The nuclear hormone receptor co‐activator CARM1 has the potential to methylate histone H3 at arginine residues in vitro. The methyltransferase activity of CARM1 is necessary for its co‐activator functions in transient transfection assays. However, the role of this methyltransferase in vivo is unclear, given that methylation of arginines is not easily detectable on histones. We have raised an antibody that specifically recognizes methylated arginine 17 (R17) of histone H3, the major site of methylation by CARM1. Using this antibody we show that methylated R17 exists in vivo. Chromatin immunoprecipitation analysis shows that R17 methylation on histone H3 is dramatically upregulated when the estrogen receptor‐regulated pS2 gene is activated. Coincident with the appearance of methylated R17, CARM1 is found associated with the histones on the pS2 gene. Together these results demonstrate that CARM1 is recruited to an active promoter and that CARM1‐mediated R17 methylation on histone H3 takes place in vivo during this active state.
Covalent post‐translational modifications of histone N‐termini, such as acetylation, phosphorylation and methylation, play a fundamental role in chromatin structure and transcriptional regulation (Strahl and Allis, 2000; Jenuwein and Allis, 2001; Marmorstein, 2001). Although histone methylation was first discovered more than 35 years ago (Murray, 1964), only recent studies have started to elucidate its biological significance. Recent work on methylation of lysines has shown a role for this modification in heterochromatic silencing (Rea et al., 2000; Bannister et al., 2001; Lachner et al., 2001; Noma et al., 2001) and euchromatic gene expression (Nielsen et al., 2001). The function of methylation of arginine residues is much less clear. This may be partly because unlike lysine methylation, which is found to occur in vivo on histones H3 and H4 (Strahl et al., 1999; Strahl and Allis, 2000), the occurrence of arginine methylation has been difficult to detect on mammalian histones (Gary and Clarke, 1998).
Different sets of enzymes carrying out either arginine or lysine methylation activity have been identified recently (Chen et al., 1999; Rea et al., 2000; Stallcup, 2001). Protein arginine (R) methyltransferases (PRMTases) share a highly conserved domain encompassing the methyltransferase activity. They have a broad spectrum of substrates, including RNA‐processing proteins, RNA‐transporting proteins, protein phosphatase 2A, G‐proteins and histones (Aletta et al., 1998). Three of the five mammalian PRMTases, PRMT1, JBP1 and CARM1 [cofactor associated arginine (R) methyltransferase 1], have been demonstrated to have histone methyltransferase activity (McBride and Silver, 2001; Stallcup, 2001). PRMT1 specifically methylates arginine 3 of histone H4 in vivo (Strahl et al., 2001; Wang et al., 2001). CARM1 was isolated through a yeast two‐hybrid screen searching for proteins that interact with the p160 nuclear hormone receptor co‐activator GRIP1 (Chen et al., 1999). Recombinantly expressed CARM1 preferentially methylates histone H3 within a bulk histone preparation in vitro (Chen et al., 1999). Both PRMT1 and CARM1 co‐activate nuclear hormone receptor regulated gene expression in transient transfection assays, and their co‐activator potential is dependent on an intact methyltransferase domain (Chen et al., 1999; Koh et al., 2001; Wang et al., 2001). These studies suggest that methylation of arginines within histone tails may represent an activating step in mammalian gene transcription.
A recent study revealed that PRMT1 also methylates the transcription factor STAT1 and that this methylation is required for STAT1‐mediated transcriptional activation (Mowen et al., 2001). This raises the question whether histones are indeed the major physiological substrate of arginine methyltransferases or whether methylation of other proteins in the transcription complexes results in a transcriptionally active promoter status, as it has already been found to be the case for acetylation (Kouzarides, 2000).
To evaluate the link between histone arginine methylation by CARM1 and active gene expression, we generated an antibody that specifically recognizes the major site of CARM1 methylation in histone H3 in vitro. This antibody allowed us to establish that this modification takes place in vivo in mammalian cells. We use chromatin immunoprecipitation analysis (ChIP) to demonstrate for the first time that an arginine methyltransferase (CARM1) is indeed recruited to an endogenous target promoter upon gene activation, a process which coincides with the appearance of arginine methylation on histone H3.
Arginine 17 on histone H3 is methylated in vivo
The nuclear hormone receptor co‐activator CARM1 is able to methylate arginine residues within histone H3 in vitro. However, evidence that such methylation takes place in vivo has been difficult to obtain (Gary and Clarke, 1998; Stallcup, 2001). We therefore raised an antibody that recognizes the methylation site for CARM1 in histone H3, to establish (i) if such methylation takes place on CARM1 regulated promoters in vivo and (ii) whether this methylation occurs when a gene is actively transcribed.
The major site for CARM1 methylation was mapped, by radiosequencing analysis of 3H‐methylated recombinant H3, to arginine 17 (R17) (Figure 1A). A peptide methylated at R17 was used to immunize rabbits. Figure 1B shows that the resulting antibody (Me‐R17H3) recognizes recombinant histone H3 only when it is methylated by CARM1 (compare lanes 1 and 2). Purified calf thymus histone H3 is also recognized by this antibody (lane 3) indicating that R17 is indeed methylated in vivo. The Me‐R17H3 antibody does not recognize recombinant histone H4 methylated by PRMT1 or when H4 is purified from calf thymus (lanes 4–6), indicating that the antibody only recognizes methyl arginine in the correct amino acid context (i.e. methyl R17 of histone H3).
A recent study (Schurter et al., 2001) reports that arginines in the C‐terminal domain of H3 are methylated by CARM1 in vitro. To confirm the specificity of the Me‐R17H3 antibody for R17 and the N‐terminal tail of H3, we analysed tail‐less histone H3 methylated in vitro by CARM1 and showed that this antibody does not recognize the C‐terminal methylation sites of CARM1 (data not shown). The high specificity for methylated R17 on H3 was further confirmed by peptide competition with unmethylated H3, R17 methylated H3 and R3 methylated H4 peptides (Figure 1C). When total U2OS cell extract is probed with antiMe‐R17H3, the only protein recognized is histone H3 (Figure 1D). Taken together these results indicate for the first time that methylation at R17 in histone H3 indeed occurs in vivo.
The methyltransferase CARM1 co‐activates the ER‐regulated pS2 promoter in a methyltransferase activity dependent manner
To establish whether methylation on R17 takes place on a promoter regulated by CARM1, we chose to study the promoter of the estrogen receptor regulated pS2 gene. It has recently been demonstrated that p160 co‐activators are recruited to the pS2 gene promoter in vivo (Shang et al., 2000). As shown in Figure 2A, this promoter can be stimulated by CARM1 in transient transfection assays in a manner dependent on the SRC1 co‐activator to which CARM1 binds. To analyse whether the methyltransferase activity of CARM1 is necessary for this stimulation, a CARM1 mutant was generated that lacks the methyltransferase domain (CARM1 284–608). This mutant is not able to stimulate the pS2 gene promoter (Figure 2A) even though it is still able to bind SRC1 with an affinity comparable to full‐length CARM1 (Figure 2B). These data identify the pS2 gene as a target for CARM1 co‐activation.
pS2 gene activation is correlated with CARM1 recruitment and histone H3 methylation at arginine 17
The Me‐R17H3 antibody and an antibody against CARM1 were used in chromatin immunoprecipitation analysis to probe the role of arginine methylation by CARM1 in vivo at the estrogen‐dependent pS2 gene in the human breast cancer cell line MCF‐7. The pS2 gene is a well‐characterized target of the estrogen receptor in these estrogen responsive cells (Shang et al., 2000). We first used our Me‐R17H3 antibody to monitor changes in R17 methylation by ChIP analysis following CARM1 transfection. However, these changes were minimal, probably as a result of the low stimulatory effect of transfected CARM1 on this promoter (Figure 2A). We therefore decided to monitor the changes in chromatin methylation at the pS2 promoter after the gene is activated by the addition of two stimuli: estradiol (E2) and tetradodecanoyl phorbol acetate (TPA) (Figure 3). When each of these stimuli is added independently to MCF‐7 cells, transcription of the endogenous pS2 gene is slightly induced, as shown by northern blot analysis (Figure 3A). However, when added together, E2 and TPA have a synergistic effect on mRNA levels (Figure 3A).
When the Me‐R17H3 and CARM1 antibodies are used in ChIPs, Figure 3B shows that the chromatin around the pS2 promoter has no detectable R17 methylation or CARM1 association under non‐induced conditions (lane 1). In contrast, when pS2 gene expression is stimulated by a combination of E2 plus TPA, methylation of R17 and recruitment of CARM1 are both dramatically increased (Figure 3B, lane 4). Individually, each of the stimuli have a small effect on methylation and CARM1 association, consistent with the small increase seen on mRNA levels.
In order to ensure that the enhanced histone H3 methylation associated with the stimulated pS2 promoter did not reflect a global nucleosome modification, we examined the methylation of a gene promoter that is not an estrogen receptor target, the cdc25 gene promoter. ChIP analysis of this promoter shows a very weak R17 methylation of H3 but there is no alteration in the methylation pattern when comparing unstimulated and stimulated MCF7 cells (Figure 3B). These data strongly argue that methylation at R17 is highly specific and takes place only on selected promoters, for example the activated pS2 gene, and is consistent with the recruitment of the CARM1 methyltransferase to the promoter.
The ability of CARM1 to methylate histone H3 in vitro and to co‐activate nuclear hormone receptor regulated transcription in transient transfection (Chen et al., 1999) supports the idea that arginine methylation of histones may have a stimulating effect on transcription. Our results show for the first time that an arginine methyltransferase is recruited to an endogenous promoter and that this event correlates with histone methylation and activation of transcription in vivo.
In contrast to a recent study, which showed that R17 and R26 are approximately equally methylated by CARM1 in vitro (Schurter et al., 2001), our sequencing reveals R17 of histone H3 as the major methylation site for CARM1. The usage of different experimental conditions and enzyme preparations might account for the variations in the sequencing results.
To probe the role of arginine methylation by CARM1 in vivo, we raised an antibody that specifically recognizes the major CARM1 methylation site in histone H3, R17, based on our sequencing results. This antibody detects arginine methylation on histone H3 in vivo, even though methylation has previously been difficult to detect by the sequencing of bulk purified histones.
Only recently Schurter et al. (2001) reported the in vivo occurrence of arginine methylation of H3 by amino acid sequencing. However, our results go beyond the previous data because we show site specificity of an arginine methyltransferase on histones in vivo using antibodies.
This raises a cautionary note for the assumption that histone sequencing recognizes all the modifications that take place in vivo. It may well be that many modifications are present on a small subset of promoter‐bound histones, or that the modifications will only occur on the induced conditions. In either case the amounts of the modifications present on purified histones may be too small to be detected by sequencing. This then raises the possibility that many more modifications exist on histones than are currently suspected.
It has been shown recently that another arginine methyltransferase, PRMT1, methylates R3 in histone H4 and that this modification can be detected in vivo (Strahl et al., 2001; Wang et al., 2001). However, the role of this modification in transcriptional regulation has not been clarified so far. With the aid of ChIP analysis we have demonstrated for the first time that recruitment of the CARM1 methyltransferase onto chromatin, and concomitant R17 methylation on histone H3 takes place when an endogenous, hormonal responsive gene is activated.
Analysis of the cdc25 control gene confirms that the estrogen induced increase in R17 methylation by CARM1 is gene specific. The fact that a cell cycle regulated gene shows a low level and constant amount of methylated R17 in its promoter opens the possibility of a broader role for arginine methylation in transcription regulation.
These data are consistent with the notion that CARM1 methyltransferase activity is essential for co‐activated transcription, as predicted by transfection experiments using CARM1‐methyltransferase deficient mutants (Figure 2A; Chen et al., 1999). Our results clearly establish that CARM1 is indeed an arginine methyltransferase whose activity is directed towards histones. The possibility also exists that CARM1 may methylate non‐histone substrates such as transcription factors. Indeed, recent data show that PRMT1 is able to methylate the transcription factor STAT1 (Mowen et al., 2001), establishing that arginine methylation of non‐histone proteins also has an important role in regulation of gene expression.
The combination of two inducers, E2 and TPA, is necessary for both maximal R17 methylation and maximal levels of mRNA expression from the pS2 gene. The mechanism of this cooperation is unclear. The effect of E2 stimulation is likely to be the recruitment of the p160 co‐activator (Shang et al., 2000). The TPA stimulus may affect a distinct signalling pathway which is necessary for maximal activity. It has been postulated that pre‐existing modifications on histones may influence further modifications (Strahl and Allis, 2000), e.g. phosphorylation of S10 in histone H3 promotes acetylation of the neighbouring K14 by GCN5 (Lo et al., 2000; Berger, 2001). Recent studies, which show cooperativity between CBP/p300 and CARM1 in nuclear hormone receptor signalling, suggest that there might be a similar cross‐talk between arginine methylation and lysine acetylation (Chen et al., 1999). Interestingly, a similar cooperativity between E2 and TPA observed here for arginine methylation has been observed for acetylation of histone H3 at the pS2 promoter (Sewack et al., 2001). This opens up the possibility that the cooperativity may involve a cross‐talk between acetylation and arginine methylation.
It has recently become clear that the covalent modifications of histones is of central importance in the transcription regulation field. Our understanding of how these modifications are regulated is crucial towards our understanding of differential gene expression involved in many cellular processes. Our data give new insights into how arginine methylation of histones is involved in gene activity. They provide direct evidence that CARM1‐mediated methylation at R17 of histone H3 takes place during transcriptional activation in vivo.
Plasmids and recombinant proteins.
Full‐length CARM1 cDNA and the mutant CARM1 cDNA fragment (284–608) deleted of its 283 first amino acids containing the methyltransferase domain, were generated by PCR and cloned into pcDNA3‐Gal4DBD. The ERα‐expression plasmid (pMT2‐MOR) and the SRC1e expression plasmid (pSG5‐SRC1e) were gifts from David Heery and have been described previously (Kalkhoven et al., 1998). The pS2 gene promoter CAT reporter (pS2 CAT) and pGEX2TK‐SRC1a (984–1441) constructs were gifts from Malcolm Parker. GST and GST–SRC1 fusion proteins were expressed in and purified from Escherichia coli XA90 according to standard procedures.
Cell culture, transfections and reporter gene assay.
MCF‐7, HEK‐293 and U2OS cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum at 37°C and 5% CO2. HEK‐293 cells were transfected using the calcium phosphate co‐precipitation method with 2 μg of pS2 CAT reporter plasmid, 1 ng pMT2‐MOR, 1 μg pSG5‐SRC1e and 0.5 μg of either wild‐type full‐length CARM1 (CARM1 fl) or mutant CARM1 (CARM1 284–608) Gal4DBD‐expression plasmids. Approximately 36 h post‐transfection, the cells were washed in PBS and harvested. CAT assays were performed as described previously (Hagemeier et al., 1993).
Histone methyltransferase assays and protein sequencing.
Two micrograms of either recombinant histone H3 or recombinant H4 were incubated with either purified GST–CARM1 or GST–PRMT1 and [3H‐Me]S‐adenosyl methionine (NEN, 80 Ci mmol−1) in PBS at 30°C for 1 h. The reaction products were then resolved by SDS–PAGE and blotted onto nitrocellulose. For N‐terminal sequencing, radiolabelled recombinant histone H3 was blotted to PVDF membrane and sequenced by Edman degradation (Protein Sequencing Facility, University of Cambridge, UK). Amino acid fractions were analysed for the presence of tritium by scintillation counting. In addition, non‐radioactively methylated recombinant H3 was sequenced by Edman degradation and amino acid content in each fraction determined by HPLC.
Rabbits were immunized with a histone H3 N‐terminal peptide corresponding to amino acids 11–24, in which arginine 17 was asymmetrically dimethylated. Immunoreactive serum was applied sequentially to a H3 R17‐unmethylated peptide and a H3 R17‐methylated peptide column to affinity purify specific antibodies, since the antiserum crossreacted with unmethylated recombinant histone H3. The antibody is available from Abcam.com.
In vitro translations and GST pull‐down assays.
In vitro‐translated 35S‐labelled wild‐type full‐length (CARM1 fl) or mutant CARM1 (CARM1 284–608) were incubated with equal amounts of GST fusion proteins essentially as described previously (Hagemeier et al., 1993). Bound proteins were resolved by SDS–PAGE and viewed by fluorography.
Chromatin immunoprecipitation (ChIP) and northern blot analysis.
MCF‐7 cells were grown to 95% confluence in Phenol Red‐free DMEM supplemented with 5% charcoal‐dextran‐stripped fetal calf serum and 1 nM 4‐hydroxy‐tamoxifen for at least 3 days. Following the addition of DMSO/ethanol as control (no induction), 200 nM 17β‐estradiol (E2), or 100 ng/ml tetradecanoyl phorbol acetate (TPA), or both (E2+TPA) for 3 h, cells were either harvested for northern blot analysis, or cross‐linked with 1% formaldehyde (Sigma) at room temperature for 15 min. Cells were rinsed twice with ice‐cold PBS (pH 7.4) and collected in PBS and centrifuged for 5 min at 2000 g. ChIPs were then performed as described (Dedon et al., 1991; Orlando et al., 1997) with 20 μl of antiMe‐R17H3 and 2.5 μl of antiCARM1 (Upstate) antibody. Immunoprecipitates were analysed for the presence of pS2 promoter fragment by PCR using previously described primers (Mazumdar et al., 2001) that amplify the promoter region –463 to –159, which includes the ER‐responsive element. PCR products were resolved in a 2% agarose gel and visualized with ethidium bromide.
For northern blot analysis 20 μg of total MCF‐7 RNA per sample were electrophoresed in a 1.2% formaldehyde‐agarose gel and transferred to a nylon membrane (Amersham Pharmacia). The pS2 and GAPDH RNA were detected using a 32P‐labelled DNA probe from the corresponding gene sequence and subsequent autoradiography.
We thank Mike Weldon for Edman degradation of labelled proteins, Ulla Hansen for providing MCF‐7 cells, Michael Stallcup for providing the pGEX4T1‐CARM1 plasmid, Malcolm Parker and Roger White for the pGEX2TK‐SRC1a (984‐1441) and the pS2 CAT plasmid and David Heery for the pMT2‐MOR and the pSG5‐SRC1e plasmid. We would like to thank Andy Bannister for suggestions and critical reading of the manuscript. S.D. and S.J.N. were funded by a grant from the Cancer Research Campaign (SP2081‐0104) and U.M.B. by an HFSP grant (RG‐196/98). K.N is supported by the Wellcome Trust.
- Copyright © 2002 European Molecular Biology Organization