Oncogene (2001) 20, 7998 ± 8008 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Opposing functions of ATF2 and Fos-like transcription factors in c-Jun-mediated myogenin expression and terminal di€erentiation of avian myoblasts

Laetitia Daury1, Muriel Busson1, NikolaõÈ Tourkine1, FrancËois Casas1, Isabelle Cassar-Malek1, Chantal Wrutniak-Cabello1, Marc Castellazzi2 and Ge rard Cabello*,1

1Unite d'Endocrinologie Cellulaire, UMR Di€eÂrenciation Cellulaire et Croissance, Institut National de la Recherche Agronomique (INRA), 2 place Viala, 34060 Montpellier Cedex 1, France; 2Unite de Virologie Humaine, Institut National de la Sante et de la Recherche MeÂdicale (INSERM U412), Ecole Normale SupeÂrieure, 46 alleÂe d'Italie, 69364 Lyon Cedex 07, France

With the aim to identify the oncoprotein partners and MRF4 appear to be positive e€ectors of major implicated in the c-Jun myogenic in¯uence, we carried importance. These proteins belong to a bHLH out stable transfection experiments of c-Jun and/or ATF2, transcription factors family, and activate transcription Fra2, c-Fos overexpression in avian myoblasts. Before of genes speci®cally expressed in muscle tissue. In vivo induction of di€erentiation, c-Jun repressed myoblast experiments using disruption of genes encoding these withdrawal from the cell cycle, as did a TPA treatment. myogenic factors in mice have led to the conclusion However, after serum removal, unlike TPA, c-Jun that whereas Myf5 and MyoD are more involved in signi®cantly stimulated myoblast di€erentiation. In search cell commitment to the myogenic lineage, Myogenin for speci®c partners involved in this dual in¯uence, we plays a key role in induction of terminal di€erentiation found that a reduction in the amounts of c-Fos and Fra2 (Braun et al., 1992; Rudnicki et al., 1992; Hasty et al., and an increase in c-Jun proteins occurred at cell 1993; Nabeshima et al., 1993). con¯uence, a situation likely to favor cooperation between A number of negative regulators of myoblast c-Jun and ATF2 during terminal di€erentiation. Whereas di€erentiation have progressively been described, such c-Fos and Fra2 cooperated with c-Jun to abrogate as Id, Twist, ZEB, Mist1 or MyoR, acting principally myoblast withdrawal from the cell cycle and terminal by interfering with the transcriptional activity of di€erentiation, ATF2 co-expression potentiated the posi- myogenic factors (Benezra et al., 1990; Hebrok et al., tive myogenic c-Jun in¯uence. In addition, myogenin 1997; Postigo and Dean, 1997; Lemercier et al., 1998; expression was a positive target of this cooperation and Lu et al., 1999). Particular attention was also paid to this regulation occurred through a stimulation of myogenin cellular oncogenes, such as c-Src, c-Ras or c-Fos which promoter activity: (1) whereas c-Fos or Fra2 co-expression prevent myoblast di€erentiation by inhibiting myogenic abrogated c-Jun stimulatory activity on this promoter, factors expression and/or activity (Falcone et al., 1991; ATF2 co-expression potentiated this in¯uence; (2) using a Lassar et al., 1989; Pedraza-Alva et al., 1994). dominant negative ATF2 mutant, we established that Amongst the oncogenes, since the initial study of Su c-Jun transcriptional activity required functionality of et al. (1991), c-Jun is also considered as a major endogenous ATF2. These data suggest that through this myogenic repressor. However, the molecular basis of dual myogenic in¯uence due to cooperations with di€erent this in¯uence is not fully understood. In particular, it is partners, c-Jun is involved in the control of duration of well established that c-Jun activity involves cooperation myoblast proliferation and thereafter of fusion eciency. with partners interacting in a range of AP-1 complexes Oncogene (2001) 20, 7998 ± 8008. including Fos-like and ATF2 proteins. The involve- ment of such cooperations in myogenesis regulation is Keywords: Jun; Fos; ATF; myoblasts; di€erentiation; suggested by previous studies indicating that stimula- Myogenin tion of Fos-like and c-Jun proteins expression by Tetradecanoyl Phorbol Acetate (TPA) or okadaic acid fully abrogates myoblast di€erentiation (Sulakhe et al., Introduction 1985; Park et al., 1992). The interest in identi®cation of such c-Jun partners involved in this myogenic in¯uence The regulation of myoblast di€erentiation involves has been strengthened by the observation that, complex interactions between positive and negative depending on the nature of this cooperation, di€erent e€ectors. Myogenic factors Myf5, MyoD, Myogenin genetic programs could be induced. In particular, studies concerning the in¯uence of c-Jun on chicken embryonic ®broblast transformation clearly established that anchorage-independent growth needs cooperation *Correspondence: G Cabello; E-mail: [email protected] Received 6 July 2001; revised 24 August 2001; accepted 13 between c-Jun and Fos-like proteins, whereas serum- September 2001 independent growth depends on a c-Jun/ATF2 co- Regulation of myoblast differentiation by c-Jun L Daury et al 7999 operation (van Dam et al., 1998; Huguier et al., 1998; during di€erentiation, the c-Jun level increased at cell van Dam and Castellazzi, 2001). con¯uence. No signi®cant change was detected for the Taking all these bibliographic data into account, the ATF2 level (Figure 1). major aim of this study was to identify the partners These data pointed out a possible involvement in involved in the myogenic activity of c-Jun, using myogenesis regulation of the rise in c-Jun expression constitutive overexpression of c-Jun on the one hand, occurring just before induction of terminal di€erentia- and/or c-Fos, Fra2 or ATF2 on the other, in avian tion. This led us to study the in¯uence of c-Jun myoblasts. overexpression on myoblast withdrawal from the cell cycle and on myotube formation.

Results Dual influence of c-Jun overexpression on avian myoblast differentiation Changes in c-Jun, c-Fos, Fra2 and ATF2 protein levels This study was performed by stably overexpressing during avian myoblast differentiation wild-type c-Jun or a mutant deleted of the major part This work was performed using the quail myoblast cell of the transactivation domain (c-Jun D176) in QM7 line QM7, established by Antin and Ordahl (1991). myoblasts. Control cells were obtained after cotrans- This in vitro system reproduces all major in vivo events fection of pPD18 empty vector and pSV2-neoR occurring during muscle di€erentiation: myoblast plasmid, and selection of neomycine-resistant cells. proliferation, withdrawal from the cell cycle at cell Experiments were carried out on polyclonal popula- con¯uence characterized by an early synthesis of post- tions of transfected cells. c-Jun and c-Jun D176 mitotic markers including connectin (Hill et al., 1986) overexpression was assessed in Western blot experi- and, after lowering serum level in the culture medium, ments (Figure 2). terminal di€erentiation characterized by myoblast Two sets of data were recorded. First, we studied fusion and expression of numerous muscle-speci®c myoblast withdrawal from the cell cycle at cell proteins. con¯uence, just before serum removal, by assessing In an initial step, using Western blot experiments, we the number of cells expressing connectin, a post-mitotic characterized our myogenic system by assessing the marker. Second, the extent of myoblast di€erentiation spontaneous changes occurring in the levels of c-Jun, c- was studied using morphological criteria in cytoimmu- Fos, Fra2 and ATF2 proteins during critical stages of no¯uorescence experiments and calculation of the QM7 myoblast di€erentiation: proliferation, cell con- fusion index (FI). ¯uence and terminal di€erentiation. Whereas c-Fos and At cell con¯uence, c-Jun overexpression signi®cantly Fra2 protein levels decreased at cell con¯uence and decreased the number of cells expressing connectin,

Figure 1 Changes in the expression of AP-1 family members during avian myoblast di€erentiation. One hundred mg of proteins extracted from QM7 myoblasts during proliferation, at cell con¯uence or 72 h after the induction of di€erentiation were analysed by Western blot using speci®c antibodies raised against avian c-Fos, Fra2, c-Jun and ATF2 proteins. Quanti®cation was performed using a phosphorimager (Molecular Dynamics). Results are expressed as percentages of the value recorded in proliferating myoblasts. Data are the mean+s.e.m. of four separate experiments. A typical Western blot is presented

Oncogene Regulation of myoblast differentiation by c-Jun L Daury et al 8000

Figure 2 Stable overexpression of the di€erent AP-1 family members in avian myoblasts. One hundred mg of proteins extracted from pool of clones of QM7 myoblasts overexpressing c-Jun D176, c-Jun and/or ATF2, ATF2 D276, c-Fos or Fra2 were analysed by Western blot using speci®c antibodies raised against avian c-Fos, Fra2, c-Jun and ATF2 proteins. Quanti®cation was performed using a phosphorimager (Molecular Dynamics). Results are expressed as percentages of the value recorded in control myoblasts; for protein mutants, the results are expressed as percentages of the value recorded for the wild-type protein in control myoblasts. Data are the mean+s.e.m. of three separate experiments carried out on three di€erent pools of clones; they represent the average overexpression level obtained in the di€erent clones. A typical Western blot is presented

relative to control cultures (1.23+0.09% vs Unexpectedly, after induction of myoblast di€eren- 2.99+0.31%, P50.01; Figure 3b,a), thus indicating a tiation, c-Jun overexpression signi®cantly increased the signi®cant inhibition of myoblast withdrawal from the extent of myoblast fusion (FI: 21.05+1.12% vs cell cycle. This in¯uence was not recorded after 16.46+1.29%, P50.005; Figure 3e,d). As observed overexpression of the c-Jun D176 mutant for myoblast withdrawal from the cell cycle, deletion of (3.13+0.29%; Figure 3c), thus establishing that the the c-Jun transactivation domain (c-Jun D176) abro- transactivation domain of the oncoprotein was directly gated this myogenic in¯uence (FI: 15.20+1.87%; involved in the repression of myoblast withdrawal from Figure 3f), again underlining the involvement of the the cell cycle. transactivation domain of the oncoprotein.

Oncogene Regulation of myoblast differentiation by c-Jun L Daury et al 8001

Figure 3 In¯uence of c-Jun and c-Jun D176 mutant on avian myoblast withdrawal from the cell cycle and terminal di€erentiation. Cytoimmuno¯uorescence staining with an antibody raised against connectin of control or QM7 myoblasts overexpressing c-Jun or c-Jun D176. (a,b,c): staining at cell con¯uence. (d,e,f): 72 h after induction of di€erentiation. (a,d): control cells. (b,e): c-Jun overexpression. (c,f): c-Jun D176 overexpression. These results are representative of three separate experiments (6100)

Figure 4 In¯uence of expression or co-expression of AP-1 family members on avian myoblast withdrawal from the cell cycle. Cytoimmuno¯uorescence staining with an antibody raised against connectin, at cell con¯uence. (a): control cells. (b): c-Jun overexpression. (c): ATF2 overexpression. (d): Fra2 overexpression. (e): c-Fos overexpression. (f): TPA treatment. (g): c-Jun+ATF2 co-expression. (h): c-Jun+Fra2 co-expression. (i): c-Jun+c-Fos co-expression. These results are representative of three separate experiments (6100)

These data establish a surprising dual myogenic previously shown that signi®cant changes in the activity of c-Jun, clearly dependent on the induction relative amounts of di€erent c-Jun partners occurred of di€erentiation, at cell con¯uence. As we have at this stage of the culture, we assessed the

Oncogene Regulation of myoblast differentiation by c-Jun L Daury et al 8002 possibility that these opposite in¯uences could result 1.13+0.25%, P50.001). In agreement with previous from related changes in the equilibrium of c-Jun data (Sulakhe et al., 1985), we simultaneously cooperations with Fos-like or ATF2 proteins. observed that addition of TPA to the culture medium induced a potent block to myoblast withdrawal from the cell cycle (0.09+0.03%, P50.001; Figure 4f). By Opposite influences of Fos-like or ATF2 overexpression contrast, co-expression of c-Jun and ATF2 increased on c-Jun myogenic influence the number of post-mitotic myoblasts relative to To study the in¯uence of Fos-like or ATF2 coopera- control cultures (4.50+0.42% vs 2.99+0.31%, tion with c-Jun on the myogenic in¯uence of the P50.05; Figure 4g) and some small myotubes were oncoprotein, we stably overexpressed c-Jun and/or generally observed despite the presence of high serum ATF2, Fra2, c-Fos, in QM7 myoblasts (overexpression concentration in the culture medium. All these results was assessed in Western blot experiments; Figure 2). A already suggest that the inhibition of myoblast TPA treatment, increasing c-Jun/c-Fos and c-Jun/Fra2 withdrawal from the cell cycle induced by c-Jun was activity (Greenberg and Zi€, 1984; Lamph et al., 1988; driven by cooperation with c-Fos or Fra2. Yoshida et al., 1993) was also performed. Furthermore, after serum removal, ATF2 or Fra2 In cytoimmuno¯uorescence experiments, assessment overexpression did not signi®cantly in¯uence myoblast of the number of connectin-expressing myoblasts at terminal di€erentiation (control cells, FI: 16.46+1.29% cell con¯uence, just before serum removal, led us to vs 18.31+1.92% for ATF2 or 15.41+1.89% for Fra2; conclude that Fra2 or c-Fos overexpression inhibited Figure 5a,c,d), whereas c-Fos overexpression exerted a QM7 myoblast withdrawal from the cell cycle relative potent negative in¯uence (FI: 1.95+1.22%, P50.001; to control myoblasts (respectively 1.18+0.11% and Figure 5e). Finally, co-expression of Fra2 or c-Fos with 1.13+0.25% vs 2.99+0.31%, P50.01; Figure 4d,e,a), c-Jun abrogated myoblast di€erentiation as eciently as as did c-Jun overexpression (Figure 4b). ATF2 did not a TPA treatment (FI: respectively 2.10+1.54%, exert any signi®cant in¯uence on the number of post- 0.00+0.00% and 0.00+0.00%; Figure 5h,i,f). In mitotic myoblasts (Figure 4c). In addition, co- particular, c-Fos and c-Jun co-expression induced a expression of c-Jun and Fra2 or c-Jun and c-Fos more drastic di€erentiation block than c-Fos alone induced a more drastic inhibition of myoblast with- (P50.001). Interestingly, co-expression of ATF2 with drawal from the cell cycle (Figure 4h,i) than that c-Jun potentiated the stimulation of di€erentiation recorded for each protein alone (respectively induced by c-Jun alone (FI: 29.57+1.11% vs 0.32+0.06% vs 1.18+0.11% and 0.28+0.07% vs 21.05+1.12%, P50.005; Figure 5g,b).

Figure 5 In¯uence of expression or co-expression of AP-1 family members on avian myoblast terminal di€erentiation. Cytoimmuno¯uorescence staining with an antibody raised against connectin, 72 h after induction of di€erentiation. (a): control cells. (b): c-Jun overexpression. (c): ATF2 overexpression. (d): Fra2 overexpression. (e): c-Fos overexpression. (f): TPA treatment. (g): c-Jun+ATF2 co-expression. (h): c-Jun+Fra2 co-expression. (i): c-Jun+c-Fos co-expression. These results are representative of three separate experiments (6100)

Oncogene Regulation of myoblast differentiation by c-Jun L Daury et al 8003 This set of data strongly suggests that in avian Expression of myogenic factors is a target of myoblasts, c-Jun and ATF2 cooperate to stimulate AP-1 family members terminal di€erentiation, whereas cooperation between c-Jun and Fos-like proteins irreversibly inhibits with- In an attempt to determine the mechanisms involved in drawal from the cell cycle and consequently terminal the dual myogenic activity of c-Jun, we studied the di€erentiation. in¯uence of c-Jun and/or ATF2, Fra2, c-Fos stable

Figure 6 In¯uence of expression or co-expression of AP-1 family members on myoD mRNA levels. Thirty mg of total RNAs isolated from control or overexpressing c-Jun and/or ATF2, Fra2 or c-Fos QM7 myoblasts, at cell con¯uence (A) and 48 h after the induction of di€erentiation (B) were analysed by Northern blot. Quanti®cation was performed using a phosphorimager (Molecular Dynamics) and normalized in relation to S26 level. Results are expressed as percentages of the value recorded in control myoblasts. Data are the mean+s.e.m. of three separate experiments. A typical Northern blot is shown at each stage of the culture

Figure 7 In¯uence of expression or co-expression of AP-1 family members on myogenin mRNA levels. Thirty mg of total RNAs isolated from control or overexpressing c-Jun and/or ATF2, Fra2 or c-Fos QM7 myoblasts, at cell con¯uence (A) and 48 h after induction of di€erentiation (B) were analysed by Northern blot. Quanti®cation was performed using a phosphorimager (Molecular Dynamics) and normalized in relation to S26 level. Results are expressed as percentages of value recorded in control myoblasts. Data are the mean+s.e.m. of three separate experiments. A typical Northern blot is shown at each stage of the culture

Oncogene Regulation of myoblast differentiation by c-Jun L Daury et al 8004 overexpression on the level of avian myoD and myogenin mRNAs in Northern blots experiments. At cell con¯uence, just before induction of di€erentiation, no signi®cant in¯uence of c-Jun, ATF2 or c- Jun+ATF2 overexpression was detected on myoD mRNA levels (Figure 6A). However, Fra2, c-Jun+ Fra2, c-Fos and c-Jun+c-Fos signi®cantly decreased myoD mRNA levels (1.4-, 1.5-, 1.7- and 3-fold decrease respectively; Figure 6A), as did a TPA treatment (threefold decrease). Similar data were obtained 48 h after serum removal, with the exception of Fra2 and c-Jun/Fra2 which did not exert their Figure 8 In¯uence of c-Fos, Fra2 and ATF2 on the stimulation inhibitory in¯uence during di€erentiation (Figure 6B). of myogenin promoter activity by c-Jun. Cells were transfected with 1 mg/dish of the 7131/+40 avian myogenin-CAT reporter As expected (Marchal et al., 1995), myogenin gene and 2 mg of c-Jun, ATF2, c-Fos or Fra2 expression vectors. mRNA levels were at the brink of detection at cell TPA (50 ng/ml) was added in the culture medium when indicated. con¯uence and were up-regulated during terminal Results are expressed as percentages of CAT activity in control di€erentiation. At cell con¯uence, ATF2 did not cells. Data are the mean+s.e.m. of four separate transfection signi®cantly in¯uence myogenin mRNA levels, whereas experiments c-Jun, Fra2 and c-Fos decreased these levels (1.7-, 2.3- and 2.2-fold decrease respectively; Figure 7A); this reduction was even greater in c-Jun+Fra2 and c- mRNA levels (correlation between the fusion index and Jun+c-Fos overexpressing myoblasts (4.2- and 3.2-fold myogenin mRNAs, r=0.953, P50.001; correlation decrease respectively; Figure 7A), or in TPA-treated between the fusion index and myoD mRNAs, NS). cells (®vefold decrease). By contrast, c-Jun+ATF2 In addition, the in¯uence of c-Fos on myoD expression moderately increased myogenin mRNA levels (1.4-fold has already been documented (Pedraza-Alva et al., increase; Figure 7A). Forty-eight hours after induction 1994), thus leading us to bring a particular attention to of di€erentiation, c-Jun overexpression signi®cantly myogenin expression. increased myogenin mRNA levels (twofold increase; Figure 7B) whereas c-Jun+ATF2 co-expression in- c-Jun/ATF2 cooperation activates myogenin promoter duced a fourfold increase. On the other hand, we activity found that c-Fos, and to a lesser extent Fra2, continued to inhibit myogenin mRNA levels (11-fold In transient transfection experiments, we assessed the and twofold inhibition respectively; Figure 7B). In in¯uence of AP-1 family members on myogenin addition, c-Jun+c-Fos and c-Jun+Fra2 inhibited promoter activity. In agreement with data concerning myogenin mRNA levels more than c-Fos or Fra2 myogenin mRNA levels, c-Jun expression induced a alone (16-fold and fourfold inhibition respectively; fourfold stimulation of the basal promoter activity Figure 7B). Moreover, a TPA treatment also greatly (P50.005; Figure 8). Whereas devoid of any signi®cant reduced myogenin expression (25-fold decrease; Figure in¯uence on myogenin promoter activity by themselves, 7B). c-Fos or Fra2 co-expression with c-Jun fully abrogated All data concerning myogenic factors expression and its transcriptional in¯uence (Figure 8). Moreover, it di€erentiation are summarized in Table 1. Analysis of appeared that c-Jun stimulation was potentiated by these results led to the conclusion that the in¯uence of ATF2 co-expression (sevenfold stimulation relatively to all AP-1 family members on myoblast di€erentiation basal activity, P50.001; twofold relatively to c-Jun- was essentially related to their e€ect on myogenin induced activity; P50.01; Figure 8), despite the observation that ATF2 did not by itself in¯uence myogenin promoter activity. In addition, the c-Jun/ Table 1 Respective in¯uence of AP-1 family member expression ATF2 cooperation was abrogated by c-Fos or Fra2 co- or co-expression on myogenic factor expression and terminal di€er- expression (Figure 8). entiation myoD myogenin Fusion mRNAs (%)a mRNAs (%)b indexc ATF2 functionality is required for the positive myogenic influence of c-Jun Control 100 100 16.46 c-Jun 111 130 21.05 In order to assess the requirement of ATF2 function- ATF2 102 108 18.31 ality for the stimulation of myoblast di€erentiation by Fra2 84 40 15.41 c-Fos 63 6 1.95 c-Jun, we studied the in¯uence of an ATF2 negative c-Jun/ATF2 103 146 29.57 dominant mutant lacking the transactivation domain c-Jun/Fra2 147 21 2.10 of the protein (ATF2 D276) on the ability of c-Jun to c-Jun/c-Fos 48 8 0.00 stimulate myogenin promoter activity and terminal TPA 51 4 0.00 di€erentiation. a % of control value, see Figure 6; b % of control value, see Figure 7; In transient transfection experiments, we estab- c see Figure 5 lished that ATF2 D276 expression greatly reduced

Oncogene Regulation of myoblast differentiation by c-Jun L Daury et al 8005

Figure 10 In¯uence of c-Jun Leucine Zipper mutants on the myogenin promoter activity. Cells were transfected with 1 mg/dish of the 7131/+40 avian myogenin-CAT reporter gene and 2 mgof c-Jun, c-Jun m1, c-Jun m0, and c-Jun eb1 expression vectors. ATF2 expression vector was cotransfected when indicated. Results are expressed as percentages of CAT activity in control cells. Data are the mean+s.e.m. of three separate transfection experiments

Western blot experiments; Figure 2) repressed terminal di€erentiation (FI: 10.13+1.53% vs 15.43+1.01% for control cells, P50.025; Figure 9B). Constitutive expression of this mutant in stably c-Jun overexpressing myoblasts fully repressed the ability of the oncoprotein to stimulate myoblast di€erentiation (FI: 14.97+1.21% vs 15.43+1.01% for control cells, NS and vs 20.21+1.08% for c-Jun, P50.01; Figure 9b). All these data indicate that ATF2 functionality is required for the positive c-Jun myogenic in¯uence.

The functional c-Jun/ATF2 interaction does not involve formation of an heterodimerical complex To assess the possibility that stimulation of myogenin promoter activity was driven by formation of a c-Jun/ ATF2 heterodimerical complex, we studied the in¯uence of ATF2 coexpression with several c-Jun mutants : c-Jun eb1 only able to form homodimers, c-Jun m0 and c-Jun m1 displaying a preferential anity respectively for Fos- like proteins, and ATF2 (van Dam et al., 1998). Figure 9 In¯uence of a dominant negative ATF2 mutant on Surprisingly, all c-Jun mutants stimulated myogenin positive c-Jun myogenic activity. (A) Expression of ATF2 D276 promoter activity as wild-type c-Jun did (about threefold mutant abrogates c-Jun and c-Jun+ATF2 in¯uences on myogen- increase, P50.005, Figure 10). In addition, their in promoter activity. Cells were transfected with 1 mg/dish of the 7131/+40 myogenin-CAT reporter gene and 2 mg of c-Jun, in¯uence was similarly potentiated by ATF2 coexpres- ATF2 or ATF2 D276 expression vectors. The results are sion (5 ± 6-fold increase, P50.001). expressed as percentages of CAT activity in control cells. Data These data rule out the involvement of c-Jun Leucine are the mean+s.e.m. of three separate transfection experiments. Zipper in the cooperation with ATF2. In addition, they (B) In¯uence of ATF2 D276 mutant on avian myoblast terminal di€erentiation. Cytoimmuno¯uorescence staining with an anti- suggest that this cooperation does not involve forma- body raised against connectin of control or QM7 myoblasts stably tion of a c-Jun/ATF2 heterodimerical complex. In overexpressing c-Jun and/or ATF2 D276, 72 h after induction of agreement with this possibility, in gel retardation di€erentiation. (a): control cells. (b): c-Jun overexpression. (c): experiments, we detected ATF2 binding to the minimal ATF2 D276 overexpression. (d): c-Jun+ATF2 D276 co-expres- myogenin promoter sequence, but we failed to detect a sion. These results are representative of three separate experi- ments (6100) signi®cant c-Jun binding (data not shown).

Discussion the basal promoter activity and abolished the c-Jun- or c-Jun+ATF2-induced stimulation (Figure 9A). Although it is generally assumed that c-Jun is a potent Accordingly, ATF2 D276 stable overexpression in repressor of myoblast di€erentiation, some aspects of QM7 myoblast (overexpression was assessed in this myogenic in¯uence remain unclear. In particular,

Oncogene Regulation of myoblast differentiation by c-Jun L Daury et al 8006 identi®cation of partners possibly involved in this The dual in¯uence of c-Jun reported in this study, myogenic in¯uence is still lacking, leading us to related to the occurrence of changes in predominant characterize the c-Jun cooperations with proteins cooperations with ATF2 or Fos-like proteins, has to be included in the diversity of AP-1 complexes. Un- compared to results concerning its in¯uence on chicken expectedly, we found that, in relation to the nature of embryonic ®broblast transformation. Whereas c-Jun/ the cooperation, c-Jun exerts a dual in¯uence on Fos-like protein cooperation induces anchorage-inde- myoblast di€erentiation : repression through coopera- pendent cell growth, serum growth independency de- tion with Fos-like proteins, stimulation through pends on a c-Jun/ATF2 cooperation (for review van Dam cooperation with ATF2. and Castellazzi, 2001). Therefore, as in the present work, the nature of the c-Jun partner plays a crucial role in the induction of speci®c genetic programs. c-Jun exerts a dual myogenic influence In this work, we observed that c-Jun stable over- Mechanisms involved in the c-Jun dual influence on expression exerts two opposite in¯uences on quail myoblast differentiation myoblast di€erentiation, according to the stage of cell progression in the myogenic program. Whereas it We searched for the in¯uence of c-Jun on myoD and represses myoblast withdrawal from the cell cycle myogenin expression, paying particular attention to the before induction of di€erentiation by reducing serum cooperations of the oncoprotein with ATF2 or Fos-like concentration in the culture medium, it signi®cantly proteins. As myoD mRNA levels are not in¯uenced at stimulates myoblast terminal di€erentiation after its any stage of the culture by c-Jun and/or ATF2 induction. In the same line, a recent report indicates expression, we conclude that myoD expression is not that c-Jun also exerts two opposite functions before a target for c-Jun/ATF2 cooperation. By contrast, c- and after neuronal di€erentiation of PC12 cells (Leppa Jun/c-Fos co-expression or a TPA treatment signi®- et al., 2001). Whereas c-Jun expression protects cantly reduces myoD expression, in agreement with undi€erentiated cells against apoptosis and triggers previous data obtained in murine myoblasts (Pedraza- neurite formation, it induces apoptosis in di€erentiated Alva et al., 1994). PC12 cells. In addition, the present study provides a set We were particularly interested in the regulation of of convincing data indicating that this dual in¯uence in myogenin expression by c-Jun. We ®rst observed that, myoblasts is mediated through cooperation with for each kind of expression or co-expression experi- di€erent partners. Firstly, we established that the ment, myogenin mRNA levels are highly signi®cantly negative c-Jun myogenic in¯uence is linked to high related to the extent of myoblast fusion. In association endogenous levels of c-Fos or Fra2 proteins, whereas with the major importance of myogenin for the the positive e€ect is recorded after a spontaneous induction of terminal di€erentiation established in decrease in these levels, favoring c-Jun/ATF2 coopera- gene disruption experiments (Hasty et al., 1993; tion. Secondly, we observed that co-expression of c-Jun Nabeshima et al., 1993), this observation suggests and c-Fos or Fra2 inhibits myoblast withdrawal from that this gene is an important target for c-Jun the cell cycle and terminal di€erentiation, whereas co- interactions. Interestingly, the stimulation of terminal expression of c-Jun and ATF2 exerts the opposite di€erentiation by c-Jun is associated with a rise in in¯uence. Therefore, these results clearly suggest that myogenin mRNA levels, and potentiated by ATF2 co- the repression of myoblast withdrawal from the cell expression. In conjunction with the observations cycle is induced by a c-Jun/Fos-like protein coopera- indicating that c-Jun increases myogenin promoter tion, in agreement with previous data indicating that activity, and that this in¯uence is enhanced by ATF2 TPA-inducible AP-1 activity is a potent repressor of co-expression, this set of data leads us to conclude myoblast di€erentiation (Sulakhe et al., 1985), also that c-Jun/ATF2 cooperation increases transcription con®rmed in the present study. They also are in of the myogenin gene, and consequently myoblast agreement with the work of Hirayama et al. (1995) terminal di€erentiation. Furthermore, this possibility reporting that TPA a€ects the step before commitment is also supported by the observation that the c-Jun- for fusion of quail myoblasts. induced rise in myogenin promoter activity is By contrast, the surprising stimulation of myoblast abrogated by expression of an ATF2 negative di€erentiation by c-Jun overexpression recorded after dominant. This last result establishes that endogenous serum removal appears to be driven by a c-Jun/ ATF2 functionality is required for the c-Jun stimula- ATF2 cooperation. This possibility is well sustained tion of myogenin expression. However, the exact by the observation that a negative dominant form of nature of the c-Jun/ATF2 cooperation remains to be ATF2 not only abrogates the in¯uence of c-Jun on established. The use of Leucine Zipper mutants of c- terminal di€erentiation, but also inhibits by itself Jun suggests that such cooperation does not involve c- myoblast di€erentiation. As ATF2 is devoid of any Jun/ATF2 heterodimer formation. This result is in¯uence by itself, these data suggest that ATF2 substantiated by EMSA experiments: we system- D276 exerts this negative in¯uence by abrogating the atically observed a signi®cant binding of two ATF2- stimulation of terminal di€erentiation by endogenous containing complexes to a sequence of the minimal c-Jun, thus underlining the physiological relevance of myogenin promoter, but we have not, to date, this regulation. detected c-Jun binding to DNA (data not shown).

Oncogene Regulation of myoblast differentiation by c-Jun L Daury et al 8007 Conversely, cooperation between c-Jun and Fos-like corresponding quail proteins (Vial et al., 2000). The pDP18 proteins results in a reduction in myogenin mRNA c-Jun D176 expression vector (Castellazzi et al., 1993) led to levels: (1) the inhibition of myogenin expression the expression of a c-Jun mutant deleted from almost all the induced by c-Jun before serum removal is potentiated transactivating domain. The pDP18 ATF2 D276 expression by Fra2 or c-Fos co-expression; (2) after the induction vector (Huguier and Castellazzi, unpublished data) led to the expression of an ATF2 mutant deleted from all the of myoblast di€erentiation, co-expression of Fos-like transactivating domain. The pDP18 c-Jun m1 and pDP18 proteins turns the stimulatory activity of c-Jun on c-Jun m0 expression vectors (van Dam et al., 1998) led to the myogenin expression into a repressive one. These data expression of c-Jun mutants which preferentially hetero- suggest that cooperation between c-Jun and Fos-like dimerize with respectively ATF2 and Fos-like proteins. The proteins are speci®cally involved in the reduction of pDP18 c-Jun eb1 expression vector (Castellazzi et al., 1993) myogenin mRNA levels. Study of myogenin promoter led to the expression of a c-Jun mutant in which the natural activity indicates that this regulation partly occurs at dimerization domain was replaced by the dimerization the transcriptional level. The observation that c-Fos domain of the viral EB1 protein (c-Jun eb1 only forms and Fra2 do not by themselves in¯uence myogenin homodimers). The 7131/+40 avian myogenin-CAT reporter promoter activity, but abrogate the stimulation induced plasmid contains a fragment of the chicken minimal myogenin promoter upstream the CAT reporter gene (Malik by c-Jun, clearly suggests an indirect in¯uence. In et al., 1995). particular, as ATF2 cooperation is needed for the increase in myogenin expression induced by c-Jun, it could be put forward that overexpression of c-Fos or Stable transfections Fra2 competes eciently with ATF2 for cooperating Cells were transfected using the calcium phosphate co- with c-Jun. precipitation method as previously described (Cassar-Malek et al., 1996). QM7 myoblasts constitutively expressing c-Jun were obtained by co-transfection of 10 mg of pDP18 c-Jun Conclusions expression vector with 2 mg of pSV2-neoR plasmid. Pools of In this work, we have provided evidence that c-Jun stable clones were selected in medium containing Neomycin (G418, 500 mg/ml) for 12 days. Cells expressing c-Jun D176, exerts a dual in¯uence on avian myoblast di€erentia- ATF2, c-Fos and Fra2 were obtained with the same tion. By acting through TPA-inducible interactions procedure using the corresponding plasmids previously involving Fos-like proteins, c-Jun could prevent an mentioned. anticipated myoblast di€erentiation able to induce a de®ciency in muscle development. Thereafter, by acting Transient transfections and CAT assays through c-Jun/ATF2 cooperations at the onset of di€erentiation, characterized by a decrease in the As for stable transfections, transient transfections were expression of c-Fos and Fra2, c-Jun stimulates terminal performed using the calcium phosphate co-precipitation di€erentiation, leading to an optimal muscle develop- procedure as previously described. One mg of pCMV b- ment. Such a mechanism is interesting because it galactosidase expression vector was cotransfected to provide an internal control of transfection eciency. After cell demonstrates that the same transcription factor could exposure to precipitates for 24 h, the DNA-containing exert two opposite in¯uences depending on the nature medium was replaced with a fresh medium containing TPA of its partner, however leading to a common (Tetradecanoyl Phorbol Acetate, 50 ng/ml) when indicated, developmental bene®t. It remains now to understand and the cells were grown for a further 24 h. b-galactosidase the mechanisms inducing the changes observed in the activity was measured as previously described (Nielsen et al., expression of c-Jun partners. 1983). CAT enzymatic activity was measured by following the kinetics of chloramphenicol acetylation (Cassar-Malek et al., 1996). Results are expressed as the percentage of control values after b-galactosidase normalization.

Materials and methods Cytoimmunofluorescence studies Myoblast di€erentiation was assessed by morphological Cell cultures changes and accumulation of connectin, a muscle speci®c Quail myoblasts of the QM7 cell line (Antin and Ordahl, marker only expressed in myoblast from withdrawal from the 1991) were seeded at a plating density of 7000 cells/cm2. They cell cycle (Hill et al., 1986). After methanol ®xation and were grown in medium 199 supplemented with tryptose appropriate washings, cells were stained with an antibody phosphate broth (0.2%), L-glutamin (2 mM), gentamicin raised against connectin (kindly provided by Dr F Pons, (50 mg/ml), and fetal calf serum (10%). Terminal di€erentia- INSERM Montpellier, France) and a ¯uorescein-conjugated tion was induced after 96 h of culture at cell con¯uence by antibody raised against mouse immunoglobulins. Myoblast lowering the medium's fetal calf serum concentration to withdrawal from the cell cycle was estimated by the per cent 0.5%. of cells expressing connectin at cell con¯uence (number of nuclei in stained myoblasts6100/total number of nuclei). To quantify terminal di€erentiation, the fusion index (FI) was Plasmids and reporter genes calculated by counting total nuclei and nuclei incorporated in The pDP18 c-Jun, pDP18 ATF2, pDP18 c-Fos and pDP18 myotubes (FI=number of nuclei in myotubes6100/total Fra2 expression vectors led to the expression of the number of nuclei).

Oncogene Regulation of myoblast differentiation by c-Jun L Daury et al 8008 Protein preparation and Western blot analysis Statistical analysis One hundred mg of nuclear proteins were resolved by 10 or Statistical analyses were performed using the paired t-test 12% SDS ± PAGE mini-gels and blotted onto nitrocellulose (Snedecor, 1961). membranes that were further incubated with anti-c-Jun, anti- c-Fos, anti-Fra2 or anti-ATF2 polyclonal antibodies (Vial et al., 2000). Signals were further detected by ECF (Amersham Pharmacia Biotech.). Quanti®cation was performed using a Storm was performed using a Storm phosphoimager (Molecular Dynamics). Acknowledgments We would like to thank Dr M Lemonnier for the gift of b- RNA preparation and Northern blot analysis tropomyosin-CAT reporter plasmid, Dr F Pons for the gift Total RNAs were isolated after 96 h (cell con¯uence) and of connectin antibody, Dr C Dechesne for the gift of 144 h (terminal di€erentiation) of culture, as described by probes for chicken myoD and myogenin and Dr Schmidt Chomczynski and Sacchi (1987). Thirty mg of total RNAs for the gift of 7131/+40 avian myogenin-CAT reporter were loaded in each lane of 1% formaldehyde-agarose gel. plasmid. This work was supported by grants from the After electrophoresis, RNAs were transferred onto nylon Institut National de la Recherche Agronomique (INRA), membrane. RNAs were hybridized with cDNA probes Association pour la Recherche sur le Cancer (ARC), Ligue labeled with 32P-adCTP using the Megaprime DNA labeling contre le cancer and Association FrancËaise contre les system (Amersham Pharmacia Biotech.). An S26 ribosomal Myopathies (AFM). Laetitia Daury is recipient of fellow- protein cDNA probe was used to provide an invariant ships from the Ligue De partemental contre le Cancer and control as previously described (Vincent et al., 1993). the Association pour la Recherche sur le Cancer and Quanti®cation was performed using a Storm phosphoimager Muriel Busson from the MinisteÁ re de la Recherche et de (Molecular Dynamics). l'Enseignement.

References

Antin PB and Ordahl CP. (1991). Develop. Biol., 143, 111 ± Malik S, Huang CF and Schmidt J. (1995). Eur. J. Biochem., 121. 230, 88 ± 96. Benezra R, Davis RL, Lockshon D, Turner DL and Marchal S, Cassar-Malek I, Magaud JP, Rouault JP, Weintraub H. (1990). Cell, 61, 49 ± 59. Wrutniak C and Cabello G. (1995). Expl.CellRes.,220, Braun T, Rudnicki MA, Arnold HH and Jaenisch R. (1992). 1±10. Cell, 71, 369 ± 382. Nabeshima Y, Hanaoka K, Hayasaka M, Esumi E, Li S, Cassar-Malek I, Marchal S, Rochard P, Casas F, Wrutniak Nonaka I and Nabeshima Y. (1993). Nature, 364, 532 ± C, Samarut J and Cabello G. (1996). J. Biol. Chem., 271, 535. 11392 ± 11399. Nielsen DA, Chou J, MacKrell AJ, Casadaban MJ and Castellazzi M, Loiseau L, Piu F and Sergeant A. (1993). Steiner DF. (1983). Proc. Natl. Acad. Sci. USA, 80, 5198 ± Oncogene, 8, 1149 ± 1160. 5202. Chomczynsky P and Sacchi N. (1987). Anal. Biochem., 162, Park K, Chung M and Kim SJ. (1992). J. Biol. Chem., 267, 156 ± 159. 10810 ± 10815. Falcone G, Alema S and Tato F. (1991). Mol. Cell Biol., 11, Pedraza-Alva G, Zingg JM and Jost JP. (1994). J. Biol. 3331 ± 3338. Chem., 269, 6978 ± 6985. Greenberg ME and Zi€ EB. (1984). Nature, 311, 433 ± 438. Postigo AA and Dean DC. (1997). EMBO J., 16, 3935 ± 3943. Hasty P, Bradley A, Morris JH, Edmonson DG, Venuti JM, Rudnicki MA, Braun T, Hinuma S and Jaenisch R. (1992). OlsonENandKleinWH.(1993).Nature, 364, 501 ± 506. Cell, 71, 383 ± 390. Hebrok M, Fuchtbauer A and Fuchtbauer EM. (1997). Exp. Snedecor GW. (1961). Statistical Methods. p.534. Iowa State Cell Res., 232, 295 ± 303. University Press: Ames IA. Hill CS, Duran S, Lin ZX, Weber K and Holtzer H. (1986). J. Su H, Bos TJ, Monteclaro FS and Vogt PK. (1991). Cell Biol., 103, 2185 ± 2196. Oncogene, 6, 1759 ± 1766. Hirayama E, Hase H, Nakano Y and Kim J. (1995). Cell Sulakhe PV, Johnson DD, Phan NT and Wilcox R. (1985). Struct. Funct., 20, 377 ± 386. FEBS Lett., 186, 281 ± 285. Huguier S, Baguet J, Perez S, van Dam H and Castellazzi M. van Dam H, Huguier S, Kooistra K, Baguet J, Vial E, van (1998). Mol. Cell Biol., 18, 7020 ± 7029. der Eb AJ, Herrlich P, Angel P and Castellazzi M. (1998). Lamph WW, Wamsley P, Sassone-Corsi P and Verma IM. Genes Dev., 12, 1227 ± 1239. (1988). Nature, 334, 629 ± 631. van Dam H and Castellazzi M. (2001). Oncogene, 20, 2453 ± Lassar AB, Thayer MJ, Overell RW and Weintraub H. 2464. (1989). Cell, 58, 659 ± 667. Vial E, Perez S and Castellazzi M. (2000). Oncogene, 19, Lemercier C, To RQ, Carrasco RA and Konieczny SF. 5020 ± 5029. (1998). EMBO J., 17, 1412 ± 1422. Vincent S, Marty L and Fort P. (1993). Nucleic Acids Res., Leppa S, Eriksson M, Sa€rich R, Ansorge W and Bohmann 21, 1948 ± 1953. D. (2001). Mol. Cell Biol., 21, 4369 ± 4378. Yoshida T, Suzuki T, Sato H, Nishina H and Iba H. (1993). Lu J, Webb R, Richardson JA and Olson EN. (1999). Proc. Nucleic Acids Res., 21, 2715 ± 2721. Natl. Acad. Sci. USA, 96, 552 ± 557.

Oncogene