|Home | About | Journals | Submit | Contact Us | Français|
The transcriptional coactivator p/CIP(SRC-3/AIB1/ACTR/RAC3) binds liganded nuclear hormone receptors and facilitates transcription by directly recruiting accessory factors such as acetyltransferase CBP/p300 and the coactivator arginine methyltransferase CARM1. In the present study, we have established that recombinant p/CIP (p300/CBP interacting protein) is robustly methylated by CARM1 in vitro but not by other protein arginine methyltransferase family members. Metabolic labeling of MCF-7 breast cancer cells with S-adenosyl-L-[methyl-3H]methionine and immunoblotting using dimethyl arginine-specific antibodies demonstrated that p/CIP is specifically methylated in intact cells. In addition, methylation of full-length p/CIP is not supported by extracts derived from CARM1−/− mouse embryo fibroblasts, indicating that CARM1 is required for p/CIP methylation. Using mass spectrometry, we have identified three CARM1-dependent methylation sites located in a glutamine-rich region within the carboxy terminus of p/CIP which are conserved among all steroid receptor coactivator proteins. These results were confirmed by in vitro methylation of p/CIP using carboxy-terminal truncation mutants and synthetic peptides as substrates for CARM1. Analysis of methylation site mutants revealed that arginine methylation causes an increase in full-length p/CIP turnover as a result of enhanced degradation. Additionally, methylation negatively impacts transcription via a second mechanism by impairing the ability of p/CIP to associate with CBP. Collectively, our data highlight coactivator methylation as an important regulatory mechanism in hormonal signaling.
Nuclear hormone receptors are a superfamily of ligand-dependent transcription factors that mediate the effects of small lipophilic hormones such as estrogen and retinoic acid. Chromatin-based assays examining promoter occupancy of nuclear receptor (NR)-dependent genes have shown that transcriptional activation involves recruitment of a large repertoire of coactivators that interact with liganded NRs in a coordinated fashion (42). Among the first coactivators recruited in response to hormone are the steroid receptor coactivator (SRC) proteins that include the p300/CBP interacting protein (p/CIP), also known in human form as SRC-3/AIB1/ACTR/RAC3/TRAM1 (2, 8, 32, 51, 53), steroid receptor coactivator 1 (SRC-1) (23, 41), and steroid receptor coactivator 2 (SRC-2), also known as GRIP1/TIF2 (22, 56). p/CIP is a well-established NR coactivator which has been shown to mediate the transcriptional effects of liganded NRs in a variety of contexts in vitro as well as in vivo (34, 53). Downregulation of p/CIP expression in intact cells has been shown to inhibit cell growth (35, 52, 65, 66), and the targeted disruption of the p/CIP gene in mice has demonstrated a requirement for p/CIP in mammary gland development as well as several nuclear receptor-dependent and -independent pathways (4, 30, 59, 61). Importantly, the human homologue of p/CIP is amplified in a percentage of breast and ovarian cancers (2), and overexpression of p/CIP has been shown to initiate tumor development in mice (54).
Biochemical and functional studies have shown that p/CIP functions as a bridging factor that promotes the assembly of transcriptional complexes essential for NR-mediated transactivation. This assembly is mediated by several domains within the p/CIP protein that are conserved in SRC-1 and GRIP/TIF2 (17). The amino terminus contains a highly conserved basic helix-loop-helix and Per/ARNT/SIM domain which functions as an interaction surface for a wide variety of proteins (24). The central region of the SRC proteins encodes the receptor interacting domain consisting of three leucine-rich motifs (also known as NR boxes) with the consensus amino acid sequence LXXLL (where X is any amino acid and L is leucine). The LXXLL motifs are responsible for making direct contact with liganded NRs (20, 37, 53, 55). The carboxy terminus of p/CIP contains two transcriptional activation domains (AD1 and AD2) which serve as interaction surfaces for the recruitment of two major classes of coactivators. AD1 interacts directly with the acetyltransferases CBP and p300 (8, 53) and with GCN5/PCAF(1, 5, 27, 38). AD2 serves as an interaction surface for a second class of enzymatic proteins, known as protein arginine methyltransferases (PRMTs) (3). Eight mammalian PRMTs have been identified, although the coactivator arginine methyltransferase (CARM1/PRMT4), PRMT1, and, more recently, PRMT2 appear to be the most relevant with respect to signaling by NRs (28, 43). Cotransfection assays using transient or stably integrated reporter genes have shown that individual SRC proteins can synergize with histone acetyltransferases and methyltransferases to activate NR-dependent transcription (26, 50, 57). CARM1 and PRMT1 can methylate specific arginines on histones H3 and H4, respectively, and several studies have demonstrated a correlation between methylation of histone H3, recruitment of CARM1, and activation of several steroid-responsive genes (11). Disruption of the CARM1 gene in mice results in embryonic lethality (64), and estrogen receptor (ER)-dependent signaling is compromised in the CARM1 null cells. Collectively, these studies suggest that recruitment of CARM1 represents an essential activating step in mammalian gene transcription. The recent discoveries of lysine and arginine demethylases indicates that methylation, similar to acetylation (49, 58), represents a dynamic and reversible modification and adds another level of complexity to our understanding of transcriptional regulation. Importantly, arginine methylation is not restricted to posttranslational modification of histones. There is mounting evidence supporting the hypothesis that in addition to methylating histone H3, the physiological role of CARM1 may be dependent on methylation of nonhistone proteins. For example, CARM1 specifically methylates CBP/p300, and mutation of the methylation sites in CBP abolishes steroid-dependent transcriptional activation (10, 31, 62). In addition to histone H3 and p300/CBP, CARM1 methylates proteins involved at several other critical points in the transcriptional process including the poly(A)-binding protein (PABP1) (29), RNA binding molecules involved in transcript stability (HuR and HuD) (15), and the core splicing factor SmB (12).
In the present study we have characterized the association between p/CIP and the methyltransferase protein CARM1. We demonstrate that the association between p/CIP and CARM1 occurs on target genes in response to hormone and is not detectable in solution following immunopurification of p/CIP. In addition, we show that p/CIP is a substrate for CARM1 in vitro and in vivo. Using a combination of mass spectrometry and molecular mapping, we have identified several CARM1-dependent methylation sites in p/CIP. At least three sites are found within a highly conserved glutamine-rich region of the carboxy-terminus of p/CIP, and additional sites are also located in the central region of p/CIP. Our results suggest that p/CIP methylation by CARM1 may have a dual role in regulating p/CIP function. First, arginine methylation appears to reduce protein stability and causes an increase in p/CIP turnover. Second, we demonstrate that arginine methylation represents a critical regulatory signal involved in coordinating the dissociation of CBP and p/CIP. Collectively, the results of our study suggest that arginine methylation represents an important posttranslational modification which may be important for regulating coregulator stability as well as the dynamics of nuclear receptor-mediated transcription.
The cDNAs encoding full-length CARM1 were kindly provided by Michael Stallcup (University of California-Los Angeles), and PRMT1 cDNA was provided by Harvey Herschman (University of California-Los Angeles). The baculovirus vectors for p/CIP, p/CIP deletion mutants, CARM1, CBP, and PRMT1 were subcloned into a Fastbac vector using standard cloning techniques. For the Gal4 transfection assays, p/CIP fragments were generated by PCR using PCR Supermix High Fidelity (Invitrogen) and subcloned in frame into a PMgal (Stratagene) expression vector. Antibodies against CARM1 and CBP were purchased from Santa Cruz Biotechnologies. Antibody against human ERα was obtained from Bethyl Laboratories. Anti-p/CIP antibody was generated as previously described (44). The experiments shown in Fig. Fig.3C3C and Fig. Fig.77 were performed using an anti-SRC-3 antibody obtained from Bethyl Laboratories. αmR17 antibodies were purchased from UpState. The p/CIPR3A cDNA (containing substitutions of arginine to alanine at amino acids 1178, 1184, and 1195) and the p/CIPR6A (containing substitutions of arginine to alanine at amino acids 1163,1168, 1170, 1178,1184, and 1195) were generated by site-directed mutagenesis of p/CIP using a Quickchange kit (Stratagene) according to the manufacturer's instructions. S-adenosyl-L-[methyl-3H]methionine ([3H]sam) (80 Ci/mmol) was purchased from Amersham and [35S]methionine (419 mBq/ml) was purchased from Perkin Elmer. Biotin-labeled peptides were obtained commercially from Biosynthesis Incorporated.
HeLa, COS-1, CV-1, MCF-7, U2OS, CARM1+/+, and CARM1−/− mouse embryonic fibroblasts (MEFs) were maintained in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 0.1 U/ml penicillin G, and 0.1 μg/ml streptomycin sulfate in 5% CO2 at 37°C. To assess transcriptional activity of Gal4-p/CIP fusion proteins, transient transfections were performed using Lipofectamine 2000 according to the manufacturer's instructions (Invitrogen). COS-1 cells were seeded at approximately 180,000 cells/well in 24-well plates. A total of 100 ng of the pMGal vector (cytomegalovirus) directing the expression of either the Gal4 DNA binding domain (DBD) alone and either p/CIP deletions or mutants fused in frame with Gal4 DBD were cotransfected with 250 ng of luciferase reporter plasmid containing five binding sites for Gal4 upstream of a minimal promoter. Renilla luciferase (0.5 ng) was used to monitor transfection efficiency, and total DNA was adjusted to 500 ng per well with pBKS vector. Forty-eight hours after transfection, cells were harvested, and cell extracts were prepared in 75 μl of 1× reporter lysis buffer (Promega). Ten-microliter aliquots of the extract were used for subsequent determination of luciferase and Renilla activity according to the manufacturer's instructions (Promega). Luciferase values were normalized to the output of the internal Renilla control plasmid, and the relative activation (n-fold) was calculated relative to the level of pMGal alone, which was arbitrarily set to 1. Luciferase activities are shown as the means and standard errors of at least three independent experiments performed in triplicates.
To analyze the expression levels of Gal4 fusion proteins, 4 μg of each protein was transfected with Lipofectamine 2000 into COS-1 cells according to the manufacturer's protocol (Invitrogen). Prior to transfection, COS-1 cells were seeded at 90% confluence in six-well plates. Forty-eight hours posttransfection, cells were harvested and lysed with radioimmunoprecipitation assay lysis buffer (50 mM Tris, pH 8.0, 1 mM EDTA pH 8.0, 150 mM NaCl, 0.1% sodium dodecyl sulfate [SDS], 1.0% NP-40, 0.5% sodium deoxycholate, and protease inhibitors). Bradford protein assays were performed to determine the concentration of whole-cell extracts. Twenty micrograms of each extract was subjected to SDS-polyacrylamide gel electrophoresis (PAGE) followed by Western blotting with anti-Gal antibody (Santa Cruz). Antitubulin antibody (Sigma) was used as loading control.
To assess the significance of methylation in estrogen signaling, cells were maintained in phenol red-free DMEM supplemented with 10% charcoal-stripped FBS for 48 h. A day before transfection, cells were seeded at ~75,000 cells/well in 24-well plates. Twenty-five nanograms of wild-type p/CIP and of p/CIPR6A cDNA was cotransfected with 200 ng of a luciferase reporter containing three copies of an estrogen response element (ERE) upstream of a minimal promoter and 10 ng of ER. For some experiments, increasing amounts of wild-type p/CIP or p/CIPR6A (50, 100, and 200 ng) were cotransfected with 200 ng of a pS2 luciferase reporter and 10 ng of ER. Total DNA was adjusted to 800 ng/well with empty vector, and transfections were performed using Lipofectamine 2000 according to the manufacturer's protocol (Invitrogen).
Twenty-four hours after transfection, cells were stimulated with 1 μM β-estradiol for 16 h. Cells were lysed with 75 μl of 1× reporter lysis buffer (Promega), and a 20-μl aliquot of the whole-cell extract was used for luciferase activity analysis. Luciferase activities are shown as the means ± standard errors of three independent samples.
Cells were washed twice in phosphate-buffered saline (PBS), harvested, and lysed in lysis buffer (~300 μl/10-cm plate) consisting of 20 mM Tris (pH 7.9), 300 mM KCl, 0.1% NP-40, 10% glycerol, 0.1 mM dithiothreitol (DTT), 0.5 mM EDTA, 0.5 mM EGTA, and protease inhibitor cocktail. Extracts were then centrifuged for 10 min at 16,000 × g at 4 C, and the soluble extracts were retained. Samples were normalized for protein content, followed by the addition of the appropriate antibody for 2 h on ice. Samples were then diluted to 500 μl with lysis buffer, and 50 μl of a 50% slurry (vol/vol) of protein A/G agarose was added overnight with rocking. The antibody complexes were washed extensively with lysis buffer, and the bound complexes were separated by SDS-PAGE, transferred to nitrocellulose membrane, and blocked overnight in PBS containing 0.1% Tween 20 and 5% nonfat dried milk. The appropriate antibodies were then diluted in blocking buffer, and the membrane was probed for 2 h at room temperature with rocking, followed by the appropriate secondary antibody for 1 h. Proteins were detected using enhanced chemiluminescence according to the manufacturer's recommendations (Amersham). For experiments involving the analysis of 35S-labeled proteins, immunoprecipitated proteins were separated by SDS-PAGE and analyzed by fluorography. Gels were then treated with ENHANCE, dried, and exposed for several days to film. For the experiments shown in Fig. Fig.3C3C and and7,7, cells were grown for 8 days in the presence of AdOx prior to immunoprecipitation of p/CIP.
Nuclear extracts were dialyzed against buffer A (20 mM Tris [pH 7.9], 0.5 mM EDTA, 0.5 mM EGTA, 10% glycerol, 0.5 mM DTT, 0.2 mM phenylmethylsulfonyl fluoride, and 5 μg/ml of leupeptin, aprotinin, and pepstatin) containing 100 mM KCl. To purify p/CIP, the nuclear extract was loaded onto a P11 phosphocellulose column preequilibrated in the same buffer. The flowthrough was collected, and the column was washed sequentially with buffer A containing increasing concentrations of KCl. The 0.1 M fraction containing p/CIP was precipitated with 20 to 60% ammonium sulfate, and the precipitated proteins were resuspended in 4 ml of buffer A containing 100 mM KCl. This was then dialyzed against the same buffer to remove residual ammonium sulfate before the sample was applied to a Sephacryl S300 gel filtration column. The column was washed with buffer A at a flow rate of 0.4 ml/min. Fractions were collected, pooled, and analyzed for p/CIP by Western blotting. The p/CIP-containing fractions were pooled and dialyzed against buffer A containing 100 mM KCl. For immunoaffinity purification of p/CIP, affinity-purified p/CIP antibody was cross-linked to protein A-Sepharose using dimethylpalmilidate according to standard procedures (19). Fractions from the gel filtration step were pooled and precleared by passing the pooled fractions through a control affinity column containing anti-rabbit immunoglobulin G. The eluant was then loaded onto the anti-p/CIP affinity column at a flow rate of 0.2 to 0.5 ml/min. The flowthrough was collected and reloaded on the column five times prior to elution of the bound proteins with 100 mM glycine (pH 3.0). For mock-purification experiments, samples from the gel filtration step were loaded onto protein A-Sepharose cross-linked to an irrelevant antibody. Normally, protein samples were fractionated by SDS-PAGE and transferred to nitrocellulose membrane, and specific proteins were analyzed by Western blotting with the appropriate antibodies as indicated. For the purification of p/CIP from MEFs, cells were grown to confluence on 10- by 150-mm plates and harvested, and nuclear extracts were prepared as described above. Extracts were then passed directly through the anti-p/CIP immunoaffinity column as described above and then subjected to SDS-PAGE and Western blotting.
FLAG- or hemagglutinin-tagged proteins were generated using the Bac-to-Bac baculovirus expression system according to the manufacturer's instructions (Invitrogen). Proteins were subcloned into the Fastbac expression vectors and transformed into DH10 bacteria. The resulting bacmids were transfected into Sf9 cells to produce recombinant baculovirus, which was amplified and used to infect Sf9 cells at a multiplicity of infection of 5. After 48 h, cells were lysed, and recombinant proteins were purified by immunoaffinity with anti-Flag M2 affinity resin essentially as described previously. After extensive washing, proteins were eluted with 20 mM Tris buffer, pH 7.9, 100 mM KCl, 10% glycerol, 0.5 mM EDTA, and 0.2 mg/ml of the appropriate peptide competitor. Proteins were then dialyzed against elution buffer without peptide, frozen, and stored at −80 C.
Glutathione S-transferase (GST) fusion proteins containing various regions of p/CIP were purified on glutathione agarose beads following standard procedures. Regions of p/CIP were subcloned into the pGEX bacterial expression vector in frame with the GST moiety, and the resulting plasmid was transformed into the Escherichia coli strain BL-21. The overnight culture was diluted 1:100 in 100 to 300 ml of selection medium and grown at 37°C with shaking until an optical density at 600 nm of ~0.5 was reached, at which time protein expression was induced with 1.0 mM isopropyl-β-d-thiogalactopyranoside (final concentration) for 5 h at room temperature. Cells were pelleted by centrifugation at 4°C for 10 min at 10,000 × g and resuspended in 10 ml of resuspension buffer per 100 ml of starting culture volume prior to incubation on ice for 30 min with occasional mixing. The resuspended cells were then lysed by sonication prior to centrifugation of the samples at 4°C for 1 h at 30,000 × g. A total of 500 μl of GST-Sepharose beads prewashed in resuspension buffer was added to the clarified supernatant, and the samples were incubated at 4°C with rocking for 1 h. The beads were pelleted by centrifugation at 4°C for 2 min at 1000 rpm and subsequently washed five times with 12 ml of resuspension buffer. The beads were resuspended in 1 ml of resuspension buffer containing 10 mM reduced glutathione and incubated for 5 min with rocking. The supernatants containing the eluted proteins were dialyzed against a buffer consisting of 20 mM Tris, pH 7.9, 100 mM KCl, 10% glycerol, 0.5 mM EDTA, and 0.5 mM DTT and stored at −80°C.
In vitro methylation reactions were performed using 0.2 to 1 μg of either purified proteins or biotinylated peptides as substrates and 100 ng of purified CARM1 protein. Reactions were performed at 30°C for 1 h in 20 μl (final volume) of reaction buffer consisting of 50 mM Tris-HCl, pH 8.5, 20 mM KCl, 10 mM MgCl2, 1 mM β-mercaptoethanol, and 0.1 M sucrose and 5 μCi of [3H]SAM. Upon completion of the methylation reactions, biotin-labeled peptides were purified with MagnaSphere streptavidin beads (Promega) for 1 h at 4°C with rotation. Beads were washed three times with wash buffer (50 mM Tris, pH 8, 150 mM NaCl, 5 mM EDTA, 0.5% NP-40), and radioactivity was measured by scintillation counting.
For in vivo labeling of methylated proteins, MCF-7 cells were grown for 24 h in DMEM without methionine and containing stripped serum, followed by the addition of l-[methyl-3H]methionine (50 μCi/ml) for 1 h. Cells were then harvested, lysed, and p/CIP immunopurified as previously described (36).
To analyze methylation of p/CIP by mass spectrometry, a 100-ml culture of Sf9 cells (1.5 × 106 cells/ml) was infected with FLAG-p/CIP baculovirus alone or coinfected with FLAG-CARM1. At 48 h postinfection, the cells were harvested and resuspended in a 50 mM Tris buffer (pH 7.9) containing 300 mM KCl and 0.1% NP-40 and disrupted by 10 strokes with a dounce homogenizer; the cell lysate was then clarified by centrifugation at 100,000 × g for 30 min. To each of the cell extracts, anti-FLAG-agarose was added and incubated for 2 h at 4 C; cells were washed extensively with lysis buffer, with a final wash with buffer containing 100 mM KCl and 0.1% NP-40. The bound proteins were eluted with buffer containing 1.0 mg/ml FLAG peptide, and the coimmunopurified complexes were separated by SDS-PAGE. The 160-kDa band corresponding to p/CIP was excised and analyzed by mass spectrometry (MS) performed at the Southern Alberta Mass Spectrometry Center (Calgary, Alberta, Canada).
MCF-7 cells grown in phenol red-free DMEM containing 10% charcoal-stripped FBS for 48 h prior to stimulation with 10−7 M β-estradiol for the time periods indicated in Fig. Fig.1C.1C. Cells were then cross-linked with 1% formaldehyde at room temperature for 10 min. Cross-linking was terminated by adding glycine to a final concentration of 125 mM and incubating for 5 min. Cells were washed twice with ice-cold PBS containing 0.5 mM EDTA and harvested. Cells pellets were lysed in 0.3 ml of cell lysis buffer (50 mM Tris-HCl [pH 8.1], 10 mM EDTA, 1% SDS, and protease inhibitors) and incubated on ice for 10 min. Cell lysates were sonicated to yield DNA fragments ranging in size from 300 to 1,000 bp. Supernatants were diluted 10-fold in dilution buffer (20 mM Tris-HCl [pH 8.1], 1% Triton X-100, 2 mM EDTA, 150 mM NaCl, and protease inhibitors) and precleared with 60 μl of a 50% slurry of protein A-Sepharose containing 2.5 μg of sheared salmon sperm DNA for 2 h at 4°C. Immunoprecipitation was performed overnight at 4°C with 1.5 to 4 μg of the antibodies. A total of 60 μl of protein A-Sepharose containing 2.5 μg of salmon sperm DNA per ml was added to the solution and incubated for 1 h at 4°C. The beads were washed one time with buffer A (20 mM Tris-HCl [pH 8.1], 150 mM NaCl, 2 mM EDTA, 0.1% SDS, 1% Triton X-100), four times with cold buffer B (20 mM Tris-HCl [pH 8.1], 500 mM NaCl, 2 mM EDTA, 0.1% SDS, 1% Triton X-100), and one time with LiCl-detergent buffer (10 mM Tris-HCl [pH 8.1], 0.25 M LiCl, 1% NP-40, 1% deoxycholate, 1 mM EDTA). Finally, the samples were washed three times with Tris-EDTA buffer. Immunocomplexes were extracted twice with 200 μl of elution buffer (1% SDS-0.1 M NaHCO3). NaCl was added to a final concentration of 200 mM, and the cross-linking was reversed by heating at 65°C overnight. The protein was digested by adding Tris (pH 6.5) to a final concentration of 40 mM and 20 μg of proteinase K and incubating at 45°C for 1 h. The DNA was purified using QIAGEN PCR purification spin columns. PCR was performed using primers specific to the human pS2 promoter consisting of the forward primer 5-GGCCATCTCTCACTATGAATCATTCTGC-3′ and the reverse primer 5′-GGCAGGCTCTGTTTGCTTAAAGAGCG-3′. PCR conditions were as follows: initial denaturing cycle of at 94°C for 3 min, followed by 35 cycles of 94°C for 30 s, 52°C for 30 s, and 72°C for 1 min, with a final elongation step of 72°C for 10 min. The PCRs were quantified by scanning densitometry.
MEFs were grown to approximately 80% confluence and grown in DMEM methionine-cysteine-free medium for 1 h and then replaced by medium containing methionine for 40 min. Cells were then washed twice in PBS and incubated with medium containing 2 mM cold methionine-cysteine. Cells were harvested at the indicated times (see Fig. 6C and D), and p/CIP was immunoprecipitated from the cell lysates using p/CIP antibody. For cells overexpressing p/CIP, 2- by 15-cm plates were transfected with 10 μg of cDNA for 24 h. The transfected cells corresponding to each cDNA were pooled, replated (3 × 106 cells/10-cm plate), and allowed to adhere overnight, and pulse-chase experiments were performed the following day as described above.
CARM1 was initially identified based on its ability to interact with the carboxy terminus of GRIP1 (7) and has also been purified as a component of a large multiprotein complex from MCF-7 cells(63). To determine if p/CIP and CARM1 are stably associated in solution, we performed biochemical purification experiments of p/CIP and CARM1 using HeLa nuclear extracts. Gel filtration chromatography of nuclear extracts followed by Western blotting indicated that p/CIP could be found in two chromatographically distinct peaks, a large peak corresponding to a size of approximately 1 MDa and a second, more abundant peak, migrating at approximately 600 kDa (Fig. (Fig.1A).1A). Interestingly, CARM1 was identified as a broad peak partially overlapping with the smaller 600-kDa p/CIP component, whereas CBP was found in fractions corresponding to the larger 1-MDa peak. Fractions corresponding to each peak were pooled and purified using an anti-p/CIP affinity column. Immunoblotting of the affinity-purified p/CIP complexes indicated that CBP was associated with p/CIP purified from the larger 1-MDa component (Fig. (Fig.1B).1B). In contrast, CARM1 was not retained by the anti-p/CIP affinity column but was detected entirely in the flowthrough component of the 600-kDa (data not shown) fraction, suggesting that CARM1 and p/CIP do not stably associate in solution in HeLa cells. We also used ChIP to assess if p/CIP and CARM1 associate at the PS2 promoter in response to hormone (Fig. (Fig.1C).1C). Following stimulation with estradiol, the cells were fixed with formaldehyde, and chromatin was immunoprecipitated with a p/CIP antibody, followed by an immunoprecipitation using a CARM1-specific antibody. These results indicate that both p/CIP and CARM1 associate with the PS2 promoter in a hormone-dependent manner, with a peak occurring at approximately 30 min, which is consistent with earlier observations (38, 47). Collectively, these results establish that the association between CARM1 and p/CIP occurs on target genes in the context of chromatin and that they most likely do not form a stable complex in solution.
Recent studies have shown that CARM1 is capable of methylating nonhistone proteins, such as CBP, resulting in concomitant changes in transcriptional regulation (10, 28, 62). Based on the direct association between p/CIP and CARM1, we assessed whether p/CIP may also serve as a substrate for CARM1. In vitro methylation assays using purified proteins (Fig. (Fig.2A)2A) indicated that both p/CIP and CBP are robustly methylated by CARM1 (Fig. (Fig.2B).2B). The p/CIP methylation is directly dependent on the methyltransferase activity of CARM1 as a point mutation that abolishes methyltransferase activity was incapable of methylating either p/CIP or CBP (not shown). In contrast, neither p/CIP nor CBP is methylated by the related family member PRMT1 (Fig. (Fig.2B).2B). Furthermore, the addition of increasing amounts of free histones in the methylation reaction had no effect on p/CIP methylation, suggesting that p/CIP serves as a very efficient substrate for CARM1 (Fig. (Fig.2C2C).
To further define the specificity of the methylation reaction, we utilized whole-cell extracts from MEFs isolated from CARM1 null mice as a source of methyltransferase activity. Purified recombinant p/CIP could not be methylated in vitro using extracts from CARM1−/− MEFs, indicating that that no alternative pathway exists for p/CIP methylation (Fig. (Fig.3A).3A). Experiments were also performed to assess whether p/CIP was specifically methylated in intact cells. First, MCF-7 cells were metabolically labeled with [3H]SAM, followed by immunoprecipitation using a p/CIP antibody. The results of the fluorographic image indicated specific incorporation of the radioisotope into the immunoprecipitated p/CIP protein (Fig. (Fig.3B).3B). Second, p/CIP was immunoprecipitated from cells grown in the absence or presence of the methylation inhibitor periodate oxidized adenosine (AdOx). The immunoprecipitated proteins were then immunoblotted with p/CIP antibody or with an antibody which specifically recognizes asymmetric dimethylarginines found on arginine 17 of histone H3 (αmR17). The antibody, αmR17, has previously been shown to not only recognize the methylated H3 but also display immunoreactivity with a number of other proteins in a CARM1-dependent fashion (64). Both the anti-p/CIP antibody and the αmR17 antibody were able to recognize the immunoprecipitated p/CIP protein (Fig. (Fig.3C).3C). However, for cells grown in the presence of AdOx, a significant decrease in immunoreactivity was detected using the αmR17. Taken together, these results establish that p/CIP is a substrate for CARM1-dependent methylation both in vitro and in intact cells.
To identify the specific amino acids methylated by CARM1, we overexpressed p/CIP, alone or in combination with CARM1, in Sf9 cells using a baculovirus expression system. Forty-eight hours after infection, p/CIP was affinity purified, isolated by SDS-PAGE, and analyzed by matrix-assisted laser desorption ionization-time of flight MS. The masses of 15 peptides were identified and found to match the predicted peptide masses of p/CIP. Four were identified with increased masses, consistent with the presence of dimethyl groups (Table (Table1).1). Three of the peptides were found to contain low levels of an arginine dimethylated form, regardless of the presence of CARM1. Significantly, it was found that the relative abundance of the methylated form of a peptide corresponding to amino acids (aa) 1176 to 1201 of p/CIP was directly correlated with the coexpression of CARM1. This peptide was then analyzed further by liquid chromatography-tandem MS, and it was determined that the arginines located at positions 1178, 1184, and 1195 were all asymmetrically dimethylated in a CARM1-dependent manner.
To validate the results obtained by MS, C-terminal truncation mutants of p/CIP were generated using baculovirus, purified, and tested as substrates for methylation by CARM1 in vitro (Fig. (Fig.44 A). Whereas a fragment spanning residues 1 to 1398 was strongly methylated, deletion of residues 1085 to 1398 dramatically decreased methylation. This is consistent with the indicated sites of methylation determined by MS. In addition, we also employed biotinylated synthetic peptides as substrates for in vitro methylation reactions. The peptides were bound to streptavidin-coated beads, washed, and analyzed by liquid scintillation counting. The peptide corresponding to aa 1176 to 1201, which contained the CARM1-dependent methylation sites identified by MS, was robustly methylated in vitro (Fig. (Fig.4B).4B). Further inspection of this region indicated the presence of three additional arginines at aa 1163, 1168, and 1170 which lie immediately N-terminal to the region identified by MS. Although not identified by MS, in vitro methylation reactions using a synthetic peptide indicated that this region is also methylated by CARM1 in vitro (Fig. (Fig.4B4B).
Finally, mapping studies were performed using recombinant GST proteins fused to regions of p/CIP spanning the full-length protein. Using the GST fusion proteins as substrates, two additional regions of p/CIP, corresponding to aa 819 to 853 and 920 to 968, were found to serve as highly efficient substrates for CARM1 in vitro (Fig. (Fig.5A).5A). Further mapping using biotinylated peptides indicated that aa 844 to 865, containing arginines at aa 839 and 844, was robustly methylated by CARM1 in vitro (Fig. (Fig.5B).5B). A second peptide containing an arginine at aa 961 was also methylated, although to a lesser degree. To determine which arginine is specifically methylated within aa 844 to 865, individual arginines were mutated to alanine. Mutation of the amino acid at position 839 completely abolished CARM1-dependent methylation, whereas mutation of arginine to alanine at aa 844 diminished methylation by approximately 50%, suggesting that arginine 844 may be involved in substrate recognition by CARM1 (Fig. (Fig.5C,5C, center panel). Collectively, the results of our biochemical analysis indicate two methylated regions in p/CIP: methylation domain 1 (MD1) stretching from R1163 to R1195 and MD2 stretching from R839 to R961.
Amino acids at positions 839 and 844 are found directly adjacent to a region of p/CIP containing several conserved serines, at aa 847 and 850, which have recently been identified as physiologically relevant phosphorylation sites (60). The proximity of these phosphorylation sites to R839 suggests a potential for cross talk between methylation and phosphorylation of p/CIP. To test this hypothesis, we generated a peptide phosphorylated at aa 847 which has been shown to be phosphorylated by IκB kinase in response to hormone treatment (60). Phosphorylation at aa 847 decreased the methylation at aa 839 by approximately fivefold, indicating that prior phosphorylation antagonizes methylation at aa 839 in vitro (Fig. (Fig.5C,5C, right panel).
To investigate the functional relevance of p/CIP methylation of the carboxy terminus (MD1), we generated substitutions of arginine to alanine at R1178, R1184, and R1195, yielding (p/CIPR3A, and at R1163, R1168, R1170, R1178, R1184, and R1195, yielding p/CIPR6A. Methylation reactions using purified recombinant p/CIP3A and p/CIP6RA proteins indicated a significant decrease in CARM1-dependent methylation in vitro (Fig. (Fig.6A).6A). Surprisingly, transient expression of the methylation site mutants in COS-1 cells consistently yielded higher protein levels compared to wild-type p/CIP, suggesting that methylation of p/CIP may play a role in p/CIP turnover (Fig. (Fig.6B).6B). This was also reflected in the amount of mutant protein immunoprecipitated relative to wild-type p/CIP. In order to examine the effect of CARM1-mediated methylation on p/CIP stability, we conducted pulse-chase experiments in CARM1−/− and CARM1+/+ MEFs. Cells were labeled with [35S]methionine for 40 min and then chased with cold methionine for various time periods, followed by immunoprecipitation of endogenous p/CIP and analysis by immunoblotting and fluorography. The level of radiolabeled p/CIP relative to the total cellular pool was substantially reduced at 8 h in CARM1+/+ cells, while no noticeable reductions were observed in CARM1−/− cells (Fig. (Fig.6C).6C). These results suggest that the turnover of p/CIP in the CARM1 wild-type cells was significantly faster, consistent with the notion that methylation affects protein stability. To determine whether this effect is directly dependent on the amino acid residues recognized by CARM1, pulse-chase experiments were also conducted using cells transfected with p/CIP mutants. Cells were transfected with either p/CIP or p/CIPR6A, and after 24 h cells were replated and labeled with [35S]methionine. In agreement with the above data, the stability of the mutant was significantly increased compared to wild-type p/CIP (Fig. (Fig.6D).6D). To examine if p/CIP turnover is directly dependent on CARM1 methylation, MCF-7 cells were grown in the presence of AdOx to hypomethylate p/CIP; subsequently, p/CIP was immunoprecipitated and incubated with extracts derived from CARM1−/− or CARM1+/+ MEFs, in the presence of SAM, for various time periods, followed by immunoblotting with p/CIP antibody. A dramatic reduction in p/CIP protein levels was observed following incubation with extracts derived from the CARM1 wild-type MEFs (Fig. (Fig.7).7). In contrast, p/CIP incubated with CARM1−/− extracts was relatively resistant to degradation. These results suggest that CARM1-dependent methylation may be required for p/CIP turnover.
To determine a functional role of p/CIP methylation, we examined the effects of p/CIP on transcriptional activity by the estrogen receptor. Transient transfection assays in COS-1 cells were performed using a reporter vector containing three copies of an ERE upstream of a minimal promoter, as well as the PS2 proximal promoter-driven luciferase gene. Transfection of limiting amounts of wild-type p/CIP resulted in a slight increase in transcriptional activity of the ERE luciferase in the presence of β-estradiol (Fig. (Fig.8A).8A). In contrast, an equivalent concentration of p/CIPR6A mutant resulted in a significantly larger response in all cell types examined. An enhanced transcriptional response by the p/CIPR6A mutant was more evident on the PS2 promoter. Overexpression of wild-type p/CIP resulted in an approximately 2.8-fold increase in activity, whereas transfection of the p/CIPR6A mutant resulted in a concentration-dependent effect with an additional threefold increase in luciferase activity relative to the wild type (Fig. (Fig.8A).8A). These results suggest that a lack of p/CIP methylation may cause p/CIP to accumulate within the cell, which in turn results in enhanced transcription in response to β-estradiol.
We also examined the transcriptional activity of various regions of p/CIP fused to the Gal4 DBD. p/CIP (aa 819 to 1160) contains the major activation domain (AD1), which overlaps directly with the CBP/p300 interaction domain, but does not include the MD1 (5). When expressed in COS-1 cells, this Gal fusion produced a robust transcriptional response (Fig. (Fig.8B).8B). Surprisingly, extension of this construct to include the MD1 (aa 819 to 1215) resulted in a dramatic reduction in transcriptional activity, indicating that this region may impart a negative regulatory influence on p/CIP-mediated transcriptional activation. Analysis of Gal4-p/CIP (aa 1158 to 1215) encompassing the MD1 alone revealed no significant transcriptional activity. Experiments were also performed using the entire carboxy terminus of p/CIP fused to Gal4 (aa 819 to 1398) containing point mutations in the MD1, which should prevent (Fig. (Fig.8B,8B, lower panel) the effects on transcriptional activity imparted by methylation. Remarkably, these mutants displayed increased transcriptional activity when overexpressed in COS-1 cells, consistent with the effects obtained on the ERE-driven reporter constructs. However, under these conditions we did not observe significant differences in expression of the various Gal4 fusion proteins. These results suggest that the reduced transcriptional activity associated with methylation is not simply a result of increased p/CIP degradation but that the methylation may affect activity by an additional mechanism.
The effects of p/CIP methylation on transcription and the proximity of the methylation sites to the CBP/p300 interaction domain suggested that methylation may regulate the association with CBP. To investigate this possibility, we prepared nuclear extracts from both wild-type and CARM1−/− MEFs and purified endogenous p/CIP protein complexes using an immunoaffinity column. Initial experiments indicated that p/CIP levels were elevated in the CARM1−/− MEFs by approximately twofold (Fig. (Fig.9A).9A). In order to compensate for the increase in p/CIP levels, the amount of starting material in the input was normalized so that each had approximately the same amount of p/CIP, based on Western blotting. Immunoblotting of the purified proteins with a CBP-specific antibody indicated that the amount of CBP stably associated with p/CIP is significantly enhanced in the CARM1 knockout MEFs (Fig. (Fig.9B).9B). Indeed, in the wild-type MEFs, CBP was not detectable in the purified p/CIP complexes. These results indicate that in the absence of CARM1, the association between p/CIP and CBP is facilitated.
Previous overexpression studies in intact cells and in vitro reconstitution assays using chromatinized templates have shown that CARM1 can interact and synergize with both p/CIP and CBP/p300 to activate NR-mediated transcription (6, 26, 62). This synergy is dependent on the methyltransferase activity of CARM1 and has been attributed to the ability of CARM1 to methylate specific arginines on histone H3 tails (50). Our data suggest that recruitment of CARM1 by p/CIP may play a more complex role in transcription than previously anticipated. In the present study we demonstrate that the transcriptional coactivator p/CIP/SRC-3 is methylated by CARM1. Using mass spectrometric analysis, we have shown that the predominant CARM1-dependent methylation sites in p/CIP are localized to a glutamine-rich region in the carboxy terminus, which we refer to as MD1.
PRMT1, a related methyltransferase, has also been shown to synergize with SRC and CBP/p300 proteins to activate NR signaling in transient transfection assays (26). The substrate specificity of CARM1 and PRMT1 is different, suggesting that the mechanism of coactivation of these two related methyltransferases may also differ (50). Indeed, our data suggest that PRMT1 is unable to methylate p/CIP in vitro, suggesting that p/CIP is a specific substrate for CARM1. Additionally, p/CIP could not be methylated in vitro by extracts derived from CARM1−/− MEFs which still express PRMT1. This indicates that CARM1 is required for methylation of p/CIP in vivo. Importantly, immunoprecipitation of p/CIP from cell extracts, followed by Western blotting using the αR17H3 antibody which recognizes a dimethylated CARM1 recognition site on histone H3, clearly recognized methylated p/CIP. Taken together, these results, in conjunction with the mass spectrometry data, demonstrate that p/CIP is a direct substrate for CARM1 in vivo.
By in vitro mapping experiments we have also identified a second region in p/CIP that can be methylated (R839 and R961) which we refer to as MD2. Although methylation of MD2 is extremely robust using peptide substrates and recombinant GST proteins, the functional significance of this finding is currently unclear as we did not detect methylation at this site using MS. Furthermore, deletion of MD2 did not have a significant impact on overall methylation of p/CIP. Our inability to detect methylation at R839 through mass comparisons may be due to the fact that R839 in Sf9 cells is already occupied by methyl groups and would not be accessible to CARM1-dependent methylation in Sf9 cells or in vitro using baculovirus-generated protein. Additionally, we have shown that phosphorylation of S847 antagonizes methylation at R839. This finding is significant as this site has recently been shown to be phosphorylated on endogenous SRC-3 and is functionally required for coactivation in response to hormonal signaling (60). Perhaps p/CIP produced in Sf9 cells may already be phosphorylated at S847, which may render R839 or R844 inaccessible to subsequent methylation by CARM1.
A major finding of this study is that methylation regulates the stability of p/CIP. p/CIP mutants containing substitutions of arginine to alanine accumulate to a much greater degree when overexpressed in various cell types compared to wild-type p/CIP. A significantly higher concentration of p/CIP was also detected in the CARM1−/− MEFs relative to the wild-type MEFs. This finding is also reflected in the pulse-chase experiments, which indicate that the methylation status of p/CIP is an important determinant regulating its stability, both in the CARM1−/− MEFs and in cells expressing p/CIP methylation site mutants. Furthermore, we demonstrate that reconstituting hypomethylated p/CIP with extracts derived from wild-type MEFs resulted in a more rapid degradation of p/CIP compared to reconstitution of p/CIP using CARM1−/− MEF extracts. These results suggest that the methylation status of p/CIP may serve as a prerequisite for rapid degradation. A critically unanswered question is the mechanism through which methylation affects p/CIP stability. Studies have shown that arginine methylation promotes protein-protein interactions in a variety of signal transduction pathways (3). Thus, one possibility is that methylation of p/CIP by CARM1 allows specific proteins to interact selectively with p/CIP, which, in turn, facilitates proteolytic degradation.
An additional finding in this study is that p/CIP methylation may play a role in the transcriptional process by regulating complex assembly. This assertion is based on the observation that extension of the carboxy terminus AD1 to include the MD1 resulted in a significant decrease in transcriptional activity when tethered to the Gal4 DBD. Conversely, mutation of the CARM1 recognition sites to nonmethylatable alanines enhanced the transcriptional response of the carboxy terminus when tethered to Gal4. Under these experimental conditions, however, we did not observe a significant difference in the expression levels of the various fusion proteins. This suggested that in addition to its role in protein stability, methylation may promote a conformational change that weakens the association of accessory proteins such as CBP/p300 and consequently contributes to complex dissociation. In support of this hypothesis, we observed that CBP forms a stable association with p/CIP and copurified with p/CIP from extracts derived from CARM1−/− MEFs, compared to CARM+/+ MEFs where no CBP was detected by Western blotting. Interestingly, a recent study has shown that CARM1-dependent methylation of the GRIP1-binding domain within CBP may also promote coactivator complex disassembly (31). Alternatively, methylation may result in a conformational change in p/CIP which allows another protein to bind, thereby preventing interaction with CBP/p300.
The increase in p/CIP turnover and dissociation of CBP in response to methylation by another transcriptional coactivator seem paradoxical as these activities would typically be associated with the negative control of transcription. However, they are consistent with the notion that transcriptional activation by nuclear hormone receptors is a dynamic process involving repetitive cycles of association and dissociation of receptors and a large number of coregulators with target genes (38, 39). A number of mechanisms may contribute to this process, such as interaction with molecular chaperones which can associate with liganded nuclear receptors and disassemble nuclear receptor/coregulator complexes (14). In addition, several studies have established a link between ligand-dependent transcriptional activation by nuclear hormone receptors and the ubiquitin-proteasome pathway (25). For example, it has been shown that the degradation of ERα, in response to β-estradiol, is dependent on the presence of AIB1 (48). Coactivator turnover is also dependent on interactions with the proteasome and may be directly regulated by phosphorylation. The phosphorylation of GRIP1 by protein kinase A induces its degradation via the ubiquitin-proteasome pathway, and a recent study has shown that SRC-3 is phosphorylated and then degraded in response to retinoic acid (16, 21). Interestingly, Li et al. have identified a novel interaction between the carboxy terminus of SRC-3 and Regγ, an activator of the 20S proteasome. Loss of Regγ results in the accumulation of SRC-3 protein levels (33). Studies examining promoter occupancy have also established a link between proteasomal activity and transcriptional activation (25). It has been shown that recruitment of the proteasomal machinery is required for gene activation, presumably to allow cyclic clearance and continuous recycling of NR and coregulators at target genes (46, 48). The exact mechanism for coordinating the events involved in coactivator association and dissociation are poorly understood although increasing evidence suggests that posttranslational modifications may be instrumental to the overall process. For example, treatment of cells with hormones or various growth factors stimulates the phosphorylation of p/CIP/SRC-3 at specific sites and appears to be a prerequisite for hormone-dependent transcription activation that facilitates binding to CBP (13, 60). Acetylation of SRC-3 also appears to play a role in influencing transcription by facilitating coactivator release from hormone-bound NR (9).
The human homologue of p/CIP was first isolated from human breast tumors, and it is located within a region of chromosome 20 that is often amplified in breast and ovarian cancer (2). Several follow-up studies have confirmed that p/CIP is amplified in a fraction of breast tumors, with amplification frequencies ranging from 5 to 10% (18, 35, 40, 45). Moreover, compared to normal breast epithelium, p/CIP is overexpressed at the RNA level in 31 to 64% of breast tumors examined. One question arising from the present study is whether deregulated arginine methylation of p/CIP plays a role in cellular transformation. Based on the data presented, we would predict that the methylation status of p/CIP is an important determinant of p/CIP transcriptional activity and may affect its oncogenic capacity. Alternatively, the accumulation of hypomethylated p/CIP could also adversely affect the cycling process inherent in NR-mediated transcription and impair transcription at selective promoters.
In summary, the findings presented in this study link CARM1-dependent methylation of p/CIP to both protein turnover and coactivator complex dissociation, establishing arginine methylation as a critical regulatory signal in p/CIP-mediated transcription. In addition, this type of regulation could serve to integrate diverse signaling pathways at the level of coactivator proteins and may have important implications for understanding the role of p/CIP in breast cancer. Further delineation of the role of p/CIP methylation and its downstream physiological effects is essential to decipher the complex regulatory circuits in nuclear receptor signaling.
We thank M. Stallcup for the CARM1 cDNA and H. Herschman for the PRMTI cDNA.
J.T. is supported by National Cancer Institute of Canada grant 015428. M.T.B. is supported by NIH grant DK62248 and in part by an institutional center grant (ES07784).
Published ahead of print on 16 October 2006.