The Polycomb Repressive Complex 2 Is a Potential Target of SUMO Modifications

Background The Polycomb Repressive Complex 2 (PRC2) functions as a transcriptional repressor through a mechanism that involves methylation of Histone H3 at lysine 27. The PRC2 complex activity is essential for cellular proliferation, development, and cell fate decisions. PRC2 target genes include important regulators of development and proliferation as well as tumor suppressor genes. Consistent with this, the activity of several Polycomb group (PcG) proteins is deregulated in human cancer suggesting an important role for PcGs in tumor development. Whereas the downstream functions of PcGs are well characterized, the mechanisms of their recruitment to target genes and the regulation of their activity are not fully understood. Principal Findings Here we show that the two PRC2 components SUZ12 and EZH2 are sumoylated in vitro and in vivo. Among several putative sumoylation sites we have mapped the major site of SUZ12 sumoylation. Furthermore, we show that SUZ12 interacts with the E2-conjugating enzyme UBC9 both in vitro and in vivo and that mutation of the SUZ12 sumoylation site does not abolish this binding. Finally, we provide evidence that the E3-ligase PIASXβ interacts and enhances the sumoylation of SUZ12 in vivo suggesting that PIASXβ could function as an E3-ligase for SUZ12. Conclusions Taken together, our data identify sumoylation as a novel post-translational modification of components of the PRC2 complex, which could suggest a potential new mechanism to modulate PRC2 repressive activity. Further work aimed to identify the physiological conditions for these modifications will be required to understand the role of SUZ12 and EZH2 sumoylation in PcG-mediated epigenetic regulation of transcription.

PRC2 is a smaller and highly conserved complex. The core of the PRC2 complex is formed by the three PcG proteins EZH2, EED and SUZ12 and by the histone binding protein RbAp48/46 [13][14][15][16]. The PRC2 complex catalyzes the tri-methylation (me3) of histone H3 on Lysine (K) 27 [13][14][15][16]. The activity of PRC2 is required for the recruitment of PRC1 to target genes [11] through a mechanism that most likely involves the binding of PRC1 to H3K27me3 [15,16]. Although EZH2 is the catalytic component of PRC2, all three PcG components of the PRC2 complex are essential for EZH2 Lysine histone Methyl Transferase (KMT) activity and for mouse embryonic development [3,17].
The expression of EZH2, EED, and SUZ12 is controlled by the pRB/E2F pathway and PRC2 activity is essential for the proliferation of primary and cancer cell lines [3,18]. Consistent with this, direct deregulation of different PcGs have been reported in human cancers [3,19]. We and others have demonstrated that BMI1, CBX7, CBX8 and EZH2 have growth promoting and oncogenic effects, which in part can be ascribed to the ability of PcGs to repress the expression of the tumor suppressor proteins p16 and ARF [9,18,20,21].
Despite the key role of the PcGs in regulating cellular homeostasis, we know relatively little about the mechanisms by which PcG-mediated silencing is established and maintained. Recent studies have shown that AKT-dependent EZH2 phos-phorylation inhbits EZH2 recruitment to chromatin and thereby indirectly its repressive activity [22]. It is likely that other posttranslational modifications could be involved in the regulation of PcG activity. For example, it was shown that CBX4/HPC2, a component of the PRC1 complex, is sumoylated and functions as a SUMO E3 ligase facilitating the SUMO modification of the transcriptional repressor CtBP and of the kinase HIPK2 [23,24].
The conjugation of SUMO to target proteins involves four enzymatic steps after which SUMO is covalently attached by an isopeptidic bond to a lysine residue [25][26][27]. The target lysine is often found in a consensus sequence YKXE/D. A recent study has revealed an extended consensus motif consisting of a cluster of acidic residues downstream of the YKXE/D motif leading to a more accurate prediction of SUMO modification sites [28]. An ATP dependent cascade of the E1 activating enzyme (SAE1-SAE2), the E2 conjugating enzyme (UBC9) and of E3 ligases, covalently binds the C-terminal cystein of the small SUMO protein to the target lysine [26]. Protein modification by sumoylation can have several different outcomes. Reported effects of sumoylation include change of conformation, stability, interactions and localization of the target proteins [25][26][27]29]. Several transcription factors have been reported to be SUMO modified and such modifications mainly correlate with repression of transcription [27,29].
The identification of novel post-translational modifications of PcG proteins could improve our understanding of the mechanisms of PcG-mediated epigenetic regulation of transcription. Here we demonstrate that two subunits of the PRC2 complex (SUZ12 and EZH2) are sumoylated in vitro and in vivo. Moreover, we map the site of sumoylation in SUZ12 and we show that SUZ12 can interact with UBC9 in vivo and in vitro. Furthermore, we have investigated the upstream regulatory pathways to SUZ12 sumoylation and found that the E3 ligase PIASXb, but not CBX4, enhances the sumoylation of SUZ12. Together, our data identifies sumoylation as a novel post-translational modification of different PRC2 components, which could open new possibilities to understand the regulation of PcG activities.

EZH2 and SUZ12 are sumoylated in vivo and in vitro
To understand if members of the PRC2 complex are potential targets for sumoylation, we analyzed the amino acid sequence of SUZ12, EZH2, EED and RbAp48 with the SUMOplot TM Analysis Program (Abgent; www.abgent.com/doc/sumoplot). This analysis revealed that all four proteins contain one or more SUMO consensus motifs (data not shown), suggesting that all components of the PRC2 complex could be targets of sumoylation. Thus, we tested if endogenous SUZ12, EZH2, EED and RbAp48 could be SUMO modified in vivo. Most of the sumoylated proteins reported in the literature present a very low pool of protein conjugated with SUMO [25]. In order to raise the proportion of modified endogenous substrate we transfected 293T cells with SUMO-1 and UBC9 (Fig. 1A). Western blot analysis detected slower migrating forms of SUZ12 and EZH2 upon SUMO and UBC9 expression with a molecular weight consistent with SUMOconjugation (Fig. 1A). EED is present in the cells in at least 4-5 different isoforms [30,31] and it is difficult to determine the existence of SUMO-conjugated EED forms. Despite this, we were unable to detect any modifications of the slowest migrating EED isoforms suggesting that these isoforms are not SUMO modified in vivo (Fig. 1A). Finally, SUMO and UBC9 expression never led to any RbAp48 modifications indicating that RbAp48 does not undergo SUMO conjugation in vivo (Fig. 1A). Together, these data suggest that EZH2 and SUZ12 can be targets for SUMO modifications in vivo.
To confirm the SUZ12 and EZH2 sumoylation, we tested if the components of the PRC2 complex could be SUMO modified in vitro. For this, we reconstituted an in vitro SUMO system by purifying from bacteria recombinant SAE1/SAE2 heterodimer (the E1 activating enzyme), UBC9 (the E2 conjugating enzyme) and SUMO-1. As substrate for the reaction we used in vitro translated ( 35 S-radio-labeled) components of the PRC2 complex. The reaction products were separated by SDS-PAGE and radiolabeled proteins were visualized by autoradiography. As shown in Fig. 1B, SUZ12 can be efficiently attached to SUMO in vitro, while EZH2 cannot (Fig. 1B). EED did not show substantial modifications, but a weak modified band could be detected (Fig. 1B). This may suggest that EED can be sumoylated in vitro with low efficiency. Finally, RbAp48 did not show any form of SUMO modification in agreement with the in vivo data (Fig. 1B).
A possible explanation for the lack of EZH2 sumoylation in vitro could be that EZH2 is not folded correctly, when in vitro translated, and therefore not recognized efficiently by the sumoylation machinery. Alternatively, other components of the PRC2 complex could be required for EZH2 sumoylation. To test this, we performed the in vitro sumoylation reaction using recombinant PRC2 complex purified from insect cells [3] as substrate. As shown in Figure 1C, both SUZ12 and EZH2 were SUMO modified under these reaction conditions, confirming the in vivo results presented in Figure 1A. EED was also sumoylated, confirming that EED can be sumoylated in vitro. Finally, in agreement with the results presented in Figure 1A and 1B, we did not observe any modifications of RbAp48. Together, these results demonstrate that SUZ12 and EZH2 can be SUMO modified both in vitro and in vivo, whereas we have been unable to detect sumoylated forms of EED in vivo and of RbAp48 in vitro and in vivo. In addition, SUZ12 always presents a single SUMO modified form ( Figure 1A-C), suggesting that it contains a single site of sumoylation. Differently, EZH2 shows multiple bands of modification (Fig. 1C), suggesting that EZH2 can have different sites of SUMO modification and or multiple sumoylations on a single site.

Identification of SUZ12 SUMO conjugation site
Since SUZ12 appears to have one single site of SUMO modification, we chose to focus on characterizing this. To test if the slower migrating band of SUZ12 corresponds to a SUMOconjugated form, we expressed Flag-tagged SUZ12 in 293T cells together with SUMO-1 and UBC9. Subsequently, we immunoprecipitated SUZ12 from these lysates using an antibody specific for the Flag tag epitope and analyzed the precipitates by western blotting using SUMO-1 and Flag specific antibodies. As shown in Figure 2A, the slower migrating form appearing when SUMO-1 and UBC9 are expressed, corresponds to SUZ12 conjugated to SUMO.
The prediction of potential SUMO sites in the SUZ12 protein identifies more than 10 different potential SUMO consensus motifs. To identify the sites of SUMO modification experimen-tally, we expressed SUMO-1 and UBC9 together with three different SUZ12 fragments: a N-terminal fragment (amino acids, aa 1-283), a middle fragment (aa 250-550) and a C-terminal fragment (aa 500-739). The different fragments were expressed in 293T cells, immunoprecipitated with a Flag specific antibody and used as substrate for the in vitro sumoylation reaction while immobilized on Flag-beads (Fig. 2B). This experiment demonstrates that only the N-terminal fragment of SUZ12 is SUMOmodified in vitro, and suggests that the site of sumoylation is within the first 283 amino acids (Fig. 2B). This fragment still contains 6 of the predicted SUMO consensus sites, but only the site predicted at lysine (K) 75 contains an upstream acidic patch consisting of an aspartic and a glutamic acid residue (Fig. 2C). This cluster of acidic residues has been shown to function as an extended SUMO recognition motif that enhances the specificity of SUMO conjugation [28]. Interestingly, two additional lysine residues (K72 and K73) are found adjacent to the K75 (Fig. 2C). In order to test if K75 is the site of SUZ12 SUMO modification, we substituted K72, K73 and K75 with arginines (R) by site-directed mutagenesis and tested the ability of the single K75R, the double K72R, K73R and the triple K72R, K73R, K75R mutants to be SUMO modified. As shown in Fig. 2D, mutation of K75 strongly reduces the sumoylation of SUZ12, whereas mutation of K72 and K73 only lead to a slight reduction in sumoylation. This suggests that K75 is the preferential site of modification, however that also K72 and K73 can function as alternative sites for SUZ12 sumoylation. Consistent with this, mutations of all three lysines completely abrogate the sumoylation of SUZ12.
To understand if the same lysines are used as in vivo sumoylation sites, we expressed the different SUZ12 mutants as full-length proteins together with SUMO-1 and UBC9 in 293T cells. As shown in Fig. 2E these experiments showed that the 3 lysines are used as sites of sumoylation in vivo and confirmed that K75 is the major site of SUMO modification, also in vivo. Importantly, the alignment of the amino acid sequences between SUZ12 orthologues from different organisms revealed that K75 and the residues that form the acidic patch are highly conserved throughout evolution (Fig. 2F). In order to exclude any possible SUZ12 sumoylation in the following experiments we decided to use the triple K72R, K73R, K75R mutations and we refer to this mutant as SUZ12 3KR.

SUZ12 interacts with UBC9
Differently from the ubiquitylation pathway, the SUMO conjugating enzyme UBC9 can interact directly with its substrate and, at least in vitro, the activity of E3 ligases is limited to enhancing the efficiency of the sumoylation reaction [32,33]. Therefore, we tested if UBC9 can bind directly to SUZ12. For this, we performed an in vitro binding assay using 35 S-radio-labeled in vitro translated SUZ12 together with recombinant GST-UBC9 or GST-SUMO purified from bacteria (Fig. 3A). As shown in Fig. 3B, UBC9, but not SUMO, can bind to SUZ12 in vitro.
To test if this interaction can occur in vivo, we co-expressed SUZ12 together with UBC9 alone or in combination with SUMO and UBC9. To exclude that K75 (and K72, K73) is involved in the interaction between SUZ12 and UBC9, we also included the SUZ12 3KR mutant in this experiment. As demonstrated in Fig. 3C, UBC9 can bind efficiently to SUZ12, but only when it is co-expressed with SUMO. Moreover, this interaction is not affected by the 3KR mutation (Fig. 3C). These results confirm that SUZ12 can interact with UBC9 in vivo and that the K72, K73 and K75 are not required for this interaction. Furthermore, the fact that efficient UBC9 binding was only observed when co-expressed with SUMO, suggests that loading of SUMO to UBC9 leads to a switch of UBC9 binding affinity for SUZ12.
To better characterize the SUMO modification of SUZ12, we tested if SUMO-SUZ12 can be incorporated into the PRC2 complex. The fact that SUZ12 can be sumoylated in vitro when assembled into the recombinant PRC2 complex (Fig. 1C) supports this possibility. To address this in vivo, we expressed the three PRC2 members EED, EZH2 and SUZ12 together with SUMO and UBC9 in 293T cells. The result shows that the sumoylated form of SUZ12 could be co-immunoprecipitated by an antibody specific to EZH2 demonstrating that SUZ12 sumoylation does not exclude the protein from the PRC2 complex (Fig. 3D).

PIASXb is an E3 SUMO ligase for SUZ12
In vitro sumoylation does not require the presence of E3 ligases, but in a physiological context these proteins play an essential role in regulating post-translational sumoylation of proteins [34,35]. Few proteins have been identified to function as SUMO E3 ligases and these include the members of the PIAS protein family and the PcG protein CBX4. CBX4 is a subunit of the PRC1 complex and since the PRC1 and PRC2 complexes share common target genes and regulatory pathways and since interactions between components of the two complexes have been observed [36], we speculated that CBX4 could function as E3 ligase for SUZ12. For this, we tested if overexpression of CBX4 could enhance SUMO site predicted at lysine 75 and the three different K to R substitutions to inactivate the potential target lysine. The predicted consensus site is highlighted in italics, the putative target residues are indicated in blue, the acidic patch is indicated in green and the different arginine substitutions are indicated in red. (D) In vitro sumoylation assay of the different Flag immobilised SUZ12 mutants shown in panel C, showing that K75 is the preferential site of sumoylation. (E) In vivo sumoylation as in panel A of the different SUZ12 mutants shown in panel C, expressed as full-length SUZ12 proteins, demonstrating that K75 is the preferential SUMO site in vivo. (F) Alignment of the sequences surrounding K75 between different SUZ12 orthologues. The K75 residue is shown in bold. The full-length SUZ12 homology is indicated on the right as percentage (%) of identities. doi:10.1371/journal.pone.0002704.g002 SUZ12 sumoylation in vivo. We performed this experiment either upon expression of both SUMO and UBC9 (Fig. 4A) or upon expression of SUMO alone (Fig. 4B). In both cases CBX4 overexpression did not enhance SUZ12 sumoylation suggesting that CBX4 does not function as a SUMO E3 ligase for SUZ12. Consistent with this, we did not detect any interactions between SUZ12 and CBX4 (Fig. 4C).
Next, we focused our attention on the PIAS protein family. PIAS are proteins with RING-finger-like domains that have been associated with the sumoylation of a number of different targets. First, we tested if any of the five different PIAS proteins (PIAS 1, Xa, Xb, 3 and PIASc) could interact with endogenous SUZ12.
Immunoprecipitations of the different PIAS proteins expressed in 293T cells showed that SUZ12 can interact with all members of the PIAS family (Fig. 4C). This interaction is specific since CBX4 expression did not co-precipitate SUZ12 (Fig. 4C). To test if any of these proteins could enhance endogenous SUZ12 sumoylation, we co-expressed the five different PIAS proteins together with SUMO in 293T cells and assayed the levels of SUZ12 sumoylation (Fig. 4D). This experiment demonstrates that PIASXb expression can specifically enhance the sumoylation of endogenous SUZ12 (Fig. 4D). Together these results suggest that all PIAS proteins have binding affinity for SUZ12, but that only PIASXß can enhance endogenous SUZ12 sumoylation in these conditions.

SUZ12 3KR mutant does not affect localization and PRC2 activity
To identify a physiological function for SUZ12 sumoylation, we first tested if this modification could affect SUZ12 cellular localization. For this, we stably expressed SUZ12 WT or 3KR mutant in U2OS cells and showed by immunofluorecence staining that mutation of SUZ12 sumoylation site did not affect SUZ12 nuclear localization (Fig. 5A). Secondly, we tested if SUZ12 sumoylation is required for the enzymatic activity of the PRC2 complex. For this, we stably expressed SUZ12 WT or 3KR mutant in mouse Suz12 2/2 Embryonic Stem (ES) cells [37] and analyzed H3K27me3 levels by western blotting (Fig. 5B). As shown in Figure 5B, re-expression of both WT and 3KR mutant SUZ12 restores physiological H3K27me3 levels suggesting that SUZ12 sumoylation is not required for PRC2 enzymatic activity in vivo. Finally, we tested if SUZ12 sumoylation could be important for the repressive activity of PRC2. For this we tested if the reintroduction of the 3KR mutant, like wild type SUZ12, could lead to the repression of the PRC2 target gene, Gata4 [6]. Expression analyses by real-time quantitative PCR show that re-expression of SUZ12 WT and 3KR mutant lead to comparable repression of Gata4 expression (Fig. 5C). Taken together these data suggest that, under the tested conditions, SUZ12 sumoylation is not required for SUZ12 localization and the catalytic activity of PRC2.

Discussion
The role of Polycomb group proteins in controlling the expression of a wide range of genes has been extensively studied. These studies have proposed that PcG activity is required for the establishment of specific transcription programs that control cell fate decisions during development [6][7][8]37]. Consistent with this, loss of PcG function results in embryonic lethality during the gastrulation stage [2,3,5,38]. Although much is now known about the processes controlled by Polycomb complexes, the molecular mechanisms that regulate their recruitment and activity are still not well understood. One possibility could be that posttranslational modifications of different PcG proteins may be involved in regulating their activity. For example, phosphorylation of EZH2 mediates its displacement from chromatin and fluctuation of BMI1 phosphorylations during the cell cycle correlates with BMI1 chromatin association [22,39]. The identification of novel post-translational modifications of PcG proteins is an important challenge that may help to better understand the mechanisms that regulate PcG activity.
In this study we demonstrate that two members of the PRC2 complex, SUZ12 and EZH2, can be sumoylated in vitro and in vivo. We have shown that SUZ12 contains one single modification site, that EZH2 contains multiple sites of modification and that EZH2 in vitro modification requires the assembly of the PRC2 complex. Furthermore, we have mapped the SUZ12 sumoylation site to lysine 75, shown that SUZ12 can interact with UBC9 and identified PIASXß as a potential SUMO E3 ligase for SUZ12.
We have not been able to identify a physiological function for the SUMO modification of SUZ12. Our results suggest that SUZ12 sumoylation is not essential for PRC2 activity in ES cells, but the general low abundance of SUMO modified proteins together with the difficulties in identifying processes where SUZ12 and EZH2 undergo physiological sumoylation make this question extremely difficult to address. Our results indicate that SUZ12 sumoylation is not essential for PRC2 activity in ES cells, however the sumoylation of SUZ12 could be required for specific processes and or in other cellular contexts. Moreover, previous data have shown that protein sumoylation can be transient and that this can be required to mediate specific protein-protein interactions [25][26][27]. A possibility could be that sumoylation of PRC2 is required to provide specific binding affinities. Furthermore, it is possible that sumoylation of the entire PRC2 complex may be required for its activity and that the SUZ12 mutation alone may not be sufficient to lead to a clear phenotype.
In this work we have also identified PIASXß as a potential E3 ligase for SUZ12 suggesting that PIASXß could function as a regulator of PcG activity. The two splice variants of PIASX (Xa and Xb) are mainly expressed in adult testis in the Sertoli cells, in the germ cells and in the spermatocytes [40,41]. It is possible that SUZ12 sumoylation may have a role in this specific process and it would be interesting to analyze the PRC2 activity at target genes during spermatogenesis in the absence of functional PIASX proteins. Furthermore, it would also be interesting to analyze the role of SUZ12 sumoylation during spermatogenesis, as well as in other developmental processes, by creating mouse models harboring the SUZ12 3KR mutation.
The fact that K75 is conserved in SUZ12 orthologues with low overall identity (Fig. 2F) suggests that SUZ12 sumoylation could be a conserved feature during evolution. Moreover, the Glutamic (E) and Aspartic acid (D) residues forming the acidic patch Cterminal to the SUZ12 sumoylation site are also highly conserved and, when lost, are substituted with an equivalent acidic residue (E to D substitution in Drosophila melanogaster and in Drosophila pseudoobscura, Fig. 2F). This could suggest that the acidic patch plays an important role in the sumoylation of SUZ12. SUZ12 expression is regulated by the pRB/E2F pathway and is required for cellular proliferation of normal and cancer cell lines [3,18]. Consistent with the role of PcG proteins in cancer formation, the SUZ12 gene locus is translocated in human Endometrial Stromal Sarcomas (ESS) with high frequency leading to the expression of an uncharacterized fusion protein [42]. Recently it was further proposed that the selection for the expression of the SUZ12 fusion protein in ESS occurs during the benign to malignant transition of these tumors [43]. It is interesting to note that the part of SUZ12 that is involved in this translocation exclude K75 suggesting that the physiological functions of SUZ12 sumoylation will be lost in cancers harboring this translocation. This does not link SUMO directly with the oncogenic effect of this translocation, but highlights that, if SUMO has a role in regulating PRC2 activity, this will be lost in the development of ESS expressing this translocation. Therefore the understanding of the biological role of SUZ12 sumoylation could also contribute to the understanding of the role of this translocation in ESS.
It was reported that a small proportion of the PRC2 complex is present in the cytoplasm of different cell types [44]. In this work it was shown that cytosolic PRC2 regulates Actin polymerization through its ability to bind the guanine nucleotide exchange factor VAV1 and that this activity is linked to Ezh2 requirement for T-cell development in vivo [44]. An additional possibility could be that PRC2 sumoylation may play a role in regulating non-chromatin associated PRC2 functions. The fact that we detect sumoylation of nuclear SUZ12 upon SUMO and UBC9 overexpression (data not shown) does not support this but further experiments in this direction will be required to validate such hypothesis.
In conclusion, we have identified a novel potential posttranslational modification of components of the PRC2 complex, which may lead to a better understanding of the mechanisms of PcG-mediated regulation of transcription during processes such as proliferation, development, cell fate decision and tumorigenesis.

Plasmids and antibodies
All constructs encoding the PRC2 complex components have been described previously [3]. All constructs for SUMO, UBC9, SAE1/SAE2 expression were described [45]. Expression constructs for PIAS proteins were described [46]. SUZ12 3KR mutations were obtained by site-directed mutagenesis using the Quick-Change Site Directed Mutagenesis kit (Stratagene). SUZ12 WT and 3KR ORFs were inserted into the HA/Flag EF1-ires-PURO lentiviral expression construct [47] by BamHI subcloning from the pCMV expression constructs described above.

Western blotting and immunoprecipitations
For Western blotting cells were lysed in high salt buffer S300P (50mM Tris-HCl, 300mM NaCl, 0.5% Igepal, 1mM EDTA, 1mM DTT, 1mM PMSF, 1 mg/ml leupeptin, 1 mg/ml aprotinin). For immunoprecipitations, protein G-agarose beads (Zymed) were pre-coupled O/N with the indicated antibodies. Alternatively, mouse anti-Flag M2 beads (Sigma) were used when indicated. Equal amounts of protein lysates (S300P buffer) were used for each immunoprecipitation. Immunoprecipitates were eluted from beads and analyzed by western blotting with the indicated antibodies.

Expression of recombinant proteins and in vitro sumoylation assay
GST-UBC9, GST-SUMO-1 and heterodimeric GST-SAE1-SAE2 fusion proteins were expressed in E. coli BL21 for 4 hours at 37uC by addition of 0.2 mM IPTG (Sigma). Recombinant proteins were purified on glutathione-Sepharose beads (Amersham). SUZ12, RbAp48, EED and EZH2 in vitro translation was performed with TNTH coupled reticulocyte lysate system (Promega) in the presence of [ 35 S]Methionine (Promix, Amersham). The recombinant PRC2 complex expressed in insect cells was prepared as described [3].
The sumoylation reaction was carried out as follows in a total volume of 20 ml at 30uC for 2 hours: 56reaction buffer (500 mM Tris pH 7.5, 50 mM MgCl 2 , 20 mM ATP, ddH 2 O) 2-5 ml 35 Slabeled in vitro translated target protein or 0.1 mg of recombinant PRC2 complex, 1 mg UBC9, 2 mg SUMO-1 and 0.4 mg SAE1-SAE2 recombinant proteins. The reaction was analyzed by autoradiography for radiolabeled substrates or by western blotting.

Quantitative PCR and primers
cDNA preparation and real-time quantitative PCR were performed following manufacturer's instructions (Applied Biosystems). The analysis was performed as described [18]. The primer sequences used for the real-time quantitative PCR were described previously [37].