Type I protein arginine methyltransferases catalyze the formation of asymmetric ω-N G,N G-dimethylarginine residues by transferring methyl groups fromS-adenosyl-l-methionine to guanidino groups of arginine residues in a variety of eucaryotic proteins. The predominant type I enzyme activity is found in mammalian cells as a high molecular weight complex (300–400 kDa). In a previous study, this protein arginine methyltransferase activity was identified as an additional activity of 10-formyltetrahydrofolate dehydrogenase (FDH) protein. However, immunodepletion of FDH activity in RAT1 cells and in murine tissue extracts with antibody to FDH does not diminish type I methyltransferase activity toward the methyl-accepting substrates glutathione S-transferase fibrillarin glycine arginine domain fusion protein or heterogeneous nuclear ribonucleoprotein A1. Similarly, immunodepletion with anti-FDH antibody does not remove the endogenous methylating activity for hypomethylated proteins present in extracts from adenosine dialdehyde-treated RAT1 cells. In contrast, anti-PRMT1 antibody can remove PRMT1 activity from RAT1 extracts, murine tissue extracts, and purified rat liver FDH preparations. Tissue extracts from FDH(+/+), FDH(+/−), and FDH(−/−) mice have similar protein arginine methyltransferase activities but high, intermediate, and undetectable FDH activities, respectively. Recombinant glutathioneS-transferase-PRMT1, but not purified FDH, can be cross-linked to the methyl-donor substrateS-adenosyl-l-methionine. We conclude that PRMT1 contributes the major type I protein arginine methyltransferase enzyme activity present in mammalian cells and tissues. Type I protein arginine methyltransferases catalyze the formation of asymmetric ω-N G,N G-dimethylarginine residues by transferring methyl groups fromS-adenosyl-l-methionine to guanidino groups of arginine residues in a variety of eucaryotic proteins. The predominant type I enzyme activity is found in mammalian cells as a high molecular weight complex (300–400 kDa). In a previous study, this protein arginine methyltransferase activity was identified as an additional activity of 10-formyltetrahydrofolate dehydrogenase (FDH) protein. However, immunodepletion of FDH activity in RAT1 cells and in murine tissue extracts with antibody to FDH does not diminish type I methyltransferase activity toward the methyl-accepting substrates glutathione S-transferase fibrillarin glycine arginine domain fusion protein or heterogeneous nuclear ribonucleoprotein A1. Similarly, immunodepletion with anti-FDH antibody does not remove the endogenous methylating activity for hypomethylated proteins present in extracts from adenosine dialdehyde-treated RAT1 cells. In contrast, anti-PRMT1 antibody can remove PRMT1 activity from RAT1 extracts, murine tissue extracts, and purified rat liver FDH preparations. Tissue extracts from FDH(+/+), FDH(+/−), and FDH(−/−) mice have similar protein arginine methyltransferase activities but high, intermediate, and undetectable FDH activities, respectively. Recombinant glutathioneS-transferase-PRMT1, but not purified FDH, can be cross-linked to the methyl-donor substrateS-adenosyl-l-methionine. We conclude that PRMT1 contributes the major type I protein arginine methyltransferase enzyme activity present in mammalian cells and tissues. protein arginine N-methyltransferase S-adenosyl-l-methionine glutathioneS-transferase fibrillarin glycine-arginine domain fusion protein polyacrylamide gel electrophoresis heterogeneous nuclear ribonucleoprotein 10-formyl-5,6,7,8-tetrahydrofolate 10-formyltetrahydrofolate dehydrogenase adenosine dialdehyde coactivator-associated arginine methyltransferase 1 phosphate-buffered saline S-adenosylhomocysteine Arginine methylation in proteins was discovered over 30 years ago (1.Paik W.K. Kim S. J. Biol. Chem. 1968; 243: 2108-2114Abstract Full Text PDF PubMed Google Scholar, 2.Paik W.K. Kim S. Biochem. Biophys. Res. Commun. 1967; 29: 14-20Crossref PubMed Scopus (121) Google Scholar). At least two types of protein arginineN- methyltransferase (PRMT)1 activities that transfer methyl groups fromS-adenosyl-l-methionine (AdoMet) to the guanidino group of arginine residues exist in mammalian cells (3.Lee H.W. Kim S. Paik W.K. Biochemistry. 1977; 16: 78-85Crossref PubMed Scopus (74) Google Scholar). Type I PRMT enzymes catalyze the formation of ω-monomethylarginine and asymmetric ω-N G,N G-dimethylarginine. Type I substrates include many RNA binding and transporting proteins, transcription factors, nuclear matrix proteins, and cytokines (4.Gary J.D. Clarke S. Prog. Nucleic Acids Res. Mol. Biol. 1998; 61: 65-131Crossref PubMed Google Scholar). Functions of type I arginine methylation in proteins may include regulation of transcription, modulation of the affinity of nucleic acid-binding proteins, regulation of interferon signaling pathways, and targeting of nuclear proteins (4.Gary J.D. Clarke S. Prog. Nucleic Acids Res. Mol. Biol. 1998; 61: 65-131Crossref PubMed Google Scholar, 5.Abramovich C. Yakobson B. Chebath J. Revel M. EMBO J. 1997; 16: 260-266Crossref PubMed Scopus (155) Google Scholar, 6.Altschuler L. Wook J.O. Gurari D. Chebath J. Revel M. J. Interferon Cytokine Res. 1999; 19: 189-195Crossref PubMed Scopus (40) Google Scholar, 7.Pintucci G. Quarto N. Rifkin D.B. Mol. Biol. Cell. 1996; 7: 1249-1258Crossref PubMed Scopus (61) Google Scholar, 8.Shen E.C. Henry M.F. Weiss V.H. Valentini S.R. Silver P.A. Lee M.S. Genes Dev. 1998; 12: 679-691Crossref PubMed Scopus (255) Google Scholar). Type II enzymes catalyze the formation of ω-monomethylarginine and symmetric ω-N G,N′G-dimethylarginine (9.Ghosh S.K. Paik W.K. Kim S. J. Biol. Chem. 1988; 263: 19024-19033Abstract Full Text PDF PubMed Google Scholar, 10.Young P.R. Waickus C.M. Biochem. J. 1988; 250: 221-226Crossref PubMed Scopus (9) Google Scholar). Myelin basic protein is the only known substrate for type II arginine methyltransferase activity (4.Gary J.D. Clarke S. Prog. Nucleic Acids Res. Mol. Biol. 1998; 61: 65-131Crossref PubMed Google Scholar). The type III enzyme, discovered in yeast, catalyzes the monomethylation of the internal δ-guanidino nitrogen atom of arginine residues (11.Zobel-Thropp P. Gary J.D. Clarke S. J. Biol. Chem. 1998; 273: 29283-29286Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). Four enzymatically active type I protein arginineN-methyltransferases have been reported: PRMT1 (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar), PRMT3 (13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar), and coactivator-associated arginine methyltransferase 1 (CARM1) (14.Chen D. Ma H. Hong H. Koh S.S. Huang S.M. Schurter B.T. Aswad D.W. Stallcup M.R. Science. 1999; 284: 2174-2177Crossref PubMed Scopus (1020) Google Scholar) from mammalian cells and arginine methyltransferase I (RMT1) from yeast (15.Gary J.D. Lin W.J. Yang M.C. Herschman H.R. Clarke S. J. Biol. Chem. 1996; 271: 12585-12594Abstract Full Text Full Text PDF PubMed Scopus (185) Google Scholar). Knockout of RMT1, the only type I PRMT gene in yeast, has no obvious phenotype. However, a mutant allele ofRMT1 is synthetically lethal to yeast in combination with a temperature-sensitive mutant allele of NPL3 (8.Shen E.C. Henry M.F. Weiss V.H. Valentini S.R. Silver P.A. Lee M.S. Genes Dev. 1998; 12: 679-691Crossref PubMed Scopus (255) Google Scholar). NPL3 is an RMT1 substrate involved in nuclear protein import, pre-RNA processing, and export of mRNA from the nucleus (8.Shen E.C. Henry M.F. Weiss V.H. Valentini S.R. Silver P.A. Lee M.S. Genes Dev. 1998; 12: 679-691Crossref PubMed Scopus (255) Google Scholar). PRMT1, the first protein arginine N-methyltransferase in mammalian cells to be cloned, was discovered as a protein interacting with the immediate-early gene products BTG1 and TIS21 (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar). BTG1 and TIS21 are negative regulators of cell growth whose overexpression in cells can lead to cell growth arrest (16.Rouault J.P. Rimokh R. Tessa C. Paranhos G. Ffrench M. Duret L. Garoccio M. Germain D. Samarut J. Magaud J.P. EMBO J. 1992; 11: 1663-1670Crossref PubMed Scopus (274) Google Scholar, 17.Rouault J.P. Falette N. Guehenneux F. Guillot C. Rimokh R. Wang Q. Berthet C. Moyret-Lalle C. Savatier P. Pain B. Shaw P. Berger R. Samarut J. Magaud J.P. Ozturk M. Samarut C. Puisieux A. Nat. Genet. 1996; 14: 482-486Crossref PubMed Scopus (359) Google Scholar). BTG1 and TIS21 interact with PRMT1 and regulate its enzymatic activity (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar). PRMT1 also associates with the interferon α/β receptor (5.Abramovich C. Yakobson B. Chebath J. Revel M. EMBO J. 1997; 16: 260-266Crossref PubMed Scopus (155) Google Scholar, 6.Altschuler L. Wook J.O. Gurari D. Chebath J. Revel M. J. Interferon Cytokine Res. 1999; 19: 189-195Crossref PubMed Scopus (40) Google Scholar). PRMT1, a predominantly nuclear protein, exists in a large complex of 300–400 kDa (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar) and methylates arginine residues in RGG and RXR motifs of many RNA-binding proteins and other proteins (4.Gary J.D. Clarke S. Prog. Nucleic Acids Res. Mol. Biol. 1998; 61: 65-131Crossref PubMed Google Scholar, 18.Smith J.J. Rucknagel K.P. Schierhorn A. Tang J. Nemeth A. Linder M. Herschman H.R. Wahle E. J. Biol. Chem. 1999; 274: 13229-13234Abstract Full Text Full Text PDF PubMed Scopus (136) Google Scholar). PRMT2 was identified because of its sequence similarity to PRMT1 (19.Scott H.S. Antonarakis S.E. Lalioti M.D. Rossier C. Silver P.A. Henry M.F. Genomics. 1998; 48: 330-340Crossref PubMed Scopus (144) Google Scholar). To date no methyltransferase activity has been demonstrated for PRMT2. PRMT3 is a monomeric cytoplasmic protein whose activity overlaps with that of PRMT1 (13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar). CARM1, the third active mammalian arginine methyltransferase to be discovered, was cloned as a protein interacting with the carboxyl-terminal region of p160 coactivator (14.Chen D. Ma H. Hong H. Koh S.S. Huang S.M. Schurter B.T. Aswad D.W. Stallcup M.R. Science. 1999; 284: 2174-2177Crossref PubMed Scopus (1020) Google Scholar). PRMT1, PRMT2, PRMT3, CARM1, and yeast RMT1 all contain signature regions (I, post-I, -II, and -III) that constitute the core of the AdoMet-binding site (20.Kagan R.M. Clarke S. Arch. Biochem. Biophys. 1994; 310: 417-427Crossref PubMed Scopus (429) Google Scholar). In a recent study, the predominant protein arginineN-methyltransferase was purified from rat liver (21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar). Sequence analysis identified the major polypeptide in this preparation as 10-formyltetrahydrofolate dehydrogenase (FDH, EC 1.5.1.6), suggesting that the major protein arginine methyltransferase may be encoded by a gene encoding an enzyme involved in folate metabolism. FDH catalyzes (i) NADP+-dependent oxidation of 10-formyltetrahydrofolate (10-FTHF) to tetrahydrofolate, NADPH, and CO2; (ii) NADP+-independent hydrolysis of 10-FTHF to formate and tetrahydrofolate; and (iii) NADP+-dependent oxidation of 2-propanal (22.Cook R.J. Lloyd R.S. Wagner C. J. Biol. Chem. 1991; 266: 4965-4973Abstract Full Text PDF PubMed Google Scholar,23.Cook R.J. Wagner C. Arch. Biochem. Biophys. 1995; 321: 336-344Crossref PubMed Scopus (13) Google Scholar). FDH purified from rat liver exists as a tetramer of identical 99-kDa subunits (22.Cook R.J. Lloyd R.S. Wagner C. J. Biol. Chem. 1991; 266: 4965-4973Abstract Full Text PDF PubMed Google Scholar, 24.Scrutton M.C. Beis I. Biochem. J. 1979; 177: 833-846Crossref PubMed Scopus (29) Google Scholar, 25.Min H. Shane B. Stokstad E.L. Biochim. Biophys. Acta. 1988; 967: 348-353Crossref PubMed Scopus (52) Google Scholar). The cDNA-deduced FDH amino acid sequence contains several domains (22.Cook R.J. Lloyd R.S. Wagner C. J. Biol. Chem. 1991; 266: 4965-4973Abstract Full Text PDF PubMed Google Scholar, 26.Schirch D. Villar E. Maras B. Barra D. Schirch V. J. Biol. Chem. 1994; 269: 24728-24735Abstract Full Text PDF PubMed Google Scholar), including the amino-terminal phosphoribosyl-glycinamide formyltransferase homologous domain (amino acids 1–203) (27.Krupenko S.A. Wagner C. Cook R.J. J. Biol. Chem. 1997; 272: 10273-10278Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar) and the carboxyl-terminal aldehyde dehydrogenase homologous domain (amino acids 417–902) (28.Krupenko S.A. Wagner C. Cook R.J. J. Biol. Chem. 1997; 272: 10266-10272Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar). However, FDH does not contain any methyltransferase signature sequences (20.Kagan R.M. Clarke S. Arch. Biochem. Biophys. 1994; 310: 417-427Crossref PubMed Scopus (429) Google Scholar). To understand further the relationship between the PRMT and FDH gene products, we sought to identify the major type I protein arginineN-methyltransferase in cells and tissues. We used cultured rat cells and tissues from FDH(+/+), FDH(+/−), and FDH(−/−) mice to determine (i) which methyltransferase is the predominant type I protein arginine methyltransferase and (ii) whether the protein arginine methyltransferase activity in FDH enzyme preparations is catalyzed by the FDH enzyme or by a copurified, but distinct, methyltransferase. S-Adenosyl-l-[methyl-3H]-l-methionine (specific activity about 75 Ci/mmol) was obtained from NEN Life Science Products or Amersham Pharmacia Biotech. 10-Formyl-5,8-dideazafolate was obtained from Dr. John B. Hynes, Department of Pharmaceutical Chemistry, Medical University of South Carolina. 5-Formyltetrahydrofolate (FTHF) affinity gel was prepared by covalently linking 5-FTHF to AH-Sepharose 4B (29.Cook R.J. Wagner C. Methods Enzymol. 1986; 122: 251-255Crossref PubMed Scopus (14) Google Scholar). Recombinant GST-PRMT1, GST-GAR, and hnRNP A1 were purified as described previously (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar). PRMT1 and PRMT3 antibodies were described previously (13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar). Anti-FDH antibody was generated in rabbits using purified rat liver FDH as the antigen (27.Krupenko S.A. Wagner C. Cook R.J. J. Biol. Chem. 1997; 272: 10273-10278Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar). Methyltransferase activity was assayed at 37 °C for times and in final volumes as specified in figure legends. The methyl-donor substrate was [3H]AdoMet. The methyl-accepting substrates were GST-GAR, hnRNP A1, or hypomethylated proteins present in lysates from adenosine dialdehyde-treated RAT1 cells (13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar). Methylation reactions were stopped by adding SDS-PAGE sample buffer and resolved on SDS-PAGE. The methylated proteins were visualized by fluorography, as described previously (13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar). A separate set of assays on murine liver extracts were performed as described by Kim et al. (21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar, 29.Cook R.J. Wagner C. Methods Enzymol. 1986; 122: 251-255Crossref PubMed Scopus (14) Google Scholar). FDH activity was determined as described previously by Cook (30.Cook R.J. Methods Enzymol. 1997; 281: 129-134Crossref PubMed Scopus (5) Google Scholar). The 1.0-ml reaction mixture contained 100 mm HEPES pH 7.8 buffer, 100 μm 2-mercaptoethanol, 61.6 μm10-formyl-5,8-dideazafolate, 100 μm NADP+, and cell extract. The production of 5,8-dideazafolate was monitored by the increase in absorbance at 295 nm for 15 min at 23 °C. Rat liver FDH was purified, as described previously (29.Cook R.J. Wagner C. Methods Enzymol. 1986; 122: 251-255Crossref PubMed Scopus (14) Google Scholar), through the 5-FTHF-Sepharose substrate affinity column step. The final purified FDH preparation contains only a major band on SDS-PAGE gel at 110 kDa, when proteins on the gel are visualized by Coomassie Blue staining. Protein samples were subjected to SDS-PAGE and immunoblotting analysis or silver staining as described previously (13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar). Specific conditions for immunoprecipitation with anti-PRMT1, anti-PRMT3, anti-FDH, and control antibodies are described in the relevant figure legends. In general, cell lysates were incubated with antibodies and protein A-Sepharose 4B at 4 °C for 90–120 min. Supernatant and pellet fractions were recovered. The pellet fractions were further washed for 10 min with PBS containing 0.5% Triton X-100 and 0.5% Tween 20 for 3 times and then with PBS to remove detergents. The precipitated proteins on protein A-Sepharose beads were used directly in in vitromethyltransferase assays. RAT1 cells are cultured in Dulbecco's modified Eagle's medium, 10% fetal bovine serum in the presence of 20 μm Adox for 48 h. Cells are washed twice with PBS and harvested in PBS with protease inhibitors (Roche Molecular Biochemicals). Cells are lysed by brief low power sonication, and cell lysate was subjected to 5 min of centrifugation. The supernatant fraction was collected as the Adox-treated RAT1 cell lysate (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar). Mouse tissues were homogenized in phosphate-buffered saline (PBS) containing protease inhibitor mix (Roche Molecular Biochemicals). The crude extract was first centrifuged at 3,000 × g for 10 min at 4 °C to remove tissue debris. The supernatant fraction was recovered and subjected to centrifugation at 100,000 × g for 30 min at 4 °C. The supernatant fraction was used as the cell lysate from mouse tissues. Photoaffinity labeling of methyltransferase enzymes with S-adenosyl-l-methionine was performed as described previously by Hurst et al. (31.Hurst J.H. Billingsley M.L. Lovenberg W. Biochem. Biophys. Res. Commun. 1984; 122: 499-508Crossref PubMed Scopus (32) Google Scholar). FDH was previously purified as the predominant arginine methyltransferase from rat liver (21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar). To test whether FDH activity and protein arginine methyltransferase activity in this preparation can be further fractionated, we prepared FDH protein through the substrate affinity chromatography step. The FDH protein preparation contains a predominant polypeptide of 110 kDa. With extensive overloading of FDH proteins, some minor contaminating proteins are visible on SDS-PAGE when the gel is stained with Coomassie Blue dye (data not shown). Among these minor contaminating proteins, a polypeptide band of 45 kDa was detected. This purified FDH preparation is active in in vitromethyltransferase assays, using GST-GAR as the methyl-accepting substrate (Fig. 1, panel A, lane 7). However, the specific activity of the purified FDH preparation is less than 1% of the specific activity of purified GST-PRMT1. To determine whether the methyltransferase activity in the purified FDH protein preparation is intrinsically associated with FDH protein, we used antisera against PRMT1 and FDH, as well as control antiserum, to immunoprecipitate the PRMT activity present in the FDH protein preparation. Neither anti-FDH antibody (panel A, lanes 5 and6) nor the control antibody (panel A, lanes 3 and4) can precipitate significant methyltransferase activity in the purified FDH protein preparation. In contrast, anti-PRMT1 antibody precipitates most of the GST-GAR-methylating activity in the FDH enzyme preparation (panel A, lanes 1 and 2). To confirm the ability of the anti-FDH antibody to bind and immunoprecipitate FDH, we used anti-FDH antibody to immunodeplete FDH activity from the FDH preparation (Fig. 1, panel B). Anti-FDH antibody immunodepletes 90% of the FDH activity. In contrast, anti-PRMT1 antibody does not immunoprecipitate FDH activity. We conclude that anti-PRMT1 and anti-FDH specifically immunoprecipitate PRMT1 and FDH activities; the PRMT activity present in the purified FDH protein preparation can be separated from the FDH activity. In our previous studies we identified two PRMTs (PRMT1 and PRMT3) in cultured RAT1 fibroblast cells (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar). PRMT1 is by far the most active protein arginine methyltransferase we have tested, using GST-GAR as the methyl-accepting substrate (13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar, 18.Smith J.J. Rucknagel K.P. Schierhorn A. Tang J. Nemeth A. Linder M. Herschman H.R. Wahle E. J. Biol. Chem. 1999; 274: 13229-13234Abstract Full Text Full Text PDF PubMed Scopus (136) Google Scholar). Because FDH was reported as the predominant PRMT in rat liver (21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar), we were interested in determining which protein arginine methyltransferase is the predominant PRMT in RAT1 cells and contributes most to the total arginine methyltransferase activity. We first generated a hypomethylated RAT1 cell lysate by treating cells with Adox, which inhibits the enzyme that metabolizesS-adenosylhomocysteine, a potent endogenous methyltransferase inhibitor (32.Najbauer J. Johnson B.A. Young A.L. Aswad D.W. J. Biol. Chem. 1993; 268: 10501-10509Abstract Full Text PDF PubMed Google Scholar). Western blotting with anti-PRMT1 and anti-FDH antibodies indicates that both FDH and PRMT1 are present in this hypomethylated RAT1 cell lysate (Fig.2, panel A). The endogenous methyltransferases present in the hypomethylated RAT1 cell lysate methylate many substrates after [3H]AdoMet is added (Fig.2, panel B, lane 1). The control antibody-protein A Sepharose complex (lane 5) as well as the protein A-Sepharose complex with anti-PRMT3 (lane 3) and anti-FDH (lane 4) all deplete some PRMT activity from the cell lysate. This is presumably due to nonspecific binding of protein A-Sepharose with arginine methyltransferases and/or the hypomethylated substrates. However, of the three antisera (anti-PRMT1, anti-PRMT3, and anti-FDH), only anti-PRMT1 (lane 2) can deplete a greater degree of PRMT activity than the control serum. Only the anti-PRMT1 immunoprecipitate (lane 6), but not the anti-PRMT3 (lane 7), anti-FDH (lane 8), or control (lane 9) immunoprecipitates, can methylate the arginine methyltransferase substrate GST-GAR. When the anti-PRMT1 immunoprecipitate is used to reconstitute the methyltransferase activity in the hypomethylated cell lysate, the protein arginine methylation pattern is restored (lane 10). In contrast, PRMT3 (lane 11), FDH (lane 12), and control (lane 13) immunoprecipitates cannot restore methylation of the PRMT substrates in hypomethylated, PRMT1-immunodepleted cell lysates. These results suggest that PRMT1 is the predominant endogenous protein arginine methyltransferase for substrates present in RAT1 cells. If FDH is the predominant protein arginine methyltransferase in liver, then livers from FDH(−/−) mice should contain much less protein arginine methyltransferase activity than livers from FDH(+/−) mice. The FDH(−/−) mouse was characterized previously (33.Champion K.M. Cook R.J. Tollaksen S.L. Giometti C.S. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 11338-11342Crossref PubMed Scopus (56) Google Scholar). To test whether FDH(−/−) mice have any significant deficiency in PRMT activity, we generated mouse liver lysates from FDH(+/−) and FDH(−/−) mice and compared the protein arginine methyltransferase activities in these mouse liver lysates (Fig.3, panel A, lanes 1 and10). When quantitated, the methyltransferase activities against GST-GAR in the liver lysates from FDH(+/−) or FDH(−/−) mice are not significantly different (panel B). The GST-GAR-methylating activity in the liver lysates can be immunoprecipitated by anti-PRMT1 antibody (lanes 7 and16). After immunoprecipitation with anti-PRMT1 antibody, little PRMT activity is left in the supernatant fraction (lanes 3 and 12). In contrast, anti-PRMT3 (lanes 4, 8, 13, and 17), anti-FDH (lanes 5, 9, 14, and18), or control antibodies (lanes 2, 6, 11, and15) do not precipitate GST-GAR-methylating activities from the FDH(+/−) or FDH(−/−) mouse liver lysates. To confirm the genotype of the mouse livers and the FDH antibody specificity, the FDH activities from mouse liver lysates and from the supernatant fraction of the immunoprecipitation reaction were assayed (panel C). Liver extracts from FDH(−/−) mice contain no detectable FDH activity (data not shown). The FDH activity analysis confirmed that only anti-FDH, but not the anti-PRMT1, anti-PRMT3, or control antibodies, immunoprecipitates FDH activity. The genotype of the mouse livers is further confirmed by Western blot using the anti-FDH antibody (panel D). The results presented in Fig. 3suggest that, although the FDH(−/−) mouse does not contain FDH enzyme in its liver, it has protein arginine methyltransferase activity levels similar to those found in the liver of the heterozygote FDH(+/−) mouse. The GST-GAR-methylating activity in mouse liver can only be immunoprecipitated by anti-PRMT1 antibodies. FDH activity can be immunodepleted by the anti-FDH antibody, but the immunoprecipitated FDH has no detectable GST-GAR-methylating activity. We also analyzed whether PRMT1 contributes the bulk of the hnRNP A1 and GST-GAR-methylating activity in other tissues from FDH(+/+), FDH(+/−), and FDH(−/−) mice (Fig. 4). In this experiment, cell lysates were prepared from brains, livers, and testes of FDH(+/+), FDH(+/−), and FDH(−/−) mice. Protein arginineN-methyltransferase activities were analyzed, using both GST-GAR and hnRNP A1 as methyl-accepting substrates (Fig. 4,panel A). Brain, testis, and liver lysates from FDH(+/+), FDH(+/−), and FDH(−/−) mice have similar methyltransferase activities against both GST-GAR and hnRNP A1. Although liver has substantially more FDH protein and activity than brain or testis, the latter two organs have substantially more PRMT activity (panel A). These methyltransferase activities can only be immunoprecipitated by the anti-PRMT1 antibody but not by anti-FDH or control antibodies. In contrast, FDH enzyme activity (Fig. 4,panel B) and antigen (Fig. 4, panel C) are detectable only in FDH(+/−) and FDH(+/+) mice. The results from these experiments indicate that it is PRMT1, not FDH, that correlates with GST-GAR- and hnRNP A1-methylating activities in tissues. We also compared the hepatic protein arginine methyltransferase activities from FDH(−/−), FDH(+/−), and FDH(+/+) mice, using the “protein methylase I” assay described by Kim et al.(21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar). The values, in pmol/min/mg protein ± S.D., are 4.50 ± 0.28, 3.19 ± 0.50, and 2.80 ± 0.28 for liver extracts of FDH(−/−), FDH(+/−), and FDH(+/+) animals, respectively. Although the reasons for increased protein arginine methyltransferase activity in FDH(−/−) mice are not clear and bear further investigation, these data clearly demonstrate that FDH does not provide a major contribution to hepatic protein arginine methyltransferase activities. Protein arginine N-methyltransferases transfer methyl groups from AdoMet to arginine residues in proteins (4.Gary J.D. Clarke S. Prog. Nucleic Acids Res. Mol. Biol. 1998; 61: 65-131Crossref PubMed Google Scholar, 34.Clarke S. Curr. Opin. Cell Biol. 1993; 5: 977-983Crossref PubMed Scopus (201) Google Scholar). We would expect that protein arginine methyltransferases should bind and perhaps cross-link to AdoMet. We used a UV irradiation cross-linking method to determine whether GST-PRMT1 and/or FDH, a NADP+-binding dehydrogenase (25.Min H. Shane B. Stokstad E.L. Biochim. Biophys. Acta. 1988; 967: 348-353Crossref PubMed Scopus (52) Google Scholar), can bind and cross-link to AdoMet. Affinity-purified FDH and recombinant GST-PRMT1 were incubated with [3H]AdoMet and UV cross-linked for 20 min. In Fig. 5, panel A, FDH and GST-PRMT1 proteins cross-linked with [3H]AdoMet are subjected to SDS-PAGE and fluorography. Proteins that have the ability to bind and cross-link to AdoMet are labeled with [3H]AdoMet and give bands in the fluorograph. Under the conditions used, only PRMT1 (panel A, lanes 1–4), but not FDH (panel A, lanes 5–8), can be cross-linked to AdoMet. Binding and cross-linking of PRMT1 to [3H]AdoMet can be competed away by unlabeled 10 μm AdoMet (lane 2), and more completely by 10 μm AdoHcy (lane 3), but not by 10 μm NADP+ (lane 4). The amount and purity of the FDH and GST-PRMT1 preparations used in these cross-linking experiments are analyzed by Coomassie Blue staining in panel B. Since type I protein arginine methylation was discovered over 30 years ago, there have been extensive efforts to characterize the methyltransferases responsible (4.Gary J.D. Clarke S. Prog. Nucleic Acids Res. Mol. Biol. 1998; 61: 65-131Crossref PubMed Google Scholar). Previous purification studies have shown that the predominant native activity is a large complex of 300–400 kDa (9.Ghosh S.K. Paik W.K. Kim S. J. Biol. Chem. 1988; 263: 19024-19033Abstract Full Text PDF PubMed Google Scholar, 35.Rawal N. Rajpurohit R. Paik W.K. Kim S. Biochem J. 1994; 300: 483-489Crossref PubMed Scopus (44) Google Scholar, 36.Liu Q. Dreyfuss G. Mol. Cell. Biol. 1995; 15: 2800-2808Crossref PubMed Scopus (272) Google Scholar). The polypeptide composition of this complex has not been well characterized. In one study, the purest fraction demonstrated at least 8 protein bands on SDS-PAGE analysis. Two of the most prominent bands were at 100 and 45 kDa (36.Liu Q. Dreyfuss G. Mol. Cell. Biol. 1995; 15: 2800-2808Crossref PubMed Scopus (272) Google Scholar). The 45-kDa protein matches the molecular weight of PRMT1. In a more recent study, the predominant protein arginine methyltransferase from rat liver was purified, and the major polypeptide in the preparation, with a molecular weight of 110 kDa, was identified as FDH, suggesting that this protein may be bifunctional (21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar). In this current study, however, we demonstrate that PRMT1 contributes most of the type I protein arginine methyltransferase activity, both in RAT1 cells and in murine tissues, for GST-GAR, hnRNP A1, or the hypomethylated RAT1 cell lysate (Figs. Figure 2, Figure 3, Figure 4). Anti-FDH antibody does not immunoprecipitate detectable protein arginineN-methyltransferase activity from cells and tissues, using as substrates GST-GAR, hnRNP A1, or the endogenous methyl-accepting substrates in hypomethylated cell lysates. It is, therefore, unlikely that FDH contributes a significant portion of the total protein arginine N-methyltransferase activity in cells or tissues. The native molecular mass of the cellular complex that contains PRMT1 is 300–400 kDa (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar). Because (i) immunodepletion experiments demonstrate that PRMT1 contributes most of the type I protein arginineN-methyltransferase activity in cells and tissues and (ii) the native molecular weight of the PRMT1 complex (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar) matches the molecular weight of the predominant type I protein arginineN-methyltransferase activity (9.Ghosh S.K. Paik W.K. Kim S. J. Biol. Chem. 1988; 263: 19024-19033Abstract Full Text PDF PubMed Google Scholar, 21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar, 35.Rawal N. Rajpurohit R. Paik W.K. Kim S. Biochem J. 1994; 300: 483-489Crossref PubMed Scopus (44) Google Scholar, 36.Liu Q. Dreyfuss G. Mol. Cell. Biol. 1995; 15: 2800-2808Crossref PubMed Scopus (272) Google Scholar), we conclude that PRMT1 is a catalytic subunit of the predominant protein arginineN-methyltransferase in tissues. This predominant protein arginine methyltransferase activity, which contains PRMT1 as its major catalytic subunit, has also been termed protein methylase I (9.Ghosh S.K. Paik W.K. Kim S. J. Biol. Chem. 1988; 263: 19024-19033Abstract Full Text PDF PubMed Google Scholar, 21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar,35.Rawal N. Rajpurohit R. Paik W.K. Kim S. Biochem J. 1994; 300: 483-489Crossref PubMed Scopus (44) Google Scholar). Several lines of evidence indicate that FDH is unlikely to be a protein arginineN-methyltransferase, even though FDH was identified as the predominant protein in a protein arginineN-methyltransferase purification procedure (21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar). The evidence includes the following observations: (i) PRMT activity can be further fractionated from FDH activity (Fig. 1), (ii) tissue extracts from FDH(−/−) mice contain normal levels of type I protein arginineN-methyltransferase activity against both GST-GAR and hnRNP A1 (Figs. 3 and 4) and normal levels of protein methylase I activity, and (iii) purified FDH does not cross-link to AdoMet, the methyl-donor substrate of protein arginine N-methyltransferase, in contrast to recombinant GST-PRMT1 protein (Fig. 5). If FDH does not contribute significantly to protein arginine methyltransferase activity in cell lysates, why was it purified as the predominant protein arginine N-methyltransferase from rat liver? One possible explanation is that PRMT1 and FDH physically interact to form a large multisubunit protein complex. However, our results indicate that if this is the case, such a complex is difficult to demonstrate. The GST-GAR- or hnRNP A1-methylating activities in cells can be completely separated away from FDH proteins by immunodepletion (Fig. 3). No physical interactions between FDH and any GST-GAR- or hnRNP A1-methylating protein arginineN-methyltransferase have been detected in coimmunoprecipitation experiments (Figs. Figure 1, Figure 2, Figure 3, Figure 4). Some common features shared between the predominant protein arginineN-methyltransferase and FDH proteins exist, such as their large native molecular mass of 300–400 kDa (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 22.Cook R.J. Lloyd R.S. Wagner C. J. Biol. Chem. 1991; 266: 4965-4973Abstract Full Text PDF PubMed Google Scholar, 24.Scrutton M.C. Beis I. Biochem. J. 1979; 177: 833-846Crossref PubMed Scopus (29) Google Scholar, 25.Min H. Shane B. Stokstad E.L. Biochim. Biophys. Acta. 1988; 967: 348-353Crossref PubMed Scopus (52) Google Scholar) and their tight association to 5-formyltetrahydrofolate-Sepharose gels (21.Kim S. Park G.H. Joo W.A. Paik W.K. Cook R.J. Williams K.R. J. Biol. Chem. 1998; 273: 27374-27382Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar). These common characteristics probably contributed to the misidentification of FDH as the predominant protein arginineN-methyltransferase in rat liver. However, until recombinant FDH can be purified from a source that does not have endogenous PRMT activity (such as Escherichia coli), we cannot rule out the remote possibility that FDH has some intrinsic PRMT activity. Multiple arginine methyltransferases exist in mammalian cells (4.Gary J.D. Clarke S. Prog. Nucleic Acids Res. Mol. Biol. 1998; 61: 65-131Crossref PubMed Google Scholar). In vitro methylation assays indicate that three known type I protein arginine methyltransferases (PRMT1, PRMT3, and CARM1) have overlapping, but distinct, substrate specificities (12.Lin W.J. Gary J.D. Yang M.C. Clarke S. Herschman H.R. J. Biol. Chem. 1996; 271: 15034-15044Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar, 14.Chen D. Ma H. Hong H. Koh S.S. Huang S.M. Schurter B.T. Aswad D.W. Stallcup M.R. Science. 1999; 284: 2174-2177Crossref PubMed Scopus (1020) Google Scholar). PRMT1 and PRMT3 both methylate GST-GAR, hnRNP A1, hnRNP A2, and poly(A)-binding protein II. However, the specific activity of PRMT1 for these substrates is much greater than that of PRMT3 (13.Tang J. Gary J.D. Clarke S. Herschman H.R. J. Biol. Chem. 1998; 273: 16935-16945Abstract Full Text Full Text PDF PubMed Scopus (277) Google Scholar, 18.Smith J.J. Rucknagel K.P. Schierhorn A. Tang J. Nemeth A. Linder M. Herschman H.R. Wahle E. J. Biol. Chem. 1999; 274: 13229-13234Abstract Full Text Full Text PDF PubMed Scopus (136) Google Scholar). The activities of PRMT1 and CARM1 overlap as well. When histone proteins were used as methyl-accepting substrates, PRMT1 prefers H4 and H2A, whereas CARM1 prefers H3, H2A, and H2B (14.Chen D. Ma H. Hong H. Koh S.S. Huang S.M. Schurter B.T. Aswad D.W. Stallcup M.R. Science. 1999; 284: 2174-2177Crossref PubMed Scopus (1020) Google Scholar). Yeasts contain only one type I protein arginine N-methyltransferase, whereas mammalian cells contain multiple type I PRMT enzymes with overlapping activities but distinct regulation mechanisms and subcellular localization. The physiological significance, if any, of the difference between these methyltransferases is not known. It is possible that mammalian cells require additional regulation beyond that required in yeast. This hypothesis is supported by the observation that PRMT1 and PRMT3 have different subcellular localizations and regulatory mechanisms, despite their overlapping substrate specificities. Identification of the substrates of each protein arginine N-methyltransferase will contribute to our understanding of the functions of asymmetric arginine methylation and the role of each type I protein arginineN-methyltransferase. We thank the members of the Herschman and Clarke labs for helpful discussions. We also thank Rick B. Dye (Vanderbilt University School of Medicine) for raising the FDH-deficient mice and purifying rat liver FDH.