MyoD Can Induce Cell Cycle Arrest but not Muscle Differentiation in the Presence of Dominant Negative SWI/SNF Chromatin Remodeling Enzymes

of dominant negative versions of the mammalian SWI/SNF ATPase subunits, BRG1 and BRM, abolished the ability of MyoD to activate transcription of two muscle specific genes, myogenin and myosin heavy chain. Inhibition of myogenin gene expression was correlated with suppression of chromatin remodeling in the myogenin promoter region (22). To better understand the role of SWI/SNF chromatin remodeling enzymes during muscle differentiation, we looked at the effects of dominant negative BRG1 and BRM on expression of a number of other skeletal muscle markers and cell cycle regulatory proteins. We found that in the absence of functional SWI/SNF, MyoD can promote normal cell cycle regulation but not the expression of skeletal muscle markers. Though MyoD mediated cell cycle arrest is coordinated with activation of muscle specific genes, the two processes are temporally separable. We show here that at least one basis for the distinct mechanisms that regulate cessation of cell proliferation and muscle specific gene expression is that SWI/SNF mediated chromatin remodeling is required specifically for the induction of muscle specific genes.

Skeletal muscle differentiation involves the activation of skeletal muscle genes by the myogenic proteins (23)(24)(25). The myogenic helix loop helix (bHLH) family of transcription factors (MyoD, Myf 5, myogenin, and Mrf 4) heterodimerize with ubiquitously expressed E proteins to bind a consensus E box found in the regulatory regions of many muscle specific genes (26,27). These heterodimers cooperate with members of the myocyte enhancer factor 2 (Mef 2) family to activate myogenesis (28,29). Ectopic expression of the bHLH proteins into non-muscle cells can activate transcription of muscle specific genes (30). Genetic and biochemical evidence indicate that MyoD and Myf 5 establish the myogenic lineage while myogenin promotes terminal differentiation (31)(32)(33)(34)(35)(36). MyoD and Myf 5 are expressed in dividing myoblasts which must exit the cell cycle for muscle specific genes to be activated.
Exit from the cell cycle is accomplished by the down-regulation of cyclins except for cyclin D3, which is upregulated during muscle differentiation and contributes to cell cycle 4 of p21 or p57 is essential for muscle differentiation during embryonic development in mice (42). Ectopic expression of p21 can activate muscle gene expression even in high concentrations of serum (37).
The activity of cell cycle regulatory proteins is important for coordinating myogenesis with arrest in the G0/G1 phase of the cell cycle, however muscle differentiation and cell cycle arrest are not concommitant events. Introduction of MyoD in transformed and tumor cell lines inhibits growth even when muscle specific genes are not activated (43). Similiarly, a MyoD construct in which the basic domain was substituted with the corresponding domain from the E12 transcription factor could inhibit growth but could not activate differentiation (44). Differentiation of C2C12 myoblasts is an ordered process such that expression of myogenin precedes expression of p21, which is then followed by activation of the muscle structural genes with the subsequent appearance of the contractile phenotype and lastly cell fusion. Induction of p21 correlates with the postmitotic state and failure to reinitiate DNA synthesis upon stimulation with growth factors (45).
Expression of Rb is also upregulated by MyoD and is critical for both activation of muscle specific gene expression and cell cycle withdrawal. Rb activates the Mef 2 factors that are necessary for expression of the muscle specific genes, and it prevents cell cycle progression by repressing the E2F family of transcription factors (46). The activity of RB in turn is modulated by the cyclins and their CDK partners. Recent work has demonstrated that the ATPase chromatin remodeling activity of SWI/SNF is required for Rb mediated repression of cyclin A and possibly cyclin E and arrest in the G1 phase of the cell cycle (47,48).
Activation of myogenesis also involves the reorganization of repressive chromatin structure on previously silent muscle specific loci and requires SWI/SNF enzymes. Expression abolished the ability of MyoD to activate transcription of two muscle specific genes, myogenin and myosin heavy chain. Inhibition of myogenin gene expression was correlated with suppression of chromatin remodeling in the myogenin promoter region (22). To better understand the role of SWI/SNF chromatin remodeling enzymes during muscle differentiation, we looked at the effects of dominant negative BRG1 and BRM on expression of a number of other skeletal muscle markers and cell cycle regulatory proteins. We found that in the absence of functional SWI/SNF, MyoD can promote normal cell cycle regulation but not the expression of skeletal muscle markers. Though MyoD mediated cell cycle arrest is coordinated with activation of muscle specific genes, the two processes are temporally separable. We show here that at least one basis for the distinct mechanisms that regulate cessation of cell proliferation and muscle specific gene expression is that SWI/SNF mediated chromatin remodeling is required specifically for the induction of muscle specific genes. 6

Cell Culture
Cells were grown as previously described (14). Dominant negative BRG1 and BRM expression were induced by passing cells in media lacking tetracycline and differentiated as described (22).

Protein Extracts and Western Analysis
Isolation of proteins and western blotting were as described (14). Antibodies against the flag epitope, p27, cyclin A, cyclin E (M20), RB (C15), and rel B were from Santa Cruz. The anticyclin D1 and D3 antibodies were from Pharmingen/Signal Transduction Laboratories.

Northern Analysis
Total cellular RNA was isolated by Trizol (Invitrogen) as described by the manufacturer. For Northern analysis, 10 to 20 µg of total cellular RNA was electrophoresed on a 1% formaldehyde gel and transferred to Nytran Plus (Schleicher and Schuell). The membranes were baked for 2 hours and prehybridized for at least 2 hours at 42 0 C in 6X SSC, 50%formamide, 50mM NaPO 4 , pH 6.8, 5X Denhardt's solution, 0,.5% SDS, and 100ug/ml salmon sperm DNA (not boiled). The buffer was then changed to 6X SSC, 50% formamide, 50 mM NaPO4, pH 6.8, 1X Denhardt's solution, 0.1% SDS, and 100ug/ml salmon sperm (which was first boiled with a 32 P labeled probe generated by random priming) and membranes were hybridized overnight at 42 0 C. Probes for myogenin and myosin heavy chain were described (22). A 1.7Kb Xho I restriction fragment was used to detect p21, a 0.7Kb EcoR I restriction by guest on March 24, 2020 http://www.jbc.org/ Downloaded from 7 fragment was used to detect myosin light chain, a 0.9Kb Pst I restriction fragment was used to detect troponin T, and a 700 bp EcoR I-Hind II, restriction fragment was used to detect GAPDH. Blots were washed sequentially in 2X SSC, 0.1%SDS at room temperature, then in 0.2X SSC, 0.1% SDS at 42 o C, and were exposed to a PhosphorImager screen and analyzed with ImageQuant.
The gels were dried and exposed to a PhosphorImager screen. 9

Transcription of muscle specific genes is inhibited by dominant negative BRM and BRG1
Mammalian SWI/SNF complexes contain either the Brm or Brg1 ATPase subunit.
When overexpressed in cells, BRM or BRG1 with mutations in the ATPase domains act as dominant negatives and have been shown to inhibit gene activation events that normally require SWI/SNF function. Stable cell lines in which the genes encoding flag tagged dominant negative BRM or BRG1 placed under control of the Tet-VP16 activator were previously described (14). These cells were grown in the presence or absence of tetracycline for 96 hours, then infected with a retrovirus containing the MyoD gene and forced to differentiate in low serum conditions. In the presence of tetracycline, when dominant negative BRM or BRG1 is not expressed, the expression of two muscle specific genes, myosin heavy chain and myogenin, was activated. However, when these cells were differentiated in the absence of tetracycline, dominant negative BRM and BRG1 inhibited activation of these muscle specific markers (22). This indicated that SWI/SNF complexes play an important role in muscle differentiation. To determine how extensively SWI/SNF is required for muscle differentiation, we looked at the effect of dominant negative BRG1 and BRM expression on a number of muscle specific genes. inducibly express dominant negative hBRM, and B22 cells, which inducibly express dominant negative BRG1. We observed that myosin heavy chain, myosin light chain and troponin T gene expression was induced by MyoD when cells were grown and differentiated in the presence of tetracycline. When cells were grown and differentiated in the absence of tetracycline, dominant negative protein expression inhibited activation of these genes. 10 Activation of these muscle specific markers was not inhibited when a control Tet-VP16 cell line, which expresses only the Tet-VP16 activator, was grown and differentiated in the absence of tetracycline. Interestingly, expression of the glyceraldehyde-3-phosphate dehydrogenase gene (GAPDH) was upregulated during myogenesis but was not affected by dominant negative BRM and BRG1. RT-PCR of two additional muscle specific markers shows that α-skeletal actin and desmin expression is also inhibited by dominant negative BRM and BRG1 (Fig. 1B).
The levels of flag tagged dominant negative BRM and BRG1 expression in this experiment were detected by Western analysis (Fig. 1C).
Muscle differentiation involves the activity of the basic helix loop helix family of transcription factors, which heterodimerize with E proteins to bind E boxes on the promoters of muscle specific genes, and the Mef2 family of transcription factors, which bind AT rich sequences on these promoters. We previously showed that myogenin expression is induced by MyoD and inhibited by expression of dominant negative BRM and BRG1 in H17 and B22 cell lines respectively ( Fig. 2A and (22) (Fig. 2B). Taken together, these results indicate that SWI/SNF plays an extensive role in MyoD mediated myogenesis and is critical for the expression of muscle specific regulatory and structural proteins.

Cell cycle arrest occurs in the presence of dominant negative BRM and BRG1
Terminal cell cycle arrest is closely coupled to muscle differentiation and is required for These data differ from other reports that have shown that SWI/SNF complexes were required for pRb mediated cell cycle arrest (47,48). In those studies, proliferating cells expressing alleles of pRb that could not be phosphorylated, and thus are constitutively active, did not arrest unless functional BRG1 was present. Similar results were obtained when the cyclin dependent kinase (CDK) inhibitor p16/ink4a was expressed. In addition, proliferating cells expressing dominant negative BRG1 did not arrest when treated with cisplatin, a DNA damaging agent, whereas cisplatin treatment of proliferating cells grown in the presence of 12 tetracycline (dominant negative off) did undergo cell cycle arrest. Clearly, the experiments presented in Figure 3 differ methodologically from those described in earlier reports. We suggest that while BRG1 and SWI/SNF complexes are required for cell cycle arrest under some conditions, they are not required universally. To further address the role of SWI/SNF complexes in MyoD mediated cell cycle withdrawal and muscle differentiation, we looked more closely at whether dominant negative BRM and BRG1 affect the levels of cell cycle regulatory proteins.

p21, Cyclin and Rb expression are unaffected by dominant negative BRM and BRG1
One class of cell cycle regulators, the CDK inhibitors, is generally upregulated during cell differentiation and may play a general role in promoting and maintaining the terminally differentiated phenotype by binding to and inactivating cyclin-CDK complexes (reviewed in (53)). The CIP1/KIP1 family includes p21, p57, and p27, all of which have been implicated in cell cycle withdrawal during muscle differentiation. p21 and p57 mRNAs are both induced during differentiation in C2C12 cells and the presence of one of the two genes is required for myogenesis in mice (42,54,55). We observed that MyoD strongly induced p21 expression and that dominant negative BRM or BRG1 had no effect on its expression (Fig. 4A). RT-PCR showed that p57 mRNA was not expressed in our cells (Fig. 4B). This is consistent with previous reports that showed p57 is not activated by MyoD in 10T1/2 fibroblasts (54). p27 protein levels were similar in mock and MyoD differentiated cells and also were unaffected by expression of dominant negative BRM or BRG1 (Fig. 4C). High levels of p27 have been Cyclin levels are generally downregulated during differentiation. One exception is cyclin D3, which was previously shown to be transcriptionally upregulated during muscle differentiation and which contributes to irreversible withdrawal from the cell cycle by forming inactive complexes with unphosphorylated Rb, cdk4, p21, and PCNA (38). We therefore looked at whether dominant negative BRM and BRG1 could inhibit induction of cyclin D3 during muscle differentiation by MyoD. Fig. 5A shows that cyclin D3 protein levels were upregulated by MyoD during differentiation in the presence or absence of tetracycline. Cyclin E levels during differentiation can vary depending on cell type and the presence of extracellular factors (reviewed in (57)). In differentiated C2C12 cells, cyclin E was reported both to remain constant and to decline and to associate with p21 to form inactive complexes during differentiation (39,58,59). In our experiments, cyclin E protein levels were higher in MyoD differentiated cells than in mock differentiated or proliferating cells and were unaffected by expression of dominant negative BRM and BRG1 (Fig. 5A). The increase in cyclin E levels may reflect increased protein stability due to the increase in p21 levels. The hypophosphorylated state of Rb in these cells (Fig. 5C) suggests the cyclin E is inactive.
Our results indicate that SWI/SNF complexes are not required for the normal expression of the CIP1/KIP1 CDK inhibitors or cyclins D3 or E under these differentiation conditions.  (57)). Cyclin D1 was present at low to undetectable levels in both mock and MyoD differentiated cells in the presence or absence of dominant negative BRM or BRG1 expression. The levels in these growth arrested cells is contrasted to the higher levels present in proliferating cells (Fig. 5A).
Cyclin A is also downregulated when cells exit the cell cycle and its forced expression results in phosphorylation of pRb and in inhibition of muscle differentiation (60). Fig. 5A shows that cyclin A levels were low during growth arrest, even in the presence of the dominant negative proteins. Rb mediated down regulation of cyclin A during cell cycle arrest requires BRG1 (47,48). However, our data demonstrates that in both the mock differentiated as well as the MyoD differentiated cells, cyclin A is downregulated even in the presence of non-functional SWI/SNF enzymes. As discussed above, this suggests that the requirement for SWI/SNF complexes during cell cycle arrest induced by different experimental conditions is not absolute.
Instead, the multiple signals received by confluent cells placed under low serum conditions may overcome or bypass the need for SWI/SNF enzymes during exit from the cell cycle. In conclusion, it appears that there are likely to be multiple mechanisms for achieving cell cycle withdrawal, some of which do not require SWI/SNF remodeling enzymes.
Rb has been shown to interact with MyoD and to promote muscle gene activation and cell cycle arrest (61). Rb has also been demonstrated to interact with BRM and BRG1 and has by guest on March 24, 2020 http://www.jbc.org/ Downloaded from 15 been implicated in both cell cycle arrest and induction of gene expression by the glucocorticoid receptor (7,47,48,62,63). In C2C12 cells, MyoD upregulates Rb gene expression (64). The phosphorylation state of Rb is also regulated during the cell cycle and during differentiation.
Cells in G1 contain mostly the hypo-phosphorylated form while cells approaching S phase accumulate the hyperphosphorylated form of Rb. Cells lacking Rb are competent for induction of early differentiation markers and for cell cycle arrest but fail to induce late muscle specific marker genes (65). We therefore examined Rb mRNA and protein levels in MyoD differentiated cells. Fig. 5B shows that Rb mRNA levels were induced in the MyoD differentiated cells and that dominant negative BRM and BRG1 did not inhibit activation. Fig.   5C shows that the corresponding increase in pRb protein levels and the phosphorylation state of Rb in MyoD differentiated cells also were not affected by dominant negative BRM or BRG1.
Withdrawal from the cell cycle during myogenesis is intricately coordinated with activation of muscle specific gene expression, yet the two events are separable. Here, we demonstrate that functional SWI/SNF chromatin remodeling complexes are required for the induction of muscle specific regulators and structural genes but are not required to induce the expression of a number of cell cycle regulators associated with cell cycle arrest. Thus cells expressing non-functional SWI/SNF complexes can withdraw from the cell cycle upon MyoD mediated muscle differentiation.
Existing data suggests that muscle regulatory factors and cell cycle regulators cooperate to promote muscle differentiation. In most models, MyoD and/or Myf5 stimulate the expression of "early" differentiation markers such as myogenin, MEF2 and P21. This step in the differentiation process promotes cell cycle withdrawal and can occur in the absence of pRb (65). However, additional functions mediated by pRb, including activating the transcriptional 16 competence of the MEF2 activators, help promote the induction of "late" differentiation markers (46). The involvement of SWI/SNF chromatin remodeling enzymes during MyoD mediated muscle differentiation does not appear specific for either the "early" or "late" differentiation events. Instead, we suggest that the requirement for SWI/SNF complexes is likely traced to its enzymatic properties that enable it to alter chromatin structure. Genes that are off or that require remodeling of chromatin at or near promoter sequences to be upregulated will be dependent on SWI/SNF enzymes for expression during muscle differentiation. In contrast, genes that are upregulated without requiring ATP dependent structural changes in chromatin at or near promoter sequences will be independent of the function of SWI/SNF and related enzymes. According to this hypothesis, we would predict that genes that are off prior to muscle differentiation, such as myogenin, muscle heavy and light chains, troponin T, desmin and α-actin, would require SWI/SNF mediated chromatin remodeling to induce gene expression. In fact, we have previously shown that failure to induce myogenin expression in the absence of functional SWI/SNF complexes correlates with inhibition of chromatin remodeling at the endogenous myogenin promoter (22). Additionally, we would predict that a gene such as Mef2C, which is on in undifferentiated and mock differentiated cells but is upregulated in MyoD differentiated cells, would also require SWI/SNF mediated remodeling of its promoter structure in order for expression to be induced. In contrast, our hypothesis predicts that upregulation of p21, cyclin D3, and pRb does not involve SWI/SNF mediated changes in promoter structure. A schematic model indicating the SWI/SNF dependency of different steps leading to muscle differentiation is presented in Figure 6. Analysis of the endogenous promoter structure of these and other genes, before and after differentiation and in the presence and absence of functional SWI/SNF enzymes, should permit us to directly test our hypothesis.