EZH2 variants differentially regulate polycomb repressive complex 2 in histone methylation and cell differentiation
Polycomb repressive complex 2 (PRC2) is responsible for establishing and maintaining histone H3K27 methylation during cell differentiation and proliferation. H3K27 can be mono-, di-, or trimethylated, resulting in differential gene regulation. However, it remains unknown how PRC2 specifies the degree and biological effects of H3K27 methylation within a given cellular context. One way to determine PRC2 specificity may be through alternative splicing of Ezh2, PRC2’s catalytic subunit, during cell differentiation and tissue maturation.
We fully characterized the alternative splicing of Ezh2 in somatic cells and male germ cells and found that Ezh’s exon 14 was differentially regulated during mitosis and meiosis. The Ezh2 isoform containing exon 14 (ex14-Ezh2) is upregulated during cell cycle progression, consistent with a role in maintaining H3K27 methylation during chromatin replication. In contrast, the isoform lacking exon 14 (ex14D-Ezh2) was almost exclusively present in spermatocytes when new H3K27me2 is established during meiotic differentiation. Moreover, Ezh2’s transcript is normally controlled by E2F transcription activators, but in spermatocytes, Ezh2’s transcription is controlled by the meiotic regulator MYBL1. Compared to ex14-EZH2, ex14D-EZH2 has a diminished efficiency for catalyzing H3K27me3 and promotes embryonic stem cell differentiation.
Ezh2’s expression is regulated at transcriptional and post-transcriptional levels in a cellular context-dependent manner. EZH2 variants determine functional specificity of PRC2 in histone methylation during cell proliferation and differentiation.
KeywordsPolycomb-group proteins EZH2 Spermatogenesis Alternative splicing Histone methylation Cell differentiation
Polycomb repressive complex 2 (PRC2) catalyzes mono-, di-, and trimethylation of histone H3 lysine 27 (H3K27) and functions as a key epigenetic regulator for normal development and diseases. The three forms of H3K27 methylation are mutually exclusive and form spatially defined genomic domains. Monomethylated H3K27 (H3K27me1) is enriched in the gene bodies of active genes, indicative of its involvement in gene activation [1, 2, 3, 4]. In contrast, dimethylated H3K27 (H3K27me2) primarily resides within large intergenic and intragenic chromatin domains (50–70% of H3 tails) [5, 6], possibly exerting a protective role by preventing non-cell-type-specific enhancer use . Lastly, trimethylated H3K27 (H3K27me3) is widely present in the promoter regions of silenced genes and thought to provide PRC2 with a role in transcriptional repression [7, 8]. PRC2 catalyzes H3K27me1 and H3K27me2 relatively efficiently compared to H3K27me3 . During cell cycle progression, H3K27me1 and me2 are rapidly incorporated into chromatin after DNA synthesis, but the restoration of normal H3K27me3 levels is delayed [10, 11, 12]. In light of PRC2’s versatility in histone methylation and gene regulation, it is important to understand how this complex is spatiotemporally orchestrated to load the appropriate methyl groups onto targeted genomic loci during cell proliferation and differentiation.
The PRC2 complex is composed of core subunits (EED, EZH1/2, SUZ12, and RBBP4/7) required for enzymatic activity, and accessory subunits (PCLs, JARID2, AEBP2) that regulate the recruitment and activity of PRC2 . Recent studies suggest that accessory PRC2 components regulate DNA binding, methyltransferase activity and the spread of methylation marks on the genome. The selective incorporation of accessory subunits leads to the formation of PRC2 subcomplexes, PRC2.1, characterized by the incorporation of AEBP2 and JARID2, and PRC2.2, characterized by mutually exclusive binding of one of the three polycomb-like homologs (PCLs) PHF1, PHF19, and MTF2 . However, these PRC2 variants only exhibit subtle differences in histone methylation and do not explain the locus-specific H3K27 methylation across the whole genome. In contrast, with the exception of EED, little is known about the roles that the core subunits may play in determining PRC2’s flexibility in histone methylation. EED undergoes alternative translation to yield four protein products; however, this diversity does not control the number of methyl groups added to H3K27 since each isoform is sufficient to generate all three H3K27 methylation marks .
PRC2 plays two major roles in histone methylation: it can maintain existing H3K27 methylation marks so that they are propagated to daughter cells and it can establish new H3K27 methylation sites. During cell proliferation, the E2F family of cell cycle transcription factors enhances the expression of Ezh2, which duplicates H3K27 methylation during cell proliferation . During cell differentiation, PRC2 establishes new H3K27 methylation sites, especially in male germ cells. These new H3K27 methylation marks are introduced into the genome to determine the cell-type-specific transcriptome . In this case, PRC2’s recruitment to specific loci for methylation appears to be a more complicated process, potentially involving noncoding RNAs, sequence-specific transcription factors, and/or PRC2-interacting proteins with affinity for CpG-rich DNA elements . Our previous results demonstrate that EED, EZH2, and SUZ12 are dramatically upregulated in pachytene spermatocytes , suggesting the germ cell-specific PRC2 complex has a role in establishing H3K27 methylation during meiotic progression. Therefore, it is crucial to understand how PRC2 can distinguish these two different types of histone methylation coupled with cell proliferation and differentiation.
EZH2 may have a major role in determining PRC2’s differential methylation roles. It possesses multiple interaction domains for EED and SUZ12, facilitating the methyltransferase activity conveyed by its SET domain [18, 19, 20]. EZH1, a homolog of EZH2 encoded by a separate locus , has much less methyltransferase activity and cannot substitute for EZH2 in histone methylation and related biological functions in many tissues . Because other PRC2 subunits only have a subtle effect on EZH2 methyltransferase specificity, we speculate that EZH2’s variants themselves diversify PRC2’s functional roles in distinct methylation processes during cell proliferation and differentiation.
Here, we identified multiple Ezh2 isoforms derived from alternative transcriptional splicing in various tissues and cell types. Expressions of EZH2 variants that include or exclude exon 14 are differentially regulated via cell cycle or meiotic regulators, respectively, during mitosis and meiosis. The EZH2 isoform without exon 14 (ex14D-EZH2) has a disrupted CXC domain and is the primary isoform found in spermatocytes. This isoform is responsible for the establishment of H3K27me2, but is less efficient at catalyzing H3K27me3. Moreover, exclusive expression of ex14D-EZH2 in ES cells promotes their differentiation, indicated by precocious and enhanced expression of mesoderm genes. In contrast, the EZH2 isoform with exon 14 (ex14-EZH2) is the most common isoform in proliferating cells and more efficient at catalyzing H3K27me3. Our study suggests that the incorporation of specific EZH2 variants into the PRC2 complex controls the appropriate level and extent of H3K27 methylation in polycomb target loci during the establishment and maintenance of these epigenetic marks.
Ezh2 pre-mRNA splicing is differentially regulated during meiosis and mitosis
First, we examined Ezh2 and Ezh1 transcript levels during spermatogenesis. Ezh2 transcripts without exon 14 (ex14D-Ezh2) were at a relatively low level in mitotic germ cell populations such as progenitors (E13.5 and E15.5) and spermatogonial stem cells (p5). However, this isoform was dramatically upregulated at the time of entry into meiosis (p10) and maintained at high levels during the meiotic progression as shown in p19, 1-month and 4-month testes. In contrast, the Ezh2 transcripts containing exon 14 (ex14-Ezh2) were downregulated during meiosis (Fig. 1b). The dynamics of exon 14-related alternative transcripts were verified by qPCR analysis (Fig. 1c). Transcription of PRC2’s core components, Eed, Suz12, and Rbbp4, also increased during germ cell maturation (Fig. 1d), suggesting ex14D-EZH2 functions within the PRC2 complex during meiotic differentiation. In contrast, Ezh1 levels were consistent throughout germ cell development (Fig. 1c), indicating its expression is independent of meiotic differentiation.
Because ex14-Ezh2 is abundant in mitotic germ cells and rapidly dividing ES cells and primary MEFs (Fig. 1b), we wanted to determine the dynamics of Ezh2 and Ezh1 transcripts during mitosis. Thus, we synchronized primary MEFs at the G0/G1 phase by serum starvation and then released them into the S and G2/M phases with serum supplementation. In comparison with meiosis, ex14-Ezh2, but not ex14D-Ezh2, was remarkably elevated during mitotic progression (Fig. 1e). The Ezh2 transcripts with a full exon 3 (ex3-Ezh2) or a truncated exon 3 (ex3T-Ezh2) were both upregulated. EZH2, EED, and SUZ12 protein levels also increased in mitotic cells (Fig. 1f). Ezh1 expression decreased with the cell cycle activation, which is consistent with the high expression of Ezh1 in mature tissue but low in proliferating tissues . These results indicate that Ezh variants are differentially regulated during meiosis and mitosis.
ex14D-Ezh2 transcription is independent of E2F regulation and responsible for establishing H3K27me2 in spermatocytes
Meiotic ex14D-Ezh2 expression occurs at a time when new histone marks are established during cell differentiation, suggesting that it may have a role in this process. Thus, we characterized H3K27 methylation patterns in different populations of spermatocytes during meiosis. There were no significant differences in H3K27me1 and H3K27me3 levels between early stage (leptonema) and late stage (pachynema) spermatocytes. However, H3K27me2 levels were strikingly elevated during the zygonema to pachynema transition (Fig. 2g), which is consistent with the upregulation of EZH2 in pachytene spermatocytes  and with ex14D-Ezh2 transcripts in pachynema-enriched p19 testes (Fig. 1b, c). These results suggest that ex14D-Ezh2 transcription and processing occur in a cellular context-dependent manner for establishing H3K27me2.
MYBL1 activates ex14D-Ezh2 expression in spermatocytes
Because ex1b encodes the 5′ UTR of ex14D-Ezh2, we performed motif analysis with HOMER to search for transcription factors that could bind near it. This analysis predicted that MYBL1, a master regulator of male meiosis , had a DNA binding site upstream of ex1b, and binding was confirmed by MYBL1 ChIP-seq analysis (Fig. 3a) . Consistent with Ezh2 expression in meiosis, MYBL1 was at a low level in leptotene and zygotene, and significantly elevated in pachytene and diplotene spermatocytes (Fig. 3c). MYBL1’s role as a transcriptional activator of Ezh2 in germ cells is further supported by the evidence that a Mybl1 knockout in testis caused the downregulation of Ezh2 transcription . MYBL1 was also enriched in the promoter regions of Eed and Suz12 (Additional file 2: Fig. S2c, d), indicative of its functioning as a PRC2 regulator during spermatogenesis. Thus, all of these data suggest that MYBL1 activates ex14D-Ezh2 expression from ex1b in spermatocytes.
ex14D-EZH2 has preferential activity toward H3K27me2 but not H3K27me3 and promotes ES cell differentiation
PRC2 can establish new H3K27me3 marks for reprogramming the transcriptome during cell lineage specification. De Novo trimethylation of H3K27 occurs during retinol acid-induced ES cell differentiation . In RA-treated ES cells, we found that ex14D-EZH2 introduced new H3K27me3 marks to Fgf4’s promoter and intron, Utf1, and Nphs1, but at reduced levels (Fig. 4g). In contrast, Ezh2 knockout cells failed to establish those marks during ES cell differentiation, suggesting that EZH1 cannot establish new H3K27me3 marks even though it can partially rescue the maintenance of H3K27me3 during cell proliferation as shown in Fig. 4f.
Because EZH2 is required for mouse ES cell differentiation , we tested if ex14D-Ezh2 had a specific role in this process. Withdrawal of leukemia inhibitory factor (LIF) induces differentiation of mouse ESCs into various embryonic and extraembryonic lineages . To understand to what degree ex14D-EZH2 contributes to ES cell differentiation, we measured the expression of mesoderm marker genes activated by LIF withdrawal in Ezh2KO, ex14D-Ezh2, and control cells. As expected, in Ezh2 knockout cells, Brachyury, Eomes, and Mixl1 failed to activate. In contrast, Brachyury, Eomes, and Mixl1 were all upregulated in ex14D-Ezh2, and control cells. Intriguingly, exclusive expression of ex14D-Ezh2 led to precocious expression of these marker genes at higher levels (Fig. 4h). In conclusion, these data suggest that ex14D-Ezh2 promotes ES cell differentiation.
EZH1 and EZH2 variants have differential activity in catalyzing trimethylation of H3K27
Since EZH variants have different abilities with respect to H3K27 methylation, we asked if the physical interactions between EZH variants and core subunits or accessory subunits regulate their catalytic specificity and activity. EZH2 in ES cells and ex14D-EZH2 in spermatocytes can interact with SUZ12 and EED (Additional file 4: Fig. S4a). In HEK293T cells cotransfected with EZH variants and PRC2 subunits, we found that PHF1, PHF19, JARID2, and RBBP4 were part of the protein complexes containing either ex14D- or ex14-Ezh2 (Additional file 4: Fig. S4b, c). These results suggest that EZH variants do not selectively exclude these subunits when assembled into PRC2. Taken together, the alteration of EZH2’s CXC domain determines its SET domain capacity in H3K27me3.
For most histone methylation processes, such as H3K9me and H3K4me, multiple methyltransferases participate in loading a specific number of methyl groups onto histones, resulting in gene regulation . However, the three forms of H3K27 methylation are mostly executed through EZH2-PRC2. H3K27me1 and H3K27me2 are found throughout the chromosomes, but without the coexistence of PRC2 by ChIP-seq analysis . In contrast, chromatin domains marked by H3K27me3 are also abundant in PRC2 binding . This suggests that PRC2 uses differential mechanisms to control the methylation reaction on its targeting loci. Although other PRC2 subunits either regulate or are essential for EZH2’s enzymatic activity, they do not determine the level of methylation (i.e., H3K27me1, me2 and me3). EZH1 and EZH2 variants contain multiple domains that mediate physical interactions with most of PRC2’s subunits and ncRNA . We hypothesize that the diversity in EZH1 and EZH2 proteins regulate H3K27 methylation and establish and maintain these epigenetic marks. By examining the dynamics of Ezh2 transcripts during cell proliferation and differentiation, we showed that EZH2 variants with and without exon 14 mechanistically regulated PRC2’s ability to load specific numbers of methyl groups onto histones.
Ezh2’s exon 14 encodes a CXC domain which precedes the SET catalytic domain and is characterized by three C-X(6)-C-X(3)-C-X-C motifs [31, 32]. Mutation of the cysteine in CXC domains impaired the methyltransferase activity in PRC2  and overexpression of the protein interacting with a CXC domain in EZH2 can enhance H3K27 methylation . Consistent with these two studies, ex14D-EZH2, which lacks two C-X(6)-C-X(3)-C-X-C motifs in the CXC domain, exhibited diminished catalytic activity in H3K27me3. But this disruption of the CXC domain did not affect PRC2’s ability to mono- and di-methylate H3K27. We also found that nearly exclusive expression of ex14D-Ezh2 is coincident to acquiring H3K27me2 in spermatocytes. This led us to speculate that Ezh2’s alternative splicing serves as a switch to control the methylation status on modified nucleosomes. Because EZH2 can convert H3K27me2 to H3K27me3 [35, 36], ex14D-EZH2 in spermatocytes may help to maintain the H3K27me2 marks which fail to be trimethylated at genomic loci where silencing normally involves H3K27me3-independent mechanisms. Previous studies showed that the H3K27me3 marks accumulated on key developmental genes were not replaced by protamines in sperm [37, 38] and could be transmitted to the next generation for regulating embryonic gene expression . Thus, it is important to maintain the genomic distribution of H3K27 methylation states in germ cells for epigenetic inheritance.
EZH1 was thought to compensate for the absence of EZH2 in H3K27 methylation. But in our study, EZH1 was unable to generate new H3K27me3 marks on target genes during cell differentiation. Also, overexpression of EZH1 did not rescue H3K27me3 in proliferating cells as ex14-EZH2 did, even though EZH1 and EZH2 are nearly identical in their CXC and SET domains. However, compared to EZH2, EZH1 does not have a cell cycle-regulated expression pattern, it shares less similarity in the N-terminal domains, and it promotes mRNA transcription during cellular differentiation . This indicates that the methylation activity of EZH1-PRC2 is regulated differently from EZH2-PRC2. In contrast to EZH2, EZH1 is abundant in highly differentiated tissues and low in actively proliferating cells. It will be interesting to explore EZH1’s cell-type-specific interacting partners, which facilitate the recruitment onto chromatin and the activation of its methyltransferase activity.
EZH2 level is a determinant for its appropriate biological functions. Deficiency in EZH2 caused cell cycle arrest, senescence, and differentiation defects . EZH2 overexpression has been reported to promote cell proliferation and neoplastic transformation . Our findings show that EZH2 expression is finely controlled at both transcriptional and post-transcriptional levels for producing suitable doses and isoforms of EZH2. To regulate transcription, cell-type-specific TFs and cell cycle-related TFs cooperate to adjust EZH2 expression levels according to the cell’s differentiation and proliferation status. The levels of ex14D-EZH2 for generating new H3K27me in spermatocytes are much lower than EZH2 for duplicating H3K27 methylation in highly proliferating ES cells (Additional file 4: Fig. S4d). The relatively low levels of ex14D-EZH2 in differentiating cells could serve as a mechanism for recovery of H3K27 methylation due to histone turnover occurring at a low rate  and, meanwhile, introducing new marks at specific genomic loci. For pre-mRNA processing, each alternative splicing event is controlled by multiple RNA binding proteins (RBPs), the combined action of which creates a distribution of alternatively spliced products in a given cell type . In renal cancer cells, the splicing factor SF3B3 stimulates the inclusion of exon 14 in Ezh2 transcripts . During spermatogenesis, the splicing landscape is globally reprogrammed in the mitotic-to-meiotic transition during the germ cell cycle . We reason that spermatocytes express a meiosis-specific array of RBPs, which give rise to the skewed production of ex14D-Ezh2 transcripts.
We revealed that Ezh2’s expression was differentially regulated at transcriptional and post-transcriptional levels during mitosis and meiosis. EZH2 isoforms have different catalytic activity in mono-, di-, and trimethylation of H3K27. The incorporation of specific EZH2 variants into the PRC2 complex regulates the fidelity in the establishment and maintenance of these epigenetic marks across the genome.
Materials and methods
RT-PCR and qPCR
Total RNA was extracted from whole testis or isolated spermatocytes using TRIzol reagent (Invitrogen) followed by an RNA cleanup step and on-column DNA digestion using Direct-zol™ RNA MiniPrep Kit (Zymo Research) according to the manufacturers’ instructions. cDNA was synthesized using ProtoScript II Reverse Transcriptase (NEB). Real-time qPCR analysis was performed using SsoFast EvaGreen supermix (Bio-Rad) and CFX96 thermocycler (Bio-Rad) (Additional file 5: Table S1).
The ChIP-seq was performed as described . Five million mouse ES cells and 2.5 μg of H3K27me3 antibody (ab6002; Abcam) or 5 μl of EZH2 antibody (#5246, cell signaling) were used per IP. ChIP-seq libraries were prepared according to Illumina instructions and sequenced on Illumina’s Genome Analyzer IIx or HiSeq2000 instrument. Sequence reads were aligned to genomic sequence with Bowtie . All mm10 genome annotations were obtained from the UCSC genome browser. HiddenDomains  and MACSv2  identified ChIP-seq enrichment using 200 bp bins. Peak annotation was created by ChIPseeker  using only the peaks with posterior probabilities > 0.9. Reads at promoters, defined as a 1 kb window centered on the 5′ most TSS, were counted using htseq-count .
Mouse embryonic fibroblasts (MEFs) were cultured as described previously . Primary MEFs were synchronized by serum starvation (0.2% FBS) for 2 days and then released into cell cycle by supplement with 10% FBS. The cells were harvested at the time points indicated for RNA and protein extraction. HEK293 were cultured in the same medium as MEFs and transfected with FLAG- or HA-tagged PRC2 subunits using calcium phosphate. The cells were collected 48 h post-transfection for protein extraction. Mouse E14 ES cells were cultured in Glasgow Minimum Essential Medium supplemented with 15% fetal bovine serum, 1.0 mM l-glutamine, 0.1 mM minimal essential medium-nonessential amino acids, 0.1 mM β-mercaptoethanol, and leukemia inhibitory factor. The differentiation of ES cells with retinol acid treatment was performed as previously described .
Protein extraction, histone extraction, immunoprecipitation, and Western blotting
Protein or histone was prepared from MEFs or ESCs for Western blot analysis as described . For immunoprecipitation assay, 200 μg protein extract from HEK293T cells was incubated with anti-FLAG (F3165, Sigma) or anti-HA (A190-108A, Bethyl Labs) conjugated protein A/G agarose beads (sc-2003, Santa Cruz Biotechnology). FLAG- or HA-tagged full-length coding fragments of EZH1, EZH2, PHF1, PHF19, JARID2, and RBBP4 were amplified from their cDNAs and cloned into an expression vector bearing CAG promoter.
Immunostaining on cryosections and mouse ES cells
The immunostaining on testis cryosections is as described . Ezh2KO and EzhDKO ES cells were transfected with FLAG-tagged EZH variants using Xfect™ mESC Transfection Reagent (631320, Clontech) according to manufacturer’s instruction. The cells were trypsinized and resuspended in ES cell medium 48 h post-transfection. ES cells were fixed on slides by adding three volumes of 2% PFA (containing 0.1% Triton X-100 and 100 mM sucrose in 1 × PBS) at 4 °C for 30 min. The slides were air-dried and stored at 4 °C for immunofluorescence assays.
Genome editing of Ezh variants by CRISPR-Cas9
sgRNAs targeting Ezh1 and Ezh2 were cloned into pX330-U6-Chimeric_BB-CBh-hSpCas9 (42230, Addgene) using Golden Gate assembly cloning strategy . The modification of Ezh genes in ES cells followed the procedure as described . Briefly, 5 × 104 E14 ES cells were cultured on 60 mm dishes for 1 day and then transfected with plasmids expressing Cas9 and sgRNAs, along with a plasmid expressing PGK-PuroR (Addgene, Cat. No. 31937) using FuGENE HD reagent (Promega) according to the manufacturer’s instructions. The cells were treated with 2 μg/ml puromycin for 2 days and recovered in normal culture medium until ES cell colonies grew up. Ezh1 targeted colonies were genotyped by PCR and verified by DNA sequencing. Ezh2 targeted colonies were screened by Western blot analysis and confirmed by DNA sequencing.
WM and TM conceived the experiments; WM and DY performed the experiments; WM, JS, and TM analyzed the data and wrote the manuscript. All authors read and approved the final manuscript.
We thank all members of the Magnuson laboratory for helpful comments on manuscript preparation.
The authors declare that they have no competing interests.
All mice were maintained at the University of North Carolina at Chapel Hill Animal Facility using standard techniques in accordance with protocols approved by the Institutional Animal Care and Use Committee.
Availability of data and materials
H3K27me3 and EZH2 ChIP-seq data are deposited to the Gene Expression Omnibus (GSE123174).
Consent for publication
Ethics approval and consent to participate
This work was supported by National Institutes of Health Grants RO1 GM101974 and U42 OD010924.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 9.McCabe MT, Graves AP, Ganji G, Diaz E, Halsey WS, Jiang Y, Smitheman KN, Ott HM, Pappalardi MB, Allen KE, et al. Mutation of A677 in histone methyltransferase EZH2 in human B-cell lymphoma promotes hypertrimethylation of histone H3 on lysine 27 (H3K27). Proc Natl Acad Sci USA. 2012;109(8):2989–94.CrossRefGoogle Scholar
- 36.Yap DB, Chu J, Berg T, Schapira M, Cheng SW, Moradian A, Morin RD, Mungall AJ, Meissner B, Boyle M, et al. Somatic mutations at EZH2 Y641 act dominantly through a mechanism of selectively altered PRC2 catalytic activity, to increase H3K27 trimethylation. Blood. 2011;117(8):2451–9.CrossRefGoogle Scholar
- 55.Bauer DE, Canver MC, Orkin SH. Generation of genomic deletions in mammalian cell lines via CRISPR/Cas9. J Vis Exp JoVE. 2015;95:e52118.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.