm-related genes

Meis2 treatment % of PAX6 relative to control Map2k6 Meis3 0 Sox3 Serf1 Day 0 Day 2 Day 4

Klf4 Log2 FPKM Tet-Express treatment Esrrb

Sox2 Day 4 -5 -5 -4 -3 -2 -1 0 1 2 3 4 5

Nanog treatment Pou5f1 25μm Pluripotency Figure S3

a b d Tet-Express + - +

4 Adnp Adnp-/- + Tet-Express inducible 3XFLAG-ADNP bioRxiv preprint doi: https://doi.org/10.1101/761064Day 0 day; this2 version dayposted September4 day 9, 2019. 6 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. Pax6 150 KD 3 Nestin ADNP 3xFLAG-ADNP Olig2 150 KD 37 KD 2 GAPDH GAPDH 37 KD 1 Relative mRNA levelsRelative mRNA 0 Day 0 2 4 6

c Day 3 Adnp-/- Day 3 control e

Rep #1 #2 #1 #2 Tet-Express treatment Tet-Express treatment Tet-Express treatment Phox2b from day 0 from day 2 from day 4 En1 Prdm11 Sox9 Hes5 Gli3 Nrcam Six5 Nptx2 Map2k6 250μm Zfp423 Meis3 Serf1 Nes Neuroectoderm-related genes Sox3 Zic5 Sall2

-5 -4 -3 -2 -1 0 1 2 3 4 5 Log2 FPKM Figure 4 Wnt signaling is impaired in the absence of ADNP

a b Day 6 Adnp-/- Day 6 control c d control Rep #1 #2 #1 #2 Day 3 2.0 Adnp-/- Wnt9b Dusp5 1.5 TopFlash-luc bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. The copyright holder for Wnt10bthis preprint (which 1.0 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.Wnt1 control Wnt3a 0.5 Log2 FPKM Adnp-/- Wnt4 0 150 Wnt6 Wnt5a Adnp Axin2 Dkk1 Ctnnb1 Lef1 Ccnd1 100 Tcf712 0 1 2 3 4 5 6 7 8 9 L-Myc 50 Day 9 2.0 -log10 (P value) Dkk1 levelsRelative mRNA Wnt-related genes 0 Tcf4 1.5 Axin2 Relative luciferase activities DMSO Wnt3a Wnt8a 1.0 Lef1 0.5 Ccnd1

Tcf3 -5 -4 -3 -2 -1 0 1 2 3 4 5 0 Adnp Axin2 Dkk1 Ccnd1 e 1.5 1.5 f Control 1.0 Adnp-/- 1.0 * + CHIR DAPI NESTIN ADNP Merge 0.5 + Wnt3a 0.5 + 3xFLAG-ADNP

0 0 control 3 3 3 Nes 2 * 2 2 1 1 1 Adnp-/-

Relative mRNA levelsRelative mRNA 0 0 0

2.0 2.0 Pax6 2.0

* * + CHIR 1.5 1.5 1.5 25μm

1.0 1.0 1.0 * 0.5 0.5

0 0 0 Figure S4

a b c Day 3 Adnp-/- Day 3 control

Rep #1 #2 #1 #2 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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. Nos2 Lgr5 300 TopFlash-luc Fgf18 Sox9 Cd44 200 Dusp5 Wnt4 Wnt5b 100 Nrcam Wnt3 0

Irx3 Relative luciferase activities Pitx2 ADNP 0 0.1 0.3 0.5 ug Wnt8a

Wnt-related genes Axin2 Cdc25b Mmp2 0 50 100 150 200 250 300 350 400 L-Myc Number of genes Ccnd1 Msl1 Sfrp2 d Tcf3 Nanog DAPI PAX6 β-Catenin Merge

-5 -4 -3 -2 -1 0 1 2 3 4 5 log2 FPKM

control e Adnp-/-+ CHIR Adnp-/-+ CHIR Adnp-/-+ CHIR control Adnp-/- Day 0 Day 2 Day 4 Adnp-/- +CHIR 10μm Figure 5 Identifying β-Catenin as a ADNP interacting protein

a b IP c CHD4 IgG ADNP Input ADNP ADNP 150 KD IP IgG ADNP Input TRIM28 β-Catenin 92 KD bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. The copyright holder for this preprint (which 150 KD was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. ADNP TRIM27 IP β-Catenin 92 KD LIN28A IgG β-Catenin Input

β-Catenin ADNP 150 KD

β-Catenin 92 KD HP1 γ（CBX3）

d e f IP IP 244 446 685 1108 1 Input IgG 3x FLAG-ADNP ADNP Input IgG FLAG-ADNP (1-685) ADNP (1-244) 150 KD FLAG-ADNP 75 KD 3x FLAG-ADNP ADNP-NAP (1-685) 55 KD ADNP (446-685) MYC-β-Catenin 92 KD β-Catenin 92 KD ADNP (686-1108) ADNP (1-685)

g IP: FLAG h IP i FLAG-ADNP (1-685) + - + - Input Armadillo repeats HA--β-Catenin - + + - IgG FLAG-ADNP-NAP 1 151 666 781 HA-β-Catenin β-Catenin 75 KD 92 KD FLAG-ADNP (1-685) β-Catenin (1-151) FLAG-ADNP-NAP HA--β-Catenin 92 KD 27 KD β-Catenin (151-666)

β-Catenin (666-781)

j IP k DAPI β-Catenin ADNP Merge IgG FLAG-β-Cat (151-666) Input

ADNP 150 KD

FLAG-β-Catenin 57 KD 10μm (151-666) Figure S5

a IP c IP Input FLAG-β- IgG HA-β-Catenin Input IgG bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. The copyright holder for this preprint (which Catenin (1-150) was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 92 KD HA-β-Catenin ADNP 150 KD

FLAG-ADNP (686-1108) 45 KD FLAG-β-Catenin (1-150) 17 KD

IP

b Input HA IgG IP ADNP 150 KD FLAG-ADNP IgG Input (1-244) HA-β-Catenin (667-827) 18 KD HA-β-Catenin 92 KD

FLAG-ADNP (1-244) 28 KD d

IP DAPI ADNP Merge FLAG-ADNP IgG Input (446-685) Day 0 HA-β-Catenin 92 KD

FLAG-ADNP (446-685) 25μm 27 KD DAPI NESTIN ADNP Merge IP

FLAG-ADNP IgG (1-685) ∆NAP Input Day 6

HA-β-Catenin 92 KD 25μm

DAPI TuJ1 ADNP Merge differentiation neural ESC FLAG-ADNP(1-685) ∆NAP 75 KD Day 20 25μm Figure 6 ADNP stabilizes β-Catenin during ESC neural differentiation

a b c Cytoplasmic Nuclear control Adnp-/- control Adnp-/- D0 D4 D6 D0 D4 D6 control Adnp-/- control Adnp-/- ADNP 150 KD ADNP 150 KD 150 KD 92 KD ADNP β-Catenin NESTIN 200 KD bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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. β-Catenin 92 KD 92 KD GAPDH 37 KD Total β-Catenin GAPDH 37 KD pβ-Catenin 92 KD PCNA 36 KD GAPDH 37 KD

d e DAPI β-Catenin Merge h control Adnp-/- control Adnp-/-

CHIR - - + - - + MG132 - - - + + + control

ADNP 150 KD

β-Catenin 92 KD

/- GAPDH 37 KD Adnp-

10μm

f IP: β-Catenin g control Adnp-/- ADNP 0 1 3 5 ug 150 KD ADNP

92 KD Total β-Catenin

92 KD Ubiquitin p-β-Catenin GSK3β 46 KD

46 KD pGSK3β

β-Catenin 92 KD GAPDH 37 KD Figure S6

bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. The copyright holder forb this preprint (which c awas not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. IP: β-Catenin

shRNA shRNA control Adnp-/- control Adnp Adnp-/- +FLAG-DNP Gfp A ADNP 150 KD ADNP 150 KD

β-Catenin 92 KD β-Catenin 92 KD Ubiquitin GAPDH 37 KD GAPDH 37 KD

β-Catenin 92 KD

d e

Day 3 neurosphere

FLAG-ADNP 1 2 3ug control Adnp-/- HA-β-Catenin 2 2 2ug ADNP 150 KD FLAG-ADNP 150 KD

92 KD HA-β-Catenin GSK3β 46 KD 46 KD GSK3β IP: HA pGSK3β 46 KD Total β-Catenin 92 KD GAPDH 37 KD Figure 7 ADNP stabilizes β-Catenin by peventing the degradation complex formation

a b c Armadillo domain IP IP 151 320 666 FLAG-β-Catenin Input IgG IgG FLAG-β-Catenin Input 1 2 3 4 5 6 7 8 9 10 11 12 (151-320) (320-666) 150 KD 150 KD bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. The copyright holder for this ADNPpreprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. ADNP Arm1 Arm2 FLAG-β-Catenin 19 KD FLAG-β-Catenin 38 KD Axin (151-320) (320-666) APC

d e f Tet-Express + - +

FLAG-ADNP 0 1 2 3 ug FLAG-ADNP 0 1 2 ug MYC-APC MYC-AXIN1 2 2 2 2 ug 2 2 2 ug Adnp-/- + HA- control Adnp-/- Catenin HA-β-Catenin 2 2 2 2 ug HA-β-Catenin 2 2 2 ug β-

150 KD HA-β-Catenin 95 KD FLAG-ADNP FLAG-ADNP 150 KD

MYC-APC Total β-Catenin 92 KD MYC-AXIN2 110 KD 300 KD IP: HA IP: HA HA-β-Catenin 92 KD GAPDH 37 KD HA-β-Catenin 92 KD

g DAPI TuJ1 Merge control /- Adnp-

+ β-Catenin 50μm Figure 8 ADNP plays key role in cerebellar neuron development

a c 15 Adnp * 12 En2 Math1 10 10 30 8 6 20 5 4 10 2 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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. 0 0 0

Wnt1 50 Pax6 * 40 * pax2 * 40 30 60 30 20 20 40 b 10 10 20 control Adnp-/- 0 0

Relative mRNA levelsRelative mRNA 0 60 150 Zic1 Six3 * 30 Wnt8 40

Day7 100 20 50 20 250μm 10

0 0 0 Day 0 Day 6 Day 0 Day 6 Day 0 Day 6 Control Adnp-/- d e f Adnp Math1 Ncam 10 600 * 600 * 8 400 6 400

4 200 200 DAPI β-Catenin MATH1 Merge 2 0 0 0 DAPI L7 HuC/D Merge

Six3 8 Pax6 300 Wnt1 control 10 8 6 200 control 6 4 4 2 100 Adnp-/- 2 /- 0 0 0 Relative mRNA levelsRelative mRNA Adnp- Pax2 Otx2 25μm 25 15 100 Zic1 + CHIR 20 80 * 10μm 15 10 60 10 5 40 5 20 0 0 0 Day 0 Day 10 Day 0 Day 10 Day 0 Day 10 Figure S7

a b Day6? a' d' b' c' DAPI NESTIN Merge bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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' conrol

1mm 1mm

b' c' d' Adnp-/- 25μm

1mm 1mm 1mm

c d control Olig2 15 S100b Adnp-/- DAPI L7 HuC/D Merge 12 10 8 5 4

control 0 0

150 Gfap 2.0 Adnp /-

100 1.0 Adnp- Relative mRNA levelsRelative mRNA 10μm 50

0 0 Day 0 Day 10 Day 0 Day 10 Figure 9 ADNP regulates the proliferation of ESC-derived neural progenitor cells

a b c DAPI PH3 Merge control Adnp-/- 150 bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. The copyright holder for this preprint (whichWT + CHIR was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

100

Adnp-/- 50

control Adnp-/- Number of PH3+ cells + CHIR Adnp-/- 50μm 0

d e Day 6 neurospheres f Day 6 neurospheres control Adnp-/- -/- control Adnp-/- control D0 D2 D4 D6 D0 D2 D4 D6 Adnp + 3xFLAG-ADNP ADNP 150 KD 17 KD Caspase-3 32 KD PH3 PH3 17 KD cleaved-Caspase3 17 KD GAPDH 37 KD β-Catenin 92 KD GAPDH 37 KD GAPDH 37 KD g

control Adnp-/- Adnp-/-+ CHIR Adnp-/-+ β-Catenin +3x FLAG-ADNP Figure 10 loss of adnp leads to reduced Wnt signaling and defective brain development in zebrafish embryos

a b c d gRNA gRNA DAPI HuC/D Merge adnpa adnpb control 160/190 control bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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. 19/20

WT: Mutant: 24/25 30/190 adnpa-/- adnpb-/- adnpa-/- adnpb-/-

e f g elavl3: GFP dlx5a neurod1 phox2a

control 2 adnpa-/- adnpb-/- adnpa mRNA control control 15/15 19/19 29/30

1

50μm Relative luciferase actvities 0 adnpa-/- adnpb-/- adnpa-/- adnpb-/- 14/22 17/20 24/30

h

1.0

control 0.5 adnpa-/- adnpb-/- + adnpa mRNA

Relative mRNA levelsRelative mRNA 0.0 axin2 lef1 ccnd1 C-myc Figure S8

a b

bioRxiv preprint doi: https://doi.org/10.1101/761064; this version posted September 9, 2019. 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. adnpa

24 h 36 h 48 h 72 h 96 h adnpb adnpa

adnpb d e

C1 150/208 c 15/110

control adnpa-/- adnpb-/-

V1 80/110 OE C2 27/208

V2 10/110 Adnpa C3 11/208 adnpa-/- adnpb-/-

V3 5/110 C4 20/208