THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry a d Molecular Biology, Inc.

Vol. 268, No. 13, Issue of May 5, pp. 9869-9878,1993 Printed in U.S.A.

CAMP-dependent Protein Kinase Represses Myogenic Differentiation and the Activity of the Muscle-specific Helix-Loop-Helix Transcription Factors Myf-5 and MyoD*

(Received for publication, November 6, 1992, and in revised form, December 31,1992)

Barbara Winter, Thomas Braun, and Hans Henning Arnold$ From the Department of Cell and Molecular Biology, University of Braunschweig, Konstantin- Uhde-Strasse 5, 3300 Braunschweig, Federal Republic of Germany

Myf-5 and MyoD are members of a family of muscle- specific basic helix-loop-helix (bHLH) proteins that are fundamental for myogenic cell differentiation and transcriptional activation of muscle-specific genes. Here we report that elevated levels of the intracellular signaling molecule CAMP and overexpression of CAMP-dependent protein kinase (PKA) inhibit my- ogenic differentiation. PKA represses the transcrip- tional activation of muscle-specific genes by the my- ogenic regulators Myf-5 and MyoD. The repression is directed at the basic HLH domain and is mediated through the E-box DNA consensus motif to which these proteins bind. However, phosphorylation of Myf-5 and MyoD by PKA in vitro does not affect their ability to bind to DNA. PKA specifically inhibits the activity of myogenic bHLH proteins, but not of other HLH pro- teins, such as the ubiquitously expressed E2A gene products E12 and E47 (E2-5). Our results demonstrate that PKA mediates the CAMP-induced inhibition of muscle cell differentiation by repressing the activity of Myf-5 and MyoD. The inhibition by PKA occurs post-translationally and presumably affects the trans- activation process at a step following DNA-binding. The regulation of Myf-5 and MyoD function by a CAMP-dependent pathway may partly explain how ex- ternal signals generated by serum and certain peptide growth factors can be transduced to the nucleus and inhibit dominant-acting factors that are responsible for myoblast differentiation.

Differentiation of skeletal muscle cells is characterized by the coordinate activation of an array of genetically unlinked muscle-specific genes and the arrest of cell proliferation. Serum components and peptide growth factors, such as basic fibroblast growth factor and transforming growth factor-@, specifically inhibit the differentiation of muscle cells and the transcriptional activation of muscle-specific genes (for review, see Florini et al. (1991)). These and other observations have led to the concept that differentiation and growth are mu- tually exclusive events in myogenic cell lines. Little, however, is known about the mechanisms underlying the antagonism between molecules that promote growth and those that regu-

~

* This work was supported by grants of the Deutsche Forschungs- gemeinschaft, Deutsche Muskelschwundhilfe e. V., and the European Community. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed: Dept. of Cell and Molecular Biology, University of Braunschweig, Konstantin-Uhde- Str. 5, 3300 Braunschweig, FRG.

late cell differentiation. Therefore, the elucidation of net- works that link signals at the cell surface to transcriptional events in the nucleus requires the identification of intracel- lular signaling molecules as well as the characterization of trans-acting factors that serve as nuclear targets for the regulatory pathways.

Myoblasts offer an excellent model system for studying mechanisms whereby differentiation may be controlled, be- cause a family of muscle-specific nuclear factors that can activate the complete differentiation program has recently been identified (for reviews see, Emerson (19901, Olson (1990), and Weintraub et al. (1991a)). This family of myogenic control proteins which includes MyoD (Davis et ul., 1987), myogenin (Wright et al., 1989; Edmondson and Olson, 1989), Myf-5 (Braun et al., 1989a), and MRF4 (Rhodes and Konieczny, 1989), also called herculin (Miner and Wold, 1990) or Myf-6 (Braun et al., 1990a), shares extensive sequence homology within a basic region and an adjacent putative helix-loop- helix (HLH)’ domain which mediate sequence-specific DNA binding and dimerization, respectively (Murre et ul., 1989; Davis et al., 1990; Brennan et al., 1991; Winter et al., 1992). High affinity binding to the DNA consensus sequence, CANNTG, referred to as E-box, requires heterodimerization of the myogenic bHLH proteins with the more widely ex- pressed HLH proteins E12 or E47 (E2-5). Transactivator domains have been identified outside of the bHLH region in the NH, terminus of MyoD (Weintraub et al., 1991b) and the NH2 and COOH termini of myogenin (Schwartz et al., 1992) and Myf-5 (Braun et al., 1990b; Winter et al., 1992). Numerous muscle-specific genes containing one or more E-box motifs within their control regions can be transactivated through the interaction with the myogenic HLH proteins (for review, see Olson (1990), Emerson (1990), and Weintraub et al. (1991a)). Thus, these proteins constitute cell type-specific transcription factors and may be potential targets for signals generated by serum components and peptide growth factors.

An approach taken to identify signaling pathways through which growth factors may inhibit myogenesis has been to transfect myoblasts with activated oncogenes known to be involved in signal transduction and growth control. These investigations have revealed that activated ras proteins (Las- sar et al., 1989; Olson et al., 1987; Payne et al., 19871, mem- brane-associated tyrosine protein kinases, such as v-src (Fal- cone et al., 1985), and nuclear protooncogene products, such

The abbreviation used are: HLH, helix-loop-helix; bHLH, basic HLH; PKC, protein kinase C; PKA, protein kinase A DMEM, Dulbecco’s modified Eagle’s medium; CAT, chloramphenicol acetyl- transferase; MHC, myosin heavy chain; kb, kilobase pair(s); MLC, myosin light chain; EMSA, Electrophoretic mobility shift assay; DM, differentiation medium.

9869

9870 Inhibition of Myogenesis by CAMP and Pk%

as C-myc (Miner and Wold, 1991; Schneider et at., 1987), C- fos, and jun (Rahm et af, 1989; Li et al., 1992a; Bengal et ui., 1992), and adenovirus E l a protein (Webster et ai., 1988; Enkemann et al., 1990; Braun et ai., 1992) can disrupt myo- genesis and inhibit muscle-specific gene expression. Some of these oncogene products act in pathways regulated by protein kinase C (PKC), which by itself can inhibit myogenesis. Phosphorylation of myogenin by PKC in vitro prevents DNA binding and constitutively expressed PKC inhibits transacti- vation by myogenin.2 Other than PKC, no pathways that may link components acting close to the cell membrane with those that function in the nucleus have yet been established. The availability of cloned regulatory proteins that induce muscle differentiation and directly activate transcription of muscle- specific genes offers the unique oppo~unity to explore mech- anisms whereby negative regulators of myogenesis prevent muscle-specific transcription.

CAMP is one of the important intracellular messengers that may be involved in the control of cellular events in response to external signals. It has been reported previously that di- butyryl cAMP and other agents that increase intracellular cAMP levels inhibit muscle cell differentiation and the expression of muscle-specific genes (Hu and Olson, 1988; Salminen et al., 1991). Regulation of gene transcription by CAMP is frequently mediated through CAMP-dependent pro- tein kinase (PKA) which phosphorylates nuclear targets (Kemp and Pearson, 1990) when cAMP is bound to the regulatory subunits of the holoenzyme, thereby releasing and activating the catalytic subunits (Clegg et al., 1987).

To begin to dissect mechanisms whereby CAMP-dependent pathways may inhibit the muscle differentiation program, we investigated the role of PKA in the activation of muscle- specific gene expression. In the present report, we demon- strate that transactivation of endogenous and exogenous mus- cle-specific genes by Myf-5 and MyoD can be specifically repressed by PKA. The basic HLH domain as well as the DNA E-box motif to which these proteins bind are important for the inhibition. Both MyoD and Myf-5 are substrates for phosphorylation by PKA in vitro; however, this modification does not affect their ability to bind to the E-box sequence, suggesting that PKA interferes with transactivation at a step distal to DNA binding.

MATERIALS AND METHODS

Cell Cultures, Transfections, and Plasmid Constructs-C3H 10T1/ 2 fibroblasts and mouse C2 and rat L6 myoblasts were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. To induce differentiation cells were shifted to differentia- tion medium containing DMEM supplemented with 10% horse serum and 5 pg/ml insulin. Dibutyryl CAMP was used at 3 pmol/ml final concentration. Transfections were performed using calcium phos- phate precipitation as described previously (Braun et al., 1989~). In transient assays cells were shifted to differentiation medium 24 h after transfection and cultured for additional 48 h. Chloramp~lenicol acetyltransferase (CAT) and @-g~actosidase activities were deter- mined in cell extracts according to standard procedures (Gorman, 1985). All CAT activities were standardized to @-galactosidase ob- tained from 5 pg of cotransfected plasmid RSV-@gal.

Immunohistochemical Staining of Cells-Immunohistochemical staining of cells expressing sarcomeric myosin heavy chain (MHC) was performed with the monoclonal antibody MF-20 (Bader et al., 1982) and the Vectastain ABC Kit as described previously (Braun et al., 1989a).

For transactivation assays, 5 or 10 pg of the following reporter genes were used; MLC-CAT plasmid containing the muscle-specific enhancer of the rat myosin light chain 1/3 (MLC) gene and the proximal promoter of the MLCl gene linked to the CAT gene (Ro- senthal et al. 1990); 4R-tk CAT plasmid containing a multimerized

"...".__I

* E. Olson, personal communication.

E-box of the MLC enhancer upstream of the basal thymidine kinase itk) promoter linked to CAT (Weintrauh et al., 1990); Myf41-CAT plasmid containing a 1.1 kb 5' upstream fragment of the human myogenin (Myf-4) promoter linked to CAT (Salminen et al,, 1991); plasmid (E2+5) TATA-CAT containing six tandem copies of pE2 and pE5 sites fused to the alkaline phosphatase TATA box (Henthorn et aL, 1990); and plasmid pG5ElB-CAT containing five copies of the GAL4-responsive element linked to the adenovirus Elb TATA box (Martin et al., 1990). The following expression vectors were cotrans- fected at varying concentrations: pEMSV-MyoD and pEMSV-Myf5 expressing the mouse MyoD and human Myf-5 cDNAs, respectively (Davis et al., 1987; Braun et ai., 1989b); Myf5-VP16 and Myf5N10- VP16 expressing chimeric prot,eins between the VP16 transactivator domain (Cress and Triezenberg, 1991) and the Myf-5 open reading frame of amino acids 1-149 and Myf-5 amino acids 75-135, respec- tively (Braun et at., 1992; Winter et a[., 1992); GALmyf5 (amino acids 135-255) expressing a fusion protein between the GAL4 DNA binding domain (amino acids 1-147) and the carboxyl terminus of Myf-5 (amino acids 135-255) (Braun et al., 1990b; Winter et al., 1992); plasmid pSV2ABE2-5 expressing the E2-5 (E47) cDNA (Henthorn et al., 1990). The expression vectors for the cataiytic and regulatory subunits of PKA were MT-CEV,,, and MT-REVAB-neo, respectively (Uhler and McKnight, 1987). The control plasmids @-actin CAT and GAL-VP16 have been described previously (Lohse and Arnold, 1988; Martin et al., 1990).

Isolation of RNA and Northern Blot Analysis-RNA was isolated from tissue culture cells by the method described by Chomzynski and Sacchi (1987). Purified RNA was separated on denaturing agarose gels, transferred to a nylon membrane, and hybridized as described previously (Braun et al., 1989b). The following cDNA probes were used for specific hybridization: MyoD represented by the 0.8-kb ~paII/EcoRI fragment of the 3'-noncoding region of the mouse cDNA (Davis et a/., 1987); myogenin encoded by a 0.7-kb fragment of the rat cDNA as described previously (Sassoon et al., 1989); Myf-5 represented by a 0.3-kb ApaLl/BaZI genomic fragment encompassing part of the first exon of the mouse Myf-5 gene (Ott et ai., 1991; Bober et al., 1991); glyceraldehyde phosphate dehydrogenase containing 1.1 kb of the mouse glyceraldehyde phosphate dehydrogenase cDNA.

Purification of Bacterialb ExpressedpGEX-Myf5 andpGEX-MyoD Fusion Proteins-The construction of the vectors pGEX-Myf5 and pGEX-MyoD fusion proteins between glutathione S-transferase (GST) and Myf-5 or MyoD has been described previously (Braun et al., 1990a). The expression and purification were performed as re- ported by Smith and Johnson (1988) and Lassar et al. (1989). Purified protein preparations were stored in 20% glycerol a t -20 "C.

DNA Binding Assays and in Vitro Phospho~~ation-DNA binding of GST-Myf fusion proteins was measured by gel retardation assays as described (Braun et al., 1989c, 1991). Typical band shift experi- ments were performed with 500 ng of purified fusion proteins and 100 ng of poly(d1-dC) in a reaction volume of 30 pl. End-labeled oligonucleotide corresponding to the high affinity binding site of the MLC1/3 enhancer (GATCAAGTAACAGCAGGTGCAAAATAAAG T) was used as target DNA. The specificity of binding was assessed

AGTAAGTAACTGTGCAAATAAAGT). Protein-DNA complexes by competition with wild-type and mutant oligonucleotides (GATCA

were resolved on 5% nondenaturing polyacrylamide gels in 0.25 X TBE running buffer at 160 V for 5 h at 4 "C. In vitro phosphorylation of bacterially produced fusion proteins was done in 10 p1 of kinase reaction buffer containing 17.5 mM Tris-HC1, pH 7.4, 2,5 mM mag- nesium chloride, 40 p~ ATP plus 10 pCi of [Y-~'P]ATP (specific activity, 3000 Ci/mmol), and 500 ng of purified fusion protein. The reaction was st.arted by adding 10 units of PKA, catalytic subunit (Sigma fnc.) and incubated at 37 "C for 15 min.

Prepuru~~on o ~ N u c ~ a r E ~ t ~ a c t s ~ r o m Tissue Culture Cells-Nuclear extracts from undifferentiated and different.iated C2C12 cells were prepared from 20 dishes containing approximately 3 X lo6 cells each and processed as described previously (Braun et al., 1992). Briefly, cells were grown for 3 days in DMEM plus 10% fetal calf serum and 3 mM dibutyryl CAMP (Sigma) or in differentiation medium alone (DMEM, 10% horse serum, 4.5 g/liter glucose). Cells were washed twice in PBS, harvested, pelleted at 1400 rpm, and resuspended in 8 ml of buffer containing 10 mM HEPES, pH 7.9, 10 mM KC1, and a mix of protease inhibitors containing 0.5 mM phenylmethylsulfonyl fluoride, 1 pg/ml leupeptin, 1 pg/ml pepstatin, 1 pg/ml aprotinin, and 40 pg/ml bestatin. Cells were lysed with 0.7% Nonidet P-40. The nuclear pellet was resuspended in 0.8 ml of ice-cold buffer containing 20 mM NEPES, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, and protease inhibitor mix. Nuclei were extracted

Inhibition of Myogenesis by cAMP and PKA 9871

FIG. 1. cAMP inhibits differen- tiation of mouse C2 and rat L6 myo- blasts. Cells were cultured in differen- tiation medium with or without Bt2cAMP (3 pmol/ml) for 3 days, fixed, and immunostained with anti-MHC monoclonal antibody MF-20. Essentially all myotubes stained positive. The upper two panels show C2 cells, and the lower panels show L6 cells.

1 I L 1 myogenin

MyoD 1 @ 'ei kr, I I * * + Q a I Myf-5 1 2 3 4 5 6

FIG. 2. Northern blot analysis for MyoD, Myf-5, and my- ogenin mRNAs in C2 and L6 cells grown in the presence of BtZcAMP. Approximately 20 pg of total RNA isolated from cells that have been cultured for 3 days under the indicated conditions were analyzed as described under "Materials and Methods." Differ- entiated control RNA was obtained from cells cultured in differentia- tion medium for 8 days. All hybridization probes were labeled to a specific activity of 1 X 10' cpmlpg. Films were exposed overnight a t -80 "C using an intensifying screen.

by gently shaking on a rotatory platform for 15 min a t 4 "C. Insoluble material was removed by centrifugation for 10 min. The supernatant was adjusted to 25 mM HEPES, pH 7.9, 100 mM NaCl, 1 mM dithiothreitol, 0.2 mM EDTA, 20% glycerol and loaded onto a small heparin-Sepharose column. The column was washed with loading buffer, and proteins were eluted with the same buffer containing 0.8

TABLE I Myogenic conversion of IOT1/2 fibroblasts by MyoD and Myf-5 i s

inhibited by PKA

Plasmids" Myosin Inhibition positive cellsb by PKA

x pEMSV-a scribe 0

+PKA (10 pg) 36 74 +PKA (15 pg) 23 83

+PKA (10 pg) 16 70 +PKA (15 pg) 8 85

PEMSV-MYOD 136

pEMSV-Myf5 52

~. ~

Transfections were performed with 10 pg of activator plasmid and 10 or 15 pg of MT-CEVneo plasmids on 5 X 10" cells/plate.

*Number of cells per viewing field stained with the monoclonal antibody MF20. Numbers represent the mean of 30 fields counted for each transfection.

M NaCI. The eluate was dialyzed against 50 mM HEPES, pH 7.9, 50 mM NaCl, 1 mM dithiothreitol, 0.2 mM EDTA, 10% glycerol, and stored in aliquots a t -80 "C. 5-20 pg of nuclear extracts were used for mobility shift assays as described previously (Braun et al., 1991b).

RESULTS

cAMP Attenuates Myotube Formation of Mouse C2 and Rat L6 Myogenic Cell Lines-It has been reported previously (Hu and Olson, 1988) that cAMP analogs and compounds that increase intracellular cAMP levels inhibit the activation of muscle gene expression in the fusion-defective cell line BC3H1. To analyze the effect of elevated cAMP on the differentiation program of other myogenic cell lines, mouse C2 and rat .L6 myoblasts were cultured in medium which

9872 Inhibition of Myogenesis by CAMP and PKA A B

FIG. 3. The catalytic subunit of PKA inhibits transactivation of the muscle-specific reporter genes Myf41-CAT and MLC- CAT by Myf-5 and MyoD. A, 10T1/2 fibroblasts were cotransfected with 10 pg of Myf41-CAT reporter plasmid, '7.5 pg of pEMSV-Myf5 transactivator plasmid, and varying amounts of MT-CEV., plasmid expressing the catalytic subunit of PKA or MT-REVAB-neo plasmid expressing the regulatory subunit of PKA as specified. pRSV-Ela plasmid was used as a control for a known suppressor of Myf-5 transactivation (Braun et al., 1992). Myf41-CAT plasmid alone is inactive in 10T1/2 fibroblasts as shown previously (Salminen et al., 1991). Cells were cultured in differentiation medium for 48 h, and CAT activity was determined according to standard procedures. 5 pg of pRSV- Bgal plasmid was included in each transfection, and CAT values were standardized according to 0-galactosidase activity. B, cotransfections of 15 pg of MLC-CAT or &actin CAT reporter plasmids with 10 pg of pEMSV-Myf5 or pEMSV-MyoD transactivator plasmids were performed as described in A. The specified amounts of MT-CEV,, inhibitor plasmid were included where indicated. pEMSVa-2 plasmid containing no insert cDNA was used as control vector. CAT activities are given relative to the activity obtained in the absence of inhibitor plasmid. A representative result of three independent experiments is shown.

A B

reporter1 E5+E2. tkCAT

activator pSV2AB E24

inhibitor(ug)

FIG. 4. The inhibition by PKA is directed at the E-box DNA binding motif but does not affect the nonmyogenic bHLH protein E2-5 (E47). 10T1/2 fibroblasts were transfected with 5 pg of 4R-tkCAT ( A ) or (E2+5)-tkCAT ( R ) reporter plasmids and 10 pg of the indicated activator plasmids pEMSV-Myf5, pEMSV-MyoD, and pSV2AP-E2-5. The inhibitor plasmid MT-CEV., was included as specified. pEMSVa-2 was used as the vector control. The efficiency of each transfection was controlled by cotransfection of 5 pg of RSV- @gal, and all CAT values were calibrated accordingly. Columns indicate CAT activities relative to control transactivation (without inhibitor). The shown result is representative for three independently performed transfections.

Inhibition of Myogenesis by cAMP and PKA 9873

A

reporter MLC - CAT

activator p~3ISvMfl5 MflS-VPIB M y f S N I O - W I B

inhibitor(ug)

reporter I 1 pg GSEIB-CAT I

e

FIG. 5. The basic HLH domain of Myf-5 constitutes the target for the repression by PKA. A, transactivation of 5 pg of MLC- CAT reporter plasmid by pEMSV-Myf5 (10 pg), Myf5-VP16 (1 pg), and Myf5NlO-VP16 (0.5 pg) was measured in transient transfections of 10T1/2 fibroblasts. The inhibitor plasmid MT-CEV,, was cotransfected as indicated. B, transactivation of 1 pg of G5ElB-CAT reporter plasmid by 1 pg of GALmyf5 (amino acids 135-255) transactivator in the presence of varying amounts of MT-CEV., or MT-REVAB-neo plasmids expressing the catalytic and regulatory subunits of PKA, respectively. The experiments were performed as described in the legend to Fig. 3 and under “Materials and Methods.” The result is representative for three independent transfections.

supports differentiation in the absence and presence of dibu- tyryl cAMP (Bt2cAMP). Although both cell lines in medium differentiation (DM) readily fused into multinucleated myo- tubes, a marked reduction of morphological differentiation was observed in the presence of Bt2cAMP (Fig. 1). The expression of several muscle-specific marker genes, such as MHC measured by immunostaining with the monoclonal antibody MF20, muscle creatine kinase, myosin light chains, and sarcomeric actin was also suppressed (data not shown). At the applied concentration of Bt2cAMP, the cells remained firmly attached to the substratum and appeared morphologi- cally normal with no signs for general toxicity. We therefore conclude that the differentiation of skeletal muscle cells is subject to specific negative regulation by the cAMP signal transduction pathway.

Myoblasts in the Presence of Bt2cAMP Express Normal Levels of Myf-5 and MyoD but No Myogenin mRNA-Since cAMP inhibits not only the expression of single muscle- specific genes but seems to affect the complete myogenic differentiation program, we examined the expression of the myogenic regulatory genes of the bHLH family of proteins which are believed to be essential for establishing the muscle phenotype. C2 and L6 myoblasts are known to constitutively express MyoD and Myf-5, respectively, and accumulate my- ogenin at the onset of differentiation (Braun et al., 1989b). Therefore, RNA from both cell lines grown in DM medium in the absence or presence of Bt2cAMP was analyzed on Northern blots using cDNA hybridization probes specific for Myf-5, MyoD, and myogenin. As shown in Fig. 2, expression of Myf-5 and MyoD mRNA was essentially unaltered by BtZcAMP, whereas the expression of myogenin mRNA was almost completely inhibited in both cell lines. These results suggest that the expression of the various myogenic bHLH

genes is differentially sensitive to CAMP-dependent regula- tion. Although transcription of the MyoD and Myf-5 genes was not affected by CAMP, expression of the myogenin gene was nearly totally suppressed.

PKA Can Substitute for CAMP in the Inhibition of Myogen- esis and Represses Transactivation by Myf-5 and MyoD- Many effects of cAMP on gene regulation are mediated through CAMP-dependent PKA. To investigate whether PKA would be involved in the mechanism by which cAMP inhibits muscle differentiation, we overexpressed the catalytic subunit of PKA in 10T1/2 fibroblasts and examined its effect on myogenic conversion by MyoD or Myf-5. Transient transfec- tion of the expression plasmids pEMSV-MyoD and pEMSV- Myf5 converted a substantial number of 10T1/2 cells into myosin-positive myoblasts, but cotransfection of the vector MT-CEVneo expressing the catalytic subunit of PKA (Uhler and McKnight, 1987) resulted in a significant reduction of myosin containing cells as measured by immunostaining with the monoclonal antibody MF-20 which recognizes sarcomeric MHC (Table I). Comparable inhibition was also observed with 10 independently derived C2 clones which had been stably transfected with the PKA expressing vector MT-CEV- neo (data not shown). These observations indicate that PKA can substitute for cAMP in the repression of muscle cell differentiation, suggesting that PKA may mediate the inhib- itory effect of CAMP.

It has been shown that some of the oncogene products and external mitogenic stimuli inhibit the activity of constitu- tively expressed myogenic factors (from viral LTR promotors) by a post-translational mechanism without affecting their synthesis (Vaidya et al., 1989; Lassar et al., 1989; Li et al., 1992a; Bengal et al., 1992; Braun et al., 1992). In order to examine whether PKA would also interfere with the trans-

9874 + PKA

Inhibition of Myogenesis by CAMP and PKA

FIG. 6. In vitro phosphorylation of bacterially produced pGEX-MyoD and pGEX-Myf5 by PKA. 500 ng of affinity-puri- fied pGEX fusion proteins were phosphorylated in vitro with ["PI ATP and PKA as described under "Materials and Methods." 10-pg aliquots were separated on 10% SDS-polyacrylamide gel and exposed on x-ray film overnight. The positions of the chimeric proteins were determined by Coomassie Blue staining and are indicated by arrow- heads. pGEX-Myf5 contains a faster migrating degradation product. The weaker bands are primarily due to residual bacterial proteins. The glutathione S-transferase (pGEX) alone is not phosphorylated.

activation function of constitutively expressed Myf-5 or MyoD, we cotransfected 10T1/2 fibroblasts with the transac- tivator plasmid pEMSV-Myf5 together with the reporter plas- mid Myf41-CAT containing the muscle-specific promoter se- quence of the human myogenin gene linked to the CAT gene (Salminen et al., 1991) and the PKA-expressing plasmid MT- CEVneo. The Myf41-CAT reporter plasmid is silent in 10T1/ 2 cells but can be transactivated by overexpression of Myf-5 (Salminen et al., 1991). This transactivation was markedly suppressed when increasing amounts of the catalytic subunit of PKA were coexpressed (Fig. 3A) . We also transfected the cells with plasmid MT-REVA~."- expressing the regulatory subunit of PKA as a convenient vector control, which had no effect on transactivation by Myf-5. When regulatory and catalytic subunits were coexpresssed, most of the inhibition exerted by the catalytic PKA subunit alone was reversed, which we interpret as an indication that the inhibition was specifically dependent on PKA activity. Similar transactiva- tion experiments were also performed with the MLC-CAT reporter gene containing the muscle-specific enhancer of the myosin light chain 1/3 gene (Rosenthal et al., 1990). As shown in Fig. 3B, transactivation of the MLC enhancer by Myf-5 and MyoD was also strongly suppressed by PKA. Other reporter genes that are constitutively expressed in 10T1/2 fibroblasts, such as the cytoplasmic @-actin promoter driving CAT (@-actin-CAT) or RSV-CAT (data not shown), were not inhibited by PKA. To rule out the possibility that the repres-

sion by PKA was due to a direct inhibition of the MSV-LTR promoter driving the expression of the transactivators MyoD and Myf-5, we determined the accumulation of both transac- tivators in the presence and absence of PKA. No decrease in levels of MyoD and Myf-5 mRNAs were observed in PKA expressing cells (data not shown). These results indicate that PKA specifically prevents the activation of genes that are dependent on myogenic bHLH factors but not the transcrip- tion of constitutively active promoters such as the @-actin promoter.

P K A Inhibition of Muscle-specific Transactiuation by Myf- 5 Is Directed at the E-box but Does Not Affect Other E-box Binding Factors Such as the E2A Gene Products-The acti- vation of complex muscle-specific promoters and enhancers by myogenic bHLH proteins is dependent on the consensus DNA binding motif CANNTG, referred to as E-box, but also involves other cis-acting DNA elements to which cooperating transcription factors may bind. In order to analyze whether the inhibition by PKA would be targeted at the E-box element, we used the CAT reporter gene 4R-tkCAT, which is solely controlled by the basal thymidine kinase (tk) promoter and four copies of the E-box motif derived from the muscle crea- tine kinase enhancer (Buskin and Hauschka, 1989; Weintraub et al., 1990). Similar to the more complex control elements of the myogenin promoter and the MLC enhancer, transactiva- tion of this test gene by pEMSV-Myf5 or pEMSV-MyoD in 10T1/2 fibroblasts was also inhibited by PKA (Fig. 4A). The regulatory subunit of PKA had no effect on transactivation and reversed the repression by the catalytic subunit (data not shown). From these results we conclude that transcriptional repression by PKA is directed at the muscle-specific E-box DNA binding motif. Whether other cis-acting elements pres- ent in the more complex muscle-specific enhancers may be involved additionally remains possible.

The widely expressed bHLH gene products of the E2A gene, E12 and E47 or E2-5, are part of the active muscle- specific DNA binding complex (Lassar et al., 1991). They are required for high affinity DNA binding of the myogenic bHLH factors and enhance their transactivating function (Winter et al., 1992). Although both of these E2A proteins contain tran- sactivator domains, they are unable to activate muscle-spe- cific genes, but readily activate genes containing the distinct pE-boxes present in the immunoglobulin heavy chain en- hancer (Henthorn et al., 1990). In order to investigate whether PKA might also repress transactivation by the E2A gene products, we cotransfected 10T1/2 cells with the plasmid pSV2AP E2-5 expressing the E2-5 transactivator protein, the CAT reporter plasmid (E2+5)-TATA-CAT containing the cognate E-boxes pE2 and pE5, and the PKA-expressing vector MT-CEVneo. As shown in Fig. 4B, transactivation by E2-5 was not inhibited by the catalytic subunit of PKA, in marked contrast to the results obtained with the muscle-specific bHLH proteins. Similarly, no inhibition was observed when E12 was used as a transcriptional activator (data not shown). We therefore conclude that PKA selectively represses the transactivating function of the myogenic bHLH proteins MyoD and Myf-5 but does not inhibit their dimerization partners E2-5 or E12.

Repression of Transactivation by P K A Is Targeted at the bHLH Domain of Myf-5"To determine which part of the Myf-5 molecule was necessary and sufficient to serve as target for the inhibition by PKA, fusion proteins were constructed in which the transactivator domains of Myf-5 were substituted by the VP-16 activator domain. The construct Myf5-VP16 was derived by replacing sequences encoding the COOH- terminal Myf-5 transactivator domain (amino acids 150-255)

A

Inhibition of Myogenesis by CAMP and PKA

B 9875

PRCYI'EIK

PHOSPH

pCM-MyoD

~GnU-Myf5

C E12 + E12 +

pCEX-MyG pCEX-MyoD

PHOSPH. [ - + - + I

E12KUyoD

retie. lyute

FIG. 7. Sequence-specific DNA binding of in vitro phosphorylated pGEX-MyoD and pGEX-Myf5. A , Gel mobility shift assays (EMSA) were performed with 500 ng of unphosphorylated pGEX-Myf5 protein and pGEX-Myf5 protein phosphorylated in vitro by PKA. Synthetic oligonucleotide encompassing the high affinity E-box of the MLC1/3 enhancer was used as the DNA binding site (for sequence, see "Materials and Methods"). Complexes formed with unphosphorylated control protein in binding buffer (-) or in phosphorylation buffer (buffer) and with phosphorylated protein (phosph.) are shown. The same experiment performed with pGEX-MyoD is shown in lanes 4-6. B, EMSA performed with "P-labeled proteins and unlabeled oligonucleotide. Only phosphorylated complexes can be detected in this assay. The positions of the pGEX-Myf5 and pGEX-MyoD complexes were determined in parallel with labeled oligonucleotides as indicated. Unbound proteins migrated faster on the gel (data not shown). C, EMSA performed with ElZ/pGEX-MyoD and E12/pGEX-Myf5 heterodimeric complexes on the E-box oligonucleotide. 50 ng of unphosphorylated (-) or PKA phosphorylated (+) fusion proteins were mixed with E12 protein produced in reticulocyte lysate (10 pl) as described previously (Braun and Arnold, 1991). Only parts of the gels containing the relevant complexes are shown. The faster moving Myod/E12 complex was seen inconsistently and may be a degradation product during the phosphorylation reaction.

by the VP16 activator region as described previously (Braun et al., 1992). Plasmid Myf5NlO-VP16 was derived by deletion of the NH2-terminal transactivator region (amino acids 14- 74) from Myf5-VP16 (Winter et al., 1992). Both of these Myf5-VP16 chimeric proteins retained the bHLH region of Myf-5 and proved to be strong transcriptional activators of the endogenous MHC gene (Braun et al., 1992) and the muscle-specific reporter construct MLC-CAT in 10T1/2 fi- broblasts. Coexpression of increasing concentrations of MT- CEVneo plasmid expressing the catalytic subunit of PKA resulted in a severe repression of MLC-CAT activation (Fig. 5 A ) . In a control experiment, we showed that the GAL4-VP16 transactivator tested on the appropriate reporter plasmid G5ElB-CAT was not affected by PKA, indicating that the VP16 transactivator domain itself is not inhibited by PKA (data not shown). These results then suggested that within the Myf-5 molecule the bHLH domain was sufficient to mediate the inhibitory effect of PKA.

To control whether the bHLH region of Myf-5 was also required, we replaced this motif by the DNA binding domain of the yeast transcription factor GAL4 (GALmyf5). Transfec- tion of the plasmid expressing the GALmyf5 fusion protein

together with G5ElB-CAT reporter plasmid into 10T1/2 fi- broblasts showed strong activation, which was not inhibited by coexpression of the catalytic or regulatory subunits of PKA. From these results we conclude that the transactivator domain of Myf-5 in conjunction with a heterologous DNA binding domain constitutes no target for the repression by PKA, whereas the Myf-5 bHLH region can confer this regu- lation onto a heterologous transactivator, such as VP16. Therefore, the bHLH domain represents the primary target for the inhibitory action of PKA.

MyoD and Myf-5 Can Be Phosphorylated by PKA in Vitro but This Does Not Affect Their Ability to Bind to DNA-It has been reported previously that the myogenic bHLH factors are nuclear phosphoproteins (Tapscott et al., 1988; Brennan and Olson, 1990). To examine whether MyoD and Myf-5 might serve as substrates for phosphorylation by PKA in uitro, we incubated bacterially produced fusion proteins be- tween glutathione S-transferase and human MyoD (pGEX- MyoD) or Myf-5 (pGEX-Myf-5) with PKA enzyme and ["PI ATP and separated the reaction products on SDS-polyacryl- amide gels. As shown in Fig. 6, both chimeric proteins pGEX- MyoD and pGEX-Myf-5 were highly phosphorylated, whereas

9876 Inhibition of Myogenesis by cAMP and PKA

c2c12 c2c12 +db-cA" differ. extract:

competitor: ( x-fold)

Mef 1 Mef-Mut Mef 1 Mef-Mut - 100 200 loo 200 - 100 200 100 200

FIG. 8. Gel mobility shift assay performed with nuclear extracts from differentiated and BtZcAMP- treated C2 muscle cells. Nuclear ex- tracts were prepared as described under "Materials and Methods" and used for EMSA on the MLC oligonucleotide en- compassing the high affinity E-box. Wild-type and mutant oligonucleotides carrying a mutated E-box were used as competitors as specified. Complexes C2a and C2b correspond to MEF-1 binding (bHLH proteins), whereas C1 corre- sponds to the putative MEF-2 complex as determined by specific competition (data not shown).

- c1

- C2a

- C2b

unbound

the glutathione S-transferase protein alone (pGEX) was not. We observed frequently a degradation product of pGEX-Myf5 which was also phosphorylated. Because the inhibition of MyoD and Myf-5 activity by PKA was directed at the bHLH region and was mediated through the E-box DNA sequence, we next investigated whether protein phosphorylation might alter the DNA binding properties. Electrophoretic mobility shift assays (EMSA) using control and in vitro phosphoryl- ated pGEX-MyoD and pGEX-Myf5 proteins on a synthetic oligonucleotide containing the muscle creatine kinase E-box revealed similar DNA binding between control proteins and their phosphorylated counterparts (Fig. 7 A ) . The same result was obtained a t protein concentrations ranging from 200 to 800 ng (data not shown). Although we estimated from the amount of radioactive phosphate residues incorporated into the proteins that the majority of molecules had been phos- phorylated and therefore the observed DNA binding should not be due to remaining unphosphorylated protein, we wanted to confirm this in EMSAs using 32P-labeled protein and unlabeled oligonucleotides. In this experiment, a band shift can only be observed when the phosphorylated protein species actually binds to the DNA. As shown in Fig. 7B, both labeled pGEX-MyoD and pGEX-Myf5 proteins formed DNA com- plexes that migrated to the same position on the gel as the control complexes of unlabeled fusion proteins and radioac- tively labeled oligonucleotide but slower than free proteins. These results suggest that phosphorylation of MyoD and Myf- 5 by PKA in vitro does not prevent sequence-specific DNA binding.

Similar experiments performed in the presence of in vitro synthesized E12 which favors the binding of heterodimers also revealed no alterations in the DNA binding capacity of phosphorylated MyoD and Myf-5 (Fig. 7 C ) . Nuclear extracts prepared from differentiated control and Bt2cAMP-treated C2 cells were tested for binding at an oligonucleotide encom- passing the high affinity E-box of the myosin light chain enhancer, including the adjacent MEF-2 binding site (Rosen- thal et al., 1990). As shown in Fig. 8, nuclear extracts from

control and BtncAMP-treated cells formed similar complexes in EMSAs, suggesting that DNA binding of the cognate transcription factors was not affected by CAMP. Because the shifted complexes were successfully competed by excess of wild-type oligonucleotide, but not by oligonucleotide carrying a mutated E-box, we believe that they correspond to bHLH factors. This result provides preliminary evidence that ele- vated intracellular levels of CAMP, which presumably activate phosphorylation by PKA, probably do not interfere with DNA binding of HLH proteins in muscle cell lines.

DISCUSSION

CAMP-dependent Protein Kinase Conveys Inhibitory Sig- nals to the Myogenic Regulators Myf-5 and MyoD-Serum and peptide growth factors have long been known to suppress differentiation of muscle cells in culture and prevent the activation of muscle-specific genes (reviewed by Florini et al. (1991)). I t was also shown that MyoD, as well as myogenin and Myf-5, are phosphorylated in proliferating myoblasts (Tapscott et al., 1988; Brennan and Olson, 1990). In an attempt to identify intracellular pathways that may transduce signals generated at the cell surface to the nucleus, we exam- ined the role of cAMP and PKA during myogenesis. Our results demonstrate that stimulation of the PKA pathway inhibits myoblast differentiation and prevents transcriptional activation of muscle-specific genes by the myogenic activators Myf-5 and MyoD. The repression of Myf-5 and MyoD func- tion by PKA is mediated through a post-transcriptional mech- anism that is directed at the bHLH domain of the myogenic factors. Other bHLH proteins such as the E2A gene products, E12 and E2-5 (E47), are not inhibited by PKA, indicating that the effect is specific for myogenic bHLH proteins. Al- though the repression is mediated through the E-box DNA binding site, phosphorylation of Myf-5 or MyoD by PKA in vitro does not affect sequence-specific DNA binding. We therefore conclude that some other events necessary for trans- activation are inhibited. Whether the repression of muscle differentiation is a direct result of Myf-5 or MyoD phos-

Inhibition of Myogenesis by CAMP and PKA 9877

phorylation remains to be determined. Taken together, our observations provide evidence that members of the MYOD family of regulatory proteins constitute effective nuclear tar- gets for the control by which CAMP-dependent pathways attenuate muscle differentiation. Although we do not know how the levels of cAMP and PKA might change during differentiation and how this may affect the phosphorylation of the myogenic factors in uiuo, PKA conceivably participates in regulatory networks that convey inhibitory signals in myo- blasts to the important myogenic control factors present in the nucleus. Whether high PKA activity may also partly account for the poor responsiveness of certain cell types to the myogenic conversion by the muscle regulatory bHLH proteins (Schiifer et al., 1990; Weintraub et al., 1989) is currently under investigation.

Possible Mechanisms for the Repression of Myogenesis by PKA-What types of mechanisms may account for the inhi- bition of Myf-5 and MyoD function by PKA? In most cases in which phosphorylation by PKA affects transactivation, the exerted control is positive. The only example to date in which transactivation is inhibited by phosphorylation relates to the yeast transcription factor ADRl which enhances the expres- sion of the ADHZ gene via an upstream activating sequence (UAS) (Cherry et al., 1989). When ADRl is phosphorylated by PKA, its ability to interact with the general transcriptional machinery appears to be abolished, whereas it still binds to DNA normally (Taylor and Young, 1990). Similar to ADR1, PKA also suppresses transactivation by Myf-5 and MyoD without affecting their ability to bind to DNA. Significantly, however, the inhibition of Myf-5 activity does not require the transactivator domains which would be expected to be the candidate regions for interactions with the basal transcrip- tional machinery. Although we have shown that Myf-5 and MyoD can be phosphorylated by PKA in vitro, we have no evidence that direct phosphorylation of the bHLH proteins causes the block in transactivation. It is equally possible that proteins which interact with Myf-5 or MyoD for transcrip- tional activation constitute the critical targets for PKA. In fact, Li et al. (1992b) have recently demonstrated by muta- genesis of one PKA site in the basic region of myogenin that phosphorylation of this site is not important for the inhibition of myogenin activity by PKA. It is unlikely, however, that the known dimerization partners E12 or E47 mediate the inhibi- tory effect as their transactivating capacity was not affected by PKA. Inspection of the primary amino acid sequence within the basic HLH domain of Myf-5 that is required and sufficient for the repression by PKA revealed two regions that resemble the consensus sequence for phosphorylation by PKA. These putative PKA sites are located outside of the basic clusters that presumably mediate DNA contacts but are contained within a sequence that is highly conserved among the myogenic bHLH proteins and therefore may serve some fundamental function. Mutational analysis to assess the im- portance of the putative phosphorylation sites in Myf-5 is in progress. It will be interesting to see whether the loss of Myf- 5 activity might be caused by phosphorylation-induced con- formational changes or by interference with additional factors that may be necessary as coactivators of Myf-5.

An alternative scenario for the repression of myogenesis by PKA might be that PKA induces the expression of one or several negative regulators of muscle-specific transcription. Indeed, it has been shown that cAMP activates the expression of the immediate early response genes c-fos and c-jun which both inhibit myogenesis and the function of myogenic bHLH proteins (Lassar et ul., 1989; Bengal et al., 1992; Li et al., 1992a). Although the exact mechanism whereby these factors

interfere with the activity of the myogenic regulators has not been elucidated, it is clear that the repression requires the bHLH domain and is directed at the E-box (Li et al., 1992a). Therefore, Fos and Jun or both proteins conceivably are candidates for effectors of PKA-induced inhibition of myo- genesis. Whether the cAMP response element-binding protein (CREB) known to be activated by PKA-dependent phos- phorylation may be involved in the induction of negative regulators remains a possibility. However, CREB cannot play a direct role in the repression of muscle-specific gene activa- tion, because it does not bind to the E-box sequence that is sufficient to mediate the inhibition by PKA.

The results of this study identify one of numerous signal transduction pathways that regulate the activity of the im- portant myogenic HLH transcription factors whereby the differentiated phenotype of muscle cells can be modulated. Although we have not yet defined the precise events by which PKA inhibits myogenesis, we have provided evidence for a cellular control component that can regulate the transacti- vation capacity of members of the MyoD family of proteins at the post-translational level. We suggest a mechanism which probably acts at a step downstream of DNA binding.

Acknowledgments-We thank T. Kadesh for providing the plas- mids (E2+5)TATA-CAT and pSV2APE2+5, and G. S. McKnight for the plasmids MT-CEV,, and MT-REVAB-neo. We acknowledge the expert technical assistance by A. Gaser and H. Eberhardt, and the secretarial help by A. Broecker-Nagel.

REFERENCES Bader, D., Basak, T., and Fishmann, D. A. (1982) J. Cell Biol. 9 5 , 763-770 Bengal, E., Ransone, L., Scharfmann, R., Dwarki, V. J., Tapscott, S. J.,

Bober, E., Lyons, G. E., Braun, T., Cossu, G., Buckingham, M., and Arnold, H.

Braun, T., and Arnold, H. H. (1991) Nucleic Acids Res. 19,5645-5651 Braun, T., Buschhausen-Denker, G., Bober, E., Tannich, E., and Arnold, H. H.

Braun, T., Bober, E., Buschhausen-Denker, G., Kohtz, S., Grzeschik, K. H.,

Braun, T., Tannich, E., Buschhausen-Denker, G., and Arnold, H. H. (1989~)

Braun, T., Bober, E., Winter, B., Rosenthal, N., and Arnold, H. H. (1990a)

Braun, T., Winter, B., Bober, E., and Arnold, H. H. (1990b) Nature 4 6 , 663-

Braun, T., Gearing, K., Wright, W. E., and Arnold, H. H. (1991) Eur. J.

Braun, T., Bober, E., and Arnold, H. H. (1992) Genes & Deu. 6,888-902 Brennan, T. J., and Olson, E. N. (1990) Genes & Deu. 4,582-595 Brennan, T. J., Chakraborty, T., Olson, E. N. (1991) Proc. Natl. Acad. Sci.

Buskin, J. N., and Hauschka, S. D. (1989) Mol. Cell. Biol. 9,2627-2640 Cherry, J. R., Johnson, T. R., Dollard, C., Shuster, J. R., and Denis, C. L.

Chomczynski, P., and Sacchi, N. (1987) Anal. Bhchem. 162,156-159 Clegg, C. H., Correll, L. A., Cadd, G. G., and McKnight, G. S. (1987) J. Biol.

Cress, W. D., and Triezenberg, S. J. (1991) Science 251,87-90 Davis, R. L., Weintraub, H., and Lassar, A. B. (1987) Cell 51,987-1000 Davis, R. L, Cheng, P. F., Lassar, A. B., and Weintraub, H. (1990) Cell 60,

Edmondson, D. G., and Olson, E. N. (1989) Genes & Deu. 3,628-640 Emerson, C. P. (1990) Curr. Opin. Cell Biol. 2 , 1065-1075 Enkemann, S. A., Konieczny, S. F., and Taparowsky, E. J. (1990) Cell Growth

Falcone, G., Tatao, F., and Alema, S. (1985) Proc. Natl. Acad. Sci. U. S. A. 8 2 ,

Florini, J. R., Ewton, D. 2.. and Magri, K. A. (1991) Annu. Reu. Physiol. 5 3 ,

Gorman, C. (1985) in DNA Cloning ZI (Glover, D. M., ed) pp. 143-190, IRL

Henthorn, P., Kiledjian, M., and Kadesch, T. (1990) Science 2 4 7 , 467-470 Hu, J.-S., and Olson, E. N. (1988) J. Biol. Chem. 2 6 3 , 19670-19677

Lassar, A. B., Thayer, M. J., Overell, R. W., and Weintraub, H. (1989) Cell 5 8 , Kemp, B. E., and Pearson, R. B. (1990) Trends Biochem. Sci. 15 , 342-346

Lamar, A. B., Davis, R. L., Wright, W. E., Kadesch, T., Murre, C., Voronova,

Li, L., Chambard, J. C., Karin, M., and Olson, E. N. (1992a) Genes & Deu. 6,

Li, L., Heller-Harrisson, R., Czech, M., and Olson, E. N. (1992b) Mol. Cell.

Lohse, P., and Arnold, H. H. (1988) Nucleic Acids Res. 1 6 , 2787-2803 Martin, K. J., Lillie, J. W., and Green, M. R. (1990) Nature 346 , 147-152

Weintraub, H., and Verma, I. M. (1992) Cell 68,507-519

H. (1991) J. Cell Biol. 6,1255-1265

(1989a) EMBO J. 8 , 701-709

and Arnold, H. H. (1989b) EMBO J. 8,3617-3625

Mol. Cell. Biol. 9 , 2513-2525

EMBO J. 9,821-831

665

Biochem. 198 , 187-193

U. S. A. 8 8 , 5675-5679

(1989) Cell 56,409-419

Chern. 2 6 2 , 13111-13119

733-746

& Differ. 1,375-382

426-430

201-216

Press, Oxford

659-667

A., Baltimore, D., and Weintraub, H. (1991) Cell 66, 305-315

676-689

Bwl. 12,4478-4485

9878 Inhibition of Myogenesis by CAMP and PKA Miner, H. J., and Wold, J. B. (1990) Proc. Nutl. Acud. Sci. U. S. A. 8 7 , 1089-

Miner, H. J., and Wold, J. B. (1991) Mol. Cell. Biol. 11 , 2842-2851 Murre, C., McCaw, P. S., Vaessin, H., Caudy, M., Jan, L. Y., Jan, Y. N.,

Cabrera, C. V., Buskin, J. N., Hauschka, S. D., Lassar, A. B. Weintraub, H., and Baltimore, D. (1989) Cell 68,537-544

Olson, E. N. (1990) Genes & Deu. 4 , 1454-1461 Olson, E. N., Spizz, G., and Tainsky, M. A. (1987) Mol. Cell. Biol. 7, 2104-2111 Ott, M.-O., Bober, E., Lyons, G., Arnold, H. H., and Buckingham, M. (1991)

Pa ne, P. A , Olson, E. N., Hsiau, P., Roberts, R., Perryman, M. B., and

Rahm, M., Jin, P., Sumegir, J., and Sejersen, T. (1989) J. Cell. Physiol. 139 ,

Rhodes, S. J., and Konieczny, S. F. (1989) Genes & Deu. 3,2050-2061 Rosenthal, N., Berglund, E. B., Wentworth, B. M., Donoghue, M., Winter, B.,

Bober, E., Braun, T., and Arnold, H. H. (1990) Nucleic Acids Res. 18,6239-

1093

Development (Camb.) 111 , 1097-1107

&hneider,'M. D. (1987) Proc. Nutl. Acud. Sci. U. S. A. 84,8956-8960

237-244

Salminen, A,, Braun, T., Jiirs, S., Winter, B., and Arnold, H. H. (1991) J. Cell 6245

Biol. 116.905-917 ~~~ ~~~

Sassoon, D., Lyons, G., Wright, W. E., Lin, V., Lassar, A,, Weintraub, H., and

Schiifer, B. W., Blakely, B. T., Darlington, G. J., and Blau, H. M. (1990) Nature

I ~~~ ~~

Buckingham, M. (1989) Nature 341,303-307

344,454-458

Schneider, M. J., Perryman, M. B., Payne, P. A,, Spizz, G., Roberts, R., and

Schwartz, J., Chakraborty, T., Martin, J., Zhou, J., and Olson, E. N. (1992)

Smith, D. B., and Johnson, K. S. (1988) Gene (Amst.) 67,31-40 Tapscott, S. J., Davis, R. L., Thayer, M. J., Cheng, P.-F., Weintraub, H., and

Taylor, W. E., and Young, E. T. (1990) Proc. Nutl. Acud. Sci. U. S. A. 87,4098-

Vaidya, T. B., Rhodes, S. J., Taparowsky, E. J., and Konieczny, S. F. (1989) Uhler, M. D., and McKnight, G. S. (1987) J. Bid. Chem. 262,15202-15207

Webster, K. A., Muscat, G. E. O., and Kedes, L. (1988) Nature 337,553-557 Weintraub, H., Tapscott, S. J., Davis, R. L., Tayer, M. J., Adam, M. A,, Lassar,

Weintraub, H., Davis, R., Lockshon, D., and Lassar, A,, (1990) Proc. Nutl. A. B., and Miller, A. D. (1989) Proc. Nutl. A d . Sci. U. S. A. 86,5434-5438

Weintraub, H., Davis, R., Tapscott, S., Thayer, M., Krause, M., Benezra, R., Acud. Sci. U. S. A. 87,5623-5627

Blackwell, T. K., Turner, D., Rupp, R., Hollenberg, S. Zhuang, Y., and Lassar, A. (1991a) Science 2 5 1 , 761-766

Weintraub, H., Dwarki, V. J., Verma, I., Davis, R., Hollenberg, S., Snider, L., Lassar, A,, and Tapscott, S. J. (1991b) Genes & Deu. 5 , 1377-1386

Winter, B., Braun, T., and Arnold, H.-H. (1992) EMBO J. 11 , 1843-1855 Wright, W. E., Sassoon, D. A., and Lin, V. K. (1989) Cell 56,607-617

Olson, E. N. (1987) Mol. Cell. Biol. 7 , 1973-1977

Mol. Cell. Biol. 1 2 , 266-275

Lassar, A. B. (1988) Science 242,405-411

4102

Mol. Cell. Biol. 9, 3576-3579