The catalytic core of PRC1 is a dimer consisting of really interesting new gene 1A or B (RING1A/B) which functions as ubiquitin E3 ligases and one of the six polycomb group RING finger (PCGF) orthologs 6 which regulates PRC1 enzymatic activity (Dobrinić et al., 2021). PRC1.1-1.6, the six major groups of PRC1 complex, are defined by the different associated PCGF orthologs (Gao et al., 2012). RING1A/B and all PCGF proteins contain two conserved protein domains: Ring finger domain and WD40-associated ubiquitin-like (RAWUL) domain (Geng and Gao, 2020; de Potter et al., 2023). The similar RING domain mediates the dimerization of RING1 and PCGF, and the catalytic core of PRC1 is formed. This RING domain dimer facilitates the binding of the E2-conjugating enzyme to PRC1. Notably, PCGF2 and PCGF4 specifically comprise Proline Serine rich (PS) domains. And PCGF proteins can interact with various auxiliary subunits through their RAWUL domains (Geng and Gao, 2020; de Potter et al., 2023). Depending on the subunit associated with the RING-PCGF core, PRC1 can be further categorized as canonical PRC1 (cPRC1) and noncanonical PRC1 (ncPRC1).

cPRC1 complexes compromise RING1A/B, PCGF2/4 and one of the five chromobox (CBX) proteins (CBX2, CBX4, CBX6, CBX7 or CBX8) (Gao et al., 2012). The CBX proteins contain N-terminal chromodomains which are responsible for recruitment and stabilization of cPRC1 to specific regions of the chromatin, especially H3K27me3-rich regions. And the AT-hook of CBX proteins facilitates their binding to AT-rich major satellites DNA sequence (Bernstein et al., 2006; Gao et al., 2012). Moreover, cPRC1 also contains one of the polyhomeotic-like proteins (PHC1, PHC2 or PHC3) and sex comb on midleg homolog 1 or 2 (SCMH1/2), which facilitate the polymerization of PRC1 complexes via the sterile alpha motif (SAM) domain (Di Croce and Helin, 2013; Wani et al., 2016; Geng and Gao, 2020). Furthermore, PHC proteins can bind to the RAWUL domains of PCGF2/4 but not PCGF1/3 (Kim and Kingston, 2022).

Unlike cPRC1, the ncPRC1 complexes are made up of any of the six PCGF proteins (PCGF1-6), RING1/B and YY1 binding protein (RYBP) or its homolog YY1-associated factor 2 (YAF2) (Chan et al., 2018). RYBP and YAF2 can competitively bind to RING1B and stimulate the enzymatic activity of RING1B through a positive feedback model (Gao et al., 2012; Chagraoui et al., 2018; Liu et al., 2023). Furthermore, recent research revealed that RYBP/YAF can bind to H2AK119ub1 to further promote the spreading of H2AK119ub1 to neighbor nucleosomes (Zhao et al., 2020).

Interestingly, some subunits are found to specially exist in one of the ncPRC1-6 complexes. For example, the CXXC domain of the Lysine Demethylase 2B (KDM2B), a component of ncPRC1.1, can recognize CpG islands, and further contribute to the recruitment of ncPRC1.1 to hypomethylated CpG-rich promoters (Farcas et al., 2012; Sugishita et al., 2021). BCL6 corepressor (BCOR) and BCL6 corepressor like 1 (BCORL1) subunits are required for the stability of PCGF1 and are essential for coupling KDM2B to the enzymic core of ncPRC1.1 (Wong et al., 2016; Schaefer et al., 2022). Fibrosin (FBRS) and Autism susceptibility candidate 2 (AUTS2) are two paralog proteins of ncPRC1.3 and ncPRC1.5 (Collier et al., 2022). AUTS2 can recruit histone acetyltransferase EP300 to PRC1 complexes (Castanza et al., 2021; Pauli et al., 2021). While the role of FBRS in ncPRC1.3/5 remains elusive. Several proteins associated with PCGF6 like L3MBTL histone methyl-lysine binding protein 2 (L3MBTL2), E2F transcription factor 6 (E2F6) and MAX gene-associated protein (MGA) are common subunits of ncPRC1.6. Generally, these proteins act collaboratively to facilitate the loading of ncPRC1.6 to its target sites (Huang et al., 2018; 2022; Dahlet et al., 2021).

The catalytic core of PRC2 comprises four subunits: enhancer of Zeste homologue 1/2 (EZH1/2), embryonic ectoderm development (EED), suppressor of Zeste 12 (SUZ12) and RB-binding protein 4 or 7 (RBBP4/7) (Vijayanathan et al., 2022). EZH proteins contain a SET domain, endowing PRC2 with histone lysine methyltransferase activity (Shen et al., 2008; Lee et al., 2018, 2). EED stabilizes EZH1/2 in the PRC2 and stimulates the methylation activity of EZH1/2 (Margueron et al., 2009; Hsu et al., 2020). The WD40 repeat domain (WDR) of EED is responsible for the location of H3K27me3 on PRC2 (Oksuz et al., 2018). SUZ12 functions as a structured platform interacting with other three core proteins, which contributes to the stability of the complex (Kasinath et al., 2018). WDR-containing proteins RBBP4/7 are another core component of PRC2 and modulate the recruitment of PRC2 to chromatin (Schapira et al., 2017; Huang et al., 2021).

The PRC2 core assembling with auxiliary proteins forms two distinct homocomplexes-PRC2.1 and PRC2.2 (Petracovici and Bonasio, 2021; de Potter et al., 2023). These auxiliary subunits, including Zinc finger protein AE binding protein 2 (AEBP2), Polycomb-like homologues (PCLs), PRC2-associated LCOR isoform 1 or 2 (PALI1/2), Polycomb repressive complex 2-associated protein (EPOP) and Jumonji and ATrich interaction domain containing 2 (JARID2), contain DNA binding domains which are able to recruit PRC2 to specific genomic loci and allow the adaption of PRC2 to altered cellular states (Conway et al., 2018; Huang et al., 2023). PRC2.1 is defined by the presence of one of the three PCLs (PCL1/2 PCL3) and either EPOP or PALI1/2. The C-terminal chromo-like region of PCL can directly interact with PRC2.1 core subunit SUZ12. The N-terminal domain of PCL targets the binding of PRC2.1 to specific chromatin regions (Guo et al., 2021). Interestingly, it was found that PCL proteins can bind to unmethylated DNA. The binding of the winged-helix domain of PCL1 to DNA contributes to the prolonged residence time of PCL1-PRC2 on chromatin and thereby facilitates H3K27 methylation (Choi et al., 2017). And the Tudor domains of PCL1-3 have high affinity to chromatin regions modified by H3K36me3 and give rise to the DNA-driven PRC2.1 recruitment to new target genes (Ballaré et al., 2012; Musselman et al., 2012). PCL2 can recognize and bind the regions with a high density of unmethylated CpGs in a relatively unwound helix, which is required for the precise targeting and binding of PCL2-PRC2 to the developmental genes (Perino et al., 2018). Moreover, methylated PALI1 can bind to PRC2.1-core subunit EED to allosterically activate PRC2.1 and also facilitate its DNA binding (Zhang et al., 2021). Importantly, the binding of the C-terminal domain of EPOP with the ZnB-Zn domain of SUZ12 is required for stable interactions between EPOP and PRC2.1 (Guo et al., 2021).

Notably, PRC2.2 contains two zinc-finger-containing subunits, AEBP2 and JARID2 (Hauri et al., 2016; Kasinath et al., 2021). Both AEBP2 and JARID2 can directly bind to the ZnB-Zn domain of SUZ12 to stabilize the interactions of PRC2.2 with genomic targets (Chen et al., 2018). AEBP2 but not JARID2, competes with PCL3 to bind with the C2 domain of SUZ12 (Chen et al., 2018). JARID and EPOP share a binding domain of SUZ12 (Chen et al., 2018). Moreover, these binding modules contribute to the structural organizing of different classes of PRC2 holo complexes (Chen et al., 2018). Similar to PALI1, methylated JARID2 can also allosterically activate PRC2.2 by directly interacting with EED protein (Sanulli et al., 2015). JARID2 regulates the recruitment and activation of PRC2 by recognizing and binding H2AK119ub1-containing nucleosomes. Interestingly, cryo-electron microscopy (cryo-EM) results indicated that JARID2 and AEBP2 joint can localize to CpG-rich promoter regions with active transcription markers (H3K4me3 and H3K36me3) (Kasinath et al., 2021). JARID2 additionally contains a RNA-binding region, and the binding of JARID2 with noncoding RNAs (ncRNAs) facilitates the JARID-PRC2 interactions, which promotes the recruitment of PRC2 to chromatin (Kaneko et al., 2014) (Figure 1).

PcG proteins are mainly present in repressed genes which indicates their role in gene silencing. KDM2B-PRC1.1 and PCL-PRC2.1 can bind to the CGIs of silenced gene promoters and facilitate the formation of Polycomb chromatin domains which were characterized by the occupancy of H3K27me3, H2AK119ub1 as well as polycomb proteins. The interplay between PRC1 and PRC2 is necessary for the formation of polycomb domains (Dobrinić et al., 2021). In detail, H2AK119ub1 catalyzed by PRC1 can be recognized by JARID2 which recruits PRC2.2 and promotes the deposition of H3K27me3. Similarly, CBX-PRC1 can bind to the H3K27me3 and stimulate PRC1 enzymatic activity. RYBP/YAF-PRC1 and EED-PRC2 can recognize and bind to H2AK119ub1 and H3K27me3 respectively, which further promote the propagation of Polycomb domains. Furthermore, it is found that PRC2 occupancy at polycomb domains is primarily relied on the PRC1/H2AK119ub1 in mouse embryonic stem cells (Dobrinić et al., 2021). Mechanically, the occupancy of H2AK119ub1 and H3K27me3 interferes with the recruitment and activity of RNA Polymerase II (RNA Pol II), therefore largely prevents the transcription elongation by RNA Pol II. For bivalent genes with both H3K27me3 and H3K4me3, PcG proteins can hold poised RNA Pol II over the transcription start site (Flora et al., 2021). Moreover, PcG complexes can generate compacted chromatin structure, which is mainly independent on its catalytic activity. The compact state of chromatin prevents the binding of chromatin remodeling complexes such as the SWI/SNF complex, and therefore leads to transcriptional silencing. Polymerization of PRC1 which relies on SAM domains of PHC proteins is essential for polycomb-mediated chromatin structure organization (Flora et al., 2021).

PRC1 complexes also contribute to gene activation. For instance, ncPRC1.1 subunits such as PCGF1 and KDM2B, are co-localized on actively transcribed genes (van den Boom et al., 2016). ChIP-seq data revealed the presence of a certain level of H2AK119ub1 and the absence of H3K27me3 in ncPRC1.1-bound active loci. These observations suggested that the transcriptional active role of PRC1 is independent on PRC2. Moreover, no H2AK119ub1 was detected around these loci in some cell types. In neuronal cells, AUTS2 can directly interact with histone acetyltransferase EP300 through its HX repeat domain, thus endowing ncPRC1.3/5 with the ability of transcriptional activation (Gao et al., 2014a; Castanza et al., 2021; Pauli et al., 2021). Meanwhile, AUTS2 recruits Casein kinase 2 (CK2) to phosphorylate and inactivate RING1B. As a consequence, the H2A ubiquitination activity of PRC1 was blocked (Barbour et al., 2020). A recent study further indicated that Nuclear respiratory factor 1 (NRF1) is essential for AUTS2-ncPRC1-mediated gene activation by guiding its binding to specific locus (Liu et al., 2021). There must be other factors like NRF1 that contribute to the PRC1-dependent gene activation. Therefore, how PRC1 is involved in gene activation needs to be further explored.

XCI is a developmental process in that one of the two X chromosomes becomes silent in female cells to equalize the dosage imbalance of X-chromosome-linked genes between XY males and XX females. Mechanically, the initiation of XCI is dependent on the binding of the future inactive X chromosome (Xi) with the X-inactive specific transcript (Xist) which is a long noncoding RNA expressed on this chromosome in female cells (Borsani et al., 1991; Brockdorff et al., 1991; Loda et al., 2017). The Xist coated on Xi can recruit various complexes and allow them to bind and spread across the chromosome directly or indirectly (Markaki et al., 2021). Meanwhile, the chromatin structure of Xi is changed to a compacted state with the accumulation of DNA methylation, loss of active histone markers (like methylation of H3K4 and pan-acylation) and the deposition of repressive histone markers (like methylation of H3K9) (Keniry et al., 2016; Li et al., 2022). Furthermore, there is an accumulation of PRC1-associated H2AK119Ub and PRC2-associated H3K27me3 during this process (Brockdorff, 2017). Using female mouse embryonic stem cells, Żylicz et al. (2019) found that the enrichment of H2AK119Ub is prior to H3K27me3, and the marked regions of them are largely overlapped on the X chromosome. Interestingly, both of them are firstly deposited at the intergenic regions around the Xist RNA entry sites which are marked by PcGs prior to Xist upregulation. It seems that the initial chromatin landscape of the X chromosome instructs the alteration of the chromatin structure, which leads to the specific pattern of Xist spreading and transcriptional silencing (Żylicz et al., 2019). Subsequently, PcGs spread into gene bodies after deacetylation and gene silencing occurrence, which indicates that propagation of PcG-dependent markers may be not the trigger for gene repression during the XCI process (Żylicz et al., 2019). While the role of the PcG complex in XCI coordination during embryonic development in vivo is largely unexplored.

In mice, a subset of imprinted genes is controlled by PRC2-mediated H3K27me3, which is termed noncanonical imprinting (Inoue, 2023). And the H3K27me3-dependant noncanonical imprinting is inherited from oocytes. Indeed, H3K27me3-mediated imprinting regulates the repression of expression of maternal Xist, which is responsible for protecting maternal X (Xm) from being silenced in imprinted XCI (Chen and Zhang, 2020). In autosomes, maternal H3K27me3 also contributes to maternal-specific silence of imprinted genes. After implantation, DNA methylation compensates H3K27me3 to maintain the imprinting state of these genes in the embryonic lineage. However, H3K27me3-mediated imprinting can be maintained in some genes important for placenta development in extraembryonic lineages (Chen and Zhang, 2020). While such maternal-H3K27me3 dependent noncanonical imprinting is not conserved in humans (Chen and Zhang, 2020).

Xist is a maternal imprinted gene in mice. It was found that H3K27me3, but not DNA methylation, is allelic specifically deposited at the maternal Xist locus and contributes to its imprinting state in oocytes and early embryos (Kobayashi et al., 2012; Zheng et al., 2016; Inoue et al., 2017a; 2017b). H3K27me3-dependent imprinting of maternal Xist is responsible for the safeguard of the maternal X chromosome from XCI (Inoue et al., 2017b). Similar to what was found in embryos injected with Kdm6b at the Zygotic stage, loss of the H3K27me3 domain at the Xist locus led to the reactivation of maternal Xist and maternal XCI in maternal Eed knockout morula embryos (Inoue et al., 2017b). Aberrant XCI can be largely restored by E4.0 in both female and male maternal Eed knockout embryos, which is consistent with the results that some embryos survive to term. Specially, in female embryos, XCI occurs in a random manner in extraembryonic cells, with variable parental biases of X-linked gene expression (Inoue et al., 2018; Harris et al., 2019). Further studies are needed to clarify the mechanism underlying the conversion from Xi/Xi to Xa/Xi in Eed maternal KO ExEs (Inoue et al., 2018; Harris et al., 2019). And it remains to be determined whether Xist or autosomal H3K27me3 imprinting loss plays a major role in the male-biased lethality of EED maternal deficiency embryos. Zygotic Eed transcription was carried out at the 4-cell stage in mouse embryos. Different from maternal EED, the absence of zygotic EED had limited impacts on the initiation and establishment of imprinted X-inactivation but resulted in the downregulation of a subset of X-linked gens (Harris et al., 2019).

In mice, the XCI occurs with two waves. The first XCI wave occurs shortly after fertilization with a parent-of-origin bias, a systematic inactivation of the paternal X chromosome, which is called imprinted XCI (Mak et al., 2004; Okamoto et al., 2004; Ravid Lustig et al., 2023). And maternal H3K27me3 inherited from oocytes is required for the genomic imprinting of Xist during XCI (Inoue et al., 2017b). At the late blastocyst stage, this form of X-chromosome inactivation is reversed in cells from the inner cell mass (ICM), a process known as X-chromosome reactivation (XCR) (Patrat et al., 2009; Min et al., 2017). While the established XCI in extraembryonic lineage is maintained during the subsequent developmental stage. A random XCI (rXCI) is initiated again in embryonic lineage cells around the time of implantation. And once occurs, the status of X-inactivation will be inherited by all the progeny cells (Pacini et al., 2021).

Recently, Masui et al. (2023) induced PRCs deletion after the establishment of imprinted XCI (Ring1a/b deletion at E5.5 or Eed deletion at E3.5), and especially focused on the role of PRC1 and PRC2 in the maintenance of XCI (Borensztein et al., 2017). A disrupted suppression of Xist-linked genes was observed in extraembryonic lineage at E7.5, which demonstrated that PRC1/2 has a substantial impact on the maintenance of imprinted XCI in extraembryonic lineages. Notably, PRC1 and PRC2 seem to function independently since the loss of PRC1 or PRC2 did not affect the accumulation of H3K27me3 or H2AK119ub in both extraembryonic lineages and embryo at this stage (Masui et al., 2023). While previous study reported that the knockout of Pcgf3/5 gene impaired the deposition of both H2AK119ub1 and H3K27me3 on Xi in mouse embryonic stem cells (Almeida et al., 2017). One possible reason for this discrepancy is that PRC1 may no longer be required for the propagation of PRC2 on Xi in all lineages during XCI maintenance phase (Almeida et al., 2017). In addition, the H3K27me3 accumulated on Xi during the initiation phase of XCI may sufficiently contribute to the subsequential recruitment and spread of PRC2 during the maintenance stage.

Allele-specific RNA seq revealed a partial overlap between PRC1- and PRC2-dependent X-linked genes, confirming there is a synergy between PRC1 and PRC2. More genes were sensitive to PRC1 depletion than PRC2 depletion, which implies that PRC1 is the main player in the maintenance of XCI (Masui et al., 2023). Furthermore, these PRC1-sensitive genes have a CGI at their promoter regions (Masui et al., 2023). In line with this notion, Andergassen et al. generated the PRC1 or PRC2 depleted zygotes and found that PRC2 but not PRC1 is dominant in XCI of extraembryonic tissues (Andergassen et al., 2021). The predominant impact of PRC2 on the initiation of XCI may contribute to this debate. Further studies are needed to explore the mechanisms underlying how PcGs are involved in the initiation and maintenance of XCI in mammalian embryos.

Interestingly, inconsistent with what was observed in extraembryonic tissues, the transcriptional silencing of X-linked genes still can be observed in PRC1-null or PRC2-null embryonic lineages at E7.5 (Masui et al., 2023). These results indicated that PRCs are redundant for random XCI in embryonic lineages and there must be other dominant mechanisms of XCI in embryo but not extraembryonic tissues, like DNA methylation or H3K9me3.

PCGF1-PRC1 is essential for the balanced output of hematopoietic stem and progenitor cells (HSPCs). Pcgf1-deletion caused myeloid-biased differentiation of HSPCs, mainly due to the derepression of C/EBPα and Hox family genes (Takano et al., 2022; Nakajima-Takagi et al., 2023). Lymphoid differentiation was suppressed in Pcgf1-deficient cells. As a result, deletion of Pcgf1 in hematopoietic cells led to mild anemia and leukopenia in mice (Takano et al., 2022; Nakajima-Takagi et al., 2023). Mice with a functional insufficiency BCOR that failed to interact with PCGF1 also display myeloid-biased differentiation (Tara et al., 2018). Upregulation of Cebp and Hox genes was also observed in BCOR insufficient hematopoietic cells (Tara et al., 2018). KDM2B also governs the self-renewal capacity of HSCs. KDM2B deficiency caused a significant reduction of HSPCs and compromised lymphoid specification (Andricovich et al., 2016). These results indicate that ncPRC1.1 is essential for definitive hematopoiesis and lineage commitment of HSPCs. However, PCGF4 was essential for self-renewal capacity and multipotency of HSCs through repression of the expression of cell cycle regulator INK4A/ARFPAX5 and B cell lineage developmental regulator EBF1 and PAX5 (van der Lugt et al., 1994; Oguro et al., 2010).

PRC2 also plays a pivotal role in hematopoietic development. EED is required for normal hematopoiesis during the postnatal period. Conditional knockout of Eed in mouse hematopoietic cells by VavCre had no visible impact on pups at birth but resulted in severe leukopenia, anemia and early lethality shortly after birth. Impaired differentiation of neonatal bone marrow (BM) hematopoiesis may account for the decreased matured blood cells (Xie et al., 2014). For adult BM HSC, loss of Eed led to HSC exhaustion which indicates EED is required for the maintenance of adult BM HSCS (Xie et al., 2014). For fetal hematopoiesis, EED loss via VavCre did not affect the development of fetal liver (FL) hematopoietic stem cells (HSCs) (Xie et al., 2014). Controversially, EED deletion resulted in disrupted HSC homeostasis and postimplantation lethality at mid-gestation are found in Tie2Cre (EED^CKO) embryos (Yu et al., 2017). The earlier onset of Tie2Cre deletion in hemangioblasts may partially account for these different results (Yu et al., 2017).

Unlike EED, EZH2 is dispensable for the maintenance of both FL and BM HSCs (VavCre-mediated Ezh2 excision) (Xie et al., 2014). Conversely, TieCre-mediated EZH2 deletion resulted in largely reduced FL HSCs and embryonic lethality at the mid-gestation stage, which indicates the indispensable role of EZH2 in fetal hematopoiesis (Mochizuki-Kashio et al., 2011). It seems that EZH1 partially compensated for the deficiency of EZH2 in BM hematopoiesis, but not in fetal liver (Mochizuki-Kashio et al., 2011; Xie et al., 2014). Further, the compromised vascular integrity observed in EZH2^CKO embryos was lacking in EED^CKO embryos. EZH2 may play an important role in vascular development through EED-independent non-canonical PRC2 functions. In addition to hematopoiesis, EZH2 was reported to be required for postnatal cardiac homeostasis (Delgado-Olguín et al., 2012). Deletion of Ezh2 in cardiac progenitors impaired postnatal cardiomyocyte differentiation and proliferation and eventually led to myocardial hypertrophy and fibrosis after birth (Delgado-Olguín et al., 2012).

SUZ12 is also required for both fetal hematopoiesis and adult HSC maintenance. SUZ12^CKO mice generated by VavCre died before weaning. It seems that SUZ12 may regulate fetal hematopoiesis through PRC2-independent actions. This hypothesis is further supported by the existence of a noncanonical subcomplex which contains EZH1 and SUZ12 but lacks EED (Xu et al., 2015). Moreover, lymphoid development was largely compromised in mice with a lymphoid-specific deletion of Suz12, displaying lymphopenia and significantly reduced spleen and thymus cellularity (Lee et al., 2015).

Unlike the core units of PRC2 expressed abundantly in all tissues, the accessory proteins are only expressed in certain tissues (Rothberg et al., 2018). The non-core units of PRC2.1-PCL2 are also found to play an important role in definitive erythroid development. The embryos lacking PCL2 were died by E15.5, displaying growth defects and anemia (Rothberg et al., 2018). Pcl2 knockout resulted in a significantly decreased core PRC2 proteins level and a global loss of promoter H3K27me3. Mechanically, loss of PCL2-mediated H3K27me3 abnormally activated the Wnt/β-catenin signaling pathway, which resulted in impaired maturation and differentiation of erythroid (Rothberg et al., 2018) (Table 2).

PcGs also play an essential role in the regulation of mammalian neuronal development (Desai and Pethe, 2020). For instance, Pcgf4-knockout in mice also caused neurological abnormalities, like ataxic gait and sporadic seizures (van der Lugt et al., 1994). Similarly, the Auts2-knockout induced abnormality in nervous system (Hori et al., 2015). And Auts2-deletion especially in the mouse central nervous system caused a phenotype similar to that of AUTS2 syndrome described in humans (Gao et al., 2014b). Contrary to the typical role of PRC1 in gene repression, AUTS2-PRC1 acts as a transcriptional activator in neuronal cells, through its recruitment of CK2 and interaction with EP300 (Gao et al., 2014b). Notably, NRF1 was required for AUTS2-PRC1 recruitment to target sites (Liu et al., 2021).

Some subunits of PRC2 are also involved in neuronal differentiation and proliferation. It was reported that EZH2 plays an important role in the fate transition of both cortical progenitor cells in the cerebral cortex and GABAergic neurons in cerebellum (Pereira et al., 2010; Feng et al., 2016). Loss of EZH2 significantly altered the timing of cortical development. Owing to the removal of H3K27me3 which was caused by EZH2 deletion, the balance between self-renewal and differentiation of cortical progenitor cells was destroyed and turned to differentiation (Pereira et al., 2010). Cerebellar-specific deletion of EZH2 led to an increase in cerebellar interneurons and a reduction in Purkinje cells and granule precursor cells in the embryonic cerebellum, and ultimately led to cerebellar hypoplasia (Feng et al., 2016). Moreover, PRC1 subunit Pcgf4 was also essential for cerebellar development and contributes to the expansion of granule precursor cells (Leung et al., 2004). EED regulated neuronal differentiation and proliferation of neural stem/progenitor cells. Conditional knockout of Eed in the brain led to postnatal lethality, with impaired neural differentiation and proliferation and malformation of the dentate gyrus (Liu et al., 2019). Overexpression of SOX11, the downstream target of EED, can rescue EED-ablation-induced neuronal differentiation defect. EED/PRC2-driven H3K27me1 deposition was indicated to be required for transcriptional activation of Sox11 (Liu et al., 2019). Importantly, EED mutation in humans was related to the occurrence of Weaver syndrome which is characterized by intellectual disability (Cooney et al., 2017). Furthermore, brain malformations were also observed in mice with a heterozygous mutation of the Suz12 gene (Miró et al., 2009).