Selective Repression of Myod Transcription by V-Myc Prevents

Home , Myogenesis

Oncogene (2002) 21, 4838 – 4842 ª 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00 www.nature.com/onc SHORT REPORTS Selective repression of myoD transcription by v-Myc prevents terminal diﬀerentiation of quail embryo myoblasts transformed by the MC29 strain of avian myelocytomatosis virus

Severina A La Rocca1,3,5, Serena Vannucchi1,4,5, Monica Pompili1, Deborah F Pinney2, Charles P Emerson Jr2, Milena Grossi*,1 and Franco Tato` 1,6

1Istituto Pasteur-Fondazione Cenci-Bolognetti, Dipartimento di Biologia Cellulare e dello Sviluppo, Sezione di Scienze Microbiologiche, Universita’ di Roma ‘La Sapienza’, 00185-Roma, Italy; 2Department of Cell and Developmental Biology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, PA 19104-6058, USA

We have investigated the mechanism by which expression In vitro, after an initial proliferative phase, myogenic of the v-myc oncogene interferes with the competence of cells withdraw from the cell cycle, enter the post- primary quail myoblasts to undergo terminal differentia- mitotic stage and activate the transcription of a battery tion. Previous studies have established that quail of muscle-specific genes. Terminal differentiation of myoblasts transformed by myc oncogenes are severely skeletal myogenic cells is activated and maintained by impaired in the accumulation of mRNAs encoding the the concerted action of ubiquitous transcription myogenic transcription factors Myf-5, MyoD and factors, transcriptional coactivators (Puri and Sartor- Myogenin. However, the mechanism responsible for such elli, 2000) and muscle-specific transcription factors a repression remains largely unknown. Here we present belonging to two major groups: the Muscle Regulatory evidence that v-Myc selectively interferes with quail Factors (MRFs) of the MyoD family (Myf-5, MyoD, myoD expression at the transcriptional level. Cis- Myogenin and MRF-4) and members of the MEF2 regulatory elements involved in the auto-activation of family (Lassar et al, 1994; Li and Olson, 1992; qmyoD are specifically targeted in this unique example Molkentin and Olson, 1996). Thus, it is hardly of transrepression by v-Myc, without the apparent surprising that the interference exerted by many participation of Myc-specific E-boxes or InR sequences. oncogenes on the differentiation potential of myogenic Transiently expressed v-Myc efficiently interfered with cells in vitro is often associated with reduced expression MyoD-dependent transactivation of the qmyoD regula- and/or abnormal functioning of MRFs (Alema` and tory elements, while the myogenin promoter was Tato` , 1994; Maione and Amati, 1997). Independent unaffected. Finally, we show that forced expression of investigations on the effects of Myc proteins on MyoD in v-myc-transformed quail myoblasts restored myogenesis have led to different results that include: myogenin expression and promoted extensive terminal (i) complete inhibition of terminal differentiation in differentiation. These data suggest that transcriptional primary quail myogenic cells (Falcone et al, 1985; La repression of qmyoD is a major and rate-limiting step in Rocca et al, 1989, 1994); (ii) inhibition of fusion into the molecular pathway by which v-Myc severely inhibits myotubes, but not of biochemical differentiation and terminal differentiation in myogenic cells. cell cycle withdrawal, in the murine myogenic cell line Oncogene (2002) 21, 4838 – 4842. doi:10.1038/sj.onc. C2C12 (Crescenzi et al, 1994); (iii) severe inhibition of 1205586 MyoD- and/or Myogenin-initiated myogenic conver- sion of murine NIH3T3 cells (Miner and Wold, 1991). Keywords: Myc; qmyoD; myogenic differentiation; cell Indeed, Myc proteins can act as powerful modula- transformation tors of transcription through interaction with Max and/or other factors (Amati et al, 2001; Gartel et al, 2001; Grandori et al, 2000), and their action can be expected to vary, depending on the specific cellular context analysed. In this light, we decided to investigate further the mechanism by which v-Myc, *Correspondence: Milena Grossi, Dipartimento di Biologia Cellulare the product of the viral avian myc allele, severely e dello Sviluppo, Sezione di Sc. Microbiologiche, Via dei Sardi 70, perturbs the differentiation potential of primary quail 00185-Roma, Italy; E-mail: [email protected] Current addresses: 3Department of Immunology and Pathology, myoblasts, on the assumption that such a cellular Institute for Animal Health, Pirbright Surrey, GU24 ONF UK; model ought to be more likely to reflect muscle cell 4Istituto Superiore di Sanita` , Laboratory of Virology, 00161-Roma, physiological behavior, and hence, be informative on Italy the mechanism(s) of oncogenic processes in skeletal 5These authors contributed equally to this work 6Deceased on 7th July 2001 muscle tissues. Primary quail myoblasts in vitro have Received 22 January 2002; revised 4 April 2002; accepted 15 April been shown to express myoD, myf5 and myogenin in 2002 the proliferative phase before terminal differentiation; Transcriptional silencing of qmyoD by the v-myc oncogene SA La Rocca et al 4839 myogenin expression was further up-regulated upon terminal differentiation, whereas myf5 expression appeared to decline in coincidence with the progression of myotube maturation (Russo et al, 1997). In contrast, quail myoblasts infected with avian myelocytomatosis virus MC29, encoding a 110 kD Gag-v-Myc fusion protein, QM(myc), used in this study, failed to terminally differentiated when shifted to differentiation medium (DM), and displayed a strong reduction in the accumulation of myoD, myogenin and myf5 mRNA in proliferative condition, as compared to uninfected myoblasts (QMb) (data not shown; see Falcone et al, 1985; La Rocca et al, 1989, 1994). In order to elucidate the mechanism responsible for such a repression we analysed the functioning of qmyoD and myogenin control region in QM(myc), as compared to QMb. qmyoD expression is essentially dependent on a well defined, tissue-specific enhancer element, named R1, which is located 11.5 to 17.5 Kb upstream of the transcription start site. The R1 enhancer activity localizes to two adjoining fragments, P3 and P4, both containing myogenic E-boxes, putative MEF-2 sites and sequences identical to the corresponding human distal enhancer (Pinney et al, 1995). Transient transfec- tion assays were performed with CAT constructs driven by the following elements of the qmyoD control region: 72400 (promoter), R1 (entire enhancer) and its Figure 1 Transcriptional repression of myoD by v-Myc in quail myoblasts. (a) Schematic representation of the qmyoD R1 distal sub-elements P3, P4, sc1, sc3 and sc5 (Figure 1a), as enhancer. (b) Transient expression of CAT reporters driven by well as with a CAT construct controlled by the 5’ the qmyoD 72400 promoter region, the R1 control region and re- regulatory region 71565 of mouse myogenin, which lative sub-elements, and the myogenin regulatory region 71565, has a strong homology to the corresponding 5’ flanking analysed in QM(myc) versus QMb and (c) in QM(mycER) treated with b-estradiol (+) or left untreated (7). Quail embryo myo- region of avian myogenin (Malik et al, 1995). Figure 1b blasts (QMb) were prepared as previously described (Falcone et shows that in QM(myc), the expression level elicited by al., 1985). Quail cells were propagated in Growing Medium (D- the tissue specific enhancer R1 was significatively lower MEM containing 10% fetal calf serum (FCS), 10% tryptose (about eightfold) than in QMb. The analysis of R1 phosphate broth and 1% chicken serum) (GM). QM(myc) cells sub-elements P3 and P4 indicated that v-Myc could were obtained by high-multiplicity infection with the subgroup A pseudotype MC29(RAV1) and were propagated in GM on col- affect transcription from both these regions, particu- lagen-coated dishes. QM(mycER) were obtained by infection with larly from the P3 sub-element. Interestingly, the the mycER(RAV1) retroviral vector, generated by transfecting analysis of the P3 sub-fragments sc1, sc3 and sc5 chick embryo fibroblasts, producing the helper virus Rous asso- showed that the strongest inhibition (about 12-fold) in ciated virus 1 (RAV1), with the pmycER plasmid, derived by de- letion of the sea oncogene from pmycER-ts-sea vector (Pollerberg QM(myc) was associated with the construct controlled et al., 1995), and propagated in GM on collagen-coated dishes. by sc1 element, that is principally responsible for the Equimolar amounts of CAT constructs were cotransfected over- activity and muscle specificity of R1 enhancer in QMb night with the control plasmid RSVLacZ by the DNA-calcium phosphate precipitation method. Transfected cells were fed with (Pinney et al, 1995). Expression of the positive, but not 77 tissue-specific, sc5 control element was not appreciably GM, with the exception of QM(mycER)+, where 10 M b-estradiol was also added, and harvested after 48 h. Expression levels of affected in QM(myc). In order to compare more CAT protein were determined in cell extracts using an enzymatic isogenic recipient cells, quail myoblasts transformed immuno-assay (Boeringher Mannheim). For all the experiments, by the conditional, chimaeric v-mycER oncogene RSVLacZ was also cotransfected with the plasmid RSVCAT in (QM(mycER)), where the hormone binding domain identical dishes, to normalize the results of independent experiments. Relative CAT amounts are presented as percentage of of the human estrogen receptor (ER) is tethered to v- the RSVCAT control construct expression and represent the aver- Myc and confers b-estradiol dependence to the age+s.d. of three independent experiments transforming activity of v-MycER, were transiently transfected with selected representative constructs among which used for QM(myc). As shown in Figure 1c, b-estradiol-activated v-MycER inhibited the activity the qmyoD R1 enhancer, and particularly with its sc1 of the qmyoD R1 enhancer and the inhibition pattern sub-element. The low activity of the myogenin was consistent with that observed in QM(myc). These regulatory region (71565 construct) in QM(myc)and data suggest that v-myc-induced transformation of QM(myc-ER)+b-estradiol indicates a transcriptional quail myoblasts is associated with the repression of interference also with myogenin expression. We also qmyoD transcription, through interference with the investigated the kinetics of qmyoD mRNA down- activity of the major tissue-specific control elements of regulation upon v-MycER activation. Results from

Oncogene Transcriptional silencing of qmyoD by the v-myc oncogene SA La Rocca et al 4840 Northern blot analysis indicated that repression of qmyoD mRNA by v-Myc was evident within 2 h of b- estradiol treatment (data not shown), suggesting that interference with qmyoD expression is a very early event after v-MycER activation, rather than an epiphenomenon, secondarily occurring during the propagation of myc-transformed cells. An important feature of MRFs is their ability to transcriptionally activate their own and each other’s expression. To assess whether the v-Myc oncoprotein could interfere with qmyoD autoactivation we analysed the activity of R1 enhancer control elements, regulated by the exogenous qmyoD, in uninfected or v-myc transformed quail fibroblasts, QEF and QEF(myc) respectively. Cells were transiently cotransfected with a qmyoD expression vector and various CAT reporters driven by the following elements: R1, P3, sc1, sc3, sc5, and the myogenin regulatory regions, 71565 and 784. Expression from R1 and its muscle specific subfrag- ments P3, sc1 and sc3, in QEF, strictly required the presence of exogenous qMyoD (Figure 2a), confirming the role of the sc1 and sc3 elements in qmyoD autoactivation (Pinney et al, 1995). Interestingly, qMyoD was severely impaired in activating expression Figure 2 v-Myc selectively interferes with MyoD/qmyoD auto- from all R1-tissue-specific elements in QEF(myc), in regulatory loop in quail fibroblasts. Early passage quail embryo particular, from the sc3 sub-fragment. Unexpectedly, in fibroblasts (QEF) were propagated in GM. QEF(myc) were ob- the same conditions, the myogenin regulatory region tained by high-multiplicity infection with the subgroup A pseudotype MC29(RAV1) and were propagated in GM on collagen- resulted equally activated by exogenous qMyoD either coated dishes. (a) CAT reporters driven by various qmyoD control in QEF or in QEF(myc), thus indicating that elements and by the myogenin regulatory regions 71565 and transformation by v-myc does not affect MyoD 784, were cotransfected with qmyoD expression vector Xc307, transactivating function per se; rather, v-Myc interferes or the corresponding empty vector (pECE), in QEF or in QEF specifically with the proper functioning of MyoD on (myc). (b) CAT reporters, driven by the qmyoD control elements sc1 and sc3, the myogenin regulatory regions 71565 and 784, qmyoD 5’ regulatory sequences. and by an oligomerized Myc responsive element M4 (Kretzner To demonstrate that the effect of v-Myc on the et al., 1992), were cotransfected with qmyoD expression vector qmyoD autoactivation circuit was specific and direct, Xc307 or its empty version (pECE), and myc expression vectors, CAT constructs containing the qmyoD sc1 and sc3 encoding wild type (RCANMC29) or mutant Myc (RCAND48), in QEF. Equimolar amounts of plasmids were utilized, with the elements or the two myogenin control region 71565 exception of the Myc vectors, used in a fivefold molar excess over and 784 were cotransfected in QEF along with the the MyoD expression vector. CAT expression was analysed as qmyoD expression vector and with myc expression described in Figure 1. Mean values+s.d. of three independent vectors, encoding either wt-Myc or the D48-Myc experiments are shown mutant, which lacks the entire leucine zipper domain. Such a mutant is unable to heterodimerize with Max (Crouch et al, 1993), and can neither cause cell mediated qmyoD repression, and thus activate endo- transformation nor inhibit myogenic differentiation genous myogenin expression; this, in turn, might allow (La Rocca et al, 1994). The results in Figure 2b show the execution of the terminal differentiation program. that sc1 and sc3 activation by qMyoD was strongly Accordingly QM(myc) were super-infected with an inhibited by transient co-expression of Myc, whereas empty adenoviral vector (Ad-null) or with an adeno- myogenin regulatory regions were activated by qMyoD viral vector encoding murine MyoD (Ad-MyoD) and, independently from functional Myc expression. Inter- after cultivation in DM, super-infected cultures were estingly, the D48-Myc mutant could not repress examined for the expression of Myogenin, skeletal transactivation by MyoD on either sc1 or sc3, Myosin Heavy Chain (MHC) and fusion into multi- suggesting that Max binding was required, and that nucleated myotubes. Figure 3 shows that Ad-MyoD the aminoterminal transmodulatory domain was not superinfection triggered quantitative expression of sufficient to interfere with qmyoD autoregulatory endogenous Myogenin as well as MHC in QM(myc), circuit. These results strongly indicate that suppression as compared to Ad-null super-infected cells (Figure of endogenous myoD transcription in quail myoblasts 3a), and that differentiated cells could fuse into is a direct and specific effect of v-Myc that might multinucleated myotubes (Figure 3b). represent a major, rate-limiting step in Myc-induced All together, the results here presented are block of terminal differentiation in QM(myc). On this consistent with the following molecular pathway to basis, it would be predicted that forced expression of explain how transcription of myogenin is severely MyoD in QM(myc) should override the v-Myc- reduced and competence for terminal differentiation is

Oncogene Transcriptional silencing of qmyoD by the v-myc oncogene SA La Rocca et al 4841 Myc proteins are known to affect transcription by dimerization with Max and binding to DNA via the specific E-box ‘CACGTG’; in addition, Myc can also interfere with gene transcription via the InR elements and by interacting with several transcription factors (Amati et al, 2001; Gartel et al, 2001; Grandori et al, 2000). Surprisingly, the sc1/sc3 enhancer elements of qmyoD regulatory region do not ostensibly contain myc-specific E boxes or InR elements, although the presence of an intact leucine zipper was required for v-Myc inhibitory activity. Here we have presented evidence that MyoD can transactivate the minimal myogenin promoter, which contains a single myogenic E-box, even in the presence of v- Myc. This finding strongly suggests that myogenic E- boxes, present also in the sc1/sc3 elements, are not involved in the transcriptional silencing by v-Myc, Figure 3 Rescue of myogenic differentiation by exogenous and confirms the contention that v-Myc is not MyoD in QM(myc). QM(myc) cells were super-infected with an empty (Ad-null) or a MyoD-carrying adenoviral vector under simply interfering with MyoD transactivating func- the control of Rous sarcoma virus LTR (Ad-MyoD) (Murry et tion per se, but does so in the context of quail al., 1996). Multiplicity of infection ranged between 100 and myoD autoregulatory circuit. 1000. Differentiation Medium (F14 medium containing 2% FCS Although Myc and MyoD are unable to hetero- and 0.5 mg/ml bovine insulin) (DM) was added the day after infection and cells were analysed after further 3 days. (a) Western dimerize with each other, we cannot exclude the blot analysis using the anti-Myc polyclonal antibody 237 (Crouch possibility that Myc and MyoD share a transcriptional et al., 1993), to reveal the P110gag-v-myc protein, the anti-skeletal or co-transcriptional partner(s), specifically required in muscle Myosin Heavy Chain (MHC) monoclonal antibody the MyoD autoregulatory circuit. Alternatively, an MF20, and an anti-Myogenin polyclonal antibody. (b) Ad-MyoD unknown v-Myc target gene might be activated that super-infected cells were fixed with 3% paraformaldehyde, per- meabilized with 0.25% Triton X-100, and incubated with the interferes with MyoD transactivating function on the polyclonal antiserum raised against murine MyoD expressed in sc1/sc3 elements. bacteria R4B4 (Russo et al., 1997) and with the above referred The implication of the myoD autoregulatory circuit antibodies against MHC and Myogenin. Nuclei were stained with as a specific Myc target might reflect the existence of a the Hoechst 33258 dye (Sigma), that revealed also the abundant occurrence of apoptotic figures (A,D), as expected for myc-trans- similar, reversible mechanism during muscle formation formed cells in a low serum medium. Left micrographs represent in the embryo. Proliferation and migration of the same field where MHC (B) and MyoD (C) were visualized. committed muscle precursor cells from the somites to Right micrographs represent the same field, from a sister plate, distal sites is required during embryogenesis, and these where MHC (E) and Myogenin (F) expression were analysed. processes are under the control of several genes Bar=20 mm including HGF/SF and c-met (Birchmayer and Gher- ardi, 1998). In migrating precursor cells, presumably derived from somitic cells expressing Myf-5 and/or lost in QM(myc): v-Myc-induced silencing of qmyoD MyoD, myogenic bHLH factor expression is repressed transcription would lead to inactivation of myogenin during the migratory phase and yet the myogenic transcription; in the absence of functional Myogenin, identity is retained (Tajbakhsh and Buckingham, 1994). neither muscle-specific gene expression nor permanent We have shown that QM(myc) can be induced to withdrawal from cell cycle can be activated in terminally differentiate when co-cultured with various QM(myc). This model is supported by the observa- normal cell types, thus implying that v-Myc-induced tion that forced expression of an exogenous MyoD in transformation does not erase the myogenic identity QM(myc) could rescue endogenous myogenin expres- (La Rocca et al, 1989). A c-Myc-dependent, transcrip- sion and the execution of the myogenic differentiation tional repression circuit, akin to the one here illustrated program. Although at the moment we cannot exclude for v-Myc, might be responsible for the reversible the possibility that other defects in the myogenic suppression of MyoD family member expression in regulatory circuit may contribute to v-myc-induced migrating myogenic cells. Consistent with this view, interference with differentiation of quail myoblasts, expression of c-myc is known to be elevated in the present results strongly indicate that transcrip- embryonic and proliferating cells (Zimmerman et al, tional repression of qmyoD is a critical, rate-limiting 1986), and the consequences of high c-Myc levels on step that can largely account for the inhibition of quail embryo myoblasts are fully comparable to those myogenic differentiation by v-Myc, and led to the of v-Myc (La Rocca et al, 1994). Accordingly, identification of qmyoD distal enhancer components QM(myc) represent an attractive model system to (sc1, sc3), previously shown to be involved in qmyoD identify MyoD-specific target genes in myogenic cells autoactivation (Pinney et al, 1995), as the regions as well as factors, acting upstream of the MyoD sensitive to v-Myc induced transcriptional down- family, involved in the establishment and maintenance regulation. of the myogenic identity.

Oncogene Transcriptional silencing of qmyoD by the v-myc oncogene SA La Rocca et al 4842 Acknowledgements USA) for the anti-Myogenin antibody, G Piaggio (Istituto WethankSAlema` (CNR, Roma, Italy) for the adenoviral Regina Elena, Roma, Italy) for the M4 construct and M vectors, D Fischman (Cornell University, New York, USA) Zenke (MDC, Berlin, Germany) for the pmycER-ts-sea for the MF20 antibody, D Gillespie (CRC Beatson vector. This work was supported by grants from Associa- Laboratories, Glasgow, UK) for the 237 antibody, E zione Italiana per la Ricerca sul Cancro (AIRC), MURST Olson (UT Southwestern Medical Center, Dallas, USA) (Coﬁn 1999) and CNR/MURST (Legge 95/95). for the myogenin constructs, B Paterson (NIH, Bethesda,

References

Alema` SandTato` F. (1994). Sem. Cancer Biol., 5, 147 – 156. Maione R and Amati P. (1997). Biochim. Biophys. Acta, Amati B, Frank SR, Donjerkovic D and Taubert S. (2001). 1332, 19 – 30. Bioch. Bioph. Acta,, 1471, M135 – M145. Malik S, Huang CF and Schmidt J. (1995). Eur. J. Biochem., Birchmayer C and Gherardi E. (1998). Trends Cell Biol., 8, 230, 88 – 96. 404 – 410. Miner JH and Wold BJ. (1991). Mol. Cell. Biol., 11, 2842 – CrescenziM,CrouchDHandTato` F. (1994). J. Cell Biol., 2851. 125, 1137 – 1145. Molkentin JD and Olson EN. (1996). Curr. Opin. Genes Dev., Crouch DH, Fisher F, Clark W, Jayaraman PS, Goding CR 6, 455 – 453. and Gillespie DAF. (1993). Oncogene, 8, 1849 – 1855. Murry CE, Kay MA, Bartosek T, Hauschka SD and Falcone G, Tato` FandAlema` S. (1985). Proc. Natl. Acad. Schwartz SM. (1996). J. Clin. Invest., 98, 2209 – 2217. Sci. USA, 82, 426 – 430. PinneyDF,delaBrousseFC,FaermanA,ShaniM, Gartel AL, Ye X, Goufman E, Shianov P, Hay N, Maruyama K and Emerson Jr CP. (1995). Dev. Biol., Najmabadi F and Tyner L. (2001). Proc. Natl. Acad. Sci. 170, 21 – 38. USA, 98, 4510 – 4515. Pollerberg GE, Kuschel C and Zenke M. (1995). J. Neurosci. Grandori C, Cowley SM, James LP and Eisenman RN. Res., 41, 427 – 442. (2000). Annu. Rev. Cell Dev. Biol., 16, 653 – 699. Puri PL and Sartorelli V. (2000). J. Cell. Physiol., 185, 155 – Kretzner L, Blackwood EM and Eisenman RN. (1992). Curr. 173. Top. Microbiol. Immunol., 182, 435 – 443. Russo S, Tato` F and Grossi M. (1997). Oncogene, 14, 63 – 73. La Rocca SA, Grossi M, Falcone G, Alema` SandTato` F. Tajbakhsh S and Buckingham ME. (1994). Proc. Natl. Acad. (1989). Cell, 58, 123 – 131. Sci. USA, 91, 747 – 751. La Rocca SA, Crouch DH and Gillespie DAF. (1994). Zimmerman K, Yancopoulos GD, Collum RG, Smith RK, Oncogene, 9, 3499 – 3508. KohlNE,DenisKA,NauMM,WitteON,Toran- Lassar AB, Skapek SX and Novitch B. (1994). Curr. Opin. Allerand D, Gee CE, Minna JD and Alt FW. (1986). Cell Biol., 6, 788 – 794. Nature, 319, 780 – 783. Li L and Olson EN. (1992). Adv. Cancer Res., 58, 95 – 120.

Oncogene