A Proteomic Analysis of Arginine-methylated Protein Complexes*

Arginine methylation is a post-translational modification that results in the formation of asymmetrical and symmetrical dimethylated arginines (a- and sDMA). This modification is catalyzed by type I and II protein-arginine methyltransferases (PRMT), respectively. The two major enzymes PRMT1 (type I) and PRMT5 (type II) preferentially methylate arginines located in RG-rich clusters. Arginine methylation is a common modification, but the reagents for detecting this modification have been lacking. Thus, fewer than 20 proteins have been identified in the last 40 years as containing dimethylated arginines. We have generated previously four arginine methyl-specific antibodies; ASYM24 and ASYM25 are specific for aDMA, whereas SYM10 and SYM11 recognize sDMA. All of these antibodies were generated by using peptides with aDMA or sDMA in the context of different RG-rich sequences. HeLa cell extracts were used to purify the protein complexes recognized by each of the four antibodies, and the proteins were identified by microcapillary reverse-phase liquid chromatography coupled on line with electrospray ionization tandem mass spectrometry. The analysis of two tandem mass spectra for each methyl-specific antibody resulted in the identification of over 200 new proteins that are putatively arginine-methylated. The major protein complexes that were purified include components required for pre-mRNA splicing, polyadenylation, transcription, signal transduction, and cytoskeleton and DNA repair. These findings provide a basis for the identification of the role of arginine methylation in many cellular processes.

Protein arginine methylation is a post-translational modification that adds monomethyl or dimethyl groups to the guanidino nitrogen atoms of arginine (1). The enzymes responsible for protein arginine methylation have been classified in two major classes; type I enzymes promote the formation of asymmetrical -N G ,N G -dimethylated arginines (aDMA), 1 and type II enzymes catalyze the formation of symmetrical -N G ,NЈ G -dimethylated arginines (sDMA) (1). -N G -Monomethylarginine is thought to be an intermediate formed by both enzyme types. The metabolic cost of methylation is high, requiring the use of 12 ATP molecules/methylation event (1). The fact that evolution has retained such an "expensive" reaction underscores the biological importance of this posttranslational modification (2). There are now at least five type I protein-arginine methyltransferases in mammals; PRMT1 (3), PRMT2 (4), PRMT3 (5), CARM1 (PRMT4) (6), and PRMT6 (7); and one type II, PRMT5 (8). Recently, a new arginine methyltransferase was identified and by homology is likely a type I PRMT (9).
Myelin basic proteins (MBPs) and histones are among the first proteins shown to contain dimethylated arginines (10,11). MBP has been shown to contain sDMA, but the enzyme or the function of this post-translational modification remains unknown. Histones have been shown to be methylated by PRMT1 and CARM1 in vivo (12)(13)(14)(15)(16). Histone-arginine methylation is thought to contribute to the histone code (17). Another class of dimethylated proteins includes RNA-binding proteins (18,19). Arginine-glycine-rich sequences have been suggested and have been shown to contribute to the RNA binding activity (20 -22). Thus, the methylation of these arginines would be predicted to alter or regulate their RNA binding activity, but the evidence for this has been lacking. It has been shown that arginine methylation regulates protein localization (2). It was first shown by Shen and co-workers (23) that the removal of the yeast methyltransferase Hmt1p causes the nuclear retention of two hnRNPs, Npl3p and Hrp1p. Subsequently it was shown that arginine methylation regulates the import of Npl3p (24) and the protein localization of hnRNP A2 (25), Sam68 (26), and p80-coilin (27,28).
Arginine methylation has been shown to regulate protein-protein interactions (29). The discovery that Sm proteins are methylated (30) and the discovery that the product of the spinal muscular atrophy gene product SMN associates with methylated Sm proteins (31) led to the proposal that arginine methylation may be a signal that targets proteins. In the case of small nuclear ribonucleoprotein (snRNP) particle assembly, arginine methylation by the PRMT5 methylosome (32,33) has been proposed to be the signal for the recognition and targeting to the SMN protein complexes (31). Although the role of arginine methylation in signal transduction was proposed in 1998 (34), few signaling proteins have been identified to be arginine-methylated. The inhibition of PRMT1 by antisense quenches the interferon signaling (35), and similarly arginine methylation of STAT1 has been shown to be required for interferon signaling (36). Moreover, the methylation of Sam68 peptides prevents their association with SH3 domains (29).
In the present manuscript, we utilized four arginine dimethyl-specific antibodies to purify arginine-methylated protein complexes. The proteins were identified by LC/MS/MS, and the protein sequences were searched for neighboring RG sequences that match the epitopes used to generate the peptide antibodies. We report the identification of over 200 proteins that are putatively arginine-methylated. These include RNA-binding proteins, transcription and polyadenylation factors, cytoskeleton proteins, as well as proteins involved in signal transduction and DNA repair. These data will help elucidate the role of arginine methylation in many cellular processes.

EXPERIMENTAL PROCEDURES
Antibodies-SYM10 and ASYM24 have been described previously (26,27) and are distributed by UBI Inc. (Lake Placid, NY). SYM11 and ASYM25 were generated by immunizing rabbits with the peptides KAAILKAQVAAR sDMA GR sDMA GR sDMA GMGR sDMA G and KFGGR aDM -AGGGR aDMA GGGR aDMA GGFGGR aDMA GGR aDMA G, respectively. Polyclonal antibodies were generated by using New Zealand white rabbits injected with peptides coupled to keyhole limpet hemocyanin (Sigma).
Mass Spectrometry-HeLa-S3 (5 ϫ 10 8 ) cells were obtained from Biovest International Inc./National Cell Culture Center (Minneapolis, MN) and lysed in a buffer containing 1% Triton X-100 (Roche Applied Science), 20 mM Tris, pH 7.4, 150 mM NaCl, 1 g/ml aprotinin, 1 g/ml leupeptin, and 0.01% phenylmethanesulfonyl fluoride. Endogenous methylated proteins were immunopurified from the cell lysate using 1 mg of the respective polyclonal methyl-specific antibody coupled to 1 g of protein A-Sepharose (Sigma). After extensive washings with lysis buffer and phosphate-buffered saline, pH 7.4, the bound proteins were eluted with 500 l of 1ϫ phosphate-buffered saline containing 250 M corresponding immunogenic peptide. Eluted proteins were dialyzed against water overnight, lyophilized, and identified using trypsin digestion and LC/MS/MS sequencing of peptides. Specifically, proteins were directly digested (not reduced or alkylated) with 20:1 protein:trypsin (mass ratio) for 16 h at 37°C. LC/MS/MS experiments were conducted on a Qstar Pulsar i instrument configured with a Protana nanospray source to which an Ultimate LC system was coupled (LC Packings/Dionex). A 75-m inner diameter by 15 cm reverse-phase C 18 PepMap TM column (LC Packings) operated at a flow rate of 200 nl/min with a gradient of 5-15% B (0-5 min), 15-50% B, 5-50% B (5-40 min), and 50 -80% B (40 -50 min). (Solvent A is 0.05% aqueous formic acid, and Solvent B is 0.05% formic acid in acetonitrile.) All solvents were high pressure liquid chromatography grade (Fisher). An information-dependent acquisition experiment was conducted using a 1-s survey scan and two 2-s pendant MS/MS scans incorporating mass-dependent ion bunching for sensitivity enhancement (pulsar mode). Some of the eluted proteins were resolved by SDS-PAGE and revealed by Coomassie Blue R-250 staining. Bands corresponding to appropriate molecular weights were excised, in-gel digested with trypsin, and analyzed on a Voyager DE-STR reflectron MALDI-TOF using a-cyano-4-hydroxycinnamic acid as the matrix in a standard dried droplet sample preparation protocol.

RESULTS
To identify dimethylated arginine-containing cellular proteins, immunoprecipitations were performed with dimethylspecific antibodies. SYM10 is an antibody that we generated previously against peptide R sDMA GR sDMA GR sDMA GR sDMA G. Using enzyme-linked immunosorbent assay, we have shown that SYM10 could effectively distinguish sDMA versus aDMA, even at low dilutions of 1:500 (27). Further analysis by enzyme-linked immunosorbent assay helped us define the SYM10 epitope as at least two, preferentially non-contiguous, sDMA-Gs in a given peptide, which is consistent with the fact that SYM10 does not recognize a peptide derived from MBP that harbors a single sDMA-G (27). The epitope for SYM10 diminishes in PRMT5 small interfering RNA-treated cells demonstrating that PRMT5 is an enzyme that contributes to the SYM10 epitope (27). We have shown previously that one of the complexes immunoprecipitated by SYM10 was the spliceosomal snRNP core proteins, which include known sDMAcontaining proteins SmB/BЈ, D1, and D3 (27). Among the additional proteins present in the SYM10 immunoprecipitate, we identified p80-coilin as being an sDMA-containing protein (27). A second antibody, SYM11, was generated against KAAILKAQVAAR sDMA GR sDMA GR sDMA GMGR sDMA G and also specifically recognizes sDMA-containing peptides (data not shown), as determined by the methodology described above for SYM10. SYM10 and SYM11 each recognize specific protein patterns as determined by immunoblotting (Fig. 1). The recognition patterns are distinct but partially overlap as the Sm proteins are both recognized by these antibodies (Fig. 1). ASYM24 was generated by using the peptide KGR aDMA -GR aDMA GR aDMA GR aDMA GPPPPPR aDMA GR aDMA GR aDMA -GR aDMA G as an antigen as reported previously (26). ASYM24 recognizes aDMA specifically and was shown to recognize Sam68, an RNA-binding protein (26). The epitope for ASYM24 diminishes significantly in PRMT1Ϫ/Ϫ cells, demonstrating that the major enzyme that contributes to the ASYM24 epitope is PRMT1 (26). ASYM25 was generated by immunizing rabbits with the peptide KFGGR aDMA GGGR aDMA -GGGR aDMA GGFGGR aDMA GGR aDMA G. ASYM25 is aDMAspecific and has a different specificity than ASYM24 as demonstrated by immunoblotting ( Fig. 1).
At least two large scale immunopurifications were performed with HeLa-S3 cell extracts using each of the dimethyl-arginine-specific antibodies. One control immunoprecipitation was performed using the protein A-Sepharose resin alone (data not shown). The bound protein complexes were eluted with the respective immunogenic peptides. The eluted proteins were digested with trypsin, and the peptides were separated and identified by LC/MS/MS, which is a mass spectrometry technique particularly suitable to the analysis of complex protein mixtures. In the case of SYM10, predominant bands visualized by SDS-PAGE were also identified by using MALDI-TOF mass spectrometry (as indicated by a ϩ sign in Table I). The resulting peptide mass maps were internally calibrated using three molecular weight standards and were submitted to Mascot (Matrix Science Ltd., London) search algorithms (with a 30-ppm mass tolerance with acrylamide adduction chemistry for cysteines) to query the National Center for Biotechnology Information non-redundant (NCBI-nr) database. These results were then confirmed using the Protein Prospector tools at University of California-San Francisco (prospector.ucsf.edu), and the identifications given a MOWSE score greater than 1000 were considered positive. All proteins that were scored positive using this procedure were also identified using LC/MS/MS, validating this approach. Hence, only LC/MS/MS was used for the remainder of the analysis. LC/MS/MS data were searched using Mascot against the National Center for Biotechnology Information non-redundant (NCBI-nr) database (release date of April 5, 2003) with a peptide mass tolerance of 100 ppm and a fragment mass tolerance of 0.1 Da as well as allowing for two missed cleavages. A protein was scored positive by LC/MS/MS if it was not found in the experiment performed with protein A-Sepharose alone and if 1) it was identified in more than one analysis, 2) the number of unique peptides was greater than three, or 3) the amino acid sequence contained two RG repeats separated by less than 10 amino acids. The identified proteins were grouped according to their known or putative functions (Tables I-IV). Proteins that were previously reported to be methylated are marked with q in Tables I-IV. The proteins that contain the RG repeats represent the epitopes for our DMAspecific antibodies and are likely to contains sDMA or aDMA (shown in bold characters in Tables I-IV). The proteins that are devoid of RG motifs are likely co-purifying proteins.
The SYM10 and SYM11 antibodies purified the splicing factor "KH-type splicing regulatory protein" (KSRP), a homolog of zipcode-binding proteins ZBP1 and ZBP2 (40 -42). The presence of RG repeats suggests that they contain sDMAs. The gene for TLS (Translocated in LipoSarcoma) or FUS is rearranged in human myxoid liposarcoma (43). The TLS protein has been shown to function as a splicing factor and is also implicated in DNA repair (44,45). TLS was shown to be a substrate of PRMT1 (46) and shown to contain aDMA (47). Thus, TLS most likely represents a co-purifying protein for SYM11. Another gene product often associated with chromosomal translocation was identified, namely, the Ewing sarcoma protein EWS (48). The EWS protein has been shown to contain dimethylated arginines; however, it has not been shown whether the modification was symmetric or asymmetric (49). Our data suggest that EWS contains sDMA because of its presence in SYM10 immunoprecipitations (Table I). P80coilin is the marker for Cajal bodies, and this nuclear structure is associated with snRNP assembly and thus splicing (50). We and others have shown that the RG repeats of p80-coilin are symmetrically dimethylated on arginines (27,28).
ASYM24 and ASYM25-Three RNA helicases containing RG repeats were purified with ASYM25 (Table IV). One of these helicases, RH70, co-purifies with the U1 snRNP (51). In addition, several RNA-binding proteins were identified including Sam68, the Src substrate in mitosis, which has been shown to function in various aspects of RNA metabolism (26). HnRNP Q2 protein was purified with ASYM24 (Table III) suggesting that it contains aDMA. The RG sequences of hnRNP Q interact with SMN and may link SMN to the spliceosome (52).

Protein Translation
A protein complex that was affinity-purified by SYM10 but not SYM11 was the cleavage and polyadenylation specificity factor (CPSF) complex including the 25-, 68-, and 100-kDa subunits ( Table I). The poly(A)-binding proteins PABP1 and PABP2 as well as eIF-4G were purified (53). CPSF6, a protein of 68 kDa, has several RG repeats that are likely the epitope for SYM10, and the other components are likely co-associating proteins (54). PABP1 has been shown to be asymmetrically dimethylated by CARM1 at arginine 455 and 460, which is not an epitope for SYM10 (46). Thus, PABPs and eIF-4G are likely purified as co-associating proteins with the CPSF com-plex. SYM11 purified the elongation factor 1 ␦ isoform, which is a guanine nucleotide exchange factor (Table II). ASYM24 and ASYM25 purified several ribosomal proteins and tRNA components (Tables III and IV).

DNA Transcription
The role of arginine methylation in transcriptional regulation is well established with histones being a major target of protein-arginine methyltransferases (see the Introduction). As histones do not contain RG-rich sequences, we did not expect to identify them, and indeed histones were not identified. SYM11 purified transcriptional proteins involved in all aspects of tran-  (Table I) (55) and histone-lysine methyltransferase DOT1L were identified (Table II) (56). Two proteins with methyl-binding domains were identified (Table II) including methyl-CpG-binding protein 2 (MBD2), which plays a role in transcriptional control (57,58), and the transcription terminator factor I-interacting protein 5. A protein involved in transcriptional elongation, DSIF p160, and a serine phosphatase TFIIF-associated CTD phosphatase I, which is involved in the recycling of RNA polymerase II (59), were purified with SYM11. Several transcription factors were also identified including the basic helix-loop-helix leucine zipper protein TFEB, p98 Rel homolog, interleukin enhancer-binding factor 1 and 2 (ILF1 and ILF2), zinc finger DNA-binding proteins, homeodomain protein PBX4, the C/EBP-induced protein, and ZNP9. Interestingly, ILF2 was identified with SYM11 and ASYM24 suggesting that it contains both modifications. The amino acid sequence of TFEB has RG repeats that are likely candidates for SYM10 recognition. A chromosomal translocation leads to a promoter switching, giving rise to an ␣TFEB fusion gene, which leads to renal cell carcinomas (60,61). The ZNF9, RING zinc finger protein 9, was identified by LC/MS/MS, and the presence of RG repeats in its sequence suggests that it is a direct epitope of SYM10. The function of ZNF9 is unknown, but it is a nucleic acid-binding protein with an AIR1 (arginine methyltransferase-interacting protein 1) domain (62). The presence of CCTG repeats in intron 1 of ZNF9 gene leads to myotonic dystrophy type 2 (63). Thus, arginine methylation may regulate the transcriptional initiation, elongation, and termination.

Receptors and Signaling
Six G-coupled receptors were identified with the DMAspecific antibodies. Both the ␣2CII adrenergic receptor and the urotensin II receptor were purified by SYM10 (Table I); the ␣-2A adrenergic receptor was identified with SYM11 (Table II), and three unclassified receptors were identified with ASYM25 (Table IV). This analysis is the first indication that G-coupled receptors may contain dimethylated arginines, and the presence of RG repeats suggests that the receptors were directly recognized by the antibodies. The ␣2CII adrenergic receptor is present in the axon terminals of neurons of spinal cord origin and may mediate nociceptive information (64). The urotensin II receptor has vasoactive properties, and its increased expression in cardiomyocytes correlates with cardiac dysfunction (65).
Other cell surface receptors that were identified include Roundabout 1 (Table II), which is an axon guidance receptor that controls axon crossing of the central nervous system midline (66). The platelet-derived growth factor receptor ␤ was identified with ASYM25, and it contains RG repeats (Table IV). A phosphatidylserine receptor was identified with ASYM25. The phosphatidylserine receptor is required for recognition of the asymmetrical phosphatidylserine distribution of a dying cell (67). A potassium voltage-gated channel was identified with SYM10 (Table I), and a calcium channel was identified with ASYM24 (Table III).
Intracellular proteins involved or associated with signaling proteins were also identified including the Crk-associated substrate p130cas, the breast cancer antiestrogen resistance 3 (BCAR3), the STE20-like kinase from prostate (PSK), ataxia telangiectasia (ATM), protein kinase N ␤, FLASH homolog RIP25, phosphatidylinositol-4-phosphate 3-kinase, and NOXO1. CAS is an SH3 domain-docking protein that participates in FAK-dependent cell migration (68). BCAR3 contains a putative SH2 domain and a guanine nucleotide exchange activity and has been shown to associate with CAS (69). The absence of RG repeats in CAS and BCAR3 indicates that they were likely co-purified. PSK contains several RG-rich repeats and was identified with SYM11. PSK is known to activate mitogen-activated protein kinase pathways (c-Jun NH (2)terminal kinase, p38, or extracellular signal-regulated kinase) (70). ATM is a serine kinase that is involved in DNA repair signal transduction; the absence of RG repeats indicates that it was co-purified (71). Protein kinase N ␤ is an unassigned kinase identified with ASYM24. The FLASH homolog RIP25 is known to be involved in signaling to the interleukin 2 gene expression and was identified with ASYM25. Phosphatidylinositol-4-phosphate 3-kinase C2 ␤ represents a class II PI3K and has been shown to be recruited to tyrosine kinase complexes (72). NOXO1 ␥ is an RG-containing protein purified with ASYM24 (Table III). NOXO1 was identified as a homolog of p47phox and has been shown to regulate superoxide production (73).
Several proteins implicated in apoptosis were also obtained including HtrA2, apoptosis inhibitor protein (IAP), and mortalin-2 (74). These proteins are co-purified as they are devoid of RG repeats. The identification of the phosphatidylserine receptor with ASYM25 suggests that arginine methylation may play a role in cell death.

DNA Repair
The MRE11-Rad50-NBS1 complex (75) was purified by using ASYM25. The MRE11-Rad50-NBS1 complex localizes to radiation-induced foci upon double strand DNA breaks (76). The presence of RG repeats in MRE11 suggests that MRE11 harbors an epitope for ASYM25 and that Rad50, NBS1, and ATM are co-purifying proteins with ASYM25 and SYM11.

Cytoskeleton
Many cytoskeletal proteins including keratin type II cytoskeletal 1, ␣ 1 type VII, XI, and XXII collagen precursors, spectrin, ankyrin, and dynein were purified (Tables I-IV). It is tempting to speculate that some of the components like keratin are contaminants, but the presence of abundant RG repeats makes these proteins likely targets for the DMAspecific antibodies. DISCUSSION In the present study we have identified over 200 proteins that contain RG-rich repeats and are putative substrates of protein-arginine methyltransferases. The identification and purification of proteins known to contain dimethylated arginines such as SmB/BЈ, SmD1/D3, p80-coilin, EWS, TLS, and Sam68 confirm our strategy. Most of these proteins are associated with RNA metabolism and contain RG-rich motifs. Gary and Clarke (1) searched for PRMT substrates by using databases with glycine-arginine-rich consensus sequences. The glycine-arginine-rich consensus sequence was derived from the methylation sites of fibrillarin, nucleolin, hnRNP A1, basic fibroblast growth factor, and MBP (1). Putative substrates identified included mammalian EWS, hTAFII68, hnRNPs, TLS, and ribosomal S2. Thus, the major category of putative substrates was RNA-binding proteins. By using antibodies that are specific for sDMA or aDMA, we were able to minimize the amount of RNA-binding proteins identified. Each antibody was designed with unique surrounding sequences; thus, we were able to identify proteins that could not be predicted by using a glycine-arginine-rich domain search.
The presence of Sm proteins and p80-coilin in SYM10 and SYM11 immunoprecipitations indicates the specificity of these antibodies for sDMA-containing proteins in cellular lysates. It has been reported that the abundant hnRNPs account for close to 65% of all arginine methylation that occurs in the cell (19). These proteins often contain long stretches of RG and RGG repeats that have been demonstrated to contain asymmetrically dimethylated arginines. The fact that the SYM10 and SYM11 immunoprecipitated only a few hnRNPs further demonstrates the specificity of the antibody for sDMA and not aDMA or non-methylated RG-rich proteins. We would have expected the ASYM24 and ASYM25 antibodies to recognize the spectrum of hnRNP proteins as well as the nucleolar proteins fibrillarin, nucleolin, and ribosomal protein S2. In addition, we have not identified the cell death regulator aven (77) and 53BP1 (78), two proteins known to contain extensive RG motifs. This demonstrates that the methyl antibodies that we generated recognize a subset of methylated proteins excluding the majority of hnRNPs. Actually, methylated peptides corresponding to hnRNP K arginine-glycine repeats were not recognized by ASYM24 in an enzyme-linked immunosorbent assay (26). In addition the mass spectrometry data did not identify the SMN and PRMT5 complexes in the SYM10 immunoprecipitations performed with HeLa-S3 cells. SMN and PRMT5 were readily identified in SYM10 immunoprecipitations in T4 HeLa as shown previously (27). It has been shown that SMN complexes behave differently in certain HeLa cell types (79).
Thus, the proteins that were identified through the large scale immunoprecipitation can only be present if 1) they contain DMA and are directly immunoprecipitated by the antibody, 2) they are present in a complex that contains DMA, or 3) they are contaminating proteins that were nonspecifically bound during the purification process. The presence of potential epitopes (RG repeats) suggests that most of the proteins were directly recognized by the methylarginine-specific antibodies. It is possible that some of the cytoskeletal proteins are contaminants, especially keratin, but the presence of RG repeats in their sequences provides a reason for their selective purification. The identification of many proteins in the same complex, such as the pre-mRNA complex and polyadenylation complex, allows us to conclude that DMA is potentially involved in regulating the function of that complex.
The identification of proteins involved in pre-mRNA splicing and transcription was expected (2). The identification of the G-coupled receptors, the polyadenylation complex, and proteins involved in DNA repair and cell death was unexpected. The role of sDMA in snRNP assembly is known and has been proposed as a signal for targeting to the SMN complex (31). Treatment of cell extracts with SYM10 or with methylase inhibitors impairs pre-mRNA splicing (26). The identification of several splicing factors including KSRP and the SR proteins suggests a new level of regulation of splicing by arginine methylation.
The role of arginine methylation in transcription has been known since the identification of CARM1 as a coactivator (6). The regulation in most of those studies involves the histonearginine methylation. The identification of transcription factors including TFEB, ILF proteins, and PBX as well as the CTD phosphatase, DSIF p160 transcription elongation factor, and the transcription terminator-interacting protein 5 will direct us toward new modes of transcription regulated by arginine methylation. The interplay between sDMA and transcription has been suggested (80); however, the substrates of PRMT5 were not identified. Finally, arginine methylation could also delineate the transition between promoter binding and transcription elongation, an example of which has been demonstrated for SPT5 (81).
The presence of six G-coupled receptors in DMA-specific immunoprecipitations leads us to believe that these proteins are likely to be dimethylated on arginines. The presence of RG-rich sequences further provides evidence that this will be the case. It will be interesting to determine whether the methylation interferes with the signaling or potentially with the recycling/desensitization of the receptors.
The purification of the cleavage-and polyadenylation-specific complex by SYM10 demonstrates that a protein within the complex contains sDMA (54). The presence of RG repeats in CPSF6, the 68-kDa subunit, indicates that this protein is putatively methylated by type II PRMTs. The fact that SMN associates with sDMA-containing proteins (31) and has been dubbed the "master" assembler (82) suggests that the CPSF complex formation, localization, and function may be regulated by SMN complex and/or the methylosome.
The MRE11-Rad50-NBS1 complex (75) was purified by using ASYM25. In addition, the ATM kinase was purified with SYM11. The MRE11-Rad50-NBS1 complex localizes to radiation-induced foci upon double strand DNA breaks (76), and the ATM kinase is involved in signaling DNA damage (71). The presence of RG repeats in MRE11 suggests that MRE11 harbors an epitope for ASYM25 and that Rad50, NBS, and ATM are co-purifying proteins. The purification of proteins involved in DNA repair indicates that arginine methylation will have a role in this cellular process. Arginine methylation may regulate the signaling of DNA damage, protein-protein interactions, or protein localization during DNA damage.
In conclusion, the identification of proteins containing DMA will allow us to shed some light on new roles for this posttranslational modification. The methylation of every protein on the list of putative methylated proteins will have to be confirmed, and their physiological roles will have to be demonstrated. The present study will facilitate the integration of arginine methylation in pre-mRNA splicing, protein translation, receptor signaling, transcription, DNA repair, and the cytoskeleton.