Phosphoinositide 3-Kinase Induces the Transcriptional Activity of MEF2 Proteins during Muscle Differentiation*

The activity of phosphoinositide 3-kinase (PI3-K) is essential for the differentiation of skeletal muscle cells by largely unknown mechanisms. Here we show that inhibition of PI3-K activity by the pharmacological agent LY294002 affects early processes of myoblast differentiation including the transcriptional activation of myogenin. Previous studies indicated that transcription of myogenin was dependent on MyoD and MEF2 proteins. We find that expression of a dominant negative form of PI3-K or growth in the presence of LY294002 inhibits cellular activity of MEF2 but not of MyoD. Evidence reveals that whereas MEF2 transcriptional activity is inhibited, its DNA binding activity remains unaffected. Recent studies demonstrated that phosphorylation by p38 mitogen-activated protein kinase (MAPK) induced transcriptional activity of MEF2 proteins. We show that the phosphorylation of MEF2 occurring during muscle differentiation is prevented if the activity of PI3-K is inhibited. However, our results also indicate that p38 MAPK is not affected by PI3-K in muscle cells. Nevertheless, p38 MAPK can substitute for PI3-K in the induction of MEF2 and muscle transcription. Together, these findings indicate that PI3-K affects skeletal muscle differentiation by inducing phosphorylation and transcriptional activity of MEF2 proteins in a parallel but distinct route from p38 MAPK.

The differentiation of skeletal muscle cells involves two major stages: the withdrawal of myoblasts from the cell cycle and subsequent expression of myotube-specific genes. Proliferating myoblasts express two myogenic transcription factors from the basic helix-loop-helix family, MyoD and Myf5, prior to the onset of muscle differentiation (1)(2)(3). Once activated, MyoD and Myf5 induce the withdrawal of myoblasts from the cell cycle and the expression of another myogenic basic helix-loop-helix factor, myogenin, as well as transcription factors from the MEF2 family. Together, myogenin and MEF2 family members cooperate in the activation of many muscle structural genes (4,5).
Insulin-like growth factors (IGFs) 1 have been implicated in the control of skeletal muscle growth and differentiation both in embryonic development and in muscle regeneration (6 -9). Unlike other growth factors, IGF-I stimulates myoblast proliferation as well as differentiation (10 -12). Studies have indicated that the effects of IGF-I on proliferation and differentiation are temporally separated. In proliferating myoblasts, IGFs increase the expression of factors involved in cell cycle progression (10,13). After myoblasts are withdrawn from the cell cycle, IGFs promote muscle differentiation by inducing the expression or activity of myogenic regulatory factors (14). It was recently suggested that IGFs induce these two functions by activating separate signal transduction pathways (15,16). Proliferation is mediated by the mitogen-activated protein (MAP) kinase pathway, whereas differentiation is mediated by phosphoinositide 3-kinase (PI3-K). Several studies proved the involvement of PI3-K in differentiation by inhibition of its activity. Two approaches were used: treatment of cells with chemical inhibitors like LY294002 and wortmannin or expression of a mutated form of the regulatory subunit of PI3-K, p85, functioning as a dominant negative (15,(17)(18)(19). These studies indicated that PI3-K activity was involved in the induction of muscle gene expression.
PI3-K activates downstream molecules like protein kinase B (PKB). The involvement of PKB in muscle differentiation was recently demonstrated. PKB induces muscle-specific transcription (17) and functions as a positive mediator of muscle cell survival (20). No other molecules that function downstream to PI3-K and PKB have so far been identified as being involved in myogenesis.
Some earlier studies suggested that IGF-I induced the expression of myogenin in L6 myoblasts (21). The induction of myogenin expression did not occur immediately following administration of IGF-I, suggesting that the effect was not direct and involved intermediary molecules.
The present study aimed at understanding how PI3-K affects muscle-specific transcription. We suggest that PI3-K affects skeletal muscle differentiation by inducing the transcriptional activity of MEF2 proteins. Our results explain why PI3-K is necessary for the expression of myogenin and of late muscle differentiation genes.

Materials
LY294002 and SB203580 were supplied by Calbiochem. PKB (catalog no. 9272) and phosphospecific PKB (Ser 473 ) (catalog no. 9271) antibodies were supplied by New England BioLabs. An antibody to MEF2A and MEF2C proteins (SC-313) was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). MyoD polyclonal antibody (SC-760) was from Santa Cruz Biotechnology, and a monoclonal antibody to myosin heavy chain (MF-20) was a gift from Dr. S. Tapscott. p38 MAPK anti-body was from Santa Cruz Biotechnology (SC-535). Protein A-Sepharose was supplied by Sigma.

Cell Culture
L8 cells were a gift of Dr. David Yaffe (32). 10T1/2 cells were obtained from ATCC (Manassas, VA). 10T1/2 cells that expressed the MyoDestrogen receptor (ER) chimera protein were described by Hollenberg et al. (33). Cell lines were maintained in Dulbecco's modified Eagle's medium supplemented with 15% calf serum (Hyclone), penicillin, and streptomycin (growth medium (GM)). To induce differentiation, we used Dulbecco's modified Eagle's medium supplemented with 10 g/ml insulin and 10 g/ml transferrin (differentiation medium (DM)). Differentiation of 10T1/2 cells that expressed MyoD-ER protein was induced by the addition of DM that contained 10 Ϫ7 M estradiol.

Growth of Cells in the Presence of LY294002 or SB203580
LY294002 and SB203580 were dissolved in dimethyl sulfoxide to a concentration of 10 mM that was added directly to the differentiation medium to a final concentration of 10 -20 M, as indicated. Control cells were incubated with the same volumes of dimethyl sulfoxide without the inhibitors. The medium was replaced every 24 h by medium containing fresh inhibitor.

Transfections
Transfections were performed by calcium phosphate precipitation as described (34) or using the TransFast reagent of Promega according to the recommended protocol. Cells in 6-cm TC dishes (Corning) were transfected with a total amount of 10 g (or 5 g, using TransFast) of the following plasmid DNA: 1 g of pCMV-LacZ or pMSV-LacZ, 3 g of CAT reporter gene, 3 g of the various expression vectors. "Empty vectors" of pCDNA3 or EMSV containing promoters of cytomegalovirus and murine sarcoma virus long terminal repeat, respectively, were added to equalize the promoter content in each transfection experiment. Following transfection, the medium was replaced with either GM or DM for another 24 -48 h. Transfection efficiencies were tested in soluble X-gal assays, and amounts of extracts used for the CAT assays were adjusted accordingly (35).

In Situ ␤-Galactosidase Staining of Cells
Transfected cells were washed three times in phosphate-buffered saline and then fixed in 0.5% gluteraldehyde and washed again in phosphate-buffered saline. Cells were then incubated for 3 h at 37°C in the dark with the following staining solution: 1 mg/ml X-gal, 5 mM K 4 Fe(CN) 6 , 5 mM K 3 Fe(CN) 6 , 2 mM MgCl 2 . Stained cells were photographed.

Immunohistochemical Staining
Cells were fixed and immunostained as described (36). The primary antibodies used were polyclonal anti-MyoD (Santa Cruz Biotechnology) and monoclonal anti-MHC (MF-20). The immunochemically stained cells were viewed at ϫ 200 magnification in a fluorescence microscope (Olympus, model BX50) and photographed.

Western Analysis
Cells were lysed as described for the kinase assays, and equal amounts of extracted proteins (30 -100 g) were loaded and separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocel-lulose filters. Membrane was incubated in blocking buffer (1ϫ TBS, 0.1% Tween 20, 5% (w/v) nonfat dry milk) and then with the same buffer with the first and secondary antibodies. Between the second and third incubations, membranes were washed three times in 0.1% Tween 20 and 1ϫ TBS. Immunoblotting was conducted with the following antibodies: anti-PKB, anti-phospho-PKB (New England BioLabs), anti-p38 MAPK (Santa Cruz Biotechnology), and anti-phospho-p38 MAPK (Thr 180 /Tyr 182 ) (New England BioLabs) (1:1000). Proteins were visualized using the enhanced chemiluminescence kit of Pierce.
Labeling with [ 32 P]Orthophosphate-L8 cells were incubated in Dulbecco's modified Eagle's medium without sodium phosphate, to which [ 32 P]orthophosphate was added at 0.75 mCi/ml for 3 h before proteins were extracted in radioimmune precipitation buffer.
Immunoprecipitation of MEF2-Equal amounts of labeled proteins were incubated with antibody to MEF2 for 2 h and an additional 1 h with protein A-Sepharose beads. Beads were washed four times with radioimmune precipitation buffer containing 0.5 M NaCl and once in radioimmune precipitation buffer. Beads were resuspended in SDS sample buffer, and proteins were separated over 10% SDS-polyacrylamide gel electrophoresis.

Gel Shift Analysis of MEF2
The double-stranded DNA probe contained a binding site of MEF2 from the MCK enhancer region. The sequence of the sense strand was 5Ј-AGCTCGCTCTAAAAATAACCCTGGATCC-3Ј.
Proteins were extracted from 10T1/2 MyoD-ER cells that were induced to differentiate in DM and estradiol (10 Ϫ7 M) during 24 h in the absence or presence of 20 M of LY294002. 20 g of proteins in binding buffer (10% glycerol, 2 mM spermidine, 0.1 mg/ml bovine serum albumin, 2 mM MgCl 2 , 0.02% Nonidet P-40, 0.1 mM EDTA, 50 mM KCl, 10 mM HEPES, pH 7.9) were added to 500 ng of poly(dI-dC)-poly(dI-dC) and labeled probe (50,000 cpm, ϳ1 ng). Proteins and DNA were incubated for 30 min at room temperature. Unlabeled competitor DNA or a specific antibody to MEF2 was incubated with proteins on ice prior to the addition of labeled probe. DNA-protein complexes were separated over 4% native polyacrylamide gel (0.25ϫ TBE).

RESULTS
PI3-K is involved in early stages of myoblast cell differentiation. The involvement of PI3-K in skeletal muscle differentiation was previously demonstrated using different approaches to inhibit this kinase (15,(17)(18)(19). In order to study the involvement of PI3-K in the differentiation of L8 myoblasts, we added insulin to these cells at concentrations known to bind the IGF-I receptor (37). To find out whether the PI3-K pathway was activated, we monitored the phosphorylation state of PKB, which is downstream to PI3-K in the pathway. We found that phosphorylation of PKB was induced immediately after the addition of insulin to cells and gradually declined to its basal levels after 24 h (Fig. 1A). Phosphorylation of PKB was prevented in cells treated with LY294002, an inhibitor of PI3-K, suggesting that PKB phosphorylation was dependent on the activity of PI3-K (Fig. 1B, lane 3). Phosphorylation of PKB was only marginally reduced in cells treated with SB203580, an inhibitor of p38 MAPK (Fig. 1B, lane 4). Like insulin, treatment of myoblasts with IGF-1 induced the phosphorylation of PKB (data not shown). The immediate transient phosphorylation of PKB suggests that the pathway may be involved in the differentiation process during this time window. To test this possibility, L8 cells were induced to differentiate for 48 h, during which LY294002 was added for different periods of time. An inhibitor that was added during the first 24 h and then removed prevented differentiation as reflected by the low expression of MLC2 (Fig. 1C, lane 4). An inhibitor added for 24 h after a delay of 24 h did not affect differentiation (Fig. 1C, lane 5). These results suggest that PI3-K is involved during early stages of the differentiation process.
Inhibition of PI3-K Affects the Expression of the Myogenin Gene-The involvement of PI3-K during the initial 24 h of muscle differentiation prompted us to analyze the expression of "early induced genes" that are turned on during that period of time. We chose to study a 10T1/2 cell line expressing a chimera protein of MyoD and the hormone-binding domain of estrogen receptor (MyoD-ER) (33). The chimera protein resides in an inactive form in the cytoplasm of cells. The addition of estradiol induces its translocation to the nucleus, where it activates the expression of muscle genes. Induction of myogenesis in this cell line follows a very precise timetable, since it depends on the activation of the MyoD-ER protein. Cells were induced to differentiate in the absence or presence of LY294002, and the expression of several genes was analyzed ( Fig. 2A). The expression of myogenin and myosin light chain 2 was significantly repressed in cells grown in the presence of the PI3-K inhibitor compared with cells that were induced to differentiate in the absence of the inhibitor (compare lanes 2-4 with lanes 5-7). However, the expression of p21-WAF1 and MEF2A was not affected in cells treated with the inhibitor. It was previously suggested that the expression of these two genes is dependent on the MyoD protein (38,39). Therefore, the results in Fig. 2A indicate that the activity of the MyoD-ER protein was not affected, while other transcription factor(s) needed for the expression of myogenin and MLC2 genes were inhibited. To test this possibility, cells were induced to differentiate in the presence of the protein synthesis inhibitor cycloheximide (Fig. 2B, lane 4). Under these conditions, genes that are directly induced by MyoD independently of newly synthesized transcriptional mediators can be identified. The expression of p21-WAF1 and MEF2A was not affected, while that of myogenin and MLC2 was totally prevented in the presence of cycloheximide (compare lane 3 with lane 4). These results indicate that the activity of MyoD is not sufficient to induce the transcription of myogenin. The transcription of myogenin is affected by an additional transcription factor(s) that may be inhibited in cells grown in the presence of LY294002. To test whether the inhibitor of PI3-K affected the promoter activity of myogenin, we studied a reporter gene that contained the myogenin promoter driving the expression of the lacZ gene. The promoter was active in 10T1/2 cells expressing the MyoD-ER protein after their treatment with estradiol, as visualized by ␤-galactosidase staining (Fig. 3). The expression of this reporter gene was inhibited in cells grown in the presence of LY294002, as observed by the lesser number of ␤-galactosidase-stained cells (Fig. 3). The addition of LY294002 to cells transfected with another reporter gene driven by the cytomegalovirus promoter did not affect the activity of this promoter. Hence, LY294002 specifically inhibited the expression of the myogenin promoter, suggesting that PI3-K activity was involved in the transcription of myogenin.
Inhibition of PI3-K Represses the Activity of MEF2 Transcription Factors-The myogenin promoter contains binding sites for MyoD and MEF2 family members (40). In transgenic mice models, it was shown that both binding sites were necessary and sufficient for muscle-specific transcription of this gene (26). To analyze which of these transcription factors was affected by the PI3-K pathway, the activities of MyoD and MEF2 transcription factors were independently analyzed. 10T1/2 cells were transfected with an expression vector of MyoD and a specific reporter gene containing a minimal promoter and MyoD binding sites (4R-tk-CAT) or an expression vector of MEF2C and a reporter gene containing MEF2 binding sites (MEF2-MCK-CAT). Each transcription factor induced the expression of its respective reporter gene (Fig. 4, lanes 2 and 6). Treatment of the transfected cells with the inhibitor of PI3-K, LY294002, or transfection of a dominant negative form of PI3-K (p85⌬iSH2-N) significantly inhibited the activity of MEF2 (lanes 7 and 8) while only marginally affecting the activity of MyoD (lanes 3 and 4). These findings indicate that PI3-K is involved in the activity of the MEF2C transcription factor. The inhibition of myogenin expression by the PI3-K inhibitor observed in Figs. 2 and 3 might be mediated by suppression of MEF2 activity.
The Transactivation Potential of MEF2 Proteins Is Repressed by the PI3-K Inhibitor-Classical transcriptional activators contain two separate activities: DNA binding and transactivation. Using a gel shift assay, we examined whether inhibition of PI3-K affected the DNA binding activity of MEF2 in the 10T1/2 cells expressing the chimera MyoD-ER protein. MEF2 binding activity was significantly induced in extracts of cells treated with estradiol (Fig. 5A, lane 2 versus lane 3), confirming previous results and those shown in Fig. 2 according to which MEF2 expression was induced during muscle differentiation (38,41). Treatment of the cells with LY294002 during the period at which they were induced to differentiate did not affect the DNA binding activity of MEF2 proteins (Fig. 5A, lane 4). The binding was specific, as demonstrated by homologous competition and supershift of the complex with a specific MEF2 antibody (Fig. 5A, lanes 5 and 6). Therefore, the DNA binding activity of MEF2 is not affected when PI3-K is inhibited.
The transactivation potential of MEF2 was studied in a GAL4 activator/reporter system. The transcriptional activators were chimeric proteins consisting of the transactivation domain (TAD) of MEF2 and the DNA-binding domain of the yeast GAL4 protein. The reporter gene contained GAL4 binding sites in the promoter. Using this system, we analyzed the activity of MEF2 TAD independently of its DNA binding activity. The transcriptional activities of several chimera activators were analyzed, all of which activated the transcription from the GAL4 reporter gene (Fig. 5B). Treatment of the transfected cells with LY294002 inhibited the activity of GAL4-MEF2 proteins (Fig. 5B, lanes 2 and 4) but did not affect the activities of GAL4-MyoD or GAL4-VP16 (lanes 6 and 8). We suggest therefore that the effect of PI3-K inhibitor is specific to the TAD of MEF2 proteins.
Expression of a MEF2C-VP16 Protein Partially Substitutes for the Function of PI3-K in Muscle-specific Transcription-Given that the TAD of MEF2 protein is a target for PI3-K, we assumed that artificial activation of MEF2 could bypass the requirement for PI3-K activity. We tested a chimera protein containing the MEF2 DNA-binding domain (amino acids 1-117) linked to the VP16 TAD (MEF2C-VP16) in the activation of MEF2-responsive genes (28). Both MEF2C and MEF2C- VP16 activated transcription from a muscle creatine kinase reporter gene (peϩAT-CAT) (Fig. 6A, lanes 2 and 6). Treatment of cells with LY294002 or the expression of a mutated form of the regulatory subunit of PI3-K (p85⌬iSH2-N) significantly inhibited MCK transcription in cells transfected with MEF2C but did not affect transcription in cells transfected with MEF2C-VP16 expression vector (compare lanes 3 and 5 with lanes 7 and 9). Similar results were obtained when we analyzed the transcription of a minimal promoter containing MEF2 binding sites (data not shown).
To analyze whether MEF2C-VP16 can rescue the expression of endogenous muscle genes, we transfected expression vectors of MyoD and MEF2C, or alternatively MEF2C-VP16, to 10T1/2 cells that were induced to differentiate in the absence or presence of LY294002. We followed the expression of endogenous myosin heavy chain in the transfected cells by immunostaining with specific antibodies to MyoD and MHC (Fig. 6B). Treatment of cells transfected with expression vectors of MyoD and MEF2C with LY294002 reduced the percentage of MHC-expressing cells (Fig. 6C, lanes 1 and 2). Expression of MHC in cells that were transfected with expression vectors of MyoD and MEF2C-VP16 and treated with LY294002 was also inhibited, but to a lesser extent (Fig. 6C, lanes 3 and 4). Therefore, expression of MEF2C-VP16 could partially substitute for the activity of PI3-K needed for the expression of endogenous MHC. Activity of PI3-K Affects the Phosphorylation State of MEF2 Proteins in L8 Muscle Cells-Several kinases phosphorylate MEF2 proteins at their transactivation domain and induce their activity (29,(42)(43)(44). Since the results presented here suggest that PI3-K affects the transactivation potential of MEF2, we decided to examine whether this activity was also involved in the phosphorylation of MEF2. L8 cells at different stages of differentiation, grown in the presence or absence of LY294002, were incubated with orthophosphate. MEF2 proteins were immunoprecipitated from extracts of these cells and analyzed (Fig. 7). MEF2 proteins were not phosphorylated in myoblasts (lane 1) but were phosphorylated in cells grown for 24 h in differentiation medium (lane 2) (25). Treatment of cells with PI3-K inhibitor (LY294002) or p38 MAPK inhibitor (SB203580), while grown in differentiation medium, blocked the phosphorylation of MEF2 proteins (lanes 3 and 4). The similar steady state levels of MEF2 proteins that were metabolically labeled with methionine suggest that the differences in phosphate labeling were not due to differences in the amount of the proteins (Fig. 7, lower panel). These findings indicate that phosphorylation of MEF2 during muscle differentiation is affected by PI3-K as well as by p38 MAPK.
p38 MAPK Is Not Affected by PI3-K, but It Can Replace PI3-K Activity in Muscle-specific Transcription-Previous studies indicated that p38 MAPK activity was induced in differentiating L8 cells and activated MEF2 proteins (25). The similarities in the effects of p38 MAPK and PI3-K on MEF2 activity suggest that these kinases share the same pathway. To test this possibility, L8 cells were grown in the presence of PI3-K inhibitor, LY294002, and the phosphorylation state of p38 MAPK was analyzed using an antibody recognizing the dually phosphorylated p38 MAPK. As was demonstrated before (25) phosphorylation of p38 MAPK was induced after 24 h in DM (Fig. 8A,  lanes 1 and 2). The amount of phosphorylated p38 MAPK was even higher in cells grown for 24 h in DM and in the presence of PI3-K inhibitor (lane 3). Therefore, p38 MAPK is not a target of PI3-K in muscle cells. Phosphorylation of p38 MAPK was unaffected in cells growing in the presence of its own inhibitor, SB203580, as expected for a compound that inhibits activity but not the activation of p38 MAPK (45). Although these results suggest that PI3-K and p38 MAPK are not functioning in the same pathway, similarities in their activities (see Figs. 6 and 7) prompted us to analyze whether p38 MAPK can substitute for PI3-K in the induction of muscle-specific transcription. For that purpose, expression vectors of MyoD and an activated form of MKK6 were transfected to 10T1/2 cells induced to differentiate in the absence or presence of LY294002 (Fig. 8B). The expression of endogenous myosin heavy chain in the transfected cells was analyzed by immunostaining with specific antibodies to MyoD and MHC. MyoD induced the expression of endogenous MHC in 65% of the transfected cells grown in the  3 and 4). These data suggest that p38 MAPK activity can replace PI3-K in the induction of endogenous MHC expression.
Next, we analyzed the transcriptional activity of a MCK reporter gene containing two MEF2 binding sites in the enhancer region (Fig. 8C). The transcription of MCK was induced by the expression of MyoD (Fig. 8B, lane 2). MCK transcription was repressed if cells were also treated with LY294002 or SB203580 or transfected with an expression vector of p85 functioning as a dominant negative of PI3-K (⌬p85) (lanes 3-5). Inhibition of PI3-K in transfected cells did not affect MCK transcription if the activated form of MKK6 was expressed (lanes 7 and 9). However, inhibition of p38 MAPK with SB203580 inhibited MCK transcription (lane 8), probably because this inhibitor blocks the activity of p38 MAPK downstream to the ectopically expressed MKK6 protein. Thus, the results presented in Fig. 8 indicate that although PI3-K does not affect the activity of p38 MAPK, these two kinases function in a similar way in the activation of muscle-specific transcription. DISCUSSION Insulin-like growth factors have been shown to induce muscle transcription via PI3-K (7), but how PI3-K affects muscle transcription has not been shown. In this study, we suggest that PI3-K affects muscle-specific transcription by regulating the activity of MEF2 proteins.
First, we have shown that the PI3-K pathway is active for 24 h after myoblasts are induced to differentiate. This is demonstrated by the immediate and transient phosphorylation of PKB, a kinase in the PI3-K pathway (Fig. 1A). Inhibition of PI3-K during the same period is sufficient to block the tran- scription of such muscle structural genes as MLC2 (Fig. 1C). The transcription of several "early genes" is induced within this time window. Analysis of the expression of three "early induced genes," p21-WAF1, MEF2A, and myogenin, reveals that only myogenin is affected by the activity of PI3-K ( Fig. 2A). Inhibition of PI3-K activity does not affect the expression of p21-WAF1 and MEF2A. p21-WAF1 is involved in the withdrawal of myoblasts from the cell cycle. This raises the possibility that, although functioning during an early stage, PI3-K may not be involved in the exit of myoblasts from the cell cycle. Musaro and Rosenthal elegantly demonstrated that PI3-K affector, IGF-I, expressed postmitotically from the MLC promoter-induced myogenic differentiation (14). However, PI3-K may also control the exit of myoblasts from the cell cycle. Inhibition of PI3-K in chicken myoblasts blocked the expression of MyoD involved in the withdrawal of myoblasts from the cell cycle (18). In this work, the effects of PI3-K on differentiation were studied in a cell line expressing the MyoD protein under a constitutive promoter (10T1/2 MyoD-ER). Therefore, this cell line enabled us to study effects of PI3-K in stages subsequent to the expression of MyoD. Therefore, PI3-K may affect processes occurring during and after the exit of myoblasts from the cell cycle.
Previous studies indicated that IGFs induce the expression of myogenin (3). Inhibition of myogenin expression by antisense oligonucleotides blocked the myogenic effect of IGF-I in myoblasts (46). Our work indicates a correlation between the expression of myogenin and the activity of PI3-K (Figs. 2 and  3). Hence, it appears that IGFs affect the expression of myogenin via the activation of PI3-K. Members of two families of transcription factors, MyoD and MEF2, regulate transcription of myogenin. We aimed to define which of the transcription factors was affected by PI3-K. Studies with the inducible MyoD-ER cell line suggested that the activity of the MyoD-ER protein was not affected by PI3-K. In these cells, the expression of p21-WAF1 and MEF2A was still induced in the presence of the inhibitor of PI3-K (Fig. 2). MyoD-ER also induced the expression of these genes in the presence of cycloheximide, suggesting that these genes were activated directly by MyoD. The promoter of the waf1 gene contains E boxes necessary for the induction of transcription by the MyoD protein (39).
MyoD-ER protein alone was not sufficient to induce the expression of myogenin if protein synthesis or PI3-K activity was inhibited. This raised the possibility that the PI3-K pathway could affect the newly synthesized MEF2 proteins (see Fig. 2). This assumption was substantiated by further studies. In a reporter gene assay, the transcriptional activity of MEF2C protein was significantly inhibited, while that of MyoD was marginally affected when cells were grown in the presence of PI3-K inhibitor or transfected with a dominant negative form of PI3-K (Fig. 4). It is possible that, although MyoD is not the prime target of PI3-K, its activity may have been affected as a result of physical and functional interactions with MEF2 proteins (28). DNA binding activity was not affected (Fig. 5A), but the transactivation potential of MEF2 proteins was inhibited in cells grown in the presence of LY294002 (Fig. 5B). The effect of PI3-K on the TAD of MEF2 proteins was demonstrated using chimeras of MEF2. A protein containing the TAD of MEF2A or MEF2C and the DNA-binding domain of the yeast GAL4 was affected by PI3-K (Fig. 5B), while a protein containing the TAD of VP16 and the DNA-binding domain of MEF2C was not affected by PI3-K (Fig. 6A). Therefore, it seems likely that the TAD of MEF2 proteins is the main target of PI3-K. Moreover, the fact that MEF2C-VP16 protein can partly restore musclespecific transcription when PI3-K activity is crippled strengthens the suggestion presented above that MEF2 proteins are involved in the PI3-K pathway.
Residues at the TAD of MEF2 are phosphorylated by p38 MAPK and big MAP kinase 1, also known as ERK5 (29,43,47). This phosphorylation leads to the transcriptional activation of the protein. In a previous study, we found that MEF2 proteins were phosphorylated during muscle differentiation and suggested that p38 MAPK was involved in their phosphorylation and subsequent activation (25). In this study, we found that inhibition of PI3-K as well as of p38 MAPK in L8 cells prevented the phosphorylation of MEF2 (Fig. 7). This observation raises the possibility that both p38 MAPK and PI3-K are involved in the phosphorylation of MEF2 proteins. The membrane-localized PI3-K is probably not directly phosphorylating nuclear MEF2 proteins. Therefore, PI3-K may cause MEF2 phosphorylation via another kinase. Results presented in Fig.  8A indicate that p38 MAPK is not the kinase mediating PI3-K activity. Two candidates that may function as mediators of PI3-K in the phosphorylation of MEF2 proteins are PKB and big MAP kinase 1. PKB mediates PI3-K activity in phosphorylating many substrates, while big MAP kinase 1 phosphorylates MEF2 proteins but is not related yet to one of the well established signaling pathways. Although p38 MAPK and PI3-K do not share the same signaling pathway, our data strongly indicate that both regulate the transcriptional activity of MEF2 proteins. This is shown by the ability of activated MKK6 to substitute PI3-K in the induction of MEF2 and muscle transcription (see Fig. 8). Interestingly, we observed that SB203580 was consistently more potent than LY294002 in the inhibition of MEF2 (see Figs. 6A and 8C). This activity of SB203580 may be explained by the recent observation that at micromolar concentrations this inhibitor blocks not only p38 MAPK but also the phosphorylation of PKB by phosphoinositide-dependent protein kinase 1 (48). Therefore, at the concentrations used in this work, SB203580 blocked both pathways affecting MEF2, while LY294002 blocked only one.
Two recent studies indicated that IGF-1 induced hypertrophy of differentiated myotubes through the calcium-activated phosphatase, calcineurin (49,50). In another work, two independent pathways of calcium signaling, one of which was calcineurin, were shown to activate MEF2 transcription factors (51). These data and our results raise the possibility that IGFs affect MEF2 activity by inducing several signaling pathways.
In addition to their role in muscle differentiation, members of the MyoD family are also involved in controlling cell cycle progression (39,52). The activity of the pRb protein is also required for the withdrawal of myoblasts from the cell cycle (53)(54)(55). Cells that do not express the pRb protein (RbϪ/Ϫ) cannot undergo late events of muscle differentiation. A recent study by Lassar and colleagues (56) demonstrated that MEF2C was functionally inactive in RbϪ/Ϫ cells. The DNA binding activity of MEF2 was normal, but its TAD was not functional, and the requirement for pRb was partially bypassed by fusing MEF2C to the VP16 TAD. In their study, the authors showed that MEF2 requires both MyoD and pRb for its full activity. Taken together, the above studies and our results suggest that pRb and PI3-K affect the transactivation potential of MEF2 proteins. We propose a model in which the PI3-K and/or p38 MAPK pathways are activated only in the presence of MyoD and pRb, which induce cell cycle withdrawal. Consequently, the IGF signal is transmitted and induces muscle differentiation only after myoblasts withdraw from the cell cycle. If proven correct, this is the first model that explains why withdrawal of myoblasts from the cell cycle is a prerequisite for activation of the differentiation program.