|Home | About | Journals | Submit | Contact Us | Français|
Histone H3 lysine 27 trimethylation (H3K27me3) catalyzed by the enzymatic subunit EZH2 in the Polycomb repressive complex 2 (PRC 2) is essential for cells to ‘memorize’ gene expression patterns through cell divisions and plays an important role in establishing and maintaining cell identity during development. However, how the epigenetic mark is inherited through cell generations remains poorly understood. Recently, we and others demonstrate that CDK1 and CDK2 phosphorylate EZH2 at threonine 350 (T350) and that T350 phosphorylation is important for the binding of EZH2 to PRC 2 recruiters, such as noncoding RNA s (ncRNAs) HOTA IR and XIST , and for the effective recruitment of PRC 2 to EZH2 target loci in cells. These findings imply that phosphorylation of EZH2 by CDK1 and CD K2 may provide cells a mechanism that enhances EZH2 function during S and G2 phases of the cell cycle, thereby ensuring K27me3 on de novo synthesized H3 incorporated in nascent nucleosomes before sister chromosomes are divided into two daughter cells. Additionally, a potential role of T350 phosphorylation of EZH2 in differing EZH2 from its homolog EZH1 in catalyzing H3K27me3 as well as the interplay between phosphorylation at T350 and other residues [e.g., phosphorylation by p38 at threonine 372 (T372)] in governing EZH2 activity in proliferating versus non-dividing cells are also discussed. Together, CDK phosphorylation of EZH2 at T350 may represent a key regulatory mechanism of EZH2 function that is essential for the maintenance of H3K27me3 marks through cell divisions.
Epigenetics was proposed by C.H. Waddington in the early 1940s.1 In contrast to genetics, epigenetics is the study of inherited changes in phenotype (appearance) or gene expression caused by mechanisms other than changes in the DNA sequence. Dependent on cell types and environment cues, epigenetic changes could be ‘memorized’ through cell divisions for the rest of a cell's life or may last only for a few generations.
Epigenetic, although they are heritable but do not involve mutations of the DNA itself. The molecular basis of epigenetics is complex. Two molecular mechanisms are commonly believed to mediate epigenetic phenomena: DNA methylations and histone modifications.2 One involves DNA modifications resulting in gene inactivation, but not the basic structure of DNA, and the other deals with changes in proteins associated with chromatin, which lead to either gene activation or inactivation.
Chromatin is constituted of DNA, histone and other proteins. Chromatin structure imposes significant obstacles on all aspects of DNA transcription mediated by RNA polymerase II.3 The dynamics of chromatin structure are tightly regulated by various mechanisms, which include histone modification, chromatin remodeling, histone variant incorporation and histone eviction.4
Histones are alkaline proteins found in eukaryotic cell nuclei. They regulate chromatin by organizing the DNA into structural units called nucleosomes. Histones are the major protein components of chromatin. They act as spools around which DNA winds and are essential for gene transcription regulation. Active genes have less bound histones while inactive genes are highly packed with histones during the interphase of the cell cycle. Histones are subject to various post-translational modifications primarily on their NH2-terminal tails, but also in their globular domains. The H3 and H4 histones have long tails protruding from the nucleosome that can be covalently modified at several residues. Modifications of the histone tails include methylation, acetylation, phosphorylation, ubiquitination, SUMOylation and ADP-ribosylation. Histone modification either disrupts chromatin contacts or affects the recruitment of nonhistone proteins to chromatin. For example, acetylation of histones loosens chromatin and lends DNA to replication and transcription. In contrast, methylation at certain residues on histones (e.g., lysine 27 on H3, H3K27) restricts access of the DNA to various enzymes, including RNA polymerase II, thereby causing gene silencing and transcription repression. Overall, histone modifications have significant impact on many fundamental biological processes.5
Lysine residues 4, 9, 27 and 36 in the NH2-terminal tail of H3 are methylated by distinct histone methyltransferases.6 In a residue-dependent manner, lysine methylation of H3 results in either activation or repression of gene transcription.6 The Polycomb group (PcG) protein enhancer of zeste homolog 2 (EZH2) is a methyltransferase primarily responsible for H3 lysine 27 trimethylation (H3K27me3) from fruitflies to mammals.7–12 EZH2 protein contains a signature SET domain at the COOH-terminus that provides the methyltransferase activity. Surprisingly, EZH2 lacks enzyme function on its own. Instead, EZH2 must partner with at least other three noncatalytic proteins, including EED, SUZ12 and RbAp48/46, to form a protein complex called Polycomb repressive complex 2 (PRC2) to attain robust histone methyltransferase activity.7–10,13 Caused by alternative translation four distinct isoforms of EED are detected.14 It is likely that different EZH2-containting PRC2 (EZH2-PRC2) complexes exist.14 In addition, EZH1, a homolog protein of EZH2, also forms a PRC2 complex with other components. The EZH1-PRC2 complex functions differently from the EZH2-PRC2 complex in cells.11,12 The underlying mechanism is not fully understood.15
H3K27me3 plays an important role in epigenetic gene silencing, particularly in homeobox (Hox) gene loci. Trimethylation of H3K27 facilitates the recruitment of Polycomb repressive complex 1 (PRC1) which keeps genes in a repression state. H3K27me3 might play a role in chromatin looping, binding to other protein complexes, altering nucleosome-nucleosome interactions and influencing elongation factors.15 H3K27me3-mediated gene silencing is important for X-chromosome inactivation, developmental patterning and maintenance of stem cell pluripotency.16–18 Recent studies also link EZH2-mediated H3K27me3 to oncogenesis.19,20
Although the biological functions of EZH2 and H3K27me3 have extensively been studied, relatively few studies have been performed to understand how the function of EZH2 is itself regulated. EZH2 gene expression is negatively regulated by the pocket proteins RB, p130 and the microRNA miR-101.21–23 Akt phosphorylates EZH2 at serine 21 and inhibits its methyltransferase activity.24 However, how the function of EZH2 is positively regulated and maintained in normal and cancer cells is largely unclear.
We report recently that EZH2 possesses one perfectly matched (Thr 350, T350) and two imperfectly matched (Thr 421, T421 and Thr 492, T492) CDK consensus phosphorylation motifs (K/RS/TPXK/R), where X represents any amino acid.25 By performing in vitro kinase assays we show that cyclin-dependent kinase 1 (CDK1) and 2 (CDK2) phosphorylate EZH2. Mutation of T350 to alanine (T350A) results in approximately 60% reduction in CDK1-mediated phosphorylation of EZH2. In contrast, approximately 30% or no reduction in phosphorylation is observed when T421A and T492A mutants are used as substrates.25 Thus, our in vitro studies demonstrate that T350 is a major CDK phosphorylation site.
To determine whether T350 on EZH2 is also phosphorylated by CDK1 and CDK2 in cells, we generated a T350 phosphorylation-specific antibody. We show that ectopic expression of CDK1 and cyclin B1 or CDK2 and cyclin E increases T350 phosphorylation in 293T and prostate cancer LNCaP cells.25 This effect of CDK1 and CDK2 is abolished by co-transfection of the CDK inhibitor p21WAF1.25 In contrast, knockdown of endogenous CDK1 and CDK2 results in a decrease in T350 phosphorylation of endogenous EZH2 in LNCaP cells, suggesting that CDK1 and CDK2 also phosphorylate EZH2 at T350 in LNCaP cells. By using a microarray-based gene profiling approach, we further show that T350 phosphorylation is important for EZH2-mediated global gene silencing. Additionally, we demonstrate that T350 phosphorylation is important for the maintenance of H3K27me3 levels at EZH2 target loci in cells. Surprisingly, further studies show that EZH2 phosphorylation at T350 neither affect the assembly of the PRC2 complex nor the intrinsic HMTase activity of PRC2.25 Instead, T350 phosphorylation is important for the recruitment of EZH2 to the promoters of its target genes. In agreement with previous reports,19,22 we demonstrate that overexpression of EZH2 promotes prostate cancer cell proliferation and migration.25 Blockage of T350 phosphorylation diminishes the oncogenic activity of EZH2. Thus, CDK1 and CDK2 phosphorylate EZH2 at T350 in vitro and in vivo, and CDK phosphorylation affects EZH2's function in regulating gene silencing, cell proliferation and migration.
Notably, T350 residue is located in a consensus CDK phosphorylation motif that is evolutionally conserved from fruit-flies to humans.25 Indeed, T345 residue in mouse Ezh2, a homolog site of T350 in human EZH2, is phosphorylated by CDK1 in vitro and in vivo.26 Consistent with the oscillation of CDK activity during the cell cycle, Ezh2 phosphorylation at T345 is higher in cells at S and G2/M phases where CDK1 and CDK2 activity is high than that in cells at G1 phase where CDK1 and CDK2 activity is low. Similar to T350 phosphorylation of human EZH2,25 mouse Ezh2 T345 phosphorylation neither affects complex assembly nor the intrinsic HMTase activity of PRC2.26 Instead, phosphorylation of Ezh2 at T345 increases the binding of Ezh2 to HOTAIR noncoding RNA (ncRNA), which plays an important role in recruitment of PRC2 into its target loci.27,28 Thus, studies with both human and mouse EZH2 show that CDK-mediated phosphorylation of EZH2 is important for PRC2 recruitment to its target loci in cells.
In an effort to identify proteins whose phosphorylation is cell cycle regulated, phosphorylation of the T492 residue in a peptide (476-VKE SSI IAP APA EDV DTP PR-495) of human EZH2 (NM_004456) is identified by mass spectrometry analyses.29 The same peptide phosphorylated by CDK1/cyclin B1 in vitro is also identified via mass spectrometry by an independent study,30 suggesting that T492 on human EZH2 can be phosphorylated by CDK1. T487 in mouse Ezh2, a homolog site of T492 in human EZH2, is also phosphorylated in cells.26 Like human EZH2,30 T487 on mouse Ezh2 can also be phosphorylated by CDK1 in vitro.26 However, T487 phosphorylation on mouse Ezh2 is not affected by treatment of cells with two different CDK inhibitors, although T487 phosphorylation is cell cycle-regulated. It is hence suggested that a kinase distinct from CDK1 might be responsible for Ezh2 phosphorylation at T487 in cells.26 However, an independent study shows that phosphorylation of T492 on human EZH2 increases in cells transfected with CDK1 and cyclin B1, but decreased in cells when endogenous CDK1 is knocked down. Moreover, similar to the effect of T345 phosphorylation mediated by CDK1, T487 phosphorylation on mouse Ezh2 does not affect the HMTase activity of Ezh2-containing PRC2.26 In contrast, T492 phosphorylation of human EZH2 promotes disassociation of EZH2 from SUZ12 and EED and suppresses EZH2 HMTase activity.30 Together, both T492 in human EZH2 and T487 in mouse Ezh2 are shown to be phosphorylated under both in vitro and in vivo conditions.26,29,30 However, which kinase(s) responsible for phosphorylation of this residue in cells are debatable.26,30 Also, the biochemical consequence of phosphorylation at this residue differs from one study to the other.26,30 Further studies are needed to sort out the differences between these studies.
From a mass spectrometry analysis, T372 in a peptide (363-LPN NSS RPS TPT INV LES K-381) of human EZH2 is found phosphorylated in cells synchronized at the mitotic phase of the cell cycle.29 A recent study shows that mitogen-activated protein kinase p38α is responsible for T372 phosphorylation, at least under in vitro conditions.31 While mutation of T372 to alanine does not impair the ability of EZH2 to bind to SUZ12 and EED in PRC2 complex, activation of p38α by its upstream activator MKK6 enhances the interaction of endogenous yin and yang 1 (YY1), a mammalian recruiter of PRC2,15 with exogenous wild-type EZH2, but not with the phosphorylation-resistant mutant T372A.31 Activation of p38α by the inflammation cytokine tumor necrosis factor α (TNFα) results in T372 phosphorylation that enhances the interaction between YY1 and PRC2 and promotes the formation of repressive chromatin in the promoter of Pax7,31 a gene important for muscle stem (satellite) cell proliferation. Genetic knockdown or pharmacological inhibition of p38α and EZH2 invariably promote Pax7 expression and expansion of satellite cells.31 Thus, p38α phosphorylation of EZH2 at T372 provides a mechanism that links inflammation to the epigenetic control of muscle regeneration.
The central theme of epigenetics is that DNA-independent phenotypic changes or gene expression alterations are inheritable through cell divisions. However, how epigenetic marks are transmitted from parental cells to daughter cells remains poorly understood. Increasing evidence indicates that PRC2-mediated H3K27me3 is important for cells to ‘memorize’ gene expression patterns during embryogenesis and development.16 It is believed that maintenance of the H3K27me3 mark during cell division cycles is essential for preservation of cell identity.32,33 Results from a recent study suggest that the PRC2 complex binds to the preexisting H3K27me3 mark at the sites of DNA replication and catalyzes K27 trimethylation on newly synthesized H3 incorporated in the daughter strands of chromatin.34 According to this model, H3K27me3 taking place on de novo synthesized H3 in nascent nucleosomes must occur after S (DNA replication) but before M (cell division) phases. Thus, there seems to be a demand for an active EZH2 during the S-G2 phases of the cell cycle. Interestingly, CDK1 and CDK2 are highly activated at these two stages, and we and others demonstrate recently that activation of CDK1 and CDK2 results in EZH2 phosphorylation at T350, which augments EZH2-mediated H3K27me3 in cells.25,26 Based upon these findings, we envision a model that in addition to their roles in driving DNA replication and mitosis, activation of CDK1 and CDK2 in S and G2 phases also promotes EZH2 T350 phosphorylation. The phosphorylation of EZH2 triggers the binding of PRC2 to recruiters such as HOTAIR and XIST ncRNAs and facilitates H3K27 trimethylation on newly synthesized H3, thereby maintaining H3K27me3 marks in daughter cells (Fig. 1). It is worth noting that Hox gene clusters, the inactive X-chromosome and paternally imprinted loci are extensively studied models for H3K27me3-mediated epigenetic silencing. Intriguingly, they invariably produce ncRNAs and accumulate methylated H3K27.15 Therefore, phosphorylation and activation of EZH2 by CDK1 and CDK2 may play a key role in the maintenance of the H3K27me3 marks in the inactive X-chromosome, paternally imprinted promoters and other PRC2 target loci through cell divisions.
EZH1 is another SET domain-containing protein. It is 65% identical to EZH2.15 EZH1 and EZH2 are partnered with the same core subunits of the PRC2 complex. As demonstrated by in vitro HMTase assays, they exhibit comparable histone methyltransferase activity.12 While knockdown of EZH2 affects global H3K27me3 levels in cells, knockdown of EZH1 is ineffectual on global H3K27me3.11 Also, EZH1 only partially complements EZH2 in executing pluripotency during embryonic stem cell differentiation.12 Notably, EZH1 lacks all the threonine residues on EZH2 that are phosphorylated by CDKs and p38 (Fig. 2). These residues are important for the binding of EZH2 to PRC2 recruiters either ncRNAs or YY1.26,31 It would be interesting to examine whether EZH2 phosphorylation mediated by CDKs and p38 contributes to the functional differences between EZH1 and EZH2 in cells. In line with the finding that the effect of EZH2 on global gene silencing largely relies on CDK1 and CDK2-mediated phosphorylation,25 EZH2 is mainly expressed in proliferating cells whereas EZH1 predominantly expressed in adult tissues and non-dividing cells.11,12,21 Thus, the functional difference between EZH1 and EZH2 may due to the distinct mechanisms that regulate their expression and activity.
While CDK phosphorylation of EZH2 at T350 augments EZH2-mediated H3K27me3 in proliferating cells,25,26 EZH2 is also functional in resting cells.31 Activation of p38α in differentiating muscle satellite cells promotes cell cycle arrest and inhibits proliferation.31 Concordance with the essential role of CDKs in activation of EZH2 in cells,25 many muscle genes that are PRC2 targets are expectedly derepressed in non-proliferating myotubes.31 However, activation of p38 signaling seems to specifically direct PRC2 to repress genes that are typically downregulated during muscle differentiation, such as Pax7.31 This effect of p38 is mediated by p38 phosphorylation of EZH2 at T372 and by T372 phosphorylationenhanced interaction of PRC2 and YY1.31 While activation of most PRC2 target genes inhibits cell proliferation, expression of Pax7 is important for satellite cell proliferation,31 suggesting that Pax7 is a ‘noncanonical’ PRC2 target gene. In line with this observation, YY1, which is required for p38-mediated repression of Pax7,31 is not a general recruiter of PRC2 because there is only limited overlap of YY1 and PRC2 in mouse embryonic stem cell chromatin.35 Based on these findings, we hypothesize that in proliferating cells, CDK1 and CDK2 phosphorylates EZH2, which facilitates the interaction of PRC2 with ‘common’ recruiters such as ncRNAs and promotes H3K27me3 in the promoters of canonical PRC2 target loci. Upon cell cycle exit at certain stages of development such as in terminal differentiated muscles, CDK-mediated phosphorylation of EZH2 at T350 declines, which may trigger the disassociation of PRC2 from ncRNAs and thereby dismiss PRC2 from their canonical target loci. In contrast, environment cues from adult tissues produce tissue-specific signaling, such as activation of p38 in muscle cells. Phosphorylation of EZH2 at T372 due to p38 activation triggers the binding of PRC2 to the gene-specific recruiter YY1, thereby promoting context-dependent gene silencing in non-dividing cells.
In summary, recent findings from us and others demonstrate that EZH2 is phosphorylated at the T350 residue by CDK1 and CDK2. T350 phosphorylation is required for the effective binding of EZH2 to PRC2 target loci in cells. Accordingly, EZH2 is phosphorylated at T350 during S and G2/M phases of the cell cycle where CDK1 and CDK2 are highly activated. Thus, CDK phosphorylation of EZH2 at T350 represents a mechanism that augments EZH2 function during S and G2 phases of the cell cycle. This mechanism may play a role in preservation of cell identity by maintaining H3K27me3 marks through cell divisions.
This work is supported by funds from NIH (CA134514) and Stony Brook University School of Medicine.