PRMT7, a New Protein Arginine Methyltransferase That Synthesizes Symmetric Dimethylarginine*

The cDNA for PRMT7, a recently discovered human protein-arginine methyltransferase (PRMT), was cloned and expressed in Escherichia coli and mammalian cells. Immunopurified PRMT7 actively methylated histones, myelin basic protein, a fragment of human fibrillarin (GAR) and spliceosomal protein SmB. Amino acid analysis showed that the modifications produced were predominantly monomethylarginine and symmetric dimethylarginine (SDMA). Examination of PRMT7 expressed in E. coli demonstrated that peptides corre-sponding to sequences contained in histone H4, myelin basic protein, and SmD3 were methylated. Furthermore, analysis of the methylated proteins showed that symmetric dimethylarginine and relatively small amounts of monomethylarginine and asymmetric dimethylarginine were produced. SDMA was

Protein methylation has been shown to occur in a diverse number of biological processes such as signal transduction (1,2), chromatin remodeling (3), RNA splicing (4,5), RNA decay (6,7), gene regulation (8,9), nuclear export of proteins (10), inhibition of oncogenic ras signal transduction (11), and bacterial chemotaxis (12,13). The methylation of protein substrates is catalyzed by enzymes that transfer a methyl group from Sadenosylmethionine to a protein acceptor, a process that occurs in many organisms. The residues that are modified by methyla-tion include arginine, histidine, lysine, and aspartic acid.
Protein-arginine methyltransferases (PRMTs) 1 catalyze the formation of methylarginine residues. Four types of protein arginine methyltransferases have been described (14). Type I PRMTs form -N G -monomethylarginine and asymmetric -N G ,N G -dimethylarginine (ADMA) residues; type II PRMTs form -N G -monomethylarginine and symmetric -N G ,N G Ј-dimethylarginine (SDMA) residues; type III and type IV PRMTs synthesize only -N G -monomethylarginine (MMA) and ␦-N Gmonomethylarginine, respectively. Although the presence of SDMA in eukaryotic cells has been known for a number of years (4,15), an enzyme that could synthesize SDMA was first identified by Pollack et al. (16) and subsequently characterized (17,18) and designated PRMT5. PRMT5 was found to methylate histones H2A and H4, myelin basic protein (MBP), and several spliceosomal Sm proteins (SmD1, SmD3 and SmB) in vitro (4, 5, 16 -21). Methylation of spliceosomal components by PRMT5 is a prerequisite for their assembly into the spliceosome (20 -22). PRMT5 also has been shown to be associated with the cyclin E gene and to be involved with other transcriptional events as well (8,23).
Although a large number of proteins have been found to contain SDMA residues (24), for several years PRMT5 was the only known Type II PRMT. Here we report the discovery of another Type II protein arginine methyltransferase, PRMT7, that can synthesize SDMA. This protein was initially characterized in hamster cells as a protein that modulates drug sensitivity to DNA-damaging agents (25).

EXPERIMENTAL PROCEDURES
Plasmid Constructs-The cDNA of PRMT7 was cloned with primers specific to the open reading frame of FLJ10640, a cDNA we identified with homology to PRMT5 and other PRMTs ( Fig. 1 and Table I). Of the three splice variants of this cDNA, the longest variant was chosen. The cDNA was amplified by PCR at 95°C for 2 min followed by 30 cycles of denaturing at 95°C for 45 s and then annealing at 55°C for 45 s and elongation at 72°C for 2 min. After 30 cycles, the reactions were then incubated at 72°C for 10 min. The PCR product was digested with EcoRI and XbaI restriction endonucleases, gel-purified, and ligated at 16°C for 21 h into vector pEF2 (16), which had been digested with EcoRI and XbaI restriction endonucleases. The primers used were 5Ј-GGAATTCCATGGACTACAAGGACGACGATGACAAGAAGAAGATC-TTCTGCAG-TCGGGCC and 5Ј-GCTCTAGAGCTCAGTCTGGGGTAT-CTGCATGC. The first primer encodes the FLAG epitope. The template FIG. 1. Pileup of the known human PRMTs. Protein sequences of PRMTs 1-7 were aligned by the GCG (Exceleris) pileup program. The Boxshade Web site was used to format the results. Comparisons were made with full-length proteins. The asterisk over position 163 of PRMT1 indicates the methionine residue in Type I PRMTs, which is replaced by cysteine and alanine in PRMT5 and PRMT7, respectively. White letters in black boxes represent identity; white letters in gray boxes represent similarity; dots represent gaps in the sequence. GenBank TM accession numbers for the PRMT proteins are as follows: PRMT1, Q99873; PRMT2, P55345; PRMT3, O60678; PRMT4, NP_954592; PRMT5, AF167572, O14744; PRMT6, Q96LA8; PRMT7, AAH00146.
for the PCR reaction was a plasmid obtained from the American Type Culture Collection (ATCC, Manassas, VA; catalog number MGC-5331) that contains the FLJ10640 cDNA. The FLJ10640 open reading frame was also amplified by PCR and digested with EcoRI restriction endonucleases and cloned into pGEX-3X at the EcoRI site to form pGEX-PRMT7. The primers used were 5Ј-GGAATTCCGATGAAGATCTTCT-GCAGTC and 5Ј-GGAATTCTCAGTCTGGGGTATCTGCATG. PCR conditions were described above.
To create the plasmid pEF2-Myc-PRMT7, the pEF2-FLAG-PRMT7 plasmid was amplified by PCR at 95°C for 2 min followed by 20 cycles of denaturing at 95°C for 45 s, annealing at 55°C for 45 s, and elongation at 72°C for 10 min with primers that encode the Myc epitope: 5Ј-GTAACGGCCGCCAGTGTGCTGGGACATGAAGATCTTCTGCAG-TCGGGCC and 5Ј-GGCCCGACTGCAGAAGATCTTCATGTCCTCCTC-AGAGATCAGCTTCTGCTCTTCCATGGAATTCCAGCACACTGGCG-GCCGTTAC. The reactions were then incubated at 72°C for 10 min. The PCR product was digested with KpnI and XbaI restriction endonucleases, gel-purified, and ligated into the vector pEF2-FLAG-PRMT7, which had been digested with KpnI and XbaI.
Purification of GST-PRMT7-Selected Escherichia coli transformants in the Rosetta strain (Novagen) were grown in TB medium (11.8 g of SELECT Peptone 140 (pancreatic digest of casein), 23.6 g of yeast extract, 9.4 g of dipotassium hydrogen phosphate, and 2.2 g of potassium dihydrogen phosphate; Invitrogen) plus 150 g/ml ampicillin to an optical density at A 600 of 0.5 at which time isopropyl 1-thio-␤-D-galactopyranoside was added to a final concentration of 1 mM. The GST-PRMT7 fusion protein was purified with glutathione-Sepharose 4B beads (Amersham Biosciences) as described (16). The bound protein was eluted with 10 mM glutathione, 50 mM Tris-HCl, pH 8.0, and 5 mM EDTA. Protein was quantified with the Bio-Rad protein assay. The SDS-PAGE of the GST-PRMT7 fusion protein is shown in Fig. 2.
Immunopurification of FLAG-PRMT7-Dishes of adherent HeLa or COS cells transfected with pEF2-FLAG-PRMT7 as described (26,27) were washed twice with ice-cold PBS and scraped with a silicone policeman in 1 ml of lysis buffer containing 50 mM sodium phosphate, pH 7.6, 1% Nonidet P-40, 0.8 mM phenylmethylsufonyl fluoride, 1 g/ml leupeptin, 2 g/ml antipain, 10 g/ml benzamidine, 10 4 KIU/ml of aprotinin, 1 g/ml chymostatin, and 1 g/ml pepstatin. The cells were incubated on ice for 30 min and then centrifuged at 14,000 rpm for 10 min in a microcentrifuge at 4°C. Supernatants were transferred to a new 1.5-ml microcentrifuge tube. To immunoprecipitate epitope-tagged PRMT7, 1.0 g of the anti-FLAG monoclonal antibody M2 (Sigma) was added to 1 ml of each lysate (supernatant) that was then incubated at 4°C overnight with rocking. Thirty l of Protein A/G Plus beads (Santa Cruz Biotechnology) were then added, and the mixture was incubated for an additional 3-4 h at 4°C with rocking. The beads were pelleted and then washed three times with 1 ml of ice-cold 50 mM sodium phosphate buffer, pH 7.6. The beads with attached immunopurified protein were stored at Ϫ80°C until used.
Preparation of Monomethylarginine (-N G -Monomethylarginine)-Conversion of Fmoc-OCOCl to Fmoc-OSu was performed as described (28,29). HOSu (0.633 g, 5.5 mmol) was added to a solution of Fmoc-OCOCl (1.29 g, 5.0 mmol) in dioxane (12.5 ml) in a 50-ml roundbottomed flask with a magnetic stirrer. It was cooled in an ice bath, and triethylamine (0.70 ml, 5.0 mmol) was added over a 5-min period with stirring. Stirring then continued at room temperature for 2 h. The precipitated triethylamine hydrochloride was filtered off (weight ϭ 0.63 g), and the filtrate was evaporated to ϳ5 ml and added portionwise to an Erlenmeyer flask containing 25 ml of diethyl ether. An oil separated out and was placed in the refrigerator overnight. The next day the supernatant solution was decanted and separated from the oil. This filtrate immediately began to crystallize and afforded 430 mg of white crystals (26% yield). Additional product left in the oil was not isolated. The product had an m.p. 143-144°C (uncorrected) (theory: m.p. 150 -153°C) and was used directly in the next step.
Preparation of Fmoc-Arg(Me)-OH-A solution of N G -methyl-L-arginine (acetate salt) (124 mg, 0.50 mmol, ICN Biomedicals, Inc., catalog number 155470) in water (1 ml) in a 10-ml round-bottomed flask was treated with triethylamine (140 l, 2.0 eq) followed by the addition of Fmoc-OSu (152 mg, 0.45 mmol, 0.90 eq) in acetonitrile (1 ml). An additional 0.30 ml of acetonitrile was required for rinsing. Magnetic stirring continued at room temperature for 30 min, and the pH was maintained at pH 8 -9 by addition of triethylamine. The resultant clear solution was evaporated to dryness with a rotary evaporator, and a white tacky semisolid residue was obtained. Water was added (5 ml), and the pH was determined to be 6 -7 and then the mixture was placed in the refrigerator overnight (4°C). The oily product that solidified overnight was isolated by filtration, washed with water, and dried in vacuo. Further drying in the dessicator gave 60 mg (32% yield) of white solid. The product was insoluble in water and ethyl acetate; soluble in acetonitrile and N,N-dimethylformamide. Deprotection of an aliquot with 15% piperidine in N,N-dimethylformamide followed by a ninhydrin test was positive and in agreement with the expected structure. The peptide, SGR Me GKGGKGLGKGGAKRHRK, was prepared by solid phase synthesis as described below for the other peptides (R Me represents -N G -monomethylarginine).
Peptides-The following peptides were synthesized chemically as substrates for PRMT7: P-SmD3, AGRGRGKAAILKAQVAARGRGRG-MGRGN-NH 2 ; P-MBP, SQGKGRGLSLSRFSWGAE-NH 2 ; M1, SGRGK-GGKGLGKGGAKRHRK-NH 2 ; GRG; M8, SGAGKGGKGLGKGGAKA-HAK-NH 2 with all arginines substituted with alanine; MM, SGR Me G-KGGKGLGKGGAKRHRK-NH 2 (R Me , monomethylarginine). Briefly, the peptides were prepared on solid phase by the Fmoc/t-butyl-strategy and O-benzotriazole-N,N,NЈ,NЈ-tetramethyluronium hexafluorophosphate-activation on a SYRO multiple peptide synthesizer (MultiSyn-Tech GmbH, Witten, Germany) as reported (30). The methylated peptide MM was synthesized until the previous glycine (position 4 of peptide MM) on a 433A batch synthesizer (Applied Biosystems, Weiterstadt, Germany), and the remaining three residues were coupled manually with 20 mg of peptide resin. After cleavage with trifluoroacetic acid the crude peptides were purified by reversed-phase high pressure liquid chromatography with an aqueous acetonitrile gradient with 0.1% trifluoroacetic acid. The purified peptides were characterized by matrix-assisted laser desorption/ionization mass spectrometry and stored dry at Ϫ20°C until used.
In Vitro Methylation Reactions-The 30-l reaction mixtures included 1-10 g of substrate protein or peptide as noted in the figure legends, 5 l of [ 3 H]S-adenosylmethionine (81 Ci/mmol, PerkinElmer Life Sciences), 0.1-1 g of GST-PRMT7, or 15 l of protein A/G PLUS beads (containing the immunopurified FLAG-PRMT7 isolated from ϳ1 ϫ 10 7 HeLa or COS cells), plus buffer (50 mM sodium phosphate, pH 7.6) to make the final concentration 8.3 mM sodium phosphate, pH 7.6. The 30-l reaction mixtures were incubated at 37°C for 5-21 h as described in the figure legends. To determine the incorporation, aliquots of the reaction mixtures were precipitated with cold 10% trichloroacetic acid onto 0.45 m nitrocellulose filters (HA, Millipore) as described (31).
Gel Electrophoresis of Methylated Proteins-Aliquots of the in vitro methylation reactions were electrophoresed on precast 15% polyacryl-

FIG. 2. Gel electrophoresis of GST-PRMT7.
GST-PRMT7 was purified as described under "Experimental Procedures," and 2.5 g of the preparation was analyzed by SDS-PAGE on 12.5% gels. The band at just over 100 kDa is GST-PRMT7. The band at 27 kDa is GST.
amide gels (Novex, Invitrogen). Gels were stained with 50% methanol, 10% acetic acid, and 0.25% Coomassie Brilliant Blue in water and destained with 30% methanol and 10% acetic acid in water and then dried and exposed to Biomax film at Ϫ80°C for 14 days.
TLC-In vitro methylation reactions (30 l) were hydrolyzed with 250 l of 6 N HCl at 110°C for 21 h in a sealed glass ampule. The hydrolyzed amino acids were dried in an oven after opening the ampule. Thirty l of water was added to the dried residue, and then 10 l of the solution was applied to each lane of a Silica 60 TLC plate (Whatman). The solvent of 30% ammonium hydroxide, chloroform, methanol, and water (2:0.5:4.5:1) was used for the chromatographic separations of the amino acids (5). Color was developed with a ninhydrin spray (Sigma). Standards (MMA, ADMA, and SDMA) were purchased from Calbiochem. Chromatographs were coated with three applications of EN 3 HANCE Spray (PerkinElmer Life Sciences) and then exposed to film at Ϫ80°C for 7-21 days.
Analysis of the Distribution of PRMT7 in Cells-COS-1 cells were plated at 3 ϫ 10 4 cells per chamber in LabTek II 4-well chamber slides (VWR) on the day before transfection. Plasmid pEF2-Myc-PRMT7 was transfected into COS-1 cells with Superfectamine transfection reagent (Qiagen) as described above under "Immunopurification of FLAG-PRTM7." Plasmid pEF2-Myc-PRMT5 or pEF2-Myc vector was transfected at the same time as positive and negative controls, respectively. 24 -36 h after transfection, the cells were washed three times with PBS, pH 7.4 and fixed in 3.7% formaldehyde in PBS for 15 min at room temperature (32). The cells were then washed three times in PBS and permeabilized in 0.2% Triton X-100 plus 1% normal goat serum in PBS  for 5 min on ice. The cells were washed three times in PBS plus 1% normal goat serum and incubated in 1:200 diluted rabbit anti-Myc antibody (Santa Cruz Biotechnology, Inc.) for 1 h at room temperature. The cells were then washed three times in PBS plus 1% normal goat serum followed by incubation in rhodamine-conjugated goat anti-rabbit antibody (Santa Cruz Biotechnology, Inc.) at a dilution of 1:150 for 1 h at room temperature. The cells were subsequently washed three times in PBS. The chamber was removed from the slide and then mounting medium (Vector Laboratories, Inc., Burlingame, CA) containing 4Ј,6diamidino-2-phenylindole was added to cells. A coverslip was placed over the cells on the slide and sealed with clear nail polish. The stained cells were visualized under a Nikon eclipse TE200 microscope with ϫ200 magnification, and photographs were taken by a CoolSNAP Pro digital camera (Media Cybernetics, Houston, TX) with Image Pro Plus software.

RESULTS
The Sequence Motif Search Program (Kyoto University) was used to search for proteins containing consensus S-adenosylmethionine binding sites. One protein identified was encoded by the cDNA designated FLJ10640, which contained a motif for an S-adenosylmethionine binding site and other methyltransferase motifs suggesting that it might be a PRMT. We designated the protein encoded by the FLJ10640 cDNA PRMT7. When the protein sequence of PRMT7 was compared with the other known human PRMTs, PRMT7 seemed to be a member of the human PRMT family (Fig. 1). Table I shows a summary of the homology of all the known protein arginine methyltransferases (PRMT1 through PRMT7) with respect to each other. PRMT7 is most similar to PRMT4.
To determine whether PRMT7 contains methyltransferase activity, FLAG-PRMT7 was expressed in HeLa cells and immunopurified with anti-FLAG antibody. The immunopurified protein was assayed in vitro for methylation activity with four different substrate proteins. Fig. 3A shows that the FLAG-PRMT7 preparation has significant methyltransferase activity with histones, MBP, GAR (a fragment of human fibrillarin), and SmB (a spliceosomal Sm protein) as substrates. Fig. 3B shows both a Coomassie Blue-stained gel and the autoradiograph of the proteins methylated in these reactions. Of the five histones (H1, H2A, H2B, H3, and H4) in the preparation that was assayed only two were methylated. Based on their size and the results of a control experiment in which the five histones were methylated individually (data not shown), we concluded that the methylated (labeled) histones are H2A and H4. MBP was also methylated. Although low quantities of GAR and SmB were present in the methylation reactions, their methylation was nevertheless appreciable judging from the intensity of the autoradiographic bands. Because PRMT5 methylates these same proteins in vitro, these data suggested that there might be some functional similarity between PRMT7 and PRMT5 (16,18). Alternatively, the similarity of substrates for PRMT5 and PRMT7 could be caused by association of PRMT5 with PRMT7. To rule out this possibility, hemagglutinin-PRMT5 and FLAG-PRMT7 were expressed in COS cells. After the cells were lysed, FLAG-PRMT7 was immunoprecipitated as described under "Experimental Procedures." The immunoprecipitated FLAG-PRMT7 was analyzed by blotting and probing with anti-hemagglutinin antibody. The results did not show any detectable hemagglutinin-PRMT5 associated with PRMT7, thereby eliminating the possibility that the activity of immunopurified PRMT7 was due to the presence of PRMT5 (Fig. 4).
Next we determined the type of methylation produced by the immunopurified PRMT7. In this experiment, COS cells were transiently transfected with pEF2-FLAG-PRMT7, and the protein was assayed for activity by in vitro methylation. As shown in Fig. 5, SDMA, ADMA, and MMA were not found in detectable quantities in the absence of substrate (lane 1). However, in the presence of histones, SDMA was clearly produced (lane 2).
No ADMA or MMA was detectable in this case. When the MBP peptide (P-MBP) was methylated, no detectable SDMA was produced, and the major product was MMA (lane 3). Nevertheless, the formation of SDMA in histone substrates suggested that PRMT7 might be a Type II PRMT.
PRMT7 was expressed in E. coli and purified by affinity chromatography as described under "Experimental Procedures." The GST-PRMT7 fusion protein was used to methylate various peptides in vitro. Fig. 6A shows that methylation occurred with peptides M1, P-SmD3, and P-MBP as substrates. Because the GRG tripeptide was not trichloroacetic acid-precipitated because of its small size, it was analyzed with the other peptides after hydrolysis and TLC. As is shown in Fig. 6B  less ADMA and MMA, indicating that PRMT7 is in fact a Type II PRMT. The fact that the GRG tripeptide was methylated by PRMT7 was surprising. Nevertheless, this result demonstrates that the GRG motif is sufficient for methylation by PRMT7.
To confirm that SDMA is synthesized by PRMT7, we used GST-PRMT7 produced in E. coli with protein (Fig. 7A) and peptide (Fig. 7B) substrates. GST-PRMT7 methylated the proteins H2A, MBP, and SmD1, although SmD1 was methylated at a low level. GST-PRMT7 also methylated the peptides M1 and MM. Peptide MM contained a G-R Me -G (where R Me is methylarginine) instead of the usual GRG motif present in peptide M1. PRMT7 methylated MM significantly greater than it labeled M1 (Fig. 7B). When TLC was performed to determine the products of the methylation, it was evident that GST-PRMT7 alone produced no detectable SDMA, ADMA or MMA (Fig. 7C, lane 1). However, PRMT7 yielded mostly SDMA with the M1 peptide (lane 2). When the monomethylated peptide (MM) was tested, once again SDMA was synthesized (lane 3), indicating that PRMT7 is a Type II PRMT. When all three of the arginines in peptide M1 were substituted with alanines, this peptide designated M8 was not methylated (lane 4) showing that PRMT7 methylates only arginine residues to synthesize SDMA predominantly.
Because Miranda et al. (33) found that PRMT7 synthesizes MMA but no dimethylarginines with a peptide substrate, we tested the effects of peptide concentration on the relative levels of MMA and SDMA produced by PRMT7. Whereas Miranda et al. (33) used 50 g of peptide we used 3 g or less of several peptides. Therefore we hypothesized that higher concentrations of peptides would reduce the quantity of SDMA, whereas the level of MMA would increase. As shown in Fig. 8, the ratio of SDMA to MMA was progressively reduced as the quantity of P-SmD3 substrate was increased from 1 to 50 g in the methylation reaction. As the increased concentration of substrate increased the amount of MMA and decreased the level of SDMA, we propose a model of peptide substrate methylation by PRMT7 in which the monomethyl peptide is released from PRMT7 after methylation (Fig. 9). Increasing the substrate concentration would increase monomethylation if the substrate concentration was not saturating. However, because the ratio of monomethylated to unmethylated peptides would be smaller at higher substrate concentrations, the methylation of monomethylated peptides is reduced because the unmethylated peptides compete with the released monomethyl peptides. Thus, less dimethylation occurs at increasing concentrations of peptide substrate. In the case of PRMT7, 10 g or more of P-SmD3 reduced the formation of SDMA to undetectable levels (Fig. 8) consistent with our model (Fig. 9).
To determine the distribution of PRMT7 in cells, Myc-PRMT7 was expressed in COS cells (Fig. 10). The data show that PRMT7 is present in the nucleus and the cytosol in comparison to PRMT5 that is present primarily in the cytosol (19,34). DISCUSSION Of the six known human PRMTs, only PRMT5 has thus far been identified as a Type II methyltransferase capable of synthesizing SDMA in proteins. In this report, we describe a new protein arginine methyltransferase, PRMT7. Our data demonstrate that PRMT7 is also capable of forming SDMA so that it is the second PRMT able to synthesize SDMA. While this paper was in preparation, we became aware of a publication in press that also described PRMT7 (33). However, the authors reported that PRMT7 synthesized only -N G -monomethylarginine but did not observe any ADMA or SDMA. In addition, the report noted that a peptide was methylated but that no protein substrate tested was methylated. In contrast, our results with PRMT7 demonstrated robust methylation of proteins and significant methylation of peptides, even the tripeptide GRG. Histone H2A, MBP, and GST-GAR were methylated by PRMT7 that we prepared (Figs. 3, 5, and 7) but were not observed to be methylated by Miranda et al. (33). Furthermore, we observed that the methylation yielded SDMA predominantly with peptides and proteins consistent with the designation of PRMT7 as a Type II PRMT. The fact that the tripeptide GRG alone can be methylated indicates that at least one motif for methylation by PRMT7 is the GRG sequence. The discrepancies between our observations and the report of Miranda et al. (33) are likely because of the concentrations of the peptides used. We demonstrated that at high concentrations of substrate there was no formation of SDMA but only MMA (Fig. 8), whereas at low peptide concentrations SDMA formation was significant. This result is consistent with our model (Fig. 9). It is thus likely that after formation of MMA, the peptide is released from PRMT7. This may not apply to protein substrates that have binding sites that keep them attached specifically to PRMT7 during methylation. Furthermore, assessing proteins that complex with PRMT7 may assist in the recognition of protein substrates. It should be noted that previously no purified PRMT was shown to synthesize SDMA or ADMA with small peptide substrates as we have shown here, although Brahms et al. (20) observed SDMA synthesis with a BSA-peptide conjugate by a cytosolic lysate of HeLa cells. The ability to use peptide substrates readily enables the definition of the motifs and the residues that are methylated as we showed above.
SDMA has been found in Sm proteins (4,20) and coilin (35). Because PRMT5 was found in these complexes and was the only known Type II PRMT, it was concluded that PRMT5 was the enzyme that carried out synthesis of SDMA residues in the relevant proteins. However, the discovery that PRMT7 is a Type II enzyme suggests that it is likely responsible for synthesis of SDMA residues in some proteins. In addition, PRMT5 and PRMT7 could assist each other, substitute for each other, or possibly function in different pathways by methylating different sets of proteins in vivo. Possibly PRMT5 and PRMT7 may methylate the same proteins under different physiological circumstances or in different cellular compartments. Isolation of proteins with antibodies to SDMA has identified over 100 proteins that contain SDMA residues (24). It was assumed that PRMT5 was responsible for the SDMA post-translational modifications because no other Type II enzyme had been identified. However, it is likely that PRMT7 is responsible for methylation of some of these proteins.
The activity of GST-PRMT7 is lower than that of immunopurified PRMT7. A similar phenomenon was noted for PRMT5 when we found that a great deal of methyltransferase activity was associated with the immunoprecipitated protein but very little or no methylation could be obtained with GST-PRMT5 isolated from E. coli. 2 This difference is likely because of the requirement of accessory proteins for full activity of PRMT5 prepared from mammalian cells, and the lack of these proteins in preparations from bacteria could account for the lower activity of PRMT5 expressed in E. coli. It is possible that PRMT7 also requires one or more associated proteins for activity. The methylation of the GRG tripeptide by GST-PRMT7 (Fig. 6B) is strong evidence that PRMT7 methylates substrates at arginine residues that are flanked by two glycine residues. Methylation of GRG-containing proteins (MBP, GAR, SmB, H2A, and H4, Fig. 3, A and B) and peptides (P-SmD3, P-MBP, Fig. 6A; and M1, MM, Fig. 7, B and C) but not a peptide without an intact GRG (M8, Fig. 7, B and C) is further evidence in support of this conclusion. Additional work will be required to determine whether the residues flanking the arginine residues contribute to the extent of methylation or to the type of modification produced.
As PRMT5 and PRMT7 are the only known Type II PRMTs, it is informative to compare their sequences to begin to determine which residues in the protein are required to form SDMA. In a previous study, we and colleagues hypothesized that differences in the residues residing in the catalytic pocket of Type I and Type II PRMTs may account for the different modifications produced (18). It was predicted that substitution of a serine for methionine in the active site would allow SDMA to be synthesized rather than ADMA (18). In our sequence alignments of the PRMTs, including PRMT7 (Fig. 1), however, a cysteine is substituted for the methionine in PRMT5 and an alanine in PRMT7. Both of these substitutions could allow for SDMA synthesis because of the shorter side chains of cysteine and alanine compared with the methionine that is present in all Type I PRMTs. Other sequence differences between the Type I and Type II PRMTs may contribute to the formation of SDMA (Fig. 1). These structure/function relationships could be defined further by site-specific mutagenesis.
PRMT7 predominantly produces SDMA and small amounts of MMA and ADMA. The same phenomenon has been observed with PRMT5 (36). This indicates that the production of ADMA, albeit in small quantities, might be a consistent property of Type II PRMTs. Because the GRG tripeptide becomes methylated with ADMA slightly and because only one arginine is available to methylate, PRMT7 synthesizes both SDMA and ADMA on the same arginine residue. Thus, it appears that synthesis of SDMA and ADMA is a function of the enzyme rather than the substrate. One consequence of the production of both ADMA and SDMA by Type II PRMTs is that the presence of ADMA in proteins is not a guarantee that the methylation was produced by a Type I methyltransferase. Our experiments with PRMT1 have not detected any SDMA in proteins methylated by this Type I protein arginine methyltransferase in vitro. 3 Elevated plasma SDMA has been implicated in kidney disease (37). In addition, plasma ADMA levels have been shown to be elevated in various cardiovascular diseases (37)(38)(39)(40)(41). These observations suggest that the relative amounts of SDMA and ADMA may have significant ramifications in health and disease and that both PRMT5 and PRMT7 may have specific roles in controlling these levels.