Myod Prevents Cyclina/Cdk2 Containing E2F Complexes Formation in Terminally DiErentiated Myocytes

MyoD prevents cyclinA/cdk2 containing E2F complexes formation in terminally dierentiated myocytes

Pier Lorenzo Puri1,3, Clara Balsano2, Vito Lelio Burgio3, Paolo Chirillo1,3, Gioacchino Natoli1,3, Letizia Ricci4, Elisabetta Mattei4, Adolf Graessmann5 and Massimo Levrero1

1Fondazione A Cesalpino and 3I Clinica Medica, University of Rome `La Sapienza', Viale del Policlinico 155, 00161 Roma, Italy; 2Dipartimento di Medicina Interna, University of L'Aquila; 4Istituto di Tecnologie Biomediche CNR, Roma; 5Institute fur Molekularbiologie und Biochemie der Freien Universitat Berlin, 14195 Berlin, Germany

Withdrawal from the cell cycle of dierentiating for myogenic dierentiation is the irreversible with- myocytes is regulated by the myogenic basic helix ± drawal from the cell cycle (Nadal-Ginard, 1978; loop ± helix (bHLH) protein MyoD and the pocket Stockdale and Holtzer, 1961). Master regulators of proteins pRb, p107 and pRb2/p130. Downstream muscle cell commitment and dierentiation are the eectors of `pocket' proteins are the components of the members of the MyoD family of myogenic basic E2F family of transcription factors, which regulate the Helix ± Loop ± Helix (bHLH) proteins (Lassar and

G1/S-phase transition. We analysed by EMSA the Munstemberg, 1994; Olson and Klein, 1994), the composition of E2F complexes in cycling, quiescent retinoblastoma protein (pRb) (Gu et al., 1993) and undierentiated and dierentiated C2C12 skeletal muscle its related proteins p107 and pRb2/p130 (Mayol et al., cells. An E2F complex containing mainly E2F4 and 1993; Schneider et al., 1994; Shin et al., 1995). pRb2/p130 (E2F-G0/G1 complex) appears when DNA In conditions permissive for dierentiation, pRb synthesis arrests, replacing the cyclinA/cdk2 containing binds MyoD via a region overlapping the `pocket'

E2F complex of proliferating myoblasts (E2F-G1/S region (Gu et al., 1993), a conserved domain, also complex). Serum stimulation reinduces DNA synthesis present in the pRb-related proteins p107 and pRb2/ and the re-appearance of E2F-G1/S complexes in p130 and required for binding with viral transforming quiescent myoblasts but not in dierentiated C2C12 proteins and the members of the E2F/DP family myotubes. In dierentiating C2C12 cells, E2F complexes (DeCaprio et al., 1988; Whyte et al., 1989; Dyson et switch and DNA synthesis in response to serum are al., 1990; LaThangue, 1994; Wolf et al., 1995). pRb also prevented when MyoD DNA binding activity and the binds and inactivates the ubiquitously expressed HLH cdks inhibitor MyoD downstream eector p21 are protein Id2 (Iavarone et al., 1994), which upon high induced. Thus, during myogenic dierentiation, forma- mitogens concentration sequestrates the E proteins and tion of E2F4 and pRb2/p130 containing complexes is an inhibits MyoD binding to the E-box (Benezra et al., early event, but not enough on its own to prevent the re- 1990). Like pRb, also p107 interacts with MyoD in a activation of DNA synthesis. Using a subclone of similar fashion, suggesting that redundance might exist C3H10T1/2 mouse ®broblasts stably expressing Estro- to ensure the dierentiation program under appropriate gen Receptor-MyoD (ER-MyoD) chimerae, we found conditions. However, pRb de®cient cells ± where p107 that estrogen directed MyoD activation prevents the re- myogenic functions are recruited, undergo phenotypic association of cyclinA/cdk2 to the E2F4 containing myogenic dierentiation, although retaining their ability complex following serum stimulation and this correlates to re-enter into the cell cycle in response to serum with suppression of E2F activity and the inability of cells stimulation (Schneider et al., 1994). This suggests that to re-enter the cell cycle. Our data indicate that, in pRb is critical to prevent DNA synthesis in differ- dierentiating myocytes, one mechanism through which entiated myotubes in response to mitogens. MyoD induces permanent cell cycle arrest involves p21 pRb's and the other `pocket' proteins role in the upregulation and suppression of the proliferation-asso- permanent exit from the cell cycle and terminal ciated cdks-containing E2F complexes formation. dierentiation is also suggested by the evidence that the products of the viral oncogenes, i.e. polyomavirus Keywords: MyoD; E2F; pRb2/p130; p21; cell cycle large T antigen, adenovirus E1A and SV40 large T antigens are all able to prevent dierentiation of both primary myoblasts and myogenic cell lines (Graess- mann et al., 1973; Webster et al., 1988; Braun et al., Introduction 1992; Maione et al., 1992, 1994; Caruso et al., 1993) and to reinduce DNA synthesis in C2C12 myotubes In vitro skeletal muscle dierentiation consists in (Endo et al., 1986; Cardoso et al., 1993; Crescenzi et mononucleated myoblasts fusion into multinucleated al., 1995). myotubes (phenotypical dierentiation) which express pRb and pRb-related proteins bind and negatively a coordinate set of muscle-speci®c genes (biochemical regulate the activity of E2F (Chellapan et al., 1991; dierentiation) (Olson and Klein, 1994). Prerequisite Hiebert et al., 1992; Weintraub et al., 1992; Cobrinik et al., 1993; Helin et al., 1993a; Zamanian and LaThangue, 1993; Tommasi and Pfeier, 1995). E2F Correspondence: PL Puri is a family of heterodimeric transcription factors made- Received 31 July 1996; revised 8 November 1996; accepted 8 up of E2F-like and DP-like subunits (LaThangue, November 1996 1994). Many genes involved in DNA replication and MyoD-mediated repression of E2Finmyocytes PL Puri et al 1172 cell proliferation that are sequentially activated during Results the cell cycle ± i.e. c-myc,c-myb, cdc2, thymidine kinase (tk), dihydrofolate reductase (DHFR) ± contain E2F complexes in cycling, G /G arrested and E2F binding sites in their regulatory region (LaThan- 0 1 dierentiated C2C12 cells gue, 1994; Nevins, 1992). At least seven genes encoding for components of the E2F family have been cloned C2C12 skeletal muscle cells were used to study the E2F (E2F from 1 to 5, DP1 and DP2) (Helin et al., 1992, complexes composition in dierent culture conditions. 1993b; Kaelin et al., 1992; Shan et al., 1992; Ivey- In high serum (20% FCS) C2C12 myoblasts actively Hoyle et al., 1993; Beijersbergen et al., 1994; Ginsberg proliferate, as demonstrated by high levels of 5-Bromo- et al., 1994; Li et al., 1994; Hijmans et al., 1995; Zhang 2 deoxy-Uridine (BrdU) incorporation (Figure 1A). and Chellapan, 1995). Multiple evidences indicate that Full dierentiation into mature multinucleated myo-

E2F1 plays a critical role in regulating G1/S phase tubes is obtained by culturing con¯uent C2C12 cells in transition (Johnson et al., 1993; Sala et al., 1994; low serum (2% FCS; DM, dierentiation medium). Kowalik et al., 1995) and deregulated expression of C2C12 myotubes do not incorporate BrdU (Figure 1F) E2F1 leads to apoptosis in certain cell types (Shan and and display several muscle-speci®c markers, such as Lee, 1994; Qin et al., 1994). In muscle cells, E2F1 myosin heavy chain (MHC) (Figure 1C). When C2C12 overexpression inhibits myogenic dierentiation (Wang cells are kept in the absence of serum (0.1% FCS) et al., 1995). In association with DP1, E2F1 binds before con¯uence is reached, quiescent undierentiated underphosphorylated pRb and forms a complex which mononucleated myoblasts can be observed between 24 retains its DNA-binding property but is not able to and 48 h after starvation. These quiescent myoblasts activate transcription (Weintraub et al., 1992; Helin et do not incorporate BrdU (Figure 1E) and do not al., 1993a). In many cell types the pRb-E2F complex is express muscle speci®c markers (Figure 1B), although

a minor component of the G0 E2F-DNA binding retaining the ability to proliferate in response to activity (Chittenden et al., 1993; Dou et al., 1994; mitogens (Puri et al., 1995). Schwarz et al., 1993). In unstimulated human T cells For gel retardation assays, cell extracts were and in growth arrested murine ®broblasts an E2F prepared from noncon¯uent C2C12 cells cultured in complex, formed by E2F4 and pRb2/p130, has been 20% serum enriched medium (cycling myoblasts), from detected (Cobrinik et al., 1993; Vairo et al., 1995). This noncon¯uent C2C12 cells cultured in 0.1% serum complex can play a role in inducing and/or maintaining (quiescent myoblasts) and from C2C12 dierentiated cellular quiescence (Claudio et al., 1994; Vairo et al., myotubes. We detected four E2F complexes whose 1995). Functional studies have demonstrated that formation is dierentially regulated along the cell cycle pRb2/p130 suppress E2F4 but not E2F1 mediated and during dierentiation (Figure 2a). The binding trans-activation and that E2F4 coexpression is able to speci®city of each E2F complex was con®rmed by

overcome pRb2/p130 mediated G1 arrest more competition experiments with excess unlabelled wild

eciently than Rb-induced G1 blockade (Vairo et al., type and mutant E2F speci®c oligonucleotides (Figure 1995). E2F4 also forms heterodimers with members of 2b) and unrelated unlabelled DNA (data not shown).

the DP family and p107 (Beijersbergen et al., 1994; The slower migrating band (E2F-G1/S complex), Ginsberg et al., 1994), an association occurring in late detected in cycling myoblasts, disappears when cells

G1 or at the G1/S boundary (Schwarz et al., 1993). In enter G0 (C2C12 quiescent) being replaced by a faster

addition, both p107 and pRb2/p130 have been migrating band (the E2F-G0/G1 complex) also found in described to associate with cyclin E and A and their dierentiated C2C12 myotubes. Two additional faster dependent kinase cdk2 to form E2F-containing migrating bands were observed in all cell cycle phases,

complexes in late G1 and S-phase, respectively but were more evident in cycling myoblasts then in (Devoto et al., 1992; Lees et al., 1992; Pagano et al., quiescent cells. Fully dierentiated myotubes display 1992; Hannon et al., 1993; Li et al., 1993). However, faster migrating bands similar in abundance to those the functional consequences of these multimeric detected in quiescent cells. These E2F complexes were interactions have not been fully de®ned. named E2F-free 1 and 2, based on the results obtained These observations prompted us to analyse the using deoxycolate (DOC) treatment to dissociate E2Fs composition of E2F-associated complexes in C2C12 from their associated proteins (Figure 2c). DOC skeletal muscle cells at critical time points of myogenic abolished in a dose dependent manner both the E2F-

dierentiation. Changes in E2F/DP complexes were G1/S (Figure 2c, lanes 3 and 4) and the E2F-G0/G1

correlated with cell cycle progression, G0 entry, (Figure 2c, lanes 6, 7 and 9 ± 10) complexes, but not the irreversible withdrawal from the cell cycle as well as free E2F forms, which simultaneously increased in with MyoD DNA binding activity and its downstream abundance, indicating the release of free E2F from the target in muscle, the cdks inhibitor p21 (Guo et al., multimeric complexes. The ®fth retarded band (Figure 1995; Halevy et al., 1995). We found that, as cells enter 2a) is nonspeci®c, being abolished by both wild type the quiescent state, an E2F4-pRB2/p130 containing and mutant E2F oligonucleotides (Figure 2b), as well complex appears. The same complex is present in as by unrelated unlabelled DNA (data not shown). dierentiated myotubes. In absence of MyoD activity During the transition from cycling to quiescent C2C12

and p21 up-regulation, the E2F4-pRb2/p130 complex myoblasts, a mixed population of G0/G1 arrested cells is unable to induce the permanent withdrawal from the together with dividing myoblasts can be observed. E2F cell cycle, thus suggesting that its appearance only gel retardation assays, performed with extracts from re¯ects growth arrest, and that MyoD/p21 activity is this mixed-cell population, reveal the coexistence of the

required to prevent re-formation of cyclinA/cdk2 E2F-G0/G1 and E2F-G1/S bands in addition to the E2F containing E2F complexes and S-phase entry in free forms (data not shown). During myogenic dierentiated myocytes upon serum stimulation. dierentiation, cells asyncronously fuse into myo- MyoD-mediated repression of E2F in myocytes PL Puri et al 1173

Figure 1 BrdU and MHC immuno¯uorescence staining of C2C12 cells. Noncon¯uent C2C12 cells growing either in high serum (20% FCS) (a and d) or in absence of serum (0.1% FCS) (b and e) for 24 h and myotubes deriving from con¯uent C2C12 cells cultured in DM for 48 h (c and f) were stained with either anti-MHC antiserum (a, b and c), or with anti-BrdU-antibody (d, e and f) as described in Materials and methods. The inset (e) in panel E displays MHC/DAPI staining to better recognize nuclei from MHC negative myoblasts. Secondary antibodies were coniugated with rhodamine (anti-MHC) and ¯uorescein (anti-BrdU), respectively

tubes. This process starts after around 18 h of culture eect was demonstrable on the other E2F bands in dierentiation medium (early dierentiation) and present in either cycling or quiescent myoblasts as well lasts for up to 96 h (late dierentiation). At this time as in dierentiated myotubes (Figure 3c, lanes 2 and the vast majority of cells are dierentiated myotubes 5). The same E2F-G1/S complex was unaected by an and only very few mononucleated satellite MHC antibody to the G2/M speci®c cyclin B1 (Figure 3a, negative myoblasts are observed (Figure 1C and F). lane 3). A monoclonal anti-p107 antibody also

We monitored E2F complexes along the entire time of supershifted the G1/S speci®c band (Figure 3a, lane dierentiation and we found that either in early or in 4). In addition, an anti-pRb2/p130 monoclonal anti- late dierentiation the electrophoretic mobility of E2F- body was found to supershift the G1/S complex

G0/G1 and free E2F detected bands was identical (data (Figure 3b, lane 4), according to the previously not shown). reported association between pRb2/p130 and the G1/ The composition of the dierent E2F complexes was S speci®c cyclins A and E and their associated kinase determined by gel retardation supershifting assays cdk2 (Cobrinik et al., 1993; Lees et al., 1992). Using a using a panel of speci®c antibodies directed against panel of antibodies against the dierent members of proteins that have been shown to form complexes with the E2F family, we identi®ed E2F4 as the major

E2Fs. Since in many cell types an E2F complex component of the E2F-G1/S complex(es) (Figure 3b, containing E2F-p107-cyclinA-cdk2 appears around the lane 2). Thus, in C2C12 proliferating cells the G1/S time of DNA synthesis, we ®rst tested anti-cyclinA and speci®c E2F complex(es) contains p107 and/or pRb2/ anti-cdk2 antibodies. Polyclonal anti-cdk2 antibody p130, cyclinA, cdk2 and E2F4. E2F4 and its negative supershifted (Figure 3a, lane 2) and a monoclonal anti- regulator pRb2/p130 (Figure 3d, lanes 2 and 4), but cyclinA antibody abolished the E2F-G1/S band present not cyclinA and cdk2 (Figure 3c, lanes 2 and 5), were in cycling C2C12 myoblasts (Figure 3a, lane 5). No also part of the faster migrating complex which, during MyoD-mediated repression of E2Finmyocytes PL Puri et al 1174 a b C2C12 cycling C2C12 myotubes C2C12 cycling C2C12 quiescent C2C12 myotubes – 100X wtE2F 200X wtE2F 200X mtE2F – 100X wtE2F 200X wtE2F 200X mtE2F

E2F-G1/S L L

L E2F-G1/S

E2F-G0/G1 L

L E2F-G0/G1 L L L

E2F-free L

L E2F-free

c C2C12 cycling C2C12 quiescent C2C12 myotubes

DOC DOC DOC

1 2 3 4 5 6 7 8 9 10 Figure 2 E2F EMSAs (a) in cycling, quiescent and dierentiated C2C12 cells. Cell culture, dierentiation conditions and extracts preparation are described in the Results and in Materials and methods. To assess the binding speci®city of the dierent E2F complexes competition experiments (b) were performed by adding 100- to 200-fold molar excess of wild type (wt) and mutant (mt) E2F DNA. Competitor wt consisted of unlabelled E2F sequences from the E2 promoter. Mutant competitor consisted of unlabelled mutated sequences from the adenovirus E2 promoter. The wild type but not the mutant competitor should abolish binding of 32P- labelled E2F speci®c probe. (c) DOC release of E2F activities. DOC treatment and EMSAs from cycling C2C12 (lanes 2 to 4), quiescent C2C12 (lanes 5 to 7) cells and C2C12 myotubes (lanes 8 to 10) were performed as described in Materials and methods. DOC was used at the concentration of 0.4% (lanes 3, 6 and 9) and 0.8% (lanes 4, 7 and 10) (wt/vol)

dierentiation, replaces the E2F-G1/S band, in combination (data not shown), would imply for the

coincidence with DNA synthesis arrest (E2F-G0/G1 presence of other(s) component(s) in this complex. complex). The failure to completely supershift this This is also suggested by the observation that both band with both anti-E2F4 and anti-pRb/p130 anti- SV40 TAg and adenovirus E1A, which are able to bodies addition, either alone (Figure 3d) or in disrupt `pocket proteins'-E2F/DP interactions, reduce MyoD-mediated repression of E2F in myocytes PL Puri et al 1175 a b c d C2C12 cycling C2C12 cycling C2C12 myotubes C2C12 myotubes -cdk2 -cyclin B -p107 -cyclin A -cdk2 -p107 -cyclin B -cyclin A – α α α α – α α α α -E2F4 -E2F1 -pRb2/p130 -E2F4 -E2F1 -pRb2/p130 – α α α – α α α

* * *

1 2 3 4 1 2 3 4 5 1 2 3 4 1 2 3 4 5 e C2C12 cycling C2C12 myotubes – XZ133 G3-245 C36 – XZ133 G3-245 C36

L E2F1

L E2F4

1 2 3

1 2 3 4 5 6 7 8

Figure 3 Antibody shift EMSAs of E2F complexes from C2C12 cycling myoblasts (a and b) and from dierentiated C2C12 myotubes (c and d). In (e): antibody shift EMSA E2F complexes from both C2C12 cycling myoblasts and dierentiated C2C12 myotubes were performed using a panel of antibodies speci®c for pRb. In (f) Immunoblot for E2F1 and E2F4 in cycling myoblasts (lane 1) and in either quiescent myoblasts (lane 2) or myotubes (lane 3) the abundance of this complex without completely 1993). None caused a clear supershift and/or disrup- abolishing it (manuscript in preparation). We also tion of either the G0-speci®c complex or any of the observed that E2F4 is the E2F family member other E2F complexes detected in cycling (Figure 3e) predominantly expressed in both cycling and quiescent and quiescent myoblasts (data not shown) as well as in myoblasts as well as in myotubes, being E2F1 early myotubes (Figure 3e). Moreover, in agreement expression restricted to the G1/S-phase cells (Figure with the observation made by Corbeil et al. (1995), the 3f). To verify the presence of pRb in E2F complexes, XZ anti-pRb monoclonal was found to supershift the four dierent anti-pRb antibodies were used, including E2F-G0/G1 band in 5 days old myotubes (data not the anti-pRb XZ ± which is able to aect the MyoD- shown). A minor supershift of the G0/G1 band in early pRb complex present in C2C12 myotubes (Gu et al., myotubes was also observed with an anti-p107 MyoD-mediated repression of E2Finmyocytes PL Puri et al 1176

a 1 2 3 4 5 6 7 8 9

10 11

Figure 4 E2F EMSAs (a) and DNA synthesis (b) in response to serum. Extracts from cycling myoblasts (lane 1), quiescent myoblasts before (lane 2) and after (lane 3) serum stimulation, semicon¯uent undierentiated myoblasts before (lane 4) and after (lane 5) serum stimulation, early myotubes (cultured 18 h in DM) before (lane 6) and after (lane 7) serum stimulation, and late myotubes (cultured 4 days in DM) before (lane 8) and after (lane 9) serum stimulation were assayed for E2F EMSA as described in Materials and methods. In lanes 10 and 11 extracts were prepared from dierentiated myotubes before and after 24 h serum stimulation, respectively MyoD-mediated repression of E2F in myocytes PL Puri et al 1177 monoclonal antibody (Figure 3c, lane 3). These results a b suggest that E2F4 and pRb2/p130 are the major B components of the E2F complexes found in both G0- κ arrested C2C12 cells and dierentiated myotubes. p107 and pRb can also be found in E2F-G0/G1 complexes as -MyoD minor components, with a mutually exclusive kinetics 1 2 3 – α control 100X E-box 100X NF of appearance during the dierentiation. * L L E-box

Irreversible loss of cyclins/cdks association to E2F-DNA L binding complexes in dierentiating C2C12 cells The interaction between E2F4 and pRb2/p130 has been described to regulate negatively E2F activity and to determine the inhibition of cell cycle progression in certain cell types (Claudio et al., 1994; Vairo et al., 1995). In muscle cells, this interaction has been implicated in the irreversible cell cycle withdrawal

(Shin et al., 1995). Since the E2F-G0/G1 complex appears after only few hours of serum removal, both in con¯uent and noncon¯uent C2C12 cells, well before Figure 5 E-box DNA-binding activity in C2C12 cells during cell phenotypical and biochemical dierentiation is ob- cycle and dierentiation. In (a) extracts from C2C12 cycling cells (lane 1), from C2C12 con¯uent myocytes cultured 18 h in DM served, we examined the behaviour of this complex in (lane 2) and myotubes (cultured 3 days in DM) (lane 3) were used response to serum stimulation in both quiescent and to perform EMSAs as described in Materials and methods. In (b) dierentiated C2C12 cells. As shown in Figure 4a, the binding speci®city of E-box complexes. Competition experiments were performed by adding a 100-fold molar excess of unlabelled E2F-G0/G1 complex (lane 2) disappears upon serum stimulation in C2C12 noncon¯uent cells starved for E-box or unrelated NFkB oligonucleotide. Antibody supershift EMSA of E-box complexes was performed using a polyclonal 24 h in 0.1% FCS, and is replaced by a slower E2F antiserum against MyoD and the corresponding preimmune complex (lane 3) which comigrates with the E2F-G1/S serum as a control complex speci®c of cycling C2C12 myoblasts (lane 1). C2C12 semicon¯uent cells cultured for 12 h in absence of serum display the same behaviour (lanes 4 and 5). In C2C12 con¯uent cells ± cultured in dierentiation FCS 20% 0.1% 2% medium for either 18 (early dierentiation) (lanes 6 and 7) or 96 h (late dierentiation) (lanes 8 and 9) ± serum stimulation is unable to aect the E2F-G0/G1 complex. In addition both quiescent and con¯uent C2C12 cells cultured in dierentiation medium, as well as mature L p21 myotubes, responded to serum with a detectable increase of the slower migrating band of the E2F free species (lanes from 2 to 9, and 10, 11). The ability to 1 2 3 respond to serum stimulation with a switch from the Figure 6 p21 levels in C2C12 cycling myoblasts FCS (lane 1), E2F-G0/G1 complex to the E2F-G1/S complex tempo- quiescent myoblasts (lane 2) and dierentiating myocytes rally correlated with the ability to incorporate BrdU (con¯uent cells cultured 18 h in DM) (lane 3) (Figure 4b). Cells unresponsive to serum were mostly positive for muscle speci®c markers, such as MHC and exibited strong immunoreactivity for MyoD and p21, when compared to BrdU positive myoblasts (data not correlate MyoD activity with E2F complexes changes shown). Although a very small proportion of satellite during dierentiation, we determined MyoD DNA unfused myoblasts, surrounding the myotubes, can binding levels in either cycling or quiescent myoblasts respond to serum with BrdU incorporation (Figure 4b, and in myotubes. EMSA experiments using an E-box lanes 7 and 9), should cyclins/cdks reassociation to the oligonucleotide (Figure 5a) demonstrated that con- E2F complex occurs in these cells it is undetectable by ¯uent C2C12 cells cultured in dierentiation medium EMSA. display increasing amounts of E-box DNA binding activity, starting at 18 h (lane 2), and reaching higher levels in mature myotubes (lane 3). The E-box The role of MyoD in the control of E2F-complexes and complex detected in myotubes includes two bands the irreversible cell cycle exit with dierent electrophoretic mobility. The slower Permanent cell cycle arrest in muscle cells results from migrating band appears in early dierentiation and a combination of events. Among these, MyoD- was speci®cally supershifted by an anti-MyoD anti- dependent p21 upregulation has been proposed as a serum but not by the preimmune serum (Figure 5b). key mechanism to block cdks activity in terminally The faster migrating band appears later during the dierentiated myocytes (Guo et al., 1995; Halevy et dierentiation and is supershifted by an anti-myogenin al., 1995). When culture conditions are permissive for antiserum (data not shown). This increase in MyoD dierentiation, MyoD becomes able to bind the DNA DNA binding activity, temporally correlates with p21 at the E-box sites and to promote muscle speci®c levels upregulation (Figure 6) and the absence of E2F genes transcription (Olson and Klein, 1994). To complexes-switch and BrdU incorporation in response MyoD-mediated repression of E2Finmyocytes PL Puri et al 1178 to serum (Figure 4b). Taken together, our results (Figure 7a), p21 and myogenin transcription (Figure

suggest that the E2F-G0/G1 complex is indeed 7b) and ®nally, in the myogenic conversion of associated with quiescence, but would not be by itself ®broblasts showing high nuclear p21 expression sucient to maintain permanent withdrawal from the (Figure 7c). When MyoD is activated by estradiol, cell cycle. In addition, our data would also imply that cells undergo cell cycle arrest, E2F-driven transcription MyoD, possibly through its downstream target, the is suppressed, as expected (Figure 8a), association of cdks-inhibitor p21, is responsible for preventing re- cyclinA/cdk2 with E2F complexes is prevented (Figure establishment of the proliferation-associated E2F 8b, lane 4), the E2F4-pRb2/p130 complex becomes complex(es). refractory to the addition of serum (Figure 8b, lane 5) Since we failed to supershift E2F complexes with an and cells are unable to re-enter the cell cycle (Figure anti-MyoD antiserum or to detect any E-box binding 8c). The link between MyoD/p21 induction and the activity when extracts from myotubes were anity control of proliferation in muscle cells was further puri®ed with biotinilated E2F oligonucleotides (data reinforced by the observation that MyoD overexpres- not shown), MyoD ability to in¯uence E2F complexes sion in cycling myoblasts, by direct microinjection of a composition would not involve a direct physical MyoD expression vector, results in both the inhibition association at the E2F DNA sites. To further of cell proliferation and the increase of p21 nuclear investigate the role of MyoD activity induction in the expression (Figure 9 and Table 1). modulation of E2F DNA binding complexes composition and activity, we used a subclone of C3H10T1/2 mouse ®broblasts stably expressing Estrogen Receptor- Discussion MyoD (ER-MyoD) chimerae (Hollemberg et al., 1993). In these cells, MyoD and myogenesis become hormone In this study we analysed the composition of E2F inducible. Upon treatment with estradiol ER-MyoD complexes along the cell cycle and during myogenic localizes into the nuclei (Hollemberg et al., 1993) and dierentiation of C2C12 mouse skeletal muscle cells. this results in the induction of E-box binding activity We de®ned two prominent complexes: the ®rst contains

a DM Estradiol – + + hours 48 48 72 10% FCS DM

Estradiol + + – + – myogenin hours 48 48 48 72 72

p21 E-box L

rRNA

1 2 3 4 5

Figure 7 Eect of estradiol treatment on C3H10T1/2 ®broblasts stably expressing Estrogen Receptor-MyoD (ER-MyoD) chimerae. E-box DNA binding activity (a) and p21 and myogenin mRNAs (b) are induced in con¯uent cells cultured for 48 h and 72 h in DM. Nuclear p21 expression (c, panel a) and MHC cytoplasmic expression (c, panel c) were detected in cells after 72 h in DM. In c (panel b) cells are double stained for p21 and MHC MyoD-mediated repression of E2F in myocytes PL Puri et al 1179 a E2F4, p107 and/or pRb2/p130 together with cyclinA and cdk2 and is present in cycling myoblasts; the latter

appears upon serum withdrawal (G0/G1 phase) and contains E2F4 and pRb2/p130, as the major components. In addition, we demonstrate that, when MyoD DNA binding activity and p21 levels increase, the reassociation of cyclins/cdks with E2F is prevented, the E2F4-pRb2/p130 containing complex become refractory to the addition of serum and the cells are unable to re-enter the cell cycle. E2F/DP family has been involved in transcription, cell cycle control and dierentiation (Nevins, 1992; Philpott and Friend, 1994; Sherr, 1994). Recently, Shin and coworkers (1995) have suggested that the accumulation of E2F-pRb2/p130 complexes is directly involved in the induction and in the maintainance of the dierentiative state of myogenic cells. From our results a more complex picture of E2F DNA binding activity during cell cycle and dierentiation of myogenic cells emerges. We found an oscillatory

b pattern of E2F changes from G0 to S-phase in C2C12 cells similar to those present in 3T3 ®broblasts. In both cell types, during proliferation E2F4 associates with the

G1/S speci®c cyclinA and its dependent kinase cdk2 in

20% FCS DM a complex that also contains p107. Since this complex Serum is also aected by anti-pRb2/p130 antibodies, a – + – + (2% ¨ 20%) separate complex containing E2F4, pRb2/130, its associated cyclin(s) and cdk(s) likely exists in C2C12 Estradiol – – – + + proliferating myoblasts. The E2F-G1/S complex dis-

appears as the cells enter the G0/G1 phase, whether C2C12 cells are in a con¯uent or a noncon¯uent state, and is replaced by a faster migrating complex. The major components of this latter complex are E2F4 and pRb2/p130. An E2F-p107 complex could be detected in the early phases of serum starvation-induced quiescence and gradually decreases along the dierentiation

process (PLP, unpublished observations). The E2F-G0/ 1 2 3 4 5 G1 complex persists after phenotypic and biochemical dierentiation into myotubes. In quiescent cells this c complex disappears upon serum addiction, being

replaced by the E2F-G1/S complex, coinciding with DNA synthesis. On the opposite, in myotubes serum fails to reinduce DNA synthesis and to aect the composition of this complex. Although a third faster migrating E2F complex, reported to contain pRb, can be detected in other cell lines (Shirodkar et al., 1992; Schwarz et al., 1993), we were not able to supershift any E2F binding retarded band in extracts from myotubes in the ®rst days of dierentiation medium culture, with the panel of anti-pRb antibodies we used. Starting from the ®fth day of culture in dierentiation

(lane 1) or with (lanes 2 to 4) increasing concentrations of 17-b Estradiol, harvested and then assayed for reporter gene expression, as previously described (Natoli et al., 1994). As a control, cells were also transfected with a similar CAT reporter carrying a deletion of the E2F sites without (lane 5) or with (lane 6) 17-b Estradiol. (b) Estradiol treatment prevents the formation of cyclin-A/cdk2 containing E2F-complexes upon serum stimula- Figure 8 E2F-driven transcription, E2F EMSA and BrdU tion of con¯uent ER-MyoD C3H10T1/2 cells cultured in DM incorporation in ER-MyoD C3H10T1/2 cells. Estradiol treat- (lanes 4 and 5). Cell extracts are prepared as described in the ment decreases E2F-driven transcription (a). Cells were plated in Materials and methods section. (c) Estradiol treatment prevention a medium containing 20% serum and tranfected with the E2F of the cyclin-A/cdk2 containing E2F-complexes formation reporter plasmids using the calcium phosphate method. After a correlates with absence of BrdU incorporation of con¯uent Er- 24 h incubation with the DNA precipitate, cells were washed and MyoD C3H10T1/2 cells stimulated by serum. BrdU incorporation maintained for 72 h in a medium containing 2% serum without was measured as described in Materials and methods MyoD-mediated repression of E2Finmyocytes PL Puri et al 1180 and myotubes, and its levels do not change upon serum stimulation. In contrast, pRb speci®c partner E2F1, is present in proliferating muscle cells and its levels decrease in abundance in quiescent cells and in myotubes. Thus, in muscle cells E2F activity mainly results from the interactions between the p107 and pRb2/p130 pocket proteins and their partner(s). The limited contribution of E2F1 to E2F activity in these cells, would explain the failure to supershift the E2F complexes with antibodies against either E2F1 or its speci®c partner, pRb. However, pRb is easily detectable in E2F complexes from 5 old days myotubes in the absence of a clear supershift with anti-E2F1 antibodies. pRb up-regulation and his role during muscle dierentiation have been previously de®ned (Martelli et al., 1994; Schneider et al., 1994). Our results support the notion that pRb can play a pivotal role in muscle dierentiation independently from its binding to E2F1. E2F free forms were detected in both cycling and quiescent C2C12 cells, although their abundance was clearly reduced in the latter. Dierently from Shin and coworkers (1995), we found low levels of E2F free forms in dierentiated myotubes. The persistence of

low levels of E2F free forms in G0 arrested cells as well as in C2C12 myotubes has been shown by others (Dou et al., 1994; Halevy et al., 1995; Hertzinger et al., 1995; Wolf et al., 1995). The presence of free E2F/DP Figure 9 BrdU incorporation in C3H10T1/2 ®broblasts growing activity is not directly involved in cell cycle progression in 20% FCS containing medium, after microinjection with a because DNA synthesis is almost undetectable in MyoD encoding plasmid (pCMV-MyoD) or the control empty quiescent cells. Free E2F/DP levels increase in vector (pcDNA3). The ability of pCMV-MyoD to encode for response to serum and this correlates with induction MyoD in injected cells was preliminarly controlled by immunostaining using an anti-MyoD polyclonal antiserum. To easily of cell cycle progression into G1/S-phase. Interestingly, identify productively injected cells, the pCMV-MyoD plasmid or also in myotubes the slower migrating E2F/DP free the correspondent control vector, was coinjected together with a band increases in response to serum, even in the pCMV-CD20 expression vector, which encodes for the lympho- absence of DNA synthesis. A possible explanation for cytes dierentiation membrane antigen CD20. After 36 h cells were ®xed and stained for CD20 expression (rhodamine) and this phenomenon relates to the presence of E2F sites BrdU incorporation (¯uorescein). BrdU was added 3 h before required for transcriptional repression of genes, i.e.

®xation cdc2, E2F1, in G0 and early G1 (Lam and Watson, 1993; Adams and Kaelin, 1995). An alternative

explanation could be that E2F free forms might be

wyoh overexpression indues pPI nuler umultion

le I involved in the activation of transcription of genes up-

nd suppression of hxe synthesis in gQrIHIGP ells regulated by growth factors stimulation of terminally

frd pPI dierentiated cells and linked to processes other than

DNA synthesis (i.e. hypertrophy) (Endo et al., 1986; pgwEwyoh WS7 SS7

SS7 VS7 phxeQ Puri et al., 1995).

A critical dierence between quiescent C2C12 gQrIHIGP ells re ultured in PH7 pg supplemented medium

nd miroinjeted with either the pgwEwyoh expression plsmid myoblasts and C2C12 myotubes resides in their ability

por qunti®tion of frd

or the ontrol phxeQ empty vetorF to respond to growth factors stimulation with DNA

inorportionD ells were oEinjeted with either the pgwEwyoh synthesis (hyperplasia) and protein synthesis (hyper- plsmid or the orrespondent ontrolD together with pgwEghPH trophy) respectively (Puri et al., 1995). It is conceivable

expression vetorD whih enodes for the lymphoytes dierentition that at least two separate steps, the G arrest and the memrne ntigen ghPHD s mrker for produtively injeted ellsF 0

efter QT h ells were ®xed nd stined for memrne ghPH irreversible withdrawal from the cell cycle, should exist

expression @rhodmineAD pPI nuler umultion @rhodmineA during myogenic dierentiation. Since the E2F-G0/G1

nd frd inorportion @¯uoreseinAF frd ws dded Q h efore complex appears as the cell enter G0 phase, it might be

®xtionF pigures represent the perentge of pPI or frd positively directly involved in inducing cell cycle arrest. The stined ells

ability of serum to abolish the E2F-G0/G1 complex and

to recruit the E2F-G1/S complex in quiescent myoblasts, but not in myotubes, favours the hypoth- medium, extracts from myotubes can be supershifted esis that additional events play a key role in by both anti-pRb and anti-pRB2/p130 antibodies (data maintainance of this complex in dierentiated cells not shown). These ®ndings ®t well with the E2F even upon serum stimulation, thereby keeping cells in

members expression levels. Indeed, according to the permanent G0 and shifting their response to growth results described by Moberg et al. (1996), we found factors from proliferative to hypertrophic. In agree- that the speci®c p107-pRb2/p130 partner E2F4 is ment with previous studies reporting a critical role of expressed in proliferating as well as arrested myoblasts MyoD/p21 in the induction of permanent cell cycle MyoD-mediated repression of E2F in myocytes PL Puri et al 1181 arrest (Guo et al., 1995; Halevy et al., 1995), we also the association between E2F4 and pRb2/p130 is likely provide evidence that, when MyoD DNA binding to be p53- and MyoD-independent. From our data it activity and p21 levels increase, C2C12 cells lose their appears that the association between E2F4 and pRb2/ ability to respond to serum with DNA synthesis and to p130 re¯ects processes taking place prior to the maintain the E2F-G0/G1 complex after serum stimula- irreversible cell cycle arrest and myogenic differentia- tion. Since MyoD has been described to induce both tion. Indeed, the E2F4-pRb2/p130 complex becomes cell cycle arrest and myogenic conversion in many cell unresponsive to serum only when MyoD activity types (Crescenzi et al., 1990; Sorrentino et al., 1990; increases and p21 upregulation is induced. As cells

Hollemberg et al., 1993), we studied the behaviour of proliferate in response to serum stimulation, G1 cyclins/ E2F complexes in nonmuscle cells ectopically expres- cdks are timely activated to phosphorylate key sing a hormone regulated ER-MyoD functional substrates such as pocket proteins and associate to chimerae. In these cells, when MyoD is activated by E2F4-p107 and E2F4-pRb2/p130 in multimeric com- estrogens, a temporally ordered sequence of events plexes found at the G1/S boundary. More recently, it takes place: E-box binding, myogenin and p21 has been described the ability of p21 to disrupt the expression are induced, E2F4-pRb2/p130 complex interaction between cdk2 and the E2F-pRb2/p130 become refractory to the addition of serum, re-entry complex (Shiyanov et al., 1996). Taken together, the in G1/S-phase in prevented and, ®nally, fusion into available data suggest that MyoD induces the multinucleated myotubes can be observed. irreversible withdrawal from the cell cycle in muscle

The E2F4-pRb2/p130 complex is unable to activate cells by upregulating p21 which, in turn, blocks G1 transcription from promoters containing E2F sites cyclins/cdks activity. This would result in the (Tommasi and Pfeier, 1995). Indeed, we found a accumulation of pRb and its related proteins in their strong reduction of E2F driven reporter gene activity unphosphorylated form and in the prevention of upon MyoD activity induction in C3H10T1/2 MyoD- cyclins/cdks association with E2F4 and its inhibitory ER cells. Taken together, these results support the constraints (pRb2/p130 and p107). The ®nal outcome hypothesis that E2F might be a downstream target of is the permanent arrest of cells in the G0 phase that is MyoD. It would have been interesting to assess reverted by those oncoviral proteins which aect whether MyoD, which is able to bind the E2F MyoD function and/or directly dissociate and E2F/ partners pRB and p107 (Gu et al., 1993), might Rb-related proteins interactions, but not by mitogens. directly in¯uence the E2F activity by physical interaction with components of the multimeric E2F-

G0/G1 complex formed in response to serum withdrawal. The failure to detect any MyoD/E2F com- Materials and methods plexes interaction, would suggest that MyoD does not control E2F-complexes composition by physically Cell lines and culture conditions interacting with E2F/DP proteins. The control of C2C12 mouse cell line (Yae and Saxel, 1977) was MyoD on E2F complexes composition and activity obtained from G Cossu. Cells were grown in DMEM might be explained by the ability of MyoD to prevent supplemented with 20% foetal bovine serum (cycling the association of the predominant component of E2F myoblasts) until they reached 80 ± 90% con¯uence, when complexes in muscle cells, E2F4, and its inhibitory dierentiation was induced by switching cell cultures to DMEM containing 2% foetal bovine serum (DM, partners, pRb2/p130 and p107, with G1 cyclins/cdks, thus hampering the formation of the E2F complexes dierentiation medium). Quiescent C2C12 myoblasts are noncon¯uent cells cultured in 0.1% foetal bovine serum for associated with the G /S phase transition. It has been 1 48 h. They are negative for muscle speci®c markers shown that a key mechanism through which MyoD (Myosin Heavy Chain and Desmin), and for BrdU, but induces cell cycle arrest, and thereby triggers myogenic still able to incorporate BrdU following serum stimulation dierentiation, involves the p53-independent induction (Puri et al., 1995) of the cell cycle inhibitor p21 (Guo et al., 1995; Halevy Stable lines in C3H10T1/2 were generated by co- et al., 1995). p21 prevents the entry into S-phase by transfecting 0.5 mg of the puromycin resistance gene vector pBABEpuro (kindly provided by H Land, ICRF, London, inhibiting G1 cyclins/cdks activity (Sherr, 1994). Cdks recruitment results in phosphorylation of pRb and Rb- UK) together with 5 mg of MyoD-ER construct and 4.5 mgof related proteins (Beijersbergen et al., 1995; Mayol et mouse genomic DNA as a carrier, using the calcium al., 1995; Suzuki-Takahashi et al., 1995; Wolf et al., phosphate precipitation method (Natoli et al., 1994). The hybrid MyoD-ER construct, obtained from the fusion of the 1995; Claudio et al., 1996). In C2C12 cells a direct MyoD expressing vector EMC11S with the hormone-binding correlation between phosphorylation of pRb and the domain of the human estrogen receptor, was supplied by H ability to enter the cell cycle has been described (Gu et Weintraub and previously characterized (Hollemberg et al., al., 1993). Most of pRb is phosphorylated in cycling 1993). Cells were kept in DMEM w : o phenol red and 20% myoblasts, but rapidly becomes unphosphorylated double stripped de®ned supplemented calf serum in the when the cells are induced to dierentiate by growth presence of 2.5 mg/ml puromycin. After 1 week, resistant factors removal. In myotubes pRb is refractory to clones were picked and ampli®ed in duplicate. Clones that rephosphorylation in response to mitogens (Gu et al., exhibited a morphological dierentiation after 48 to 72 h in 1993). Similarly, in L6 myoblasts induced to differ- the presence of estradiol (1077 M) and 2% serum were further entiate, both pRb2/p130 and p107 become underpho- expanded. sphorylated (Kiess et al., 1995). Since we found the Oligonucleotides same E2F/G0/G1 complex in C2C12 quiescent cells, before MyoD/p21 induction, and in C2C12 cells The wt E2F oligonucleotide probe sequences were derived hyperexpressing the wild type as well as the mutant from the adenovirus E2 promoter and consisted of one inactive form of p53 (PLP, unpublished observation), E2F site. The mutant E2F oligonucleotide contains a C-to- MyoD-mediated repression of E2Finmyocytes PL Puri et al 1182 A mutation which determines a strongly reduced E2F from the adenovirus E2 promoter. After 20 min at room binding activity. The E-box oligonucleotide probe sequence temperature, samples were subjected to electrophoresis in derived from the MCK enhancer was 5'-CCCCAA- 4% nondenaturing polyacrylamide gel, as described (Puri CACCTGCTGCCTGA-3'. et al., 1995). For competition studies a 100- to 200-fold molar excess of either unlabelled E2F wild type or E2F mutant oligonucleotide as well as of NFkB control DNA, transfections and CAT assay oligonucleotide, were included in the reaction before ER-MyoD C3H10T1/2 cells were plated in medium speci®c labelled probe was added. DOC released E2F containing 10% serum and transfected by the calcium EMSAs were performed as described previously (Shirodkar phosphate method. After 24 h of incubation with the DNA et al., 1992). For E-box EMSAs, cell extracts were precipitate, cells were washed and maintained for 72 h in prepared as previously described by Gu et al. (1993). medium containing 2% serum with or without 17-b 20 mg of cell extract were combined in 30 mlofbinding Estradiol (Sigma), harvested and assayed for reporter reaction containing 1 mg of poly(dIdC) in 20 mM HEPES gene expression as previously described (Natoli et al., (pH 7.6), 50 mM KCl, 1 mM dithioreitol, 1 mM EDTA, 5% 1994). For these experiments we used a CAT reporter glycerol and 32P-end labelled E-box probe derived from the plasmid containing three E2F sites derived from the c-Myc muscle creatine kinase promoter. P2 promoter (a gift from D Hall). For antibody shift EMSAs antibodies were added to extracts at either room temperature or in ice for 30 min to 1 h (depending on the antibody or antiserum used) before the Northern blot DNA binding reaction was started. Antibodies against E2F1 Total cellular RNA was extracted using the Chomczynski (C20 from Santa Cruz), E2F4 (C-108 from Santa Cruz), pRb and Sacchi method (1987). Ten micrograms of RNA from (XZ133, G3-245, C36 from Pharmigen), p107 (SD9 from undierentiated MyoD-ER C3H10T1/2 cells or from cells Pharmigen), pRb2/p130 (C-20 from Santa Cruz), cyclin A harvested at dierent times after estradiol-induced differ- (BF683 from Pharmigen), cyclin B (from Oncogene Science), entiation were size-fractionated on 1.2 agarose ± 6.7% cdk2 (from UBI) and MyoD (a gift from M Crescenzi and S formaldehyde gels and transferred to nylon ®lters AlemaÁ ) were used. (Quiabrane, Quiagen), as described (Chomczynski and Sacchi, 1987). RNA integrity and quantity was checked by ethidium bromide staining of ribosomal RNA. DNA Immunohistochemistry and DNA synthesis assay 32 fragments, puri®ed on agarose gel, were P-labelled by ForDNAsynthesisassayBrdU(10mM) was added to cells random priming and used to probe the ®lters. The 3 h before cell ®xation. After washing, cells were ®xed in myogenin probe was EcoRI fragment from a pUC19 paraphormaldehyde 1% for 5 min at room temperature, myogenin plasmid. The mouse p21 probe was obtained rinsed in TBS (Tris Buer/NaCl 0.05 M pH 7.6), treated by C Schneider (Trieste, Italy). Hybridization was carried with TRITON-X 100 0.1% in TBS and incubated in 2 N outfor18hat428C in 50% formamide, 56 SSPE (SSPE HCl for 1 h at 378C. After neutralization in 0.1 M borate 16 is 0.18 M sodium chloride, 0.01 M sodium phosphate, buer pH 8.5 cells were washed in TBS and processed for 0.001 EDTA, pH 7.7), 56 Denhardt's solution (16 the immunohistochemical staining with anti-BrdU mono- Denhardt's is 0.02% Ficoll, 0.02 polyvynylpyrrolidone, clonal antibody (Dako) according to the APAAP com- 0.02% BSA), 0.5% SDS, 20 mg of salmon sperm DNA and plexes method (Cordell et al., 1984). Nuclei were 26106 c.p.m. ml71 of 32P-labelled probes. Filters were counterstained with hematoxilin. washed thrice in 0.16 SSC (16 SSC is 150 mM sodiul chloride, 15 mM sodium citrate) 0.1% SDS for 20 min at 658C, and then autoradiographied. Immuno¯uorescence Cells were washed in PBS, ®xed in a methanol/acetone 1 : 2 Western blot v/v solution, dried and incubated with either the anti-MHC Total cell lysates were loaded and then resolved on SDS ± polyclonal antiserum MF20 (a gift from G Cossu) or a PAGE. After overnight electrophoresis, gels were equili- polyclonal anti-p21 antibody (C19 from Santa Cruz) at brated for 30 min in transfer buer (25 mM Tris, 200 mM 378C for 30 min. Speci®cally bound antibody was Glycine) and transferred to PDVF membranes (Millipore) at visualized by incubation with rhodaminated second-step 0.25 Ampere for 5 h at 48C. Membranes were ®rst incubated antibody against immunoglobulin of the relevant species in 16 TBS (20 mM Tris, pH 6.5, 0.5 M NaCl) with 5% BSA and observed by using a ¯uorescent microscope. Immuno- for 1 h, then overnight with a speci®c antibodies ± anti-E2F1 ¯uorescence for the detection of 5-bromo-2'-deoxy-uridine (C20 from Santa Cruz), anti-E2F4 (C-108 from Santa Cruz) (BrdU), as indicator of DNA synthesis, was performed and anti-p21 antibody (C19 from Santa Cruz) at the dilution using the BrdU labelling and detection Kit (Boehringer). of 1 : 1000 in 0.56 TBS with 2.5% BSA and, ®nally, with a For double staining for MHC and BrdU, the anti-MHC secondary antibody conjugated to horseradish peroxidase antiserum MF-20 was used ®rst, before cells were for 1 h. The antigen-antibody interaction was visualized by incubated with the HCl containing buer and a separate incubation for 30 s in a chemilluminant reagent (ECL immuno¯uorescence assay was performed. Western blotting detection from Amersham) and exposure to an X-ray ®lm. Cells microinjection

Electrophoretic mobility shift assay (EMSA) and antibody shift For microinjection experiments, C2C12 cells were grown EMSA on small glass slides subdivided into numbered squares of 2mm62 mm. Cells were microinjected with 100 molecules Electrophoretic mobility shift assays (EMSAs) for E2F of DNA per nucleus, as previously described (Graessmann were performed using 5 mg of whole cell extracts prepared and Graessmann, 1983). After injection cells were cultured as previously described (Pagano et al., 1992), mixed with in 20% FCS enriched medium for additional 36 h and then 2 ml of reaction buer (®nal concentration: NaCl 80 mM, ®xed in a methanol/acetone solution 1 : 2 v/v, dried and HEPES 4 mM,MgCl2mM, glycerol 10%, DTT 0.5 mM, subjected to immunostaining for the dierent proteins EDTA 0.2 mM), 1 ml of poly(dAdT) and a large excess of encoded by the expression plasmids and BrdU incorpora- 32P-end labelled double strand E2F oligonucleotide derived tion. MyoD-mediated repression of E2F in myocytes PL Puri et al 1183 Acknowledgements for helpful discussion and critical reading of the manu- We thank Prof Giulio Cossu for providing C2C12 cells and script. This project was supported by grants from: the the anti-MHC antiserum, Dr Marco Crescenzi and Dr Fondazione A Cesalpino to ML; grant A64 from Telethon; Stefano AlemaÁ for the anti-MyoD antiserum, Dr David the Progetto Finalizzato ACRO, Consiglio Nazionale delle Hall for the E2F-CAT reporter. We also thank Dr Mirella Ricerche to ML. EM is supported by a grant from the Falco, Dr Elisabetta De Marzio, Daniela Collepardo and Associazione Italiana per la Ricerca sul Cancro (AIRC) to Eva Guhl for their excellent technical assistance; Dr Marco Armando Felsani. Crescenzi, Dr Maurizia Caruso and Dr Armando Felsani

References

Adams PD and Kaelin Jr WG. (1995). Cancer Biol., 6, 99 ± GuoK,WangJ,AndreÂs V, Smith R and Walsh K. (1995). 108. Mol. Cell. Biol., 15, 3823 ± 3829. Beijersbergen RL, Kerkhoven RM, Zhu L, Carlee L, Halevy O, Novitch BG, Spicer DB, Skapek SX, Rhee J, Voorhoeve PM and Bernards R. (1994). Genes Dev., 8, Hannon GJ, Beach D and Lassar A. (1995). Science, 267, 2680 ± 2690. 1018 ± 1021. Beijersbergen RL, CarleeÂ L, Kerkhoven RM and Bernards Hannon GJ, Demetrick D and Beach D. (1993). Genes Dev., R. (1995). Genes Dev., 9, 1340 ± 1353. 7, 2378 ± 2391. Benezra R, Davis R, Lockston L, Turner DL and Weintraub Helin K, Lees JA, Vidal M, Dyson N, Harlow E and Fattaey H. (1990). Cell, 61, 49 ± 59. A. (1992). Cell, 70, 337 ± 350. Braun T, Bober E and Arnold HH. (1992). Genes Dev., 6, Helin K, Harlow E and Fattaey A. (1993a). Mol. Cell. Biol., 888 ± 902. 13, 6501 ± 6508. Cardoso MC, Leonhardt H and Nadal-Ginard B. (1993). Helin K, Wu CL, Fattaey A, Lees J, Dynlacht B, Ngwu C Cell, 74, 979 ± 992. and Harlow E. (1993b). Genes Dev., 7, 1850 ± 1861. Caruso M, Martelli F, Giordano A and Felsani A. (1993). Hertzinger T, Wolf DA, Eick D and Kind P. (1995). Oncogene, 8, 267 ± 278. Oncogene, 10, 2079 ± 2084. Chellapan S, Hiebert S, Mudryj M, Horawitz J and Nevins Hiebert S, Chellapan SP, Horowitz J and Nevins JR. (1992). JR. (1991). Cell, 65, 1053 ± 1061. Genes Dev., 6, 177 ± 185. Chittenden T, Livingston D and DeCaprio JA. (1993). Mol. Hijmans EM, Voorhoeve PM, Beijersbergen RL, van't Veer Cell. Biol., 13, 3975 ± 3983. LJ and Bernards R. (1995). Mol. Cell. Biol., 15, 3082 ± Chomczynski P and Sacchi A. (1987). Anal. Biochem., 162, 3089. 156 ± 159. Hollemberg SM, Cheng PF and Weintraub H. (1993). Proc. Claudio PP, Howard CM, Baldi A, De Luca A, Fu Y, Natl. Acad. Sci. USA, 90, 8028 ± 8032. Condorelli G, Sun Y, Colburn N, Calabretta B and Iavarone A, Garg P, Lasorella A, Hsu J and Israel MA. Giordano A. (1994). Cancer Res., 54, 5556 ± 5560. (1994). Genes Dev., 8, 1270 ± 1284. Claudio PP, De Luca A, Howard CM, Baldi A, Firpo EJ, Ivey-Hoyle M, Conroy R, Huber HE, Goodhart PJ, Oli A Ko A, Paggi MG and Giordano A. (1996). Cancer Res., and Heimbrook DC. (1993). Mol. Cell. Biol., 13, 7802 ± 56, 2003 ± 2008. 7812. Cobrinik D, Whyte P, Peeper D, Jacks T and Weinberg R. Johnson DG, Schwartz JK, Cress WD and Nevins JR. (1993). Genes Dev., 7, 2392 ± 2404. (1993). Nature, 365, 349 ± 352. Corbeil HB, Whyte P and Branton PE. (1995). Oncogene, 11, Kaelin Jr WG, Krek W, Sellers WR, DeCaprio JA, 909 ± 920. Ajchambaum F, Fuchs CS, Chittenden T, Li Y, Farnham CordellJL,FaliniB,ErberWN,GhoshAK,AbdulazizZ, PJ, Blanar MA, Livingston DM and Flemington EK. Macdonald S, Pulford KAF, Stein H and Mason DY. (1992). Cell, 70, 351 ± 364. (1984). J. Histochem. Cytochem., 32, 219 ± 229. KiessM,GillMandHamelP.(1995).Cell Growth Dier., 6, Crescenzi M, Fleming TP, Lassar AB, Weintraub H and 1287 ± 1298. Aaronson SA. (1990). Proc.Natl.Acad.Sci.USA,87, Kowalik TF, DeGregori J, Schwarz JK and Nevins JR. 8442 ± 8446. (1995). J. Virol., 69, 2491 ± 2500. Crescenzi M, Soddu S and TatoÁ F. (1995). J. Cell. Phisiol., Lam E W-F and Watson RJ. (1993). EMBO J., 12, 2705 ± 162, 26 ± 35. 2713. DeCaprio JA, Ludlow JA, Figge JW, Shew JY, Huang C-M, LaThangue N. (1994). Trends Biochem. Sci., 19, 108 ± 114. Lee W-H, Marsilio E, Paucha E and Livingston DM. Lassar A and Munsterberg A. (1994). Curr. Biol., 6, 432 ± (1988). Cell, 60, 275 ± 283. 442. Devoto SH, Mudryj M, Pines J, Hunter T and Nevins JR. Lees E, Faha B, Dulic V, Reed S and Harlow E. (1992). Genes (1992). Cell, 68, 167 ± 176. Dev., 6, 1874 ± 1885. Dyson N, Bernards R, Friend SH, Gooding LR, Hasel JA, Li Y, Graham C, Lacy S, Duncan A and Withe P. (1993). Major EO, Pipas JM, Van Dyke T and Harlow E. (1990). Genes Dev., 7, 2366 ± 2377. J. Virol., 64, 1353 ± 1356. Li Y, Slansky JE, Myers DJ, Drinkwater NR, Kaelin WG DouQ-P,ZhaoS,LevinAH,WangJ,HelinKandPardee and Farnham PJ. (1994). Mol. Cell. Biol., 14, 1861 ± 1869. AB. (1994). J. Biol. Chem., 269, 1306 ± 1313. Maione R, Fimia G and Amati P. (1992). Oncogene, 7, 85 ± Endo T and Nadal-Ginard B. (1986). Mol. Cell. Biol., 6, 93. 1412 ± 1416. Maione R, Fimia G, Holman P, Schahausen B and Amati Ginsberg D, Vairo G, Chittenden T, Xiao ZH, Xu G, P. (1994). Cell Growth Dier., 5, 231 ± 237. Wydner KL, De Caprio J, Lawrence J and Livingston D. Mayol X, Grana X, Baldi A, Sang N, Hu Q and Giordano A. (1994). Genes Dev., 8, 2665 ± 2679. (1993). Oncogene, 8, 2561 ± 2566. Graessmann A, Graessmann M and Fogel. (1973). Dev. Mayol X, Garriga J and Grana X. (1995). Oncogene, 11, Biol., 35, 180 ± 186. 801 ± 808. Graessmann M and Graessmann A. (1983). Methods Martelli F, Cenciarelli C, Santarelli G, Polikar B, Felsani A Enzymol., 101, 482 ± 492. and Caruso M. (1994). Oncogene, 9, 3579 ± 3590. Gu W, Schneider JW, Condorelli G, Kaushal S, Mahdavi V Moberg K, Starz MA and Lees JA. (1996). Mol. Cell. Biol., and Nadal-Ginard B. (1993). Cell, 72, 309 ± 324. 16, 1436 ± 1449. MyoD-mediated repression of E2Finmyocytes PL Puri et al 1184 Nadal-Ginard B. (1978). Cell, 15, 855 ± 864. Shirodkar S, Ewen M, DeCaprio J, Morgan D, Livingston Natoli G, Avantaggiati ML, Chirillo P, Puri PL, Ianni A, DM and Chittenden T. (1992). Cell, 68, 157 ± 166. Balsano C and Levrero M. (1994). Oncogene, 9, 2837 ± Sorrentino V, Pepperkok R, Davis RL, Ansorge W and 2843. Philipson L. (1990). Nature 345, 813 ± 815. Nevins JR. (1992). Science, 258, 424 ± 429. Stockdale FE and Holtzer H. (1961). Exp. Cell. Res., 24, Olson EN and Klein WH. (1994). Genes Dev., 8, 1±8. 508 ± 520. Pagano M, Draetta G and Jansen-Durr P. (1992). Science, Suzuki-Takahashi I, Kitagawa M, Saijo M, Higashi H, Ogini 255, 1144 ± 1147. H, Matsumoto H, Taya Y, Nishimura S and Okuyama A. Philpott A and Friend SH. (1994). Mol. Cell. Biol., 14, 5000 ± (1995). Oncogene, 10, 1691 ± 1698. 5009. Tommasi S and Pfeier GP. (1995). Mol. Cell. Biol., 15, Puri PL, Avantaggiati ML, Burgio VL, Chirillo P, 6901 ± 6913. Collepardo D, Natoli G, Balsano C and Levrero M. Wang J, Helin K, Jin P and Nadal Ginard B. (1995). Cell (1995). J. Biol. Chem., 270, 22129 ± 22134. Growth Dier., 6, 1299 ± 1306. Qin X-Q, Livingston DM, Kaelin WG and Adams P. (1994). Vairo G, Livingston DM and Ginsberg D. (1995). Genes Proc. Natl. Acad. Sci. USA, 91, 10918 ± 10922. Dev., 9, 869 ± 881. Sala A, Nicolaides NC, Engelhard A, Bellon T, Lawe DC, Webster KA, Muscat G and Kedes L. (1988). Nature Arnold A, Grana X, Giordano A and Calabretta B. (London), 332, 553 ± 557. (1994). Cancer Res., 54, 1402 ± 1406. Weintraub SJ, Prater CA and Dean DC. (1992). Nature Schneider JW, Gu W, Zhu L, Mahdavy V and Nadal-Ginard (London), 358, 259 ± 261. B. (1994). Science, 264, 1467 ± 1470. Whyte P, Williamson NM and Harlow E. (1989). Cell, 56, Schwarz JK, Devoto SH, Smith EJ, Challapan SP, Jakoi L 67 ± 75. and Nevins JR. (1993). EMBO J., 12, 1013 ± 1020. Wolf DA, Hermeking H, Albert T, Herzinger T, Kind P and ShanB,ZhuX,ChenP-L,DurfeeT,YangY,SharpDand Eick D. (1995). Oncogene, 10, 2067 ± 2078. Lee WH. (1992). Mol. Cell. Biol., 12, 5620 ± 5631. Yae D and Saxel O. (1977). Nature (London), 270, 225 ± Shan B and Lee WH. (1994). Mol. Cell. Biol., 14, 8166 ± 8173. 227. Shiyanov P, Bagchi S, Adami G, Kokontis J, Hay N, Arroyo Zamanian M and LaThangue N. (1993). Mol. Cell. Biol., 4, M, Morozov A and Raychaudhuri P. (1996). Mol. Cell. 389 ± 396. Biol., 16, 737 ± 744. Zhang Y and Chellapan S. (1995). Oncogene, 10, 2085 ± 2093. Sherr CJ. (1994). Cell, 79, 551 ± 555. Shin EK, Shin A, Paulding C, Schaausen B and Yee AS. (1995). Mol. Cell. Biol., 15, 2252 ± 2262.