INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Skeletal muscle differentiation is characterized by withdrawal of myoblasts from the cell cycle, induction of muscle-specific gene expression, and cell fusion into multinucleated myotubes. All of these events are coordinated by a family of muscle-specific transcription factors including MyoD (8), Myf5 (4), myogenin (12, 56), and MRF4 (39). These proteins show homology within a basic helix-loop-helix (bHLH) domain that mediates both heterodimerization with ubiquitous activating bHLH proteins such as E12 and E47 and DNA binding to a specific sequence, CANNTG, called the E box (9, 25, 30). One of the most remarkable properties of myogenic factors is that their ectopic expression in nonmuscle cells forces these cells into muscle differentiation, a process known as myogenic conversion (6, 8). Although capable of inhibiting cell proliferation (7, 47) and inducing differentiation, MyoD is constitutively expressed in proliferating myoblasts long before differentiation takes place, implying that its activity is regulated in replicating cells (26, 49). Indeed, when cultured myoblasts are exposed to serum or growth factors such as basic fibroblast growth factor and transforming growth factor , both muscle differentiation and MyoD activity are inhibited (34, 48). One of the inhibitory mechanisms that target MyoD in proliferative myoblasts involves the Id family of proteins. These HLH proteins, which are devoid of DNA-binding basic domains, can heterodimerize with bHLH factors, thus inhibiting their binding to DNA (3). In addition, like most transcription factors (23), MyoD is a phosphoprotein (49), and its phosphorylation could constitute an important mechanism by which mitogens negatively regulate its activity. Protein kinase C (PKC), which is activated in response to fibroblast growth factor, was first shown to inhibit the DNA binding activity of myogenin (28) by phosphorylating a site conserved in the basic region of all myogenic HLH proteins. This same site was shown not to be required for the inhibition of MRF4 by PKC (19). Protein kinase A (PKA) was also demonstrated to repress the activity of Myf5 and MyoD, albeit via an indirect mechanism (55).

Because differentiation requires withdrawal from the cell cycle, kinases involved in cell cycle control are likely candidates for the inhibition of MyoD in the proliferative state. Cyclin-dependent kinases (CDKs), in association with their regulatory partners, the cyclins, are key regulators of cell cycle progression. cdk2-cyclin A/E and cdk4-cdk6/cyclin D are involved in the G₁/S transition, whereas cdk1 (also called cdc2)-cyclin A/B is implicated in the G₂/M transition of the cell cycle (32, 40). Several lines of evidence support the involvement of CDKs in the regulation of muscle differentiation. Overexpression of cyclin D1 inhibits MyoD muscle-specific gene transactivation (38, 42, 43). Cyclins A and E have, to a lesser extent, the same effect alone or in combination with cdk2, whereas the effects observed with cyclins B, D2, and D3 remain controversial (17, 38, 42, 43). Cyclin-dependent inhibition of muscle gene transactivation requires CDK activation and can be reversed by overexpression of p21 (Waf1, Cip1), one of the general CDK inhibitors. Interestingly, induction of p21 constitutes one of the earliest markers of cell cycle exit associated with myoblast differentiation (2) and depends on MyoD (16, 18). Although CDKs appear to be involved in the inhibition of MyoD in proliferating myoblasts, no direct phosphorylation of MyoD by CDKs has been described.

In this report, we show that MyoD phosphorylation is high in proliferative C2 myoblasts and diminishes during the course of muscle differentiation. By tryptic phosphopeptide mapping and mutational analysis of MyoD, we show that a CDK consensus site comprising Ser200 is phosphorylated in vivo in myoblasts and in vitro by cdk1 (cdc2) and cdk2. We demonstrate that a nonphosphorylable Ser200 mutant of MyoD shows both higher activity in transactivating muscle-specific gene expression through the E box and greater ability to convert 10T1/2 fibroblasts to muscle cells. We also report that Ser200 phosphorylation is involved in specifying the short half-life of MyoD in proliferative myoblasts by showing that MyoD-Ala200 displays a half-life threefold higher than that of wild-type MyoD protein (MyoD-wt). These data show that direct CDK-dependent phosphorylation of MyoD on Ser200 is involved in negatively regulating MyoD activity.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell culture. C2.7 myoblasts (36) were kept in growth medium (50% Dulbecco modified Eagle medium [DMEM; ICN, Orsay, France], 50% HaM F12 [Gibco BRL, Life Technologies, Cergy Pontoise, France]) supplemented with 10% fetal calf serum (FCS; DAP, Neuf-Brisach, France). To induce terminal differentiation, myoblasts were placed in differentiation medium (DMEM, 2% FCS). A nearly complete differentiation is obtained in 60 h. Mouse 10T1/2 cells (American Type Culture Collection, Biovaley, France) were maintained in growth medium and moved to differentiation medium following transfection to induce myogenic conversion.

Purified proteins. Production and purification of full-length murine MyoD have been described elsewhere (52); MyoD-Ala5 and MyoD-Ala200 were purified by using the same protocol. The active kinase cdk1-cyclin B was purified from starfish oocytes (24).

2D gel electrophoresis. Proteins extracts from proliferating and differentiated C2.7 cells were analyzed by two-dimensional (2D) electrophoresis by the method of O'Farrell (33). First-dimension electrofocusing gels contained 9.5 M urea, 2% (wt/vol) Nonidet P-40 (NP-40), and 5% dithiothreitol (DTT). The ampholine mixture used was composed of 60% (vol/vol) ampholine pH 3 to 10 and 40% (vol/vol) ampholine pH 5 to 7. The second dimension was performed on sodium dodecyl sulfate (SDS)-12% polyacrylamide gels. Following transfer onto nitrocellulose membranes, MyoD isoforms were revealed by Western blotting with anti-MyoD monoclonal antibody 5.8A (kindly provided by P. Dias and P. Houghton, Memphis, Tenn.). For calf intestinal phosphatase (CIP) treatment, nuclear extracts from C2.7 myoblasts were treated with 20 U of CIP (Promega, Charbonnieres, France) for 30 min at 37°C.

Western blotting. Nitrocellulose membranes were blocked with phosphate-buffered saline (PBS) containing 10% dry milk and incubated either with anti-CDKs (Santa Cruz Biotechnology, Santa Cruz, Calif.) or anti-MyoD polyclonal antibody C20 (Santa Cruz Biotechnology) diluted 1/300 or with monoclonal anti--tubulin (Sigma, St. Quentin Fallavier, France) diluted 1/2,000 in PBS containing 0.5% bovine serum albumin for 1 h at room temperature. After three washes in PBS, blots were incubated with secondary antibodies (horseradish peroxidase-conjugated goat anti-rabbit or goat anti-mouse; Amersham, les Ulis, France) and developed by using the Amersham ECL (enhanced chemiluminescence) reagent.

In vivo labeling and immunoprecipitation. Cells (myoblasts and myotubes) cultured in 60-mm-diameter dishes were labeled with [³²P]orthophosphate (1 mCi/ml) for 2 h at 37°C. After three washes with PBS, cells were lysed in 100 µl of a mixture composed of (by volume) Laemmli buffer-2% NP-40, 10 mM -glycerophosphate, and 1 mM phenylmethylsulfonyl fluoride. Lysates were boiled for 3 min, diluted to 500 µl in radioimmunoprecipitation assay (RIPA) buffer (10 mM Na₂HPO₄, 100 mM NaCl, 5 mM EDTA, 50 mM NaF, 1 mM DTT, 0.1% NP-40, 5% sodium deoxycholate, 0.1% SDS), and homogenized by passages through a 21-gauge needle. Following 10 min of centrifugation at 13,000 rpm, supernatants were precleared by incubation with protein G-Sepharose beads (Pharmacia, Orsay, France) and incubated for 2 h at 4°C with anti-MyoD monoclonal antibody 5.8A; 20 µl of protein G-Sepharose beads was added for 30 min at 4°C, and the beads were washed three times with RIPA buffer and once with PBS before loading onto a SDS-12% gel for polyacrylamide gel electrophoresis (PAGE). The radioactivity was analyzed by autoradiography. The amount of immunoprecipitated MyoD was estimated by Western blotting using antibody C20 anti-MyoD polyclonal as described above.

Immunoprecipitation and CDK assays. Cells (myoblasts and myotubes) were washed twice in 1× PBS and scraped in 1 ml of PBS. After centrifugation at 3,000 rpm, pellets were resuspended in lysis buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 0.4% NP-40, 2 mM EDTA, 50 mM NaF, 10 mM -glycerophosphate, 1 mM ATP, 2 µg each of leupeptin and aprotinin per ml, 2 mM sodium vanadate, 2 mM DTT). After 10 passages through a 21-gauge needle, cell lysates were cleared by centrifugation at 13,000 rpm. Protein concentrations were determined by using a Bio-Rad DC kit. Extracts (200 µg) were immunoprecipitated with either monoclonal anti-cdk1 (C7) or polyclonal anti-cdk2 (M2) or anti-cdk5 (C8) antibodies for 2 h at 4°C. All antibodies (Santa Cruz Biotechnology) were used at a 1/50 dilution. Depending on antibody species, protein A- or G-Sepharose was added for 1 h at 4°C. After centrifugation, pellets were washed three times with lysis buffer, twice in lysis buffer containing 400 mM NaCl, and twice in kinase buffer (25 mM HEPES [pH 7.4], 25 mM MgCl₂, 25 mM -glycerophosphate, 2 mM DTT, 0.1 mM NaVO₃). Purified cdk1-cyclin B or beads containing CDKs immunoprecipitated from C2.7 cells were incubated in 20 µl of kinase buffer containing 50 µM ATP and 5 µCi of [-³²P]ATP (Kodak X-ray films) and then used for Western blot analyses.

Phosphopeptide mapping. ³²P-labeled MyoD (immunoprecipitated from myoblasts) and bacterially expressed MyoD-wt, MyoD-Ala5, and MyoD-Ala200 phosphorylated in vitro by cdk1-cyclin B were excised from SDS-gels and digested twice with 10 µg of trypsin for 12 h at 37°C in buffer containing 200 mM NH₄H₂CO₃. Digests were desalted by repeated lyophilization and loaded onto thin-layer chromatography plates (Merck-Coger, Paris, France) for 2D peptide mapping. The first dimension was run for 30 min at 1,000 V at pH 1.9 (formic acid-acetic acid-water [50:150:1,800]); second-dimension chromatography was performed in phosphochromo buffer (isobutyric acid, 1-butanol-pyridine-acetic acid-water [15:10:3:2]). ³²P-labeled peptides were subsequently visualized by autoradiography of the thin-layer chromatography plates.

Mutation of the seven proline-directed sites present on MyoD. The MyoD cDNA was mutagenized in the Moloney sarcoma virus long terminal repeat expression vector pEMSV-scribe. MyoD mutants were obtained by oligonucleotide-directed mutagenesis using a QuickChange site-directed mutagenesis kit (Stratagene, Ozyme, Montigny le Bretonneux, France) as instructed by the manufacturer. Oligonucleotides were 30 to 32 nucleotides in length, with 14 to 15 nucleotides of exact homology with MyoD in the region flanking the substitution. Mutant clones were screened with the oligonucleotide used for mutagenesis, which had been labeled with T4 polynucleotide kinase by using [-³²P]ATP. Selected clones were used for preparative plasmid isolation and then sequenced by using a Sequenase 2.0 kit (U.S. Biochemical) and [³⁵S]dATP (3,000 Ci/mmol; Amersham). The mutants resulting from a change of serine or threonine to alanine were designated MyoD-Ala5, MyoD-Ala37, MyoD-Ala200, MyoD-Ala262, MyoD-Ala277, MyoD-Ala296, and MyoD-Ala298. Substitution of alanine for serine at amino acids 5 and 200 was also performed in the T7 procaryote expression construct pET3a-MyoD (52).

Phosphorylation of MyoD wild-type and mutant proteins. MyoD-wt and MyoD alanine mutants were obtained by in vitro translation as described by the manufacturer (TnT coupled reticulocyte lysate system; Promega) and were phosphorylated by cdk1-cyclin B as described above but in the absence of [-³²P]ATP. ³⁵S-radiolabeled proteins were visualized by autoradiography.

Transfection and chloramphenicol acetyltransferase (CAT) assays. Plasmids used for transfection were pEMSV-MyoD wild type and mutants, pCMV-gal (Stratagene, Paris, France), pAch-CAT+ and pAchmutCAT+ (gifts from J. Piette, Montpellier, France) (35), and p4E-TK-CAT and pTK-CAT (gifts from H. Weintraub) (54). For Western blot analyses, transfection of 10T1/2 cells were carried out with a ratio of 5 µl of Lipofectamine to 1 µg of DNA as described by the manufacturer (Gibco BRL, Life Technologies).

Myogenic conversion. 10T1/2 cells were transfected with 1 µg of plasmid expressing either MyoD-wt or MyoD-Ala200; 24 h after transfection, cells were collected for Western blot analyses or moved to differentiation medium for 60 h and either used for Western blot analyses as described above or processed for immunofluorescence as previously described (51). Anti-MyoD polyclonal antibody C20 (Santa Cruz Biotechnology) was used to identify transfected cells, and anti-troponin T antibody JLT-12 (Sigma) was used to quantify the level of differentiation. MyoD antibodies were visualized with biotinylated anti-rabbit antibodies and Texas red-streptavidin (Amersham). Fluorescein-conjugated anti-mouse antibodies were used to detect troponin T antibodies. DNA was stained with Hoechst dye (Sigma).

Cycloheximide treatment. 10T1/2 cells were transfected with either pEMSV-MyoD-wt or pEMSV-MyoD-Ala200 in 35-mm-diameter dishes as described above. Transfected cells were treated with cycloheximide (Sigma) at 15 µg/ml for the indicated times and harvested for Western blot analyses. MyoD was stained with anti-MyoD antibody C20 as described above. For each experiment, -tubulin was used as an internal control. Western blots were scanned and quantified by using ImgCalc sensitivity software (developed by N. J. C. Lamb; details upon request) on a Silicon Graphics Indigo2 Workstation.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Hyperphosphorylation of MyoD in proliferative myoblasts. To examine the posttranslational modifications of MyoD during myogenesis, we have analyzed MyoD protein expression in the course of C2.7 differentiation by 2D gel electrophoresis followed by western blotting. As shown in Fig. 1A, four major MyoD isoforms of similar intensities are detected in proliferative myoblasts, with some other minor spots in the more acidic part of the gel. By contrast, only two major spots are visible after 60 h of differentiation, a stage when most of the cells have differentiated into myotubes. To confirm that posttranslational modifications of MyoD involve mainly phosphorylation, we treated nuclear extracts from proliferative myoblasts with CIP and analyzed the mobility of MyoD by 2D gel electrophoresis and western blotting as before. As shown in Fig. 1A, phosphatase treatment resulted in only one major MyoD isoform, which resolved at the basic side of the gel. To confirm that MyoD is more phosphorylated in myoblasts than in myotubes, C2 proliferative myoblasts and myotubes were labeled with [³²P]orthophosphate, MyoD was immunoprecipitated and separated by SDS-PAGE, and its phosphorylation was analyzed by autoradiography of the gel. MyoD phosphorylation was higher in myoblasts than in myotubes (Fig. 1B, top); Western blot analysis of the immunoprecipitate shows that the phosphorylated band corresponds to MyoD, and the amounts of immunoprecipitated MyoD were comparable between myoblasts and myotubes (Fig. 1B, bottom).

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Hyperphosphorylation of MyoD in C2.7 myoblasts. (A) Total cellular proteins were extracted from C2.7 cells and separated by 2D gel electrophoresis. The different isoforms of MyoD (arrows) were detected by Western blot analysis using extracts made from proliferative myoblasts (P) or cells placed for 60 h in differentiation medium or nuclear from extracts from proliferative myoblasts treated with CIP (P+cip). Migrations of the first dimension electrofocusing (IEF) gel (with the acid side indicated) and the second SDS-PAGE dimension are illustrated by arrows. (B) MyoD was immunoprecipitated from [32P]orthophosphate-labeled proliferative (Prol) or differentiated (Diff) C2.7 cells as described in Materials and Methods. MyoD was resolved by SDS-PAGE, its phosphorylation analyzed by autoradiography (32p), and the amount of immunoprecipitated MyoD protein was controlled by Western blot (WB) analysis.

CDK-dependent phosphorylation of MyoD. CDKs, a family of kinases implicated throughout the cell cycle, are potentially involved in both phosphorylation of MyoD and inhibition of its activity in myoblasts (16-18, 38, 42, 43). Analysis of the amino acid sequence of MyoD revealed seven putative CDK phosphorylation sites distributed in the NH₂- and COOH-terminal regions of the protein outside the bHLH domain (see Fig. 5A and below). As such, MyoD represents a potential target for direct phosphorylation by CDKs.

View larger version (63K): [in this window] [in a new window]   FIG. 2.   cdk1 and cdk2 from C2 myoblasts phosphorylate efficiently MyoD. (A) cdk1, cdk2, and cdk5 were immunoprecipitated from proliferative (P) and differentiated (72) C2.7 cells and assayed for kinase activity against H1 histone (top panel) and MyoD (bottom panel). Shown are autoradiograms of the different kinase reactions following SDS-PAGE. (B) Western blot (WB) analysis of cdk1, cdk2, and cdk5 after immunoprecipitation from proliferative (P) and differentiated (72) C2.7 cells.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Electrophoretic shift of MyoD after phosphorylation by cdk1-cyclin B. (A) Bacterially produced MyoD protein was phosphorylated by purified cdk1-cyclin B kinase for the time indicated. Shown is the autoradiogram from the 32P phosphorylation reaction. Unphosphorylated MyoD migrates as 45-kDa band that shifts to 47 kDa in the course of phosphorylation by cdk1-cyclin B. (B) Western blot showing MyoD migration after SDS-PAGE of mitotic cells extracts (M) and proliferative C2.7 nuclear extracts before (C2P) and after (C2P cip) treatment with CIP.

MyoD is phosphorylated on a CDK site in vivo. To compare the sites phosphorylated on MyoD in vivo in proliferative myoblasts with those targeted in vitro by purified cdk1-cyclin B and cdk1 or cdk2 immunoprecipitated from myoblasts, we used tryptic digestion of MyoD followed by 2D phosphotryptic peptide mapping. As illustrated in Fig. 4A, two major phosphotryptic peptides (spots 1 and 2 in the left panel) are obtained after digestion of ³²P-labeled MyoD immunoprecipitated from proliferating myoblasts. In the case of MyoD phosphorylated by cdk1-cyclin B in vitro, two major phosphotryptic peptides are also resolved (arrowed in the middle panel). We have observed the same pattern when analyzing MyoD phosphorylated in vitro by cdk1 or cdk2 immunoprecipitated from proliferating C2.7 cells (unpublished observations). When in vivo- and in vitro-phosphorylated MyoD tryptic peptides are mixed (right panel), only one of the two peptides resolved in vivo (Spot 2) comigrated with one of the phosphopeptides from cdk1-phosphorylated MyoD (the other major site phosphorylated in vitro was never observed in vivo). To estimate which sites were phosphorylated, we used the PhosPepSort program to obtain a prediction of the mobility map for the tryptic phosphopeptides expected after phosphorylation of the seven proline-directed sites on MyoD (Fig. 4B). As shown in Fig. 4B, the seven sites should lie in four phosphopeptides spanning amino acids (aa) 1 to 9 (Ser5), aa 10 to 41 (Ser37), aa 188 to 202 (Ser200), and aa 258 to 319 (Ser262, Ser277, Ser298, and Thr296). According to the mobility prediction, the two major phosphopeptides obtained after in vitro phosphorylation of MyoD by cdk1 and cdk2 would correspond to phosphorylation of Ser5 and Ser200. Of these two peptides, only one, which is predicted to contain Ser200, is common between in vitro and in vivo maps.

View larger version (31K): [in this window] [in a new window]   FIG. 4.   Phosphotryptic map analysis of MyoD phosphorylation in C2 myoblasts and following in vitro phosphorylation by cdk1-cyclin B. (A) Phosphorylation sites on MyoD were analyzed by tryptic digestion followed by 2D peptide mapping. The phosphopeptide map for in vivo MyoD phosphorylation (left panel) was obtained after immunoprecipitation of 32P-radiolabeled MyoD from dividing myoblasts. The phosphopeptide map for in vitro-phosphorylated MyoD (middle panel) was performed on bacterially produced MyoD phosphorylated in vitro by cdk1-cyclin B. In vivo and in vitro-phosphorylated MyoD were mixed following tryptic digestion, before 2D analysis (right panel). Numbers 1 and 2 indicate the two major phosphopeptides obtained from MyoD in proliferative myoblasts. The two major phosphopeptides found with MyoD phosphorylated in vitro are pointed to by arrows. (B) Mobility prediction of MyoD tryptic phosphopeptides by using the PhosPepSort program shows the pattern of migration for the four MyoD phosphopeptides that would be generated if all of the seven proline-directed sites present on MyoD were phosphorylated.

Ser200 is a major site of CDK-dependent phosphorylation. To determine precisely the site for in vivo CDK-dependent phosphorylation of MyoD, we mutated each putative CDK site in MyoD. As shown in Fig. 5A, seven putative sites are distributed in the NH₂ and COOH ends of MyoD, at positions Ser5, Ser37, Ser200, Ser262, Ser277, Thr296, and Ser298. Seven mutants were generated by site-directed mutagenesis replacing the amino acid serine or threonine by a nonphosphorylatable alanine residue and named MyoD-Ala5 to MyoD-Ala298.

View larger version (24K): [in this window] [in a new window]   FIG. 5.   Among the seven potential CDK-dependent phosphorylation sites, mutation of only Ser200 prevents the phosphorylation shift of MyoD. (A) MyoD protein exhibits seven proline-directed sites on its amino acid sequence: Ser5, Ser37, Ser200, Ser262, Ser277, Thr296, and Ser298. Each of the Ser and Thr residues was mutated to Ala as described in Materials and Methods. (B) In vitro-translated 35S-labeled MyoD-wt (WT) and individual Ala mutants of MyoD (Ala5, Ala37, Ala200, Ala262, Ala277, Ala296, and Ala298) were incubated with (circled "p") or without cdk1-cyclin B. Phosphorylation shifts were visualized after SDS-PAGE. Shown is the autoradiogram of the SDS-PAGE analysis. (C) Western blot analysis of MyoD overexpression in 10T1/2 cells transiently transfected with either MyoD-wt or each of the seven Ala mutants. Shown is the ECL detection of MyoD immunoreactivity.

View larger version (48K): [in this window] [in a new window]   FIG. 6.   Phosphotryptic map analysis of MyoD-Ala5 and MyoD-Ala200 following in vitro phosphorylation by cdk1-cyclin B. Phosphorylation sites on MyoD were analyzed by tryptic digestion followed by 2D peptide mapping. Phosphopeptide maps were obtained for bacterially produced MyoD-wt (left panel), MyoD-Ala200 (middle panel), and MyoD-Ala5 (right panel) phosphorylated in vitro by cdk1-cyclin B. The two major phosphopeptides found with MyoD-wt phosphorylated in vitro are pointed to by arrows labeled Ser5 and Ser200.

MyoD-Ala200 shows enhanced muscle gene-specific transactivating activity. To assess the consequence of Ser200 phosphorylation on MyoD activity, we initially compared the abilities of MyoD-wt and MyoD-Ala200 to transactivate muscle-specific gene expression. Plasmids expressing either MyoD-wt or MyoD-Ala200 were cotransfected with a CAT reporter gene containing the acetylcholine receptor -subunit promoter (pAch-CAT+) in 10T1/2 cells. Transfected cells were kept in proliferative medium for 36 h, and transactivation of the reporter gene estimated by CAT assay. In each case, plasmid pCMV-gal was cotransfected as an internal control for transfection efficiency. As expected (Fig. 7A), the low basal activity of the wild-type reporter gene was highly enhanced by MyoD-wt; moreover, MyoD-Ala200 further increased the level of CAT reporter activity threefold over that obtained with MyoD-wt. Such an increase was not observed with any of the other MyoD mutants (unpublished observations). To demonstrate that this increased transactivation activity of MyoD-Ala200 required the E boxes, we performed the same experiment with a mutant form of the reporter, pAchmutCAT+, where the E boxes had been mutated (35). Neither MyoD-wt nor MyoD-Ala200 could transactivate the reporter plasmid pAchmutCAT+ (unpublished observations). We also used plasmid p4E-TK-CAT, which contains a simplified enhancer comprising four tandem copies of the E-box sequence (from the muscle creatine kinase gene enhancer) upstream of the minimal thymidine kinase promoter and, as a control, plasmid pTK-CAT, devoid of E boxes. As shown in Fig. 7A, MyoD-Ala200 was again threefold more efficient than MyoD-wt in transactivating CAT expression from the p4E-TK-CAT construct, which confirms the effect observed with pAch-CAT.

View larger version (77K): [in this window] [in a new window]   FIG. 7.   MyoD-Ala200 displays enhanced transactivating and myogenic activities. (A) pAch-CAT+ and p4E-TK-CAT reporter constructs were cotransfected in 10T1/2 cells with either pEMSV or encoding plasmid pEMSV-MyoD-wt or pEMSV-MyoD-Ala200 and pCMV-gal. Transfected cells were grown for 36 h in DMEM containing 10% FCS, and CAT activity was measured and corrected with respect to -galactosidase activity. CAT activities are expressed relative to that of each reporter plasmid transfected with pEMSV-MyoD, set as 100%. (B) 10T1/2 were transfected with either pEMSV-MyoD-wt or pEMSV-MyoD-Ala200 and placed in differentiation medium (DMEM containing 2% FCS) for 60 h. Cells were fixed and stained for both MyoD and troponin T. Shown are the extents of myogenic conversion of 10T1/2 by MyoD-wt (left panels) and MyoD Ala200 (right panels) with staining for MyoD expression (a and b), troponin T expression (c and d), and DNA staining (e and f). Bar, 10 µm. The average percentages of cells expressing MyoD-wt or MyoD-Ala200 that coexpressed troponin T were calculated from two different experiments and are indicated in the bottom. The total numbers of MyoD-positive cells counted were 235 for MyoD-wt and 280 for MyoD-Ala200. (C) 10T1/2 cells were transfected with either pEMSV-MyoD-wt or pEMSV-MyoD-Ala200 as described above. Transfected cells were grown in proliferative medium (DMEM containing 10% FCS) for 24 h (P) and placed in differentiation medium for 60 h (60h). Cells were collected either before (P) or after (60h) myogenic conversion and analyzed by western blotting for MyoD and troponin T expression.

MyoD-Ala200 promotes complete myogenic conversion of 10T1/2 fibroblasts to muscle cells. Since mutation of MyoD serine 200 to alanine increased its capacity to transactivate muscle-specific gene expression, we next compared the abilities of MyoD-wt and MyoD-Ala200 proteins to trigger myogenic conversion. 10T1/2 cells were transfected with expression vectors coding for either MyoD-wt or MyoD-Ala200. Transfected cells were placed in differentiation medium for 60 h and analyzed by immunofluorescence for expression of MyoD and troponin T as a differentiation marker. The efficiency of myogenic conversion was estimated as the percentage of cells expressing MyoD that also expressed troponin T. The immunofluorescence presented in Fig. 7B reveal that MyoD-Ala200 was significantly more efficient than MyoD-wt in converting 10T1/2 cells to myotubes. After 60 h of differentiation (Fig. 7B), nearly all MyoD-Ala200-expressing cells had differentiated into troponin T-positive myotubes whereas 35% of MyoD-wt-expressing cells remained negative for troponin T. A clear difference in activity between the two MyoD proteins was also observed at the phenotypic level. As shown in Fig. 7B, MyoD-Ala200-expressing cells formed many giant interconnected myotubes. We never observed this extent of differentiation with the wild-type protein even if conversion was allowed for up to 5 days. To accurately quantify the increase in myogenic conversion ability of MyoD-Ala200 versus MyoD-wt, conversions were done as before but MyoD and troponin T expression levels were analyzed by western blotting. As shown in Fig. 7C, after 24 h (wt P and Ala200 P), similar levels of MyoD-wt and MyoD-Ala200 are expressed (upper panel), with no detactable troponin T expression (lower panel). After 60 h in differentiation medium (wt 60h and Ala200 60h), troponin T is expressed (lower panel) and is present at levels fivefold higher in MyoD-Ala200- than MyoD-wt-overexpressing cells. It is worth noting that in these culture conditions, MyoD-wt protein level appears to be about twofold lower than the MyoD-Ala200 level (upper panel, wt 60h and Ala200 60h), probably as a result of differences in protein half-life (see below).

Ser200 phosphorylation regulates MyoD protein turnover. Because phosphorylation by CDKs has been involved in the targeted degradation of several factors such as p27 (53), we next investigated a potential link between CDK-dependent phosphorylation of MyoD and its specific degradation. If phosphorylation of Ser200 is implicated in MyoD degradation, mutation of Ser200 would be expected to increase the half-life of MyoD. To test this hypothesis, we transfected MyoD-wt and MyoD-Ala200 in 10T1/2 cells and determined the half-life of MyoD following cycloheximide treatment (Fig. 8A). The half-life of MyoD-wt was found to be about 40 min (average of values obtained from two different experiments [Fig. 8B]), in agreement with a previous report from Thayer et al. (50). Expression of -tubulin, a stable protein, was not modified 2 h after cycloheximide addition. By contrast, MyoD-Ala200 was found to be more stable than MyoD-wt, with an half-life extended to 140 min.

View larger version (31K): [in this window] [in a new window]   FIG. 8.   Impeding MyoD Ser200 phosphorylation stabilizes the protein. (A) 10T1/2 cells were transfected with either pEMSV-MyoD-wt or pEMSV-MyoD-Ala200 and grown for 24 h in proliferative medium before addition of cycloheximide (15 µg/ml) to the medium for 0, 15, 30, 60, or 120 min. MyoD and -tubulin protein levels were determined by immunoblot analysis at the indicated times after cycloheximide addition. (B) Immunoblots were quantified by densitometric scanning, and MyoD protein levels (corrected with respect to tubulin expression) were expressed relative to that observed before cycloheximide treatment, set as 100%.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

An essential step during myogenesis is the reorientation of the proliferative cell cycle toward differentiation processes in which the transcription factor MyoD plays clearly a critical role. Although overexpression of MyoD can drive nontransformed fibroblasts into differentiation (8), myoblasts proliferate efficiently while expressing MyoD. A mechanism other than regulation of MyoD expression is therefore required to explain why myoblasts do not enter differentiation. In this report, we demonstrate that phosphorylation plays an active role in preventing differentiation through a negative effect on MyoD activity. We observe that MyoD phosphorylation is high in myoblasts and reduced during differentiation and show for the first time that MyoD is a direct substrate for phosphorylation by CDKs. Comparative peptide mapping combined with site-directed mutagenesis show that MyoD is phosphorylated by cdk1 (cdc2) and cdk2 on Ser200 both in vitro and in proliferating myoblasts. Indeed, substitution of Ser200 by an alanine (MyoD-Ala200) prevents the appearance of hyperphosphorylated MyoD after either its phosphorylation by cdk1-cyclin B in vitro or overexpression in 10T1/2 cells. The phosphorylation of this site by CDKs is clearly inhibitory to MyoD function, as demonstrated by the greater myogenic activity of MyoD-Ala200 than of MyoD-wt.

cdk1 and cdk2 phosphorylate MyoD on serine 200 in proliferative myoblasts. Our data show that the kinases responsible for the phosphorylation of Ser200 on MyoD in proliferative myoblasts include the mitotic activator kinase cdk1-cyclin B and cdk2-cyclin A/E kinase, which is present and active from mid-G₁ until mitosis. Overexpression of cyclin D1 was shown to promote hyperphosphorylation of MyoD (42, 43), suggesting that cdk4-cyclin D1 could directly phosphorylate MyoD. However, in contrast to the efficient phosphorylation of MyoD by cdk2 and cdk1 in vitro, we have been unable to observe an effective phosphorylation of MyoD in assays using immunoprecipitated cdk4 from C2 myoblasts (unpublished observations). This observation is in agreement with that of Skapek et al. (43), who found that baculovirus-produced cdk4-cyclin D1 fails to phosphorylate MyoD. It thus appears that the hyperphosphorylation of MyoD observed after cyclin D1 overexpression may be the result of an indirect effect rather than a direct cdk4-dependent phosphorylation of MyoD. cdk1- and cdk2-dependent phosphorylation requires the Ser/Thr-Pro (S/T-P) cluster to be followed immediately by a basic residue (Lys/Arg [46]), which is the case for the motif containing Ser200 that we have identified on MyoD. Of the 6 other S/T-P sites present on MyoD, only serine 5 is also a potential site for phosphorylation by cdk1 and cdk2. Although this site is phosphorylated in vitro, as shown by phosphopeptide map analysis (Fig. 4 and 6), it was never found phosphorylated in vivo in C2.7 myoblasts (Fig. 4), and mutation of Ser5 to alanine did not cause any significant effect on MyoD-dependent transcriptional activation of a reporter gene containing the acetylcholine receptor promoter (unpublished observations). Ser200 is thus the only cdk1- and cdk2-dependent site used in vivo. It is also the only site responsible for the electrophoretic shift in mobility seen when MyoD is phosphorylated either in vitro or in vivo. It is to be noted that the sequence immediately surrounding and including Ser200 is highly conserved in MyoD from many different species (unpublished observations). We cannot rule out the possibility that phosphorylation of Ser200 is a prerequisite for phosphorylation of MyoD at other sites. In this context, kinases other than CDKs may also phosphorylate MyoD and contribute to its inhibition in proliferative myoblasts. In addition to Ser200, a second phosphopeptide is clearly observed by 2D tryptic mapping of MyoD isolated from proliferative myoblasts. It does not correspond to any of the peptides resolved after in vitro phosphorylation of MyoD by cdk1-cyclin B (Fig. 4), implying that a kinase other than cdk1 or cdk2 also phosphorylates MyoD in vivo. PKA and PKC have been implied to negatively regulate myogenic factors, although this regulation appeared to be indirect in the case of PKA (28, 55). According to the PhosPepSort mobility analysis shown in Fig. 4, it is unlikely that PKC is responsible for this MyoD phosphorylation in myoblasts. Indeed, the predicted map of MyoD-Thr 115 phosphopeptide (equivalent to the site phosphorylated by PKC on myogenin [28]) does not correspond to the second tryptic phosphopeptide observed in vivo (spot 1 in Fig 4). The kinase responsible for this phosphopeptide remains to be identified.

Impeding Ser200 phosphorylation enhances MyoD activity. The mutant MyoD-Ala200 was more efficient than MyoD-wt in converting 10T1/2 cells to muscle cells. This augmentation was correlated to an enhanced ability of MyoD-Ala200 (about threefold higher than that of MyoD-wt) to transactivate muscle gene expression via the E box (Fig. 7A). However, this increased transactivating capability was not linked to significant alteration in MyoD DNA binding affinity. Indeed, by band shift analysis, we observed that phosphorylation of MyoD by cdk1-cyclin B did not alter the binding of MyoD homodimer to the E-box and had marginal effects on MyoD-E12 DNA binding (unpublished observations). That DNA-binding and transcriptional activities of myogenic factors are not necessarily correlated has been reported in previously. For instance, in myoblasts blocked from differentiating by transforming growth factor , myogenic factors appear to retain DNA-binding activity without activating muscle gene transcription (5). Interestingly, MyoD-containing complexes capable of binding to an E box are observable in nuclear extracts from both proliferating myoblasts and differentiated myotubes (reference 41 and our unpublished observations). It thus appears that the transcriptional activity of MyoD is not necessarily reflected by its capacity to bind to DNA. Because DNA-binding activity was not the mechanism by which phosphorylation of MyoD Serine 200 could control MyoD activity, we have compared the stabilities of MyoD-wt and MyoD-Ala200. We found, in agreement with a recent report from Song et al. (45), that MyoD-Ala200 was more stable than MyoD-wt, suggesting that phosphorylation of Ser200 decreases MyoD activity by reducing its half-life. This phosphorylation seems to be required for targeting MyoD to the ubiquitin pathway (45). A rapid turnover of MyoD may allow a fine regulation of its activity in myoblasts. High-level expression of MyoD obtained by ectopic expression into nonmuscle cells is known to stop cell cycle progression before S phase, allowing cells to engage into the differentiation process (7, 47). Controlled degradation of MyoD may be necessary to prevent MyoD from reaching a threshold that can interfere with normal cell cycle events before myoblasts have received the appropriate signal to differentiate. By isolating C2 cells that have lost MyoD expression, Horwitz (22) found that the autoactivation loop of MyoD is tightly linked to protein stability. In addition, the reduced half-life that we observed for phosphorylated MyoD may result from a change in MyoD-associated protein. Ser200 phosphorylation may reduce the association of MyoD with partners such as pRb (14), MEF-2 proteins (29), the coactivator p300 (11, 37), or proteins such as Id. cdk2-dependent phosphorylation has been shown to change the interaction specificity of the HLH protein Id3 (10). CDK-dependent phosphorylation has also been shown to change the interaction between the transcription factor E2F and pRB (21, 44). In a similar way, free MyoD could be more sensitive to degradation. Gerber et al. (13) have recently shown that MyoD, in addition to being able to bind DNA and activate muscle-specific gene expression, can remodel chromatin at binding sites in muscle gene regulatory regions and activate transcription at previously silent loci. This ability of MyoD to activate genes within inactive chromatin mapped to a cysteine- and histidine-rich region of the amino terminus and a region extending from aa 218 to 269 in the carboxy terminus of MyoD. Interestingly, deletion of a region between aa 170 and 209 (that includes Ser200) increased the ability of MyoD to initiate transcription of endogenous genes, implying a repressive role of this region in chromatin remodeling by MyoD. It is tempting to hypothesize that in addition to modulating MyoD half-life and transactivating ability, Ser200 phosphorylation may cause conformational changes of MyoD and thereby modulate intra- or intermolecular interactions involved in remodeling chromatin.