Title: Reactivation of Phosphorylated Actin Depolymerizing Factor and Identification of the Regulatory Site
Abstract: Actin depolymerizing factor (ADF) occurs naturally in two forms, one of which contains a phosphorylated Ser and does not bind G-actin or depolymerize F-actin. Removal of this phosphate in vitro by alkaline phosphatase restores full F-actin depolymerizing activity. To identify the phosphorylation site, [32P]pADF was purified and digested with endoproteinase Lys-C. The digest contained only one 32P-labeled peptide. Further digestion with endoproteinase Asp-N and mass spectrometric analysis showed that this peptide came from the N terminus of ADF. Alkaline phosphatase treatment of one Asp-N peptide (mass 753) converted it to a peptide of mass 673, demonstrating that this peptide contains the phosphate group. Tandem mass spectrometric sequence analysis of this peptide identified the phosphorylated Ser as the encoded Ser3 (Ser2 in the processed protein). HeLa cells, transfected with either chick wild-type ADF cDNA or a cDNA mutated to code for Ala in place of Ser24 or Thr25, express and phosphorylate the exogenous ADF. Cells also expressed high levels of mutant ADF when Ser3 was deleted or converted to either Ala or Glu. However, none of these mutants was phosphorylated, confirming that Ser3 in the encoded ADF is the single in vivo regulatory site. Actin depolymerizing factor (ADF) occurs naturally in two forms, one of which contains a phosphorylated Ser and does not bind G-actin or depolymerize F-actin. Removal of this phosphate in vitro by alkaline phosphatase restores full F-actin depolymerizing activity. To identify the phosphorylation site, [32P]pADF was purified and digested with endoproteinase Lys-C. The digest contained only one 32P-labeled peptide. Further digestion with endoproteinase Asp-N and mass spectrometric analysis showed that this peptide came from the N terminus of ADF. Alkaline phosphatase treatment of one Asp-N peptide (mass 753) converted it to a peptide of mass 673, demonstrating that this peptide contains the phosphate group. Tandem mass spectrometric sequence analysis of this peptide identified the phosphorylated Ser as the encoded Ser3 (Ser2 in the processed protein). HeLa cells, transfected with either chick wild-type ADF cDNA or a cDNA mutated to code for Ala in place of Ser24 or Thr25, express and phosphorylate the exogenous ADF. Cells also expressed high levels of mutant ADF when Ser3 was deleted or converted to either Ala or Glu. However, none of these mutants was phosphorylated, confirming that Ser3 in the encoded ADF is the single in vivo regulatory site. Actin depolymerizing factor (ADF),1 1The abbreviations used are: ADFactin depolymerizing factorE10embryonic day 10DTTdithiothreitolPAGEpolyacrylamide gel electrophoresisHPLChigh pressure liquid chromatographyHAPhydroxylapatiteGTPγSguanosine 5'-3-O-(thio)triphosphate. is an 18.5-kDa protein with a pH-dependent F-actin binding/depolymerizing activity (1Hayden S.M. Miller P.S. Brauweiler A. Bamburg J.R. Biochemistry. 1993; 32: 9994-10004Crossref PubMed Scopus (203) Google Scholar, 2Hawkins M. Pope B. Maciver S.K. Weeds A.G. Biochemistry. 1993; 32: 9985-9993Crossref PubMed Scopus (240) Google Scholar) very similar to that of its structural homolog, cofilin(3Nishida E. Maekawa S. Sakai H. Biochemistry. 1984; 23: 5307-5313Crossref PubMed Scopus (253) Google Scholar, 4Yonezawa N. Nishida E. Sakai H. J. Biol. Chem. 1985; 260: 14410-14412Abstract Full Text PDF PubMed Google Scholar). Proteins in the ADF/cofilin family appear to be ubiquitous in eukaryotes (for review, see (5Sun H.-Q. Kwiatkowska K. Yin H.L. Curr. Opin. Cell Biol. 1995; 7: 102-110Crossref PubMed Scopus (171) Google Scholar)). Mutations that inactivate the ADF/cofilin gene in Saccharomyces cerevisiae(6Moon A.L. Janmey P. Louie K.A. Drubin D. J. Cell Biol. 1993; 120: 421-435Crossref PubMed Scopus (201) Google Scholar), Drosophila melanogaster(7Edwards K.A. Montague R.A. Shepard S. Edgar B.A. Erikson R.L. Kiehart D.P. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4589-4593Crossref PubMed Scopus (59) Google Scholar), and Caenorhabditis elegans(8McKim K.S. Matheson C. Marra M.A. Wakarchuk M.F. Baillie D.L. Mol. & Gen. Genet. 1994; 242: 346-357Crossref PubMed Scopus (100) Google Scholar) are lethal. actin depolymerizing factor embryonic day 10 dithiothreitol polyacrylamide gel electrophoresis high pressure liquid chromatography hydroxylapatite guanosine 5'-3-O-(thio)triphosphate. Phosphorylated forms of both ADF and cofilin have been identified in vertebrate cells(9Ohta Y. Nishida E. Sakai H. Miyamoto E. J. Biol. Chem. 1989; 264: 16143-16148Abstract Full Text PDF PubMed Google Scholar, 10Bamburg J.R. Minamide L.S. Morgan T.E. Hayden S.M. Giuliano K.A. Koffer A. Methods Enzymol. 1991; 196: 125-140Crossref PubMed Scopus (21) Google Scholar, 11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). We isolated phosphorylated ADF and showed it to be inactive in depolymerizing F-actin and in inhibiting the assembly of G-actin(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). Phosphoamino acid analysis identified phosphoserine as the phosphorylated amino acid(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). Here, we demonstrate that pADF can be completely reactivated following phosphatase treatment. We also determine the location of the phosphoserine and show by site-directed mutagenesis and expression in vertebrate cells that this is the site for regulation by phosphorylation in vivo. pADF was isolated free from ADF and cofilin from embryonic day 10 (E10) or E11 chick brain as described previously(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). It was further purified for digestion and sequence analysis as described under "Results." Chick brain ADF was purified according to the method of Giuliano et al.(12Giuliano K.A. Khatib F.A. Hayden S.M. Daoud E.W.R. Adams M.E. Amorese D.A. Bernstein B.W. Bamburg J.R. Biochemistry. 1988; 27: 8931-8938Crossref PubMed Scopus (43) Google Scholar). Skeletal muscle actin was purified from rabbit muscle acetone powder(13Pardee J.D. Spudich J.A. Methods Cell Biol. 1982; 24: 271-289Crossref PubMed Scopus (340) Google Scholar). The in vitro reactivation of pADF by treatment with alkaline phosphatase (Sigma) is described under "Results." The amount of G-actin depolymerized from F-actin in the presence of ADF was quantified by using the DNase I inhibition assay(10Bamburg J.R. Minamide L.S. Morgan T.E. Hayden S.M. Giuliano K.A. Koffer A. Methods Enzymol. 1991; 196: 125-140Crossref PubMed Scopus (21) Google Scholar, 14Harris H.E. Bamburg J.R. Bernstein B.W. Weeds A.G. Anal. Biochem. 1982; 119: 102-114Crossref PubMed Scopus (24) Google Scholar). Cells in 10-cm culture dishes were scraped in 300 εl of ice-cold extraction buffer (10 mM Tris, pH 7.6, 1% SDS, 15 mM NaF, 10 mM dithiothreitol (DTT), 2 mM EGTA, 0.3 mM sodium orthovanadate, 10 εl/ml protease inhibition mixture)(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar), sonicated, and immersed 5 min in a boiling water bath. After cooling the samples on ice, proteins were precipitated with chloroform/methanol(15Wessel D. Flügge U.I. Anal. Biochem. 1984; 138: 141-143Crossref PubMed Scopus (3185) Google Scholar). The pellets were dissolved in sample preparation buffer (0.125 M Tris, pH 6.8, 1% SDS, 5% glycerol, 10% 2-mercaptoethanol, 0.01% bromphenol blue) for SDS-polyacrylamide gel electrophoresis (PAGE) or in lysis buffer (9.5 M urea, 2% Nonidet P-40, 2% ampholytes, pH 3-10, 10% 2-mercaptoethanol) for two-dimensional gels (16O'Farrell P.H. J. Biol. Chem. 1975; 250: 4007-4021Abstract Full Text PDF PubMed Google Scholar). Protein concentrations were determined by the solid phase dye-binding method(17Minamide L.S. Bamburg J.R. Anal. Biochem. 1990; 190: 66-70Crossref PubMed Scopus (237) Google Scholar). SDS-PAGE was performed by the method of Laemmli (18Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207478) Google Scholar) on 15% acrylamide (2.7% cross-linker) isocratic mini-slab gels. Two-dimensional, nonequilibrium pH gradient electrophoresis (19O'Farrell P.Z. Goodman H.M. O'Farrell P.H. Cell. 1977; 12: 1133-1141Abstract Full Text PDF PubMed Scopus (2583) Google Scholar) on mini-gels, immunoblotting onto polyvinylidene difluoride membrane (Immobilon P; Millipore Corp., Bedford, MA) for 1 h at 0.3 A, and immunostaining for ADF were performed as described previously(10Bamburg J.R. Minamide L.S. Morgan T.E. Hayden S.M. Giuliano K.A. Koffer A. Methods Enzymol. 1991; 196: 125-140Crossref PubMed Scopus (21) Google Scholar). Alkaline phosphatase-conjugated secondary antibody (Amersham Corp.) was used, and blots were developed after a quick rinse in high pH buffer (100 mM Tris, pH 9.5, 100 mM NaCl, 50 mM MgCl2), first with Lumiphos substrate (Boehringer Mannheim, Indianapolis, IN) and then in 0.165 mg/ml of 5-bromo-4-chloro-3-indolylphosphate p-toluidine salt and 0.33 mg/ml of nitro blue tetrazolium chloride (Life Technologies, Inc.) diluted in the pH 9.5 buffer. Hyperfilm ECL (Amersham Corp.) exposures of the Lumiphos images and the stained blots were scanned using a Microscan 2000 image analysis system (Technology Resources, Inc., Knoxville, TN). Internal standards of chick brain ADF were included on all one-dimensional immunoblots. Chick skin fibroblasts, prepared by trypsin dissociation of skin from the dorsal side of 8-day embryos, were grown in 10-cm culture dishes in Dulbecco's modified Eagle's medium (Sigma) containing 10% fetal bovine serum. When 70% confluent, cells were transferred to a low phosphate medium (Dulbecco's modified Eagle's medium without phosphate) (Sigma) containing 10% fetal bovine serum that had been exhaustively dialyzed against 0.15 M NaCl. [32P]Orthophosphate (5 Ci/mmol; 100 εCi/ml culture medium) was added for 12 h. The cells were washed with 3 × 10 ml of 4°C phosphate-buffered saline, and then harvested in extraction buffer as described above. The cell lysate was added to a homogenate prepared from 6 g of 10-11-day chick embryo brain in 10 ml of the same buffer, and the mixture was further homogenized with five to six more strokes of a Teflon-glass homogenizer. The 32P-labeled pool of pADF was then purified as described under "Results." For endopeptidase digestion, approximately 60 εg of precipitated ADF or pADF were solubilized in 20 εl of 8 M urea, 20 mM Tris-Cl buffer, pH 9.0. The volume was adjusted to 100 εl with 20 mM Tris-Cl, pH 9.0. Endopeptidase Lys-C (Wako Bioproducts, Richmond, VA) was added at an enzyme/substrate molar ratio of 1:50, and the mixture was incubated at 30°C for 8 h. Aliquots were removed periodically to evaluate the progress of digestion by SDS-PAGE. The digested samples were either immediately fractionated or frozen in liquid nitrogen and stored at −80°C for later use. For endopeptidase Asp-N digestion of the N-terminal fragment from Lys-C digestion, the appropriate fractions were lyophilized and resolubilized in 10 εl of 8 M urea, 50 mM sodium phosphate, pH 8.0. The volume was adjusted to 60 εl with 50 mM sodium phosphate, pH 8.0. Endopeptidase Asp-N (Boehringer Mannheim) was added at an enzyme/substrate weight ratio of 1:20. The mixture was incubated for 8 h at 36°C and then lyophilized. Peptides from Lys-C digests of ADF or [32P]pADF were fractionated by HPLC on a column of Microsorb-MV C18, 5 εm, 300-Å pore (5.6 mm × 25 cm) (Rainin Inst., Woburn, MA) eluted with a linear step gradient (2.0-37.5% B, 60 min; 37.5-75% B, 30 min; 75-90% B, 15 min; solvent A, 0.1% trifluoroacetic acid; solvent B 80% acetonitrile, 0.1% TFA) (Millipore, Waters Chromatography, Milford, MA). The flow rate was 0.5 ml/min, and the column effluent was monitored at 220 nm. Fractions (0.25 ml) were collected every 30 s, and Cerenkov radiation was measured. HPLC electrospray mass spectrometry and tandem mass spectrometry were performed at the Beckman Institute, City of Hope, Duarte, CA on a TSQ-700 triple quadrupole mass spectrometer (Finnigan-MAT, San Jose, California), equipped with an electrospray ion source operating at atmospheric pressure. Mass spectra were recorded in the positive ion mode. The solvent delivery system, fused silica capillary columns, and UV detection system have been described in detail(20Swiderik K.M. Lee T.D. Shively J.E. Methods Enzymol. 1995; (in press)Google Scholar). A 656-base pair EcoRI/AvrII fragment containing the full ADF cDNA was excised from pUC 19 (21Adams M.E. Minamide L.S. Duester G. Bamburg J.R. Biochemistry. 1990; 29: 7414-7420Crossref PubMed Scopus (50) Google Scholar) and cloned into the EcoRI-XbaI sites of M13mp18. Oligonucleotide-directed point mutagenesis (22Kunkel T.A. Roberts J.D. Zakour R.A. Methods Enzymol. 1987; 154: 367-382Crossref PubMed Scopus (4560) Google Scholar) was used to mutate the Ser24 and Thr25 codons to Ala using the mutagenic primers 5'-CCTCAGGCGCTGCGCATTTCCG-3', 5'-CCTCAGGCGTTGCGCCATTT-3', 5'-CCTCAGGCGCAGAGCATTTCCG-3' giving the ADF(A24,25), ADF(A24), and ADF(A25) mutant cDNAs, respectively. NcoI/HindIII fragments encoding ADF(A24,25), ADF(A24), or ADF(A25) were subcloned into the modified pBR 322 bacterial expression vector (pET) as described previously(21Adams M.E. Minamide L.S. Duester G. Bamburg J.R. Biochemistry. 1990; 29: 7414-7420Crossref PubMed Scopus (50) Google Scholar, 23Studier F.W. Rosenberg A.H. Dunn J.J. Dubendorff J.W. Methods Enzymol. 1990; 185: 60-89Crossref PubMed Scopus (6005) Google Scholar). BamHI/AvrII fragments (560 base pairs) from pET containing the full ADF or ADF(A24,25) cDNAs were cloned into the multi-cloning site of the eukaryotic expression vector pcDNA3 (Invitrogen, San Diego, CA), giving the vectors pcDNA3ADF and pcDNA3ADF(A24,25). Site-directed mutagenesis was performed on pcDNA3ADF using the Transformer Mutagenesis Kit (Clontech, Palo Alto, CA)(24Deng W.P. Nickoloff J.A. Anal. Biochem. 1992; 200: 81-88Crossref PubMed Scopus (1079) Google Scholar). Three mutagenic primers were used, encoding mutations in the Ser3 codon of the ADF cDNA (Ser → Ala, 5'-CTTGTACTCCAGCTGCCATGGATC-3'; Ser → Glu, 5'-CGGCAACTTGTACTCCTTCTGCCATGGATCCGAGC-3'; Ser deletion, 5'-CGGCAACTTGTACTCCTGCCATGGATCCGAGC-3'). A selection primer, eliminating a unique pcDNA3 HindIII restriction site, was also synthesized (5'-GGTACCAAGCAAGGGTCTCCC-3'). All mutant cDNAs were confirmed by DNA sequencing (Sequenase Version 2.0 DNA Sequencing Kit, U. S. Biochemical Corp., Cleveland, OH). All oligonucleotides were synthesized by Macromolecular Resources, Colorado State University, Fort Collins, CO. Bacterial expression of the ADF(A24,25) and ADF(E3) mutants was performed as described previously for the expression of ADF (21Adams M.E. Minamide L.S. Duester G. Bamburg J.R. Biochemistry. 1990; 29: 7414-7420Crossref PubMed Scopus (50) Google Scholar). The postinduction bacterial pellets were resuspended in 50 mM Tris, pH 8.0, 2.0 mM EGTA, 2.4 mM phenylmethylsulfonyl fluoride, 20 εl/ml protease inhibition mixture(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar), and then lysed in 2.0 mg/ml lysozyme on ice for 15 min, followed by addition of 15 mM MgCl2, 0.2 mM MnCl2, and 8 εg/ml DNase I. After incubation at room temperature for 10 min, 20 mM DTT and 10 εl/ml protease inhibition mixture were added. After centrifugation at 60,000 × gav for 30 min at 4°C, the supernatant was passed through a column of DEAE-cellulose (DE-52; Whatman), and the flow-through was applied to a column of Green A-agarose (Amicon Corp., Danvers, MA). The ADF was eluted with 175 mM NaCl in 10 mM Tris, pH 7.6, 5 mM DTT, concentrated on a Centricon-10 (Amicon Corp.), frozen in liquid nitrogen, and stored at −80°C. HeLa cells were cultured in minimal essential medium (Sigma) containing 10% fetal bovine serum, 50 units/ml penicillin, and 50 εg/ml streptomycin. Cells at 30-50% confluency on 10-cm dishes were transfected by the calcium phosphate method (25Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989: 16.33-16.36Google Scholar) with 10 εg of pcDNA3 vector containing ADF cDNA or one of the site-directed mutants of ADF cDNA. The precipitate was removed after 12 h by washing 3 times with phosphate-buffered saline, and extracts of cells were prepared for SDS-PAGE 48 h after transfection as described above. We previously demonstrated that the phosphorylated form of ADF, partially purified from chick embryo brain, had no actin depolymerizing activity(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). Partially purified preparations of pADF contain proteases that degrade ADF when incubated at 37°C(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). While the SDS-denatured pADF could be dephosphorylated by alkaline phosphatase, we did not demonstrate reactivation. We found that brain ADF, denatured by boiling in 1% SDS, could be completely reactivated after removal of SDS and renaturation. We used these conditions to renature ADF produced from denatured pADF following alkaline phosphatase treatment. Complete ADF activity was regained after dephosphorylation, whereas no ADF activity was found in the pADF control (Fig. 1). Electrophoresis of both samples confirmed that dephosphorylation occurred in the alkaline phosphatase-treated sample, as previously reported(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). The combined homogenate from the 32P-labeled chick skin fibroblasts and chick embryo brain was centrifuged at 100,000 × g. pADF was partially purified from the supernatant by chromatography on DEAE-cellulose, Green A-agarose, and hydroxylapatite (HAP). SDS-PAGE of HAP fractions (Fig. 2A) and the corresponding autoradiograph (Fig. 2B) demonstrate that [32P]pADF eluted in the position expected for pADF, as previously identified by immunoblotting(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). The pADF-containing fractions were boiled in 0.1% SDS, combined, and concentrated to less than 100 εl in a Centricon-10 microconcentrator. The volume was adjusted to 1 ml with 0.1 M Tris-Cl, pH 7.6, and the protein was reduced, and alkylated with 4-vinylpyridine(26Ozols J. Methods Enzymol. 1990; 182: 587-601Crossref PubMed Scopus (185) Google Scholar). Proteins were precipitated (15Wessel D. Flügge U.I. Anal. Biochem. 1984; 138: 141-143Crossref PubMed Scopus (3185) Google Scholar) and redissolved in sample preparation buffer for separation on a preparative SDS-PAGE mini-slab gel (0.5 mm). After electrophoresis, proteins were visualized by KCl precipitation(27Nelles L.P. Bamburg J.R. Anal. Biochem. 1976; 73: 522-531Crossref PubMed Scopus (82) Google Scholar), and the pADF band was excised from the gel. pADF was electroeluted into a Centricon-10 chamber in 0.5 × SDS-PAGE running buffer (18Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207478) Google Scholar) and concentrated to about 0.1 ml by centrifugation. An aliquot of this material was rerun on SDS-PAGE to check for homogeneity (Fig. 2C). Electroeluted ADF and pADF were precipitated(15Wessel D. Flügge U.I. Anal. Biochem. 1984; 138: 141-143Crossref PubMed Scopus (3185) Google Scholar), dissolved in buffer, and digested with endopeptidase Lys-C. The resulting peptides were separated by HPLC and Cerenkov radioactivity in each fraction determined (Fig. 3). A single phosphopeptide, unique to the pADF digest, was identified (Fig. 3, arrow). This single peptide contained about 60% of the total radioactivity loaded onto the column, the remainder being accounted for in the unbound fraction and a few small peaks visible in Fig. 3. No sequence could be obtained by direct sequence analysis of the single phosphopeptide, whereas neighboring peptides from the same digest were readily sequenced. This suggested that the labeled peptide contained a blocked N terminus and might arise from the N terminus of ADF, a peptide containing only a single Ser residue. The N-terminal Lys-C peptide from pADF contains only a single Ser and has the sequence Ac-ABGVQVADEVJRIFYDMK, where B is phosphoserine, and J is pyridylethylcysteine. This sequence suggests that the N-terminal Met is removed and that the penultimate Ala is acetylated. This sequence was confirmed by mass spectrometric analysis as follows. The isolated Lys-C phosphopeptide (containing some of the contaminating trailing peak) was digested with Asp-N and rerun on HPLC. The undigested contaminating peptide and two major peptide fragments of the phosphopeptide were obtained (Fig. 4). The parent ion masses of the two major Asp-N peptides are 753.6 and 1150.3, as determined by electrospray mass spectrometric analysis. These correspond to the masses expected for peptides derived from the sequence shown above: A, Ac-ABGVQVA (mass 753); B, DEVJRIFY (mass 1150). In addition, the C-terminal peptide (C) DMK was identified by its mass of 393. Analysis of the N-terminal peptide by tandem mass spectrometry produced the spectrum in Fig. 5A. The parent ion (P) at m/z 753 matches the value predicted for that of the N-terminal peptide. The predicted fragment ions, type b and y are shown in Fig. 5C, and those ions observed are labeled in Fig. 5A. Fragment ions b3 through b6 yield the C-terminal sequence -VQVA, confirming the origin of the peptide. The b3 ion m/z corresponds to the N-terminal fragment Ac-ABG-. Also apparent in the spectrum are ion species representing the loss of phosphate as phosphoric acid (98 mass units) due to collision. The N-terminal peptide (A) in Fig. 4 was isolated by HPLC and treated with alkaline phosphatase. Analysis of the resulting peptide by tandem mass spectrometry yielded the spectrum in Fig. 5B. The parent ion at m/z 673 is that of the dephosphorylated peptide (−80 mass units). Fragment ions b3 through b6 yield an identical C-terminal sequence but are decreased by 80 mass units due to the loss of phosphate. Additionally, there are no peaks representing the loss of phosphoric acid due to collision. The b3 fragment (AcASG-) contains only one phosphorylatable residue, Ser, and therefore, the phosphorylation site on pADF is Ser2 (Ser3 in the encoded protein).Figure 5:Identification of the phosphorylation site by tandem mass spectrometry. A, N-terminal Asp-N peptide from pADF. B, same peptide after alkaline phosphatase treatment. C, peptide fragments expected from the N-terminal end (b) and C-terminal end (y) of the peptide. Those that are readily identified in the mass spectra are labeled in A and B. p, parent ion. Amino acid B is phosphoserine (serine in the dephosphorylated peptide).View Large Image Figure ViewerDownload Hi-res image Download (PPT) Extracts from HeLa cells, transiently transfected with pcDNA3 carrying wild-type ADF, ADF(A24,25), ADF(A3), ADF(E3), or ADF(Δ3) expressed the transfected forms of ADF as shown on two-dimensional immunoblots (Fig. 6). Endogenous HeLa cell ADF only weakly cross-reacts with the chick ADF antiserum (Fig. 6, 1a), eliminating the need to epitope tag the exogenously expressed protein. Two-dimensional immunoblots of extracts from cells expressing wild-type ADF (Fig. 6, 2a) or the ADF(A24,25) mutant (Fig. 6, 1b) typically show two species, the more acidic one arising from the in vivo phosphorylation of ADF. Extracts of cells transfected with ADF(A3) (Fig. 6, 2b), ADF(E3) (Fig. 6, 3a), or ADF(Δ3) (Fig. 6, 3b) do not contain the phosphorylated forms of these proteins, confirming that Ser3 of the encoded protein is the regulatory site for phosphorylation in vivo. Bacterially expressed forms of wild-type ADF, ADF(A24), ADF(A25), and ADF(E3) were purified from supernatants of bacterial lysates. Proteins were homogeneous on SDS-PAGE (not shown). Wild-type ADF, ADF(A24) and ADF(A25) were identical to brain ADF in depolymerizing F-actin. The ADF(E3) mutant, a structural homolog of the phosphorylated ADF, had F-actin depolymerizing activity that was about 10% of the wild-type (Fig. 7). Regulation of proteins by phosphorylation is a well established mechanism for controlling biological activity (for review, see (28Walsh D.A. Newsholme P. Cawley K.C. van Patten S.M. Angelos K.L. Physiol. Rev. 1991; 71: 285-304Crossref PubMed Scopus (17) Google Scholar)), nuclear translocation (for review, see (29Silver P.A. Cell. 1991; 64: 489-497Abstract Full Text PDF PubMed Scopus (394) Google Scholar)), and protein degradation(30Nishizawa M. Okazaki K. Furuno N. Watanabe N. Sagata N. EMBO J. 1992; 11: 2433-2446Crossref PubMed Scopus (110) Google Scholar). Since ADF and cofilin contain identical nuclear translocation signal sequences(31Abe H. Endo T. Yamamoto K. Obinata T. Biochemistry. 1990; 29: 7420-7425Crossref PubMed Scopus (69) Google Scholar, 32Abe H. Nagaoka R. Obinata T. Exp. Cell Res. 1993; 206: 1-10Crossref PubMed Scopus (72) Google Scholar), and both can be transported to the nucleus of cells under stress(33Nishida E. Iida K. Yonezawa N. Koyasu S. Yahara I. Sakai H. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 5262-5266Crossref PubMed Scopus (211) Google Scholar, 34Ono S. Abe H. Nagaoka R. Obinata T. J. Muscle Res. Cell Motil. 1993; 14: 195-204Crossref PubMed Scopus (50) Google Scholar), their phosphorylation in cells may be involved in regulating one or more of the above events. In order to demonstrate that the reversible phosphorylation of ADF controls its cellular activity, we needed to show that it can be reactivated by removal of the phosphate. Partially purified preparations of pADF contain proteases, which degrade ADF when incubated at 37°C(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). However, by boiling the proteins with SDS, these proteases are permanently inactivated. ADF can be reactivated after denaturation and dephosphorylation. Regulation by phosphorylation/dephosphorylation may control the activity of ADF in cells. ADF is normally down-regulated in developing chick skeletal muscle in vivo but not in primary myocyte cultures of equivalent age in vitro. Interestingly, myotubes developing in vitro compensate for their inability to down-regulate ADF expression by converting the abundant ADF to the inactive form, pADF, which accumulates during the process of in vitro myogenesis(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). ADF colocalizes with actin in the membrane ruffles of motile cells and, in developing neuronal growth cones, areas containing highly dynamic actin filaments(35Bamburg J.R. Bray D. J. Cell Biol. 1987; 105: 2817-2825Crossref PubMed Scopus (158) Google Scholar). Rapid cycling of actin assembly/disassembly could be regulated by ADF phosphorylation. Recent studies show phosphorylation of both ADF and cofilin are altered by various signal transduction pathways that are activated by growth factors and pharmacological agents. Treatment of cultured rat primary astrocytes with dibutyryl-cAMP or the myosin light chain kinase inhibitor, ML-9, causes dramatic shape changes and reorganization of the actin filament network and increased dephosphorylation of both the regulatory light chain of myosin and ADF(36Baorto D.M. Mellado W. Shelanski M.L. J. Cell Biol. 1992; 117: 357-367Crossref PubMed Scopus (138) Google Scholar). Additionally, treatment of primary cultures of dog thyroid cells with thyrotropin causes distinct morphological changes and a redistribution of the actin stress fiber network, as well as rapid dephosphorylation of both ADF and cofilin(37Saito T. Lamy F. Roger P.P. Lecocq R. Dumont J. Exp. Cell Res. 1994; 212: 49-61Crossref PubMed Scopus (54) Google Scholar). Since the thyrotropin effects could be mimicked by forskolin, it seems likely that the dephosphorylation of ADF is regulated in both astrocytes and thyroid cells by a cAMP-dependent process. Additionally, cofilin dephosphorylation in human platelets could be stimulated by GTPγS or by free calcium but not by activation of protein kinase C(38Davidson M.M.L. Haslam R.J. Biochem. J. 1994; 301: 41-47Crossref PubMed Scopus (74) Google Scholar). Again, the GTPγS effects could very well be mediated through increases in cAMP via the protein Gs. Together, this evidence suggests that ADF is a major regulator of actin filament dynamics in vivo and that the regulation of ADF by phosphorylation may provide a rapid, localized, and reversible control mechanism for regulating actin filament dynamics in areas of high filament turnover. Dephosphorylation of ADF (37Saito T. Lamy F. Roger P.P. Lecocq R. Dumont J. Exp. Cell Res. 1994; 212: 49-61Crossref PubMed Scopus (54) Google Scholar) and cofilin (9Ohta Y. Nishida E. Sakai H. Miyamoto E. J. Biol. Chem. 1989; 264: 16143-16148Abstract Full Text PDF PubMed Google Scholar, 37Saito T. Lamy F. Roger P.P. Lecocq R. Dumont J. Exp. Cell Res. 1994; 212: 49-61Crossref PubMed Scopus (54) Google Scholar) accompanies nuclear transport in many cells placed under stress. In addition, the costimulatory signals for human T-cell activation induce the dephosphorylation of cofilin and its translocation to the nucleus of human T-cells(39Samstag Y. Eckerskorn C. Wesselborg S. Henning S. Wallich R. Meuer S.C. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 4494-4498Crossref PubMed Scopus (88) Google Scholar). Cofilin dephosphorylation correlates with the induction of interleukin-6 responsiveness and interleukin-2 secretion. Interestingly, cofilin dephosphorylation and translocation to the nucleus occur spontaneously in autonomously proliferating T-lymphoma cells. However, dephosphorylation of ADF and cofilin, which occurs in thyroid cells in response to thyrotropin(37Saito T. Lamy F. Roger P.P. Lecocq R. Dumont J. Exp. Cell Res. 1994; 212: 49-61Crossref PubMed Scopus (54) Google Scholar), is not accompanied by nuclear translocation, suggesting that dephosphorylation is not sufficient for nuclear uptake. Ser24 in both ADF and cofilin exists within a conserved CaM kinase II consensus sequence. This seemed a likely site for phosphorylation as it is situated next to a nuclear translocation sequence, conserved between ADF and cofilin, both of which have been shown to translocate to the nucleus. It was suggested that the dephosphorylation of cofilin at this site may be involved in its nuclear translocation(9Ohta Y. Nishida E. Sakai H. Miyamoto E. J. Biol. Chem. 1989; 264: 16143-16148Abstract Full Text PDF PubMed Google Scholar). However, ADF was not phosphorylated in vitro to any significant extent with several common kinases including CaM kinase II, protein kinase C, protein kinase A, and myosin light chain kinase(11Morgan T.E. Lockerbie R.O. Minamide L.S. Browning M.D. Bamburg J.R. J. Cell Biol. 1993; 122: 623-633Crossref PubMed Scopus (143) Google Scholar). The studies reported here in which the ADF(A24,25) mutant is expressed and phosphorylated in HeLa cells demonstrates convincingly that Ser24 is not the regulatory site. We have identified the single regulatory phosphorylation site within the chick brain ADF primary sequence as Ser3 of the encoded protein. This is a conserved residue within the ADF/cofilin family, present in mammals, birds, amphibians, and insects (Fig. 8). Echinoderms (40Takagi T. Konishi K. Mabuchi I. J. Biol. Chem. 1988; 263: 3097-3102Abstract Full Text PDF PubMed Google Scholar) and yeast (6Moon A.L. Janmey P. Louie K.A. Drubin D. J. Cell Biol. 1993; 120: 421-435Crossref PubMed Scopus (201) Google Scholar) have an extra amino acid between this Ser and the initiating Met, while plants (41Kim S.-R. Kim Y. An G. Plant Mol. Biol. 1993; 21: 39-45Crossref PubMed Scopus (70) Google Scholar) have four extra amino acids, two of which are Ser. In the amoeba protein, actophorin(42Quirk S. Maciver S.K. Ampe C. Doberstein S.K. Kaiser D.A. VanDamme J. Vandekerckhove J.S. Pollard T.D. Biochemistry. 1993; 32: 8525-8533Crossref PubMed Scopus (54) Google Scholar), this Ser is adjacent to the initiating Met. In mammals, birds, and amphibians, the only classes so far in which phosphorylation of ADF/cofilin proteins have been demonstrated, a consensus sequence for the phosphorylation site appears to be N-ASGVXVXD. Such a phosphorylation consensus sequence has not been reported(43Pearson R.B. Kemp B.E. Methods Enzymol. 1991; 200: 62-81Crossref PubMed Scopus (873) Google Scholar, 44Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). Although autophosphorylation of the N-terminal penultimate Ser residue in c-mos has been characterized, this Ser is surrounded by two Pro residues, which are likely to be involved in determining the phosphorylation specificity of this meiotic cell cycle regulatory protein(30Nishizawa M. Okazaki K. Furuno N. Watanabe N. Sagata N. EMBO J. 1992; 11: 2433-2446Crossref PubMed Scopus (110) Google Scholar). The small GTP-binding protein Rac1 is a key regulator of the rapid actin reorganization accompanying membrane ruffling during the early phases of growth factor-induced signal transduction (for review, see (45Hall A. Annu. Rev. Cell Biol. 1994; 10: 31-54Crossref PubMed Scopus (768) Google Scholar)). ADF is co-localized with actin in regions where Rac1 induces dramatic alterations in actin dynamics. Overexpression of a constitutively active mutant of Rac1 in Drosophila neurons inhibits nerve growth and increases the rhodamine-phalloidin staining of growth cones(46Luo L. Liao Y.J. Jan L.Y. Jan Y.N. Genes & Dev. 1994; 8: 1787-1802Crossref PubMed Scopus (819) Google Scholar). This finding suggests that growth inhibition may arise from a decrease in F-actin turnover. Conversely, overexpression of a dominant negative form of Rac1 also led to inhibition of nerve growth, but in this case growth cones did not stain with rhodamine phalloidin, suggesting that they contained little or no F-actin(46Luo L. Liao Y.J. Jan L.Y. Jan Y.N. Genes & Dev. 1994; 8: 1787-1802Crossref PubMed Scopus (819) Google Scholar). One downstream effector of Rac1 is the serine/threonine kinase, p65PAK(47Manser E. Leung T. Salihuddin H. Zhao Z. Lim L. Nature. 1994; 367: 40-46Crossref PubMed Scopus (1305) Google Scholar), which we believe could be the kinase that is directly or indirectly responsible for the phosphorylation of ADF. Interestingly, Drosophila ADF has the phosphorylation consensus sequence found in ADF/cofilin from mammals, birds, and amphibians (Fig. 8). These results are at least consistent with a model in which ADF plays a prominent role in growth cone actin dynamics. Although the major actin binding regions of proteins in the ADF/cofilin family have been identified as domains in the C-terminal half of the protein(48Yonezawa N. Nishida E. Ohba M. Seki M. Kumagai H. Sakai H. Eur. J. Biochem. 1989; 183: 235-238Crossref PubMed Scopus (44) Google Scholar, 49Yonezawa N. Homma Y. Yahara I. Sakai H. Nishida E. J. Biol. Chem. 1991; 266: 17218-17221Abstract Full Text PDF PubMed Google Scholar), recent evidence suggests that amino acids near the N terminus also play a role.2 2T. Obinata, personal communication. Presented as abstract (Kusano, K., Abe, H., and Obinata, T. (1993) Cell Struct. Funct.18, 639). Introduction of a phosphate group nearby an actin binding domain could directly affect the interactions, but until the structure of the entire molecule is resolved, we can only speculate on how the phosphorylation of the penultimate serine brings about the changes in the actin binding properties of ADF. It is interesting to note, however, that extensions to the N terminus (three amino acids in recombinant ADF (21Adams M.E. Minamide L.S. Duester G. Bamburg J.R. Biochemistry. 1990; 29: 7414-7420Crossref PubMed Scopus (50) Google Scholar) or the addition of a 26-kDa glutathione S-transferase fusion protein)3 3H. Abe, T. Obinata, L. S. Minamide, and J. R. Bamburg, manuscript in preparation. do not inhibit the actin depolymerizing or pH-sensitive F-actin binding of ADF. The identification of the regulatory phosphorylation site on ADF and the development of site-directed mutants will enable further investigation into the cellular role of this regulation. We are currently isolating stable, clonal populations of various cell lines, which have been transfected with vectors expressing the ADF(A3) and ADF(E3) mutants and will examine the effects of these mutations on actin-based functions including cell motility, neurite outgrowth, and nuclear rod formation. We thank Dr. K. M. Swiderik, Department of Immunology, Beckman Research Institute of the City of Hope, Duarte, CA for expert technical assistance in obtaining the mass spectrometry results presented here and Craig Miles, Macromolecular Resources, Colorado State University, for protein sequencing. We also thank Judith Sneider for technical assistance and Drs. Norman Curthoys, Barbara Bernstein, and A.-Young Woody for critical reading of the manuscript.