Cross-binding of factors to functionally different promoter elements in c-fos and skeletal actin genes

A family of muscle gene promoter element (CArG) binding activities in Xenopus embryos: CArG/SRE discrimination and distribution during myogenesis

The CArG box is an essential promoter sequence for cardiac muscle actin gene expression in Xenopus embryos. To assess the role of the CArG motif in promoter function during Xenopus development, the DNA-binding activities present in the embryo that interact with this sequence have been investigated. A family of four Embryo CArG box1 Factors (ECFs) was separated by a 2-step fractionation procedure. These factors were distinct from the previously described C-ArG box binding activity Serum Response Factor (SRF). ECF1 was the most prominent binding activity in cardiac actin-expressing tissues, and bound the CArG box in preference to a Serum Response Element (SRE). SRF was also detectable in muscle, but it bound preferentially to an SRE. The properties of ECF3 were similar to those of ECF1, but it was much less prominent in cardiac actin-expressing tissues. The properties of the two other factors were distinctive: ECF2 was of relatively low affinity and high abundance, whilst ECF4 bound non-specifically to ends of DNA. The binding activity (or activities) that interacted with the CArG box was found to be influenced by both the concentrations of the other CArG box binding activities and the sequence of the site. Although there was no evidence for a muscle-specific CArG box binding activity, the properties of ECF1 suggest that it could play a role in the expression of the cardiac actin gene during Xenopus development.

Heterodimers of myogenic helix-loop-helix regulatory factors and E12 bind a complex element governing myogenic induction of the avian cardiac alpha-actin promoter

Recent studies have shown that two genes regulating myogenesis (MyoD and myogenin) are coexpressed with cardiac alpha-actin during early stages of skeletal muscle development. Myogenin and MyoD are members of a family of regulatory proteins which share a helix-loop-helix (HLH) motif required for dimerization and DNA binding. Myogenin and MyoD form heterodimers with the ubiquitous HLH protein E12 which bind cis-acting DNA elements that have an E box (CANNTG) at their core. E boxes are present in the control regions of numerous muscle-specific genes, although their functional importance in regulating many of these genes has not yet been evaluated. In this report we examine the possibility that myogenin (or MyoD) directly transactivates the cardiac alpha-actin promoter. Heterodimers of myogenin and E12 (or MyoD and E12) specifically bound a restriction fragment extending from -200 to -103 relative to the start of cardiac alpha-actin transcription. Methylation interference footprints pinpointed the site of interaction to an E box immediately adjacent to a previously identified CArG box (CArG3). Site-directed mutations to the DNA-binding site revealed that either an intact E box or an intact CArG3 is required for induction of the cardiac alpha-actin promoter in myoblasts and for transactivation by myogenin in cotransfected fibroblasts. However, deletion and substitution experiments indicate that the complex E box/CArG3 element alone does not confer muscle-specific expression to a minimal promoter. These results suggest that direct and indirect pathways involving multiple cis-acting elements mediate the induction of the cardiac alpha-actin promoter by myogenin and MyoD.

Identification of a multiprotein complex interacting with the c-fos serum response element

The serum response element (SRE) is essential for serum and growth factor stimulation of the c-fos gene. We have examined the nuclear proteins, obtained from tissues with elevated expression of the c-fos gene (proliferating rat liver and hepatocarcinoma), that bind to the SRE sequence. A synthetic oligonucleotide containing the SRE sequence from the mouse c-fos gene promoter (-299 to -322) was radioactively labeled, used as a probe for the mobility shift assay and Southwestern (DNA-protein) blotting, and also used for sequence-specific affinity chromatography. We have identified a group of nuclear proteins of molecular sizes 36, 45, 62, 67, 72, and 112 kDa capable of interacting with the SRE sequence. The 36-, 67-, and 112-kDa proteins have DNA-binding properties, but the presence of the others in the SRE-protein complex could be the result of protein-protein interaction. All of these protein factors were present in nuclei obtained from intact and proliferating rat liver as well as from 5123tc Morris hepatoma. The DNA-binding activity (on Southwestern blots) of the 67- and 112-kDa proteins was not affected by alkaline phosphatase treatment, but the ability of the dephosphorylated nuclear proteins to form the complex with the SRE sequence under gel shift assay conditions was severely impaired. The same alkaline phosphatase treatment completely abolished the DNA-binding properties of the c-fos cyclic AMP-responsive element-specific proteins. Therefore, transcriptional activation of the c-fos gene at the SRE must require the presence of a multiprotein complex the formation of which is governed by phosphorylation. The binding of the 67- and 62-kDa proteins to the c-fos SRE has been previously reported; however, the 36-. 45-, 72-, and 112-kDa proteins are novel factors involved in the multifaceted regulation of c-fos gene expression in vivo.

Identification of a multiprotein complex interacting with the c-fos serum response element

The serum response element (SRE) is essential for serum and growth factor stimulation of the c-fos gene. We have examined the nuclear proteins, obtained from tissues with elevated expression of the c-fos gene (proliferating rat liver and hepatocarcinoma), that bind to the SRE sequence. A synthetic oligonucleotide containing the SRE sequence from the mouse c-fos gene promoter (-299 to -322) was radioactively labeled, used as a probe for the mobility shift assay and Southwestern (DNA-protein) blotting, and also used for sequence-specific affinity chromatography. We have identified a group of nuclear proteins of molecular sizes 36, 45, 62, 67, 72, and 112 kDa capable of interacting with the SRE sequence. The 36-, 67-, and 112-kDa proteins have DNA-binding properties, but the presence of the others in the SRE-protein complex could be the result of protein-protein interaction. All of these protein factors were present in nuclei obtained from intact and proliferating rat liver as well as from 5123tc Morris hepatoma. The DNA-binding activity (on Southwestern blots) of the 67- and 112-kDa proteins was not affected by alkaline phosphatase treatment, but the ability of the dephosphorylated nuclear proteins to form the complex with the SRE sequence under gel shift assay conditions was severely impaired. The same alkaline phosphatase treatment completely abolished the DNA-binding properties of the c-fos cyclic AMP-responsive element-specific proteins. Therefore, transcriptional activation of the c-fos gene at the SRE must require the presence of a multiprotein complex the formation of which is governed by phosphorylation. The binding of the 67- and 62-kDa proteins to the c-fos SRE has been previously reported; however, the 36-. 45-, 72-, and 112-kDa proteins are novel factors involved in the multifaceted regulation of c-fos gene expression in vivo.

Heterodimers of myogenic helix-loop-helix regulatory factors and E12 bind a complex element governing myogenic induction of the avian cardiac alpha-actin promoter

Recent studies have shown that two genes regulating myogenesis (MyoD and myogenin) are coexpressed with cardiac alpha-actin during early stages of skeletal muscle development. Myogenin and MyoD are members of a family of regulatory proteins which share a helix-loop-helix (HLH) motif required for dimerization and DNA binding. Myogenin and MyoD form heterodimers with the ubiquitous HLH protein E12 which bind cis-acting DNA elements that have an E box (CANNTG) at their core. E boxes are present in the control regions of numerous muscle-specific genes, although their functional importance in regulating many of these genes has not yet been evaluated. In this report we examine the possibility that myogenin (or MyoD) directly transactivates the cardiac alpha-actin promoter. Heterodimers of myogenin and E12 (or MyoD and E12) specifically bound a restriction fragment extending from -200 to -103 relative to the start of cardiac alpha-actin transcription. Methylation interference footprints pinpointed the site of interaction to an E box immediately adjacent to a previously identified CArG box (CArG3). Site-directed mutations to the DNA-binding site revealed that either an intact E box or an intact CArG3 is required for induction of the cardiac alpha-actin promoter in myoblasts and for transactivation by myogenin in cotransfected fibroblasts. However, deletion and substitution experiments indicate that the complex E box/CArG3 element alone does not confer muscle-specific expression to a minimal promoter. These results suggest that direct and indirect pathways involving multiple cis-acting elements mediate the induction of the cardiac alpha-actin promoter by myogenin and MyoD.

Identification of single-stranded-DNA-binding proteins that interact with muscle gene elements

A sequence-specific DNA-binding protein from skeletal-muscle extracts that binds to probes of three muscle gene DNA elements is identified. This protein, referred to as muscle factor 3, forms the predominant nucleoprotein complex with the MCAT gene sequence motif in an electrophoretic mobility shift assay. This protein also binds to the skeletal actin muscle regulatory element, which contains the conserved CArG motif, and to a creatine kinase enhancer probe, which contains the E-box motif, a MyoD-binding site. Muscle factor 3 has a potent sequence-specific, single-stranded-DNA-binding activity. The specificity of this interaction was demonstrated by sequence-specific competition and by mutations that diminished or eliminated detectable complex formation. MyoD, a myogenic determination factor that is distinct from muscle factor 3, also bound to single-stranded-DNA probes in a sequence-specific manner, but other transcription factors did not. Multiple copies of the MCAT motif activated the expression of a heterologous promoter, and a mutation that eliminated expression was correlated with diminished factor binding. Muscle factor 3 and MyoD may be members of a class of DNA-binding proteins that modulate gene expression by their abilities to recognize DNA with unusual secondary structure in addition to specific sequence.

Identification of single-stranded-DNA-binding proteins that interact with muscle gene elements

A sequence-specific DNA-binding protein from skeletal-muscle extracts that binds to probes of three muscle gene DNA elements is identified. This protein, referred to as muscle factor 3, forms the predominant nucleoprotein complex with the MCAT gene sequence motif in an electrophoretic mobility shift assay. This protein also binds to the skeletal actin muscle regulatory element, which contains the conserved CArG motif, and to a creatine kinase enhancer probe, which contains the E-box motif, a MyoD-binding site. Muscle factor 3 has a potent sequence-specific, single-stranded-DNA-binding activity. The specificity of this interaction was demonstrated by sequence-specific competition and by mutations that diminished or eliminated detectable complex formation. MyoD, a myogenic determination factor that is distinct from muscle factor 3, also bound to single-stranded-DNA probes in a sequence-specific manner, but other transcription factors did not. Multiple copies of the MCAT motif activated the expression of a heterologous promoter, and a mutation that eliminated expression was correlated with diminished factor binding. Muscle factor 3 and MyoD may be members of a class of DNA-binding proteins that modulate gene expression by their abilities to recognize DNA with unusual secondary structure in addition to specific sequence.

Multiple basal promoter elements determine the level of human c-fos transcription

Three cis-acting domains that contribute to the basal promoter activity of the human c-fos gene were identified. One encompasses the serum response element and has been previously described. Another spans an NF1-like site situated at -170. Mutations and in vitro protein binding assays pinpoint this site as the sole basal element of the medial domain. The third, or promoter-proximal, domain can be divided into several distinct sites, one containing a directly repeated GC-rich element and the other consisting of partially overlapping recognition sites for transcription factors ATF/CREB and MLTF/USF. Each of these sites contributes to basal activity as assayed by transient transfections and by in vitro transcription. Consistent with this, several complexes could be visualized between this region and nuclear proteins in vitro and genomic footprinting demonstrated that both elements are constitutively bound in vivo. On the basis of these results, we conclude that all three domains are necessary for full c-fos promoter function.

Multiple basal promoter elements determine the level of human c-fos transcription

Three cis-acting domains that contribute to the basal promoter activity of the human c-fos gene were identified. One encompasses the serum response element and has been previously described. Another spans an NF1-like site situated at -170. Mutations and in vitro protein binding assays pinpoint this site as the sole basal element of the medial domain. The third, or promoter-proximal, domain can be divided into several distinct sites, one containing a directly repeated GC-rich element and the other consisting of partially overlapping recognition sites for transcription factors ATF/CREB and MLTF/USF. Each of these sites contributes to basal activity as assayed by transient transfections and by in vitro transcription. Consistent with this, several complexes could be visualized between this region and nuclear proteins in vitro and genomic footprinting demonstrated that both elements are constitutively bound in vivo. On the basis of these results, we conclude that all three domains are necessary for full c-fos promoter function.

Differential regulation of skeletal alpha-actin transcription in cardiac muscle by two fibroblast growth factors

In cardiac muscle, acidic and basic fibroblast growth factors (aFGF and bFGF) regulate at least five genes in common (including alpha and beta myosin heavy chains, atrial natriuretic factor, and the sarcoplasmic reticulum calcium ATPase), provoking a generalized "fetal" phenotype similar to events in pressure-overload hypertrophy; however, aFGF and bFGF differentially control the striated alpha-actins. bFGF stimulates and aFGF inhibits skeletal alpha-actin transcripts associated with the embryonic heart, whereas cardiac alpha-actin mRNA is inhibited by aFGF but not bFGF. To elucidate mechanisms for these selective and discordant actions of aFGF and bFGF on cardiac muscle, chicken skeletal and cardiac alpha-actin promoter-driven reporter genes were introduced into neonatal rat cardiac myocytes by electroporation. Skeletal alpha-actin transcription was selectively stimulated by bFGF, whereas the cardiac alpha-actin promoter was unaffected. In contrast, aFGF suppressed both transfected alpha-actin genes. The differential regulation of skeletal alpha-actin transcription was equivalent with either purified or recombinant FGFs and was observed with 5' flanking sequences from either nucleotide -202 or -2000 to nucleotide -11. Positive and negative modulation of alpha-actin transcription by growth factors corresponded accurately to the endogenous genes in all permutations studied. These investigations provide a model for reciprocal control of gene transcription by aFGF vs. bFGF.

M-CAT binding factor, a novel trans-acting factor governing muscle-specific transcription

The cardiac troponin T (cTNT) promoter contains a highly muscle specific distal promoter element capable of conferring muscle-specific transcription from a heterologous TATA box-transcription initiation site. Three sequence motifs within this distal promoter element are conserved in the promoter and regulatory regions of many sarcomeric protein genes. Mutational analysis demonstrated that homologies to two of these conserved motifs (CArG/CBAR and MEF 1) were not required for activity of cTNT promoter-marker gene constructs in transfected embryonic skeletal muscle cells. In contrast, disruption of either or both copies of the conserved M-CAT motif (5'-CATTCCT-3') inactivated the cTNT promoter in these cells. Both M-CAT motifs were protected from DNase I cleavage in solution footprint assays by an M-CAT binding factor (MCBF) present in nuclear extracts from embryonic muscle tissue. M-CAT mutations that inactivated the cTNT promoter also disrupted MCBF binding, indicating that MCBF may be a key trans-acting factor required for muscle-specific expression of the cTNT promoter. MCBF also bound to the M-CAT motif in the distal promoter region of the skeletal alpha-actin gene, suggesting that it may play a role in the regulation of this and perhaps other muscle genes that contain M-CAT motifs.

M-CAT binding factor, a novel trans-acting factor governing muscle-specific transcription

The cardiac troponin T (cTNT) promoter contains a highly muscle specific distal promoter element capable of conferring muscle-specific transcription from a heterologous TATA box-transcription initiation site. Three sequence motifs within this distal promoter element are conserved in the promoter and regulatory regions of many sarcomeric protein genes. Mutational analysis demonstrated that homologies to two of these conserved motifs (CArG/CBAR and MEF 1) were not required for activity of cTNT promoter-marker gene constructs in transfected embryonic skeletal muscle cells. In contrast, disruption of either or both copies of the conserved M-CAT motif (5'-CATTCCT-3') inactivated the cTNT promoter in these cells. Both M-CAT motifs were protected from DNase I cleavage in solution footprint assays by an M-CAT binding factor (MCBF) present in nuclear extracts from embryonic muscle tissue. M-CAT mutations that inactivated the cTNT promoter also disrupted MCBF binding, indicating that MCBF may be a key trans-acting factor required for muscle-specific expression of the cTNT promoter. MCBF also bound to the M-CAT motif in the distal promoter region of the skeletal alpha-actin gene, suggesting that it may play a role in the regulation of this and perhaps other muscle genes that contain M-CAT motifs.

Functional analysis of elements affecting expression of the beta-actin gene of carp

Regulatory regions of the beta-actin gene of the common carp (Cyprinus carpio) have been examined by linking upstream, 5'-flanking sequences and regions of the first intron to a bacterial chloramphenicol acetyltransferase (CAT) reporter gene. By analysis of the mRNA products and encoded CAT activity, we have identified four putative regions that influence expression: (i) a negative regulatory region 2,300 to 1,100 base pairs (bp) ahead of the gene; (ii) a proximal promoter element, containing the highly conserved CCAAT, CC(A/T)6GG, and TATA boxes, that is within the first 204 bp upstream of the initiation site; (iii) a negative element of 426 bp in the 5' region of the first intron; and (iv) a positive 304-bp element near the end of the first intron that contains highly conserved sequences found in all characterized beta-actin genes. The positive intron element is not a classical enhancer; it is position and orientation dependent, as has been observed in other housekeeping genes in vertebrates. Depending on the elements joined together, CAT gene expression can be modulated more than 500-fold in transfected mouse cells.

Functional analysis of elements affecting expression of the beta-actin gene of carp

Regulatory regions of the beta-actin gene of the common carp (Cyprinus carpio) have been examined by linking upstream, 5'-flanking sequences and regions of the first intron to a bacterial chloramphenicol acetyltransferase (CAT) reporter gene. By analysis of the mRNA products and encoded CAT activity, we have identified four putative regions that influence expression: (i) a negative regulatory region 2,300 to 1,100 base pairs (bp) ahead of the gene; (ii) a proximal promoter element, containing the highly conserved CCAAT, CC(A/T)6GG, and TATA boxes, that is within the first 204 bp upstream of the initiation site; (iii) a negative element of 426 bp in the 5' region of the first intron; and (iv) a positive 304-bp element near the end of the first intron that contains highly conserved sequences found in all characterized beta-actin genes. The positive intron element is not a classical enhancer; it is position and orientation dependent, as has been observed in other housekeeping genes in vertebrates. Depending on the elements joined together, CAT gene expression can be modulated more than 500-fold in transfected mouse cells.

Regulation of the chicken embryonic myosin light-chain (L23) gene: existence of a common regulatory element shared by myosin alkali light-chain genes

The transcriptional regulation of the chicken myosin alkali light-chain (MLC) L23 gene was analyzed. Two different types of cis-regulatory regions were identified: one was a silencerlike region located between 3.7 and 2.7 kilobases upstream of the mRNA initiation site, and the other was essential for the expression of L23 in skeletal muscle cells and was located between 106 and 91 base pairs upstream of the cap site. This 16-base-pair cis-acting element was designated as the MLC box since it is well conserved in various muscle-specific MLC promoter regions. The activity of the MLC box showed tissue specificity. To analyze the relationship between the nucleotide sequence and the activity of the MLC box precisely, mutation analysis was performed. The 16-base-pair sequence was indispensable for the active transcription of L23 gene, and the MLC box could function in either orientation. The inverted sequence of the MLC box was similar to the sequence of the alpha-actin CArG box. By using a gel mobility retardation assay, the nuclear protein(s) that binds to both MLC box and CArG box was detected with nuclear extract prepared from chicken embryonic breast muscle. These observations imply that a common factor regulates the coordinate expression of these contractile proteins in muscle differentiation.

Regulation of the chicken embryonic myosin light-chain (L23) gene: existence of a common regulatory element shared by myosin alkali light-chain genes

The transcriptional regulation of the chicken myosin alkali light-chain (MLC) L23 gene was analyzed. Two different types of cis-regulatory regions were identified: one was a silencerlike region located between 3.7 and 2.7 kilobases upstream of the mRNA initiation site, and the other was essential for the expression of L23 in skeletal muscle cells and was located between 106 and 91 base pairs upstream of the cap site. This 16-base-pair cis-acting element was designated as the MLC box since it is well conserved in various muscle-specific MLC promoter regions. The activity of the MLC box showed tissue specificity. To analyze the relationship between the nucleotide sequence and the activity of the MLC box precisely, mutation analysis was performed. The 16-base-pair sequence was indispensable for the active transcription of L23 gene, and the MLC box could function in either orientation. The inverted sequence of the MLC box was similar to the sequence of the alpha-actin CArG box. By using a gel mobility retardation assay, the nuclear protein(s) that binds to both MLC box and CArG box was detected with nuclear extract prepared from chicken embryonic breast muscle. These observations imply that a common factor regulates the coordinate expression of these contractile proteins in muscle differentiation.

The c-fos cyclic AMP-responsive element conveys constitutive expression to a tissue-specific promoter

The c-fos and cardiac alpha-actin promoters share homologous 5' protein binding elements that are essential for serum-inducible and tissue-specific expression, respectively. Additional elements, auxiliary proteins or factor modifications, must distinguish the individual transcriptional responses of these two promoters. An element in the c-fos basal promoter that is normally responsible for transient stimulation of the fos gene in response to Ca2+ or cyclic AMP (CRE) may be able to modulate the expression of the upstream elements. We report here that this element, when inserted into the cardiac alpha-actin promoter, conveys constitutive expression to this otherwise highly restricted promoter. Additional data support the proposal that the CRE binding protein creates an alternative pathway whereby upstream regulatory elements in the cardiac alpha-actin promoter can activate transcription in a manner which circumvents the requirement for a tissue-specific environment.

The c-fos cyclic AMP-responsive element conveys constitutive expression to a tissue-specific promoter

The c-fos and cardiac alpha-actin promoters share homologous 5' protein binding elements that are essential for serum-inducible and tissue-specific expression, respectively. Additional elements, auxiliary proteins or factor modifications, must distinguish the individual transcriptional responses of these two promoters. An element in the c-fos basal promoter that is normally responsible for transient stimulation of the fos gene in response to Ca2+ or cyclic AMP (CRE) may be able to modulate the expression of the upstream elements. We report here that this element, when inserted into the cardiac alpha-actin promoter, conveys constitutive expression to this otherwise highly restricted promoter. Additional data support the proposal that the CRE binding protein creates an alternative pathway whereby upstream regulatory elements in the cardiac alpha-actin promoter can activate transcription in a manner which circumvents the requirement for a tissue-specific environment.

A new myocyte-specific enhancer-binding factor that recognizes a conserved element associated with multiple muscle-specific genes

Exposure of skeletal myoblasts to growth factor-deficient medium results in transcriptional activation of muscle-specific genes, including the muscle creatine kinase gene (mck). Tissue specificity, developmental regulation, and high-level expression of mck are conferred primarily by a muscle-specific enhancer located between base pairs (bp) -1350 and -1048 relative to the transcription initiation site (E. A. Sternberg, G. Spizz, W. M. Perry, D. Vizard, T. Weil, and E. N. Olson, Mol. Cell. Biol. 8:2896-2909, 1988). To begin to define the regulatory mechanisms that mediate the selective activation of the mck enhancer in differentiating muscle cells, we have further delimited the boundaries of this enhancer and analyzed its interactions with nuclear factors from a variety of myogenic and nonmyogenic cell types. Deletion mutagenesis showed that the region between 1,204 and 1,095 bp upstream of mck functions as a weak muscle-specific enhancer that is dependent on an adjacent enhancer element for strong activity. This adjacent activating element does not exhibit enhancer activity in single copy but acts as a strong enhancer when multimerized. Gel retardation assays combined with DNase I footprinting and diethyl pyrocarbonate interference showed that a nuclear factor from differentiated C2 myotubes and BC3H1 myocytes recognized a conserved A + T-rich sequence within the peripheral activating region. This myocyte-specific enhancer-binding factor, designated MEF-2, was undetectable in nuclear extracts from C2 or BC3H1 myoblasts or several nonmyogenic cell lines. MEF-2 was first detectable within 2 h after exposure of myoblasts to mitogen-deficient medium and increased in abundance for 24 to 48 h thereafter. The appearance of MEF-2 required ongoing protein synthesis and was prevented by fibroblast growth factor and type beta transforming growth factor, which block the induction of muscle-specific genes. A myoblast-specific factor that is down regulated within 4 h after removal of growth factors was also found to bind to the MEF-2 recognition site. A 10-bp sequence, which was shown by DNase I footprinting and diethyl pyrocarbonate interference to interact directly with MEF-2, was identified within the rat and human mck enhancers, the rat myosin light-chain (mlc)-1/3 enhancer, and the chicken cardiac mlc-2A promoter. Oligomers corresponding to the region of the mlc-1/3 enhancer, which encompasses this conserved sequence, bound MEF-2 and competed for its binding to the mck enhancer. These results thus provide evidence for a novel myocyte-specific enhancer-binding factor, MEF-2, that is expressed early in the differentiation program and is suppressed by specific polypeptide growth factors. The ability of MEF-2 to recognize conserved activating elements associated with multiple-specific genes suggests that this factor may participate in the coordinate regulation of genes during myogenesis.

A new myocyte-specific enhancer-binding factor that recognizes a conserved element associated with multiple muscle-specific genes

Exposure of skeletal myoblasts to growth factor-deficient medium results in transcriptional activation of muscle-specific genes, including the muscle creatine kinase gene (mck). Tissue specificity, developmental regulation, and high-level expression of mck are conferred primarily by a muscle-specific enhancer located between base pairs (bp) -1350 and -1048 relative to the transcription initiation site (E. A. Sternberg, G. Spizz, W. M. Perry, D. Vizard, T. Weil, and E. N. Olson, Mol. Cell. Biol. 8:2896-2909, 1988). To begin to define the regulatory mechanisms that mediate the selective activation of the mck enhancer in differentiating muscle cells, we have further delimited the boundaries of this enhancer and analyzed its interactions with nuclear factors from a variety of myogenic and nonmyogenic cell types. Deletion mutagenesis showed that the region between 1,204 and 1,095 bp upstream of mck functions as a weak muscle-specific enhancer that is dependent on an adjacent enhancer element for strong activity. This adjacent activating element does not exhibit enhancer activity in single copy but acts as a strong enhancer when multimerized. Gel retardation assays combined with DNase I footprinting and diethyl pyrocarbonate interference showed that a nuclear factor from differentiated C2 myotubes and BC3H1 myocytes recognized a conserved A + T-rich sequence within the peripheral activating region. This myocyte-specific enhancer-binding factor, designated MEF-2, was undetectable in nuclear extracts from C2 or BC3H1 myoblasts or several nonmyogenic cell lines. MEF-2 was first detectable within 2 h after exposure of myoblasts to mitogen-deficient medium and increased in abundance for 24 to 48 h thereafter. The appearance of MEF-2 required ongoing protein synthesis and was prevented by fibroblast growth factor and type beta transforming growth factor, which block the induction of muscle-specific genes. A myoblast-specific factor that is down regulated within 4 h after removal of growth factors was also found to bind to the MEF-2 recognition site. A 10-bp sequence, which was shown by DNase I footprinting and diethyl pyrocarbonate interference to interact directly with MEF-2, was identified within the rat and human mck enhancers, the rat myosin light-chain (mlc)-1/3 enhancer, and the chicken cardiac mlc-2A promoter. Oligomers corresponding to the region of the mlc-1/3 enhancer, which encompasses this conserved sequence, bound MEF-2 and competed for its binding to the mck enhancer. These results thus provide evidence for a novel myocyte-specific enhancer-binding factor, MEF-2, that is expressed early in the differentiation program and is suppressed by specific polypeptide growth factors. The ability of MEF-2 to recognize conserved activating elements associated with multiple-specific genes suggests that this factor may participate in the coordinate regulation of genes during myogenesis.