AP-1 Binds to a Putative CAMP Response Element of the MyoD1 Promoter and Negatively Modulates MyoD1 Expression in Dividing Myoblasts*

We have studied the transcriptional activity of the mouse MyoDl gene promoter in vivo and in vitro using mouse G8 myoblasts and muscle cell nuclear extracts. 5‘ deletion analysis of the promoter and transcription-com-petition analysis using oligonucleotides corresponding to several cis-acting elements revealed that the basal activity of the MyoDl promoter is conferred by two SP1 boxes, an AP-2 box, and a CAAT box. We have identified a negative regulatory sequence located between nucleo- tide position -342 to -322 with respect to the cap site. The negative regulatory element shows sequence homology with CAMP-responsive element (CRE) and AP-1 binding site (S’-GAGCACTGAGGTCAGTACAG-3’). As determined by gel mobility shift competition analysis, oligonucleotides containing AP-1 binding sites inhibit protein inter- actions with the MyoDl CRE-like element. We also show that binding to this element is down-regulated during myogenic differentiation and can be reinduced by the addition of serum. Furthermore, mutation of the CRE- like element induces MyoD promoter activity in diving myoblasts. By using anti-c-Fos antibodies we show that AP-1 is binding to the MyoDl CRE-like element. Our results indicate that AP-1 negatively modulates MyoDl ex- pression in growing myoblasts and strongly suggest that c-Fos and c-Jun inhibit myogenesis and MyoDl expres- sion by direct binding to a negative


AP-1 Binds to a Putative CAMP Response Element of the MyoD1 Promoter and Negatively Modulates MyoD1 Expression in Dividing Myoblasts*
Gustavo Pedraza-Alva, Jean-Marc Zingg, and Jean-Pierre JostS From the Friedrich Miescher Institute, l? 0. Box 2543, Base1 CH-4002, Switzerland We have studied the transcriptional activity of the mouse MyoDl gene promoter in vivo and in vitro using mouse G8 myoblasts and muscle cell nuclear extracts. 5' deletion analysis of the promoter and transcription-competition analysis using oligonucleotides corresponding to several cis-acting elements revealed that the basal activity of the MyoDl promoter is conferred by two SP1 boxes, an AP-2 box, and a CAAT box. We have identified a negative regulatory sequence located between nucleotide position -342 to -322 with respect to the cap site. The negative regulatory element shows sequence homology with CAMP-responsive element (CRE) and AP-1 binding site (S'-GAGCACTGAGGTCAGTACAG-3'). As determined by gel mobility shift competition analysis, oligonucleotides containing AP-1 binding sites inhibit protein interactions with the MyoDl CRE-like element. We also show that binding to this element is down-regulated during myogenic differentiation and can be reinduced by the addition of serum. Furthermore, mutation of the CRElike element induces MyoD promoter activity in diving myoblasts. By using anti-c-Fos antibodies we show that AP-1 is binding to the MyoDl CRE-like element. Our results indicate that AP-1 negatively modulates MyoDl expression in growing myoblasts and strongly suggest that c-Fos and c-Jun inhibit myogenesis and MyoDl expression by direct binding to a negative cis-acting element in the MyoDl promoter.
Muscle cell differentiation is presumably the end-point of a cascade of intracellular events involving progenitor cell determination to the myogenic lineage, multiplication and withdrawal from the cell cycle of the myogenic precursor cells, and terminal differentiation and modulation of the terminally differentiated state by developmental and physiological signals. This complex and multistep process appear to be directed by a hierarchy of regulatory genes. Several genes whose protein products are potentially responsible for the determination of the myogenic phenotype have been characterized, including MyoD1, MRF4, Myogenin, and Myf5 Pinney et al., 1988;Edmonson and Olson, 1989;Wright et al., 1989). Each of these muscle-specific regulatory factors was shown to convert 10T1/2 fibroblasts to the muscle phenotype. Interestingly, all of these proteins share a putative helix-loop-helix domain, which is also present in the myc proto-oncogene family Caoudy et al., 1988). The best characterized member of the myogenic factors is MyoD1, which is a nuclear payment of page charges. This article must therefore be hereby marked * The costs of publication of this article were defrayed in part by the "advertisement" in accordance with 18 U.S.C. Section 1734 solelv to indicate this fact. t To whom correspondence should be addressed. "el.: 061-6976688; Fax: 061-6973976.
phosphoprotein (Tapscott et al., 19881, binds as a homo-or heterodimer to the consensus sequence CANNTG (Davis et al., 19901, and trans-activates muscle-specific promoters (Lassar et al., 198913;Piette et al., 1990;Lin et al., 1991;Sartorelli et al., 1990;Wentworth et al., 1991). The dimerization with other regulatory proteins like E12 and E47 (Murre et al., 1989;Baldwin and Burden, 1989) or with its inhibitor ID (Benezra et al., 1990) takes place by means of its helix-loop-helix domain. Skeletal muscle differentiation is blocked by serum (Peterson et al., 1990) and growth factors, like fibroblast growth factor and transforming growth factor+; these peptides inhibit both fusion and MyoDl expression (Olson et al., 1986;Wice et al., 1987;Vaidya et al, 1989). Several oncogenes have also been shown to inhibit muscle differentiation and MyoDl expression, like ras (Olson et al., 19871, myc (Miner and Wold, 1991), fos and jun (Lassar et al., 1989a). Jun can form heterodimers with MyoDl through the leucine zipper and the helix-loop-helix motives and inhibits trans-activation of the MyoDl promoter and the muscle creatine kinase enhancer (Bengal et al., 1992). All of thejun and fos genes are immediate early genes, the transcription of which is rapidly induced in response to cell stimulation with growth factors, cytokines, and other agents Angel et al., 1987). The fos andjun gene products constitute the transcription factor AP-1, which trans-activates several promoters in response to serum and growth factors (Brenner et al., 1989;Kim et al., 1990). AP-1 can also inhibit transcription (Schule et al., 1990a(Schule et al., , 1990b; moreover, a single DNA sequence might mediate both AP-1 positive and negative effects on transcription (Diamond et al., 1990). Recently, it was shown that okadaic acid (0A)l blocks myogenesis by inhibiting MyoDl expression and MyoDl binding activity. Inhibition of MyoDl expression by OA correlates with induction of mRNA expression for the c-fos family, to a lesser extent for the jun family, and the consequent formation of active AP-1 complexes (Kim et al., 1992;Park et al., 1992). Taken together, these observations suggest that AP-1 might bind to the MyoDl promoter and negatively regulate MyoDl expression. The results presented here provide evidence that a Fos-related protein can bind specifically to a negative cis-acting element in the MyoDl promoter. This element shows sequence homology with a CAMP-responsive element (CRE) and with AP-1 recognition sequences. We also show that binding to the MyoDl CRE-like element is regulated during myoblast differentiation.

MATERIALS AND METHODS
Plasmid Construction-The template used for in vitro transcription was the plasmid pM050 (Fig. 11, which contains the MyoDl promoter The abbreviations used are: OA, okadaic acid; CRE, CAMP response element; CREB, CAMP response element binding protein; ERE, estrogen response element; CAT, chloramphenicol acetyltransferase; PCR, polymerase chain reaction; DM, differentiation medium; GM, growing medium; TPA, 12-0-tetradecanoylphorbol-13-acetate.

6978
This is an Open Access article under the CC BY license.
linked to a synthetic oligonucleotide from the cauliflower mosaic virus genome, from nucleotide 1643 to nucleotide 1671 (Franck et al., 1980). This sequence is unique and therefore suitable for specific detection. The deletion mutants ( Fig. 1) were performed taking advantage of the presence of specific restriction sites. As an internal control in the in vitro transcription assay we use the plasmid pCHllO (Pharmacia LKB Biotechnology Inc.), which contains the p-galactosidase gene under the control of the SV40 promoter. For DNA transfection, the plasmid pM054 containing the MyoDl promoter linked to the luciferase gene was used (Zingg et al., 1991), and the deletion mutants were those described above. The CRE-like and the F oligonucleotides (see below) were cloned by blunt-end ligation into the BglII site of the plasmid pCATP (Promega Corp.). Mutation of the CRE-like element was done using PCR according to Higuchi et al., 1988. Primary PCRs were performed in pairs containing the mismatched oligonucleotide 5'-GAG-CATGTC'JTCCCGTAC-3' (the mutated bases are shown in bold) and the corresponding 3' or 5' oligonucleotides. The primer used at the 5' end was the standard M13 universal primer and the primer at the 3' end was (5'-AGCCGCCCCTAGCTTCGCGCCAGGGCCCCT-3') expanding sequences of the MyoDl promoter located between nucleotide positions (-47 to -32). The target DNA was pMO, which contains the genomic clone of the MyoDl gene (Zingg et al., 1991). Reactions were performed with Taq polymerase following the manufacturer's instructions. Each cycle consisted of denaturation at 95 "C for 30 s, annealing a t 42 "C for 40 s, and extension a t 72 "C for 1 min. Twenty-five cycles were performed. Extension time at the last cycle was 10 min. After the primary PCRs were completed 20 pl of the reaction mix were separated on 1% agarose gel. Each pair fragment of the expected size was cut out of the gel, electroeluted, ethanol-precipitated, and resuspended into 25 pl of HzO. 5 pl of the solution containing the primary fragments were then used as target DNAs for the secondary PCR, and the 5' and 3' oligonucleotides mentioned above were used a s primers. The conditions for the secondary PCR were the same as for the primary reactions. The final mutated PCR product was cloned in to the vector PCRTM (Invitrogen). The mutation was confirmed by chain termination sequencing. The sequences between the Hind111 and PuuII restriction sites of the wild type MyoDl promoter were substituted in the plasmid pM054 by the HindIII-PuuII DNA fragment containing the mutated CRE-like element, to generate the plasmid pM054mut. The plasmid SomCAT the CRE (Andrisani et al., 1987).
contains the rat somatostatin basal promoter (-72 base pairs) including Oligonucleotides-Oligonucleotides were purified and equal moles of complementary strands were heated to 90 "C for 1 min and annealed by slow cooling to room temperature. The oligonucleotide F from the chicken vitellogenin gene (nucleotide position -213 to -182) is 5'-ATA-CATTGAAACTTCTGGTCAATCAGAAAAAG-3' (Vaccaro et al. 1990). The oligonucleotides used as competitors in the in vitro transcription assay are shown in Table I, and those used as competitors in gel mobility shift assays are shown in Fig. 4A.
Nuclear Extract Preparation-Skeletal muscle nuclear extract from mouse hind limbs were prepared according to Zahradka et al. (19891, buffer (10 mM Hepes pH 7.5, 5 mM KCl, 10 mM MgCl,, 5 mM 2-mer-with the following modifications. The tissue was homogenized in lysis captoethanol, and 320 mM sucrose); nuclei were purified through a 2.2 M sucrose pad in lysis buffer; the nuclei pellet was resuspended in nuclear lysis buffer (20 mM Hepes pH 7.9, 1 mM EDTA, 2 mM dithiothreitol, and 25% glycerol) and homogenized by 5-8 strokes in a Dounce glass-glass homogenizator (pestle A). The DNA concentration of the nuclei suspension was adjusted to 0.5 mg/ml with nuclear lysis buffer, and then 1/10 volume of 4 M (NH,),SO, was added. The nuclear lysate was centrifuged for 60 min at 250,000 x g, and 0.3 g of solid (NH,),SO, was added per ml of supernatant. Proteins were sedimented at 250,000 x g for 20 min, and the pellet was resuspended in dialysis buffer (20 mM Hepes pH 7.9, 1 mM dithiothreitol, 0.2 mM EDTA, 50 m M KC1, 5 mM MgC12, and 20% glycerol) and dialyzed two times for 1 h each against a large volume of the same buffer. The nuclear proteins were stored in small aliquots at -80 "C. Nuclear extracts from cell lines were prepared according to Schreiber et al. (1989). Protein determination was performed using Bio-Rad protein assay system according to the manufacturer's indications.
In Vitro Dunscription Assay-Nuclear extracts (10 pg) and DNA template (500 ng) were incubated in a final volume of 50 pl of the transcription reaction mixture (15 mM Hepes pH 7.6, 30 mM KCl, 10 mM MgC12, 150 p~ of each NTP, 2.5 mM dithiothreitol, 12% glycerol, and 40 units of RNase inhibitor) and incubated for 1 h a t 30 "C. For the competition experiments, the reaction contained increasing amounts of competing oligonucleotides, added at the assembly of the reaction. At the end of the incubation, the reaction mixture was treated with pro-teinase K (100 pg/ml) and extracted with phenokhloroform, and the RNA was precipitated. Transcripts were detected by nuclease S1 protection assay. The RNA was hybridized overnight a t 49 "C with approximately 1 ng of end-labeled oligonucleotide (lo8 c p d p g ) . For detection of the @galactosidase transcripts we used a 63-mer oligonucleotide encompassing nucleotide position 1388-1450 from the sequence of the plasmid PCHllO (Pharmacia). The hybridization mixture was digested with S1 nuclease (Boehringer Mannheim) followed by phenoV chloroform extraction (Ley et al., 1982). Protected fragments were analyzed on 12% polyacrylamide/8 M urea gels. The gels were dried and exposed for autoradiography a t -70 "C. Gels were quantified using a scanner model CS-930 (Shimadzu).
Gel Mobility Shift Assay-Oligonucleotides used for gel mobility shift assay were end-labeled with T4 polynucleotide kinase (Biofinex), and 0.1 ng of labeled oligonucleotide (lo8 cpdpg) was incubated with 10 pg of nuclear extract for 20 min a t room temperature in the presence of increasing concentrations of specific competitor oligonucleotide and 1 pg of nonspecific DNAcompetitor (sonicated Escherichia coli DNA). The reaction mixture contained 25 mM Hepes pH 7.9, 40 mM KCl, 3 mM MgC12, 0.1 mM EDTA, 1 mM dithiothreitol, and 10% glycerol in a final volume of 20 pl. The DNA-protein complexes were separated from the free oligonucleotide on a 6% polyacrylamide gel (29:l) in 0.25 x TBE buffer (22 mM Tris base, 22 mM boric acid, and 0.5 mM EDTA). Gels were dried and exposed for autoradiography. When antibodies were used, the nuclear extracts were first incubated with the antibody for 1 h a t 4 "C before the addition of the labeled probe, and then the reaction was carried out a s described above.
Cell Culture and DNA Dansfection-Mouse myoblasts (G8) were grown on collagen-coated plates in Dulbecco's modified Eagle's medium containing 4.5 gAiter glucose, 10% heat-inactivated fetal calf serum, and 10% horse serum (growing medium, GM). G8 myoblasts were induced to differentiate a t 90% confluency by changing them to Dulbecco's modified Eagle's medium containing 2% fetal calf serum (differentiation medium, DM). Transfection and luciferase assay were performed as described previously (Zingg et al., 1991). Transfection of the plasmids containing the CAT gene as reporter was performed by the CaPO, method. Cells were transfected with 5 pg of reporter plasmid, 3 pg of internal control, and 1 pg of expression vector. Expression vectors for c-Jun and c-Fos have been previously described (Schontal et al., 1988;Hirai et al., 19901, and the CAMP-response element binding protein (CREB) expression vector (Gonzalez et al., 1989) was kindly provided by Dr. Y. Nagamine at the Friedrich Miescher Institut. CAT assay was performed according to Gorman et al. (1983). For these experiments the internal control plasmid was RSV-/3-GAL (Gynheung et al., 1982) to avoid competition with the SV40 promoter. 6-Galactosidase assay was performed according to Lucibello and Mueller (1989).
Cyto-immunofluorescence-Cells were fixed with 4% paraformaldehyde in phosphate-buffered saline for 10 min a t room temperature. Permeabilization was done with 0.5% Nonidet P-40 in phosphate-buffered saline for 5 min at room temperature. Cells were then incubated for 10 min a t 37 "C with 10% goat serum in phosphate-buffered saline.
c-Fos and c-Jun expression was determined by indirect immunofluorescence, using affinity-purified rabbit polyclonal antibodies specific for c-Fos and c-Jun (Oncogene Science). Anti-c-Fos antibody was diluted 1:lOO in 10% goat serum, followed by staining with rhodamine-conjugated goat anti-rabbit antibody (Sigma). For c-Jun detection, the antic-Jun antibody was used a t 1:25 dilution in 10% goat serum.

MyoDl 5' Flanking Region Sequences Are
Required for E f fcient Dunscription Both in Vivo and in Vitro-Recently we reported a genomic clone of the MyoDl gene and 653 base pairs of its 5' sequence. We showed that the promoter is active both in vivo and in vitro (Zingg et al., 1991). To study the relationship between cis-acting elements and trans-acting factors and their effect on MyoDl transcription, several deletion mutants were constructed (Fig. lA) and tested for their transcriptional activity in vivo. As shown in Fig. lB, deletion to the XbaI site (-470, relative to the mRNA start site) had no effect on the luciferase activity when compared with the wild type. However, an increase in the activity was observed by further deletion to the AccI site (-314). Deletion to the PvuII site (-222) reduced the luciferase activity t o wild type levels. As expected, further deletion t o the SmaI site (-90) dramatically reduced luciferase activity, because this deletion removes all putative binding mutants. G8 myoblasts were transiently transfected with 5 pg of plasmids containing the MyoDl promoter deletion mutants and a promomoterless plasmid ( c ) . Luciferase activity was tested 48 h after DNA was added to the cells. Luciferase values were normalized to values obtained for P-galactosidase activity of the internal control plasmid (pCH110). C, in vitro transcription activity of the MyoDl promoter deletion mutants. Transcription was carried out as described under "Materials and Methods." Data shown represent the average of a t least three independent experiments with a standard deviation below 15%. sites for the positive ubiquitous trans-acting factors (see Fig. lA). By using the same deletion mutants in the in vitro transcription system (Fig. lC), we observed basically the same effect of the deletions on MyoDl transcription. Deletion mutant to the AccI site increased transcription (%"-fold), whereas deletion to the PuuII site decreased transcription to the wild type levels. On the other hand, deletion to the SmaI site did not decrease transcriptional activity as for the in vivo expression. These in vivo and in vitro experiments suggest that negative regulating sequences are situated between nucleotide positions -470 and -314 (AccI deletion mutant) and, furthermore, that 222 nucleotides (PuuII mutant) can confer the basal promoter activity of the MyoDl gene.

Functional Analysis of cis-Acting Elements by in Vitro Danscription Competition Assay-As
an alternative approach to confirm the above results, we used the in vitro transcription competition assay of the MyoDl wild type construct, pM050 (Fig. lA), with a series of synthetic oligonucleotides covering the upstream region of the MyoDl promoter. Table I shows the oligonucleotides tested in the competition assays. The amount of oligonucleotide needed to compete all the specific factors binding to the DNA template was determined by gel mobility shift competition assays under the same conditions as for the in vitro transcription assay. 100-fold molar excess of unlabeled oligonucleotide was sufficient to completely displace the specific protein-DNA complexes formed in the presence of 10 pg of nuclear protein (data not shown). Three different concentra-  2B). In addition, GC1, GC2, and AP-2 oligonucleotides all decreased the transcription activity by 30-50%, suggesting that these sequences confer the basal activity of the MyoDl promoter. The oligonucleotides containing SP1 binding sites also decreased SV40 transcription activity, because the SV40 promoter contains six SP1 sites. Therefore, the values given in Fig.  2B for the two SP1 oligonucleotides are relative to the transcriptional activity of the MyoDl promoter with no competitor. Transfection and in vitro transcription with the deletion mutant AccI (see Fig. 1, B and C) strongly suggest that the sequences located between the two restriction sites XbaI and AccI (nucleotide positions -470 to -314) contain a negative regulatory element. As shown previously (Zingg et al., 1991), a halfpalindrome of the estrogen response element (ERE) and a putative CRE can be found in this region. Two oligonucleotides covering this sequence were used in the transcription competition experiments. Competition with the oligonucleotide ERE (nucleotide positions -364 to -347) had no effect on transcription (Fig. 2, A and IB), whereas competition with the oligonucleotide containing the CRE-like element (nucleotide positions -342 to -322) resulted in a 2.8-fold increase in the transcriptional activity. These results clearly show that the negative effect of this sequence on MyoDl promoter activity is mediated mainly by a CRE-like element.
To test whether in vivo the CRE-like element is playing a role in the negative regulation of the MyoD1 promoter, we transfected G8 myoblast with the plasmid PM054mut, in which the CRE-like element has been mutated, and the rest of the MyoDl promoter remains intact. Fig. 3 shows that mutation of the CRE-like element induces luciferase activity by 1.5-2-fold when compared with the wild type construct, thus providing direct evidence for a functional role of this element in the negative regulation of the MyoDl promoter in dividing myoblasts.

AP-1 Oligonucleotides Inhibit Binding to the CRE-like Element-
The sequence of the CRE-like element differs only by one nucleotide from the canonical AP-1 recognition sequence (Fig. 4A). To determine whether there is a AP-1 component binding to the MyoDl CRE-like element, inhibition of the binding was examined with oligonucleotides containing AP-1 responsive sequences (Fig. 4A). Competition with a 50 X molar excess of the CRE-like element abolishes the binding (Fig. 4 B ,  lane 3 ) . Deletion of a G in the middle of the sequence creates a consensus AP-1 binding site (Fig. 4A), which also inhibits the binding to the wild type (Fig. 4B, lane 4 ) . Two other mutants myoblasts were transfected with the MyoDl promoter wild type plasmid (pM054wt) or with the plasmid carrying the mutated CRE-like element (pM054mut). Luciferase and &galactosidase assay were performed 48 h after transfection. The data represent the average of 4 independent experiments. Standard deviation was around 10%.
were able to compete for binding (lanes 5 and 6 ) . Lee et al. (1993) have recently characterized an AP-1 element from the porcine urokinase plasminogen activator promoter. Interestingly, this element shows perfect sequence homology with the MyoDl CRE-like element (Fig. 4A). Oligonucleotides contain- ing this AP-1 element and various mutants, which do not affect TPA responsiveness, were able to abolish binding (Fig. 4B,   lanes 7-9 ). On the other hand, a mutant that lacks TPA responsiveness competed to a lesser extent the binding to the MyoDl CRE-like element (Fig. 4B, lane 11 ). Finally, an oligonucleotide containing the SV40 TPA responsive element (Angel et al., 1987) competed the binding (lane 10). These results suggest that AP-1 may bind to the MyoDl CRE-like element.

MyoDl CRE-like Element binds c-Fos and c-Jun
Oncoproteins-It is well known that myoblasts in culture fuse and differentiate several days after removal of the serum from the medium. Fusion is preceded by an increase in MyoDl mRNA levels (Bengal et al., 1992)." We therefore tested the binding activity of nuclear extracts to the CRE-like oligonucleotide from G8 myoblast grown in the presence of serum and nuclear extracts from G8 myoblasts grown in low serum medium. As shown in Fig. 5A Because expression of the jun and fos gene families can be induced upon cell stimulation with serum, we tested whether the binding to the CRE-like element could be reinduced in cells grown 48 h in differentiation medium (low serum levels) and then restimulated for 4 h with growing medium (high serum levels). Fig. 5A (lane 4 ) shows clearly that serum can reinduce the binding of a factor to the CRE-like element. Binding is induced up to 3-fold upon addition of serum (compare lanes 3 and 4 ) . Fig. 5B shows that the DNA-protein complex with the CRE probe and nuclear extracts from myoblasts restimulated with serum can be competed by AP-1 oligonucleotides. Thus, AP-1 binding activity to the negative cis-acting element of the MyoDl promoter is down-regulated during myogenesis and can be restored by serum stimulation. Binding activity to an oligonucleotide, which contains the MyoDl binding site of the muscle creatine kinase enhancer (Lassar et al., 1989b), does not change under these conditions (Fig. 5 0 ) .
I t h a s been shown that the addition of anti-Fos antibodies to nuclear extracts from several cell types blocks the formation of AP-1.DNAcomplexes (Distel et al., 1987;Rauscheret al., 1988). Fig. 5C, lane 2, shows that anti-c-Fos antibodies inhibited complex formation with the MyoDl CRE-like element, on the other hand, anti-Jun antibodies had no effect on the protein binding to the CRE-like element (data not shown). It is interesting to mention that anti-c-Jun antibody is unable to immmunoprecipitate Fos in Jun-Fos complexes . As a control we used anti-Mos antibodies, which had no effect on protein-DNA complex formation (lane 1 ). The same results were obtained with extracts from dividing myoblasts (data not shown). As determined by indirect immunofluorescence using anti-c-Fos and anti-c-Jun antibodies serum levels regulate c-fos and c-jun expression in G8 myoblasts (Fig. 6). Altogether, these results clearly show a direct relationship between c-fos and c-jun expression and DNA binding activity to the MyoDl CRElike element. The results strongly suggest that the increase in MyoDl mRNA levels after serum removal is in part due to the down-regulation of the AP-1 binding activity to the negative element (CRE) of the MyoDl promoter.
MyoDl CRE-like Element Confers a n Additive Effect to the SV40 Minirnal Promoter in Response to TPA Stimulation and to c-fos a n d c-jun Overexpression-To determine whether the CRE-like element of the MyoDl promoter is able to negatively modulate a heterologous promoter, we cloned this element into the pCATP reporter plasmid containing the CAT gene under the control of the minimal SV40 promoter. As shown in Fig. 7A the CRE-like element showed no effect on the basal activity of the SV40 promoter when transfected into G8 myoblasts. In contrast, the CRE-like element induced the SV40 promoter activity in response to TPA stimulation and to c-fos and czjun expression (Bfold). No such effect was observed by cloning a n oligonucleotide containing a half-palindrome of the ERE from the chicken vitellogenin promoter (oligonucleotide F). Similar results were obtained by transfecting HeLa cells (data not shown).
Because a n oligonucleotide containing the CRE from the rat somatostatin promoter can compete for binding to the MyoDl CRE-like element (data not shown), we also tested whether this element is able to respond to the CAMP signal transduction pathway. As shown in Fig. 7B, neither treatment of the cells with forskolin nor cotransfection with a CREB expression vec- tor showed any effect on CAT activity. On the other hand, the reporter plasmid containing the somatostatin CRE (pSOM-CAT) showed an 8-fold increase in CAT activity when cotransfected with the CREB expression vector. These results clearly show that the MyoDl CRE-like element does not confer negative modulation to a heterologous promoter; however, it can respond to c-Fos and c-Jun in vivo and act cooperatively with the AP-1 element from the SV40 minimal promoter. On the other hand, it does not confer CAMP responsiveness to the SV40 promoter and is unable to induce the SV40 promoter by CREB expression.

DISCUSSION
MyoDl has been implicated as a master regulatory gene in the process of muscle determination and differentiation.
Although a great deal of information about the protein structure and its function has been generated, very little is known about MyoDl gene expression. We recently demonstrated that a fragment of about 650 base pairs from the MyoDl 5' upstream region contains promoter activity (Zingg et al., 1991). In the present study, we have shown that in vivo the basal MyoDl promoter activity is conferred by sequences downstream from the PvuII site (nucleotide position -222). Several recognition sequences for ubiquitous transacting factors are located in this region. As shown by the in vitro transcription competition assay, two SP1 elements, a CAAT box and an AP-2 sequence, are required for MyoDl promoter basal activity. The differences observed in transcriptional activity of the deletion mutant SmaI between transfection experiments using proliferating G8 myoblasts and the in vitro transcription assay using adult muscle extracts might indicate differential regulation of the MyoDl promoter, depending of the differentiation state of the cell. Preliminary data indicate that this could be the case, because nuclear extracts from myotubes show different binding patterns with sequences downstream from the TATA box when compared with extracts obtained from dividing myoblasts.:' We have shown the presence of a negative cis-acting element located between nucleotide positions -470 to -314, whose removal enhances MyoDl transcription both in vivo and i n vitro (deletion mutant AccI). In this region there are two possible regulatory elements, a half-palindromic ERE and a CRE-like element. Transcription competition experiments with these oligonucleotides and mutation of this element in an otherwise wild type promoter show that the sequences in the CRE-like element are involved in the negative modulation of the MyoDl promoter.
er Regulation 6983 blasts withdraw from the cell cycle and fuse. Fusion is preceded by a 2-3-fold induction in MyoDl mRNA levels in C2C12 (Bengal et al., 1992) and in G8 myoblasts.2 c-fos and c-jun are expressed in dividing myoblasts, but transcription is turned off when myoblasts are induced to differentiate. Here we show that the binding activity to the CRE-like element is downregulated during myogenic differentiation. Moreover, we show that this binding activity can be recovered by addition of serum to the cultures. We have also shown that the binding activity to the CRE-like element can be competed by oligonucleotides containing A P -1 recognition sequences. Direct evidence that the serum-dependent binding is due to AP-1 complexes was obtained with anti c-Fos antibodies. Anti-c-Fos antibodies inhibited DNA-protein complex formation with extracts prepared from cells induced to differentiate and restimulated with serum. These data clearly show that AP-1 is binding to the MyoDl CRE-like element and strongly suggest that AP-1 is negatively modulating MyoDl promoter activity in growing myoblasts by direct binding to a negative cis-acting element.
Recently it was shown that OA can inhibit MyoDl expression and MyoDl promoter activity (Kim et al., 1992). OAinduces the expression of c-fos and c-jun family members and activates A P -1 complexes (Park et al., 1992). Thus the effect of OA on MyoDl transcription could possible be explained by the induction ofjun and, therefore, the inhibition of the MyoDl autoregulatory feedback loop. Interestingly, an internal deletion that removes the negative regulatory sequences of the MyoDl promoter is less sensitive to OA treatment (Kim et al, 1992) indicating that the effect of OA on MyoDl promoter activity is in part due to AP-1 interactions with these sequences. Despite the fact that myogenin is not expressed until myoblasts enter the differentiation pathway in response to serum withdrawal (Edmonson and Olson, 1989), it has been shown that high serum levels inhibit the expression and the autoregulatory loop of myogenin. This negative effect was found to be directed to the myogenin protein as well a s to the myogenin promoter (Edmonson et al., 1991). Furthermore, OA also inhibits myogenin expression (Kim et al., 1992). Moreover, sequence comparison shows that the myogenin promoter contains an element like the CRE from the MyoDl promoter and a consensus AP-1 binding site, suggesting that AP-1 may bind to these elements and negatively modulate myogenin expression. However, recently it has been reported that in cardiac myocytes fos and jun repress transcription of the atrium natriuretic factor gene. Repression does not require a typical AP-1 binding site but is targeted to the cardiac-specific element of the atrium natriuretic factor promoter (McBride et al., 1993). I t is likely therefore that c-fos and ejun exert the negative effect on the expression of skeletal and cardiac myogenic factors by distinct control mechanisms.
The fact that the MyoDl CRE-like element did not show a negative effect on the basal activity of the SV40 promoter indicates that this element may work only in the context of its own promoter and may be modulating specific interactions with the transcriptional machinery. The other possibility is that MyoDl sequences flanking the CRE-like element are necessary to mediate negative modulation. It is interesting to note that there is an inverted repeat of the sequence TGAGGT located upstream of the half-palindromic ERE. Moreover, the CRE-like element contains a half-palindrome of the ERE (GGTCA) (see Fig. 8 ) , therefore i t is possible that these sequences are also required for negative modulation of the MyoDl promoter. The fact that the deletion mutant AccI shows higher activity than the CRE-like mutant favor the second possibility. Several elements with homology to TRE or CRE containing the sequences GTCA are known to induce (Gaub et al., 1990) or to repress transcription (Schule et al., 1990b). The oncogenes jun and fos and steroid hormone receptors have been implicated in both instances. Recently, it has been shown that direct protein-protein interaction between J u n andlor Fos and members of the steroid hormone receptor family is required for biological function (Yang-Yen et al., 1990;Schule et al., 1991;Doucas et al., 1991). Furthermore, a single DNA sequence may mediate both the positive and negative effects of AP-1 (Diamond et al., 1990) and confers opposite regulation between mitogenic and differentiation pathways (Schule et al., 1990a). Therefore, it is possible that the upstream half-ERE and the CRE-like element are required to form a higher order DNAprotein complex with AP-1 and a steroid hormone receptor to mediate negative modulation. Currently we are testing these possibilities.
It has previously been shown that multiple synthetic copies of the consensus AP-1 binding site can act as TPA-inducible enhancer (Lee et al., 1987). Additionally, in the case of the human P-globin gene, naturally occurring tandem AP-1 sites constitute a n inducible enhancer (Ney et al., 1990). When the MyoDl CRE-like element is placed upstream of the SV40 promoter, it can confer an additive effect upon TPA stimulation and c-fos and cjun overexpression. This suggests that AP-1 complexes can bind the MyoDl CRE-like element in vivo and act cooperatively with the SV40 AP-1 site. However, the CRElike element did not show any effect on the SV40 promoter in response to CAMP signaling. The CRE-like element is identical to an element from the major histocompatibility complex class I promoter, which can respond to TPA and CAMP stimulation (Israel et al., 1989). Thus, despite the fact that the MyoDl CRE-like element is unable to respond to CAMP levels or to CREB expression in actively dividing myoblast, it does not rule out a role of this element within the myogenic differentiation pathway in response to a CAMP-specific signaling cascade.
The CRE-like element is conserved in the rat MyoDl promoter (Vaidya et al., 1992) and only one base pair substitution is observed (Fig. 8). The sequences located between the XbaI and AccI sites (negative regulatory sequences) show 90% homology with the rat sequences suggesting a similar functional role in the MyoDl promoter regulation in rodents. Whether or not the CRE-like element is also involved in the regulation of the of MyoDl expression in other species remains to be determined.