MyoD functions as a transcriptional repressor in proliferating myoblasts.

The myogenic basic helix-loop-helix (myo-bHLH) proteins are a family of transcriptional regulators expressed in myoblasts and differentiated skeletal muscle. Ectopic expression of myo-bHLH regulators transdetermines some fibroblast cell lines into myoblasts, which exit the cell cycle and differentiate into skeletal muscle when cultured in low mitogen medium. While members of the myo-bHLH family have been shown to function as transcriptional activators in differentiating muscle, the molecular basis of their function in proliferating myoblasts has not been elucidated. In this report, we present evidence that MyoD functions as a transcriptional repressor in myoblasts. We show that transcription from a cyclin B1 promoter construct is repressed in proliferating myoblasts and that repression is mediated by a pair of MyoD binding sites. We also show that transcription from the cyclin B1 promoter is repressed in proliferating C3H10T1/2 cells by ectopic expression of MyoD. These results demonstrate that MyoD can repress transcription of specific genes in proliferating cells, a novel function that may be important to maintenance of the myogenic phenotype and to cell cycle regulation in myoblasts.

The myogenic basic helix-loop-helix (myo-bHLH) proteins are a family of transcriptional regulators expressed in myoblasts and differentiated skeletal muscle. Ectopic expression of myo-bHLH regulators transdetermines some fibroblast cell lines into myoblasts, which exit the cell cycle and differentiate into skeletal muscle when cultured in low mitogen medium. While members of the myo-bHLH family have been shown to function as transcriptional activators in differentiating muscle, the molecular basis of their function in proliferating myoblasts has not been elucidated. In this report, we present evidence that MyoD functions as a transcriptional repressor in myoblasts. We show that transcription from a cyclin B1 promoter construct is repressed in proliferating myoblasts and that repression is mediated by a pair of MyoD binding sites. We also show that transcription from the cyclin B1 promoter is repressed in proliferating C3H10T1/2 cells by ectopic expression of MyoD. These results demonstrate that MyoD can repress transcription of specific genes in proliferating cells, a novel function that may be important to maintenance of the myogenic phenotype and to cell cycle regulation in myoblasts.
Cells committed to the myogenic lineage are referred to as myoblasts, and the stability of this phenotype is evident after several passages in culture (1). The development and differentiation of competent myoblasts is dependent on expression of the myogenic basic helix-loop-helix (myo-bHLH) 1 regulators (2)(3)(4)(5)(6)(7). Four myo-bHLH regulators have been identified (MyoD, myogenin, myf-5, MRF4/myf-6/herculin; Refs. 8 -13), and ectopic expression of any one will confer the myoblast phenotype to C3H10T1/2 fibroblasts (8,14). The myo-bHLH regulators appear to perform two functions in myogenic cells: conferral and maintenance of the myogenic phenotype in proliferating myoblasts and activation of muscle gene transcription in differentiated muscle cells. Activation of muscle gene transcription by myo-bHLH regulators has been well studied and is contingent on the removal or depletion of mitogenic stimuli and withdrawal from the cell cycle (15)(16)(17)(18). In contrast, myo-bHLH regulators confer the myogenic phenotype to mitogenic cells, even though they appear to be inert as activators of transcription in this cellular context.
Myogenic bHLH regulators isolated from myoblasts blocked from differentiation by high serum or transforming growth factor ␤ display significant site-specific DNA-binding activity in electrophoretic mobility shift assays (19). Inhibition of muscle gene transcription by these mitogens involves redundant mechanisms that include sequestering of E proteins by protein Id (20) and binding of certain members of the Jun family (21,22). Inhibition of myo-bHLH regulators by bFGF also may involve redundant mechanisms. Stimulation of protein kinase C by bFGF may block the activity of some myogenic bHLH regulators through phosphorylation of a conserved threonine in the basic regions of these proteins (23). Treatment of myoblasts with bFGF also has been shown to stimulate expression of cyclin D1 (24). Cyclin D1-dependent kinase activity inhibits the activity of the myo-bHLH regulators through a mechanism that may involve phosphorylation of sites in the activation domains of these proteins 2 (25,26). Although activation of most muscle-specific genes by the myo-bHLH regulators is blocked in myoblasts, autoactivation of transcription from the myo-bHLH regulator genes is not inhibited. The mechanisms of transcriptional activation of the myo-bHLH genes are not well understood, and the role of MyoD in autoactivation may involve an indirect mechanism. Indeed, recent evidence has suggested that direct binding of myo-bHLH regulators to the upstream regions of these genes is not required for autoactivation (27).
Functional complexes of the B-type cyclins and p34 CDC2/CDK1 are necessary for entry into and progression through mitosis (28 -30). Two B-type cyclins (B1 and B2) have been identified in mammals (31) that differ in subcellular distribution (32). Accumulation of B-type cyclins and cyclin B-p34 CDC2/CDK1 kinase activity during the G 2 , G 2 /M, and M phases of the cell cycle is regulated by several mechanisms. Proteolysis of B-type cyclins in late M phase occurs after ubiquitination (33), and the subcellular location of B-type cyclins also changes dramatically during different phases of the cell cycle. In addition, transcription of B-type genes is strongly induced during late S and G 2 phases (31).
The structure and sequence of the cyclin B1 promoter has been described (34,35). Two consensus binding sites for MyoD and one for products of the E2A gene were identified upstream of a basal promoter (36 -38). Since the cyclin B1 gene is active in proliferating rather than terminally differentiated cells, this promoter presented an opportunity to study the effects of MyoD on gene transcription in proliferating myoblasts. In this report, we show that the paired MyoD binding sites in the cyclin B1 promoter mediate repression of cyclin B1 gene transcription in proliferating C2C12 myoblasts. Ectopic expression of MyoD in proliferating C3H10T1/2 cells also repressed the cyclin B1 reporter. Evidence that both MyoD binding sites function in concert to mediate repression is presented. These results reveal a novel function for MyoD in proliferating myoblasts as a specific transcriptional repressor. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

MATERIALS AND METHODS
Plasmid Constructs-Deletion mutant constructs, pCycB1(Ϫ3800)-CAT, pCycB1(Ϫ1050)-CAT, pCycB1(Ϫ290)-CAT, and pCycB1(Ϫ290 USF Mut)-CAT have been described previously (34). pCycB1(Ϫ318)-CAT and pCycB1(Ϫ169)-CAT were made by deleting fragments between StuI sites at Ϫ318 or Ϫ169 and the polylinker HindIII site of pCycB1(Ϫ1050)-CAT. pCycB1(Ϫ224)-CAT was generated by deleting the PstI-PstI fragment from pCycB1(Ϫ1050)-CAT. pCycB1(Ϫ127)-CAT was created by subcloning the XmnI to XbaI fragment of pCycB1(Ϫ1050)-CAT into SmaI-XbaI sites of pBluescript KS ϩ (Stratagene). Then, the fragment was excised with HindIII and XbaI and cloned into respective sites of pCAT-Basic (Promega). Site mutant construct, pCycB1(Ϫ290 MyoD 5Ј Mut)-CAT was made by engineering a MluI site at the 5Ј MyoD binding sequence. pCycB1(Ϫ290 MyoD 3Ј Mut)-CAT was also made by engineering an EcoRI site at the 3Ј MyoD binding sequence. pCycB1(Ϫ290 MyoD [5Ј ϩ 3Ј] Mut)-CAT was generated by engineering both MluI and EcoRI sites at the 5Ј and 3Ј MyoD binding sites. pRSV-CAT (39)  Transient Expression Assays-C2C12 myoblasts (40) were grown in high mitogen medium (Dulbecco's modified Eagle's medium supplemented with 15% fetal bovine serum (HyClone) and 0.5% chick embryo extract (Life Technologies, Inc.)). C3H10T1/2 fibroblasts were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Cells were grown to 50% confluency and transfected using a standard calcium phosphate precipitate protocol (41). After 18 h with the precipitate, new medium was added, and cells were harvested for CAT activity 16 h later. CAT assays were performed as described (42). Equal protein levels were determined by the Bradford assay. CAT activities were quantitated by phosphorimage analysis of thin layer chromatography plates. Baseline activity of pCAT-Basic was determined from transfections with pCAT-Basic and pEMSV-␣2. One hundred percent activity was determined from transfections with either pCycB1(Ϫ318)-CAT or pCycB1(Ϫ290)-CAT and pEMSV-MyoD. The values were normalized to parallel transient expression assays performed with a Rous sarcoma virus long terminal repeat expression construct (pRSV-CAT).

Repression of the Cyclin B1 Gene Minimal Promoter in C2C12
Myoblasts-The cyclin B1 minimal promoter has recently been identified, and analysis of its sequence has revealed consensus binding sites for several transcriptional regulators upstream of the transcription initiation site (34 1A). A series of deletion mutants of the cyclin B1 promoter (Ϫ3800, Ϫ1050, Ϫ318, Ϫ224, Ϫ169, and Ϫ127; Fig. 1B) were assayed for activity as CAT reporter constructs in proliferating C2C12 myoblasts (the promoter displayed only low level activity in differentiated myotubes; data not shown). Of these constructs, pCycB1(Ϫ224)-CAT displayed the greatest transcriptional activity. The activity of pCycB1(Ϫ169)-CAT and pCycB1(Ϫ127)-CAT was significantly less than that of pCycB1(Ϫ224)-CAT, due to the loss of an essential USF binding site (Fig. 1B; Refs. 34 and 53). Interestingly, retention of the pair of MyoD binding sites in constructs pCycB1(Ϫ318)-CAT, pCycB1(Ϫ1050)-CAT, and pCycB1(Ϫ3800)-CAT resulted in repression of transcription in comparison to pCycB1(Ϫ224)-CAT. Since it is expressed at significant levels in C2C12 myoblasts (8), this result suggested to us that MyoD may mediate transcriptional repression of certain genes in proliferating myoblasts.
Paired MyoD Binding Sites Are Required for Repression of the Cyclin B1 Promoter-We next asked whether the MyoD binding sites specifically mediate repression of the pCycB1-CAT constructs, and whether both sites or a single site are required for this function. Single-base mutants eliminating the left (pCycB1(Ϫ290 5Ј MyoDmut)-CAT), right (pCycB1(Ϫ290 3Ј MyoDmut)-CAT), or both (pCycB1(Ϫ290 [5Ј ϩ 3Ј] MyoDmut)-CAT) MyoD binding sites were tested for expression in C2C12 myoblasts. Expression was compared to that of a wild-type construct (pCycB1(Ϫ290)-CAT) spanning the same promoter region (this construct contains the paired MyoD binding sites; Fig. 2A). Expression of all three mutant reporter constructs greatly exceeded that of the wild-type construct (Fig. 2B). Thus, mutation of either MyoD binding site abolishes repression of the pCycB1-CAT reporters in proliferating myoblasts. This result strongly suggests that binding to these sites is required for repression. Interestingly, disruption of either site blocked repression, indicating that repression requires binding to both sites.
Ectopic Expression of MyoD Mediates Repression of the Cyclin B1 Promoter in C3H10T1/2 Fibroblasts-We next asked whether repression of transcription from the pCycB1-CAT reporters is mediated by MyoD expression. The mouse fibroblast cell line C3H10T1/2 was transfected with pCycB1(Ϫ127)-CAT, pCycB1(Ϫ169)-CAT, pCycB1(Ϫ224)-CAT, and pCycB1(Ϫ318)-CAT either with or without a MyoD expression construct (pEMSV-MyoD). The transient expression assays were performed with C3H10T1/2 cells under mitogenic conditions (subconfluence in high serum medium). Under these conditions, activation of muscle gene transcription by ectopic MyoD is inhibited as it is in myoblasts cultured similarly. Expression of pCycB1(Ϫ127)-CAT, pCycB1(Ϫ169)-CAT, and pCycB1(Ϫ224)-CAT in proliferating C3H10T1/2 fibroblasts was either unaffected or stimulated by ectopic expression of MyoD (Fig. 3). In  (8,37,38), SP1 (45)(46)(47)(48), E2A gene products (36,54), USF (52,53), AP-2 (43,44), and NF-Y (49, 50) are shown. The sizes are determined by their distances from the transcription initiation site, which is indicated with an arrow. B, 10 g of each deletion mutant was transfected as a CAT reporter construct into proliferating C2C12 myoblasts and assayed for transcriptional activity. The highest transcriptional activity was observed for pCycB1(Ϫ224)-CAT, which lacks the MyoD consensus binding sites. CAT activities of each reporter construct were normalized to CAT activities obtained with pRSV-CAT. Mean values and standard error from three independent transfection experiments are shown.
contrast, ectopic expression of MyoD inhibited expression of pCycB1(Ϫ318)-CAT, consistent with the presence of the paired MyoD binding sites on this construct. This result shows that MyoD can directly mediate repression of the minimal cyclin B1 reporter through interaction with the paired binding sites present in pCycB1(Ϫ318)-CAT. Interestingly, repression of pCycB1(Ϫ318)-CAT by ectopic expression of MyoD was not observed in CV-1 cells (data not shown), which are resistant to myogenic conversion (58).
The role of the paired MyoD binding site in mediating repression by MyoD was further investigated by transient expression assays of pCycB1(Ϫ290)-CAT site mutants in prolif-erating C3H10T1/2 cells. Ectopic expression of MyoD repressed pCycB1(Ϫ290)-CAT in proliferating C3H10T1/2 cells only when both binding sites were intact (Fig. 4). Indeed, disruption of the right (3Ј) or both binding sites resulted in stimulation of pCycB1(Ϫ290)-CAT by ectopic expression of MyoD (Fig. 4). These results are consistent with those obtained using the same mutants in C2C12 myoblasts and emphasize the functional importance of both binding sites in mediating repression by MyoD. Disruption of other E-box sequences (including the USF site) in the pCycB1-CAT constructs does not affect repression (data not shown).

DISCUSSION
The possibility that E-box binding proteins may function as transcriptional repressors has been suggested by studies of the immunoglobulin heavy chain (IgH) enhancer. The activity of a bHLH protein (TFE3) that binds a E3 site in the IgH enhancer is repressed by binding of another factor to an adjacent E5 site (59). Displacement of the repressor factor by another E5-binding protein (ITF-1) has been shown to mediate activation of the IgH enhancer (59). The IgH enhancer contains E-box sequences that can bind MyoD; however, binding of MyoD to these sites does activate transcription (36,60). Repression of transcriptional activation of the IgH enhancer by MyoD is also dependent on an adjacent E5 site; removal of this site allows MyoD to activate transcription from the mutant enhancer (61). The protein(s) binding E5 site(s) in the IgH enhancer and repressing transcriptional activation by adjacent E-box binding proteins have not been characterized as yet. Nonetheless, a functional parallel may be drawn to repression of the cyclin B1 promoter by MyoD, which may result from repression of transcriptional activation by USF. In support of this possibility, specific binding of the cyclin B1 promoter by USF has been shown to be required for transcriptional activation (34).
Although ectopic expression of MyoD in fibroblasts can induce autoactivation of the endogenous MyoD gene and activate expression of other myo-bHLH regulators, the molecular basis of these events is not clear. Electrophoretic gel mobility shift assays suggest that MyoD maintains significant sequence-specific DNA binding activity in myoblasts (36,38,62). On the other hand, activation of muscle gene transcription by all of the myo-bHLH regulators is restricted to mitogenically arrested cells. Repression of the transcriptional activity of the myo-bHLH regulators in proliferating cells has been well studied (63), and recent studies of autoactivation of the MyoD gene in myoblasts have supported an indirect mechanism (27). Binding of MyoD to its own promoter is not required for autoactivation in myoblasts, consistent with its lack of transcriptional activity in mitogenically active cells. As shown here, the repressor activity of MyoD is functional in proliferating myoblasts. From this observation, an indirect mechanism of autoactivation can be proposed in which the expression of a repressor of MyoD gene transcription is itself repressed by MyoD.
Understanding the role of MyoD in regulating expression of the cyclin B1 gene during the cell cycle will require further investigation. Similar to activation of the IgH enhancer by TFE3 described above, activation of the cyclin B1 gene by USF and/or downstream elements may require displacement of MyoD by another factor. The activity and/or expression of the displacing factor may be restricted to the G 2 and M phases of the cell cycle, when expression of the cyclin B1 gene is maximal. Recent studies by Hwang et al. (35) have suggested that binding of inhibitory protein(s) to a region Ϫ90 to ϩ1 of the cyclin B1 promoter represses transcription during late G 1 . In addition, their studies show that repression of the cyclin B1 promoter during G 1 is mediated by a region between Ϫ207 and Ϫ872, which we show here contains the binding sites that confer repression by MyoD. It is possible that in other cell types functionally analogous repressor proteins are expressed that bind the same E-box sites as MyoD in the cyclin B1 promoter. Experiments are in progress to test this hypothesis.