RING1 missense variants reveal sensitivity of DNA damage repair to H2A monoubiquitination dosage during neurogenesis

Ryan, C. W.; Regan, S. L.; Mills, E. F.; McGrath, B. T.; Gong, E.; Lai, Y. T.; Sheingold, J. B.; Patel, K.; Horowitz, T.; Moccia, A.; Tsan, Y. C.; Srivastava, A.; Bielas, S. L.

doi:10.1038/s41467-024-52292-8

Download PDF

Article
Open access
Published: 10 September 2024

RING1 missense variants reveal sensitivity of DNA damage repair to H2A monoubiquitination dosage during neurogenesis

Nature Communications volume 15, Article number: 7931 (2024) Cite this article

713 Accesses
Metrics details

Subjects

Abstract

Polycomb repressive complex 1 (PRC1) modifies chromatin through catalysis of histone H2A lysine 119 monoubiquitination (H2AK119ub1). RING1 and RNF2 interchangeably serve as the catalytic subunit within PRC1. Pathogenic missense variants in PRC1 core components reveal functions of these proteins that are obscured in knockout models. While Ring1a knockout models remain healthy, the microcephaly and neuropsychiatric phenotypes associated with a pathogenic RING1 missense variant implicate unappreciated functions. Using an in vitro model of neurodevelopment, we observe that RING1 contributes to the broad placement of H2AK119ub1, and that its targets overlap with those of RNF2. PRC1 complexes harboring hypomorphic RING1 bind target loci but do not catalyze H2AK119ub1, reducing H2AK119ub1 by preventing catalytically active complexes from accessing the locus. This results in delayed DNA damage repair and cell cycle progression in neural progenitor cells (NPCs). Conversely, reduced H2AK119ub1 due to hypomorphic RING1 does not generate differential expression that impacts NPC differentiation. In contrast, hypomorphic RNF2 generates a greater reduction in H2AK119ub1 that results in both delayed DNA repair and widespread transcriptional changes. These findings suggest that the DNA damage response is more sensitive to H2AK119ub1 dosage change than is regulation of gene expression.

Gain and loss of function variants in EZH1 disrupt neurogenesis and cause dominant and recessive neurodevelopmental disorders

Article Open access 11 July 2023

Two competing mechanisms of DNMT3A recruitment regulate the dynamics of de novo DNA methylation at PRC1-targeted CpG islands

Article 13 May 2021

Simultaneous disruption of PRC2 and enhancer function underlies histone H3.3-K27M oncogenic activity in human hindbrain neural stem cells

Article 22 July 2021

Introduction

Waves of genetic discoveries identified polycomb proteins and their functions following the initial discovery of homeotic transformation phenotypes¹. Mouse models revealed the functional conservation of Polycomb repressive complex 1 (PRC1) orthologues, with the phenotype of biallelic loss-of-function (LOF) or hypomorphic alleles categorized based on embryonic or perinatal lethality^{2,3,4,5,6,7,8}. Phenotypes, even lethal ones, that did not display overt skeletal homeotic transformations highlighted the plethora of cellular processes impacted by polycomb mechanisms. Widespread use of research and clinical exome sequencing has identified rare de novo dominant missense variants in PRC1 components that add an additional layer of complexity to our understanding of PRC1 genetic and functional mechanisms^{9,10,11,12,13,14,15,16,17}.

The heterogeneity of clinical features attributed to pathogenic variants in PRC1 core components implicate more functional diversity between distinct PRC1 complexes than was previously appreciated¹⁸. Leveraging human pathogenic variants to understand the associated polycomb mechanisms has revealed novel functions of heterogeneous PRC1 complexes. Pathogenic variants in RING1 and RNF2 exemplify this scenario. RING1 and its paralogue RNF2 are E3-ubiquitin ligases that function interchangeably within PRC1 to catalyze the monoubiquitination of histone H2A (H2AK119ub1). The molecular functions of RING1 were overlooked based on the mild phenotype associated with knockout of its mouse ortholog, Ring1a^4,5,6,7,8. In contrast, knockout of the mouse ortholog of RNF2, Ring1b, was embryonically lethal⁸. Functionally redundant paralogues can mask LOF phenotypes and obscure important molecular functions, particularly in the case of RING1. Conversely, stable expression of missense variant proteins permits discrete molecular functions to be revealed. This is supported by the recent identification of a de novo dominant RING1 c.284G > A;p.R95Q missense variant associated with microcephaly, growth restriction, intellectual disability and schizophrenic symptoms, which implicates critical yet unknown functional contributions of RING1 to PRC1 activity¹⁵. RING1 c.284G > A;p.R95Q produces a stable protein that does not catalyze H2AK119ub1. The RING1 arginine 95 residue is required to mediate the electrostatic interaction between PRC1 and the acidic patch on the H2A tail, providing a mechanistic rationale for loss of H2A monoubiquitination¹⁹. The clinical phenotype associated with this hypomorphic RING1 allele implicates a role for RING1-dependent H2AK119ub1 during human corticogenesis, however the pathogenic mechanism is not apparent based on the current understanding of RING1 biology.

H2AK119ub1 is an abundant histone modification, distributed across genic and intergenic regions genome wide²⁰. This pervasive distribution of H2AK119ub1 is highly conserved across species, suggesting the functional importance of its abundance to polycomb biology. Investigating the functional significance of this broad deposition mandates interpreting the relative dosage of H2AK119ub1 at functional domains, rather than its presence or absence at a locus. H2AK119ub1 is enriched at polycomb targets, where it has a repressive effect on transcription^3,7,21. H2AK119ub1 is also necessary for efficient replication of pericentromeric heterochromatin²². Additionally, PRC1 is recruited to sites of DNA damage to catalyze local H2AK119ub1, which promotes repair by recruiting the downstream factors necessary to silence transcription around the break site^{1,23,24,25,26,27,28}. The mechanistic relationship between relative H2AK119ub1 dosage across chromosomes, genes and functional elements and the specificity of these H2AK119ub1 functions is a highly significant unanswered question.

Genetic studies of inherited syndromes established a clear connection between the DNA damage response and neurodevelopmental disorders characterized by primary microcephaly and primordial growth restrictions that persist throughout development (Seckel syndrome; OMIM 210600)^29,30. DNA damage slows the rate of DNA replication through replication fork stalling, which can limit neural progenitor cell (NPC) proliferation and ultimately result in reduced brain size^29,30,31. Thus, efficient DNA repair is required for rapid proliferation of the NPCs during neurogenesis and brain development. There is mounting evidence that PRC1-dependent H2AK119ub1 contributes to the repair cascade by silencing local transcription and recruiting downstream repair machinery; however a contribution of RING1 to these processes has not been established^{9,23,24,25,27,28}.

In this study, we established a human in vitro model of neurodevelopment to investigate an allelic series of clinically relevant RING1 and RNF2 missense variants. Our observations reveal that these missense variants function according to a dominant-negative genetic mechanism. Catalytically inactive RING1^MS- and RNF2^MS-PRC1 compete with catalytically active wild-type PRC1 complexes for occupancy at their target loci. We further demonstrate that RING1 and RNF2 bind and ubiquitinate overlapping target loci, suggesting unbiased incorporation of both proteins into PRC1 complexes. This allelic series creates a graded reduction of H2AK119ub1, revealing a nuanced relationship between H2AK119ub1 dosage and its functions of transcriptional regulation and DNA repair. We found that the timing of DNA damage repair and S-phase progression of NPCs are most sensitive to H2AK119ub1 dosage, while altered transcriptional regulation is more tolerant to modest changes. These findings provide insights into RING1 neurobiology, and the molecular underpinnings of the microcephaly associated with pathogenic RING1 variants.

Results

RING1^MS binds polycomb domains and reduces H2A monoubiquitination

RING1 and RNF2 are core components of PRC1 and interchangeably incorporated into PRC1 complexes to effectuate H2AK119ub1 catalysis (Fig. 1A). Population sequencing studies indicate that RING1 exhibits greater tolerance to loss of function (LOF) variants than RNF2, as more LOF variants are observed in RING1 (loss-of-function observed/expected upper bound fraction score (LOEUF) = 0.41) than in RNF2 (LOEUF = 0.32) in healthy populations (gnomAD v4.0.0)³². While RING1 is more tolerant to LOF variation, some missense variants are pathogenic. Analysis of control population sequence data projected as a Metadome plot reveal highly constrained amino acid residues in the RING and RAWUL domains of RING1 (Fig. 1B)³³. The de novo RING1 c.284G > A;p.R95Q variant disrupts an amino acid that is highly intolerant to missense changes. Consistent with this selective pressure, RING1 c.284G > A;p.R95Q diminishes H2AK119Ub1 and is associated with microcephaly and growth restriction¹⁵. RING1^R95Q is stably expressed in proband-derived lymphoblasts, indicating reduced catalytic activity in the context of diminished H2AK119ub1, although the genetic mechanism has remained unclear¹⁵. We hypothesized that stably expressed catalytically hypomorphic RING1 (RING1^MS) would diminish H2AK119Ub1 through competitive occupancy at polycomb domains. To investigate this pathogenic mechanism, we used CRISPR-Cas9 genome editing to generate hetero- and homozygous RING1 c.284G > A;p.R95Q missense (RING1^+/Ms and RING1^Ms/Ms) and RING1 c.280delG; p.S96SFS10* frameshift null (RING1^−/−) human embryonic stem cell (hESC) lines (Fig. 1C and Supplementary Fig. 1A). RING1 R95 corresponds to R98 in its paralog, RNF2. Both residues are both predicted to facilitate PRC1 interaction with an acid patch on H2A¹⁹. A missense change of this conserved arginine residue in RNF2, RNF2 c.292_293AG > CA;p.R98Q (RNF2^Ms/Ms) was generated to complement the RING1 allelic series (Fig. 1C and Supplementary Fig. 1A). Of note, RNF2^−/− ESCs were also generated but this hESC line could not be maintained due to spontaneous differentiation, highlighting a functional difference between the RING1 and RNF2 paralogs.

**Fig. 1: RING1^MS binds polycomb target loci and reduces H2A monoubiquitination.**

To assess the developmental neuropathology of these variants, hESCs were neural differentiated to NPCs using the dual SMAD neural differentiation protocol³⁴. The RING1 antibody detected the two major RING1 isoforms at molecular weight 49 kDa and 59 kDa in RING1^+/+, RING1^+/Ms and RING1^Ms/Ms NPCs, which were absent in RING1^−/− NPCs (Fig. 1D). Similarly, RNF2 is detected in both RNF2^+/+ and RNF2^Ms/Ms lines, indicating that the missense alleles express full-length RING1^MS and RNF2^MS (Fig. 1E). Additionally, no significant change in the abundance of RING1^MS and RNF2^MS was observed compared to wild-type protein, indicating that the missense variants do not destabilize the expressed protein (Supplementary Fig. 1B).

Next, we assessed the effect of these alleles on global H2AK119ub1 levels in NPCs by Western blot. We observe a dose-dependent decrease of global H2AK119ub1 in RING1^+/Ms (33%), RING1^Ms/Ms (36%) and RNF2^Ms/Ms (48%) NPCs while total H2A remains level across the RING1 allelic series (Fig. 1D, F and Supplementary Fig. 1D). Notably, H2AK119ub1 was not significantly decreased in RING1^−/− NPCs (Fig. 1D, F). As prior studies have demonstrated, RING1 and RNF2 are responsible for all appreciable H2AK119ub1³. This finding suggests that RNF2 compensates for RING1 loss to maintain global H2AK119ub1 levels. RNF2 is well situated to compensate for loss of RING1 and its catalytic activity in NPCs as RNF2 expression is significantly higher than RING1 in control NPCs (Supplementary Fig. 1C). Taken together, these findings suggest that stably expressed RING1^MS is catalytically hypomorphic and its presence prevents compensation by RNF2, which is observed in the absence of RING1.

To determine whether RING1^MS binds to polycomb target loci, we performed Cleavage Under Targets and Release Using Nuclease (CUT&RUN) for RING1 in RING1^+/+ and RING1^Ms/Ms NPCs. Using the peak caller MACS2, we identified 4255 high-confidence binding sites for RING1 and/or RNF2 in NPCs, which we refer to as PRC1 targets. We found that RING1^MS occupancy overlaps with RING1, exhibiting a high correlation across PRC1 target loci (R² = 0.98) (Fig. 1G, H). We quantified RNF2 occupancy in RING1^+/+ and RING1^Ms/Ms NPCs and found that RNF2 occupancy at polycomb domains overlaps with that of RING1 and RING1^MS (Fig. 1I). We also observed the analogous relationship in RNF2^Ms/Ms NPCs, where RING1 occupancy remains similarly enriched at PRC1 target loci (Supplementary Fig. 1E). Enrichment of RING1, RING1^MS, and RNF2 occupancy at polycomb target sites implicates direct competition of catalytically active and inactive PRC1 complexes for the same loci, as is evidenced by the correlated distributions of RING1^MS with RING1 and RNF2 (R² = 0.93–1.00) (Supplementary Fig. 2B).

To visualize the changes in H2AK119ub1 that correspond to the graded reduction in dosage, we performed H2AK119ub1 CUT&RUN. Using an E. coli spike-in, we identified the genotype-dependent pattern of global H2AK119ub1 reduction in RING1^+/Ms (46%), RING1^Ms/Ms (54%), RING1^−/− (24%) and RNF2^Ms/Ms (67%) NPCs (Fig. 1J and Supplementary Fig. 1F). H2AK119ub1 abundance is reduced chromosome-wide (Fig. 1J and Supplementary Fig. 1F). The distribution of H2AK119ub1 remains highly correlated across the allelic series, suggesting a shared pattern of reduction across polycomb targets [R² ≥ 0.9] (Supplementary Fig. 2A). Taken together, these findings depict a dominant-negative genetic mechanism where hypomorphic RING1^MS and catalytically active RING1 and RNF2 are interchangeably incorporated into PRC1. Competitive binding of catalytically inactive PRC1 to polycomb targets causes a reduction of H2AK119ub1. This pattern of RING1^MS binding associated with a corresponding decrease in H2AK119ub1 is shown at three polycomb target HOX loci CIR1, HAGLR, and MNX1 (Fig. 1K).

RING1 missense and LOF variants cause minimal transcriptional dysregulation in human NPCs

To identify the effects of reduced H2AK119ub1 on transcriptional regulation, we performed bulk RNA sequencing of RING1^+/+, RING1^+/Ms, RING1^Ms/Ms, RING1^−/− and RNF2^Ms/Ms NPCs. We find only minor transcriptional changes in RING1^+/Ms (88 upregulated, 49 downregulated), RING1^Ms/Ms (9 upregulated, 3 downregulated), and RING1^−/− (70 upregulated, 30 downregulated) NPCs, whereas a far greater degree of differential expression was observed in RNF2^Ms/Ms NPCs (524 upregulated, 514 downregulated) (Fig. 2A). Genes that are the targets of polycomb regulation were significantly enriched among the differentially expressed genes (DEGs) identified in RING1^+/Ms (p < 0.0001), RING1^Ms/Ms (p = 0.032) and RNF2^Ms/Ms (p < 0. 0001) NPCs but not RING1^−/− (p = 0.61) NPCs (Fig. 2B). This disproportionate impact on polycomb target genes reveals the specificity of differential expression associated with reduced H2AK119ub1. To correlate differential expression to H2AK119ub1 dosage, its abundance across transcription start sites (TSS) at all RefGene annotated TSS was quantified. We observed H2AK119ub1 broadly decreased at TSSs in RING1^+/Ms, RING1^Ms/Ms, and RNF2^Ms/Ms NPCs, with RNF2^Ms/Ms NPCs demonstrating the greatest reduction (Fig. 2C, D). Comparable RING1 and RING1^MS localization at TSSs suggests that, as with PRC1 target loci, RING1^MS competes with RING1 and RNF2 for occupancy at TSS to reduce H2AK119ub1 (Fig. 2E, F).

Absolute decrease in H2AK119ub1 predicts de-repression in RING1^MS and RNF2^MS expressing NPCs

To shed light on the determinants of DEG de-repression, we analyzed their baseline H2AK119ub1 abundance relative to all genes in control NPCs. Interestingly, upregulated DEGs exhibit H2AK119ub1 enrichment. The 95 upregulated genes in RING1^+/Ms and RING1^Ms/Ms NPCs, assessed in combination, revealed a 355% H2AK119ub1 enrichment above the average baseline detected at all genes (Dunn’s p < 0.0001) (Fig. 3A). Likewise, the 524 upregulated DEGs in RNF2^Ms/Ms NPCs exhibit 231% more H2AK119ub1 abundance (Dunn’s p < 0.0001) (Fig. 3A). This trend is even more pronounced at the upregulated DEGs that coincide with PRC1 target loci (Supplementary Fig. 2C). This suggests that genes that are upregulated in association with reduced H2AK119ub1 tend to be those that exhibit enrichment.

To determine whether upregulated DEGs are disproportionately impacted, we quantified the percent H2AK119ub1 decrease at these loci across the graded reduction of the allelic series. Relative to all genes, the 524 RNF2^Ms/Ms upregulated DEGs exhibit a small but significant change in the percent H2AK119ub1 reduction across the allelic series (DEGs vs. all genes for genotype; 47% vs. 45% in RING1^+/Ms, 54% vs. 53% in RING1^Ms/Ms, 25% vs. 22% in RING1^−/− and 69% vs. 66% in RNF2^Ms/Ms, Dunn’s p < 0.0001 for all comparisons) (Fig. 3B). A similar trend was observed for the smaller combined list of 95 DEGs upregulated in RING1 missense conditions (Fig. 3B).

Given that H2AK119ub1 is enriched above baseline at upregulated DEGs, we hypothesized that even modest percent changes in H2AK119ub1 reduction would produce substantial absolute reductions at these loci. To investigate this possibility, we compared the absolute H2AK119ub1 decrease at upregulated DEGs to that measured for all genes. We found a highly significant difference in the absolute decrease in H2AK119ub1 at the 524 RNF2^Ms/Ms upregulated DEGs (2.42 fold greater reduction in RING1^+/Ms, 2.35 fold greater reduction in RING1^Ms/Ms, 2.64 fold greater reduction in RING1^−/− and 2.42 fold greater reduction in RNF2^Ms/Ms) and at the 95 upregulated RING1^+/Ms and RING1^Ms/Ms DEGs (3.88 fold greater reduction in RING1^+/Ms, 3.76 fold greater reduction in RING1^Ms/Ms, 4.64 fold greater reduction in RING1^−/− and 3.68 fold greater reduction in RNF2^Ms/Ms) compared to the average of all genes (Dunn’s p < 0.0001 for all comparisons) (Fig. 3C). This demonstrates that modest percent changes can be amplified at enriched loci to produce a substantial range of H2AK119ub1 reduction.

These data reaffirm the impact of H2AK119ub1 on transcriptional repression and reveal the relationship of dosage to differential expression. Despite the sheer abundance of this histone modification there are loci of enrichment that exhibit substantial absolute dosage reduction. The final level of H2AK119ub1 reduction is due to the compounded effects of both genotype and locus enrichment above baseline. As such, the modest H2AK119ub1 reduction in RING1^+/Ms and/or RING1^Ms/Ms NPCs suggests that only genes with the most enrichment will reach the absolute reduction threshold for de-repression. In contrast, the greater global reduction in RNF2^Ms/Ms NPCs is sufficient to bring lesser enriched genes to this threshold. By assessing these changes across the genotype-dependent graded reduction, a threshold of absolute H2AK119ub1 reduction beyond which expression is derepressed can be inferred (Fig. 3C). This has important implications and sheds light onto the genotypic pattern of differential expression across our allelic series. This relationship between the level of H2AK119ub1 and differential expression is depicted at the homeobox ALX4 locus. H2AK119ub1 abundance at the ALX4 locus progressively decreases in RING1^+/Ms, RING1^Ms/Ms, and RNF2^Ms/Ms samples. However, transcriptional repression is lost only when H2AK119ub1 is reduced to RNF2^Ms/Ms levels (Fig. 3D–F). This illustrates the observation that the more severe RNF2^Ms/Ms H2AK119ub1 reduction allows for derepression of several polycomb target genes, while milder reductions of H2AK119ub1 associated with the RING1^MS do not meet this threshold.

Normal neural differentiation of RING1 ^Ms/Ms human forebrain organoids reveals altered DNA damage repair pathways

To evaluate RING1^MS neuropathology we used a human in vitro model of neural development, RING1^+/+ and RING1^Ms/Ms hESC lines were differentiated to forebrain organoids using the dual SMAD neural differentiation protocol and characterized by single-cell RNA sequencing (scRNAseq) (Fig. 4A). Organoids were cultured to 42 days of differentiation (DD), as at this stage of development forebrain organoids are characterized by rings of proliferating SOX2 NPCs surrounded by differentiated BCL11B deep layer cortical neurons. Organoids at this stage therefore allow characterization of NPC proliferation and differentiation of mature neuronal subtypes. No differences in organoid cytoarchitecture were detected between RING1^+/+ and RING1^Ms/Ms samples by immunohistochemistry (IHC) (Fig. 4B). Of note, DD42 organoids differentiated from RNF2^Ms/Ms hESCs exhibit few SOX2 rings and BCL11B cortical neurons, indicating impaired differentiation consistent with the broad transcriptional changes in this genotype (Supplementary Fig. 3A). Forebrain organoids from four independent differentiations of RING1^+/+ and RING1^Ms/Ms hESCs were dissociated and analyzed by scRNAseq. Sequencing depth, unique molecular identifiers (UMIs) per cell, detected genes per cell, and cell number were similar across samples (Supplementary Fig. 3B–E). Unbiased clustering followed by cluster annotation with canonical markers identified proliferative NPCs, intermediate progenitors, subplate neurons, immature postmitotic neurons, and deep layer neurons (Fig. 4C, D). In line with the important role of PcG complexes in governing cell fate and their clinical association with microcephaly, we hypothesized that reduced H2AK119ub1 in RING1^Ms/Ms organoids would lead to premature differentiation of NPCs, depleting the NPC pool. To assess this developmental trajectory, we compared the cellular composition between RING1^+/+ and RING1^Ms/Ms organoids. No statistically significant difference in the proportional representation of each cell type was quantified (Fig. 4E). Likewise, no significant difference in pseudotime differentiation of NPCs to mature neuronal subtypes was measured by Monocle3 between RING1^+/+ and RING1^Ms/Ms transcriptomics (Fig. 4F)³⁵. Taken together, these results suggest that reductions in H2AK119ub1 in differentiating RING1^Ms/Ms organoids were not sufficient to alter cell fate or timing of differentiation.

To assess differential expression in RING1^Ms/Ms organoids, we aggregated the transcriptomes of cells within NPC clusters into a pseudobulk vector for each RING1^+/+ and RING1^Ms/Ms replicate and performed gene set enrichment analysis (GSEA). We found that 4 of the top 10 enriched pathways were related to activation of the DNA damage response, including “Activation of ATR in response to replication stress” (FDR q = 0.02), “Homologous DNA pairing and strand exchange” (FDR q = 0.01), “HDR through single strand annealing” (FDR q = 0.09), and “HDR through homologous recombination” (FDR q = 0.09) (Fig. 4G). Furthermore, the well-characterized DNA damage response genes ATRIP and ATR, which are genetic etiologies of Seckel Syndrome, were among the most enriched genes in these pathways (Fig. 4H and Supplementary Fig. 4A–C)^29,30. To determine if increased expression of DNA damage response genes is occurring due to impaired DNA damage repair in proliferating NPCs, this analysis was extended to individual NPC clusters. We used a module of genes associated with DNA damage-induced apoptosis (Qiagen RT² Profiler PCR Array) to assign NPCs module enrichment scores. We found significant enrichment of this DNA damage accumulation module in NPCs in the S-phase (186% increase, adjusted p = 0.015), but not in NPCs in G1 (113% increase, adjusted p = 0.152) or G2/M (96.8% increase, adjusted p = 0.165) (Fig. 4I). To identify evidence of cell cycle alteration, we assigned cell cycle phases to each cell in our dataset using phase-characteristic modules of genes³⁶ (Supplementary Fig. 4D). We quantified the proportion of cells in each cell cycle phase and calculated the mid-proliferative index for each genotype by taking the sum of the proportion of cells in the S-phase and G2/M phase and dividing by the proportion in G1. The proportion of cells in the S and G2/M phase is skewed, however the mid-proliferative index did not reach significance (32% increase, p = 0.43) (Supplementary Fig. 4E, F). These findings and the associated Seckel-like clinical phenotypes implicate this inherent DNA damage repair defect as a potential pathogenic mechanism of hypomorphic RING1^MS and reduced H2AK119Ub1.

DNA damage repair delayed by hypomorphic PRC1 complexes in NPCs

To evaluate the timing of DNA damage repair in RING1^+/+, RING1^+/Ms, and RING1^Ms/Ms NPCs, we induced DNA double-strand breaks (DSBs) with ionizing radiation (IR) and assessed their resolution over time (Fig. 5A). DSBs in NPCs treated with 6 Gy IR were assessed by quantifying S139-phosphorylated H2AX (γH2Ax) levels in whole cell lysates by Western blot, γH2Ax nuclear foci by IHC, and damaged DNA via the comet assay. Comparing γH2Ax levels at 0.5 h to that at 24 h shows that the majority of IR-induced DSBs were repaired between 0.5 h and 24 h in RING1^+/+, RING1^+/Ms, and RING1^Ms/Ms NPCs (Fig. 5B). However, at both timepoints after IR γH2Ax levels are elevated in RING1^+/Ms and RING1^Ms/Ms NPCs relative to controls. Consistent with decreased H2AK119ub1, RNF2^Ms/Ms NPCs also exhibit increased γH2Ax at 0.5 h and 24 h after 6 Gy IR (Supplementary Fig. 5A). To determine whether this bulk increase in γH2Ax can be attributed to distinct nuclear DSB foci, we immunostained RING1^+/+, RING1^+/Ms, and RING1^Ms/Ms NPCs 0.5 h after 6 Gy IR. A significant increase in the number of foci per nucleus was observed in RING1^Ms/Ms NPCs, relative to controls (Bonferroni-corrected Mann–Whitney p < 0.0001). No increase was observed in RING1^+/Ms NPCs (Bonferroni-corrected Mann–Whitney p = 0.45) (Fig. 5C, D). The role of H2AK119ub1 dosage is indicated by the significant increase in γH2Ax foci per RNF2^Ms/Ms NPC nucleus (Mann–Whitney p < 0.0001) (Supplementary Fig. 5B, C). Delayed repair was accompanied by robust checkpoint activation at 0.5 h across the allelic series, which was largely resolved by 24 h (Supplementary Fig. 5D). Despite dynamic changes in γH2Ax in response to IR, total whole-cell H2Ax levels are similar across the allelic series (Supplementary Fig. 5D, E).

**Fig. 5: DNA damage repair delayed by hypomorphic RING1^MS and RNF2^MS in NPCs.**

Comet assays were performed to determine whether increased γH2Ax foci reflect bona fide DNA damage. Nuclei of NPCs were electroporated through 0.7% low melting point agarose at 0.5 h, 2 h, and 5 h following IR and comets were quantified. We found that the olive tail moment (OTM), a measure of DNA fragmentation due to DSBs, is increased 0.5 hr after 6 Gy IR exposure in RING1^Ms/Ms NPCs (29%) and to a greater extent in RNF2^Ms/Ms NPCs (57%) (Fig. 5E, F). At 2 h after IR, a similar pattern was observed with a 45% and 77% increase in OTM quantified for RING1^Ms/Ms and RNF2^Ms/Ms NPCs, respectively. Thus, we directly observed delayed DSB repair in RING1^Ms/Ms and RNF2^Ms/Ms NPCs. These results indicate that reduced H2AK119ub1 due to hypomorphic RING1^MS or the more severe RNF2^MS both delay DSB repair. At 5 h after IR, RNF2^Ms/Ms NPCs continued to display a significant increase in DNA damage (106%), while RING1^Ms/Ms NPCs are comparable to wild type (7%) (Supplementary Fig. 5F, G). This suggests that the greater reduction in H2AK119ub1 in RNF2^Ms/Ms NPCs is associated with a greater delay in DSB repair.

Hypomorphic PRC1 complexes increase S-phase stalling following DNA damage

Disruptions to DNA damage repair and cell cycle progression are pathogenic mechanisms of primary microcephaly and growth restriction^29,30. Given that these phenotypes are associated with the clinically relevant RING1 c.284G > A;p.R95Q variant, we assessed whether delayed DSB repair alters NPC cell cycle progression in proliferating NPCs dissociated from neural rosettes at 14 DD. Neural rosettes recapitulate pseudostratified neuroepithelium³⁷. The size and thickness of neural rosettes is determined by NPC proliferation dynamics. Rosette cytoarchitecture was assessed for altered NPC proliferation (Fig. 6A). The diameter of RING1^+/Ms, RING1^Ms/Ms, and RING1^−/− neural rosettes was unchanged relative to RING1^+/+, whereas RNF2^Ms/Ms rosette diameter was significantly larger (68% increase, Dunnett’s adjusted p = 0.0053) (Fig. 6A, B). The average NPC diameter was also larger in RNF2^Ms/Ms samples, suggesting that the increased cell size may contribute to rosette size (14.92 μm vs. 10.52 μm, Dunnet’s p < 0.0001 (Supplementary Fig. 6A).

**Fig. 6: RING1^MS and RNF2^MS increase S-phase stalling following DNA damage.**

The cell cycle dynamics of NPCs dissociated from rosettes were assessed by flow cytometry (Supplementary Fig. 7A, B). Total DNA content was stained with FxCycle Violet, and DNA synthesis was measured by EdU incorporation following a 2-h EdU pulse. Before IR, the proportion of cells in G1, S, and G2/M is similar between RING1 and RNF2 genotypes (Fig. 6C, D). The mid-proliferative index was calculated for each genotype. A small but significant decrease in the mid-proliferative index was calculated for RING1^+/Ms NPCs (15% decrease, Dunnett’s adjusted p = 0.006), while an increase was detected in RNF2^Ms/Ms NPCs (20% increase, Dunnett’s adjusted p = 0.0008) (Fig. 6E). Taken together, these results suggest that RING1^MS does not significantly hinder unirradiated NPC proliferation, however, unirradiated RNF2^Ms/Ms NPCs do exhibit an inherent increase in proliferation. Interestingly, the increased proliferative activity of RNF2^Ms/Ms NPCs corresponds to their broad differential expression, as upregulated DEGs are enriched in several pro-proliferative gene-ontology pathways (Supplementary Fig. 6B).

To detect subtle differences in cell cycle progression due to delayed DSB repair, neural rosettes were treated with 2 Gy IR and cell cycle phases were again assayed after a 1 hour incubation (Fig. 6F). Following IR, an increase in the proportion of cells in the S phase or stalled S phase was measured relative to RING1^+/+ controls in RING1^+/Ms (23%), RING1^Ms/Ms (48%), and RNF2^Ms/Ms (13%) NPCs (Fig. 6G). This correlates with a significant increase in the mid-proliferative index for RING1^Ms/Ms NPCs (79% increase, Dunnett’s adjusted p = 0.029), but does not reach statistical significance in RNF2^Ms/Ms NPCs (36% increase, Dunnett’s adjusted p = 0.46) (Fig. 6H). This shift towards a prolonged S phase in irradiated RING1^Ms/Ms NPCs suggests that delayed DSB repair in RING1^Ms/Ms NPCs impedes cell cycle progression. To assess the impact of altered cell cycle dynamics on NPCs proliferation, we quantified the accumulation of NPCs for 72 h following irradiation. A 32% decrease in RING1^Ms/Ms NPC proliferation was quantified in the first 24 h following 6 Gy IR relative to controls (two-way ANOVA, Dunnett’s corrected p = 0.016) (Supplementary Fig. 6C). Net proliferation is the product of cell cycle dynamics and apoptosis, so the fraction of apoptotic cells was also measured via a flow cytometry-based Annexin V/7-AAD apoptosis assay 24 h after 6 Gy IR (Supplementary Fig. 7C). No significant difference in the fraction of Annexin-V/7-AAD negative NPCs was identified, suggesting that differences in proliferation are primarily the result of cell cycle dynamics (Supplementary Fig. 6D, E).

Discussion

We find that catalytic hypomorph RING1^MS functions in a dominant-negative manner by incorporating into PRC1 complexes and binding genome-wide functional domains without catalyzing H2AK119ub1 (Fig. 7). This genotype-dependent reduction of H2AK119ub1 dosage preferentially impacts DNA damage repair resulting in prolonged S phase of NPCs. These findings are consistent with the molecular pathology associated with primary microcephaly, a prominent clinical feature associated with the pathogenic RING1 c.284G > A;p.R95Q variant^29,30. Our findings reveal functional redundancies between RING1 and RNF2, including a shared pattern of occupancy across the chromatin landscape, which implicates similar PRC1 complex incorporation. Taken together, this suggests that efficient DNA repair is required for NPC proliferation during human brain development and this critical developmental biology is sensitive to changes in PRC1-dependent H2AK119ub1.

**Fig. 7: Dominant-negative missense variants generate graded reduction of H2AK119ub1 to impact functions.**

The accumulation of human population genetics data in public databases is a powerful tool to reveal LOF tolerance, as well as amino acids required to maintain critical biological functions and intolerant to missense variation^32,33. In combination with clinical genetics, human sequencing efforts are painting a more complex picture of polycomb genetics than revealed by transgenic mouse models¹⁸. RING1 and RNF2 exhibit a similar pattern of missense intolerance, specifically in domains with demonstrated roles in binding to H2A histones and catalysis of H2AK119Ub1^15,19,33. This correlation highlights the reliability of functional biology implicated by human genetic data and constrained domains not yet represented by clinical genetics findings. The biology revealed by missense variants is influenced by the tolerance of a gene to LOF variants. RING1 is more tolerant to LOF variation than RNF2, allowing the RING1 allelic series to reveal RING1 biology obscured in LOF models^4,32. Conversely, the missense RNF2 model allows hypomorphic functions to be investigated despite high intolerance to LOF variants³³. In combination, these alleles produced a graded effect on H2AK119Ub1 allowing the functional implications of these variants to be investigated.

We observe a widespread reduction in H2AK119ub1 in both the RING1 and RNF2 genetic models. This finding aligns with the surprising discovery that paralogues RING1 and RNF2 exhibit an overlapping genome-wide occupancy pattern by CUT&RUN, with highly correlated binding at polycomb targets. This critical observation sheds light on the genetic and compensatory mechanisms implicated by our study. The binding pattern of RING1 and RNF2 indicates that both proteins are interchangeably incorporated into PRC1 complexes, without apparent bias for canonical or variant PRC1 complex subtypes. It is plausible that preferential incorporation into a PRC1 subtype would result in divergent occupancy patterns for RING1 and RNF2, as demonstrated for other PRC1 components such as the PCGFs^38,39. This relationship offers insights into RNF2’s compensatory mechanism in the homozygous RING1 LOF (RING1^−/−) model. In the absence of RING1 protein, catalytically active RNF2 is incorporated into the same PRC1 complexes and targeted to overlapping loci to provide H2AK119ub1 above the pathogenic threshold. Of note, we did not see the same redundancy in RNF2^−/− hESCs, which were unstable in culture and spontaneously differentiated. We show that RNF2 exhibits higher expression than RING1 in NPCs (Supplementary Fig. 1C). An outstanding question is whether the inability of RING1 to perform the reciprocal compensation in RNF2^−/− cells is solely due to differences in expression level.

The dominant-negative genetic mechanism revealed by this study is based on the expression of hypomorphic RING1^MS and RNF2^MS proteins and their unbiased incorporation into PRC1 complexes. We detected expression of RING1^MS and RNF2^MS from the RING1 and RNF2 missense alleles. The RING1 and RNF2 antibodies did not differentiate between control proteins and their missense counterparts. Thus, CUT&RUN performed in RING1^Ms/Ms and RNF2^Ms/Ms NPCs indicates that RING1^MS and RNF2^MS can be incorporated into PRC1 complexes and bind to polycomb targets on par with control RING1 and RNF2. The dominant-negative effect is generated when hypomorphic missense-PRC1 complexes bind PRC1 genomic targets but ultimately fail to ubiquitinate H2A tails. This competitive inhibition between ligase active and inactive PRC1 complexes, for the same genome-wide binding sites, is the basis for the progressive reduction of H2AK119ub1 levels from RING1^+/Ms to RING1^Ms/Ms to RNF2^Ms/Ms NPCs (Fig. 7). Ultimately, RNF2 does not encounter such competition for PRC1 incorporation or domain binding in RING1^−/− NPCs, as no RING1 protein is expressed from the frameshift RING1 LOF alleles. As a result, RNF2 can effectively compensate for the loss of RING1, as demonstrated by the slight reduction of H2AK119ub1 in RING1^−/− NPCs. This finding has important implications for the Ring1a mouse knockout model as well. This suggests that Ring1a missense mouse models of highly constrained amino acids critical for H2A ubiquitination will reveal important polycomb biology through dominant-negative mechanisms.

Our allelic series revealed a graded decrease in H2AK119ub1 dosage, resulting in a distinct combination of phenotypes in NPCs. The reduction of H2AK119ub1 observed in RING1^Ms/Ms NPCs is sufficient to disrupt timely DNA damage repair but has only minor effects on transcriptional regulation. In contrast, the greater decrease of H2AK119ub1 in RNF2^Ms/Ms NPCs disrupts both processes (Fig. 7). This suggests that DNA damage repair is more sensitive to decreased H2AK119ub1 than transcriptional regulation. Thus, the milder H2AK119ub1 decrease seen in RING1 missense genotypes is enough to delay repair in NPCs. The significant differential expression in RNF2^Ms/Ms NPCs, contrasted with the minor transcriptional changes in RING1^+/Ms and RING1^Ms/Ms NPCs, suggests that mild decreases in H2AK119ub1 are insufficient to effectuate altered expression. We observe that transcriptionally de-repressed loci in RING1^+/Ms, RING1^Ms/Ms and RNF2^Ms/Ms NPCs exhibit significantly greater absolute H2AK119ub1 reduction. Therefore, we propose that the absolute reduction in H2AK119ub1 must surpass a tolerated threshold quantity for escape of PRC1 repression to occur.

Prior work on the mouse orthologues supports this threshold model. Ring1a^+/+;Ring1b^I53A/I53A mouse embryonic stem cells (mESCs) homozygously expressing hypomorphic RING1B^I53A demonstrate an approximate 50% reduction in H2AK119ub1 but minimal differential expression (60 upregulated genes in mouse ESCs)^2,5. In contrast, mESCs homozygously expressing Ring1a^I50A/D53K and Ring1b^I53A/D56K alleles reveal the hypomorphic function of RING1A^I50A/D53K and RING1B^I53A/D56K to demonstrate complete loss of appreciable H2AK119ub1 and widespread de-repression (2828 upregulated genes)³. These data demonstrate that a ~50% H2AK119ub1 reduction causes only minor transcriptional changes while complete loss drives widespread de-repression. By generating a graded H2AK119ub1 reduction in our allelic series, our data puts these prior studies into context. Our isogenic NPCs allowed the effects of H2AK119ub1 dosages to be directly compared and support a threshold hypothesis for H2AK119ub1 transcriptional repression. Taken together, our allelic series suggests that a magnitude H2AK119ub1 decrease beyond a minimum threshold is necessary for gene de-repression, however functional differences between RING1 and RNF2 beyond H2AK119ub1 dosage cannot be ruled out. This work challenges the notion that H2AK119ub1 is dispensable for PRC1 repression, rather revealing that partial H2AK119ub1 reduction is tolerated.

H2AK119ub1 is required for timely DNA damage repair, but the unique sensitivity of this process to H2AK119ub1 dosage is not immediately evident. PRC1 is recruited to sites of DNA damage within 10 min to catalyze H2AK119ub1 within a 10–30 kb genomic window²⁷. H2AK119ub1 is necessary for assembly of repair machinery at break sites and transcriptional repression at loci with active transcription¹. Failure to establish H2AK119ub1 decreases cell viability following IR, highlighting the significance of this chromatin biology²³. Delayed DNA damage repair increases time spent in the S phase of the NPC cell cycle, a molecular mechanism of primary microcephaly^29,30. These findings provide mechanistic insight into the neuropathology associated with other pathogenic PRC1 variants. Impaired DSB repair and cell cycle progression following ionizing radiation were demonstrated for recessive PHC1 c.2974C > T;p.L992F variants, which resulted in delayed DNA damage repair and replication stalling⁹. Taken together, this suggests that PRC1-dependent H2AK119ub1 is necessary for the efficient DNA repair required for NPC proliferation during human growth and neurogenesis. These findings are consistent with our observations of defective DNA repair and cell cycle progression in RING1^Ms/Ms NPCs. In contrast, while we do observe DNA repair defects in RNF2^Ms/Ms NPCs, these cells appear hyperproliferative at baseline. This is likely due at least in part to the confounding effect of the pro-proliferative transcripts enriched among the de-repressed genes in these NPCs. Consistent with this proposed mechanism, increased pro-proliferative gene expression has been demonstrated to drive uncontrolled proliferation in spite of defective DNA repair in BRCA1-deficient cancer cells^40,41,42,43. Interestingly, RNF2 missense variants impacting arginine 98 have been associated with non-small cell lung cancer, breast cancer, and colorectal carcinoma in the COSMIC database^44,45,46. This supports the idea that RNF2 variants are compatible with proliferative states despite defects in DNA repair. Overall, these findings suggest that the molecular pathology of PRC1 disorders will be due to some combination of altered DNA repair and transcriptional regulation. More work will be required to determine the fidelity of H2AK119ub1 dosage in the pathogenic mechanisms of these disorders.

The diagnosis of schizophrenia in individuals with a comorbid neurodevelopmental disorder is challenging and may be distinct from adult-onset schizophrenia in otherwise neurotypical individuals. Thus, the pathogenesis of schizophrenia associated with the RING1 catalytic hypomorph is likely part of a broader cognitive syndrome and generalizations should be made cautiously. Nevertheless, DNA damage accumulation has been determined to contribute to the molecular pathology of schizophrenia⁴⁷. Increased neuronal 8-hydroxy, 2′ deoxyguanosine (8-OHdG) has been measured in the hippocampus of postmortem brains from individuals with features of schizophrenia relative to controls, suggesting increased oxidative DNA damage⁴⁸. Significantly higher baseline γH2Ax levels were also observed in lymphoblast lines derived from individuals with schizophrenia relative to controls⁴⁹. However, while IR-treated lymphoblasts derived from individuals with schizophrenia exhibited blunted γH2Ax induction, no difference in the rate of γH2Ax resolution was observed⁴⁹. In addition, a subsequent study found a significantly higher burden of DNA damage in drug naïve schizophrenia patients via comet assay of peripheral lymphocytes compared to age and gender-matched controls⁵⁰. These findings further support increased baseline DNA damage in schizophrenia but do not clearly implicate a DNA damage repair etiology. The schizophrenia phenotype associated with defective DNA damage repair highlighted by our study may offer clues; however further work is needed to support a DNA damage etiology of schizophrenia.

Our data demonstrate that the RING1^MS acts in a dominant-negative manner by binding PRC1 target loci but not catalyzing H2AK119ub1. This genetic mechanism revealed the functional relevance of H2AK119ub1 dosage during neurogenesis. The sensitivity of DNA damage relative to the effects on transcription was unanticipated and surprising. These findings have implications for future studies to systematically understand how different PRC1 subcomplexes contribute to H2AK119ub1 dosage and how this mechanism will contribute to the neuropathology of other PRC1 disorders. This dominant-negative genetic mechanism suggests that pathogenic missense variants that are identified in neurodevelopmental disorders will be responsive to knockdown therapy with allele-selective antisense oligonucleotides, however future studies will be needed to directly observe and elucidate the kinetics of integration into heterogeneous PRC1 complexes.

Methods

Ethics statement

This work complies with all relevant ethical regulations and was approved by University of Michigan Institutional Review Boards (HUM00070195). All experiments carried out on edited H9 hESCs that were obtained from WiCell and are of female sex (WiCell Lot#WB0090).

Statistics and reproducibility

Based on recommendations made by the ENCODE consortium, CUT&RUN experiments were performed in duplicate, and RNAseq experiments were performed in triplicate. For other assays, including western blot (4 replicates), flow cytometry (3 replicates), and the comet assay (4 replicates), the number of replicates for each assay was determined according to commonly accepted standards in the field. For single cell RNAseq data, genes that were detected in less than 3 cells were removed, and cells with fewer than 200 detectable genes, greater than 2500 genes, or greater than 5% mitochondrial genes were removed. These criteria were pre-established. Otherwise, no data was excluded from analysis.

CRISPR-Cas9 genome editing was used to generate an experimental mutation in isogenic human embryonic stem cells. Following validation of edits, wild type and edited cells were treated the same. Thus, the isogenic background and controlled environmental variables minimized confounds. For RNAseq experiments, the differentiation that neural progenitor cells were derived from was included into the DESeq2 model to control for batch effects resulting from different differentiation days. Cells were always divided by genotype during data analysis.

During comet assay image acquisition and analysis, researchers were blinded to experimental condition. Other experiments did not involve human influenced data collection and measurement and thus blinding is not relevant.

Human embryonic stem cell culture

H9 hESCs (WiCell Lot#WB0090) were cultured feeder free on matrigel with mTeSR™ Plus (Stem Cell Technologies Catalog #100-0276). Media was changed every 24 h. Culture conditions were 37 °C and 5% CO₂. When cells reached ~80% confluency, they were split 1:50. Detachment was achieved through application of Versene (Gibco 15040066) for 7 min at 37 °C, followed by resuspension and plating in mTeSR™ Plus with 10 µM ROCK inhibitor (Tocris).

CRISPR-Cas9 genome editing

Integrated DNA technologies (IDT) Alt-R system

To generate RING1^+/Ms (2 independent lines), RING1^Ms/Ms (1 line, see below for approach used to generate a second line), RING1^−/− (2 independent lines) and RNF2^Ms/Ms (2 independent lines) ESCs, we employed the IDT Alt-R genome editing protocol. A 35 mm plate of H9 hESCs at ~80% confluence was dissociated with Versene and pelleted by centrifugation at 300 × g for 5 min. Oligo sequences to generate crRNA and donor template were used using the Alt-R gRNA design tool from IDT. Donor oligos were designed to introduce missense changes in addition to synonymous change to the PAM site. Five µl 100 µM Alt-R CRISPR-Cas9 crRNA was mixed with 5 µl 100 µM Alt-R CRISPR-Cas9 trRNA and incubated at 95 °C for 5 min. The resulting gRNA was mixed with 6.67 µl Alt-R S.p. Cas9 Nuclease V3 and incubated at room temperature for 20 min. The Cas9:gRNA solution was added to 75.6 µl electroporation solution (consisting of 62 µl human stem cell nucleofector solution and 13.6 µl supplement from the Lonza Human Stem Cell Nucleofector Kit 2 VPH-5002), 4 µl HDR donor oligo (if generating missense variants), and 4 µl 100 µM Alt-R Cas9 Electroporation Enhancer. The H9 cell pellet was resuspended in this 100 µl CRISPR electroporation cocktail and electroporated in an Amaxa Nucleofector using program A-23. After electroporation, Cells were immediately resuspended in mTeSR™ Plus containing 1x CloneR^TM (Stem Cell Technologies Catalog #05888) and ~0.69 mM Alt-R HDR Enhancer V2 (if generating missense variants) and plated on 35 mm Matrigel dishes. Media was then changed every 24 h until colonies were formed.

crRNA and donor sequences used (* denotes a phosphorothioate modified oligo):

RING1 crRNA

(rUrGrGrUrGrUrCrCrArArGrCrGrArUrCrCrCrUrArGrUrUrUrUrArGrArGrCrUrArUrGrCrU, targeted DNA sequence: TGGTGTCCAAGCGATCCCTA).

RING1 c.284G > A;p.R95Q donor template (G*A*ACAAGGAGTGTCCTACCTGCCGAAAGAAGCTGGTGTCCAAGCAATCCCTAAGACCAGACCCCAACTTTGATGCCCTGATCT CTAAGATCTA*T*C)

RING1 wild type donor template (mixed equimolar with RING1 c.284G > A;p.R95Q donor template generate heterozygous missense) (G*T*GTCCTACCTGCCGAAAGAAGCTGGTGTCCAAGCGATCCCTAAGACCAGACCCCAACTTTGATGCCCTGATCTCTAAGATCTA*T*C)

RNF2 crRNA: (rGrUrCrUrGrGrCrCrUrUrArGrUrGrArUrCrUrUrUrGrUrUrUrUrArGrArGrCrUrArUrGrCrU, targeted DNA sequence GTCTGGCCTTAGTGATCTTT)

RNF2 c.292_293AG > CA;p.R98Q Donor:

(CAAAGAATGTCCTACCTGTCGGAAAAAACTAGTTTCAAAACAATCACTAAGGCCAGACCCAAACTTTGATGCACTCATCAGCA)

PX330-plasmid CRISPR protocol

A second RING1^Ms/Ms line was generated using the pX330 plasmid-based protocol⁵¹. Sense and antisense oligonucleotides were ordered from IDT and annealed to form double stranded DNA encoding for the gRNA with flanking BbsI restriction sites. This fragment and pX330 were digested and annealed to form a pX330 plasmid containing the gRNA. The plasmid was expanded in chemically competent Stbl3 E. coli (Invitrogen) and extracted via Mini-prep (Qiagen). Purified gRNA-containing pX330 and a single stranded donor oligonucleotide ordered from IDT were added to 100 µl electroporation solution (consisting of 82 µl human stem cell nucleofector solution and 18 µl supplement from the Lonza Human Stem Cell Nucleofector Kit 2 VPH-5002. One million H9 hESCs were dissociated with Versene, pelleted by centrifugation at 300 × g for 5 min, and resuspended in the electroporation cocktail containing pX330 and the donor oligonucleotide. Electroporation was done in an Amaxa Nucleofector using program A-23, and cells were immediately resuspended in mTeSR™ Plus containing 10 µM ROCK inhibitor (Tocris) and plated onto 35 mm Matrigel dishes. Media was then changed every 24 h until colonies were formed.

RING1 gRNA

(GGUGUCCAAGCGAUCCCUACGGCCA, targeted DNA sequence: GGTGTCCAAGCGATCCCTACGGCCA)

RING1 c.284G > A;p.R95Q donor template

(CTACCCCAGGAACAAGGAGTGTCCTACCTGCCGAAAGAAGCTGGTGTCCAAGCAATCCCTAAGACCAGACCCCAACTTTGATGCCCTGATCTCTAAGATCTATCCTA)

Sanger sequencing of CRISPR-Cas9 clones

Colonies were dissociated and plated at low densities by serial dilution onto 60 mm matrigel plates to form clonal colonies. Single colonies were picked and cultured in a 48-well plate. Once confluent, 48-well plates were split, and DNA was extracted for PCR amplification and Sanger sequencing (Eurofins Genomics) to confirm genotypes. Two separate clones were identified for each genotype and expanded for use.

PCR and sanger primers:

RING1 fwd: 5’ GCCATAATGGATGGCACAGAGA 3’

RING1 rev: 5’ TCCAGCCCTAATGATCAGTCTG 3’

RNF2 fwd: 5’ GGTATAGAGGGTTTGAGGTTTCC 3’

RNF2 rev: 5’ TCCTGGCTAATACTCTCTCTTG 3’

Neural differentiation of hESCs to neural rosettes and NPC monolayer cultures

To generate rosettes of NPCs from ESCs, we first form aggregates of ESCs by sedimentation as previously described⁵². Briefly, 600 cells in 30ul droplets containing mTeSR™ Plus with 10 µM ROCK inhibitor were added to each well of a 96-well non-adherent V-bottom plate. Twenty-four hours later, cultured hESCs were then differentiated to NPCs and dorsal forebrain organoids via a previously described dual SMAD inhibition protocol³⁴. First, neural induction is initiated by changing ESC culture media to N2 + SMADi medium [49.5 mL DMEM/F12 (Invitrogen), 500 µl N2 supplement (Invitrogen), 83 µl 600uM Dorsomorphin (Tocris), and 5 µl 20 mM A 83-01 (Tocris)]. Medium was changed every other day until day 7, when it was switched to N2 + B27 + Dorsomorphin + FGF medium [49 mL DMEM/F12, 250 ul N2 supplement, 500 ul B27 supplement (Invitrogen), 83 ul 600 uM Dorsomorphin, and 100 µl of 10 ng/ul bFGF] and aggregates were transferred onto 35 mm Matrigel plates using a 1000 µl pipette. Medium was then changed every other day until day 14, when medium was switched to a maintenance N2 + B27 + bFGF media [49 mL DMEM/F12, 250 ul N2 supplement, 500 ul B27 supplement (Invitrogen), and 100 µl of 10 ng/ul bFGF]. To generate NPCs from visible clusters of NPC rosettes, rosettes were removed from the dish manually with a pipette tip under the guidance of an Invitrogen™ EVOS™ microscope and subjected to chemical dissociation via 10-min Accutase incubation (Invitrogen). Dissociated cells were pelleted at 300 × g for 5 min and plated in N2 + B27+ bFGF media on 15ug/mL poly-L-ornithine (Sigma) and 10ug/mL laminin (Sigma) coated dishes.

Neural differentiation of hESCs to forebrain organoids

Neural differentiation was achieved through a dual SMAD inhibition protocol³⁴. On day 1, ESCs at ~80% confluence were switched to N2 + SMADi medium [49.5 mL DMEM/F12 (Invitrogen), 500 ul N2 supplement (Invitrogen), 83 ul 600 uM Dorsomorphin (Tocris), and 5 ul 20 mM A 83-01 (Tocris)]. On day 3, the plates were scored using a sterile needle and colonies were removed from the plate using a sterile cell lifter. Resulting pieces of tissue were cultured in N2 + SMADi medium under continuous motion at 95 rotations per minute (rpm) to form embryoid bodies. N2 + SMADi medium was changed every other day until day 7, when embryoid bodies were plated onto matrigel and cultured in N2 + B27 + Dorsomorphin + FGF medium [49 mL DMEM/F12, 250 ul N2 supplement, 500 ul B27 supplement (Invitrogen), 83 ul 600 uM Dorsomorphin, and 100 µl of 10 ng/ul bFGF]. N2 + B27 + Dorsomorphin + FGF medium was changed every other day until day 14, when rosettes were removed from the dish manually with a pipette tip under the guidance of an Invitrogen™ EVOS™ microscope and cultured in suspension at 95 rpm in a low-adherence 6-well plate in N2 + B27 + bFGF media [49 mL DMEM/F12, 250 ul N2 supplement, 500 ul B27 supplement, and 100 µl of 10 ng/ul bFGF]. N2 + B27 + bFGF media was changed every other day until day 42, when organoids were processed for downstream analysis.

Nuclear extractions

Nuclei from 2.5–5 × 10⁵ cells were placed in accutase for 2 min at 37 degrees then neutralized with DMEM. Cells were spun down for 30 cells at 500 × g and resuspended in 1 ml PBS (Gibco, 14190-144). Cells were then spun for 5 min at 200 × g at 4 degrees Celsius. After removal of PBS, cells were resuspended in 500 µl cold EZ nuclei lysis buffer (Sigma, NUC101-1KT) with 1 µl protease inhibitor, and dounce homogenized in 200 µl EZ buffer. Cells were then spun at 500 × g in 4 degrees Celsius for 5 min. Remove supernatant and resuspended in 400 µl EZ nuclei buffer (Sigma, NUC101-1KT). Centrifuged again at 500 × g Celsius for 5 min. Nuclei were resuspended in 40 ul wash buffer. Once counted, nuclei were then used for CUT&RUN.

CUT&RUN

CUT&RUN experiments were conducted utilizing the EpiCypher CUTANA™ ChIC/CUT&RUN kit v3 protocol (Epicypher). Nuclei were captured with BioMagPlus Concanavalin A beads and incubated with 1 μg primary antibody in 200 μL wash buffer for 2 h. Primary antibodies used were: RING1B (Cell Signaling, D22F2), RING1A (Cell Signaling, 2820S), H2AK119ub1 (Cell Signaling, D27C4), and IGG (Epicypher). After 24 h, unbound antibody was washed away with 100 μL wash buffer twice. Then pA-MN was added at 2.5 μL/reaction and incubated for 10 min at room temperature. The nuclei were then washed and resuspended in 200 μL cell permeable buffer for two washes. Cells were resuspended with 50 ul cold cell perm buffer. One µl of 100 mM CaCl₂ was next added to activate the enzyme, and incubated for 2 h at 4 °C. Next, 0.5 ng of E. coli spike-in DNA and 33 ul of Stop Buffer were added to each sample. Samples were then incubated at 37 °C for 10 min to release the protein-DNA complex. To initiate DNA purification, DNA binding buffer was added to each sample, and then placed in a DNA cleanup column. After centrifugation, samples were washed with DNA wash buffer twice, and then eluted from the column and into a fresh 1.5 mL tube. DNA was quantified using the Qubit fluorometer with the 1x dsDNA HS Assay Kit (Thermo Fisher Scientific) per the manufacturer’s instructions. Samples were then stored at −20 °C for future processing.

CUT&RUN Library Prep

Library preparation was done with the NEBNext Ultra II DNA Library Prep Kit. Briefly, 6 ng of CUT&RUN DNA were treated with endprep module at 20 °C for 30 min and 50 °C for 1 h. Ligation was performed by adding 5 pmol of NEB adapter and ligation mix and incubated at 20 °C for 15 min. To clean up the reaction, 1.75× volume of Agencourt AMPure XP beads were added (Beckman Coulter) to capture short ligation products. PCR amplification was performed for 12 cycles. The resulting libraries were purified with 1.2× volume of AMPure beads then analyzed and quantified by Qubit and Tapestation. For H2AK119ub1, paired-end sequencing was performed by the University of Michigan Advanced Genomics Core using a Next Seq P3 100 cycle kit (Illumina) and sequenced on a NextSeq platform with an average of 23 million reads/sample. For RING1 and RNF2, paired-end sequencing was performed by the University of Michigan Advanced Genomics Core using a Next Seq P2 100 cycle kit (Illumina) and sequenced on a NextSeq platform with an average of 12 million reads/sample.

Data processing and normalization

FASTQ files were aligned to the human reference genome (hg38) and an E. coli genome (NC_012967) using bwa-aln and paired reads were merged into SAM file format using bwa-sampe. Aligned SAM files were converted to BAM format with samtools. BAM files were then cleaned, sorted by coordinate, and PCR duplicates were removed using picard-tools/2.8.1 (https://broadinstitute.github.io/picard/). Unmapped reads were then filtered out using samtools. Reads aligning to annotated blacklist regions (ENCFF356LFX) were removed using the bedtools intersect function. Reads with map quality scores below 30 were removed using samtools. Correlation of CUT&RUN signal at PRC1 target loci across biological replicates was determined using deepTools functions multiBamSummary and plotCorrelation⁵³. All biological replicates were highly correlated, with all having a Pearson’s correlation coefficient >0.9, and were merged for subsequent analysis (Supplementary Fig. 2A, B). Data were entered into GraphPad Prism to generate heatmap plots of correlations. Merged BAM files were then converted to BIGWIG format for visualization using the coverage function in the R package rtracklayer⁵⁴. When converting to BED format, merged H2AK119ub1 BAM files were weighted by a scaling factor (SF). The SF was calculated according to the number of reads aligning to the E. Coli spike-in and the IGG control using the following equation⁵⁵:

$${{{{\rm{SF}}}}}= {{\mathrm{10,000,000}}} \\ \times \frac{{{\#\; {{{\rm{IGG}}}}\; {{{\rm{control}}}}\; {{{\rm{reads}}}}\; {{{\rm{mapping}}}}\; {{{\rm{to}}}}}}\,{E.}{coli}\, \,{{{\rm{genome}}}}}{{{\#\; {{{\rm{sample}}}}\; {{{\rm{reads}}}}\; {{{\rm{mapping}}}}\; {{{\rm{to}}}}}}\,{E.}{coli}\, \,{{{\rm{genome}}}}\times {{\#\; {{{\rm{IGG}}}}\; {{{\rm{control}}}}\; {{{\rm{reads}}}}\; {{{\rm{mapping}}}}\; {{{\rm{to}}}}\; {{hg}}}}38}$$

For RING1 and RNF2 peaks, samples were normalized to Counts Per 10 Million mapped reads (CPM). The number of reads used for these calculations can be found in Supplementary Data 1. CUT&RUN tracks were then visualized using the integrated genome viewer⁵⁶.

Identification of PRC1 target domains

The 4255 PRC1 targets highlighted in our analysis were determined through identification of RING1 and RNF2 CUT&RUN peaks that were present in both biological replicates of wild type NPCs. First, the MACS2 callpeak function used to call peaks on wild type RING1 and RNF2 samples relative to their IGG controls with the following parameters: -t sample.bam, -c IGG_control.bam, -f BAMPE, -g 2913022398, -q 5.00e-02, --broad⁵⁷. Sample.bam files were the merged bam file from two highly correlated biological replicates. IGG_control.bam files were the merged IGG files from the two experiments corresponding to the Sample.bam files. Resulting peaks were then analyzed using the R packages GenomicRanges and rtracklayer to identify 305 RING1 and 4239 RNF2 peaks present in both biological replicates^54,58. Of the 305 RING1 peaks, 289 were also present in the RNF2 peaks. BED files were exported from R, and bedops -m was used to combine BED files containing the RING1 and RNF2 peaks reproduced in both biological replicates. Bedtools merge function was then used to consolidate overlapping peaks. This resulted in the identification of 4255 RING1 and/or RNF2 bound loci that we define as PRC1 targets.

Generation of CUT&RUN heatmaps and profile plots

For analysis at 4255 PRC1 target loci, deeptools computematrix was used with parameters scale-regions, -p 10, -b 5000, -a 5000⁵³. For analysis at transcription start sites, deeptools computematrix was used with parameters reference-point, –referencePoint TSS, -p 10, -b 3000, -a 5000. Genomic transcription start sites were downloaded from RefSeq transcripts aligned by UCSC (hg38.refGene.gtf, downloaded from https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/genes/). To avoid counting a gene multiple times for each of its annotated isoforms in the RefSeq database, the AGAT agat_sp_keep_longest_isoform.pl function was used to keep only the longest isoform in the GTF file⁵⁹. Next, this GTF file was loaded into R using the read.delim base R function, and rows without the indicator “transcript” in the gene column were removed in order to eliminate redundant overlapping regions such as internal annotated introns and exons. A bed file containing the chromosome number, start location, and end location for each transcript was then generated using the base R function write.delim and used for analysis of transcription start sites. Profile plots were generated using deeptools plotProfile. Heatmaps were generated using deeptools plotHeatmap. Input bigWig files for profiling H2AK119ub1 abundance were E. coli spike-in normalized. For RING1 and RNF2 profile plots, bigWig input files were normalized to counts per 10 million mapped reads (CPM).

Generation of chromosome-wide distribution plots

H2AK119ub1 bigWig files that were normalized to the E. coli spike in were imported into R using the import.bw function in rtracklayer⁵⁴. Next, the distribution of the signal across each chromosome was plotted using the kpPlotDensity function in karyoploteR⁶⁰. The height of each density plot was scaled to match the fold change in the calculated global H2AK119ub1 abundance relative to wild type using the following formula⁵⁵:

$$ {{{{\rm{Global}}}}\; {{{\rm{Occupancy}}}}} \\ =\frac{{{\#\; {{{\rm{sample}}}}\; {{{\rm{reads}}}}\; {{{\rm{mapping}}}}\; {{{\rm{to}}}}}}\,{{{hg}}}38\times {{\#\; {{{\rm{IGG}}}}\; {{{\rm{control}}}}\; {{{\rm{reads}}}}\; {{{\rm{mapping}}}}\; {{{\rm{to}}}}}}\,{E.}{coli}\,{{{\rm{genome}}}}}{{{\#\; {{{\rm{sample}}}}\; {{{\rm{reads}}}}\; {{{\rm{mapping}}}}\; {{{\rm{to}}}}}}\,{E.}\,{coli}\,{{{\rm{genome}}}}\times {{\#\; {{{\rm{IGG}}}}\; {{{\rm{control}}}}\; {{{\rm{reads}}}}\; {{{\rm{mapping}}}}\; {{{\rm{to}}}}\; {{hg}}}}38}$$

These calculations can be found in Supplementary Data 1.

Comparison of H2AK119ub1 CUT&RUN abundance at all genes vs. upregulated genes

Counts from BAM files were aggregated across transcribed genic loci (pruned hg38.refGene.gtf file described above) using GenomicAlignments v1.34.1. Counts were multiplied by the spike-in normalization factor (described above) to generate normalized counts at each transcribed locus. Normalized counts at genes that were significantly upregulated in RING1^+/Ms and/or RING1^Ms/Ms NPCs were collected. Normalized counts at genes that were significantly upregulated in RNF2^Ms/Ms NPCs were also collected. Data were scaled by dividing each normalized count by the average normalized count of all genes in wild type NPCs. To generate percent decrease for each locus, the resulting scaled normalized count for each gene was divided by the scaled normalized count for the corresponding gene in wild type NPCs and the result was subtracted from 1. For the absolute difference, each scaled normalized count was subtracted from the corresponding gene in the wild type.

Bulk RNA sequencing and differential expression analysis

NPC monolayer cultures from three biological replicates were generated for each genotype across two independent differentiations. Cultured NPCs at approximately 80% confluence in 24-well plates were harvested for RNA using a PureLink^TM RNA Mini Kit (Invitrogen, #12183018 A). Lysates were homogenized by passage 10 times through a 21-gauge needle. RNA library preparation was performed by Azenta Life Sciences with PolyA selection (Illumina). Libraries were sequenced on an Illumina 2x150bp instrument, targeting approximately 15 million paired-end reads per sample. FASTQ files were aligned to HG38 refGene transcripts using STAR⁶¹. Quality control metrics for each sample can be found in Supplementary Data 2. Resulting gene counts were analyzed for statistically significant changes using DESeq2, with the batch effect (independent hESC differentiation to NPCs) as a covariate⁶². MA plots were generated from DESeq2 results in R. To determine which DEGs are direct targets of PRC1 we cross referenced them with the PRC1 target loci identified from CUT&RUN. To do this, BED files containing the 4255 PRC1 target loci were loaded into R and annotated with gene symbols using CHIPseeker to generate a list of direct PRC1 target genes (TxDb.Hsapiens.UCSC.hg38.knownGene database and annotations from org.Hs.eg.db)^63,64,65. To test for statistical enrichment of PRC1 targets among DEGs for each genotype, a chi-squared test was performed. The observed number of PRC1 target DEGs for each genotype is the number of DEGs that are found within the PRC1 target gene list. To determine the expected number of PRC1 target DEGs for each genotype, we first calculated the baseline probability that a randomly selected gene symbol will overlap with the PRC1 target gene list. Since we assessed 29463 total genes and identified 3725 of them as PRC1 target genes, we calculated that a randomly selected gene has a 12.7% chance of being a PRC1 target. Thus, for each genotype, the expected frequency of PRC1 target DEGs was 12.7%. We therefore multiplied the number of DEGs for each genotype by 0.127 to calculate the expected number of PRC1 target DEGs.

Seq-well single cell RNA-sequencing

Four independent differentiations were conducted to generate batches of day 42 dorsal forebrain organoids. For each sample processed for scRNAseq, 5 organoids were pooled ranging from 1 mm to 1.5 mm in diameter. Pooled organoids were first mechanically dissociated by repeated parsing with a razor blade. Chemical dissociation of the resulting organoid fragments was then done via treatment with Accutase (StemCell Technologies) for 10 min. Cells were then passed 5 times through a fire-polished glass pipette for further mechanical dissociation, and then single cells were isolated by passage through a 40 µm cell strainer and counted using a LUNA cell counter (Logos Biosystems). Seq-Well was then performed as described^66,67. Briefly, functionalized Seq-Well arrays, containing 90,000 picowells, were loaded with barcoded beads (ChemeGenes). In total, 20,000 cells were loaded onto the arrays and incubated for 15 min. To remove residual BSA and excess cells, arrays were washed with PBS. Functionalized membranes were applied to the top of arrays, sealed in an Agilent clamp, and incubated at 37 for 45 min. Sealed arrays were incubated in a lysis buffer (5 M guanidine thiocynate, 1 mM EDTA, 0.5% sarkosyl, 1% BME) for 20 min followed by a 45-min incubation with hybridization buffer (2 M NaCl, 1X PBS, 8% PEG8000). Beads were removed from arrays by centrifuging at 2000 × g for 5 min in wash buffer (2 M NaCl, 3 mM MgCl₂, 20 mM Tris-HCl pH 8.0, 8% PEG8000). To perform reverse transcription beads were incubated with the Maxima Reverse Transcriptase (Thermo Fisher Scientific) for 30 min at room temperature followed by overnight incubation at 52 °C. Reactions were treated with Exonuclease 1 (New England Biolabs, M0293S) for 45 min at 37 °C. Whole transcriptome amplification was performed using the 2X KAPA Hifi Hotstart Readymix (KAPA Biosystems, KK-2602). Beads were split to 1500–2000 per reaction and run under the following conditions 4 Cycles (98 °C, 20 s; 65 °C, 45 s; 72 °C, 3 m) 12 Cycles (98 °C, 20 s; 67 °C, 20 s; 72 °C, 3 m) final extension (72 °C, 3 m, 4 °C, hold). Products were purified with Ampure SPRI beads (Beckman Coulter, A63881) at a 0.6X volumetric ratio then a 1.0X volumetric ratio. Libraries were prepared using the Nextera XT kit (Illumina) and libraries were sequenced on an Illumina NextSeq 75 cycle instrument.

scRNA-seq data processing and sample integration

A gene expression matrix was generated using the Drop-seq software⁶⁸. FASTQ files were converted to bam format to be tagged with cell and molecular barcodes and trimmed. After converting back to FASTQs, reads were aligned to hg38 with STAR. BAM files were then sorted, merged, and tagged with gene exons. Bead synthesis errors were corrected as described and digital gene expression matrices were generated. We excluded the poor-quality cells in the gene-cell data matrix using the Seurat package (v4.0.4) for the samples. Genes that were detected in less than 3 cells were removed. Next, cells with fewer than 200 detectable genes, greater than 2500 genes or greater than 5% mitochondrial genes were removed. We next normalized the data using the Seurat (v4.0.4) sctransform function, with employs a regularized negative binomial regression with cell sequencing depth and % mitochondrial genes as covariates⁶⁹.

Next, the 3000 most variable features across all samples were identified using the Seurat function “SelectIntegrationFeatures” ⁷⁰. We next used these variable genes to integrate the samples using the FindIntegrationAnchors and IntegrateData functions of Seurat (v4.0.4). For downstream differential expression analysis, we normalized the raw unique molecular identifier (UMI) counts in each sample for each cell by dividing each feature count by the total counts for that cell multiplied by 10,000. This is then natural-log transformed using log1p to generate depth normalized counts for all features that can then be used to compare gene expression across groups of cells.

Dimensionality reduction, visualization and clustering

We used the Seurat package (v4.0.4) to perform dimensionality reduction. We used the integrated and normalized data as the input to the RunPCA function of Seurat (v4.0.4) to compute the first 30 PCs. Visualizations in a two-dimensional space were done using RunUMAP function of Seurat (v4.0.4) for the integrated data using 30 PCs. We performed a graph-based clustering approach using FindNeighbors (using dimensions 1:10) and FindClusters (using resolution 0.3) functions of Seurat (v4.0.4). We then collected cluster marker genes using the Wilcoxon rank-sum test between the cells in a single cluster and all other cells using the Seurat (v4.0.4) FindAllMarkers function with log fold change threshold of 0.25 for genes expressed in at least 25% of cells (min.pct = 0.25). To assign identities to clusters, we cross-referenced the marker genes with previously described cortical cells^71,72,73,74. The expression of cell type markers was visualized using the FeaturePlot function in Seurat. Of the 10 clusters, two only demonstrated enrichment of mitochondrial genes and thus no cell type could be assigned, so cells in these two clusters were removed for subsequent analysis. The number of cells in each annotated cell type were noted for each sample and used to calculate the proportion of cells in each sample from each cell type.

Pseudotime analysis

We used Monocle 3 to conduct single-cell pseudotime analysis³⁵. First, data was imported from the Seurat object using the as.cell_data_set function. Next, size factors were estimated using the estimate_size_factors function. Cells were clustered within Monocle 3 with the cluster_cells function, using the UMAP reduction method. The learn_graph function was then applied, with use_partition = TRUE. Next, the order_cells function was used, with the UMAP reduction method and selecting cells in NPC clusters as the root nodes. Finally, plot_cells was used to visualize the trajectory. Using the proportion of cells present at each value in pseudotime for each independent sample, the mean and standard error of the proportion of cells at each moment in pseudotime for forebrain organoids derived from RING1^+/+ and RING1^Ms/Ms hESCs was determined.

Gene set enrichment analysis

To perform gene set enrichment analysis, scRNAseq data was collapsed to pseudobulk by summation of transcripts within NPC clusters. DESeq2 was then used to normalize raw counts, and the normalized counts for each sample were exported as a .csv. This data (17745 genes across 7 samples) was then input into .gct format and loaded into the gene set enrichment analysis software for unbiased comparison of expression of genes in shared biological pathways (UCSD, Broad Institute, downloaded from (https://www.gsea-msigdb.org/gsea/index.jsp)⁷⁵. The reactome database was tested with 1000 permutations of type gene_set, with collapse activated (https://ftp.broadinstitute.org://pub/gsea/Msigdb/human/gene_sets/c2.cp.reactome.v2023.2.Hs.symbols.gmt)⁷⁶. A weighted enrichment statistic was used, with ranking determined with signal2noise. Gene list sorting mode was real, and gene list ordering was descending.

Module scores

To test for evidence of DNA damage in our single cell dataset, we used the Seurat AddModuleScore function to evaluate the aggregate expression of a list of genes enriched expression of genes known to be expressed in response to DNA damage to induce apoptosis. This gene list was assembled by Qiagen as part of the RT² Profiler PCR Array (GeneGlobe ID - PAHS-3012Z) and included the following list of genes: ABL1, AIFM1, BCL3, BRCA1, CHEK2, CIDEA, CIDEB, DYRK2, IFI16, MBD4, MSH6, PCBP4, PML, SFN, TP53, and TP73. The module score function compares the expression of genes in the module to a list of control genes with similar overall expression to generate an aggregate enrichment score. The average enrichment score for each sample was used to calculate a mean and standard deviation module score for each genotype.

To assign a cell cycle phase to each cell in our scRNAseq samples, we utilized the CellCycleScoring function using the Seurat package (v4.0.4) and a list of transcripts associated with cell cycle phases as previously described³⁶. This function is built upon the AddModuleScore function. Genes with the highest positive enrichment scores for modules of genes known to be enriched during the S or G2/M phase are assigned to those phases, and those with negative enrichment of both modules were assigned to G1. Mid-replicative index was calculated by adding the fraction of cells in the S and G2/M phase and dividing that quantity by the fraction of cells in the G1 phase.

Western blot analysis

NPCs were washed in PBS and subsequently homogenized in RIPA buffer supplemented with protease inhibitor cocktail and phosphatase inhibitor cocktail 3 obtained from Sigma-Aldrich (P8340 and P0044; St Louis, MO, USA). Protein concentrations were determined using Pierce™ BCA Protein Assay Kits (Thermo Fisher Scientific, 23225). Cell lysates were separated by electrophoresis on 4–20% Tris-Glycine gels and transferred to PVDF membrane using iBlot 2 PVDF Regular Stacks (Invitrogen IB24001) transferred on an Invitrogen iBlot 2 Gel Transfer Device (IB21001). The PVDF membrane was blocked with 5% milk in TBST and incubated with primary antibodies diluted in 5% milk in TBST overnight (except for anti-γH2Ax which was incubated in 5% BSA in TBST). Primary antibodies used were: anti-RING1 (Cell Signaling #2820, 1:1000), anti-ubiquityl-Histone H2A (Cell Signaling #8240, 1:2000), anti-Histone H2A (Cell Signaling, D6O3A, #12349, 1:1000), anti-Vinculin (Cell Signaling, E1E9V, 1:1000), anti-RNF2 (Cell Signaling #5694, 1:1000), anti-γH2Ax (Cell Signaling #2577, 1:1000), anti-H2Ax (Cell Signaling #2595, 1:1000), anti-pChk2 (Cell Signaling, Thr68, #2661, 1:1000), and anti-Chk2 (Cell Signaling, D9C6, #6334, 1:1000). Donkey anti-rabbit HRP-conjugated (Cytiva, NA9340V, 1 to 5000) and goat anti-mouse HRP-conjugated (Invitrogen, 32430, 1 to 10000) were used for 1 h incubation in 5% milk TBST at room temperature (except for secondary antibody targeting anti-γH2Ax, which was incubated in 5% BSA in TBST). Chemiluminescence detection was performed using SuperSignal™ West Femto Maximum Sensitivity Substrate according to manufacturer’s instruction [Thermo Fisher Scientific, cat no. 34095].

Immunostaining and quantification of γH2Ax in NPC cultures

NPCs were plated in 4-well chamber slides and cultured to ~50% confluency, when they were treated with 6 Gy ionizing radiation (at 2.27 Gy/min) in an orthovoltage unit through the University of Michigan Experimental Irradiation Core. Each chamber was treated independently throughout this protocol. NPCs were then allowed to recover for 30 min at 37 °C and 5% CO₂ and subsequently fixed with 4% paraformaldehyde for 5 min at room temperature. Next, cells were treated for with CSK buffer (100 mM NaCl, 300 mM Sucrose, 3 mM MgCl2, 0.7% Triton X-100 in deionized water) 10 min at room temperature. NPCs were then treated again with 4% paraformaldehyde for 5 min at room temperature. NPCs were next washed 2X with PBS and then permeabilized with 0.1% TRITONX-100 in PBS for 10 min at room temperature. After two more PBS washes, NPCs were saturated with 5% BSA in PBS for 30 min at room temperature. Next, anti-γH2AX S139 (Cell Signaling [20E3] #2577, S139) diluted 1:100 in PBS was added to the cells for 2 h at room temperature. Next, NPCs were washed 2x for 5 min in PBS, and then Alexa Fluor™ 488 anti-Rabbit IgG Secondary Antibody (Thermo Fisher Scientific, Catalog # A-21206) diluted in PBS was added for 30 min. After 2 more 5-min washes, NPCs were incubated in Hoechst 33342 (Thermo Fisher Scientific) for 10 min. Finally, slides were mounted using ProLong™ Gold Antifade Mountant (Invitrogen P10144). Images were acquired with a NIKON N-SIM + A1R microscope and processed with LAS X software. For each genotype, 2–4 images were acquired from each independently stained well of the 4-well chamber slide using a 60x objective at 4096 × 4096 pixel resolution using a NIKON N-SIM + A1R microscope. To count γH2Ax foci, we used ImageJ software. For each image, blinded to the γH2Ax channel, 15 nuclei that were in focus and not overlapping others were circled and added as regions of interest. Then, the “find maxima” function was applied to γH2Ax signal within regions of interest using a prominence of 100 and number of maxima identified was recorded for each nucleus. The Mann–Whitney test was used followed by Bonferroni correction for multiple testing to assess statistical significance.

Immunohistochemistry of organoids

Cells were kept in 4% PFA at 4 °C overnight. Cells were then preserved by submersion in 15% then 30% sucrose solutions. Next, samples were cryopreserved by embedding in OCT cryosectioning media (Tissue-Tek). Embedded cells were cryosectioned at 13 μm. Sections were incubated with PBS for 15 min to wash away OCT. For antibodies that required antigen retrieval, cryosections were heated in 10 mM Sodium citrate for 20 min at 95 °C followed by incubation at room temperature for 20 min. Incubation with a normal donkey serum blocking buffer [5% NDS (Jackson ImmunoResearch), 0.1% Triton X-100, 5% BSA] was performed for 1 h. Sections were stained with primary antibodies in blocking buffer at 4 °C overnight, washed with PBS, and stained with secondary antibodies at room temperature for 1 h. Slides were washed with PBS, incubated with DAPI for 5 min and cover slipped with MOWIOL. Images were acquired with a NIKON N-SIM + A1R microscope and processed with LAS X software. The following antibodies and dilutions were used: SOX2 (Neuromics, GT15098, 1:500) and BCL11B (Abcam, ab18465, 1:500). Alexa Fluor-conjugated secondaries were Alexa Fluor™ 488 Donkey anti-Rabbit IgG Secondary Antibody (Thermo Fisher Scientific Catalog # A-21206) and Alexa Fluor™ 555 Donkey anti-Mouse IgG Secondary Antibody (Thermo Fisher Scientific Catalog # A-31570).

Comet assay

To assay double strand breaks, the neutral comet assay was performed as previously described⁷⁷. Precoated slides were prepared by dipping slides into molten 1% agarose regular melting point agarose, wiping one side clean, and allowing agarose to air dry overnight. NPCs were cultured to ~80% confluency in 24-well plates and treated with 6 Gy ionizing radiation (at 2.27 Gy/min) in an orthovoltage unit through the University of Michigan Experimental Irradiation Core. NPCs were given 30 min, 2 h, and 5 h at 37 °C and 5% CO₂ to recover. Next, NPCs were chemically dissociated with Accutase (StemCell Technologies) for 10 min at 37 °C and centrifuged at 300 × g for 5 min. NPC pellets were resuspended into a single cell suspension in 4 ml 0.7% low-melting-temperature agarose in PBS to a final concentration of approximately 50,000 cells/ml. Next, 200 µl of the single cell suspension was plated onto precoated slides, a cover slip was added, and gels allowed to solidify at 4 °C for 20 min. After the agarose was gelled, slides were submerged in a covered dish containing lysis solution (1% Triton X-100 in 2.5 M NaCl, 0.1 M EDTA-Na₂, and 10.0 mM Trizma base) at 4 °C overnight. After an overnight lysis, slides were removed and washed in neutralization buffer (45 mM HEPES, 0.25 mM EDTA-Na2, 0.3 mg/mL BSA, 2% (vol/vol) glycerol) 3x for 5 min at 4 °C. Cells were next equilibrated in electrophoresis solution (0.3 M NaOH, 1 mM EDTA-Na₂) for 40 min at 4 °C. Slides were electrophoresed in a 21 cm gel box at 21 V for 30 min at 4 °C (~1 V/cm). Cells were removed from the electrophoresis chamber and neutralized in 400 ml of cold PBS for 10 min, followed by cold ddH₂O for 10 min at 4 °C. Slides were then placed on the bench to dry overnight. After drying 140ul ethidium bromide (0.01 μg/mL in water) was added to slides and cover slips were placed. Slides were imaged using a widefield microscope using a 20x objective with the user blinded to experimental genotype (Leica). For each biological replicate, a minimum of 100 comets were imaged and scored. We avoided analyzing doublets (rare because of low density of plating on slides) or comets at slide edges. Comets were scored using the Comet Assay IV software (Instem), with the user blinded to experimental genotype. For each independent experiment, the mean olive tail moment was taken for each genotype. The relative change in the mean olive tail moment for each genotype compared to the wild type for that experimental day (OTM_Experimental/OTM_{Wild Type}) was calculated. At least 3 biological replicates were completed for each genotype in independent experiments. Results were considered significant if the 95% confidence interval calculated by Graphpad Prism did not intersect 1.

Rosette sizes

Day 13 neural rosettes were cultivated from 600 cell aggregates of ESCs as described above and plated onto independent matrigel-coated wells of a 6-well plate. Each well contained a single cluster of rosettes, originating from a single 600-cell aggregate. Five rosette clusters per genotype were imaged on a brightfield EVOS microscope (Life Technologies, AMEP-VH009) with a 2x objective. For each imaged cluster, diameters of visible rosettes were calculated using ImageJ, normalized to 1000 µm scale bars. For irregular rosettes, largest diameter was taken. The mean rosette diameter for each image was then used for comparison. Dunnett’s multiple comparisons test was used to compare the diameters for each genotype across five replicates.

Cell cycle analysis by flow cytometry

Day 14 neural rosettes were cultivated from 600 cell aggregates of ESCs as described above. Replicates were cultivated in separate matrigel-coated wells. Each replicate consisted of 4–5 pooled rosette aggregates. For irradiated rosettes, 2 Gy ionizing radiation (at 2.27 Gy/min) in an orthovoltage unit was conducted through University of Michigan Experimental Irradiation Core and recovered at 37 °C and 5% CO₂ for 1 h prior to assay. To assay the cell cycle phases of neural rosettes, we performed a 2-h 10 µM EdU pulse in N2 + B27 + FGF culture media at 37 °C and 5% CO₂ and followed the Click-iT™ EdU Alexa Fluor™ 488 Flow Cytometry Assay Kit (Thermo Fisher Scientific #C10425). Total DNA content was stained with the FxCycle™ Violet Stain (Thermo Fisher Scientific #F10347). Cells (10,000 per sample) were analyzed using the 405 nm absorption (450 nm emission) and 488 nm absorption (530 nm emission) laser on a Bio-Rad ZE5 Cell Analyzer. Samples were analyzed using FlowJo™ Software (BD Life Sciences). Cells were selected by gating the highest density of cells in the forward scatter area vs. side scatter area plot. Next, doublets were removed by rectangular gating of the 450 nm area vs. 450 nm height graph. Finally, the 450 nm area vs. 530 nm area plot was used to visualize the cell cycle dynamics. Gates for cells in G1 (density of cells with low y-axis and x-axis values), S (higher y-axis across range of X-axis), stalled S (low y-axis and between G1 and G2/M on x-axis), and G2/M (low on y-axis and high on x-axis) were drawn. The number of cells in each of these 4 gates were then recorded for each sample of each genotype. The percent of cells in each cell cycle phase was calculated. Mid-replicative index was calculated by adding the fraction of cells in the S and G2/M phase and dividing that quantity by the fraction of cells in the G1 phase. Dunnett’s multiple comparisons test was used to compare the proportions for each genotype across three replicates.

Cell accumulation assay and NPC diameter assessment

NPCs (12,000 per cell) were plated into wells of a 24-well plates coated with poly-L-ornithine and laminin and cultivated in N2 + B27 + bFGF media at 37 °C and 5% CO₂. At approximately 50% confluence, untreated NPCs were collected by Accutase dissociation and cells were stained with acridine orange/propidium iodide stain (Logos Biosystems Cat# F23001) and live cells were counted using a LUNA cell counter (Logos Biosystems). The average size of live cells was recorded for each sample. Cells were then treated with 6 Gy ionizing radiation (at 2.27 Gy/min) in an orthovoltage unit through the University of Michigan Experimental Irradiation Core and returned to incubation. Wells from each condition were collected and counted each day, and media was changed in remaining wells daily.

Apoptosis flow cytometry assay

NPCs were treated with 6 Gy IR as described above and cultured for 24 h in N2 + B27 + FGF culture media at 37 °C and 5% CO₂ prior to treatment with 7-AAD/Annexin V-PE staining via the manufacturer’s protocol (SouthernBiotech Cat# 10010-09). Cells (10,000 per sample) were analyzed using a single 488 nm laser on a Bio-Rad ZE5 Cell Analyzer. Samples were analyzed using FlowJo™ Software (BD Life Sciences), using a uniform compensation matrix to deconvolve the emission signal. Cells were selected by gating the highest density of cells in the forward scatter area vs. side scatter area plot. Next, doublets were removed by rectangular gating of the 488 nm forward scatter width vs. 488 nm forward scatter height graph. Subsequent rectangular gating of the 488 nm side scatter width vs. 488 nm side scatter area further cleaned the cell population. Finally, 7-AAD area was plotted against Annexin V area, and cells positive/negative for 7-AAD/Annexin V were gated into quadrants. The proportion of each sample that was negative for 7-AAD and Annexin V was noted for each replicate.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

Single Cell RNAseq data were deposited into the Gene Expression Omnibus database under accession number GSE252449 and are available at the following URL: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE252449. Bulk RNAseq data were deposited into the Gene Expression Omnibus database under accession number GSE252447 and are available at the following URL: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE252447. CUT&RUN data were deposited into the Gene Expression Omnibus database under accession number GSE252485 and are available at the following URL: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE252485. Source data are provided with this paper.

References

Barbour, H. et al. Polycomb group-mediated histone H2A monoubiquitination in epigenome regulation and nuclear processes. Nat. Commun. 11, 5947 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Illingworth, R. S. et al. The E3 ubiquitin ligase activity of RING1B is not essential for early mouse development. Genes Dev. 29, 1897–1902 (2015).
Article CAS PubMed PubMed Central Google Scholar
Blackledge, N. P. et al. PRC1 catalytic activity is central to polycomb system function. Mol. Cell 77, 857–874.e9 (2020).
Article CAS PubMed PubMed Central Google Scholar
del Mar Lorente, M. et al. Loss- and gain-of-function mutations show a polycomb group function for Ring1A in mice. Development 127, 5093–5100 (2000).
Article PubMed Google Scholar
Eskeland, R. et al. Ring1B compacts chromatin structure and represses gene expression independent of histone ubiquitination. Mol. Cell 38, 452–464 (2010).
Article CAS PubMed PubMed Central Google Scholar
Morimoto-Suzki, N. et al. The polycomb component Ring1B regulates the timed termination of subcerebral projection neuron production during mouse neocortical development. Development 141, 4343–4353 (2014).
Article CAS PubMed Google Scholar
Tsuboi, M. et al. Ubiquitination-independent repression of PRC1 targets during neuronal fate restriction in the developing mouse neocortex. Dev. Cell 47, 758–772.e5 (2018).
Article CAS PubMed Google Scholar
Voncken, J. W. et al. Rnf2 (Ring1b) deficiency causes gastrulation arrest and cell cycle inhibition. Proc. Natl. Acad. Sci. USA 100, 2468–2473 (2003).
Article ADS CAS PubMed PubMed Central Google Scholar
Awad, S. et al. Mutation in PHC1 implicates chromatin remodeling in primary microcephaly pathogenesis. Hum. Mol. Genet. 22, 2200–2213 (2013).
Article CAS PubMed Google Scholar
Beunders, G. et al. Two male adults with pathogenic AUTS2 variants, including a two-base pair deletion, further delineate the AUTS2 syndrome. Eur. J. Hum. Genet. 23, 803–807 (2015).
Article CAS PubMed Google Scholar
Beunders, G. et al. A detailed clinical analysis of 13 patients with AUTS2 syndrome further delineates the phenotypic spectrum and underscores the behavioural phenotype. J. Med. Genet. 53, 523–532 (2016).
Article CAS PubMed Google Scholar
Ercoskun, P. et al. A new case of Turnpenny-Fry syndrome. Am. J. Med. Genet. A 188, 688–691 (2022).
Article CAS PubMed Google Scholar
Luo, X. et al. Rare deleterious de novo missense variants in Rnf2/Ring2 are associated with a neurodevelopmental disorder with unique clinical features. Hum. Mol. Genet. 30, 1283–1292 (2021).
Article CAS PubMed PubMed Central Google Scholar
Muthusamy, B. et al. Shukla-Vernon syndrome: a second family with a novel variant in the BCORL1 gene. Genes 12, 452 (2021).
Pierce, S. B. et al. De novo mutation in RING1 with epigenetic effects on neurodevelopment. Proc. Natl. Acad. Sci. USA 115, 1558–1563 (2018).
Article ADS CAS PubMed PubMed Central Google Scholar
Shukla, A. et al. Variants in the transcriptional corepressor BCORL1 are associated with an X-linked disorder of intellectual disability, dysmorphic features, and behavioral abnormalities. Am. J. Med. Genet. A 179, 870–874 (2019).
Article CAS PubMed Google Scholar
Turnpenny, P. D. et al. Missense mutations of the Pro65 residue of PCGF2 cause a recognizable syndrome associated with craniofacial, neurological, cardiovascular, and skeletal features. Am. J. Hum. Genet. 103, 786–793 (2018).
Article CAS PubMed PubMed Central Google Scholar
Ryan, C. W. et al. H2A monoubiquitination: insights from human genetics and animal models. Hum. Genet. 143, 511–527 (2023).
McGinty, R. K., Henrici, R. C. & Tan, S. Crystal structure of the PRC1 ubiquitylation module bound to the nucleosome. Nature 514, 591–596 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Fursova, N. A. et al. Synergy between variant PRC1 complexes defines polycomb-mediated gene repression. Mol. Cell 74, 1020–1036.e8 (2019).
Article CAS PubMed PubMed Central Google Scholar
Tamburri, S. et al. Histone H2AK119 mono-ubiquitination is essential for polycomb-mediated transcriptional repression. Mol. Cell 77, 840–856.e5 (2020).
Article CAS PubMed PubMed Central Google Scholar
Bravo, M. et al. Polycomb RING1A- and RING1B-dependent histone H2A monoubiquitylation at pericentromeric regions promotes S-phase progression. J. Cell Sci. 128, 3660–3671 (2015).
CAS PubMed Google Scholar
Ismail, I. H. et al. BMI1-mediated histone ubiquitylation promotes DNA double-strand break repair. J. Cell Biol. 191, 45–60 (2010).
Article CAS PubMed PubMed Central Google Scholar
Ismail, I. H. et al. A small molecule inhibitor of polycomb repressive complex 1 inhibits ubiquitin signaling at DNA double-strand breaks. J. Biol. Chem. 288, 26944–26954 (2013).
Article CAS PubMed PubMed Central Google Scholar
Kakarougkas, A. et al. Requirement for PBAF in transcriptional repression and repair at DNA breaks in actively transcribed regions of chromatin. Mol. Cell 55, 723–732 (2014).
Article CAS PubMed PubMed Central Google Scholar
Shanbhag, N. M. et al. ATM-dependent chromatin changes silence transcription in cis to DNA double-strand breaks. Cell 141, 970–981 (2010).
Article CAS PubMed PubMed Central Google Scholar
Bergink, S. et al. DNA damage triggers nucleotide excision repair-dependent monoubiquitylation of histone H2A. Genes Dev. 20, 1343–1352 (2006).
Article CAS PubMed PubMed Central Google Scholar
Ui, A., Nagaura, Y. & Yasui, A. Transcriptional elongation factor ENL phosphorylated by ATM recruits polycomb and switches off transcription for DSB repair. Mol. Cell 58, 468–482 (2015).
Article CAS PubMed Google Scholar
Deal, C. L. Abnormal fetal growth. In C. S. Kovacs & C. L. Deal (Eds.) Maternal-Fetal and Neonatal Endocrinology 913–938 (Elsevier, 2020).
Klingseisen, A. & Jackson, A. P. Mechanisms and pathways of growth failure in primordial dwarfism. Genes Dev. 25, 2011–2024 (2011).
Article CAS PubMed PubMed Central Google Scholar
Willis, N. & Rhind, N. Regulation of DNA replication by the S-phase DNA damage checkpoint. Cell Div. 4, 13 (2009).
Article PubMed PubMed Central Google Scholar
Karczewski, K. J. et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 581, 434–443 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Wiel, L. et al. MetaDome: pathogenicity analysis of genetic variants through aggregation of homologous human protein domains. Hum. Mutat. 40, 1030–1038 (2019).
CAS PubMed PubMed Central Google Scholar
Chambers, S. M. et al. Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signaling. Nat. Biotechnol. 27, 275–280 (2009).
Article CAS PubMed PubMed Central Google Scholar
Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386 (2014).
Article CAS PubMed PubMed Central Google Scholar
Nestorowa, S. et al. A single-cell resolution map of mouse hematopoietic stem and progenitor cell differentiation. Blood 128, e20–e31 (2016).
Article CAS PubMed PubMed Central Google Scholar
Wilson, P. G. & Stice, S. S. Development and differentiation of neural rosettes derived from human embryonic stem cells. Stem Cell Rev. 2, 67–77 (2006).
Article CAS PubMed Google Scholar
Endoh, M. et al. PCGF6-PRC1 suppresses premature differentiation of mouse embryonic stem cells by regulating germ cell-related genes. Elife 6, e21064 (2017).
Scelfo, A. et al. Functional landscape of PCGF proteins reveals both RING1A/B-dependent-and RING1A/B-independent-specific activities. Mol. Cell 74, 1037–1052.e7 (2019).
Article CAS PubMed PubMed Central Google Scholar
Buckley, N. E. et al. BRCA1 regulates IFN-γ signaling through a mechanism involving the type I IFNs. Mol. Cancer Res. 5, 261–270 (2007).
Article CAS PubMed Google Scholar
Gorodetska, I., Kozeretska, I. & Dubrovska, A. BRCA genes: the role in genome stability, cancer stemness and therapy resistance. J. Cancer 10, 2109 (2019).
Article CAS PubMed PubMed Central Google Scholar
Welcsh, P. L. et al. BRCA1 transcriptionally regulates genes involved in breast tumorigenesis. Proc. Natl. Acad. Sci. USA 99, 7560–7565 (2002).
Article ADS CAS PubMed PubMed Central Google Scholar
Gao, B. et al. Constitutive activation of JAK–STAT3 signaling by BRCA1 in human prostate cancer cells. FEBS Lett. 488, 179–184 (2001).
Article CAS PubMed Google Scholar
Giannakis, M. et al. Genomic correlates of immune-cell infiltrates in colorectal carcinoma. Cell Rep. 15, 857–865 (2016).
Article CAS PubMed PubMed Central Google Scholar
Rizvi, N. A. et al. Mutational landscape determines sensitivity to PD-1 blockade in non–small cell lung cancer. Science 348, 124–128 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Tate, J. G. et al. COSMIC: the Catalogue Of Somatic Mutations In Cancer. Nucleic Acids Res. 47, D941–D947 (2019).
Article CAS PubMed Google Scholar
Raza, M. U. et al. DNA damage in major psychiatric diseases. Neurotox. Res. 30, 251–267 (2016).
Article MathSciNet CAS PubMed PubMed Central Google Scholar
Nishioka, N. & Arnold, S. E. Evidence for oxidative DNA damage in the hippocampus of elderly patients with chronic schizophrenia. Am. J. Geriatr. Psychiatry 12, 167–175 (2004).
Article PubMed Google Scholar
Catts, V. S. et al. Evidence of aberrant DNA damage response signalling but normal rates of DNA repair in dividing lymphoblasts from patients with schizophrenia. World J. Biol. Psychiatry 13, 114–125 (2012).
Article PubMed Google Scholar
Muraleedharan, A. et al. Assessment of DNA damage and repair efficiency in drug naïve schizophrenia using comet assay. J. Psychiatr. Res. 68, 47–53 (2015).
Article PubMed Google Scholar
Cong, L. et al. Multiplex genome engineering using CRISPR/Cas systems. Science 339, 819–823 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Karzbrun, E., Tshuva, R. Y. & Reiner, O. An on-chip method for long-term growth and real-time imaging of brain organoids. Curr. Protoc. Cell Biol. 81, e62 (2018).
Article PubMed Google Scholar
Ramírez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res. 44, W160–W165 (2016).
Article PubMed PubMed Central Google Scholar
Lawrence, M., Gentleman, R. & Carey, V. rtracklayer: an R package for interfacing with genome browsers. Bioinformatics 25, 1841–1842 (2009).
Article CAS PubMed PubMed Central Google Scholar
Finogenova, K. et al. Structural basis for PRC2 decoding of active histone methylation marks H3K36me2/3. Elife 9, e61964 (2020).
Robinson, J. T. et al. Integrative genomics viewer. Nat. Biotechnol. 29, 24–26 (2011).
Article CAS PubMed PubMed Central Google Scholar
Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008).
Article PubMed PubMed Central Google Scholar
Lawrence, M. et al. Software for computing and annotating genomic ranges. PLoS Comput. Biol. 9, e1003118 (2013).
Article CAS PubMed PubMed Central Google Scholar
Dainat, J. AGAT: Another Gff Analysis Toolkit to handle annotations in any GTF/GFF format. (Version v0.7.0). Zenodo (2019).
Gel, B. & Serra, E. karyoploteR: an R/Bioconductor package to plot customizable genomes displaying arbitrary data. Bioinformatics 33, 3088–3090 (2017).
Article CAS PubMed PubMed Central Google Scholar
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Article CAS PubMed Google Scholar
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
Article PubMed PubMed Central Google Scholar
Yu, G., Wang, L. G. & He, Q. Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics 31, 2382–2383 (2015).
Article CAS PubMed Google Scholar
Carlson, M. org.Hs.eg.db: Genome wide annotation for Human. R package version 3.16.0. (2022).
Maintainer, B. C. T. a. B. P. TxDb.Hsapiens.UCSC.hg38.knownGene: Annotation package for TxDb object(s). R package version 3.16.0. (2022).
Aicher, T. P. et al. Seq-well: a sample-efficient, portable picowell platform for massively parallel single-cell RNA sequencing. Methods Mol. Biol. 1979, 111–132 (2019).
Article CAS PubMed PubMed Central Google Scholar
Hughes, T. K. et al. Second-strand synthesis-based massively parallel scRNA-seq reveals cellular states and molecular features of human inflammatory skin pathologies. Immunity 53, 878–894.e7 (2020).
Article CAS PubMed PubMed Central Google Scholar
Macosko, E. Z. et al. Highly parallel genome-wide expression profiling of individual cells using nanoliter droplets. Cell 161, 1202–1214 (2015).
Article CAS PubMed PubMed Central Google Scholar
Hafemeister, C. & Satija, R. Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression. Genome Biol. 20, 296 (2019).
Article CAS PubMed PubMed Central Google Scholar
Stuart, T. et al. Comprehensive integration of single-cell data. Cell 177, 1888–1902.e21 (2019).
Article CAS PubMed PubMed Central Google Scholar
Loo, L. et al. Single-cell transcriptomic analysis of mouse neocortical development. Nat. Commun. 10, 134 (2019).
Article ADS PubMed PubMed Central Google Scholar
Polioudakis, D. et al. A single-cell transcriptomic atlas of human neocortical development during mid-gestation. Neuron 103, 785–801.e8 (2019).
Article CAS PubMed PubMed Central Google Scholar
Di Bella, D. J. et al. Molecular logic of cellular diversification in the mammalian cerebral cortex. bioRxiv https://doi.org/10.1101/2020.07.02.185439 (2020).
Tanaka, Y. et al. Synthetic analyses of single-cell transcriptomes from multiple brain organoids and fetal brain. Cell Rep. 30, 1682–1689.e3 (2020).
Article CAS PubMed PubMed Central Google Scholar
Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA 102, 15545–15550 (2005).
Article ADS CAS PubMed PubMed Central Google Scholar
Gillespie, M. et al. The reactome pathway knowledgebase 2022. Nucleic Acids Res. 50, D687–D692 (2022).
Article CAS PubMed Google Scholar
Vodenkova, S. et al. An optimized comet-based in vitro DNA repair assay to assess base and nucleotide excision repair activity. Nat. Protoc. 15, 3844–3878 (2020).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

Biorender illustrations were incorporated into the manuscript. This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human development (R01AWD010411 to S.L.B.), National Institute of Neurological Disorders and Stroke (R01NS101597 to S.L.B.), Leo’s Lighthouse Foundation to S.L.B., NIH Cellular and Molecular Biology Training Grant (T32-GM007315 to C.W.R.), NRSA Fellowship Grant (F31NS127551 to C.W.R.), MSTP (T32GM007863 to C.W.R.), and Michigan Pre-doctoral Training in Genetics Program (T32-GM007544 to A.M.).

Author information

Authors and Affiliations

Cellular and Molecular Biology Program, University of Michigan Medical School, Ann Arbor, MI, USA
C. W. Ryan & S. L. Bielas
Medical Science Training Program, University of Michigan Medical School, Ann Arbor, MI, USA
C. W. Ryan
Department of Human Genetics, University of Michigan Medical School, Ann Arbor, MI, USA
C. W. Ryan, S. L. Regan, E. F. Mills, B. T. McGrath, E. Gong, Y. T. Lai, J. B. Sheingold, K. Patel, T. Horowitz, A. Moccia, Y. C. Tsan, A. Srivastava & S. L. Bielas
Sanjay Gandhi Postgraduate Institute of Medical Sciences, Department of Medical Genetics, Lucknow, India
A. Srivastava
Department of Pediatrics, University of Michigan Medical School, Ann Arbor, MI, USA
S. L. Bielas
Neuroscience Graduate Program, University of Michigan Medical School, Ann Arbor, MI, USA
S. L. Bielas

Authors

C. W. Ryan
View author publications
You can also search for this author in PubMed Google Scholar
S. L. Regan
View author publications
You can also search for this author in PubMed Google Scholar
E. F. Mills
View author publications
You can also search for this author in PubMed Google Scholar
B. T. McGrath
View author publications
You can also search for this author in PubMed Google Scholar
E. Gong
View author publications
You can also search for this author in PubMed Google Scholar
Y. T. Lai
View author publications
You can also search for this author in PubMed Google Scholar
J. B. Sheingold
View author publications
You can also search for this author in PubMed Google Scholar
K. Patel
View author publications
You can also search for this author in PubMed Google Scholar
T. Horowitz
View author publications
You can also search for this author in PubMed Google Scholar
A. Moccia
View author publications
You can also search for this author in PubMed Google Scholar
Y. C. Tsan
View author publications
You can also search for this author in PubMed Google Scholar
A. Srivastava
View author publications
You can also search for this author in PubMed Google Scholar
S. L. Bielas
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

Conception and design: C.W.R., S.L.R., Y.C.T., A.S. and S.L.B.; Development of methodology: C.W.R., S.L.R., Y.C.T., B.T.M. and S.L.B.; Acquisition of data: C.W.R., S.L.R., Y.C.T., Y.T.L., E.F.M., E.G., T.H., J.B.S., K.P., A.M. and B.T.M.; Analysis of data: C.W.R. and S.L.B.; Data interpretation: C.W.R., S.L.R., Y.C.T., E.F.M. and S.L.B.; Writing of the manuscript: C.W.R. and S.L.B.; All authors contributed to the article and approved the submitted version.

Corresponding author

Correspondence to S. L. Bielas.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Ryan, C.W., Regan, S.L., Mills, E.F. et al. RING1 missense variants reveal sensitivity of DNA damage repair to H2A monoubiquitination dosage during neurogenesis. Nat Commun 15, 7931 (2024). https://doi.org/10.1038/s41467-024-52292-8

Download citation

Received: 15 December 2023
Accepted: 30 August 2024
Published: 10 September 2024
DOI: https://doi.org/10.1038/s41467-024-52292-8

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.