Lasser M , Sun N , Xu Y , Xu Y , Wang S , Drake S , Law K , Gonzalez S , Wang B , Drury V , Castillo O , Zaltsman Y , Dea J , Bader E , McCluskey KE , State MW , Willsey AJ , Willsey HR .

U01 MH116487 NIMH NIH HHS, U01 MH115747 NIMH NIH HHS

             mitotic spindle

	Fig. 1. ASD-associated chromatin regulators localize to microtubules. (A) Human Strep-tagged constructs for ASD-associated chromatin regulators ADNP, CHD8, CHD2, POGZ and KMT5B (labeled by Strep, green) localize to the nucleus (labeled by DAPI, blue) during interphase when expressed in Xenopus. (B) However, these constructs localize to the mitotic spindle during mitosis. Negative controls do not show these localizations. Arrows indicate spindle poles. See also Figs S1-S3.
	Fig.S1. Control plasmid ETV1-Strep does not localize to the mitotic spindle. Strep-tagged ETV1 transcription factor, which is not associated with ASD, does not localize to the mitotic spin- dle during mitosis. Related to Figure 1.
	Fig. S2. ASD-associated chromatin regulators localize to microtubules in vitro. ASD-as- sociated chromatin regulators (ADNP, CHD8, CHD2, POGZ, and KMT5B) localize to the mitotic spindle or centrosome when expressed in HEK293T cells, while a control GFP-Strep construct does not. Bottom panel is a higher-magnification view of the boxed area in the panel above. Ar- rows point to the spindle poles. Related to Figure 1.
	Fig. S3. Antibody stainings for ASD-associated chromatin regulators. (A) Chd2 antibody validation by western blot of control embryos and embryos injected with antisense oligonucle-otides targeting chd2. (B) Chd2 knock-down (chd2KD) causes a reduction in Chd2 protein, relative to beta- Actin. (C) Staining of X. laevis blastula-stage embryos with ASD protein antibodies. White arrows indicate interphase cells, while red arrows indicate mitotic cells. (D) Human iPSC-derived cortical NPCs stained with the ASD protein antibodies. (E) Human iPSC-derived cortical neurons stained with the ASD protein antibodies. White arrows indicate axonal staining. Related to Figure 1.
	Fig. 2. CHD2 is required for mitotic spindle organization, cell cycle progression, genome stability and cell survival. (A-D) CHD2 CRISPRi in human iPSC-derived NPCs causes an increase in mitotic spindle defects (arrows, multipolar spindle) (A), in cyclin B (G2/M marker) fluorescence per cell (B), in pH2AX (DNA damage marker) puncta per nucleus (C), and in the proportion of CCP3 (cell death marker) positive cells (D) compared with non-targeting CRISPRi. (E) Quantification of the data shown in A (χ2 test). (F) Quantification of the data shown in B. Box is interquartile range, line is median, and whiskers are maximum to minimum values. (G) Quantification of the data shown in C. Dot is at the mean and lines are 95% confidence intervals. (H) Quantification of the data shown in D. Box is interquartile range, line is median, and whiskers are maximum to minimum values. ****P<0.0001 (non-parametric rank sum test). See also Fig. S4.
	Fig. S4. CHD2 knockdown validation, spindle defect examples, and cyclin E staining. (A) CHD2 expression is reduced following CRISPRi targeting of CHD2, compared to a non-target-ing control CRISPRi line, as determined by qPCR. (B) Examples of normal and abnormal mitotic spindles for Figure 2A. (C) CRISPRi of CHD2 causes a significant decrease in cyclin E (G2/M marker) fluorescence per cell compared to a non-targeting CRISPRi line. (D) Quantification of C. Box is 25-75% interquartile range, line is median, and whiskers are max to min. **** represents p < 0.0001 by rank sum test. Related to Figure 2.
	Fig. 3. A missense variant of CHD2 observed in an individual with ASD disrupts spindle localization. (A) Schematic of human CHD2 protein with functional domains and locations of likely pathogenic missense variants annotated. (B) Strep-tagged human CHD2 and ASD-associated variants CHD2D856G and CHD2G1174D expressed in Xenopus localize to the nucleus during interphase. (C) During mitosis, CHD2 and CHD2G1174D localize to spindles during mitosis, whereas CHD2D856G does not and instead remains localized to DNA (DAPI, blue). Arrows indicate spindle poles. (D) Quantification of the area of overlap between Strep and DAPI. Bars are mean and whiskers are interquartile range. **P<0.01 (one-way ANOVA). ns, not significant.
	Fig. 4. Enrichment of microtubule-related proteins in ASD. (A) hcASD risk proteins (taken from Fu et al., 2022) are over-represented in a microtubule-related centriolar satellite proteome (taken from Gheiratmand et al., 2019). A significant number of overlapping proteins (bold) are annotated as chromatin binding. (B) Genes encoding centriolar satellite-associated proteins (Gheiratmand et al., 2019) are more likely to carry protein-truncating variants (PTVs) in individuals with ASD (Fu et al., 2022) compared with non-network genes. (C) Model for pleiotropy of ASD-associated chromatin regulators. During interphase, they localize to nuclei, regulating gene expression. During mitosis, they regulate tubulin and organize the mitotic spindle (inspired by Yokoyama, 2016). In loss of function, the spindle is disorganized, leading to cell cycle defects, DNA damage and death. See also Table S2.
	Fig. S5. Example images of CHD2 mutant localizations in Xenopus. Additional images of Strep-tagged CHD2 construct localizations in mitotic cells, when expressed in Xenopus. CHD2 and mutant CHD2-G1174D localize to the mitotic spindle, while CHD2-D856G remains associated with DNA. Arrows indicate the spindle poles. Related to Figure 3.