(Received for publication, October 17, 1996, and in revised form, December 6, 1996)
From the Department of Pathology, Mount Sinai School of Medicine,
New York, New York 10029 and the Department of Cellular
Sciences, Glaxo Wellcome, Inc., Research Triangle Park, North Carolina
27709
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 (2-7). 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 (15-18). 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 proteins2 (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 p34CDC2/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-p34CDC2/CDK1 kinase activity during the G2, G2/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 G2 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.
Deletion mutant constructs,
pCycB1(3800)-CAT, pCycB1(
1050)-CAT, pCycB1(
290)-CAT, and
pCycB1(
290 USF Mut)-CAT have been described previously (34).
pCycB1(
318)-CAT and pCycB1(
169)-CAT were made by deleting fragments
between StuI sites at
318 or
169 and the polylinker
HindIII site of pCycB1(
1050)-CAT. pCycB1(
224)-CAT was
generated by deleting the PstI-PstI fragment from
pCycB1(
1050)-CAT. pCycB1(
127)-CAT was created by subcloning the
XmnI to XbaI fragment of pCycB1(
1050)-CAT into
SmaI-XbaI sites of pBluescript KS+
(Stratagene). Then, the fragment was excised with HindIII
and XbaI and cloned into respective sites of pCAT-Basic
(Promega). Site mutant construct, pCycB1(
290 MyoD 5
Mut)-CAT was
made by engineering a MluI site at the 5
MyoD binding
sequence. pCycB1(
290 MyoD 3
Mut)-CAT was also made by engineering an
EcoRI site at the 3
MyoD binding sequence. pCycB1(
290
MyoD [5
+ 3
] Mut)-CAT was generated by engineering both
MluI and EcoRI sites at the 5
and 3
MyoD
binding sites. pRSV-CAT (39) was given by E. Johnson (Mount Sinai
Medical School). pEMSV-MyoD and pEMSV-
2 were provided by A. Lassar
(Harvard Medical School) and H. Weintraub (Fred Hutchinson Cancer
Research Center).
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(
318)-CAT or pCycB1(
290)-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).
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). Binding sites for AP-2 (181 to
174; Refs. 43 and 44), SP1 (
263 to
258 and
144 to
139; Refs. 45-48), NF-Y (
49 to
42; Refs. 49 and 50), MyoD (
268 to
263 and
236 to
231; Refs. 36 and 51), USF (
190 to
185; Refs. 52 and
53), and E2A gene products (
224 to
218; Refs. 54-57) have been
identified (Fig. 1A). A series of deletion
mutants of the cyclin B1 promoter (
3800,
1050,
318,
224,
169,
and
127; 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(
224)-CAT displayed the greatest
transcriptional activity. The activity of pCycB1(
169)-CAT and
pCycB1(
127)-CAT was significantly less than that of
pCycB1(
224)-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(
318)-CAT,
pCycB1(
1050)-CAT, and pCycB1(
3800)-CAT resulted in repression of
transcription in comparison to pCycB1(
224)-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(290 5
MyoDmut)-CAT), right (pCycB1(
290 3
MyoDmut)-CAT), or both
(pCycB1(
290 [5
+ 3
] MyoDmut)-CAT) MyoD binding sites were tested
for expression in C2C12 myoblasts. Expression was compared to that of a
wild-type construct (pCycB1(
290)-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(127)-CAT, pCycB1(
169)-CAT,
pCycB1(
224)-CAT, and pCycB1(
318)-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(
127)-CAT,
pCycB1(
169)-CAT, and pCycB1(
224)-CAT in proliferating C3H10T1/2
fibroblasts was either unaffected or stimulated by ectopic expression
of MyoD (Fig. 3). In contrast, ectopic expression of
MyoD inhibited expression of pCycB1(
318)-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(
318)-CAT. Interestingly, repression of
pCycB1(
318)-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(290)-CAT site mutants in proliferating C3H10T1/2 cells.
Ectopic expression of MyoD repressed pCycB1(
290)-CAT in proliferating
C3H10T1/2 cells only when both binding sites were intact (Fig.
4). Indeed, disruption of the right (3
) or both binding
sites resulted in stimulation of pCycB1(
290)-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).
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 µE3 site in the IgH enhancer is repressed by binding of another factor to an adjacent µE5 site (59). Displacement of the repressor factor by another µE5-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 µE5 site; removal of this site allows MyoD to activate transcription from the mutant enhancer (61). The protein(s) binding µE5 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
G2 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 90 to +1 of the cyclin B1 promoter represses transcription during late G1. In addition, their studies show that repression of
the cyclin B1 promoter during G1 is mediated by a region
between
207 and
872, 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.