Polycomb repressive complex 2 (PRC2) is an important regulator of cellular differentiation and cell type identity. Overexpression or activating mutations of EZH2, the catalytic component of the PRC2 complex, are linked to hyper-trimethylation of lysine 27 of histone H3 (H3K27me3) in many cancers. Potent EZH2 inhibitors that reduce levels of H3K27me3 kill mutant lymphoma cells and are efficacious in a mouse xenograft model of malignant rhabdoid tumors. Unlike most SET domain methyltransferases, EZH2 requires PRC2 components, SUZ12 and EED, for activity, but the mechanism by which catalysis is promoted in the PRC2 complex is unknown. We solved the 2.0 Å crystal structure of the EZH2 methyltransferase domain revealing that most of the canonical structural features of SET domain methyltransferase structures are conserved. The site of methyl transfer is in a catalytically competent state, and the structure clarifies the structural mechanism underlying oncogenic hyper-trimethylation of H3K27 in tumors harboring mutations at Y641 or A677. On the other hand, the I-SET and post-SET domains occupy atypical positions relative to the core SET domain resulting in incomplete formation of the cofactor binding site and occlusion of the substrate binding groove. A novel CXC domain N-terminal to the SET domain may contribute to the apparent inactive conformation. We propose that protein interactions within the PRC2 complex modulate the trajectory of the post-SET and I-SET domains of EZH2 in favor of a catalytically competent conformation.
Citation: Wu H, Zeng H, Dong A, Li F, He H, Senisterra G, et al. (2013) Structure of the Catalytic Domain of EZH2 Reveals Conformational Plasticity in Cofactor and Substrate Binding Sites and Explains Oncogenic Mutations. PLoS ONE 8(12): e83737. https://doi.org/10.1371/journal.pone.0083737
Editor: Sonia Rocha, University of Dundee, United Kingdom
Received: September 20, 2013; Accepted: November 7, 2013; Published: December 19, 2013
Copyright: © 2013 Wu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The SGC is a registered charity (number 1097737) that receives funds from AbbVie, Boehringer Ingelheim, the Canada Foundation for Innovation, the Canadian Institutes for Health Research, Genome Canada through the Ontario Genomics Institute [OGI-055], GlaxoSmithKline, Janssen, Lilly Canada, the Novartis Research Foundation, the Ontario Ministry Research and Innovation, Pfizer, Takeda, and the Wellcome Trust [092809/Z/10/Z]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors are hereby declaring that AbbVie, Boehringer Ingelheim, GlaxoSmithKline, Janssen, Lilly Canada, the Novartis Research Foundation, Pfizer and Takeda are commercial funders of the Structural Genomics Consortium. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Enhancer of zeste homolog 2 (EZH2) is the catalytic component of polycomb repressive complex 2 (PRC2), an epigenetic regulator of stem cell pluripotency, and expression of tissue-specific genes involved in cellular differentiation and developmental programs [1,2,3,4,5].. EZH2 carries out a key function of the PRC2 complex, namely the sequential mono-, di- and trimethylation of Lysine 27 of histone H3 (H3K27) within chromatin. H3K27me3 is an epigenetic mark associated with transcriptional repression and contributes to repression of developmental genes, thereby participating in the maintenance of stem cell pluripotency. EZH2 contains a C-terminal SET-domain, a conserved feature of histone lysine methyltransferases [6,7,8]. However, EZH2 by itself does not demonstrate any methyltransferase activity. The catalytic activity of EZH2 requires the presence of at least two other members of the PRC2 complex, namely embryonic ectoderm development (EED) and suppressor of zeste 12 (SUZ12) [9,10,11]. Two additional subunits, the histone-binding protein RBBP4 and the Zinc finger protein AEBP2 together further stimulate EZH2 enzymatic activity .
Overexpression of EZH2 has been found in a number of human cancers [13,14]. Increased expression levels of EZH2 and increased levels of H3K27me3 are related to tumor development and are associated with poor clinical outcome [15,16,17,18,19]. Elevated expression of EZH2 has also been identified as a marker for breast cancer initiating cells, possibly reflecting its role in maintaining “stemness” [15,20]. In recent studies, missense mutations in EZH2 have been identified in a subset of lymphomas [21,22,23,24,25,26]. Mutations of Y641 (residue numbering according to GeneBank isoform C; Uniprot isoform 1) can increase trimethylase activity of EZH2, thereby leading to elevated global levels of the H3K27me3 mark in mutant cells retaining one wild type allele. In addition, another mutation within the EZH2 SET domain, A677G, has also been identified in Lymphomas and shown to have increased trimethylation efficiency. Together this data suggests a causative role for elevated catalytic activity of EZH2 in the development of cancer.
Because of the important role of EZH2 in tumorigenesis, much effort has been put into discovery of inhibitors of EZH2 catalytic activity. Recently, several groups reported novel inhibitors of EZH2 [27,28,29,30,31,32]. These compounds are selective for EZH2 over other SET domain methyltransferases, are able to inhibit catalytic activity of both wild-type and lymphoma-associated mutants of EZH2 and reduce the cellular level of H3K27me3. They also showed antiproliferative activity in a subset of lymphoma cell lines carrying EZH2 mutants, while the effects on cells carrying wild-type EZH2 were minimal. Finally, EZH2 inhibition induced regression of pediatric rhabdoid tumors, which are almost always dependant on EZH2 activity . Enzyme kinetic studies indicate that these compounds compete with the co-factor S-adenosyl-methionine (SAM), suggesting they bind to the SAM-binding pocket in the SET-domain of EZH2. Stapled peptides that disrupt the EZH2-EED interaction also have antiproliferative activity in an MLL-rearranged leukemia cell line . These results support the potential of EZH2 as a therapeutic target especially in cancers with overexpressed EZH2 or activating mutations.
To better understand the molecular mechanism of EZH2 function, we solved the crystal structure of the C-terminal region of EZH2 containing a novel CXC domain and the catalytic I-SET, SET, and post-SET domains. Although the majority of canonical SET domain characteristics were observed, our structure reveals significant differences between the arrangement of the I-SET and post-SET regions compared to that of other SET-domain containing proteins. An unusual conformation of the post-SET domain suggests a potential mechanism of activation of EZH2 by the other core PRC2 subunits. Our structure also provides insight into the gain-of-function mutations in lymphomas.
Materials and Methods
Cloning, expression and purification
A DNA fragment encoding residues 520-746 of human EZH2 (Uniprot isoform 1) was sub-cloned from cDNA (GenBank BC010858) into the baculovirus expression vector pFBOH-LIC (GenBank EF456740). The pFBOH-LIC N-terminal Hexa-His tag was removed by inserting the gene at an upstream Nco1 restriction site and a C-terminal Hexa-His tag introduced by PCR. The resulting expressed protein product is expected to have the AA sequence: MYQPCDHPRQPCDSSCPCVIAQNFCEKFCQCSSECQNRFPGCRCKAQCNTKQCPCYLAVRECDPDLCLTCGAADHWDSKNVSCKNCSIQRGSKKHLLLAPSDVAGWGIFIKDPVQKNEFISEYCGEIISQDEADRRGKVYDKYMCSFLFNLNNDFVVDATRKGNKIRFANHSVNPNCYAKVMMVNGDHRIGIFAKRAIQTGEELFFDYRYSQADALKYVGIEREMEIPHHHHHH. The protein was expressed in Sf9 cells (Invitrogen). The harvested cells were resuspended in lysis buffer containing 20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 5 mM imidazol, 2 mM β-mercaptoethanol, 5% glycerol, 0.6% NP-40, protease inhibitor cocktail (Roche), 3000 U of benzonase (Novagen). Cells were lysed by brief sonication. The clarified lysate was loaded onto a 2 mL TALON column (Clonetech). The column was washed with 50 column volumes of 20 mM Tris-HCl buffer, pH 8.0, containing 500 mM NaCl, 5% glycerol and 5 mM imidazole. The bound protein was eluted with elution buffer containing 20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 5% glycerol, 250 mM imidazole. The eluted protein was further purified to homogeneity on a Superdex200 column (GE Healthcare), equilibrated with 20 mM Tris-HCl buffer, pH 8.0, and 500 mM NaCl.
Purified EZH2 (10 mg/mL) was mixed with SAM at 1:10 molar ratio of protein:SAM and crystallized using sitting drop vapor diffusion method at 20 °C by mixing 1 µl of the protein solution with 1 µl of the reservoir solution containing 20% PEG 3,350, 0.1 M HEPES, pH 7.5, 0.2 M Li SO4. The crystals were frozen in liquid nitrogen using 10% ethylene glycol as cryo protectant.
Data Collection and Structure Determination
An initial 2.24 Å diffraction data was collected at the Canadian Light Source (CLS) beamline CMCF 08ID-1 at the Zn absorption edge and used to solve the structure of EZH2 by the single-wavelength anomalous dispersion (SAD) phasing method. A second data set (2.0 Å) collected at beam line 19ID of Advanced Photon Source (APS), Argonne National Laboratory was used to further refine of the structure. All data sets collected at a temperature of 100K and were processed using the HKL-3000 suite . REFMAC  was used for structure refinement. The graphics program COOT  was used for model building and visualization. MOLPROBITY was used for structure validation and statistics.
Methyltransferase activity assays for trimeric EZH2 complex (EZH2-EED-SUZ12) and EZH2 (520-746) were performed by monitoring the incorporation of 3H-labeled methyl group into a peptide corresponding to residues 21 to 44 of histone H3 [H3(21-44)] using Scintillation Proximity Assay (SPA). The enzymatic reactions were conducted in triplicate at 23 °C with 0.5 hour incubation of 20 µl reaction mixture in 20 mM Tris-HCl, pH 8.0 (5 mM DTT, 0.01% Triton X-100, 2 µM 3H-SAM (Cat.# NET155V250UC; Perkin Elmer; www.perkinelmer.com), 8 µM cold-SAM, 2 µM H3(21-44) peptide) and various concentrations of enzymes. Enzymatic reactions were stopped by adding 7.5 M Guanidine hydrochloride followed by 180 µl of buffer (20 mM Tris-HCl, pH 8.0), mixing and then transferring to a 96-well FlashPlate (Cat.# SMP103; Perkin Elmer; www.perkinelmer.com). After mixing, the reaction mixtures in Flash-plates were incubated for 2 hour and the CPM were measured using Topcount plate reader (Perkin Elmer, www.perkinelmer.com).
The kinetic parameters for trimeric EZH2 complex were determined using 20 nM of EZH2 at a fixed concentration of peptide (5 µM) and varying concentrations of SAM (up to 20 μM), or at a fixed concentration of SAM (10 µM) and varying concentrations of peptide (up to 5 μM) and therefore kinetic parameters are considered apparent values at the above mentioned conditions. Assays were performed in 20 mM Tris-HCl pH 8, 0.01% Triton X-100, 5 mM DTT. The reaction mixtures were incubated for 30 min at 23 °C. To stop the reactions, 7.5 M Guanidine hydrochloride was added and mixed. A total of 10 µl of the reaction mixture was spotted onto SAM2® Biotin Capture Membrane (cat# V2861, Promega) and placed at room temperature for 5 minutes. Membranes were washed with 2 M NaCl and deionized water at least twice each and were dried. Scintillation liquid was added and counts per minute (CPM) were measured. Experiments were performed in triplicate.
Purified EZH2 (520-746) was dialyzed for binding assay with SAM and peptide in 50 mM Tris-HCl buffer, pH 8.0 and 250 or 500 mM NaCl respectively. Solutions of 1 mM SAM or 0.62 mM histone peptide H3 (residues 21-44) in dialysis buffer were injected into the sample cell containing approximately 1.4 ml of 0.05 mM protein solution. ITC titrations were performed on a VP-ITC Micro Calorimeter from GE Healthcare at 25 °C by using 10 μl injections with a total of 25 injections.
Analysis of crystallographic structures was conducted with ICM (Molsoft, san Diego). “Distances” between the two edges of the substrate binding groove were calculated as the distance between the backbone carbonyl of the catalytic tyrosine, at the C-terminal extremity of the SET domain, and the backbone nitrogen of the I-SET domain which forms a conserved hydrogen bond with the backbone carbonyl of the substrate lysine in all available ternary SET domain methyltransferase structures. These two atoms are themselves linked by a hydrogen bond in the EZH2 structure.
EZH2 has a canonical SET domain scaffold with atypical subdomain features
We crystallized and solved the 2.0 Å structure of the C-terminus of EZH2 (residues 520-746; Figure 1A) (Table 1, Figure S1). Clear electron density was observed for the CXC, SET, and I-SET domains, and the first 5 residues of the post-SET domain. However, the 17 C-terminal residues including most of the post-SET domain were not observed. The cofactor SAM, present in the crystallization buffer, was also absent from the structure.
(A) Linear domain architecture of EZH2 showing the crystallized construct. Residue numbers according to GenBank isoform C (Uniprot isoform 1). (B) The catalytic SET domain (yellow) is folded as previously described for other histone methyltransferases such as EHMT1/GLP and MLL, but the post-SET domain is largely unresolved and its first five residues (blue) are oriented away from its expected position. The unique CXC domain adopts a novel conformation including two clusters of three Zn ions (light blue spheres). (C) A mesh representation of the EZH2 structure in the same orientation. The cofactor is expected to bind at the junction of the SET, post-SET and I-SET (cyan) domains. (D) Residues forming the substrate lysine-binding channel in EHMT1/GLP (beige – PDB code 2RFI) are structurally conserved in EZH2 (color coding as in A-C).
|a, b, c (Å)||45.0, 57.8, 74.5|
|Resolution (Å) (highest resolution shell)||50.00-2.00 (.03-2.00)|
|No. reflections (test set)||13414(654)|
|Rwork/ Rfree (%)||23.9/19.8|
|Bond lengths (Å)||0.010|
|Bond angles (°)||1.13|
|Ramachandran plot % residues|
Overall the EZH2 structure displays most of the canonical features of SET domain protein methyltransferases (Figure 1B-D) . First, the SET domain forms a central scaffold harboring the active site and including a characteristic pseudo-knot (Figure 1B). Second, the I-SET domain forms a β-hairpin that typically participates in formation of the substrate binding groove (Figure 1B). Third, a cluster of conserved aromatic residues line the methyl-lysine binding channel (including the catalytic tyrosine Y726) and superimpose well with ternary structures of other SET domain methyltransferases bound to cofactor and substrate (Figure 1D). This high level of structural conservation in the lysine channel/active site allows the identification of key residues that participate in catalysis (see below). Interestingly, the catalytic site of EZH2 is structurally closer to that of the human H3K9 dimethylase EHMT1 than the viral H3K27 trimethylase vSET (Figure S2), which is in agreement with the latter’s distinctly different enzymatic properties . Finally, the putative substrate binding site is electronegative (Figure S3), as expected for interactions with the highly basic H3 histone tail.
Despite the highly conserved SET domain, the EZH2 structure has several novel features (Figure 2). First, immediately N-terminal to the SET domain, a unique CXC domain, coordinated by 2 clusters of three zinc ions, differentiates EZH2 from other SET-domain PMT structures (no structural homolog of the CXC domains was found by the DALI server ). Missense mutations at residues coordinating the first and second zinc of the CXC domain were reported in acute myeloid leukemia (H525N) and myelodysplastic syndrome (C571N) suggesting that disruption of the CXC domain can be associated with specific cancer types [41,42]. Second, the conformation of the first five residues of the post-SET domain diverges drastically from previous structures of active SET-domain methyltransferases, and folds in a direction diametrically opposite to its expected position, where it would otherwise contribute to formation of the cofactor binding site (Figure 2 and Figure S4). Importantly, within the crystal lattice, each CXC domain interacts with a neighboring protein partially occupying both the SAM binding site and the region in which the post-SET domain is expected to reside (Figure 3). Thus, the conformation observed in our structure is not compatible with SAM binding and therefore reflects an inactive enzyme. The functional implications of both the conserved and unusual features of this structure are discussed below. Finally, we note the presence of a secondary pocket formed by the I-SET domain juxtaposed to the SAM binding site of EZH2 (Figure S5). It is unclear whether this pocket would be present in a non-crystalline environment or in the context of the PRC2 complex.
The substrate binding groove is too wide in MLL (right) and too narrow in EZH2 (left), compared with the catalytically competent state observed in the EHMT1/GLP ternary complex (center). Color-coding as in Figure 1.
(A) Superimposition of the EZH2 structure (colored mesh; post-SET shown as blue ribbon) with a ternary complex of EHMT1/GLP (white ribbon) shows that the cofactor binding site is only partially formed in EZH2, due to an atypical orientation of the post-SET domain. (B) The cofactor site of EZH2 is occupied by the CXC domain of a second molecule within the crystal lattice. (C) Mapping of the location of lysine-mediated cross-links detected in the purified PRC2 complex . Cross-links between Lys735 and Lys569 as well as Lys713 indicate that the post-SET domain of EZH2 (yellow) can project towards the CXC domain in solution, consistent with the conformation seen in our structure.
Structural basis for altered activity of recurrent lymphoma mutations
Recurrent mutations at Y641 and A677 have been shown to increase the trimethylase activity of EZH2 and drive the development of diffuse large B-cell and follicular lymphoma [21,22,23,24,25,26]. Superimposition of our EZH2 structure with that of another SET-domain methyltransferase, EHMT1/GLP, in complex with SAH and a dimethylated H3K9 peptide  confirms previous homology models [21,26] showing that Y641 is perfectly positioned to engage in a hydrogen-bond with the de-protonated ε-nitrogen of the substrate lysine, thereby restraining the rotational freedom of the di-methylated nitrogen atom, and disfavoring alignment of the lone pair with the scissile bond of the cofactor’s sulfonium group, necessary for the displacement of a third methyl group (Figure 4). Changing Y641 to a phenylalanine, a mutation frequently associated with lymphoma, would alleviate the conformational constraint imposed on the substrate lysine, allowing the latter to freely rotate into position for nucleophilic attack on SAM, and would provide additional space to accommodate a third methyl group. A similar mechanism was previously described for two other SET domain methyltransferases, SETD7 and EHMT2/G9a, in which mutations of the corresponding tyrosine (Y1067F in EHMT1/G9a, Y245A in SETD7) switched the predominant catalytic activity from mono- and di-methylase, respectively, to tri-methylase [43,44]. Similarly, a Tyr to Phe substitution switched SETD8 from a mono- into a di-methylase . Other mutations at Y641 to asparagine, histidine and cysteine have been described that also affect the methylation specificity of EZH2 . These mutations are also expected to disrupt the hydrogen-bond between the hydroxy group of the tyrosine and the substrate lysine, but otherwise be compatible with the structure of the lysine binding channel.
Hydrogen bonding between Tyr 641 and the substrate lysine’s ε-nitrogen, and steric envelope of the tyrosine hydroxy group impose rotational constraints that penalize proper alignment with the cofactor’s scissile bond, required for displacement of a third methyl group. A677 stabilizes the conformation of Y641 hydrogen-bonded to the substrate lysine. The cofactor and substrate lysine are from a superimposed ternary structure of EHMT1/GLP (2RFI).
Our structure also confirms a previous explanation for increased trimethylase activity of A677G mutation found in lymphoma . Y641 is sandwiched between the substrate lysine and the side-chain of A677 (Figure 4). This supports an activation mechanism whereby the absence of a side chain at residue 677 allows a conformational state of Y641 that is not accessible in the wild type protein. Such an alternate confirmation could reduce the active site hydrogen-bonding potential of Y641 with the methyl-accepting nitrogen of the substrate lysine, and remove steric hindrance associated with the trimethylated state, thereby allowing proper alignment of the substrate’s dimethylated nitrogen for displacement of a methyl group from the cofactor, resulting in increased tri-methylation activity .
Isolated EZH2 adopts a conformation that precludes SAM and substrate binding
Measurable catalytic activity for EZH2 requires the presence of PRC2 core subunits SUZ12, and EED, with incorporation of a fourth PRC2 core component, RBBP4 or RBBP7, and cofactors such as AEBP2 resulting in even greater activity [10,12,46]. We confirmed that our crystallized EZH2 construct is inactive in isolation compared to the trimeric complex of full length EZH2/EED/SUZ12 under the same conditions (Figure 5A). Since our structure indicates that the lysine-binding channel and active site appear to be competent for catalysis (Figure 1D), we tested whether the crystallized construct was capable of binding substrate or cofactor, two binding events necessary for catalysis. While kinetic analysis shows that the trimeric PRC2 complex with full-length EZH2 binds both SAM and a H3K27 peptide substrate, our crystallized EZH2 construct binds neither as measured by ITC (Figure 5D and E). This data indicates that the activity deficit in isolated, truncated EZH2 is due, at least in part, to failure to bind substrate and cofactor. There are several features of the structure, each of which can explain this lack of activity, and hint at a structural mechanism for activation of EZH2 by other PRC2 subunits.
(A) The full-length trimeric complex (●) was active, and the crystallized EZH2 construct (○) was not. Activity assay conditions were optimized for the full length EZH2 in complex with EED and SUZ12 as a control. Kinetic analysis shows that the trimeric complex binds SAM (B) and a histone peptide (C) (Km SAM: 900 ± 100 nM; Km peptide: 205 ± 25 nM; kcat: 24 ± 2 h-1). Apparent kinetic parameters are the average of three measurements ± standard deviation. ITC shows that the crystallized construct binds neither SAM (D) nor the peptide substrate (E).
First, as mentioned above, the cofactor site is occupied by the CXC domain of a second EZH2 molecule (Figure 3B). Whether this feature is a crystallographic artifact or is a cause or a consequence of post-SET misfolding is unknown. However, this intermolecular interaction cannot be the sole explanation for the inactivity of EZH2 because in solution EZH2 appears to exist both as monomer and dimer (Figure S6), and we observe no catalytic activity associated with the ~50% monomeric species (Figure 5).
A second structural feature that reflects an inactive enzyme is the absence of a peptide binding groove due to the relative position of the I-SET and the first several residues of the post-SET domain. All ternary structures of SET domain methyltransferases in complex with cofactor and substrate have the substrate peptide lying in a narrow groove formed by the I-SET domain on one side, and the post-SET domain on the opposite side. This peptide binding groove is responsible for substrate sequence specificity and positions the substrate lysine side chain into a conserved channel that goes deep into the core of the SET domain and meets the cofactor at the conserved active site as described above (Figure 2) ([38,43,44,47,48,49,50,51,52] and PDB code 4AU7). The width of the peptide-binding groove is critical for catalysis, and the distance separating the two edges is between 7.2 and 7.8 Å across all available structures of active SET domain constructs (see materials and methods section for details). Southall et al have shown that in the MLL SET domain structure, the groove is more open (9.6 Å), which results in suboptimal positioning of the substrate lysine and loss of catalytic activity (Figure 2) . Addition of RBBP5 and ASH2L, two other subunits of the active MLL complex, are necessary for activity and thought to contribute to completing or closing the peptide channel . The EZH2 structure reveals the opposite conformational state that could have similar consequences: a shift in the I-SET domain brings it in close proximity (2.9 Å) to the first 5 residues of the post-SET domain, effectively eliminating the peptide binding groove and sealing the entrance of the substrate lysine channel. This conformation is mediated by a hydrogen bond between the backbone nitrogen of N668 in the I-SET domain and the backbone oxygen of Y726 in the post-SET (Figure 2, Figure S7). This conformational state is incompatible with binding and proper positioning of the substrate for catalysis.
A third feature that may contribute to inactivity is the position of the post-SET domain. In all other SAM or SAH-bound structures of SET domain PMTs, the post-SET domain partially contributes, along with the SET and I-SET domains, to the formation of the cofactor binding site (Figure S4). However, the post-SET domain of EZH2 appears to project towards the CXC domain, away from the expected site of cofactor recruitment, resulting in an incomplete cofactor pocket (Figure 3A). Because only 5 of 22 residues of the post-SET domain have clear electron density in our structure, the actual position of the full post-SET domain is uncertain and may have multiple conformations. However, given the trajectory of the first five residues and the small size of the post-SET domain, it is unlikely to be able to “reach” the cofactor site. Similar orientations are observed at the N-terminus of the post-SET domain in the auto-inhibited, cofactor-bound structures of SUV39H2 and SETMAR [PDB codes 2R3A, 3BO5], but the full-length post-SET of these proteins is at least 18 residues longer than that of EZH2. Furthermore, using cross-linking experiments coupled with mass spectrometry in the context of the reconstituted, purified PRC2 complex, Ciferri et al. have captured physical proximity between Lys735, in the unresolved section of the post-SET domain, and Lys569, in the CXC domain, as well as neighboring Lys713 in the SET domain  (Figure 3C). This suggests that the unexpected orientation of the post-SET domain observed in our structure is not a crystallographic artifact, but is at least one of several possible conformations within the PRC2 complex.
The orientation of the EZH2 post-SET domain away from the SAM-binding site is analogous to that observed in a series of structures we previously solved for the human PRDM proteins - PRDM1 (PDB code 3DAL), PRDM2 (2QPW) , PRMD4 (3DB5), PRDM9 (4IJD) , PRDM10 (3IHX), PRDM11 (3RAY), and PRDM12 (3EP0) (Figure S8). PRDMs are methyltransferases with a core catalytic PR domain that is structurally related to the SET domain. Interestingly, the cofactor SAM is absent from all these structures, but is present in a structure of mouse PRDM9 (4C1Q) in which the post-SET domain adopts a conformation that is closer to that observed in active SET domain protein structures . However, unlike EZH2, the human PRDM9 construct that crystallized in an apo, inactive conformation is nevertheless able to bind to SAM and has significant catalytic activity .
Together, these results indicate structural features underlying suboptimal binding or positioning of both cofactor and substrate in isolated EZH2, which are expected to preclude catalysis. To test whether the specific structural features observed in our structure contribute to lack of substrate or cofactor binding we designed several EZH2 mutations that would be expected to disrupt or ‘relieve’ the apparent inactive conformations of the I-SET, post-SET and CXC domains. We mutated post-SET residue Ser729 (buried due to the altered post-SET conformation) to an aspartate to destabilize the observed conformation of the post-SET domain. We also mutated I-SET residue Phe667 (engaged in orthogonal pi stacking with F724) into a leucine to remove interactions that stabilize the shifted conformation of the I-SET domain (Figure S7). Neither of these mutants, nor the double mutant was able to bind substrate or cofactor as measured by either ITC or differential scanning light scattering . Attempts to prepare a construct that lacked the CXC domain (which might be expected to relieve the intermolecular interaction with, and disruption of the cofactor site) were unsuccessful due to instability of the truncated protein.
Finally, given the recently reported potent SAM-competitive inhibitors of EZH2 [27,28,29,30,31,32] we hypothesized that their tight binding to the cofactor binding site in the PRC2 complex may confer the ability to bind to the isolated EZH2 protein, even though SAM does not bind. However, we could detect no binding for UNC1999 by ITC (data not shown), indicating that SAM competitive inhibitors also require the minimal PRC2 complex components.
The crystal structure presented here reveals that EZH2 adopts a canonical SET domain methyltransferase fold in the absence of binding partners, and that the catalytic site is well formed (Figure 1D). We find that the structure of the cofactor site is compatible with the formation of four of six hydrogen bonds with the cofactor that are conserved across all SET domain complex structures (Figure S9) . The structure also rationalizes the increased trimethylase activity of the mutated enzyme found in lymphomas. However, unlike the isolated SET domains of many other methyltransferases, this isolated SET domain construct of EZH2 is unable to methylate its H3K27 substrate in the absence of protein interaction partners, EED and SUZ12 . Two structural features distinguish our EZH2 structure from catalytically competent conformations of other SET domain methyltransferases: the post-SET domain projects away from its expected position (Figure 3A), resulting in an incomplete cofactor binding site, and the I-SET domain is shifted towards the post-SET domain, which closes the histone binding groove and blocks the entrance of the substrate lysine channel (Figure 2 and Figure S10). While these features may be related to crystal lattice contacts which are numerous both at the post-SET and I-SET domains, there is evidence that at least some of the atypical features of our crystallized conformation (inactive trajectory of the post-SET) may be populated in the PRC2 complex in solution (Figure 3C) .
Our EZH2 structure contributes to the growing evidence for conformational plasticity of the I-SET and post-SET domains of SET domain methyltransferases and hints at potential mechanisms of regulation of catalytic activity. The post-SET domain participates in the formation of both cofactor and substrate binding sites, and is expected to adopt a catalytically competent conformation only when bound to both SAM and peptide ( for review). For instance SETD7 was captured crystallographically in apo, cofactor-bound, and cofactor- plus substrate-bound states, each with a unique conformation of the post-SET domain [44,57,58]. Similarly, (as discussed above) the related PRDM methytransferases also display extensive variability in the position of their post-SET domains . Furthermore, a recent structure of SETD8 in complex with SAM (PDB code 4IJ8) compared to the ternary structure  reveals that structural plasticity exists also at the I-SET domain. Finally, autoinhibitory conformations where the post-SET domain occludes the substrate binding groove were reported for the H3K36 methyltransferases SETD2, SETMAR and NSD1 and the H3K9 methyltransferase SUV39H2 ([43,59,60] and PDB code 3BO5). Thus, the atypical conformations of the post-SET and (to a lesser extent) I-SET domains observed in EZH2 may therefore represent one of several conformational states available to the protein and suggest a potential mechanism for modulation of catalytic activity within the PRC2 complex.
We propose that the other PRC2 subunits and/or N-terminal regions of full length EZH2 conspire to complete an active cofactor binding site by modulating the conformation of the post-SET and I-SET domains. Modulation of the trajectory of the post-SET domain toward the cofactor site would be expected to ‘release’ the I-SET from interaction with the post-SET thereby opening up the peptide-binding groove for substrate. Alternatively, interactions of PRC2 components with the I-SET may reposition it into a conformation more consistent with other active SET domain proteins, thereby releasing the post-SET to adopt an active conformation. Finally, it is entirely possible that a complex and possibly dynamic series of protein-protein and inter-domain interactions take place within the functional PRC2 complex to carry out its catalytic function.
Electron density map. Sample electron density map of the substrate lysine channel.
EZH2, EHMT1 and vSET catalytic sites. The catalytic site of EZH2 is structurally closer to that of the human H3K9 dimethylase EHMT1 than the viral H3K27 trimethylase vSET .
Electrostatic potential. The expected location of the substrate peptide binding site of EZH2, at the interface of the post-SET (blue) and I-SET (cyan) domains, is electronegative. Bottom: electrostatic potential color coding. Blue: electropositive; red: electronegative.
Post-SET domain conformation. The post-SET domain (blue) of cofactor-bound SET domain methyltransferases is structurally diverse but always participates in the formation of the cofactor site. In the EZH2 structure, it projects away from its expected position and the cofactor is absent. When present, cofactor is shown as CPK and substrate is in green.
EZH2’s secondary pocket. A mesh representation of EZH2 (color-coding as in other figures) with the cofactor of a superimposed EHMT1/GLP structure (conserved hydrogen-bonds are highlighted), reveals the existence of a secondary pocket, juxtaposed to the cofactor site.
EZH2’s dimeric state in solution. EZH2 elutes both as a monomer and dimer out of a gel filtration column.
Interactions between the post-SET and I-SET domains. The altered orientation of the post-SET domain, resulting in incomplete formation of the cofactor site, is associated with a buried conformation of Ser 729. The shifted orientation of the I-SET domain, resulting in closure of the substrate-binding groove, is stabilized by a hydrogen-bond between the backbone of N668 and Y726, and orthogonal pi-stacking between Phe 667 and Phe 724. Color coding as in other figures.
Post-SET domain in PRDM structures. The post-SET domain in all human PRDM structures (blue) is oriented away from the putative cofactor site, and the cofactor is absent from all these structures. In a mouse PRDM9 structure crystallized in complex with SAH (green sticks), the post-SET domain (green ribbon) is folded on the cofactor. Mesh representation of human PRDM9 where the post-SET domain was truncated. Post-SET domain of human PRDM1 (PDB code 3DAL), PRDM2 (2QPW Wu 20084102), PRMD4 (3DB5), PRDM9 (4IJD), PRDM10 (3IHX), PRDM11 (3RAY), and PRDM12 (3EP0), and mouse PRDM9 (4C1Q).
Conserved, but incomplete folding of the cofactor-binding site. The cofactor site of EZH2 is in a conformational state that is compatible with the formation of 4 out of 6 hydrogen bonds (black) between the SET domain and the cofactor that are conserved across all available structures of cofactor-bound SET-domain methyltransferases. Preserved hydrogen bonds are shown in cyan. Lost hydrogen bonds are shown in magenta. The EZH2 structure (color coding as in other figures) is superimposed with cofactor-bound EHMT1/GLP (beige - PDB code 2RFI). Top-right: same view, with a mesh representation of EZH2, where the EHMT1/GLP ribbon was removed.
Atypical conformations of the I-SET and post-SET domains. Superimposition of the EZH2 structure (I-SET domain: cyan; post-SET domain: blue) with ternary complexes of EHMT1/GLP (PDB code 2RFI), SETD7 (PDB code 1O9S) and SETD8 (PDB code 1ZKK) bound to cofactor (balls and sticks) and substrate (no shown) shows that the I-SET domain of EZH2 is shifted towards the post-SET domain, resulting in hydrogen-bonding between Asn 668 and Tyr 726.
We thank John Luz and colleagues for discussions and for sharing their manuscript on a similar structure before publication. Research described in this paper was performed using beamline 08ID-1 at the Canadian Light Source, which is supported by the Natural Sciences and Engineering Research Council of Canada, the National Research Council Canada, the Canadian Institutes of Health Research, the Province of Saskatchewan, Western Economic Diversification Canada, and the University of Saskatchewan and using beamline 19ID at Argonne National Laboratory, Structural Biology Center at the Advanced Photon Source, which is operated by UChicago Argonne, LLC, for the U.S. Department of Energy, Office of Biological and Environmental research under contract DE-AC02-06CH11357. CHA holds a Canada Research Chair in Structural Genomics. Atomic coordinates of the EZH2 structure were deposited at the RCSB with PDB code 4MI0.
Conceived and designed the experiments: HW MV CHA. Performed the experiments: HW HZ AD FL GS HH AS SD. Analyzed the data: HW MV CHA MS. Wrote the manuscript: HW CHA MS. Managed the research project: PJB.
- 1. Müller J, Hart CM, Francis NJ, Vargas ML, Sengupta A et al. (2002) Histone methyltransferase activity of a Drosophila Polycomb group repressor complex. Cell 111: 197-208. doi:https://doi.org/10.1016/S0092-8674(02)00976-5. PubMed: 12408864.
- 2. Czermin B, Melfi R, McCabe D, Seitz V, Imhof A et al. (2002) Drosophila enhancer of Zeste/ESC complexes have a histone H3 methyltransferase activity that marks chromosomal Polycomb sites. Cell 111: 185-196. doi:https://doi.org/10.1016/S0092-8674(02)00975-3. PubMed: 12408863.
- 3. Cao R, Wang L, Wang H, Xia L, Erdjument-Bromage H et al. (2002) Role of histone H3 lysine 27 methylation in Polycomb-group silencing. Science 298: 1039-1043. doi:https://doi.org/10.1126/science.1076997. PubMed: 12351676.
- 4. Kirmizis A, Bartley SM, Kuzmichev A, Margueron R, Reinberg D et al. (2004) Silencing of human polycomb target genes is associated with methylation of histone H3 Lys 27. Genes Dev 18: 1592-1605. doi:https://doi.org/10.1101/gad.1200204. PubMed: 15231737.
- 5. Kuzmichev A, Jenuwein T, Tempst P, Reinberg D (2004) Different EZH2-containing complexes target methylation of histone H1 or nucleosomal histone H3. Mol Cell 14: 183-193. doi:https://doi.org/10.1016/S1097-2765(04)00185-6. PubMed: 15099518.
- 6. Joshi P, Carrington EA, Wang L, Ketel CS, Miller EL et al. (2008) Dominant alleles identify SET domain residues required for histone methyltransferase of Polycomb repressive complex 2. J Biol Chem 283: 27757-27766. doi:https://doi.org/10.1074/jbc.M804442200. PubMed: 18693240.
- 7. Dillon SC, Zhang X, Trievel RC, Cheng X (2005) The SET-domain protein superfamily: protein lysine methyltransferases. Genome Biol 6: 227. doi:https://doi.org/10.1186/gb-2005-6-8-227. PubMed: 16086857.
- 8. Rea S, Eisenhaber F, O'Carroll D, Strahl BD, Sun ZW et al. (2000) Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature 406: 593-599. doi:https://doi.org/10.1038/35020506. PubMed: 10949293.
- 9. Pasini D, Bracken AP, Jensen MR, Lazzerini Denchi E, Helin K (2004) Suz12 is essential for mouse development and for EZH2 histone methyltransferase activity. EMBO J 23: 4061-4071. doi:https://doi.org/10.1038/sj.emboj.7600402. PubMed: 15385962.
- 10. Cao R, Zhang Y (2004) SUZ12 is required for both the histone methyltransferase activity and the silencing function of the EED-EZH2 complex. Mol Cell 15: 57-67. doi:https://doi.org/10.1016/j.molcel.2004.06.020. PubMed: 15225548.
- 11. Yamamoto K, Sonoda M, Inokuchi J, Shirasawa S, Sasazuki T (2004) Polycomb group suppressor of zeste 12 links heterochromatin protein 1alpha and enhancer of zeste 2. J Biol Chem 279: 401-406. PubMed: 14570930.
- 12. Margueron R, Reinberg D (2011) The Polycomb complex PRC2 and its mark in life. Nature 469: 343-349. doi:https://doi.org/10.1038/nature09784. PubMed: 21248841.
- 13. Chase A, Cross NC (2011) Aberrations of EZH2 in cancer. Clin Cancer Res 17: 2613-2618. doi:https://doi.org/10.1158/1078-0432.CCR-10-2156. PubMed: 21367748.
- 14. Simon JA, Lange CA (2008) Roles of the EZH2 histone methyltransferase in cancer epigenetics. Mutat Res 647: 21-29. doi:https://doi.org/10.1016/j.mrfmmm.2008.07.010. PubMed: 18723033.
- 15. Kleer CG, Cao Q, Varambally S, Shen R, Ota I et al. (2003) EZH2 is a marker of aggressive breast cancer and promotes neoplastic transformation of breast epithelial cells. Proc Natl Acad Sci U S A 100: 11606-11611. doi:https://doi.org/10.1073/pnas.1933744100. PubMed: 14500907.
- 16. Takawa M, Masuda K, Kunizaki M, Daigo Y, Takagi K et al. (2011) Validation of the histone methyltransferase EZH2 as a therapeutic target for various types of human cancer and as a prognostic marker. Cancer Sci 102: 1298-1305. doi:https://doi.org/10.1111/j.1349-7006.2011.01958.x. PubMed: 21539681.
- 17. Varambally S, Cao Q, Mani RS, Shankar S, Wang X et al. (2008) Genomic loss of microRNA-101 leads to overexpression of histone methyltransferase EZH2 in cancer. Science 322: 1695-1699. doi:https://doi.org/10.1126/science.1165395. PubMed: 19008416.
- 18. Varambally S, Dhanasekaran SM, Zhou M, Barrette TR, Kumar-Sinha C et al. (2002) The polycomb group protein EZH2 is involved in progression of prostate cancer. Nature 419: 624-629. doi:https://doi.org/10.1038/nature01075. PubMed: 12374981.
- 19. Wagener N, Macher-Goeppinger S, Pritsch M, Hüsing J, Hoppe-Seyler K et al. (2010) Enhancer of zeste homolog 2 (EZH2) expression is an independent prognostic factor in renal cell carcinoma. BMC Cancer 10: 524. doi:https://doi.org/10.1186/1471-2407-10-524. PubMed: 20920340.
- 20. Kunju LP, Cookingham C, Toy KA, Chen W, Sabel MS et al. (2011) EZH2 and ALDH-1 mark breast epithelium at risk for breast cancer development. Mod Pathol 24: 786-793. doi:https://doi.org/10.1038/modpathol.2011.8. PubMed: 21399615.
- 21. McCabe MT, Graves AP, Ganji G, Diaz E, Halsey WS et al. (2012) 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 U S A 109: 2989-2994. doi:https://doi.org/10.1073/pnas.1211753109. PubMed: 22323599.
- 22. Morin RD, Mendez-Lago M, Mungall AJ, Goya R, Mungall KL et al. (2011) Frequent mutation of histone-modifying genes in non-Hodgkin lymphoma. Nature 476: 298-303. doi:https://doi.org/10.1038/nature10351. PubMed: 21796119.
- 23. Pasqualucci L, Trifonov V, Fabbri G, Ma J, Rossi D et al. (2011) Analysis of the coding genome of diffuse large B-cell lymphoma. Nat Genet 43: 830-837. doi:https://doi.org/10.1038/ng.892. PubMed: 21804550.
- 24. Sneeringer CJ, Scott MP, Kuntz KW, Knutson SK, Pollock RM et al. (2010) Coordinated activities of wild-type plus mutant EZH2 drive tumor-associated hypertrimethylation of lysine 27 on histone H3 (H3K27) in human B-cell lymphomas. Proc Natl Acad Sci U S A 107: 20980-20985. doi:https://doi.org/10.1073/pnas.1012525107. PubMed: 21078963.
- 25. Wigle TJ, Knutson SK, Jin L, Kuntz KW, Pollock RM et al. (2011) The Y641C mutation of EZH2 alters substrate specificity for histone H3 lysine 27 methylation states. FEBS Lett 585: 3011-3014. doi:https://doi.org/10.1016/j.febslet.2011.08.018. PubMed: 21856302.
- 26. Yap DB, Chu J, Berg T, Schapira M, Cheng SW et al. (2011) Somatic mutations at EZH2 Y641 act dominantly through a mechanism of selectively altered PRC2 catalytic activity, to increase H3K27 trimethylation. Blood 117: 2451-2459. doi:https://doi.org/10.1182/blood-2010-11-321208. PubMed: 21190999.
- 27. Copeland RA (2013) Molecular Pathways: Protein methyltransferases in cancer. Clin Cancer Res, 19: 6344–50. PubMed: 23958745.
- 28. Knutson SK, Warholic NM, Wigle TJ, Klaus CR, Allain CJ et al. (2013) Durable tumor regression in genetically altered malignant rhabdoid tumors by inhibition of methyltransferase EZH2. Proc Natl Acad Sci U S A 110: 7922-7927. doi:https://doi.org/10.1073/pnas.1303800110. PubMed: 23620515.
- 29. Knutson SK, Wigle TJ, Warholic NM, Sneeringer CJ, Allain CJ et al. (2012) A selective inhibitor of EZH2 blocks H3K27 methylation and kills mutant lymphoma cells. Nat Chem Biol 8: 890-896. PubMed: 23023262.
- 30. Konze KD, Ma A, Li F, Barsyte-Lovejoy D, Parton T et al. (2013) An Orally Bioavailable Chemical Probe of the Lysine Methyltransferases EZH2 and EZH1. ACS Chem Biol PubMed: 23614352.
- 31. McCabe MT, Ott HM, Ganji G, Korenchuk S, Thompson C et al. (2012) EZH2 inhibition as a therapeutic strategy for lymphoma with EZH2-activating mutations. Nature 492: 108-112. doi:https://doi.org/10.1038/nature11606. PubMed: 23051747.
- 32. Qi W, Chan H, Teng L, Li L, Chuai S et al. (2012) Selective inhibition of Ezh2 by a small molecule inhibitor blocks tumor cells proliferation. Proc Natl Acad Sci U S A 109: 21360-21365. doi:https://doi.org/10.1073/pnas.1210371110. PubMed: 23236167.
- 33. Kim W, Bird GH, Neff T, Guo G, Kerenyi MA et al. (2013) Targeted disruption of the EZH2-EED complex inhibits EZH2-dependent cancer. Nat Chem Biol, 9: 643–50. PubMed: 23974116.
- 34. Minor W, Cymborowski M, Otwinowski Z, Chruszcz M (2006) HKL-3000: the integration of data reduction and structure solution--from diffraction images to an initial model in minutes. Acta Crystallogr D Biol Crystallogr 62: 859-866. doi:https://doi.org/10.1107/S0907444906019949. PubMed: 16855301.
- 35. Murshudov GN, Vagin AA, Dodson EJ (1997) Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr D Biol Crystallogr 53: 240-255. doi:https://doi.org/10.1107/S0907444996012255. PubMed: 15299926.
- 36. Emsley P, Cowtan K (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 60: 2126-2132. doi:https://doi.org/10.1107/S0907444904019158. PubMed: 15572765.
- 37. Davis IW, Murray LW, Richardson JS, Richardson DC (2004) MOLPROBITY: structure validation and all-atom contact analysis for nucleic acids and their complexes. Nucleic Acids Res 32: W615-W619. doi:https://doi.org/10.1093/nar/gkh398. PubMed: 15215462.
- 38. Schapira M (2011) Structural Chemistry of Human SET Domain Protein Methyltransferases. Curr Chem Genomics 5: 85-94. doi:https://doi.org/10.2174/1875397301005010085. PubMed: 21966348.
- 39. Swalm BM, Hallenbeck KK, Majer CR, Jin L, Scott MP et al. (2013) Convergent evolution of chromatin modification by structurally distinct enzymes: comparative enzymology of histone H3 Lys(2)(7) methylation by human polycomb repressive complex 2 and vSET. Biochem J 453: 241-247. doi:https://doi.org/10.1042/BJ20130439. PubMed: 23679895.
- 40. Holm L, Rosenström P (2010) Dali server: conservation mapping in 3D. Nucleic Acids Res 38: W545-W549. doi:https://doi.org/10.1093/nar/gkp893. PubMed: 20457744.
- 41. Wang X, Dai H, Wang Q, Xu Y, Wang Y et al. (2013) EZH2 mutations are related to low blast percentage in bone marrow and -7/del(7q) in de novo acute myeloid leukemia. PLOS ONE 8: e61341. doi:https://doi.org/10.1371/journal.pone.0061341. PubMed: 23613835.
- 42. Ernst T, Chase AJ, Score J, Hidalgo-Curtis CE, Bryant C et al. (2010) Inactivating mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat Genet 42: 722-726. doi:https://doi.org/10.1038/ng.621. PubMed: 20601953.
- 43. Wu H, Min J, Lunin VV, Antoshenko T, Dombrovski L et al. (2010) Structural biology of human H3K9 methyltransferases. PLOS ONE 5: e8570. doi:https://doi.org/10.1371/journal.pone.0008570. PubMed: 20084102.
- 44. Xiao B, Jing C, Wilson JR, Walker PA, Vasisht N et al. (2003) Structure and catalytic mechanism of the human histone methyltransferase SET7/9. Nature 421: 652-656. doi:https://doi.org/10.1038/nature01378. PubMed: 12540855.
- 45. Couture JF, Dirk LM, Brunzelle JS, Houtz RL, Trievel RC (2008) Structural origins for the product specificity of SET domain protein methyltransferases. Proc Natl Acad Sci U S A 105: 20659-20664. doi:https://doi.org/10.1073/pnas.0806712105. PubMed: 19088188.
- 46. Margueron R, Justin N, Ohno K, Sharpe ML, Son J et al. (2009) Role of the polycomb protein EED in the propagation of repressive histone marks. Nature 461: 762-767. doi:https://doi.org/10.1038/nature08398. PubMed: 19767730.
- 47. Chang Y, Levy D, Horton JR, Peng J, Zhang X et al. (2011) Structural basis of SETD6-mediated regulation of the NF-kB network via methyl-lysine signaling. Nucleic Acids Res 39: 6380-6389. doi:https://doi.org/10.1093/nar/gkr256. PubMed: 21515635.
- 48. Wang L, Li L, Zhang H, Luo X, Dai J et al. (2011) Structure of human SMYD2 protein reveals the basis of p53 tumor suppressor methylation. J Biol Chem 286: 38725-38737. doi:https://doi.org/10.1074/jbc.M111.262410. PubMed: 21880715.
- 49. Couture JF, Collazo E, Brunzelle JS, Trievel RC (2005) Structural and functional analysis of SET8, a histone H4 Lys-20 methyltransferase. Genes Dev 19: 1455-1465. doi:https://doi.org/10.1101/gad.1318405. PubMed: 15933070.
- 50. Xiao B, Jing C, Kelly G, Walker PA, Muskett FW et al. (2005) Specificity and mechanism of the histone methyltransferase PrPr-Set7. Genes Dev 19: 1444-1454. doi:https://doi.org/10.1101/gad.1315905. PubMed: 15933069.
- 51. Southall SM, Wong PS, Odho Z, Roe SM, Wilson JR (2009) Structural basis for the requirement of additional factors for MLL1 SET domain activity and recognition of epigenetic marks. Mol Cell 33: 181-191. doi:https://doi.org/10.1016/j.molcel.2008.12.029. PubMed: 19187761.
- 52. Chuikov S, Kurash JK, Wilson JR, Xiao B, Justin N et al. (2004) Regulation of p53 activity through lysine methylation. Nature 432: 353-360. doi:https://doi.org/10.1038/nature03117. PubMed: 15525938.
- 53. Ciferri C, Lander GC, Maiolica A, Herzog F, Aebersold R et al. (2012) Molecular architecture of human polycomb repressive complex 2. Elife 1: e00005. doi:https://doi.org/10.7554/eLife.00005. PubMed: 23110252.
- 54. Wu H, Mathioudakis N, Diagouraga B, Dong A, Dombrovski L et al. (2013) Molecular Basis for the Regulation of the H3K4 Methyltransferase Activity of PRDM9. Cell Rep 5: 13-20. doi:https://doi.org/10.1016/j.celrep.2013.08.035. PubMed: 24095733.
- 55. Vedadi M, Niesen FH, Allali-Hassani A, Fedorov OY, Finerty PJ Jr. et al. (2006) Chemical screening methods to identify ligands that promote protein stability, protein crystallization, and structure determination. Proc Natl Acad Sci U S A 103: 15835-15840. doi:https://doi.org/10.1073/pnas.0605224103. PubMed: 17035505.
- 56. Campagna-Slater V, Mok MW, Nguyen KT, Feher M, Najmanovich R et al. (2011) Structural chemistry of the histone methyltransferases cofactor binding site. J Chem Inf Model 51: 612-623. doi:https://doi.org/10.1021/ci100479z. PubMed: 21366357.
- 57. Kwon T, Chang JH, Kwak E, Lee CW, Joachimiak A et al. (2003) Mechanism of histone lysine methyl transfer revealed by the structure of SET7/9-AdoMet. EMBO J 22: 292-303. doi:https://doi.org/10.1093/emboj/cdg025. PubMed: 12514135.
- 58. Wilson JR, Jing C, Walker PA, Martin SR, Howell SA et al. (2002) Crystal structure and functional analysis of the histone methyltransferase SET7/9. Cell 111: 105-115. doi:https://doi.org/10.1016/S0092-8674(02)00964-9. PubMed: 12372304.
- 59. Qiao Q, Li Y, Chen Z, Wang M, Reinberg D et al. (2011) The structure of NSD1 reveals an autoregulatory mechanism underlying histone H3K36 methylation. J Biol Chem 286: 8361-8368. doi:https://doi.org/10.1074/jbc.M110.204115. PubMed: 21196496.
- 60. Zheng W, Ibáñez G, Wu H, Blum G, Zeng H et al. (2012) Sinefungin derivatives as inhibitors and structure probes of protein lysine methyltransferase SETD2. J Am Chem Soc 134: 18004-18014. doi:https://doi.org/10.1021/ja307060p. PubMed: 23043551.