Polycomb repressive-deubiquitinase complex safeguards oocyte epigenome and female fertility by restraining Polycomb activity

Jinwen Kang; Peiyao Liu; Shoko Ichimura; Lauryn Cook; Mengwen Hu; Satoshi H Namekawa; Zhiyuan Chen

doi:10.1038/s41467-026-70845-x

. 2026 Mar 19;17:4149. doi: 10.1038/s41467-026-70845-x

Polycomb repressive-deubiquitinase complex safeguards oocyte epigenome and female fertility by restraining Polycomb activity

Jinwen Kang ^1,², Peiyao Liu ^1,², Shoko Ichimura ^1,², Lauryn Cook ^1,², Mengwen Hu ³, Satoshi H Namekawa ³, Zhiyuan Chen ^1,^2,^✉

PMCID: PMC13153255 PMID: 41851118

Abstract

Mouse oocytes exhibit a unique chromatin landscape characterized by broad H3K27ac and H3K27me3 domains, demarcating euchromatin and facultative heterochromatin, respectively. However, the mechanisms underlying this non-canonical landscape remain elusive. Here we report BAP1, a core component of the Polycomb Repressive-Deubiquitinase (PR-DUB) complex, as a key negative regulator of Polycomb activity during oogenesis. BAP1 restricts pervasive H2AK119ub1 accumulation and protects oocyte-specific broad H3K27ac, particularly within gene-poor regions, from ectopic H3K27me3 deposition. While PR-DUB has been linked to gene repression, in oocytes BAP1 primarily promotes transcription and contributes minimally to Polycomb-mediated silencing. BAP1-dependent transcriptional activation during oogenesis is essential for oocyte developmental competence, maternal-to-zygotic transition, and female fertility. Notably, ectopic H3K27me3 domains established in BAP1-deficient oocytes persist in preimplantation embryos but are resolved after implantation, and loss of maternal BAP1 does not impair either canonical or non-canonical genomic imprinting. Together, these findings reveal a critical role for PR-DUB in safeguarding the oocyte epigenome by protecting euchromatin from ectopic Polycomb activity, rather than enforcing transcriptional repression.

Subject terms: Oogenesis, Epigenomics, Gene silencing

Chromatin landscape of mouse oocytes is characterized by broad H3K27ac and H3K27me3 domains, demarcating euchromatin and facultative heterochromatin. This study reports that BAP1 restricts pervasive H2AK119ub1 accumulation in mouse oocytes and protects broad H3K27ac from ectopic H3K27me3 deposition. Loss of BAP1 severely impairs oocyte developmental competence and female fertility.

Introduction

Polycomb group (PcG) proteins play a central role in establishing and maintaining epigenetic silencing during animal development^1,2. PcG proteins assemble into two major complexes, Polycomb Repressive Complex 1 (PRC1) and PRC2, which function cooperatively to mediate transcriptional repression^3,4. In mammals, PRC1 contains the E3 ubiquitin ligases RING1A and RING1B, which catalyze monoubiquitination of histone H2A at lysine 119 (H2AK119ub1)^5,6. The PRC2 comprises the methyltransferases EZH1 and EZH2, which catalyze mono-, di-, and tri-methylation of histone H3 at lysine 27 (H3K27me1/2/3)^7–9. These complexes are typically recruited to promoters of key developmental regulators where they deposit H2AK119ub1 and H3K27me3, hallmarks of facultative heterochromatin^10,11. Collectively, PRC1 and PRC2 mediate epigenetic silencing by promoting chromatin compaction¹² and counteracting transcription¹³.

Oogenesis provides a unique in vivo model to investigate chromatin-mediated gene regulation, as postnatal oocytes grow for extended periods of time without undergoing cell division. Shortly after birth, oocytes are enclosed in primordial follicles and arrested at the diplotene stage of meiotic prophase I, known as the germinal vesicle (GV) stage. After primordial follicle activation, they enter growth phase to progress from primary follicles to pre-antral follicles and pre-ovulatory follicles. During this growth phase (~3-weeks in mouse), GV oocytes undergo significant cytoplasmic and nuclear enlargement without cell cycle progression and ultimately mature into fully grown GV oocytes (FGOs). During this time, oocytes accumulate maternal RNAs and proteins essential for meiotic resumption and early embryonic development¹⁴.

In parallel with transcriptional and translational changes, oocytes acquire a distinct epigenome, including the formation of broad, non-canonical domains of H2AK119ub1 and H3K27me3^15–19. Notably, PRC1 and PRC2 have different functions during postnatal oocyte development. Loss of PRC1 in oocytes leads to widespread gene de-repression and disrupts the organization of Polycomb-associated three-dimensional chromatin domains^20,21. As a result, PRC1-null oocytes fail to support embryonic development beyond the two-cell stage²⁰. In contrast, maternal loss of PRC2 has only a modest effect on oogenesis and primarily impacts non-canonical genomic imprinting dependent on maternal H3K27me3, with developmental consequences emerging after implantation^22–28. Despite these insights, how broad Polycomb domains are established and maintained during oocyte postnatal growth remains unclear. In particular, the mechanisms that restrict the spread of facultative heterochromatin into euchromatic regions during this period are poorly understood.

In addition to Polycomb domains, FGOs exhibit pervasive acetylation of histone H3 at lysine 27 (H3K27ac), a marker of active euchromatin^29,30. This non-canonical, widespread H3K27ac is mutually exclusive with H3K27me3 domains, spatially segregating euchromatin and facultative heterochromatin in oocytes³¹. The broad H3K27ac feature is also found in early-stage human preimplantation embryos³². Remarkably, many FGO-specific putative enhancers, marked by distal H3K27ac, reside in gene-poor regions and are linked to oocyte-specific gene activation²⁹. However, the mechanisms governing the formation and function of these oocyte-specific broad H3K27ac domains remain largely unknown. These observations highlight the current knowledge gaps of how chromatin regulators coordinate the formation of spatially distinct active and repressive domains during oocyte growth.

The ubiquitin carboxy-terminal hydrolase BAP1 is the catalytic subunit of the Polycomb Repressive-Deubiquitinase (PR-DUB) complex, which removes monoubiquitin from H2AK119ub1^33–35. In Drosophila embryos, loss of BAP1 leads to elevated H2AK119ub1 levels at Polycomb target genes and disrupts gene repression by impairing PRC1-mediated chromatin compaction³⁶. In mouse embryonic stem cells (mESCs), PR-DUB broadly promotes gene activation by limiting pervasive deposition of H2AK119ub1 across the genome, yet it is also responsible for Polycomb repression via a mechanism distinct from that in Drosophila^37–40. By contrast, studies in human cell lines have revealed that BAP1 functions as a transcriptional activator, acting to restrict PRC1-mediated H2AK119ub1 accumulation at gene regulatory elements, including enhancers^41,42. These findings underscore the context-dependent role of BAP1 in gene regulation, suggesting that it may act to support both Polycomb repression and transcriptional activation, depending on cellular and developmental context. Thus, the precise mechanisms by which BAP1 regulates transcription remain enigmatic, and it is essential to dissect its in vivo function in balancing Polycomb-mediated silencing and transcriptional activation.

In this study, we aimed to determine the role of PR-DUB in regulating H2AK119ub1 in oocytes and early embryos, and to dissect how its activity contributes to the formation of facultative heterochromatin and euchromatin during oogenesis. We reveal that PR-DUB safeguards the oocyte epigenome and female fertility by protecting euchromatin from ectopic Polycomb activity, rather than by enforcing transcriptional repression.

Results

Identification of candidate H2AK119ub1 deubiquitinases in early embryos

To identify candidate DUBs responsible for H2AK119ub1 removal in vivo, we first analyzed the expression dynamics of known H2A DUBs during oogenesis and early embryonic development using publicly available RNA-seq and ribosome profiling datasets^43,44. Based on high expression levels, ribosome association in oocytes and preimplantation embryos, and lack of activity toward monoubiquitinated histone H2B, we selected five H2A DUBs for further analysis: USP16⁴⁵, USP28⁴⁶, USP21⁴⁷, BAP1³³, and MYSM1⁴⁸ (Supplementary Fig. 1a–d). To assess their functional activity toward H2AK119ub1, we microinjected zygotes with Flag-tagged mRNAs encoding each candidate immediately after fertilization and evaluated H2AK119ub1 levels by immunostaining in late zygotes (~6 h post-injection) (Fig. 1a). Subcellular localization analyses revealed that USP16 and USP21 were primarily cytoplasmic, whereas BAP1, USP28, and MYSM1 exhibited no strong preference between nucleus and cytoplasm (Fig. 1b).

Among these candidates, USP21 overexpression led to a robust depletion of H2AK119ub1 within 6 h (Fig. 1b, c). In contrast, BAP1 alone did not affect H2AK119ub1 levels, but co-expression with ASXL1 significantly depleted the H2AK119ub1 signal. This finding aligns with previous studies showing that ASXL1 is required to activate BAP1’s H2A DUB function⁴⁹. Overexpression of USP16, USP28, or MYSM1 had a mild impact, if any, on H2AK119ub1 levels. Together, these data suggest that USP21 and BAP1 potentially regulate H2AK119ub1 during early development. Since Usp21 null mice are viable and fertile^50,51, we selected BAP1 for detailed loss-of-function analysis in this study. It should be noted that we cannot exclude the possibility that USP16 (discussed further below), USP28, and MYSM1 may function as H2A DUBs at other developmental stages or require additional cofactors in zygotes. Future loss-of-function studies will be necessary to determine their contributions to H2AK119ub1 regulation during oogenesis and early embryonic development.

BAP1 limits genome wide pervasive H2AK119ub1 accumulation in oocytes

Having identified BAP1 as a potential key H2A DUB in zygotes, we next examined the expression dynamics of other PR-DUB subunits during oogenesis and early embryonic development. Reanalysis of publicly available RNA-seq and ribosome profiling datasets^43,44 revealed that key PR-DUB components, including Bap1, Asxl1, Foxk2, and Ogt, are highly expressed and associated with ribosomes throughout oocyte maturation and preimplantation stages (Fig. 1d and Supplementary Fig. 1e). Consistent with these findings, immunoblot analysis confirmed the presence of BAP1 protein in FGOs and early embryos (Fig. 1e).

To investigate the function of PR-DUB during oogenesis and early development, we generated an oocyte-specific conditional knockout (CKO) of Bap1 by crossing a floxed Bap1 line (see Methods), in which exons 6–12 are flanked by loxP sites, with a Gdf9-iCre transgenic line⁵² (Supplementary Fig. 2a). Gdf9-iCre is specifically expressed in oocytes of primordial follicles by postnatal day 3⁵², enabling targeted deletion of Bap1 in early-stage oocytes. Cre-mediated recombination is predicted to produce a truncated BAP1 protein lacking catalytic activity and the ability to interact with ASXL partners⁴⁹ (Supplementary Fig. 2b). Throughout this study, Bap1^fl/fl mice were used as controls (CTR), and Gdf9-iCre; Bap1^fl/fl mice were used as CKO. At 6-9 weeks of age, a comparable number of FGOs were recovered from CTR and CKO female mice (Supplementary Fig. 2c, d). RNA-seq and immunoblot analyses confirmed effective depletion of Bap1 at both the transcript and protein levels in CKO FGOs (Fig. 1f and Supplementary Fig. 2e).

Immunofluorescence analyses revealed that H2AK119ub1 levels in CKO FGOs were significantly elevated, approximately 2- to 3-fold, compared to those in CTR oocytes (Fig. 1g, h). To determine the genomic regions where H2AK119ub1 increases, we performed spike-in normalized Cleavage Under Targets and Release Using Nuclease (CUT&RUN) (see “Methods”; Supplementary Fig. 2f). Remarkably, this analysis uncovered a pervasive accumulation of H2AK119ub1 across the genome (Fig. 1i, j and Supplementary Fig. 2g). The extent of H2AK119ub1 elevation in BAP1-deficient oocytes appeared more pronounced than that in mESCs³⁷ (Fig. 1k). Notably, putative enhancers, defined by distal H3K27ac peaks²⁹, exhibited a particularly strong increase in H2AK119ub1 (Fig. 1k), suggesting a role of BAP1 in regulating enhancer activities. Together, these findings indicate that BAP1 functions to restrain widespread H2AK119ub1 accumulation in oocytes.

BAP1 primarily activates gene expression in oocytes

We next investigated transcriptomic changes resulting from Bap1 deletion in FGOs. Comparative analysis identified 101 upregulated and 827 downregulated genes in CKO FGOs (Fig. 1l and Supplementary Data 1). The predominance of downregulated genes suggests that BAP1 primarily acts as a gene activator in oocytes, likely by counteracting H2AK119ub1, a key player in PRC1-mediated gene silencing^53,54. Downregulated genes are enriched for gene ontology (GO) terms implicated in oogenesis, such as “cell adhesion molecule binding” (Cdh2, Dsc2, Cdh18) and “growth factor binding” (Fgf10, Egfr, Foxp2) (Supplementary Fig. 2h). To link the observed transcriptional dysregulation to changes in H2AK119ub1, we examined H2AK119ub1 levels at these genes. While downregulated genes had similar baseline H2AK119ub1 enrichment in CTR oocytes, they acquired significantly higher levels of H2AK119ub1 in CKO FGOs compared to non-differentially expressed genes (DEGs) (Fig. 1m, n). Notably, this increase occurred across both genic and intergenic regions.

It has been previously reported that the methyl-binding protein MBD6 recruits PR-DUB to retrotransposons by recognizing m⁵C-modified chromatin-associated RNAs, a process critical for retrotransposon activation⁵⁵. However, we found that BAP1 deficiency in oocytes had only a modest effect on the expression of repeat elements (Supplementary Fig. 2i and Supplementary Data 1). This suggests a limited role for MBD6-mediated PR-DUB recruitment in retrotransposon regulation during oogenesis, consistent with the low RNA-seq and Ribo-seq expression levels of Mbd6 in oocytes (Fig. 1d and Supplementary Fig. 1e). In sum, these findings indicate that the widespread accumulation of H2AK119ub1 following BAP1 loss primarily leads to gene repression, with a subset of genes being more susceptible than others (further discussed below).

BAP1 preserves H3K27ac by preventing ectopic H3K27me3 deposition in oocytes

Previous studies have shown that H2AK119ub1 guides H3K27me3 deposition^56–59, and that H3K27me3 and H3K27ac are mutually exclusive. We therefore examined how BAP1 loss affects the H3K27me3 and H3K27ac landscapes in oocytes. Unlike the dramatic increase in H2AK119ub1, immunofluorescence analyses revealed that global levels of both H3K27me3 and H3K27ac remained largely unchanged in CKO FGOs (Fig. 2a, b). This observation was further supported by spike-in normalized CUT&RUN analyses, which confirmed that most broad H3K27me3 and H3K27ac domains, known to be oocyte-specific^15,29, were retained in CKO FGOs (Fig. 2c and Supplementary Fig. 3a).

Fig. 2 — a Immunofluorescence images of H3K27ac and H3K27me3 in nuclei of *Bap1* control (CTR) and conditional knockout (CKO) fully grown oocytes (FGOs). Number of FGOs analyzed are indicated. Scale bar: 10 μm. b Quantification of H3K27ac and H3K27me3 intensities from (a). Number of FGOs analyzed are indicated in panel (a). Boxplot: center line, median; box limits, 25th–75th percentiles; whiskers, ±1.5× interquartile range. p-value: two-sided Student’s t-test. c Heatmaps showing CUT&RUN signal intensity of H3K27ac and H3K27me3 across H3K27ac (n = 10,363) and H3K27me3 (n = 15,751) domains defined by the HMM-based method in *Bap1* CTR and CKO FGOs (see “Methods”). d Genome browser view depicting H3K27ac and H3K27me3 signals at the indicated genomic locus. Regions with significantly reduced H3K27ac levels in CKO FGOs (“H3K27ac-lost regions”) are indicated. e Heatmap illustrating changes of histone modifications at the H3K27ac-lost (n = 9043) and H3K27ac-no change (n = 27,287) regions in *Bap1* CTR and CKO FGOs. f Genomic annotation of the H3K27ac-lost and H3K27ac-no change regions in *Bap1* CKO FGOs. g RNA expression changes for genes associated with the indicated H3K27ac region categories following *Bap1* deletion. Number of genes analyzed for each category are indicated. Boxplot format as in (b). p-value: two-sided Wilcoxon rank-sum test. h Stacked bar plot showing the proportions of H3K27ac-lost and unchanged regions located in gene-rich versus gene-poor genomic regions. Chi-square test: ***p = 5.28e-271. i Heatmap showing H3K27ac signal during oocyte growth across H3K27ac-lost regions. The H3K27ac data for postnatal day 7 (P7), P10 growing oocytes, and FGOs are from public datasets²⁹. j Genome browser view of H3K27ac and H3K27me3 profiles at the *Fgf10* locus. k RNA expression levels of downregulated (n = 818) and unchanged genes (n = 11,227) upon *Bap1* depletion in wild type growing oocytes (GO1 and GO2) and FGOs. GO1 and GO2 are oocytes from early secondary and secondary follicles, respectively. GO1/GO2/FGO RNA-seq are from public datasets⁹². Boxplot format as in (b). p-value: two-sided Wilcoxon rank-sum test; ns: not significant. Source data are provided as a Source data file.

However, closer inspection of genome browser tracks revealed specific loci where H3K27ac was reduced, and H3K27me3 correspondingly increased in CKO FGOs (Fig. 2d and Supplementary Fig. 3b). To systematically characterize these changes, we identified 9043 H3K27ac-lost regions (ranging from 10 to 375 kb) in CKO oocytes (see “Methods,” Supplementary Data 2). Loss of H3K27ac in Bap1 CKO FGOs is globally associated with a corresponding gain of H3K27me3 (Fig. 2e). In addition, the H3K27ac-lost regions already displayed relatively lower H3K27ac and higher H3K27me3 in CTR oocytes compared to regions with unchanged H3K27ac (Supplementary Fig. 3c). H2AK119ub1 levels increased similarly in both H3K27ac-lost and H3K27ac-unchanged regions in CKO FGOs (Fig. 2e and Supplementary Fig. 3c), suggesting that the increase in H2AK119ub1 is not the sole determining factor that dictates which regions lose H3K27ac and gain H3K27me3. Most H3K27ac-lost regions (69%) were in intergenic regions and gene bodies, suggesting potential enhancer activity (Fig. 2f). Genes associated with H3K27ac loss, either at promoters or within gene bodies, were preferentially downregulated in CKO FGOs (Fig. 2g). Together, these results indicate that BAP1 deficiency causes H3K27ac loss and H3K27me3 gain at selective loci, contributing to gene repression in oocytes.

To better understand the mechanisms underlying locus specific H3K27ac loss in CKO FGOs, we examined the genomic features of the affected regions. A striking characteristic of the H3K27ac-lost regions is their preferential localization within gene-poor regions, as opposed to gene-rich regions (Figs. 2d, h and Supplementary Fig. 3b; see “Methods”). Previous studies have shown that H3K27ac peaks in gene-poor regions are more prevalent in FGOs compared to growing oocytes²⁹. Consistently, we found that these H3K27ac-lost regions exhibit low H3K27ac levels in non-grown oocytes at postnatal day 1 (P1) and growing oocytes at P7 and P10 but gain high levels of H3K27ac during the final stage of oocyte growth, as seen in FGOs (Fig. 2i, j and Supplementary Fig. 3d). In contrast, regions that unaffected by Bap1 deletion acquire H3K27ac even before Gdf9-iCre activation on P3 (Fig. 2j and Supplementary Fig. 3d). These findings suggest that BAP1 is critical for the establishment of new H3K27ac marks during oocyte growth, rather than for maintaining pre-existing H3K27ac marks. Supporting this idea, only 9.9–13.4% of H3K27ac-lost regions overlapped with H3K27ac peaks in P7/P10 growing oocytes, whereas this overlap increased to 49–57.9% for H3K27ac-unchanged regions (Supplementary Fig. 3e). Further supporting this model, genes unaffected by Bap1 deletion are already highly expressed early during oocyte growth, whereas genes downregulated in CKO FGOs reach peak expression only at the fully grown stage and show lower expression levels than the other group (Fig. 2k and Supplementary Fig. 3f). Together, these data suggest that the timing of H3K27ac establishment during oocyte development underlies the locus specific H3K27ac loss observed in Bap1 CKO FGOs.

BAP1 plays a limited role in Polycomb-mediated silencing in oocytes

Having established that BAP1 preserves oocyte-specific broad H3K27ac domains by preventing ectopic H3K27me3 deposition, we next examined its potential role in Polycomb-mediated gene silencing. In mESCs, it has been proposed that widespread accumulation of H2AK119ub1 in PR-DUB mutants may sequester PRC2, thereby diminishing H3K27me3 deposition at canonical Polycomb group (PcG) target loci^37–39. To evaluate this model in oocytes, we assessed H2AK119ub1, H3K27me3, and H3K27ac profiles at the promoters of oocyte PcG target genes as defined previously⁶⁰. This analysis revealed a marked increase in H2AK119ub1 at PcG target promoters in Bap1 CKO FGOs, accompanied by a relatively modest decrease in H3K27me3 and a corresponding gain in H3K27ac (Fig. 3a, b).

Fig. 3 — a Metaplots displaying signal intensities of H2AK119ub1, H3K27me3, and H3K27ac at 1700 oocyte Polycomb group (PcG) target genes identified in ref. ⁶⁰ in *Bap1* control (CTR) and conditional knockout (CKO) fully grown oocytes (FGOs). b Genome browser views of EED, EZH2, H2AK119ub1, H3K27me3, and H3K27ac signals at the PcG target loci *Plxnd1* and *Gata5*. Public EED and EZH2 CUT&RUN data are included for comparison⁶⁰. c Venn diagram showing the overlap of oocyte PcG target genes that are de-repressed in PRC1, PRC2, and *Bap1* CKO FGOs. PRC1/2 RNA-seq is from public datasets²¹. d Heatmap depicting expression levels of PcG targets that are de-repressed in *Bap1* CKO FGOs, compared across PRC1 and PRC2 CKO oocytes.

To assess whether the observed chromatin alterations could lead to transcriptional de-repression of PcG targets, we re-analyzed gene expression in Bap1 CKO FGOs. For comparison, we examined publicly available RNA-seq datasets from PRC1 and PRC2 CKO oocytes²¹. Consistent with the established importance of PRC1 in postnatal oocyte development²⁰, PRC1 loss led to a greater number of DEGs (2,417 upregulated and 700 downregulated) compared to PRC2 loss (414 upregulated and 67 downregulated) (Supplementary Fig. 4a, b). Among these, 479 and 83 PcG targets were de-repressed in PRC1 and PRC2 CKO oocytes, respectively (Fig. 3c). By contrast, only 7 PcG targets were upregulated in Bap1 CKO FGOs, and the extent of their upregulation was substantially lower than that observed in PRC1/2 mutants (Fig. 3d). Analyses of PcG targets previously defined in mESCs⁶¹ also led to a similar conclusion (Supplementary Fig. 4c, d). These findings suggest that, although the global accumulation of H2AK119ub1 in BAP1-deficient oocytes may partially impair PRC2 recruitment and reduce H3K27me3 deposition at PcG targets, the overall integrity of Polycomb-mediated silencing is largely preserved. Taken together, these results indicate that BAP1 plays a limited role in the regulation of Polycomb-dependent gene silencing in oocytes.

Loss of maternal BAP1 severely impairs preimplantation development

Having established the critical role of BAP1 in safeguarding the oocyte epigenome and transcriptome, we next assessed how maternal BAP1 deficiency may affect female fertility. To this end, CTR and CKO females were co-caged with wild-type males for six months. Bap1 CKO females exhibited a significant reduction in fertility that was attributable to both fewer litters and smaller litter sizes (Fig. 4a–c). Ovarian morphologies of Bap1 CKO females appear largely normal (Supplementary Fig. 5a), which is consistent with the fact that comparable numbers of FGOs are retrieved from control and CKO females at 6–9 weeks of age (Supplementary Fig. 2c, d). However, in vitro meiotic maturation analysis revealed a mild reduction in the metaphase II rate in CKO oocytes (Supplementary Fig. 5b, c). Additionally, the number of embryos recovered at embryonic day 3.5 (E3.5) was slightly smaller in the CKO group (Fig. 4d), suggesting that ovulation is moderately impaired. Importantly, the embryos from Bap1 CKO females showed defective preimplantation development, with most (~90%) delayed at the 8–16-cell or morula stage at E3.5, whereas control embryos (~96%) progressed to the blastocyst stage by this time point (Fig. 4e, f). Although extended ex vivo culture for an additional 24 h allowed most mutant embryos to form blastocysts, these were of poor quality, exhibiting reduced total cell numbers and abnormal expression of lineage markers, including NANOG (epiblast), GATA4 (primitive endoderm), and CDX2 (trophectoderm) (Fig. 4g, h).

Fig. 4 — a Fertility assessment of *Bap1* control (CTR, n = 6) and conditional knockout (CKO, n = 5) female mice. Starting at 7 weeks of age, females were continuously housed with wild-type B6 males for six months. Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. b Total number of litters produced during the 6-month mating trial for *Bap1* CTR (n = 6) and CKO (n = 5) females. Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. c Average number of pups per litter. A total of 36 and 18 litters were analyzed for CTR and CKO groups, respectively. Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. d Number of E3.5 embryos per litter from *Bap1* CTR (n = 10) and CKO (n = 7) females. Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. e Brightfield images of E3.5 embryos collected from *Bap1* CTR and CKO females. Embryos were cultured ex vivo for an additional 24 hours. Scale bar, 100 μm. f Percentages of CTR and maternal knockout (matKO) embryos that reached the blastocyst stage at E3.5 and after 24-h culture. At E3.5, 79 embryos from 10 litters (CTR) and 41 embryos from 7 litters (matKO) were analyzed. For E3.5 + 24 h culture, 36 embryos from 4 litters (CTR) and 19 embryos from 4 litters (matKO) were analyzed. Data are presented as mean ± SD. Chi-square test. g Immunofluorescence staining of CTR and matKO blastocysts (E3.5 + 24 h culture) using antibodies against NANOG, GATA4, and CDX2. The number of blastocysts analyzed is indicated. Scale bar, 20 μm. h Quantification of cell numbers for blastocysts shown in (j). Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. i Number of implantation sites per litter at embryonic day 6.5 (E6.5) in *Bap1* CTR (n = 5) and CKO (n = 5) groups. Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. j Number of embryos recovered at E6.5 for *Bap1* CTR (n = 5) and CKO (n = 5) females. Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. k Pictures of one CTR and one matKO litter at E6.5. Scale bar, 1 mm. l Number of viable pups per litter at E18.5 in *Bap1* CTR (n = 6) and CKO (n = 8) groups. Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. m Embryonic resorption rates at E18.5 for *Bap1* CTR (n = 6) and CKO (n = 8) females. Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. n Fetal weights at E18.5. CTR: 58 pups from six litters; CKO: 31 pups from eight litters. Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. o Placental weights at E18.5. Sample sizes match those in (n). Data are presented as mean ± SD. P-value calculated by two-sided Student’s t-test. p Schematics of IVF groups. Mat: maternal; Pat: paternal. q Brightfield images of blastocysts of the indicated groups at 96 and 120 h post IVF (hpi). Scale bar, 100 μm. r Stacked bar plot showing developmental progression of IVF embryos at specified time points. 26 (CTR), 55 (matKO), 28 (patKO), and 39 (mzKO) embryos were analyzed. 1 C: one-cell stage. *p = 0.06, **p < 0.001, chi-square test. Source data are provided as a Source data file.

To further evaluate developmental competence of the blastocysts, we dissected embryos at E6.5 and found that the average number of decidua (5.2 ± 3.3) and recovered embryos (2.2 ± 3.2) in the matKO group was lower than that in controls (10.0 ± 1.2, 9.2 ± 1.8)(Fig. 4i, j). The matKO E6.5 embryos were overall smaller than CTR embryos (Fig. 4k). These data are consistent with the compromised blastocyst development of matKO embryos. We next performed cesarean section (C-section) at E18.5 and found that only 3.9 ± 2.0 matKO embryos reached term, representing just 41.4% of the number observed in CTR females (9.7 ± 1.5) (Fig. 4l, m and Supplementary Fig. 5d), indicating a high rate of post-implantation embryonic loss likely attributable to compromised blastocyst quality. Notably, both placentae and fetuses from the maternal knockout (matKO) group were slightly heavier than those of controls (Fig. 4n, o), possibly due to reduced litter size. Thus, these data indicate that BAP1-deficient oocytes could not support normal preimplantation development to give rise to competent blastocysts.

We next sought to determine when developmental delays first arise in matKO embryos during preimplantation development. By monitoring embryos derived through in vitro fertilization (IVF), we observed that developmental delays began as early as the 2- to 4-cell transition (Fig. 4p–r). The delay became more pronounced by 72 h post-IVF (hpi), at which point only ~30% of CKO embryos had reached the morula stage. Ultimately, only ~30% of CKO embryos developed into blastocysts, in contrast to the ~85% blastocyst rate observed in the control group. The more pronounced developmental arrest observed in IVF-derived matKO embryos likely reflects the increased sensitivity of mutant embryos to suboptimal ex vivo culture conditions relative to the in vivo environment.

Because the paternal Bap1 allele remains intact in matKO embryos (Fig. 4p), it is possible that Bap1 expression from the paternal genome during zygotic genome activation (ZGA) at the 2-cell stage partially compensates for the maternal loss. To test this hypothesis, we generated Prm-Cre; Bap1^fl/fl male mice, in which Bap1 is deleted during the late haploid spermatid stage. When Bap1 CKO oocytes were fertilized with sperm from these males, the resulting embryos, referred to as maternal-zygotic knockout (mzKO), displayed markedly more severe developmental defects (Fig. 4q–r). Few mzKO embryos progressed beyond the 2- to 4-cell stages. The observed phenotype was not due to intrinsic defects in Prm-Cre; Bap1^fl/f sperm, as their fertilization of control oocytes resulted in normal progression to the blastocyst stage (Fig. 4q–r). Together, these findings demonstrate that maternal BAP1 is essential for preimplantation development, and loss of both maternal and zygotic BAP1 further exacerbates the phenotype.

Loss of maternal BAP1 causes maternal-to-zygotic transition (MZT) defects

Having established that maternal BAP1 is essential for preimplantation development, we next sought to investigate the underlying mechanisms responsible for the developmental arrest and/or delay observed in Bap1 matKO embryos. The MZT is a critical developmental window during which maternal RNAs and proteins are degraded, and embryonic development becomes increasingly dependent on zygotic gene expression initiated during ZGA⁶². Minor ZGA occurs at late 1-cell (L1C) and early 2-cell (E2C) stages with the activation of hundreds of genes, whereas major ZGA takes place at late 2-cell (L2C) stage, during which thousands of genes are transcriptionally activated^63,64.

Given that the developmental defect in Bap1 matKO embryos emerges as early as the 2- to 4-cell transition, we hypothesized that loss of maternal BAP1 disrupts the MZT. To test this hypothesis, we performed total RNA sequencing of control and Bap1 mutant oocytes/embryos at the MII stage, L1C, E2C, and L2C stages (Fig. 5a and Supplementary Fig. 6a). Comparative transcriptome analyses identified 886, 869, 1096, and 2740 DEGs between CTR and mutant groups at the MII stage, L1C, E2C, and L2C stages, respectively (Fig. 5b and Supplementary Data 1). The number of DEGs was comparable across the GV, MII, and L1C stages, with a predominant trend of gene downregulation in Bap1 mutants (Figs. 1i and 5b). Notably, ~67% (511 genes) of L1C-downregulated genes were already repressed in mutant oocytes (Fig. 5c), and these genes were enriched for GO terms related to cell adhesion (e.g., Cdh2, Cdh18, Cdh20, Dsc2) and growth factor binding (e.g., Gata3, Fgf10, Egfr, Foxp2) (Supplementary Fig. 6b). Thus, these data suggest the aberrant transcriptomes in Bap1 CKO oocytes are inherited into 1-cell embryos.

Fig. 5 — a Schematic of sample collection timeline for total RNA-seq analysis. ZGA: zygotic genome activation; MII: metaphase II eggs. b Scatter plots comparing gene expression between *Bap1* control (CTR) and conditional knockout (CKO) oocytes/embryos at the indicated developmental stages. Red and blue dots indicate significantly upregulated and downregulated genes in *Bap1* mutant, respectively. Differential expression was defined by fold change ≥ 2, adjusted p < 0.05, and FPKM ≥ 0.5. The percentages of differentially expressed genes of all the detectable genes (FPKM ≥ 0.5) are indicated. c Venn diagram illustrating the overlap of downregulated genes in *Bap1* mutant samples at GV, MII, and L1C stages. d Bubble plot showing enrichment of differentially expressed genes in selected gene categories from the DBTMEE database. Statistical significance determined using a hypergeometric test; adjusted p-values are indicated. e Pie charts showing the proportion of downregulated ZGA genes and upregulated maternal genes in *Bap1* matKO embryos. Gene categories were defined according to Wang et al.³⁰. f Balloon plot depicting RNA expression dynamics of representative maternal, minor ZGA, and major ZGA genes at E2C and L2C stages. Three biological replicates were analyzed for each condition. g GO terms enriched among downregulated genes in *Bap1* matKO L2C embryos. h Venn diagrams illustrating the overlap of differentially expressed genes between *Bap1* and *Usp16* maternal KO embryos. RNA-seq data for *Usp16* samples are from public datasets⁷⁷. p value: hypergeometric test.

We next analyzed compositions of DEGs at E2C and L2C stages. Downregulated genes at these stages significantly overlapped with those classified as “minor ZGA,” “major ZGA,” and “2-cell transient” in the DBTMEE v2 database⁶⁵, while upregulated genes were enriched for transcripts categorized as “maternal RNA” (Fig. 5d), implicating defects in both zygotic activation and maternal RNA decay. To further assess the impact of maternal BAP1 loss on ZGA, we compared DEGs in Bap1 mutants with previously defined zygote and E2C minor ZGA gene sets³⁰. Among 23 zygote minor ZGA genes detected in our dataset, three (Sox6, Stk39, Lcp2) were significantly downregulated in Bap1 matKO embryos at L1C (Fig. 5e), with additional minor ZGA genes such as Dux and Usp17la/c/d showing a trend of downregulation (Supplementary Data 1). Furthermore, 22% of minor ZGA genes and 32% of major ZGA genes were downregulated at the E2C and L2C stages, respectively (Fig. 5e, f), including key regulators of early development such as Zscan4, Cdk2ap1, Nr5a2, and Obox3/6 (Fig. 5f)^66–73. Downregulated genes at L2C were also significantly enriched for GO terms related to ribosome biogenesis and rRNA processing (Fig. 5g), potentially accounting for the developmental defects at the 2- to 4-cell transition. In addition, MT2/MERVL retrotransposons, known to be essential for early development^74–76, were markedly downregulated in Bap1 mutants at the E2C stage (Supplementary Fig. 6c). Interestingly, a subset of E2C minor ZGA genes showed aberrant upregulation at L2C, possibly reflecting delayed downregulation (Fig. 5f). Lastly, 13.5% of maternal transcripts were upregulated at the L2C stage, indicating defective maternal RNA degradation (Fig. 5e). Collectively, these findings demonstrate that loss of maternal BAP1 causes defects in both ZGA and maternal transcript clearance during MZT.

Ubiquitin-specific peptidase 16 (USP16) has previously been implicated in H2A deubiquitination during the GV-to-MII transition and is essential for mouse ZGA⁷⁷. To evaluate whether USP16 and BAP1 act through similar or distinct mechanisms during oogenesis and early embryogenesis, we reanalyzed the published transcriptomic data. To our surprise, there was minimal overlap between DEGs in Usp16 and Bap1 mutants at both L1C and L2C stages (Fig. 5h and Supplementary Fig. 6d), suggesting that maternal USP16 and BAP1 function via distinct regulatory pathways. Consistent with this notion, loss of BAP1, but not USP16, causes increase in H2AK119ub1 in FGOs (Fig. 1)⁷⁷. These findings underscore the mechanistic divergence between USP16 and BAP1 in regulating H2A deubiquitination during development and highlight the need for further investigation into their individual and potentially complementary roles.

Maternal BAP1 regulates embryonic enhancer activity during ZGA

Having demonstrated MZT defects in Bap1 matKO embryos, we next investigated how the loss of maternal BAP1 affects chromatin dynamics during early embryogenesis. To this end, we performed immunostaining for H2AK119ub1, H3K27me3, and H3K27ac in CTR and Bap1 matKO embryos at the 1-cell (6 and 10 hpi), L2C (28 hpi), and 4-cell (48 hpi) stages. Since developmental arrest in Bap1 matKO embryos becomes evident as early as the 2- to 4-cell transition, only morphologically normal 4-cell embryos were included in the analysis. Similar to FGOs, H2AK119ub1 levels were significantly elevated in Bap1 matKO embryos compared to CTR at all stages examined (Fig. 6a and Supplementary Fig. 7). In contrast, H3K27me3 immunostaining levels were largely comparable between the two groups. Interestingly, H3K27ac signals were markedly reduced at the 2-cell and 4-cell stages in the absence of maternal BAP1 (Fig. 6a and Supplementary Fig. 7), suggesting impaired formation of active euchromatin that likely occurs independently of H3K27me3.

Fig. 6 — a Quantification of H2AK119ub1, H3K27me3, and H3K27ac fluorescence intensities. Number of embryos analyzed is included in Supplementary Fig. 7. Boxplot: center line, median; box limits, 25th–75th percentiles; whiskers, ±1.5× interquartile range. p-value calculated by two-sided Student’s t-test. Mat: maternal pronuclei; Pat: paternal pronuclei; hpi: hrs post IVF. b Metaplots showing H3K27ac enrichment at maternal genes, major ZGA genes, and enhancers in *Bap1* CTR and matKO late 2-cell (L2C) embryos. TSS: transcription start site. p-value: two-sided Wilcoxon rank-sum test. c Metaplots showing H3K27me3 enrichment the same as in (b). d Metaplots (top) and heatmaps (bottom) showing changes in H3K27me3 and H3K27ac profiles at the H3K27ac-lost regions in fully grown oocytes (FGOs) and L2C embryos from *Bap1* CTR and CKO groups. For L2C embryos, allelic signals are shown separately for maternal (Mat) and paternal (Pat) alleles (see Methods). e Genome browser views of H3K27me3 and H3K27ac profiles in FGOs and L2C embryos at the indicated genomic loci. f Metaplots showing allelic H3K27me3 enrichment at a subset of Polycomb group (PcG) targets in late 2-cell embryos from *Bap1* CTR and matKO groups. g Genome browser views of H3K27me3 profiles in FGOs and L2C embryos at the indicated genomic loci. Source data are provided as a Source data file.

To further explore this, we performed CUT&RUN profiling of H3K27ac and H3K27me3 at the L2C stage, when major ZGA occurs. Consistent with the immunofluorescence data, the global H3K27me3 landscape was largely unchanged between control and matKO embryos, whereas H3K27ac signals were moderately decreased (Supplementary Fig. 8a, b). Notably, putative L2C enhancers, defined by distal H3K27ac peaks²⁹, were reduced in matKO embryos (Fig. 6b and Supplementary Fig. 8c). In contrast, H3K27ac levels at promoter regions of both maternal and major ZGA genes remained largely unaffected by the loss of maternal BAP1 (Fig. 6b). These findings suggest that impaired activation of L2C enhancers contributes to defective ZGA in maternal BAP1-deficient embryos. Moreover, H3K27ac levels at the embryonic enhancers were low in FGOs and showed no difference between control and mutant oocytes (Supplementary Fig. 8d), indicating that this defect arises post-fertilization. Furthermore, H3K27me3 levels at maternal genes, major ZGA loci, and L2C enhancers were comparable between control and matKO embryos (Fig. 6c), excluding aberrant H3K27me3 deposition as a contributing factor. This also suggests that the observed upregulation of maternal transcripts (Fig. 5e, f) likely reflects impaired maternal mRNA decay rather than transcriptional derepression resulting from altered H3K27me3. Collectively, these results indicate that maternal BAP1 is required for proper embryonic enhancer activities during ZGA, independent of H3K27me3 regulation.

Aberrant H3K27me3 landscape, but not H3K27ac changes, in BAP1-null oocytes persists into early embryos

Given that H3K27me3-deficient states in Eed- or Pcgf1/6-null oocytes are irreversibly transmitted to early embryos^17,23, we next examined whether the aberrant H3K27me3 landscape observed in Bap1-null FGOs (Figs. 2–3) similarly persists post-fertilization. Leveraging the fact that our L2C embryos are derived from B6/129 × PWK F₁ hybrids, we performed allele-specific analysis of H3K27me3 distribution. We first focused on genomic regions, primarily located in gene deserts, that acquire ectopic H3K27me3 and corresponding H3K27ac loss in Bap1-deficient oocytes (Fig. 2). In control L2C embryos, these regions displayed paternally biased H3K27me3 enrichment (Fig. 6d, e and Supplementary Fig. 8e), consistent with prior reports that paternal H3K27me3 in preimplantation embryos predominantly localizes to gene-poor regions¹⁵. Remarkably, these loci acquired maternal-specific H3K27me3 enrichment in Bap1 matKO embryos, and this ectopic gain was already evident in CKO FGOs (Fig. 6d, e and Supplementary Fig. 8e), suggesting inheritance of the abnormal chromatin state from the oocyte. In contrast, H3K27ac levels at these loci declined sharply from FGO to L2C stages in both control and mutant embryos, with the CTR maternal allele showing only a slightly higher H3K27ac than that in mutants (Fig. 6d, e and Supplementary Fig. 8e). This suggests that this H3K27ac reduction reflects a normal developmental progression and is not specific to BAP1 loss. Notably, the inherited ectopic H3K27me3 may not initiate gene repression in embryos, as the downregulated genes within these domains were already transcriptionally reduced at earlier stages (Supplementary Fig. 8f, g), implying that these transcriptional defects arise during oogenesis rather than being induced post-fertilization.

We next assessed H3K27me3 and H3K27ac profiles at the typical Polycomb targets in control and Bap1 matKO L2C embryos. Previous studies have shown that H3K27me3 is largely lost from promoter regions of Polycomb targets following fertilization, with only a subset retaining residual H3K27me3 during preimplantation development^15,61. Analysis of the Polycomb targets that retain H3K27me3 at the L2C stage (n = 719; see “Methods”") revealed a maternal bias in H3K27me3 enrichment in control embryos (Fig. 6f, g). In contrast, this maternal H3K27me3 signal was either completely lost or markedly reduced in Bap1 matKO embryos (Fig. 6f, g). Notably, the reduction observed in mutant L2C embryos was more pronounced than that in Bap1 CKO oocytes, suggesting progressive loss post-fertilization. By contrast, the modest H3K27ac gain seen at Polycomb targets in CKO FGOs was not sustained in early embryos (Supplementary Fig. 8h, i). Together, these results demonstrate that the aberrant H3K27me3 landscape, but not H3K27ac changes, established in BAP1-null oocytes is inherited into early embryonic stages.

The ectopic H3K27me3 domains in Bap1 matKO embryos are resolved by early post-implantation stages

The intergenerational inheritance of aberrant H3K27me3 domains in Bap1 matKO embryos prompted us to determine how long these domains persist beyond the two-cell stage. To this end, we performed H3K27me3 CUT&RUN in morulae, the developmental stage immediately preceding the first lineage differentiation. We also dissected post-implantation E6.5 embryos into epiblast (EPI), visceral endoderm (VE), and extra-embryonic ectoderm (EXE), followed by H3K27me3 CUT&RUN profiling of each lineage. The CUT&RUN biological replicates were highly reproducible (Supplementary Fig. 9a), and hierarchical clustering revealed clear separation between morulae and post-implantation lineages (Supplementary Fig. 9b). As validation of data quality, non-canonical imprinted genes (e.g., Gab1, Jade1/Phf17, and Sfmbt2) exhibited the expected maternally biased H3K27me3 enrichment at the morula stage (Supplementary Fig. 9c). Furthermore, lineage marker genes, Pou5f1 in EPI, Gata6 in VE, and Gata3 in EXE, displayed lineage-specific deletion of H3K27me3 in their respective tissues, further validating data quality (Supplementary Fig. 9d–f).

Having established data quality, we next examined the developmental dynamics of the aberrant H3K27me3 domains induced by loss of BAP1. We found that ectopic H3K27me3 gains in Bap1-null oocytes persist through the morula stage but are resolved in both embryonic and extraembryonic lineages by E6.5 (Fig. 7a, b). This observation resonates with our previous findings on non-canonical genomic imprinting²⁶, in which the maternally inherited H3K27me3 domains are maintained during preimplantation development but are subsequently lost following implantation.

Fig. 7 — a Genome browser view of allelic H3K27me3 in late 2-cell (L2C), morula, and E6.5 control (CTR) and maternal KO (matKO) embryos. EPI: epiblast, VE: visceral endoderm, EXE: extra-embryonic ectoderm. b Metaplots showing allelic H3K27me3 enrichment at the ectopic H3K27me3 domains across different developmental stages. c Heatmaps showing allelic expression bias of imprinted genes. One control EXE RNA-seq dataset is from ref. ²⁶. MEG: maternally expressed genes. PEGs paternally expressed genes. Pat paternal, Mat Maternal, SNP single nucleotide polymorphism.

A recent study reported that knock out of Ezhip, a negative regulator of PRC2^78,79, causes excessive genome-wide H3K27me3 accumulation in early embryos and paradoxically disrupts non-canonical genomic imprinting⁶⁰. Prompted by this observation, we asked whether pervasive increase of H2AK119ub1 observed in Bap1-null oocytes might similarly affect allelic gene expression in later development. Consistent with the preserved allelic H3K27me3 enrichment at the non-canonical imprinted loci (Supplementary Fig. 9c), imprinted expression of these genes was largely maintained in the VE and EXE lineages (Fig. 7c, bolded genes). Similarly, canonical imprinting was minimally affected by maternal BAP1 loss (Fig. 7c). These data demonstrate that maternal BAP1 is dispensable for both canonical and non-canonical genomic imprinting.

Discussion

In this study, we report that BAP1 is essential for establishing spatially distinct chromatin domains during oogenesis, specifically active euchromatin marked by H3K27ac and facultative heterochromatin marked by H2AK119ub1 and H3K27me3. BAP1 prevents pervasive accumulation of H2AK119ub1 across the genome and protects broad, oocyte-specific H3K27ac domains, particularly in gene-poor regions, from ectopic H3K27me3 deposition. This chromatin regulation is critical for activating a subset of oocyte-specific genes, including those involved in cell adhesion and growth factor signaling. Consequently, loss of maternal BAP1 leads to impaired developmental competence of oocytes and female subfertility. Notably, our findings demonstrate that BAP1 functions primarily as a transcriptional activator in oocytes, rather than enhancing Polycomb-mediated gene silencing. Together, these results reveal an essential role for PR-DUB in safeguarding the oocyte epigenome and preserving reproductive potential.

This study, together with previous work in Drosophila embryos³⁶, human cell lines^41,42, and mESCs^37–40, underscores a conserved role for BAP1 in limiting widespread H2AK119ub1 accumulation, albeit to varying degrees across model systems. Notably, BAP1-null oocytes uniquely exhibit a striking loss of H3K27ac and a corresponding gain of H3K27me3 over broad genomic domains, often spanning hundreds of kilobases). Two non-mutually exclusive mechanisms may explain how increased H2AK119ub1 leads to H3K27me3 gain and H3K27ac loss in Bap1 CKO oocytes. First, elevated H2AK119ub1 may enhance recruitment of PRC2.2³, resulting in deposition of H3K27me3, which in turn could displace H3K27ac and repress transcription. Supporting this, PRC2.2 depletion prevents intergenic H3K27me3 accumulation in Bap1-null mESCs³⁸. Second, ASXLs, components of PR-DUB, interact with the MLL3/4-containing COMPASS complex, which includes the H3K27 demethylase KDM6A^41,80–82. In addition, BAP1-dependent enhancer binding of KDM6A has been observed in human cell lines⁴¹. Thus, loss of BAP1 may impair COMPASS targeting to enhancers, leading to increased H3K27me3 and reduced H3K27ac. Further investigation is needed to distinguish between these mechanisms and to elucidate how PR-DUB integrates with other chromatin-modifying complexes during oogenesis.

It is unexpected that H3K27ac loss and H3K27me3 gain in BAP1-null oocytes preferentially occur in gene-poor regions, a phenomenon not observed in other model systems. Our findings suggest that the timing of H3K27ac establishment during oocyte development underlies this gene desert-biased loss of H3K27ac in Bap1 CKO FGOs. Specifically, gene-poor regions acquire high levels of H3K27ac during the final stages of oocyte growth²⁹ and are therefore more sensitive to Gdf9-iCre-mediated Bap1 deletion than regions that gain H3K27ac early. Supporting this notion, Bap1-dependent genes reach peak expression only at the fully grown stage, whereas genes unaffected by Bap1 deletion are already highly expressed early during oocyte growth. These observations suggest that in non-dividing cells like growing oocytes, BAP1 is particularly important for the establishment of new H3K27ac marks and activation of transcription, rather than for the maintenance of existing chromatin and transcriptional states. A less likely scenario is that PR-DUB is preferentially recruited to gene-poor regions via an unknown mechanism. To further dissect these possibilities, it would be informative to delete Bap1 earlier than what is achieved with Gdf9-iCre-mediated CKO, to assess whether early BAP1 loss affects both gene-dense and gene-poor regions similarly.

Another unique aspect of BAP1 function in oocytes is its predominant role in promoting gene activation, with only a minor contribution to Polycomb-mediated gene silencing (Fig. 8a). This contrasts with findings in mESCs, where BAP1 loss leads to pervasive accumulation of H2AK119ub1, resulting in displacement of PRC1 and PRC2 from Polycomb target genes and consequent gene derepression^37–39. In BAP1-null oocytes, although H3K27me3 levels at canonical PcG targets are moderately reduced, this reduction appears insufficient to trigger derepression of these genes. One possible explanation is that the high abundance of PRC1 and PRC2 complexes accumulated during oocyte growth may buffer against sequestration by ectopic H2AK119ub1. Even if a portion of PRC1/2 is mis-localized, the remaining complexes might still be sufficient to maintain transcriptional repression at their target sites.

Fig. 8 — a BAP1 depletion causes pervasive increase of H2AK119ub1 in fully grown oocytes (FGOs), with a less extent at highly transcribed gene loci. Genes within gene deserts are preferentially down-regulated in *Bap1-*null FGOs, which is associated with increased H3K27me3 and reduced H3K27ac. Polycomb silencing is largely maintained at the target genes in FGOs despite moderate reduction of H3K27me3 and increase of H3K27ac. b The aberrant H3K27me3 landscapes, but not H3K27ac, established in *Bap1*-null oocytes persist in early embryos. H3K27ac at enhancers, but not at ZGA gene promoters, are reduced in *Bap1* maternal KO embryos. The abnormal maternal-to-zygotic transition (MZT) in *Bap1* maternal KO embryos impairs preimplantation development, leading to compromised blastocyst quality and female subfertility. Created in BioRender. Chen, Z. (2026) https://BioRender.com/9h2lqbj.

Interestingly, in Drosophila, PR-DUB preserves Polycomb silencing through a distinct mechanism compared to mESCs. In flies, H2AK118ub1 is rapidly removed by PR-DUB at embryonic stages when Polycomb repression is active and thus does not co-localize with H3K27me3³⁶. In the absence of PR-DUB, elevated H2AK118ub1 disrupts chromatin compaction, impairing gene silencing³⁶. In this context, H2AK118ub1 may function as a transcriptional activator. A similar activating role for H2AK119ub1 has also been proposed in mammalian systems^83,84. However, we did not observe such a scenario in mouse oocytes, suggesting that this mechanism is either not operated in oocytes or that H2AK119ub1 increase in this context does not reproduce the chromatin environment seen in other systems.

The role of BAP1 in chromatin regulation is critical for the activation of a subset of oocyte-specific genes during oogenesis, particularly those involved in cell adhesion and growth factor signaling. Failure to activate these genes compromises the developmental competence of Bap1-null oocytes, and the matKO embryos subsequently fail to properly complete the MZT, resulting in defective preimplantation development. Among the BAP1-dependent genes, Desmocollin 3 (Dsc3) has been shown to be essential for preimplantation development⁸⁵. Additionally, Fibroblast growth factor 10 (Fgf10) has been reported to enhance oocyte maturation and developmental competence in bovine⁸⁶. Other candidates that may contribute to the compromised oocyte and early embryonic development include N-cadherin (Cdh2) and Epidermal growth factor receptor (Egfr). These findings suggest that PR-DUB function during oogenesis is essential for supporting early embryonic development. However, we cannot exclude the possibility that maternal BAP1 also directly regulates embryonic enhancer activity after fertilization, as indicated by reduced H3K27ac at L2C enhancers in Bap1 matKO embryos.

The locus-specific gain of H3K27me3 in Bap1-null oocytes provides key insights into intergenerational epigenetic inheritance. Intergenerational transmission of H3K27me3 from the germline to early embryos has been reported across species, including C. elegans⁸⁷, Drosophila⁸⁸, and mice²⁴. In Drosophila, for example, germline-inherited H3K27me3 is required to prevent premature activation of lineage-specific genes during ZGA⁸⁸. In mice, oocyte-inherited H3K27me3 underlies germline DNA methylation-independent non-canonical imprinting²⁴, and H3K27me3-deficient states in Eed- or Pcgf1/6-null oocytes are irreversibly transmitted to early embryos, leading to defective non-canonical imprinting^17,23. Notably, Pcgf1/6 matKO embryos fail to re-establish H3K27me3 domains lost in the oocyte, despite having functional PRC2 complexes¹⁷. In this study, we observed that the aberrant H3K27me3 landscapes in Bap1-null oocytes, including ectopic gain in gene-poor regions and reduction at PcG targets, persist in early embryos (Fig. 8b). In contrast, the abnormal H3K27ac distributions in Bap1-deficient oocytes are largely reversed along with the global reduction of H3K27ac levels from FGO to L2C transition. It will be of interest in future studies to determine whether intergenerational inheritance is specific to heterochromatin-associated marks such as H3K27me3, but not euchromatin marks like H3K27ac, in mammals. The functional consequences of the inherited ectopic H3K27me3 domains in preimplantation embryos remain unclear, particularly given that these domains are resolved by early post-implantation stages. For non-canonical imprinting, allelic H3K27me3 in preimplantation embryos switches to allelic DNA methylation to maintain imprinted gene expression in placental lineages^26,89. Whether the ectopic H3K27me3 domains acquired in Bap1-null oocytes undergo a similar epigenetic switch mechanism warrants further investigation.

Methods

Mouse strains and breeding

For DUB overexpression experiments and for western blot analyses of BAP1 protein in oocytes and early embryos, we used B6D2F1/J (BDF1; JAX strain #100006) male and female mice for IVF. C57BL/6J males (JAX strain #000664) were used for Bap1 CTR and CKO breeding trials and for embryonic development analyses at E3.5 and E18.5. To enable allelic resolution in chromatin profiling and RNA sequencing, PWK/PhJ males (JAX strain #003715) were used to generate late 2-cell, morula, and E6.5 embryos.

To generate oocyte specific Bap1 KO, Bap1^flox animals (Jax strain #: 031874, B6 background) were crossed with Gdf9-iCre transgenic mice (Jax strain #: 011062, B6/129 mixed background) to produce Gdf9-iCre, Bap1 ^flox/+ males and Bap1 ^flox/+ females. These mice were intercrossed to obtain Bap1 ^flox/flox and Gdf9-iCre, Bap1 ^flox/flox females, which were used as CTR and CKO, respectively. To generate sperm specific Bap1 KO, Bap1^flox animals were crossed with Prm-Cre mice (Jax strain #: 007252, B6/129 mixed background). The resulting Prm-Cre, Bap1 ^flox/+ females were subsequently bred with Bap1 ^flox/+ males to obtain Prm-Cre, Bap1 ^flox/flox males for downstream experiments. Genotypes were confirmed by PCR using genomic DNA extracted from tail biopsies. Genotyping primers are listed in Supplementary Data 3. Mice were euthanized in CO₂ chamber followed by cervical dislocation. Mice were housed under a 12:12 h light:dark cycle with controlled ambient temperature and humidity. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at Cincinnati Children’s Hospital Medical Center (protocol IACUC2025-0089).

Mouse superovulation, IVF, embryo culture, and fertility assessment

To induce superovulation, 8–12-week-old female mice were intraperitoneally injected with either 7.5 IU Pregnant Mare Serum Gonadotropin (PMSG; BioVendor) or ~0.15 mL CARD HyperOva (Cosmo Bio). After 44–48 h, 7.5 IU Human Chorionic Gonadotropin (hCG; Sigma) was administered via intraperitoneal injection. Approximately 16–18 h post-hCG injection, cumulus-oocyte complexes (COCs) were collected from the oviductal ampullae. IVF was performed using sperm isolated from the cauda epididymis of 9–15-week-old male mice. For collection of Bap1 CTR and matKO embryos used in immunostaining, RNA-seq, and developmental rate analyses, sperm were obtained from wild-type C57BL/6 J (Jax strain #: 000664) or Prm-Cre, Bap1^flox/flox males and capacitated in Human Tubal Fluid (HTF; Millipore Sigma) medium for ~1 h at 37 °C in a CO₂ incubator with ambient air. For CUT&RUN experiments, sperm were isolated from wild-type PWK/PhJ mice (Jax strain #: 003715) and capacitated in CARD FERTIUP Preincubation Medium (Cosmo Bio).

The time of sperm addition to COCs was defined as 0 h post-insemination (hpi). At ~6 hpi, zygotes exhibiting two pronuclei were selected, washed in M2 medium (Millipore), and cultured in KSOM medium (Millipore) at 37 °C in a CO₂ incubator under ambient air conditions. Brightfield images of oocytes and embryos were acquired using EVOSII cell imaging system (Thermo Fisher Scientific).

For fertility assessment and embryo collection at E3.5 and E18.5, Bap1 CTR and CKO females were co-caged with wild-type C57BL/6J males (Jax strain #: 000664). The presence of a vaginal plug the following morning was designated as E0.5. For E3.5 collections, blastocysts were flushed from the uterus using M2 medium (Millipore). For E18.5 collections, pregnant females were euthanized, and the number and weight of fetuses and placentas were assessed.

Dissection of E6.5 embryos into epiblast (EPI), visceral endoderm (VE), and extra-embryonic ectoderm (EXE) was described previously⁹⁰. Specifically, after removing Reichert’s membrane and ectoplacental cone, embryos were incubated on ice for 5 min in a disassociation solution containing 2.5% pancreatin (Sigma), 0.25% trypsin (Fisher Scientific), and 0.5% polyvinylpyrrolidone (Irvine Scientific). The digestion was stopped by adding an equal volume of 10% fetal bovine serum (FBS, HyClone) in PBS (Gibco). The VE layer was separated by gentle pipetting through a fine glass needle. EPI and EXE were then separated by cutting at the boundary between embryonic and extraembryonic regions using 27-gauge needles. The isolated EPI, VE, and EXE were washed three times in 0.2% BSA/PBS and stored in 0.2 ml tubes at −80 °C.

Whole-mount immunostaining and histology analyses

For immunostaining, FGOs, zygotes, two-cell embryos, and blastocysts were fixed and permeabilized in 3.7% paraformaldehyde (Sigma) and 0.2% Triton X-100 (Sigma) in PBS (Gibco) at room temperature for 20 min. After four washes in 1% BSA/PBS, oocytes/embryos were incubated with primary antibodies at 4 °C overnight. The following day, samples were washed four times in 1% BSA/PBS before incubated with secondary antibody at room temperature for 1 h. Both primary and secondary antibodies were diluted in 1% BSA/PBS. Oocyte/embryos were then mounted using VECTASHIELD Antifade Mounting Medium with DAPI. Fluorescence was detected under a laser-scanning confocal microscope (Nikon A1R inverted), and images were analyzed using ImageJ (NIH).

Primary antibodies were rabbit anti-H3K27me3 (1:500, Cell Signaling, 9733), rabbit anti-H2AK119ub1 (1:2000, Cell Signaling, 8240), rabbit anti-H3K27ac (1:500, Millipore Sigma, 07360), mouse anti-Flag (1:200, Sigma, F1804), rabbit anti-NANOG (1:200, Cosmo Bio, ATL-HPA072117-25), Goat anti-CDX2 (1:200, R&D, AF3665), or Mouse anti-GATA4 (1:200, R&D MAB2606). Secondary antibodies were Alexa Fluor 488 donkey anti-mouse IgG (1:200), Alexa Fluor 568 donkey anti-rabbit IgG (1:200), or Alexa Fluor 647 donkey anti-goat IgG (1:200) (Life Technologies). For histology analyses, paraffin-embedded ovaries were sectioned at 5 μm and stained with hematoxylin and eosin.

Plasmid construction and mRNA preparation

Coding sequences for mouse Usp16 (NM_024258), Usp21 (NM_013919), Usp28 (NM_175482), Mysm1 (NM_177239), Bap1 (NM_027088), and the N-terminal region (amino acids 1–476) of Asxl1 (NM_015338) were synthesized by Twist Bioscience. Each sequence was cloned into the pcDNA3.1-Flag-poly(A)83 vector⁹¹, inserted between an N-terminal Flag tag and an 83-nucleotide polyadenylation sequence. Plasmids were linearized by restriction enzyme digestion and used as templates for in vitro transcription with the mMESSAGE mMACHINE® T7 Ultra Kit (Life Technologies), following the manufacturer’s instructions. The resulting mRNA was purified by LiCl precipitation and quantified using a NanoDrop spectrophotometer.

Micro-injection of H2A DUB mRNAs

IVF experiments using BDF1 (Jax strain #: 100006) MII oocytes and sperm were the same as in previous sections. At 2 hpi, fertilized zygotes with the presence of 2nd polar body or fertilization cone were transferred into M2 medium (Sigma), and mRNA was injected into the cytoplasm of zygotes (~2–4 hpi) using a Piezo impact-driven micromanipulator (Eppendorf). Following injection, embryos were cultured in HTF (Millipore) medium for an additional 4 h before being transferred to KSOM (Millipore). The mRNA concentrations for each H2A DUB mRNA were 2 µg µl⁻¹, except that 1 µg µl⁻¹ was used for Bap1 and Asxl1 mRNAs when they were co-injected.

Western blotting

Oocytes and embryos collection and immunoblot analysis were as previously described⁹². Specifically, oocytes and embryos were collected, and the zona pellucida was removed by a brief incubation in acid Tyrode’s solution (Sigma). After three washes with prewarmed 1% polyvinyl alcohol (PVA; Sigma) in PBS, samples in 5 µl of 1% PVA/PBS were transferred using a glass pipette into 15 µl of loading buffer containing 5 µl of 4× LDS sample buffer (Thermo Fisher Scientific) and 10 µl of water. Samples were snap-frozen in dry ice for 5 min, and the freeze-thaw cycle was repeated once before storage at −80 °C. Samples were heated at 90 °C for 5 min and loaded onto a 4–12% Bis-Tris Protein gel (Thermo Fisher Scientific). Proteins were transferred to a PVDF membrane using the Trans-Blot Turbo Transfer System (Bio-Rad). Membranes were blocked in 3% BSA/PBS for 1 h at room temperature and incubated with primary antibodies overnight at 4 °C. The following day, membranes were washed and incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Chemiluminescence was detected using an Amersham Imager 680. Primary antibodies against BAP1 (1/1000 dilution; CST 13271), Tubulin (1/1000 dilution; CST 2144), Beta-actin (1/1000, sc-47778) were used in this study. Secondary antibodies were HRP-linked anti-rabbit IgG (CST 7074P2; 1:1000) and anti-mouse IgG (CST 7076P2; 1:1000).

RNA-seq and CUT&RUN library preparation and sequencing

For oocytes and preimplantation embryos, RNA-seq libraries were generated using the SMART-Seq Stranded Kit (Takara) according to the manufacturer’s instructions. Briefly, oocytes and embryos were incubated in acid Tyrode’s solution to remove zona pellucida and then washed three times in 0.2% BSA/PBS. Samples were transferred to 0.2-ml PCR tubes using a glass needle with minimal carryover (~0.5–1 µl). After adding 7 µl of RNase inhibitor water and 6 µl of shearing mix, samples were lysed at 85 °C to fragment RNA. First-strand cDNA synthesis was performed using random primers, followed by an initial PCR amplification (10 cycles). The resulting cDNA was purified, and rRNA-derived sequences were removed by probe-mediated cleavage. Final RNA-seq libraries were prepared with an additional PCR amplification (10 cycles).

For E6.5 embryos, RNA-seq libraries were generated using SMART-Seq mRNA (Takara) in combination with the Nextera XT Sample Preparation Kit (Illumina). Reverse transcription was performed using oligo-dT according to the manufacturer’s instructions, and cDNA was subsequently amplified by PCR (13 cycles). For each sample, 400 pg of amplified cDNA was used for library preparation. The Nextera XT kit was then used for cDNA fragmentation, adaptor ligation, and library amplification, following the manufacturer’s protocols.

CUT&RUN library preparation was performed as previously described⁹³ with minor modifications. For FGOs, the zona pellucida was removed by brief incubation in acid Tyrode’s solution. For late 2-cell embryos, zona-free embryos were further incubated in TrypLE (Thermo Fisher Scientific) supplemented with 0.5% PVP (Irvine Scientific) to remove polar bodies. Oocytes and embryos were transferred to 0.2 ml CUT&RUN 8-strip tubes (EpiCypher) containing 60 µl wash buffer (20 mM HEPES-NaOH pH 7.5, 150 mM NaCl, 0.5 mM spermidine, and 1× Roche protease inhibitor) and 10 µl Concanavalin A-coated beads (Polysciences) pre-activated in bead binding buffer (20 mM HEPES-KOH pH 7.9, 10 mM KCl, 1 mM CaCl₂, 1 mM MnCl₂). For some CUT&RUN reactions, Drosophila cell nuclei (Actif motif) were used for spike-in normalization. The number of Drosophila cell nuclei used for each experiment is included in Supplementary Data 3. Beads and cells were mixed by gentle tapping and rotated on a nutator for 10–15 min at room temperature. Tubes were briefly spun down and placed on a magnetic stand to remove supernatant without disturbing the bead-cell complexes. Samples were incubated with primary antibodies diluted in antibody incubation buffer (1× wash buffer, 0.02% digitonin, 2 mM EDTA) at 4 °C overnight on a thermomixer (300 rpm). The following day, samples were washed once with 200 µl Dig-wash buffer (1× wash buffer, 0.02% digitonin) and once with 100 µl Dig-wash buffer, with each wash performed for 30 s on a magnetic stand. Beads/cells were then incubated with 5.6 ng/ml pAG-MNase (homemade) at 4 °C on a thermomixer (300 rpm) for 2–3 h. Afterward, samples were washed twice with 200 µl Dig-wash buffer and incubated with 100 µl Dig-wash buffer on ice for 3 min. To initiate pAG-MNase digestion, 2 µl of 100 mM CaCl₂ was added and mixed by gentle tapping. Digestion proceeded on ice for 30 min and was stopped by adding 100 µl 2× STOP buffer (340 mM NaCl, 20 mM EDTA, 4 mM EGTA, 50 µg/ml RNase A, 100 µg/ml glycogen, 0.02% digitonin). Samples were incubated at 37 °C for 30 min to release short DNA fragments, after which the supernatant was supplemented with 40 ng carrier RNA and purified by phenol-chloroform extraction. CUT&RUN libraries were prepared using the NEBNext Ultra II DNA Library Prep Kit (NEB) according to the manufacturer’s instructions.

For morula-stage CUT&RUN, a modified protocol was used in which primary antibody binding and pAG-MNase binding were performed prior to binding cells to Concanavalin A beads⁹⁴. Postnatal day 1 oocyte collection and H3K27ac CUT&TAG was as previously described⁹⁵.

All libraries were sequenced on the Illumina NovaSeq X Plus platform at Novogene (2 × 150 bp). The number of oocytes and embryos used for RNA-seq/CUT&RUN, sequencing depth, and read alignment are included in Supplementary Data 3.

RNA-seq data processing

Adaptor trimming, read alignment, and read count generation for both genes and repeat subfamilies were performed as previously described⁷⁴. Differential gene expression analysis was carried out using DESeq2 (v1.38.3)⁹⁶, applying the criteria: adjusted p-value < 0.05 and fold change ≥ 2. For a gene to be considered differentially expressed, the higher-expression group was additionally required to have FPKM > 0.5. Gene Ontology (GO) term enrichment among differentially expressed genes (DEGs) was assessed using the enrichGO function in the clusterProfiler R package (v4.6.2)⁹⁷. BigWig files for visualization were generated using deepTools (v2.0.0)⁹⁸ with the following parameters: --binSize 25 --normalizeUsing RPKM --scaleFactor 1. Gene sets corresponding to zygote minor ZGA genes, early 2-cell minor ZGA genes, late 2-cell major ZGA genes, and maternal transcripts were obtained from ref. ³⁰. Enrichment of DEGs across gene categories defined in the DBTMEE_v2 database⁶⁵ was analyzed using the enrichr function in clusterProfiler (v4.6.2)⁹⁷. For E6.5 stage RNA-seq data, to avoid alignment bias toward the reference genome (C57BL/6), reads were aligned to a custom genome where single-nucleotide polymorphisms (SNPs) between PWK/PhJ and B6/129 strains were masked with “N”. SNPsplit (v0.3.2)⁹⁹ was used to determine parental origin of read pairs. Imprinted gene list at E6.5 was obtained from ref. ²⁴.

CUT&RUN data processing

Raw paired-end reads from CUT&RUN experiments were trimmed using TrimGalore (v0.6.6) (https://github.com/FelixKrueger/TrimGalore) with the --paired option. For FGOs, trimmed reads were aligned to the mm10 mouse reference genome using Bowtie2 (v2.4.2)¹⁰⁰ with the parameters: --no-unal --no-mixed --no-discordant -I 0 -X 1000. Uniquely mapped reads were filtered using Sambamba (v0.6.8)¹⁰¹ with the criteria: -F “mapping_quality ≥ 30 and not (unmapped or mate_is_unmapped)”. PCR duplicates were removed using Picard (v2.18.22) (http://broadinstitute.github.io/picard/). BigWig coverage files were generated using deepTools (v2.0.0)⁹⁸ with the options: --binSize 25 --normalizeUsing RPKM --scaleFactor 1.

For late 2-cell (L2C), morula, and E6.5 stage CUT&RUN data, to avoid alignment bias toward the reference genome (C57BL/6), reads were aligned to a custom genome where SNPs between PWK/PhJ and B6/129 strains were masked with “N”. SNPsplit (v0.3.2)⁹⁹ was used to determine parental origin of read pairs. For H2AK119ub1, H3K27me3, and H3K27ac oocyte CUT&RUN datasets, spike-in normalization was performed by aligning reads to the Drosophila melanogaster reference genome (fly46), and scaling factors were calculated as described by ref. ¹⁰² (see Supplementary Data 3). Since oocyte chromatin exhibits non-canonical, broad enrichment of H2AK119ub1, H3K27me3, and H3K27ac, a hidden Markov model (HMM)-based domain caller¹⁸ was used to determine broad domain boundaries for these histone modifications.

Reproducibility between biological replicates was assessed by calculating the Pearson correlation coefficient of FPKM values computed in non-overlapping 5-kb genomic bins. For downstream analyses, biological replicates were pooled using the merge command from Sambamba (v0.6.8)¹⁰¹. The heatmaps of CUT&RUN profiles were generated using deepTools (v2.0.0)⁹⁸ computeMatrix, and plotHeatmap.

Identification of H3K27ac-lost regions in Bap1 CKO FGOs

FPKM values were computed in non-overlapping 5-kb genomic bins across the genome. Bins with FPKM ≥ 2 and fold change ≥ 1.5 between conditions were classified as differentially enriched bins. To define broader enriched regions, adjacent differentially enriched 5-kb bins were merged using BEDTools (v2.30.0)¹⁰³. Only merged regions larger than 10 kb were retained and defined as differentially enriched regions. The regions were annotated using “makeTxDbFromGFF” and “annotatePeak” function from the “ChIPpeakAnno (v4.4.1)”, “GenomicFeatures (v4.4.1)”, and “ChIPseeker (v4.4.2)” Bioconductor R packages^104–106. Note that multiple H3K27ac-lost regions could be identified from a single H3K27ac domain defined by the HMM-based method.

Analyses at PcG targets

The list of oocyte and ESC PcG target genes was obtained from refs. ^60,61, respectively. The oocyte Polycomb targets were identified based on the H3K27me3 enrichment in fully grown oocytes⁶⁰. To identify PcG targets that retain H3K27me3 at the late 2-cell stage, FPKM values at promoter regions (±2.5 kb around the transcription start site) were calculated. Genes with FPKM ≥ 2 were considered as retaining H3K27me3. This cutoff was determined based on visual inspection of representative loci, and the resulting gene list was found to be comparable to ref. ⁶¹.

Annotations and genomic intervals

The gene annotation file (GFF3 format) was downloaded from GENCODE (release M25) (https://www.gencodegenes.org/). Gene promoters were defined as ±2.5 kb surrounding the transcription start sites (TSS). Gene bodies were defined from the TSS to the annotated transcription end site (TES). Intergenic regions were defined as genomic regions that do not overlap with annotated promoters or gene bodies. Putative enhancers specific to FGOs and late 2-cell embryos were obtained from ref. ²⁹. Gene-rich and gene-poor genomic regions were defined as previously described¹⁰⁷. All mm9-based genomic coordinates were converted to the mm10 assembly using the liftOver function from the rtracklayer R package (v1.58.0)¹⁰⁸.

Visualization and statistical analyses

All statistical analyses were performed using R (v4.2.2). Statistical tests and sample sizes are indicated in the figure legends. All genomic browser tracks were visualized using the UCSC genome browser¹⁰⁹. Figure 8 was generated using BioRender.

Reanalysis of public datasets

All public ChIP-seq, CUT&RUN, and RNA-seq datasets were processed using pipelines and parameters similar to those used for in-house data.

Statistics and reproducibility

Statistical analyses were performed with R and GraphPad Prism 9.5 software. All statistical methods, sample sizes, and p-values are indicated in Figures or Figure legends. Reproducibility between RNA-seq and CUT&RUN biological replicates was evaluated by calculating Pearson’s r coefficient. No statistical method was used to predetermine sample size. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessments.

Reporting summary

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

Supplementary information

Supplementary Information^{(21.9MB, pdf)}

41467_2026_70845_MOESM2_ESM.pdf^{(280.3KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1^{(19.1MB, xlsx)}

Supplementary Data 2^{(179.6KB, xlsx)}

Supplementary Data 3^{(21.9KB, xlsx)}

Reporting Summary^{(103.4KB, pdf)}

Transparent Peer Review file^{(1.7MB, pdf)}

Source data

Source Data^{(446.7KB, xlsx)}

Acknowledgements

We thank Dr. Yueh-Chiang Hu for rederiving the Gdf9-iCre mice. We are grateful to Amanda Barbosa and Dr. Matthew Kofron for assistance with imaging analysis. We also thank Drs. S.K. Dey, Tony De Falco, Xiaofei Sun, and Maria Mikedis for helpful discussion. Confocal microscopy was performed at the CCHMC Bio-Imaging and Analysis Facility. This work was supported by Z.C.’s start-up funds and the Eunice Kennedy Shriver National Institute of Child Health and Human Development (R00 HD104902). Z.C. is also supported by R01 HD118540. S.H.N is supported by R35 GM141085.

Author contributions

Z.C. conceived the project. Z.C. and J.K. designed the experiments. With assistance from P.L., S.I., and L.C., J.K. and Z.C. performed the experiments and analyzed the data. M.H. and S.H.N. generated the P1 oocyte H3K27ac data. Z.C. carried out the bioinformatic analyses. Z.C. and J.K. wrote the manuscript with input from all authors.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

All sequencing data were deposited in the Gene Expression Omnibus under accession numbers GSE302804 and GSE302844. RNA-seq data in Fig. 1d and Ribo-seq data in Supplementary Fig. 1 were obtained from GSE165782⁴³ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE165782]. Total RNA-seq data and Ribo-seq data in Supplementary Fig. 1 were obtained from GSE169632⁴⁴ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE169632]. Oocyte H2AK119ub1 CUT&RUN data in Supplementary Fig. 2 were obtained from GSE153531¹⁸ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE153531]. ESC H2AK119ub1 ChIP-seq data in Fig. 1k were obtained from GSE161996³⁷ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE161996]. P1 oocyte H3K27ac CUT&Tag data in Figs. 2i, j, and Supplementary Fig. 3d were obtained from GSE280637⁹⁵ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE280637]. P7 and P10 oocyte H3K27ac ChIP-seq data in Fig. 2i, j and Supplementary Fig. 3d were obtained from GSE217970²⁹ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE217970]. GO1, GO2, and FGO RNA-seq data were obtained from GSE134279⁹² [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE134279]. EED and EZH2 CUT&RUN data in Fig. 3b were obtained from GSE289642⁶⁰ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE289642]. PRC1/2 KO RNA-seq data in Fig. 3c, d and Supplementary Fig. 4 were obtained from GSE118263 and GSE132156²¹ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE132156]. Usp16 RNA-seq data in Fig. 5h and Supplementary Fig. 6d were obtained from GSE154412⁷⁷ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE154412]. H3K27me3 E6.5 EXE CUT&RUN and RNA-seq data in Fig. 7 and Supplementary Fig. 9 were obtained from GSE130115²⁶ [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE130115]. Source data are provided with this paper.

Code availability

The code for this study is available in Zenodo at 10.5281/zenodo.18701494.

Competing interests

The authors declare no competing interests.

Footnotes

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

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-70845-x.

References

1.Piunti, A. & Shilatifard, A. The roles of Polycomb repressive complexes in mammalian development and cancer. Nat. Rev. Mol. Cell Biol.22, 326–345 (2021). [DOI] [PubMed] [Google Scholar]
2.Kim, J. J. & Kingston, R. E. Context-specific Polycomb mechanisms in development. Nat. Rev. Genet.23, 680–695 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Blackledge, N. P. & Klose, R. J. The molecular principles of gene regulation by Polycomb repressive complexes. Nat. Rev. Mol. Cell Biol.22, 815–833 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Tamburri, S., Rustichelli, S., Amato, S. & Pasini, D. Navigating the complexity of Polycomb repression: enzymatic cores and regulatory modules. Mol. Cell84, 3381–3405 (2024). [DOI] [PubMed] [Google Scholar]
5.Wang, H. et al. Role of histone H2A ubiquitination in Polycomb silencing. Nature431, 873–878 (2004). [DOI] [PubMed] [Google Scholar]
6.de Napoles, M. et al. Polycomb group proteins Ring1A/B link ubiquitylation of histone H2A to heritable gene silencing and X inactivation. Dev. Cell7, 663–676 (2004). [DOI] [PubMed] [Google Scholar]
7.Cao, R. et al. Role of histone H3 lysine 27 methylation in Polycomb-group silencing. Science298, 1039–1043 (2002). [DOI] [PubMed] [Google Scholar]
8.Muller, J. et al. Histone methyltransferase activity of a Drosophila Polycomb group repressor complex. Cell111, 197–208 (2002). [DOI] [PubMed] [Google Scholar]
9.Kuzmichev, A., Nishioka, K., Erdjument-Bromage, H., Tempst, P. & Reinberg, D. Histone methyltransferase activity associated with a human multiprotein complex containing the Enhancer of Zeste protein. Genes Dev.16, 2893–2905 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Margueron, R. & Reinberg, D. The Polycomb complex PRC2 and its mark in life. Nature469, 343–349 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Schuettengruber, B., Bourbon, H. M., Di Croce, L. & Cavalli, G. Genome regulation by polycomb and trithorax: 70 years and counting. Cell171, 34–57 (2017). [DOI] [PubMed] [Google Scholar]
12.Francis, N. J., Kingston, R. E. & Woodcock, C. L. Chromatin compaction by a polycomb group protein complex. Science306, 1574–1577 (2004). [DOI] [PubMed] [Google Scholar]
13.Szczurek, A. T., Dimitrova, E., Kelley, J. R., Blackledge, N. P. & Klose, R. J. The Polycomb system sustains promoters in a deep OFF state by limiting pre-initiation complex formation to counteract transcription. Nat. Cell Biol.26, 1700–1711 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Schultz, R. M., Stein, P. & Svoboda, P. The oocyte-to-embryo transition in mouse: past, present, and future. Biol. Reprod.99, 160–174 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Zheng, H. et al. Resetting epigenetic memory by reprogramming of histone modifications in mammals. Mol. Cell63, 1066–1079 (2016). [DOI] [PubMed] [Google Scholar]
16.Liu, X. et al. Distinct features of H3K4me3 and H3K27me3 chromatin domains in pre-implantation embryos. Nature537, 558–562 (2016). [DOI] [PubMed] [Google Scholar]
17.Mei, H. et al. H2AK119ub1 guides maternal inheritance and zygotic deposition of H3K27me3 in mouse embryos. Nat. Genet.53, 539–550 (2021). [DOI] [PubMed] [Google Scholar]
18.Chen, Z., Djekidel, M. N. & Zhang, Y. Distinct dynamics and functions of H2AK119ub1 and H3K27me3 in mouse preimplantation embryos. Nat. Genet.53, 551–563 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Zhu, Y. et al. Genomewide decoupling of H2AK119ub1 and H3K27me3 in early mouse development. Sci. Bull.66, 2489–2497 (2021). [DOI] [PubMed] [Google Scholar]
20.Posfai, E. et al. Polycomb function during oogenesis is required for mouse embryonic development. Genes Dev.26, 920–932 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Du, Z. et al. Polycomb group proteins regulate chromatin architecture in mouse oocytes and early embryos. Mol. Cell77, 825–839 e827 (2020). [DOI] [PubMed] [Google Scholar]
22.Inoue, A. Noncanonical imprinting: intergenerational epigenetic inheritance mediated by Polycomb complexes. Curr. Opin. Genet Dev.78, 102015 (2023). [DOI] [PubMed] [Google Scholar]
23.Inoue, A., Chen, Z., Yin, Q. & Zhang, Y. Maternal Eed knockout causes loss of H3K27me3 imprinting and random X inactivation in the extraembryonic cells. Genes Dev.32, 1525–1536 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Inoue, A., Jiang, L., Lu, F., Suzuki, T. & Zhang, Y. Maternal H3K27me3 controls DNA methylation-independent imprinting. Nature547, 419–424 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Chen, Z. & Zhang, Y. Maternal H3K27me3-dependent autosomal and X chromosome imprinting. Nat. Rev. Genet.21, 555–571 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Chen, Z., Yin, Q., Inoue, A., Zhang, C. & Zhang, Y. Allelic H3K27me3 to allelic DNA methylation switch maintains noncanonical imprinting in extraembryonic cells. Sci. Adv.5, eaay7246 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Matoba, S. et al. Noncanonical imprinting sustains embryonic development and restrains placental overgrowth. Genes Dev.36, 483–494 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Oberin, R. et al. Fetal growth delay caused by loss of non-canonical imprinting is resolved late in pregnancy and culminates in offspring overgrowth. Elife13. 10.7554/eLife.81875 (2024). [DOI] [PMC free article] [PubMed]
29.Liu, B. et al. Mapping putative enhancers in mouse oocytes and early embryos reveals TCF3/12 as key folliculogenesis regulators. Nat. Cell Biol.26, 962–974 (2024). [DOI] [PubMed] [Google Scholar]
30.Wang, M., Chen, Z. & Zhang, Y. CBP/p300 and HDAC activities regulate H3K27 acetylation dynamics and zygotic genome activation in mouse preimplantation embryos. EMBO J.41, e112012 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Burton, A. & Torres-Padilla, M. E. Epigenome dynamics in early mammalian embryogenesis. Nat. Rev. Genet. 10.1038/s41576-025-00831-4 (2025). [DOI] [PubMed]
32.Wu, K. et al. Dynamics of histone acetylation during human early embryogenesis. Cell Discov.9, 29 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Scheuermann, J. C. et al. Histone H2A deubiquitinase activity of the Polycomb repressive complex PR-DUB. Nature465, 243–247 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Ge, W. et al. Basis of the H2AK119 specificity of the Polycomb repressive deubiquitinase. Nature616, 176–182 (2023). [DOI] [PubMed] [Google Scholar]
35.Thomas, J. F. et al. Structural basis of histone H2A lysine 119 deubiquitination by Polycomb repressive deubiquitinase BAP1/ASXL1. Sci. Adv.9, eadg9832 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Bonnet, J. et al. PR-DUB preserves Polycomb repression by preventing excessive accumulation of H2Aub1, an antagonist of chromatin compaction. Genes Dev.36, 1046–1061 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Fursova, N. A. et al. BAP1 constrains pervasive H2AK119ub1 to control the transcriptional potential of the genome. Genes Dev.35, 749–770 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Conway, E. et al. BAP1 enhances Polycomb repression by counteracting widespread H2AK119ub1 deposition and chromatin condensation. Mol. Cell81, 3526–3541 e3528 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Li, R. et al. PR-DUB safeguards Polycomb repression through H2AK119ub1 restriction. Cell Prolif.56, e13457 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Kolovos, P. et al. PR-DUB maintains the expression of critical genes through FOXK1/2- and ASXL1/2/3-dependent recruitment to chromatin and H2AK119ub1 deubiquitination. Genome Res.30, 1119–1130 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Wang, L. et al. Resetting the epigenetic balance of Polycomb and COMPASS function at enhancers for cancer therapy. Nat. Med.24, 758–769 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Campagne, A. et al. BAP1 complex promotes transcription by opposing PRC1-mediated H2A ubiquitylation. Nat. Commun.10, 348 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Xiong, Z. et al. Ultrasensitive Ribo-seq reveals translational landscapes during mammalian oocyte-to-embryo transition and pre-implantation development. Nat. Cell Biol.24, 968–980 (2022). [DOI] [PubMed] [Google Scholar]
44.Zhang, C., Wang, M., Li, Y. & Zhang, Y. Profiling and functional characterization of maternal mRNA translation during mouse maternal-to-zygotic transition. Sci. Adv.8, eabj3967 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Joo, H. Y. et al. Regulation of cell cycle progression and gene expression by H2A deubiquitination. Nature449, 1068–1072 (2007). [DOI] [PubMed] [Google Scholar]
46.Li, F. et al. USP28 regulates deubiquitination of histone H2A and cell proliferation. Exp. Cell Res.379, 11–18 (2019). [DOI] [PubMed] [Google Scholar]
47.Nakagawa, T. et al. Deubiquitylation of histone H2A activates transcriptional initiation via trans-histone cross-talk with H3K4 di- and trimethylation. Genes Dev.22, 37–49 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Zhu, P. et al. A histone H2A deubiquitinase complex coordinating histone acetylation and H1 dissociation in transcriptional regulation. Mol. Cell27, 609–621 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Sahtoe, D. D., van Dijk, W. J., Ekkebus, R., Ovaa, H. & Sixma, T. K. BAP1/ASXL1 recruitment and activation for H2A deubiquitination. Nat. Commun.7, 10292 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Pannu, J. et al. Ubiquitin specific protease 21 is dispensable for normal development, hematopoiesis and lymphocyte differentiation. PLoS ONE10, e0117304 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Fan, Y. et al. USP21 negatively regulates antiviral response by acting as a RIG-I deubiquitinase. J. Exp. Med.211, 313–328 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Lan, Z. J., Xu, X. & Cooney, A. J. Differential oocyte-specific expression of Cre recombinase activity in GDF-9-iCre, Zp3cre, and Msx2Cre transgenic mice. Biol. Reprod.71, 1469–1474 (2004). [DOI] [PubMed] [Google Scholar]
53.Tamburri, S. et al. Histone H2AK119 mono-ubiquitination is essential for polycomb-mediated transcriptional repression. Mol. Cell77, 840–856 e845 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Blackledge, N. P. et al. PRC1 catalytic activity is central to polycomb system function. Mol. Cell77, 857–874 e859 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Zou, Z. et al. RNA m(5)C oxidation by TET2 regulates chromatin state and leukaemogenesis. Nature634, 986–994 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Blackledge, N. P. et al. Variant PRC1 complex-dependent H2A ubiquitylation drives PRC2 recruitment and polycomb domain formation. Cell157, 1445–1459 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Tavares, L. et al. RYBP-PRC1 complexes mediate H2A ubiquitylation at polycomb target sites independently of PRC2 and H3K27me3. Cell148, 664–678 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Cooper, S. et al. Targeting polycomb to pericentric heterochromatin in embryonic stem cells reveals a role for H2AK119u1 in PRC2 recruitment. Cell Rep.7, 1456–1470 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Kalb, R. et al. Histone H2A monoubiquitination promotes histone H3 methylation in Polycomb repression. Nat. Struct. Mol. Biol.21, 569–571 (2014). [DOI] [PubMed] [Google Scholar]
60.Zeng, Y. et al. EZHIP restricts noncanonical PRC2 binding and regulates H3K27me3 intergenerational inheritance and reprogramming. Cell Stem Cell32, 1741–1757 e1745 (2025). [DOI] [PubMed] [Google Scholar]
61.Matsuwaka, M., Kumon, M. & Inoue, A. H3K27 dimethylation dynamics reveal stepwise establishment of facultative heterochromatin in early mouse embryos. Nat. Cell Biol.27, 28–38 (2025). [DOI] [PubMed] [Google Scholar]
62.Kojima, M. L., Hoppe, C. & Giraldez, A. J. The maternal-to-zygotic transition: reprogramming of the cytoplasm and nucleus. Nat. Rev. Genet.26, 245–267 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Zou, Z., Wang, Q., Wu, X., Schultz, R. M. & Xie, W. Kick-starting the zygotic genome: licensors, specifiers, and beyond. EMBO Rep.25, 4113–4130 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Kravchenko, P. & Tachibana, K. Rise and SINE: roles of transcription factors and retrotransposons in zygotic genome activation. Nat. Rev. Mol. Cell Biol.26, 68–79 (2025). [DOI] [PubMed] [Google Scholar]
65.Park, S. J., Shirahige, K., Ohsugi, M. & Nakai, K. DBTMEE: a database of transcriptome in mouse early embryos. Nucleic Acids Res.43, D771–D776 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Falco, G. et al. Zscan4: a novel gene expressed exclusively in late 2-cell embryos and embryonic stem cells. Dev. Biol.307, 539–550 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Modzelewski, A. J. et al. A mouse-specific retrotransposon drives a conserved Cdk2ap1 isoform essential for development. Cell184, 5541–5558 e5522 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Lai, F. et al. NR5A2 connects zygotic genome activation to the first lineage segregation in totipotent embryos. Cell Res.33, 952–966 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Gassler, J. et al. Zygotic genome activation by the totipotency pioneer factor Nr5a2. Science378, 1305–1315 (2022). [DOI] [PubMed] [Google Scholar]
70.Zhao, Y. et al. Nr5a2 ensures inner cell mass formation in mouse blastocyst. Cell Rep.43, 113840 (2024). [DOI] [PubMed] [Google Scholar]
71.Festuccia, N. et al. Nr5a2 is dispensable for zygotic genome activation but essential for morula development. Science386, eadg7325 (2024). [DOI] [PubMed] [Google Scholar]
72.Guo, Y. et al. Obox4 promotes zygotic genome activation upon loss of Dux. Elife13. 10.7554/eLife.95856 (2024). [DOI] [PMC free article] [PubMed]
73.Ji, S. et al. OBOX regulates mouse zygotic genome activation and early development. Nature. 10.1038/s41586-023-06428-3 (2023). [DOI] [PMC free article] [PubMed]
74.Yang, J., Cook, L. & Chen, Z. Systematic evaluation of retroviral LTRs as cis-regulatory elements in mouse embryos. Cell Rep.43, 113775 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
75.de la Rosa, S., Del Mar Rigual, M., Vargiu, P., Ortega, S. & Djouder, N. Endogenous retroviruses shape pluripotency specification in mouse embryos. Sci. Adv.10, eadk9394 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Sakashita, A. et al. Transcription of MERVL retrotransposons is required for preimplantation embryo development. Nat. Genet.55, 484–495 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Rong, Y. et al. USP16-mediated histone H2A lysine-119 deubiquitination during oocyte maturation is a prerequisite for zygotic genome activation. Nucleic Acids Res.50, 5599–5616 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
78.Jain, S. U. et al. PFA ependymoma-associated protein EZHIP inhibits PRC2 activity through a H3 K27M-like mechanism. Nat. Commun.10, 2146 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
79.Ragazzini, R. et al. EZHIP constrains polycomb repressive complex 2 activity in germ cells. Nat. Commun.10, 3858 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
80.Zhang, Y. et al. ASXLs binding to the PHD2/3 fingers of MLL4 provides a mechanism for the recruitment of BAP1 to active enhancers. Nat. Commun.15, 4883 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
81.Zhao, Z. et al. CARM1-mediated methylation of ASXL2 impairs tumor-suppressive function of MLL3/COMPASS. Sci. Adv.8, eadd3339 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Szczepanski, A. P. et al. ASXL3 bridges BRD4 to BAP1 complex and governs enhancer activity in small cell lung cancer. Genome Med12, 63 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
83.Zhao, J. et al. H2AK119ub1 differentially fine-tunes gene expression by modulating canonical PRC1- and H1-dependent chromatin compaction. Mol. Cell84, 1191–1205 e1197 (2024). [DOI] [PubMed] [Google Scholar]
84.Li, T. et al. Non-canonical PRC1.1 licenses transcriptional response to enable Treg plasticity in immune adaptation. Mol Cell. 10.1016/j.molcel.2025.05.029 (2025). [DOI] [PubMed]
85.Den, Z., Cheng, X., Merched-Sauvage, M. & Koch, P. J. Desmocollin 3 is required for pre-implantation development of the mouse embryo. J. Cell Sci.119, 482–489 (2006). [DOI] [PubMed] [Google Scholar]
86.Zhang, K., Hansen, P. J. & Ealy, A. D. Fibroblast growth factor 10 enhances bovine oocyte maturation and developmental competence in vitro. Reproduction140, 815–826 (2010). [DOI] [PubMed] [Google Scholar]
87.Gaydos, L. J., Wang, W. & Strome, S. Gene repression. H3K27me and PRC2 transmit a memory of repression across generations and during development. Science345, 1515–1518 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
88.Zenk, F. et al. Germ line-inherited H3K27me3 restricts enhancer function during maternal-to-zygotic transition. Science357, 212–216 (2017). [DOI] [PubMed] [Google Scholar]
89.Hanna, C. W. et al. Endogenous retroviral insertions drive non-canonical imprinting in extra-embryonic tissues. Genome Biol.20, 225 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
90.Sugimoto, M. et al. A simple and robust method for establishing homogeneous mouse epiblast stem cell lines by Wnt inhibition. Stem Cell Rep.4, 744–757 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
91.Inoue, A. & Zhang, Y. Nucleosome assembly is required for nuclear pore complex assembly in mouse zygotes. Nat. Struct. Mol. Biol.21, 609–616 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
92.Zhang, C. et al. The chromatin remodeler Snf2h is essential for oocyte meiotic cell cycle progression. Genes Dev.34, 166–178 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
93.Chen, Z., Zhang, C. & Zhang, Y. Chromatin analyses in mouse oocytes and preimplantation embryos. Methods Mol. Biol.2923, 33–44 (2025). [DOI] [PubMed] [Google Scholar]
94.Hayashi, R. & Inoue, A. Low-input CUT&RUN for mouse oocytes and preimplantation embryos. Methods Mol. Biol.2577, 83–92 (2023). [DOI] [PubMed] [Google Scholar]
95.Hu, M. et al. An epigenomic roadmap primes non-growing oocytes for maturation and early embryogenesis. bioRxiv10.1101/2025.07.06.663377 (2025).
96.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). [DOI] [PMC free article] [PubMed] [Google Scholar]
97.Xu, S. et al. Using clusterProfiler to characterize multiomics data. Nat. Protoc.19, 3292–3320 (2024). [DOI] [PubMed] [Google Scholar]
98.Ramirez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res.44, W160–W165 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
99.Krueger, F. & Andrews, S. R. SNPsplit: allele-specific splitting of alignments between genomes with known SNP genotypes. F1000Res5, 1479 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
100.Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods9, 357–359 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
101.Tarasov, A., Vilella, A. J., Cuppen, E., Nijman, I. J. & Prins, P. Sambamba: fast processing of NGS alignment formats. Bioinformatics31, 2032–2034 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
102.Skene, P. J. & Henikoff, S. An efficient targeted nuclease strategy for high-resolution mapping of DNA binding sites. Elife6. 10.7554/eLife.21856 (2017). [DOI] [PMC free article] [PubMed]
103.Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics26, 841–842 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
104.Zhu, L. J. et al. ChIPpeakAnno: a Bioconductor package to annotate ChIP-seq and ChIP-chip data. BMC Bioinformatics11, 237 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
105.Lawrence, M. et al. Software for computing and annotating genomic ranges. PLoS Comput. Biol.9, e1003118 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
106.Yu, G., Wang, L. G. & He, Q. Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics31, 2382–2383 (2015). [DOI] [PubMed] [Google Scholar]
107.Liu, S. et al. From 1D sequence to 3D chromatin dynamics and cellular functions: a phase separation perspective. Nucleic Acids Res.46, 9367–9383 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
108.Lawrence, M., Gentleman, R. & Carey, V. rtracklayer: an R package for interfacing with genome browsers. Bioinformatics25, 1841–1842 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
109.Kent, W. J. et al. The human genome browser at UCSC. Genome Res.12, 996–1006 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(21.9MB, pdf)}

41467_2026_70845_MOESM2_ESM.pdf^{(280.3KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1^{(19.1MB, xlsx)}

Supplementary Data 2^{(179.6KB, xlsx)}

Supplementary Data 3^{(21.9KB, xlsx)}

Reporting Summary^{(103.4KB, pdf)}

Transparent Peer Review file^{(1.7MB, pdf)}

Source Data^{(446.7KB, xlsx)}

Data Availability Statement

The code for this study is available in Zenodo at 10.5281/zenodo.18701494.

[CR1] 1.Piunti, A. & Shilatifard, A. The roles of Polycomb repressive complexes in mammalian development and cancer. Nat. Rev. Mol. Cell Biol.22, 326–345 (2021). [DOI] [PubMed] [Google Scholar]

[CR2] 2.Kim, J. J. & Kingston, R. E. Context-specific Polycomb mechanisms in development. Nat. Rev. Genet.23, 680–695 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Blackledge, N. P. & Klose, R. J. The molecular principles of gene regulation by Polycomb repressive complexes. Nat. Rev. Mol. Cell Biol.22, 815–833 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Tamburri, S., Rustichelli, S., Amato, S. & Pasini, D. Navigating the complexity of Polycomb repression: enzymatic cores and regulatory modules. Mol. Cell84, 3381–3405 (2024). [DOI] [PubMed] [Google Scholar]

[CR5] 5.Wang, H. et al. Role of histone H2A ubiquitination in Polycomb silencing. Nature431, 873–878 (2004). [DOI] [PubMed] [Google Scholar]

[CR6] 6.de Napoles, M. et al. Polycomb group proteins Ring1A/B link ubiquitylation of histone H2A to heritable gene silencing and X inactivation. Dev. Cell7, 663–676 (2004). [DOI] [PubMed] [Google Scholar]

[CR7] 7.Cao, R. et al. Role of histone H3 lysine 27 methylation in Polycomb-group silencing. Science298, 1039–1043 (2002). [DOI] [PubMed] [Google Scholar]

[CR8] 8.Muller, J. et al. Histone methyltransferase activity of a Drosophila Polycomb group repressor complex. Cell111, 197–208 (2002). [DOI] [PubMed] [Google Scholar]

[CR9] 9.Kuzmichev, A., Nishioka, K., Erdjument-Bromage, H., Tempst, P. & Reinberg, D. Histone methyltransferase activity associated with a human multiprotein complex containing the Enhancer of Zeste protein. Genes Dev.16, 2893–2905 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Margueron, R. & Reinberg, D. The Polycomb complex PRC2 and its mark in life. Nature469, 343–349 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Schuettengruber, B., Bourbon, H. M., Di Croce, L. & Cavalli, G. Genome regulation by polycomb and trithorax: 70 years and counting. Cell171, 34–57 (2017). [DOI] [PubMed] [Google Scholar]

[CR12] 12.Francis, N. J., Kingston, R. E. & Woodcock, C. L. Chromatin compaction by a polycomb group protein complex. Science306, 1574–1577 (2004). [DOI] [PubMed] [Google Scholar]

[CR13] 13.Szczurek, A. T., Dimitrova, E., Kelley, J. R., Blackledge, N. P. & Klose, R. J. The Polycomb system sustains promoters in a deep OFF state by limiting pre-initiation complex formation to counteract transcription. Nat. Cell Biol.26, 1700–1711 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Schultz, R. M., Stein, P. & Svoboda, P. The oocyte-to-embryo transition in mouse: past, present, and future. Biol. Reprod.99, 160–174 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Zheng, H. et al. Resetting epigenetic memory by reprogramming of histone modifications in mammals. Mol. Cell63, 1066–1079 (2016). [DOI] [PubMed] [Google Scholar]

[CR16] 16.Liu, X. et al. Distinct features of H3K4me3 and H3K27me3 chromatin domains in pre-implantation embryos. Nature537, 558–562 (2016). [DOI] [PubMed] [Google Scholar]

[CR17] 17.Mei, H. et al. H2AK119ub1 guides maternal inheritance and zygotic deposition of H3K27me3 in mouse embryos. Nat. Genet.53, 539–550 (2021). [DOI] [PubMed] [Google Scholar]

[CR18] 18.Chen, Z., Djekidel, M. N. & Zhang, Y. Distinct dynamics and functions of H2AK119ub1 and H3K27me3 in mouse preimplantation embryos. Nat. Genet.53, 551–563 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Zhu, Y. et al. Genomewide decoupling of H2AK119ub1 and H3K27me3 in early mouse development. Sci. Bull.66, 2489–2497 (2021). [DOI] [PubMed] [Google Scholar]

[CR20] 20.Posfai, E. et al. Polycomb function during oogenesis is required for mouse embryonic development. Genes Dev.26, 920–932 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Du, Z. et al. Polycomb group proteins regulate chromatin architecture in mouse oocytes and early embryos. Mol. Cell77, 825–839 e827 (2020). [DOI] [PubMed] [Google Scholar]

[CR22] 22.Inoue, A. Noncanonical imprinting: intergenerational epigenetic inheritance mediated by Polycomb complexes. Curr. Opin. Genet Dev.78, 102015 (2023). [DOI] [PubMed] [Google Scholar]

[CR23] 23.Inoue, A., Chen, Z., Yin, Q. & Zhang, Y. Maternal Eed knockout causes loss of H3K27me3 imprinting and random X inactivation in the extraembryonic cells. Genes Dev.32, 1525–1536 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Inoue, A., Jiang, L., Lu, F., Suzuki, T. & Zhang, Y. Maternal H3K27me3 controls DNA methylation-independent imprinting. Nature547, 419–424 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Chen, Z. & Zhang, Y. Maternal H3K27me3-dependent autosomal and X chromosome imprinting. Nat. Rev. Genet.21, 555–571 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Chen, Z., Yin, Q., Inoue, A., Zhang, C. & Zhang, Y. Allelic H3K27me3 to allelic DNA methylation switch maintains noncanonical imprinting in extraembryonic cells. Sci. Adv.5, eaay7246 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Matoba, S. et al. Noncanonical imprinting sustains embryonic development and restrains placental overgrowth. Genes Dev.36, 483–494 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Oberin, R. et al. Fetal growth delay caused by loss of non-canonical imprinting is resolved late in pregnancy and culminates in offspring overgrowth. Elife13. 10.7554/eLife.81875 (2024). [DOI] [PMC free article] [PubMed]

[CR29] 29.Liu, B. et al. Mapping putative enhancers in mouse oocytes and early embryos reveals TCF3/12 as key folliculogenesis regulators. Nat. Cell Biol.26, 962–974 (2024). [DOI] [PubMed] [Google Scholar]

[CR30] 30.Wang, M., Chen, Z. & Zhang, Y. CBP/p300 and HDAC activities regulate H3K27 acetylation dynamics and zygotic genome activation in mouse preimplantation embryos. EMBO J.41, e112012 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Burton, A. & Torres-Padilla, M. E. Epigenome dynamics in early mammalian embryogenesis. Nat. Rev. Genet. 10.1038/s41576-025-00831-4 (2025). [DOI] [PubMed]

[CR32] 32.Wu, K. et al. Dynamics of histone acetylation during human early embryogenesis. Cell Discov.9, 29 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Scheuermann, J. C. et al. Histone H2A deubiquitinase activity of the Polycomb repressive complex PR-DUB. Nature465, 243–247 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Ge, W. et al. Basis of the H2AK119 specificity of the Polycomb repressive deubiquitinase. Nature616, 176–182 (2023). [DOI] [PubMed] [Google Scholar]

[CR35] 35.Thomas, J. F. et al. Structural basis of histone H2A lysine 119 deubiquitination by Polycomb repressive deubiquitinase BAP1/ASXL1. Sci. Adv.9, eadg9832 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Bonnet, J. et al. PR-DUB preserves Polycomb repression by preventing excessive accumulation of H2Aub1, an antagonist of chromatin compaction. Genes Dev.36, 1046–1061 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Fursova, N. A. et al. BAP1 constrains pervasive H2AK119ub1 to control the transcriptional potential of the genome. Genes Dev.35, 749–770 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Conway, E. et al. BAP1 enhances Polycomb repression by counteracting widespread H2AK119ub1 deposition and chromatin condensation. Mol. Cell81, 3526–3541 e3528 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Li, R. et al. PR-DUB safeguards Polycomb repression through H2AK119ub1 restriction. Cell Prolif.56, e13457 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Kolovos, P. et al. PR-DUB maintains the expression of critical genes through FOXK1/2- and ASXL1/2/3-dependent recruitment to chromatin and H2AK119ub1 deubiquitination. Genome Res.30, 1119–1130 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR41] 41.Wang, L. et al. Resetting the epigenetic balance of Polycomb and COMPASS function at enhancers for cancer therapy. Nat. Med.24, 758–769 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Campagne, A. et al. BAP1 complex promotes transcription by opposing PRC1-mediated H2A ubiquitylation. Nat. Commun.10, 348 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Xiong, Z. et al. Ultrasensitive Ribo-seq reveals translational landscapes during mammalian oocyte-to-embryo transition and pre-implantation development. Nat. Cell Biol.24, 968–980 (2022). [DOI] [PubMed] [Google Scholar]

[CR44] 44.Zhang, C., Wang, M., Li, Y. & Zhang, Y. Profiling and functional characterization of maternal mRNA translation during mouse maternal-to-zygotic transition. Sci. Adv.8, eabj3967 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Joo, H. Y. et al. Regulation of cell cycle progression and gene expression by H2A deubiquitination. Nature449, 1068–1072 (2007). [DOI] [PubMed] [Google Scholar]

[CR46] 46.Li, F. et al. USP28 regulates deubiquitination of histone H2A and cell proliferation. Exp. Cell Res.379, 11–18 (2019). [DOI] [PubMed] [Google Scholar]

[CR47] 47.Nakagawa, T. et al. Deubiquitylation of histone H2A activates transcriptional initiation via trans-histone cross-talk with H3K4 di- and trimethylation. Genes Dev.22, 37–49 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Zhu, P. et al. A histone H2A deubiquitinase complex coordinating histone acetylation and H1 dissociation in transcriptional regulation. Mol. Cell27, 609–621 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Sahtoe, D. D., van Dijk, W. J., Ekkebus, R., Ovaa, H. & Sixma, T. K. BAP1/ASXL1 recruitment and activation for H2A deubiquitination. Nat. Commun.7, 10292 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Pannu, J. et al. Ubiquitin specific protease 21 is dispensable for normal development, hematopoiesis and lymphocyte differentiation. PLoS ONE10, e0117304 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Fan, Y. et al. USP21 negatively regulates antiviral response by acting as a RIG-I deubiquitinase. J. Exp. Med.211, 313–328 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Lan, Z. J., Xu, X. & Cooney, A. J. Differential oocyte-specific expression of Cre recombinase activity in GDF-9-iCre, Zp3cre, and Msx2Cre transgenic mice. Biol. Reprod.71, 1469–1474 (2004). [DOI] [PubMed] [Google Scholar]

[CR53] 53.Tamburri, S. et al. Histone H2AK119 mono-ubiquitination is essential for polycomb-mediated transcriptional repression. Mol. Cell77, 840–856 e845 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Blackledge, N. P. et al. PRC1 catalytic activity is central to polycomb system function. Mol. Cell77, 857–874 e859 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Zou, Z. et al. RNA m(5)C oxidation by TET2 regulates chromatin state and leukaemogenesis. Nature634, 986–994 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Blackledge, N. P. et al. Variant PRC1 complex-dependent H2A ubiquitylation drives PRC2 recruitment and polycomb domain formation. Cell157, 1445–1459 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR57] 57.Tavares, L. et al. RYBP-PRC1 complexes mediate H2A ubiquitylation at polycomb target sites independently of PRC2 and H3K27me3. Cell148, 664–678 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Cooper, S. et al. Targeting polycomb to pericentric heterochromatin in embryonic stem cells reveals a role for H2AK119u1 in PRC2 recruitment. Cell Rep.7, 1456–1470 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] 59.Kalb, R. et al. Histone H2A monoubiquitination promotes histone H3 methylation in Polycomb repression. Nat. Struct. Mol. Biol.21, 569–571 (2014). [DOI] [PubMed] [Google Scholar]

[CR60] 60.Zeng, Y. et al. EZHIP restricts noncanonical PRC2 binding and regulates H3K27me3 intergenerational inheritance and reprogramming. Cell Stem Cell32, 1741–1757 e1745 (2025). [DOI] [PubMed] [Google Scholar]

[CR61] 61.Matsuwaka, M., Kumon, M. & Inoue, A. H3K27 dimethylation dynamics reveal stepwise establishment of facultative heterochromatin in early mouse embryos. Nat. Cell Biol.27, 28–38 (2025). [DOI] [PubMed] [Google Scholar]

[CR62] 62.Kojima, M. L., Hoppe, C. & Giraldez, A. J. The maternal-to-zygotic transition: reprogramming of the cytoplasm and nucleus. Nat. Rev. Genet.26, 245–267 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR63] 63.Zou, Z., Wang, Q., Wu, X., Schultz, R. M. & Xie, W. Kick-starting the zygotic genome: licensors, specifiers, and beyond. EMBO Rep.25, 4113–4130 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR64] 64.Kravchenko, P. & Tachibana, K. Rise and SINE: roles of transcription factors and retrotransposons in zygotic genome activation. Nat. Rev. Mol. Cell Biol.26, 68–79 (2025). [DOI] [PubMed] [Google Scholar]

[CR65] 65.Park, S. J., Shirahige, K., Ohsugi, M. & Nakai, K. DBTMEE: a database of transcriptome in mouse early embryos. Nucleic Acids Res.43, D771–D776 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] 66.Falco, G. et al. Zscan4: a novel gene expressed exclusively in late 2-cell embryos and embryonic stem cells. Dev. Biol.307, 539–550 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR67] 67.Modzelewski, A. J. et al. A mouse-specific retrotransposon drives a conserved Cdk2ap1 isoform essential for development. Cell184, 5541–5558 e5522 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR68] 68.Lai, F. et al. NR5A2 connects zygotic genome activation to the first lineage segregation in totipotent embryos. Cell Res.33, 952–966 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR69] 69.Gassler, J. et al. Zygotic genome activation by the totipotency pioneer factor Nr5a2. Science378, 1305–1315 (2022). [DOI] [PubMed] [Google Scholar]

[CR70] 70.Zhao, Y. et al. Nr5a2 ensures inner cell mass formation in mouse blastocyst. Cell Rep.43, 113840 (2024). [DOI] [PubMed] [Google Scholar]

[CR71] 71.Festuccia, N. et al. Nr5a2 is dispensable for zygotic genome activation but essential for morula development. Science386, eadg7325 (2024). [DOI] [PubMed] [Google Scholar]

[CR72] 72.Guo, Y. et al. Obox4 promotes zygotic genome activation upon loss of Dux. Elife13. 10.7554/eLife.95856 (2024). [DOI] [PMC free article] [PubMed]

[CR73] 73.Ji, S. et al. OBOX regulates mouse zygotic genome activation and early development. Nature. 10.1038/s41586-023-06428-3 (2023). [DOI] [PMC free article] [PubMed]

[CR74] 74.Yang, J., Cook, L. & Chen, Z. Systematic evaluation of retroviral LTRs as cis-regulatory elements in mouse embryos. Cell Rep.43, 113775 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR75] 75.de la Rosa, S., Del Mar Rigual, M., Vargiu, P., Ortega, S. & Djouder, N. Endogenous retroviruses shape pluripotency specification in mouse embryos. Sci. Adv.10, eadk9394 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR76] 76.Sakashita, A. et al. Transcription of MERVL retrotransposons is required for preimplantation embryo development. Nat. Genet.55, 484–495 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR77] 77.Rong, Y. et al. USP16-mediated histone H2A lysine-119 deubiquitination during oocyte maturation is a prerequisite for zygotic genome activation. Nucleic Acids Res.50, 5599–5616 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR78] 78.Jain, S. U. et al. PFA ependymoma-associated protein EZHIP inhibits PRC2 activity through a H3 K27M-like mechanism. Nat. Commun.10, 2146 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR79] 79.Ragazzini, R. et al. EZHIP constrains polycomb repressive complex 2 activity in germ cells. Nat. Commun.10, 3858 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR80] 80.Zhang, Y. et al. ASXLs binding to the PHD2/3 fingers of MLL4 provides a mechanism for the recruitment of BAP1 to active enhancers. Nat. Commun.15, 4883 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR81] 81.Zhao, Z. et al. CARM1-mediated methylation of ASXL2 impairs tumor-suppressive function of MLL3/COMPASS. Sci. Adv.8, eadd3339 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR82] 82.Szczepanski, A. P. et al. ASXL3 bridges BRD4 to BAP1 complex and governs enhancer activity in small cell lung cancer. Genome Med12, 63 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR83] 83.Zhao, J. et al. H2AK119ub1 differentially fine-tunes gene expression by modulating canonical PRC1- and H1-dependent chromatin compaction. Mol. Cell84, 1191–1205 e1197 (2024). [DOI] [PubMed] [Google Scholar]

[CR84] 84.Li, T. et al. Non-canonical PRC1.1 licenses transcriptional response to enable Treg plasticity in immune adaptation. Mol Cell. 10.1016/j.molcel.2025.05.029 (2025). [DOI] [PubMed]

[CR85] 85.Den, Z., Cheng, X., Merched-Sauvage, M. & Koch, P. J. Desmocollin 3 is required for pre-implantation development of the mouse embryo. J. Cell Sci.119, 482–489 (2006). [DOI] [PubMed] [Google Scholar]

[CR86] 86.Zhang, K., Hansen, P. J. & Ealy, A. D. Fibroblast growth factor 10 enhances bovine oocyte maturation and developmental competence in vitro. Reproduction140, 815–826 (2010). [DOI] [PubMed] [Google Scholar]

[CR87] 87.Gaydos, L. J., Wang, W. & Strome, S. Gene repression. H3K27me and PRC2 transmit a memory of repression across generations and during development. Science345, 1515–1518 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR88] 88.Zenk, F. et al. Germ line-inherited H3K27me3 restricts enhancer function during maternal-to-zygotic transition. Science357, 212–216 (2017). [DOI] [PubMed] [Google Scholar]

[CR89] 89.Hanna, C. W. et al. Endogenous retroviral insertions drive non-canonical imprinting in extra-embryonic tissues. Genome Biol.20, 225 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR90] 90.Sugimoto, M. et al. A simple and robust method for establishing homogeneous mouse epiblast stem cell lines by Wnt inhibition. Stem Cell Rep.4, 744–757 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR91] 91.Inoue, A. & Zhang, Y. Nucleosome assembly is required for nuclear pore complex assembly in mouse zygotes. Nat. Struct. Mol. Biol.21, 609–616 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR92] 92.Zhang, C. et al. The chromatin remodeler Snf2h is essential for oocyte meiotic cell cycle progression. Genes Dev.34, 166–178 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR93] 93.Chen, Z., Zhang, C. & Zhang, Y. Chromatin analyses in mouse oocytes and preimplantation embryos. Methods Mol. Biol.2923, 33–44 (2025). [DOI] [PubMed] [Google Scholar]

[CR94] 94.Hayashi, R. & Inoue, A. Low-input CUT&RUN for mouse oocytes and preimplantation embryos. Methods Mol. Biol.2577, 83–92 (2023). [DOI] [PubMed] [Google Scholar]

[CR95] 95.Hu, M. et al. An epigenomic roadmap primes non-growing oocytes for maturation and early embryogenesis. bioRxiv10.1101/2025.07.06.663377 (2025).

[CR96] 96.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). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR97] 97.Xu, S. et al. Using clusterProfiler to characterize multiomics data. Nat. Protoc.19, 3292–3320 (2024). [DOI] [PubMed] [Google Scholar]

[CR98] 98.Ramirez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res.44, W160–W165 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR99] 99.Krueger, F. & Andrews, S. R. SNPsplit: allele-specific splitting of alignments between genomes with known SNP genotypes. F1000Res5, 1479 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR100] 100.Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods9, 357–359 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR101] 101.Tarasov, A., Vilella, A. J., Cuppen, E., Nijman, I. J. & Prins, P. Sambamba: fast processing of NGS alignment formats. Bioinformatics31, 2032–2034 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR102] 102.Skene, P. J. & Henikoff, S. An efficient targeted nuclease strategy for high-resolution mapping of DNA binding sites. Elife6. 10.7554/eLife.21856 (2017). [DOI] [PMC free article] [PubMed]

[CR103] 103.Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics26, 841–842 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR104] 104.Zhu, L. J. et al. ChIPpeakAnno: a Bioconductor package to annotate ChIP-seq and ChIP-chip data. BMC Bioinformatics11, 237 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR105] 105.Lawrence, M. et al. Software for computing and annotating genomic ranges. PLoS Comput. Biol.9, e1003118 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR106] 106.Yu, G., Wang, L. G. & He, Q. Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics31, 2382–2383 (2015). [DOI] [PubMed] [Google Scholar]

[CR107] 107.Liu, S. et al. From 1D sequence to 3D chromatin dynamics and cellular functions: a phase separation perspective. Nucleic Acids Res.46, 9367–9383 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR108] 108.Lawrence, M., Gentleman, R. & Carey, V. rtracklayer: an R package for interfacing with genome browsers. Bioinformatics25, 1841–1842 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR109] 109.Kent, W. J. et al. The human genome browser at UCSC. Genome Res.12, 996–1006 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Polycomb repressive-deubiquitinase complex safeguards oocyte epigenome and female fertility by restraining Polycomb activity

Jinwen Kang

Peiyao Liu

Shoko Ichimura

Lauryn Cook

Mengwen Hu

Satoshi H Namekawa

Zhiyuan Chen

Abstract

Introduction

Results

Identification of candidate H2AK119ub1 deubiquitinases in early embryos

Fig. 1. Loss of BAP1 causes pervasive increase of H2AK119ub1 and primarily results in gene downregulation in oocytes.

BAP1 limits genome wide pervasive H2AK119ub1 accumulation in oocytes

BAP1 primarily activates gene expression in oocytes

BAP1 preserves H3K27ac by preventing ectopic H3K27me3 deposition in oocytes

Fig. 2. BAP1 preserves H3K27ac by preventing ectopic H3K27me3 deposition in oocytes.

BAP1 plays a limited role in Polycomb-mediated silencing in oocytes

Fig. 3. BAP1 plays a limited role in Polycomb-mediated gene silencing in oocytes.

Loss of maternal BAP1 severely impairs preimplantation development

Fig. 4. Loss of BAP1 severely impairs female reproductive capacity.

Loss of maternal BAP1 causes maternal-to-zygotic transition (MZT) defects

Fig. 5. Loss of maternal BAP1 causes defective maternal-to-zygotic transition.

Maternal BAP1 regulates embryonic enhancer activity during ZGA

Fig. 6. Aberrant H3K27me3 landscapes in Bap1 null oocytes, but not H3K27ac changes, persist in early embryos.

Aberrant H3K27me3 landscape, but not H3K27ac changes, in BAP1-null oocytes persists into early embryos

The ectopic H3K27me3 domains in Bap1 matKO embryos are resolved by early post-implantation stages

Fig. 7. Ectopic H3K27me3 in Bap1-null oocytes persist through the morula stage but are lost in post-implantation embryos.

Discussion

Fig. 8. A model illustrating the functions of PR-DUB in oocyte epigenome and female fertility.

Methods

Mouse strains and breeding

Mouse superovulation, IVF, embryo culture, and fertility assessment

Whole-mount immunostaining and histology analyses

Plasmid construction and mRNA preparation

Micro-injection of H2A DUB mRNAs

Western blotting

RNA-seq and CUT&RUN library preparation and sequencing

RNA-seq data processing

CUT&RUN data processing

Identification of H3K27ac-lost regions in Bap1 CKO FGOs

Analyses at PcG targets

Annotations and genomic intervals

Visualization and statistical analyses

Reanalysis of public datasets

Statistics and reproducibility

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Data availability

Code availability

Competing interests

Footnotes

Supplementary information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases