[New insight into MyoD regulation: involvement in rhabdomyosarcoma pathway?]. / Nouvelles données sur MyoD: des éléments de réponse aux rhabdomyosarcomes?

The transcription factor MyoD, member of the myogenic regulators family, induces differentiation in precursor cells by its ability to arrest cell proliferation and to activate muscle specific genes. MyoD plays a key role in the antagonism between proliferation and differentiation. The withdrawal from the cell cycle and the activation of muscle differentiation are related to the level of MyoD protein. The cyclin E-cdk2 complex, one of the key regulators of the G1/S transition is directly implicated in the degradation of MyoD by the ubiquitin-proteasome pathway, leading the myoblasts to proliferate. The display of this control in normal myoblasts suggests that its deficiency in the muscle stem cells could lead to the formation of rhabdomyosarcomas which have lost both the control of cell proliferation and the transcriptional activity of MyoD.

Cyclin E-cdk2 phosphorylation promotes late G1-phase degradation of MyoD in muscle cells.

Proliferating myoblasts already express MyoD before the induction of differentiation. Overexpression of MyoD in normal and transformed cell lines was shown to block cells from entering S phase, suggesting that the MyoD growth suppressive effect must be tightly controlled in growing myoblasts. Here we show that during G1 phase, but not in G2, MyoD abundance is down-regulated by the ubiquitin-proteasome pathway through phosphorylation of serine 200. Roscovitine, a specific inhibitor of cyclin-Cdk2 complexes, prevents both phosphorylation and degradation of MyoD in G1. Inhibition of the ubiquitin-dependent proteasome pathway by MG132 results in stabilization of MyoD-wt, with little effect on a MyoD mutant where serine 200 is replaced by an alanine. Our results show that MyoD Ser200 is the substrate for phosphorylation by cyclin E-Cdk2 stimulating its degradation by the ubiquitin-proteasome system which controls MyoD levels in G1. Phosphorylation/degradation of MyoD at the end of G1 thus represents the regulatory checkpoint in growing myoblasts allowing progression into S phase in a manner similar to the recently examplified cdk2-phosphorylation/degradation of p27(Kip1).

Mutation of MyoD-Ser237 abolishes its up-regulation by c-Mos.

Recently we have shown that Mos could activate myogenic differentiation by promoting heterodimerisation of MyoD and E12 proteins. Here, we demonstrate that MyoD can be efficiently phosphorylated by in vitro kinase assay with purified Mos immunoprecipitated from transfected cells. Comparative two-dimensional tryptic phosphopeptide mapping combined with site-directed mutagenesis revealed that Mos phosphorylates MyoD on serine 237. Mutation of serine 237 to a non-phosphorylable alanine (MyoD-Ala237) abolished the positive regulation of MyoD by Mos following overexpression in proliferating 10T1/2 cells. Taken together, our data show that direct phosphorylation of MyoD-Ser237 by Mos plays a positive role in increasing MyoD activity during myoblast proliferation.

Stabilization of MyoD by direct binding to p57(Kip2).

Recent data have demonstrated the role of Cdk1- and Cdk2-dependent phosphorylation of MyoD(Ser200) in the regulation of MyoD activity and protein turnover. In the present study, we show that in presence of p57(Kip2), MyoD(Ala200), a MyoD mutant that cannot be phosphorylated by cyclin-Cdk complexes, displayed activity 2-5-fold higher than of MyoD(Ala200) alone in transactivation of muscle-specific genes myosin heavy chain, creatine kinase, and myosin light chain 1. Furthermore, p57(Kip2) increases the levels of MyoD(Ala200) in cotransfected cells. This result implies that p57(Kip2) may regulate MyoD through a process distinct from its function as a cyclin-dependent kinase inhibitors. We report that overexpression of p57(Kip2) increased the half-life of MyoD(Ala200). This increased half-life of MyoD involves a physical interaction between MyoD and p57(Kip2) but not with p16(Ink4a), as shown by cross-immunoprecipitation not only on overexpressed proteins from transfected cells, but also on endogenous MyoD and p57(Kip2) from C2C12 myogenic cells. Mutational and functional analyses of the two proteins show that the NH(2) domain of p57(Kip2) associates with basic region in the basic helix-loop-helix domain of MyoD. Competition/association assays and site-directed mutagenesis of the NH(2) terminus of p57(Kip2) identified the intermediate alpha-helix domain, located between the Cdk and the cyclin binding sites, as essential for MyoD interaction. These data show that the alpha-helix domain of p57(Kip2), which is conserved in the Cip/Kip proteins, is implicated in protein-protein interaction and confers a specific regulatory mechanism, outside of their Cdk-inhibitory activity, by which the p57(Kip2) family members positively act on myogenic differentiation.

Dimerization of the amino terminal domain of p57Kip2 inhibits cyclin D1-cdk4 kinase activity.

Previous studies have led to the proposal that a single molecule of Cki can associate with the cyclin/Cdk complex to repress its activity. On the other hand, multiple inhibitor molecules are required to inhibit Cdks. In the present work, by using differently tagged p57Kip2 proteins we demonstrate that p57Kip2 can bind to itself in vitro and in vivo. Mutational deletion analysis showed that the NH2 terminal domain of p57Kip2 is necessary and sufficient to dimerization. Using an in vitro competition/association assay, we demonstrate that cyclin D1 alone, Cdk4 alone and/or cyclin D1/Cdk4 complexes do not compete for the p57Kip2 homodimers formation. However, a mutation in the alpha-helix domain of p57Kip2 (R33L) strongly reduced homodimer formation but did not modify interaction with cyclin D1-Cdk4 complexes. Also, increasing amounts of p57Kip2 lead in vivo to a significant augmentation in the level of p57Kip2 homodimerization associated with cyclin D1-Cdk4 complexes and to a marked inhibition of the cyclin D1-Cdk4 kinase activity. Altogether, these data suggest a model whereby p57Kip2 associates with itself by using the NH2 domain to form a homodimeric species which interacts with and inhibits the cyclin D1-Cdk4 complexes.

MyoD binds to Mos and inhibits the Mos/MAP kinase pathway.

When ectopically expressed, the serine/threonine kinase Mos can induce oncogenic transformation of somatic cells by direct phosphorylation of MAP kinase/ERK kinase (MEK1), activating the mitogen-activated protein kinases ERK1 and ERK2. On the other hand, overexpression of Mos in C2C12 myoblasts is not transforming. Mos activates myogenic differentiation by promoting heterodimerization of the MyoD/E12 proteins, increasing the expression of myogenic markers and the positive autoregulatory loop of MyoD. In this study, we show that in myogenic cells, the mitogenic and oncogenic signalling from the Mos/MEK/ERK pathway is suppressed by MyoD through the formation of a heterotrimeric complex.

p57(Kip2) stabilizes the MyoD protein by inhibiting cyclin E-Cdk2 kinase activity in growing myoblasts.

We show that expression of p57(Kip2), a potent tight-binding inhibitor of several G(1) cyclin-cyclin-dependent kinase (Cdk) complexes, increases markedly during C2C12 myoblast differentiation. We examined the effect of p57(Kip2) on the activity of the transcription factor MyoD. In transient transfection assays, transcriptional transactivation of the mouse muscle creatine kinase promoter by MyoD was enhanced by the Cdk inhibitors. In addition, p57(Kip2), p21(Cip1), and p27(Kip1) but not p16(Ink4a) induced an increased level of MyoD protein, and we show that MyoD, an unstable nuclear protein, was stabilized by p57(Kip2). Forced expression of p57(Kip2) correlated with hypophosphorylation of MyoD in C2C12 myoblasts. A dominant-negative Cdk2 mutant arrested cells at the G(1) phase transition and induced hypophosphorylation of MyoD. Furthermore, phosphorylation of MyoD by purified cyclin E-Cdk2 complexes was inhibited by p57(Kip2). In addition, the NH2 domain of p57(Kip2) necessary for inhibition of cyclin E-Cdk2 activity was sufficient to inhibit MyoD phosphorylation and to stabilize it, leading to its accumulation in proliferative myoblasts. Taken together, our data suggest that repression of cyclin E-Cdk2-mediated phosphorylation of MyoD by p57(Kip2) could play an important role in the accumulation of MyoD at the onset of myoblast differentiation.

Overexpression of Mos(rat) proto-oncogene product enhances the positive autoregulatory loop of MyoD.

The myogenic b-HLH transcription factor MyoD activates expression of muscle-specific genes and autoregulates positively its own expression. Various factors such as growth factors and oncogene products repress transcriptional activity of MyoD. The c-mos proto-oncogene product, Mos, is a serine/threonine kinase that can activate myogenic differentiation by specific phosphorylation of MyoD which favors heterodimerization of MyoD and E12 proteins. Here we show that overexpression of Mos enhances the expression level of MyoD protein in myoblasts although phosphorylation of MyoD by Mos does not modify its stability but promotes transcriptional transactivation of the MyoD promoter linked to the luciferase reporter gene. Moreover, co-expression of MyoD with Mos(wt) but not with the kinase-inactive Mos(KM) greatly enhances expression of endogenous MyoD protein and the DNA binding activity of MyoD/E12 heterodimers in 10T1/2 cells. Our data suggest that Mos increases the ability of MyoD to transactivate both muscle-specific genes and its own promoter and could therefore participate in the positive autoregulation loop of MyoD and muscle differentiation.

Mos activates myogenic differentiation by promoting heterodimerization of MyoD and E12 proteins.

The activities of myogenic basic helix-loop-helix (bHLH) factors are regulated by a number of different positive and negative signals. Extensive information has been published about the molecular mechanisms that interfere with the process of myogenic differentiation, but little is known about the positive signals. We previously showed that overexpression of rat Mos in C2C12 myoblasts increased the expression of myogenic markers whereas repression of Mos products by antisense RNAs inhibited myogenic differentiation. In the present work, our results show that the rat mos proto-oncogene activates transcriptional activity of MyoD protein. In transient transfection assays, Mos promotes transcriptional transactivation by MyoD of the muscle creatine kinase enhancer and/or a reporter gene linked to MyoD-DNA binding sites. Physical interaction between Mos and MyoD, but not with E12, is demonstrated in vivo by using the two-hybrid approach with C3H10T1/2 cells and in vitro by using the glutathione S-transferase (GST) pull-down assays. Unphosphorylated MyoD from myogenic cell lysates and/or bacterially expressed MyoD physically interacts with Mos. This interaction occurs via the helix 2 region of MyoD and a highly conserved region in Mos proteins with 40% similarity to the helix 2 domain of the E-protein class of bHLH factors. Phosphorylation of MyoD by activated GST-Mos protein inhibits the DNA-binding activity of MyoD homodimers and promotes MyoD-E12 heterodimer formation. These data support a novel function for Mos as a mediator (coregulator) of muscle-specific gene(s) expression.

The kinase inhibitor iso-H7 stimulates rat satellite cell differentiation through a non-protein kinase C pathway by increasing myogenin expression level.

We analysed the signaling pathways involved in myogenic differentiation of primary cultures of rat satellite cells using substances targeting the protein kinase C (PKC) and the cAMP protein kinase (PKA) pathways. We have previously shown that iso-H7, which putatively inhibits both PKC and PKA, strongly stimulates satellite cell differentiation, as well as the PKA inhibitor HA1004. In the study reported here, the effects of iso-H7 on satellite cell differentiation were compared to those observed in the presence of agents which reduce PKC activity. It was shown that treatments with the highly specific PKC inhibitor GF109203X or with 12-O-tetradecanoylphorbol 13-acetate (TPA) which induced a partial PKC downregulation, did not significantly alter myogenic differentiation. Northern blot analyses showed that iso-H7 activated the expression of myogenin but not that of MyoD mRNA. Concurrently, iso-H7 increased myosin light-chain mRNA expression. In contrast, TPA had no effect on these syntheses. Taken together, these results showed that iso-H7 did not act intracellularly as a PKC inhibitor but rather as a PKA inhibitor as previously suggested. Our results are compatible with the hypothesis that a reduction in PKA activity controls satellite cell myogenesis through an increased myogenin mRNA expression.

Repression of the CSF-1 receptor (c-fms proto-oncogene product) by antisense transfection induces G1-growth arrest in L6 alpha 1 rat myoblasts.

Colony Stimulating Factor (CSF-1) and the CSF-1 receptor (the c-fms product) are expressed during the proliferation of L6 alpha 1 rat myogenic cell line and both are down regulated during the formation of myotubes. In this study, we demonstrated that the expression of c-fms antisense RNA in stably transfected myoblasts repressed the CSF-1 receptor (c-fms protein) and induced a G1-growth arrest. Expression of the cyclin genes, that control passage through the G1 phase and in particular the cyclins identified as genes induced late in G1 by CSF-1 in mouse macrophages was studied in comparative Northern blot analyses of RNAs of subpopulations prepared by centrifugal elutriation of L6 alpha 1 myoblasts and induced Antifms D5 cells expressing c-fms antisense RNA. Repression of the CSF-1 receptor (c-fms product) did not affect cyclins A, B and G expression during the cell cycle. However, D-type cyclins and, at a lesser extend, cyclin E expression were dramatically altered specifically during the late G1 and early S phases, in Antifms D5 cells. These results suggest a role for the CSF-1/c-fms autocrine loop in the control of the proliferation of L6 alpha 1 rat myogenic cell line at the G1/S boundary via the D-type and E cyclins expression.

Identification of a novel regulatory element in the c-mos locus that activates transcription in somatic cells.

We have identified a DNA sequence in the 5' flanking sequence of rat c-mos gene which fulfills operational criteria for enhancers, increasing transcription from heterologous promoters in somatic cells. This new enhancer region contains some specific motifs such as two CArG boxes, two M-CAT binding sites and CCAAT consensus sequences. This Upstream Enhancer region (UER) is recognized by distinct protein complexes and particularly CArG2 and M-CAT R1 motifs, which are adjacent in the DNA sequence. Several lines of evidence indicate that none of the two CArG boxes bind to the Serum response factor (SRF). Site-directed mutations of both the CArG2 and M-CAT R1 binding sites suppress their enhancer activity. These results suggest that direct and indirect interactions involving multiple nuclear factors and distinct elements of the UER may be required for its enhancer functions in somatic cells.

Colony-stimulating factor 1 (CSF-1) is involved in an autocrine growth control of rat myogenic cells.

Colony stimulating factor-1 (CSF-1) and the CSF-1 receptor (the c-fms proto-oncogene product) are expressed during the proliferation of the L6 alpha 1 rat myogenic cell line and both are down-regulated during the differentiation to myotubes. Biologically active CSF-1 was shown to be secreted into the culture medium by L6 alpha 1 myoblasts and while they could not bind CSF-1, evidence was obtained for cell surface receptor-CSF-1 complexes. It was not possible to block the L6 alpha 1 proliferation by incubation with anti-CSF-1 antiserum or suramin. However, in L6 alpha 1 myoblasts that were stably transfected with an inducible anti-fms antisense construct, both c-fms protein expression and cell proliferation were more rapidly inhibited under induction and differentiation conditions than parental cells. Furthermore, under these conditions, the c-fms antisense transfected cells also entered myogenic differentiation more rapidly. These results suggest that autocrine regulation by CSF-1 that is intracellular may play a role in the proliferation of muscle cells and that its down-regulation leads to, and may be an obligatory step in, myogenesis.

Direct relationship between the expression of tumor suppressor H19 mRNA and c-mos proto-oncogene during myogenesis.

We have cloned and sequenced an almost complete c-DNA and the entire genomic sequence of rat the H19 gene, which is developmentally regulated in skeletal muscle. The data base comparison revealed a 92% homology with mouse gene H19. However the rat H19 ORFs do not display significant homology with the H19 ORFs from other species. In contrast to the mouse, the rat H19 mRNA is not easily detectable in fetal rat skeletal fibers. Its level increases after birth (up to 12 to 18 days) and remains stable thereafter. The pattern of H19 mRNA expression in rat muscle in vivo is very similar to the c-mos gene expression in this tissue, suggesting an interrelationship between H19 and c-mos products during muscle differentiation. Indeed, our results indicate that overexpression of c-mos protein in the muscle cell line C2C12 induces a concomitant increase of H19 mRNA expression. Furthermore, repression of c-mos protein expression by anti-sense RNAs extinguishes H19 mRNA expression and inhibits the differentiation process. These data suggest a relationship between c-mos and H19 expression and, in addition, the involvement of both gene products in the process of myogenesis.

p34cdc2 protein is complexed with the c-mos protein in rat skeletal muscle.

We have used fractionation of subcellular components of the skeletal muscle followed by Western blot analyses to study the localization of the c-mos protein in adult rat muscle. We find that p43c-mos is predominantly located in the KCl supernatant fraction. We show that immunoprecipitates of p43c-mos phosphorylate in vitro two polypeptides of about 34 kDa and 80 kDa respectively. Muscle fractionation and immunodetection studies showed that the p34 protein associated with p43c-mos is the cdc2 protein. p43c-mos is coprecipitated with p34cdc2 when using either anti PSTAIR antibody, antibody directed against the conserved COOH terminal region of the p34cdc2 and by binding to beads that contain cross-linked p13suc1, a protein known to bind p34cdc2. Likewise p34cdc2 coprecipitated with p43c-mos when using anti mos antibody. However p43c-mos is not present in histone H1 kinase active p34cdc2 complex precipitated with anti p34cdc2 COOH-terminal peptide antibody. In adult muscle tissue tubulin is not complexed with p34cdc2 and p43c-mos as previously observed in c-mos and v-mos transformed cells. Gel filtration and crosslinking experiments show that a 170 kDa complex contains c-mos and p34cdc2 proteins. In addition during postnatal development of skeletal muscle we observe modifications in the migration pattern of p34cdc2 correlated with the accumulation of p43c-mos. Our findings raise the possibility of a p43c-mos-p34cdc2 complex could play a role in the differentiation process and maintenance of myotubes in Go.

Isolation and characterization of a cDNA clone encoding for rat CSF-1 gene. Post-transcriptional repression occurs in myogenic differentiation.

A major CSF-1 (Colony-Stimulating Factor 1) mRNA 4.0 kb long was expressed during the proliferation of the L6 alpha 1 rat myogenic cells and was down-regulated after their differentiation into myotubes. A complete cDNA encoding the rat CSF-1 gene (rmCSF-1) was isolated from a cDNA library of L6 alpha 1 myoblasts and sequenced. The overall deduced amino acid sequence was 100% and 68% identical to the mouse and human CSF-1, respectively. While the previously reported mechanisms about the regulation of CSF-1 expression in TPA-treated-monocytes (Horiguchi, J., Sariban, E. and Kufe, D. (1988) Mol. Cell. Biol. 8, 3951-3954) and in fibroblasts (Falkenburg, J.H.F., Harrington, M.A., De Paus, R.A., Walsh, M.K., Daub, R., Landegent, J.E. and Broxmeyer, H.E. (1991) Blood 78, 658-665) involved a control at the transcriptional level, in contrast, the CSF-1 mRNA (half-life approximately 3 h in L6 alpha 1 myoblasts) was post-transcriptionally down-regulated during myogenesis. Inhibition of protein synthesis with cycloheximide (CHX) increased differentially the half-life of CSF-1 mRNA in L6 alpha 1 myotubes compared to L6 alpha 1 myoblasts. Finally, L6 alpha 1 myoblasts were shown to synthesize a 140 kDa homodimeric form of CSF-1. Thus, these findings, together with other results, indicate that CSF-1 gene products may play a role in the normal and neoplastic proliferation of muscular cells.

Repression of c-fos promoter by MyoD on muscle cell differentiation.

Terminal differentiation and cell proliferation are in many cases, as in muscle cells, mutually exclusive processes. While differentiating myoblasts are withdrawn from the cell cycle, myogenesis is inhibited by some mitogens and overexpression of some oncogenes, including proto-oncogene c-fos (which expresses a growth-associated protein constituting the regulatory factor AP-1 in conjunction with c-Jun). MyoD, a muscle-specific transcription factor of the basic helix-loop-helix family, acts at both levels because it triggers a muscle differentiation programme in non-muscle cells, and induces a complete block of cell proliferation. Antagonistic interaction between MyoD and c-Jun has been demonstrated. We here show that c-fos expression greatly decreases upon muscle cell differentiation, concomitant with MyoD-induced activity. We have identified a MyoD-binding site overlapping with the serum-responsive element in the c-fos promoter. We demonstrate that MyoD can act as a negative regulator for c-fos transcription by blocking serum responsiveness through this binding site. These data suggest that the MyoD negative effect on cell growth could be partly mediated by transcriptional inactivation of growth-responsive genes.

Identification of a cis acting element responsible for muscle specific expression of the c-mos protooncogene.

A series of deletion constructs of the 5' flanking region of rat c-mos gene was positioned upstream to the CAT gene and transfected into muscle and non-muscle cells. CAT activities revealed that a region located downstream of a TATA box and containing the proximal transcription start site is the muscle c-mos promoter. This promoter is more efficient in L6 alpha 1 myoblasts than in L6 alpha 1 myotubes but not in C3H10T1/2 cells. Gel shift assays demonstrated that nuclear proteins from myoblasts and myotubes formed complexes migrating differently. Footprinting analyses showed that nuclear proteins from L6 alpha 1 myoblasts protected a DNA fragment located at position nt -979 to nt -938 relative to the first ATG of the rat c-mos ORF while nuclear proteins from myotubes protected the DNA between nt -998 to nt -928. Furthermore one of protein - DNA complexes containing the proximal transcription start site, included a consensus sequence TGTC(AGT/TCG)CC(A/T)G present in the initiator element (Inr) of several genes. Southwestern blot analysis pointed to a 82kDa polypeptide as a potential candidate for trans acting factor in myoblasts. In L6 alpha 1 myotubes this polypeptide is replaced by other proteins of 40-42kDa and 82kDa. An interplay between these two complexes may constitute a developmental as well as a physiologically regulated mechanism that modulates c-mos expression during the early stages of myogenesis.