The Homeodomain Protein Barx2 Promotes Myogenic Differentiation and Is Regulated by Myogenic Regulatory Factors*

The homeobox protein Barx2 is expressed in both smooth and skeletal muscle and is up-regulated during differentiation of skeletal myotubes. Here we use anti-sense-oligonucleotide inhibition of Barx2 expression in limb bud cell culture to show that Barx2 is required for myotube formation. Moreover, overexpression of Barx2 accelerates the fusion of MyoD-positive limb bud cells and C2C12 myoblasts. However, overexpression of Barx2 does not induce ectopic MyoD expression in either limb bud cultures or in multipotent C3H10T1/2 mesenchymal cells, and does not induce fusion of C3H10T1/2 cells. These results suggest that Barx2 acts downstream of MyoD. To test this hypothesis, we isolated the Barx2 gene promoter and identified DNA regulatory elements that might control Barx2 expression during myogenesis. The proximal promoter of the Barx2 gene contained binding sites for several factors involved in myoblast differentiation including MyoD, myogenin, serum response factor, and myocyte enhancer factor 2. Co-trans-fection experiments showed that binding sites for both MyoD and serum response factor are necessary for activation of the promoter by MyoD and myogenin. Taken together, these studies indicate that Barx2 is a key reg-ulator of myogenic differentiation that acts downstream of muscle regulatory factors. myotubes paraformaldehyde and Quantitative Western Blotting— To examine protein expression in pcBarx2- and pcDNA3-transfected C2C12 cultures, proliferating cells were harvested at 80% confluence and lysed in a hypotonic buffer containing 1% Nonidet P-40 (22). Protein concentration was determined by Bradford assay (Bio-Rad) and 20- (cid:2) g aliquots were separated by SDS-PAGE, transferred to polyvinylidene difluoride membrane (In- vitrogen), and blotted using either polyclonal antibodies to Barx2 (San-ta Cruz) or monoclonal antibodies to SM- (cid:1) -actin, (cid:3) -actin, or desmin (Sigma). Blots were developed using chemiluminescent substrate (Novex) and autoradiographed. The intensity of the protein bands was determined by densitometry using NIH Image software and a ratio of either SM- (cid:1) -actin or desmin protein to (cid:3) -actin protein was determined for each sample. The average increase in SM- (cid:1) -actin expression was determined from three replicate experiments. Isolation the Murine Barx2 Gene— The murine Barx2 gene was isolated from a 129/SvJ murine genomic BACmid library in the pveloBAC vector after screening with a PCR probe derived from the 5 (cid:2) end of the Barx2 cDNA sequence (Genome Systems Inc). Several overlapping restriction fragments containing the Barx2 gene were mapped using Southern blot analysis and sequenced. promoter/lucif-erase reporter constructs were generated in the promoterless pGL3basic vector (Promega). The longest promoter construct contained Relative binding of the protein to the E-box probes was determined by measuring the intensity of the probe-protein complexes formed in gel mobility-shift experiments using a PhosphorImager (Amerhsam Bio- sciences). The relative binding data shown in Fig. were derived from four independent experiments using different C2C12 extract preparations.

Most skeletal muscle develops from mesenchymal premuscle condensations that are determined early in development (1). Several types of transcription factors control the specification and differentiation of skeletal myoblasts within these condensations. Critical among these are four basic helix-loop-helix myogenic regulatory factors (MRFs) 1 : MyoD, Myf5, myogenin, and MRF4, each of which binds to E-box sequences in the promoters of skeletal muscle genes (see Ref. 2 for review). Each of the four MRFs can induce skeletal muscle differentiation when expressed in nonmuscle cells (3)(4)(5)(6)(7). However, during normal skeletal muscle development, the MRFs function in a hierarchical fashion: MyoD and Myf5 are involved in myoblast specification, whereas myogenin and MRF4 control the terminal differentiation of myoblasts into myotubes (8,9). MyoD also promotes myoblast differentiation by inducing the expression of myogenin and MEF (myocyte enhancer factor) 2 family proteins. In addition, several homeobox genes, including Msx1, Pax3, and Pax7 influence the specification of myoblasts by regulating the expression of particular MRFs (10 -12). However, few homeobox genes have been described that act downstream of MRFs or that influence myotube fusion directly without influencing the expression of MRFs.
One component of the morphogenetic program specified by homeobox genes is the control of cell adhesion, a process that is essential to condensation, migration, and fusion of cells in several developmental contexts (13,14). Barx1 and Barx2 are two closely related homeobox proteins that were found in separate efforts to identify factors that control the expression of cell adhesion molecules (CAMs) (15,16). In an earlier study, we discovered Barx2 and showed that it is expressed in precise patterns during mouse embryonic development in the nervous system, neural crest-derived craniofacial structures, lung buds, and limb mesenchyme (16). More recently, other researchers have shown that chicken and murine Barx2 variants are expressed in both skeletal and smooth muscle (17,18). For example, the chicken Barx2 homologue (Barx2b) is expressed in the early myotome and persists after differentiation of muscle groups in the limb, neck, and cloaca (17). In addition, Herring et al. (18) showed that murine Barx2 is expressed in adult skeletal muscle and smooth muscle-containing tissues. Barx2 is also expressed in C2C12 myoblasts, and its expression increases dramatically when myoblasts fuse to form myotubes (18). Collectively, these observations suggest roles for Barx2 in both skeletal and smooth muscle development.
In this study we examined the role of Barx2 in muscle development. Inhibition of Barx2 expression in limb bud cell cultures inhibits myotube formation, whereas overexpression of Barx2 accelerates the appearance of myotubes. However, overexpression of Barx2 does not induce ectopic MyoD expression in limb bud cultures, nor is it sufficient to specify a myoblast fate in multipotent C3H10T1/2 mesenchymal cells that do not express MyoD. These results suggest that Barx2 acts downstream of MyoD to promote myogenic differentiation but cannot substitute for MyoD function. In support of this conclusion, we found that both MyoD and myogenin can bind to and activate the Barx2 promoter in C2C12 myoblasts. In C3H10T1/2 cells overexpression of Barx2 induces a smooth muscle-like appearance with increased smooth muscle-␣-actin expression and formation of stress fibers. Overall our data suggest two roles for Barx2 in muscle development: in MyoD-positive skeletal myoblasts Barx2 controls myotube fusion, whereas in MyoD-negative mesenchymal cells Barx2 may promote differ-* This work was supported by a grant from the G. Harold and Leila Y. Mathers Foundation, National Science Foundation Grant IBN-9816896 (to F. S. J.), NIH Grant NS39837, and a grant from the Charles and Mildred Schnurmacher Foundation (to D. B. E.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AY188085.
entiation of other contractile phenotypes such as smooth muscle cells or myofibroblasts.

EXPERIMENTAL PROCEDURES
Morpholino Antisense Oligonucleotide Inhibition of Barx2 Expression in Limb Bud Cultures-Morpholino oligonucleotides (ODNs) that are antisense to a region of the Barx2 mRNA near the initiation codon were designed and synthesized by Gene Tools LLC. Control sense ODNs were also prepared. Dissociated mesenchymal cells were prepared from E10.5 mouse limb buds and 2 ϫ 10 6 cells were plated in 10-l aliquots at a density of 2 ϫ 10 7 cells/ml in CMRL-1066 (Invitrogen) containing 2% fetal bovine serum (FBS) in four-well culture dishes. After 1 h incubation to allow the micromass cultures to adhere, sense and antisense ODNs were introduced using LipofectAMINE 2000 (Invitrogen). The cells were cultured for an additional 24 h before fixing with 4% paraformaldehyde and staining with either anti-MyoD or anti-Barx2 antibodies as described below.
Retroviral Transduction of Barx2 in Limb Bud Cultures-Retroviral vectors were constructed that contain either full-length mouse Barx2 cDNA or a cDNA fragment that encodes the homeodomain, Barx basic region, and COOH-terminal activation domain (HD-BBR-C) (19). A control retroviral vector was constructed that contains the green fluorescent protein gene. The retroviral vector was based on the murine embryonic stem cell virus with modifications as described previously (20). The vector contains a truncated version of the phosphoglycerate kinase promoter flanked by the chicken lysozyme insulator sequence, driving expression of the green fluorescent protein or Barx2 cDNAs. Retroviral particles were packaged in COS1 cells as described previously (21). Supernatant containing retroviral particles was collected after 48 h, filtered, and used to infect primary limb bud cells. Aliquots of 2 ϫ 10 6 dissociated E10.5 mouse limb bud cells were incubated with 2 ml of filtered retroviral supernatant for 2 h at 4°C. The cells were plated in micromass conditions as described above, incubated for 1 h, and then supplemented with 200 l of CMRL-1066 containing 10% FBS and 200 l of retroviral supernatant. After 24 h the medium was changed to CMRL-1066 containing 2% FBS and the cells were cultured for a further 24 h. Cells were then fixed with 4% paraformaldehyde and either photographed at ϫ10 magnification using phase-contrast optics, or stained with anti-MyoD antibodies as described below.
Immunohistochemistry-Primary limb bud cultures and C3H10T1/2 cell cultures were fixed with 4% paraformaldehyde and then permeabilized for 1 min with ice-cold acetone. Cultures were then stained with polyclonal antibodies to either Barx2 or MyoD (Santa Cruz), or with monoclonal antibodies to smooth muscle (SM)-␣-actin (Sigma) at 1:200 dilutions. Secondary antibodies were conjugated with either fluorescein isothiocyanate or rhodamine (Molecular Probes).
Stable Transfection of Barx2 in C2C12 Cells and Differentiation of C2C12 Cells into Myoblasts-1 ϫ 10 7 C2C12 cells were electroporated with either 10 g of linearized Barx2/pcDNA3 expression vector or pcDNA3 control plasmid. The Barx2 expression plasmid contains an in-frame NH 2 -terminal Myc tag and was described previously (16). Cells were selected for 14 days with 1 mg/ml G418 to establish a population of cells that carried the plasmid at varying integration sites. Cells were tested for expression of Barx2 after 3-4 passages by Western blotting and reverse transcriptase-PCR. Experiments were performed with low passage stocks, as Barx2 levels declined significantly after six passages. To examine differentiation of the transfected C2C12 cells, cultures were incubated in Dulbecco's modified Eagle's medium with 2% horse serum and examined for myotubes at 24, 48, and 96 h. At 48 h, the cultures were fixed for 10 min with 4% paraformaldehyde and photographed at ϫ5 magnification using phase-contrast optics.
Quantitative Western Blotting-To examine protein expression in pcBarx2-and pcDNA3-transfected C2C12 cultures, proliferating cells were harvested at 80% confluence and lysed in a hypotonic buffer containing 1% Nonidet P-40 (22). Protein concentration was determined by Bradford assay (Bio-Rad) and 20-g aliquots were separated by SDS-PAGE, transferred to polyvinylidene difluoride membrane (Invitrogen), and blotted using either polyclonal antibodies to Barx2 (Santa Cruz) or monoclonal antibodies to SM-␣-actin, ␤-actin, or desmin (Sigma). Blots were developed using chemiluminescent substrate (Novex) and autoradiographed. The intensity of the protein bands was determined by densitometry using NIH Image software and a ratio of either SM-␣-actin or desmin protein to ␤-actin protein was determined for each sample. The average increase in SM-␣-actin expression was determined from three replicate experiments.
Isolation of the Murine Barx2 Gene-The murine Barx2 gene was isolated from a 129/SvJ murine genomic BACmid library in the pveloBAC vector after screening with a PCR probe derived from the 5Ј end of the Barx2 cDNA sequence (Genome Systems Inc). Several overlapping restriction fragments containing the Barx2 gene were mapped using Southern blot analysis and sequenced.
Construction and Mutagenesis of Barx2 Promoter/Luciferase Reporter Plasmids and MRF Expression Plasmids-Barx2 promoter/luciferase reporter constructs were generated in the promoterless pGL3basic vector (Promega). The longest promoter construct contained a 1812-base pair KpnI-SacI fragment from the 5Ј end of the murine Barx2 gene. Additional reporter plasmids containing truncated versions of the Barx2 promoter, or internal deletions of promoter regions, were generated by cleavage at restriction sites or by PCR amplification. Mutation of the E-box element (E2) and the CArG box/SRE within the 0.44-kb Barx2 promoter construct was performed using the QuikChange protocol (Stratagene). All constructs were confirmed by sequencing. Synthetic promoter constructs were prepared by cloning oligonucleotides corresponding to a 58-bp segment of the Barx2 promoter (referred to here as the myogenic regulatory region (MRR)) into a modified version of the pGL3basic vector containing a minimal TATA box promoter and initiator sequence (21). In addition, two mutant versions of the MRR were generated in which either the E2 element or the CArG box/SRE were mutated. The sequences of these oligonucleotides with the E-box and CArG box motifs shown in bold, are as follows: MRR 5Ј-gctccgcacctggccctgcaggaagtgcgcgctgattgacagctgcggtgtcccaaaaaggct-3Ј; ⌬E2 5Ј-gctccgcacctggccctgcaggaagtgcgcgctgattgagtcctgcggtgtcccaaaaaggct-3Ј; ⌬CArG box 5Ј-gctccgcacctggccctgcaggaagtgcgcgctgattgacagctgcggtgttttcaaaaattct-3Ј.
To construct the MyoD and myogenin expression plasmids, we first isolated full-length cDNAs from E14 embryonic mouse limb RNA by reverse transcriptase-PCR. The cDNAs were cloned into a modified version of the pcDNA3 vector (Invitrogen) that contained an NH 2terminal Myc epitope tag. Proteins of the appropriate molecular weights were produced by in vitro translation using the Promega TNT coupled transcription/translation kit.
Cellular Transfection of C3H10T1/2 and C2C12 Cells-C3H10T1/2 were cultured in Dulbecco's modified Eagle's medium containing 10% FBS (Invitrogen) and antibiotics (penicillin and streptomycin). C2C12 cells were cultured in Dulbecco's modified Eagle's medium with 20% FBS and antibiotics. Both cell lines were obtained from American Type Culture Collection. C3H10T1/2 cells were placed in 6-well tissue culture plates at an initial density of 1 ϫ 10 5 cells/well and transfected with 0.5 g of each luciferase reporter construct. C2C12 cells were placed in 24-well tissue culture plates at an initial density of 1 ϫ 10 4 cells/well and co-transfected with 200 ng of each luciferase reporter construct and 200 ng of each pcDNA3 expression plasmid. All transfections used FuGENE reagent (Roche Molecular Biochemicals) and included the LacZ reporter CMV␤gal (Clontech) at one-tenth of the total DNA amount to provide an internal reference for transfection efficiency. Cellular transfection and assay conditions were otherwise as described previously (21). All experiments were performed in duplicate and the data shown were derived from at least three independent experiments. For co-transfection of C3H10T1/2 cells with the pEYFP1 plasmid (Clontech) and pcDNA3, pcBarx2, or pcMyoD plasmids, 1 ϫ 10 5 cells were seeded in 60-mm plates and transfected with 2 g of each plasmid using FuGENE (Roche Molecular Biochemicals).
Gel Mobility Shift Experiments-Oligonucleotide probes were synthesized that corresponded to each of the two E-box elements, and the CArG box/serum response element (SRE) from the murine Barx2 gene. In addition, mutant versions of these sequences were designed in which the core motifs were disrupted. Double-stranded probes were end-labeled with polynucleotide kinase and [␥-32 P]ATP (3000 Ci/mmol) (Du-Pont). Probes were purified by elution from an 8% polyacrylamide gel and their specific activity was determined. 25,000 cpm (ϳ10 fmol) of each probe was used in gel mobility shift experiments with C2C12 cell nuclear extracts, or in vitro translated proteins, as described previously (19). C2C12 cell extracts were prepared from proliferating C2C12 myoblasts cultured in high serum conditions (20% FBS) or from myotubes that were prepared by culture in low serum conditions (2% horse serum) for 4 days. Although C2C12 cells have been reported to differentiate in as little as 24 h after serum withdrawal, the slow rate of differentiation of our C2C12 cells is consistent with the known variation in the differentiation capacity of different stocks of C2C12 cells (23). 2 Nuclear extracts were prepared as described previously (22). For semiquantitative gel shift experiments, equal amounts of C2C12 nuclear protein from C2C12 myoblasts or myotubes were used in each binding reaction.
Relative binding of the protein to the E-box probes was determined by measuring the intensity of the probe-protein complexes formed in gel mobility-shift experiments using a PhosphorImager (Amerhsam Biosciences). The relative binding data shown in Fig. 6 were derived from four independent experiments using different C2C12 extract preparations.

Inhibition of Barx2 Expression Prevents Myotube Formation in Primary Limb Bud Cell Cultures-To determine whether
Barx2 is involved in the differentiation of skeletal myoblasts, we inhibited Barx2 expression in primary myoblasts using morpholino antisense ODNs and then examined their ability to form myotubes. Morpholino ODNs are RNase H-independent, resistant to nucleases, and have a good targeting predictability (24,25). For this study, limb bud mesenchymal cells from E10.5 mouse embryos were cultured in micromass conditions; these micromass cultures have been shown to contain cells that are committed to either a myoblast or chondrocyte fate (26). The limb buds were dissected, dissociated, and cultured in low serum under micromass conditions as described under "Experimental Procedures." The cultures were treated immediately after plating with either Barx2 antisense ODNs or control sense ODNs and examined 24 h later for formation of myotubes and for expression of Barx2 and MyoD.
As shown in Fig. 1, cultures treated with ODNs antisense to Barx2 showed negligible expression of Barx2 ( Fig. 1, top right panel) and formed few if any myotubes within the 24-h culture period. Many cells showed nuclear expression of MyoD, however, these cells were not fused and did not show the characteristic elongated nuclei of myotubes (Fig. 1, bottom right panel). In contrast, cultures that were treated with the sense ODNs contained many cells that expressed Barx2 and MyoD and that were fused into multinucleated myotubes (Fig. 1, left panels).
These results indicate that Barx2 expression is required for myoblasts to differentiate into myotubes in culture. Immunohistochemical staining showed that Barx2 antisense-treated cultures do not express Barx2, but retain appreciable, albeit diffuse, expression of MyoD. The observation that MyoD staining in antisense-treated cultures is diffuse relative to the control cultures may be due in part to the condensed and elongated morphology of nuclei in myotubes.
Barx2 Overexpression Accelerates Myotube Formation in Primary Limb Bud Cells-To determine whether overexpression of Barx2 can promote myotube formation, a retroviral system was used to overexpress Barx2 proteins in limb bud cultures. Micromass cultures were infected with retroviral constructs that express either full-length Barx2 protein or a fragment of Barx2 containing the homeodomain, Barx basic region (BBR), and the carboxyl-terminal domain (Barx2/HDBBRC). The Barx2/HDBBRC fragment contains the DNA binding and transactivation domains of Barx2 and was previously shown to activate the promoter of the Barx2 target gene N-CAM, which is a cell adhesion molecule known to influence myotube fusion (19,27). Control cultures were infected with a retroviral construct expressing green fluorescent protein. The cells were cultured in micromass conditions (28) for 24 h in media containing 10% serum to allow integration of the virus into dividing cells and the expression of Barx2 proteins. The cultures were then incubated for a further 24 h in media containing 2% serum to induce myotube formation.
As shown in Fig. 2, micromass cultures that were infected with retroviral constructs expressing either full-length Barx2 (not shown) or the Barx2/HDBBRC fragment (Fig. 2, top right panel) formed many myotubes within 24 h of culture. In contrast, cells infected with the control retrovirus produced very few myotubes within 24 h (Fig. 2, top left panel). Immunohistochemical staining with antibodies to MyoD revealed that only MyoD-positive cells fused to form myotubes. Thus Barx2 can-FIG. 1. Barx2 is required for the formation of myotubes in primary limb bud mesenchymal cultures. Limb bud cells from E10.5 embryos were cultured in micromass conditions and treated for 6 h with morpholino ODNs that were antisense to Barx2, or with control sense ODNs. Cells were then cultured for 24 h in low-serum conditions, fixed, and stained with polyclonal antibodies to either Barx2 or MyoD. Cultures treated with sense ODNs formed myotubes and expressed both Barx2 and MyoD at high levels (left panels). In contrast cultures treated with Barx2 antisense ODNs did not form myotubes and did not express Barx2, however, MyoD expression is still apparent (right panels).
FIG. 2. Barx2 promotes the formation of myotubes in primary limb bud mesenchymal cultures. Retroviral constructs that express either green fluorescent protein (control), or a fragment of Barx2 containing the homeodomain, Barx basic region, and the carboxyl-terminal activation domain (Barx2/HDBBRC), were used to infect dissociated limb bud cells from E10.5 mouse embryos. After infection, limb bud cells were cultured in micromass conditions for 24 h in high-serum media and then incubated for a further 24 h in low-serum media to induce myotube formation. Cultures that were infected with the Barx2/HDB-BRC-expressing retrovirus formed many myotubes after 24 h of culture in low-serum, whereas those infected with the control retrovirus did not (top panels). Immunostaining of cultures with polyclonal antibodies to MyoD showed that only MyoD-positive cells fused to form myotubes and that the number of MyoD-positive nuclei is similar in both control and Barx2/HDBBRC-infected cultures (middle panels). Immunostaining with monoclonal antibodies to SM-␣-actin (green) showed that cells infected with the Barx2/HDBBR-expressing retroviral construct express higher levels of SM-␣-actin. Nuclei are counterstained with 4,6diamidino-2-phenylindole (blue). not promote myotube fusion in the absence of MyoD (Fig. 2, middle panels). Moreover, when the number of MyoD-positive cells was counted in three separate fields, their numbers did not change after infection with the Barx2 construct (data not shown). Thus Barx2 does not induce ectopic expression of MyoD. Together these results indicate that Barx2 acts downstream of, but does not substitute for, MyoD activity. The myotubes that formed after infection of limb bud cells with the Barx2/HDBBRC-expressing retrovirus also showed increased expression of SM-␣-actin, an early marker of skeletal muscle differentiation (29,30) (Fig. 2, bottom panels).
Barx2 Promotes Myotube Formation in C2C12 Myoblasts but Not in C3H10T1/2 Fibroblasts-The results of both Barx2 inhibition and overexpression experiments in primary limb bud cultures suggest that Barx2 acts downstream of MyoD to promote the differentiation of cells that are already committed to the myoblast fate. To formally test whether Barx2 acts upstream or downstream of MyoD, we compared the effects of overexpressing Barx2 in MyoD-positive C2C12 myoblasts and in MyoD-negative C3H10T1/2 mesenchymal cells.
A population of C2C12 myoblasts that overexpress Barx2 was generated by stable transfection of a pcBarx2 expression plasmid, and a control population was established by stable transfection of the empty pcDNA3 vector. Overexpression of Barx2 protein in pcBarx2-transfected cells was confirmed by Western blotting with anti-Barx2 antibodies (see Fig. 3D). The effect of Barx2 overexpression on myotube formation was examined by plating pcBarx2-transfected and control cells at the same initial density and culturing in media containing 2% horse serum. Under these conditions, cultures of pcDNA3-transfected C2C12 cells differentiated relatively slowly, forming myotubes after 4 days (Fig.  3A, left panel). In contrast, C2C12 cells that were transfected with the pcBarx2 plasmid showed significantly accelerated differentiation, forming myotubes within 2 days of serum withdrawal (Fig.  3A, right panel). These results are similar to those obtained in primary limb bud cultures and provide further evidence that expression of Barx2 promotes the differentiation of MyoD-expressing myoblasts.
To test whether Barx2 is also involved in myoblast determination, we next examined the effect of overexpressing Barx2 in the mesenchymal progenitor cell line C3H10T1/2 that does not express MyoD. Ectopic expression of MyoD was previously shown to convert C3H10T1/2 cells into skeletal myoblasts that fuse into myotubes when cultured in low serum conditions (31). If Barx2 were able to induce the expression of MRFs such as MyoD it may be expected to convert C3H10T1/2 cells into myoblasts. C3H10T1/2 cells were transiently transfected with either the pcBarx2 plasmid, a pcDNA expression plasmid containing the MyoD cDNA (pcMyoD), or the empty vector. To identify transfected cells, a plasmid encoding enhanced yellow fluorescent protein (EYFP) was co-transfected with the expression vectors. The cells were then incubated in media containing 2% horse serum and the formation of myotubes by the EYFPlabeled cells was examined over the next 4 days. Consistent with previous reports (31), ectopic MyoD expression induced myotube formation in nearly all of the transfected cells after 4 days of serum deprivation (Fig. 3B, right panel). In contrast, Barx2-transfected C3H10T1/2 cells did not form myotubes (Fig.  3B, middle panel). These results indicate that, unlike MyoD, Barx2 is not sufficient to specify a skeletal myoblast fate when expressed in multipotent mesenchymal cells.
Interestingly, when pcBarx2-transfected and control pcDNA3-transfected C3H10T1/2 cells were grown in high serum conditions, pcBarx2-transfected cells showed morphological changes that are consistent with the differentiation of smooth muscle cells (SMCs) or myofibroblasts, including cell spreading and formation of stress fibers (Fig. 3C) (32,33). C3H10T1/2 cells and other fibroblastic cell lines have been previously reported to differentiate into both SMCs and myofibroblasts in response to particular signals (34 -36). Using immunohistochemistry, we found that Barx2-transfected FIG. 3. Barx2 promotes myogenic differentiation of C2C12 myoblasts but not of C3H10T1/2 fibroblasts. A, two populations of stably transfected C2C12 cells were generated by transfection with either the pcBarx2 or pcDNA3 plasmid and selection for 14 days with 1 mg/ml G418. Transfected cells were incubated in reduced serum media for 2 days, and then examined for myotube formation using phase-contrast optics. The pcBarx2 transfected cell population formed myotubes within 2 days of serum withdrawal, whereas pcDNA3 transfected cells did not. B, C3H10T1/2 cells were transiently co-transfected with either pcBarx2 or pcDNA3, and a reporter plasmid that encodes EYFP. The morphology of live transfected (fluorescent) cells was examined after 4 days of culture in reduced serum media. Under these conditions, MyoD, but not Barx2, was sufficient to induce differentiation of C3H10T1/2 cells into myotubes. C, C3H10T1/2 cells were transiently transfected with either a Barx2 expression plasmid (pcBarx2) or empty pcDNA3 plasmid, and stained with monoclonal antibodies to SM-␣-actin (red) after 4 days of culture in high serum media. Nuclei were counterstained with 4,6-diamidino-2-phenylindole (blue). Cells transfected with pcBarx2 are flattened, contain stress fibers, and stain more intensely for SM-␣-actin. D, equal aliquots of protein from C2C12 cells stably transfected with either pcDNA (Ϫ) or pcBarx2 (ϩ) were analyzed by Western blotting using Barx2 antibodies for increased expression of Barx2 (arrowhead), and for expression of the muscle markers SM-␣-actin and desmin. ␤-actin was used as a standard. The expression of SM-␣-actin in pcBarx2 (ϩ) transfected cells was an average of 1.7-fold higher than in pcDNA3 transfected cells (Ϫ) in three replicate experiments (see arrow at left).

C3H10T1/2 cells contained increased amounts of SM-␣-actin,
which is also consistent with these cellular phenotypes (Fig.  3C). These results suggest that in the absence of MyoD expression, Barx2 influences the differentiation of mesenchymal progenitors into contractile cell types other than skeletal myoblasts.
In these studies, Barx2 overexpression increased the expression of SM-␣-actin in all three cellular systems tested: primary myoblasts, C2C12 myoblasts, and C3H10T1/2 cells. To quantify this effect, we examined the expression of two cytoskeletal proteins that are regulated during myogenesis, SM-␣-actin and desmin, in Barx2-and pcDNA3-transfected C2C12 cells. Quantitative Western blotting was used to measure the amount of SM-␣actin, desmin, and ␤-actin protein in extracts of cells grown in high serum conditions (Fig. 3D). After normalization to ␤-actin, the amount of SM-␣-actin was, on average, 1.7-fold higher in pcBarx2transfected cells than in pcDNA3-transfected cells in three separate experiments. In contrast the expression of desmin did not change significantly (Fig. 3D). This indicates that Barx2 can activate particular genes that are associated with the differentiation of multiple contractile cell types. Collectively our studies in primary limb bud cultures and cell lines suggest two functions for Barx2 in muscle differentiation. In committed myoblasts that express MyoD, Barx2 promotes, and is required for, formation of myotubes, whereas in multipotent mesenchymal cells that lack MyoD, Barx2 may induce a smooth muscle or myofibroblast phenotype.
Identification of the Murine Barx2 Gene-The results of the studies described above suggest that Barx2 acts downstream of MyoD during myoblast differentiation. To determine whether MyoD or other MRFs may directly regulate the expression of Barx2, we isolated the murine Barx2 gene and characterized its promoter. The Barx2 gene was isolated by screening a BACmid murine genomic library using a probe from the 5Ј end of the Barx2 cDNA sequence. Two clones of ϳ80 kb containing the Barx2 gene were analyzed by restriction endonuclease digestion and Southern blotting. Several overlapping restriction fragments containing the Barx2 gene were mapped and sequenced, revealing four exons spanning greater than 30 kb (Fig. 4). The size of the first intron was not determined precisely; however, Southern blots indicated that it spans at least 20 kb. In addition, a 3-kb segment upstream of the first exon was isolated and sequenced.
Comparison of the sequence of the murine Barx2 gene with the recently published human Barx2 gene (37) revealed considerable similarity over a 1-kb region upstream of the translation initiation codon. In particular, the sequence that corresponds to the transcription start site (TSS) in the human gene (37) is conserved, suggesting that transcription begins at the same site in the murine gene (Fig. 5A). DNA sequence analysis of the region upstream of the predicted TSS revealed potential bind-ing sites for a number of transcription factors, including sequences that match the Sp1 consensus motif (G(A/G)GGC(A/ G)GGG(A/T)). A sequence containing a perfect match to the Sp1 consensus sequence (GGGGCGGGGT) is located ϳ340 nucleotides upstream of the TSS; an additional cluster of motifs that partially match the Sp1 consensus is located ϳ80 nucleotides upstream of the TSS. Because no TATA box elements are contained within the 3-kb region upstream of the first exon, this GC-rich region is likely to represent the core promoter. Several other conserved sequence elements were identified within a 400-bp segment upstream of the TSS. These include two E-box motifs, recognition sites for MEF and Ets family proteins, and a noncanonical CArG box/SRE (Fig. 5A).

The Murine Barx2 Promoter Contains Enhancer and Repressor Sequences That Function in Mesenchymal
Cells-To identify positive and negative regulatory regions of the Barx2 promoter, six deletion constructs were generated, either by PCR amplification or by cleaving the promoter at natural restriction sites (Fig. 5B). These promoter fragments were cloned into the promoterless pGL3basic luciferase reporter vector and their activities were examined after transfection into C3H10T1/2 cells.
The shortest Barx2 promoter construct tested (construct 0.44), containing the region between nucleotides Ϫ438 and ϩ80 relative to the predicted transcription start site, showed 7-fold greater activity than the pGL3basic plasmid suggesting that this region contains a core promoter (Fig. 5C). A fragment spanning nucleotides Ϫ1012 to ϩ80, relative to the TSS (construct 1.0), showed ϳ2-fold greater activity than construct 0.44, indicating that the region between nucleotides Ϫ438 and Ϫ1012 contains an enhancer. Addition of the segment between nucleotides Ϫ1012 and Ϫ1214 (construct 1.2) led to a 2-fold decrease in promoter activity, relative to construct 1.0, indicating that this 200-bp region contains a repressor (Fig. 5C). Two additional constructs, 1.5 and 1.7, spanning nucleotides Ϫ1446 to ϩ80 and Ϫ1732 to ϩ80, respectively, showed levels of activity similar to that of construct 1.2. However, deletion of the segment between nucleotides Ϫ1012 and Ϫ1462 from construct 1.7 (construct 1.7⌬KspI) increased activity ϳ8-fold. These results suggest that the region between Ϫ1012 and Ϫ1214 nucleotides contains a repressor that masks an enhancer activity located between Ϫ1462 and Ϫ1732 nucleotides upstream of the TSS (Fig. 5D).

MyoD and Myogenin Bind to an E-box Element within the Barx2 Promoter; Binding to this Element Is Increased during
Differentiation of C2C12 Cells-The proximal region of the Barx2 promoter contains recognition motifs for several transcription factors that are associated with muscle development. Among these elements are two sequences corresponding to the E-box consensus (CANNTG) that may be recognized by myo- genic basic helix-loop-helix factors such as MyoD or myogenin (see Fig. 5A). To determine whether these motifs can bind to MyoD or myogenin, we prepared two probes, designated E1 and E2, corresponding to each of the motifs, and tested their binding to in vitro translated MyoD and myogenin proteins in gel mobility shift experiments. MyoD and myogenin formed complexes with the E2 probe that were supershifted by antibodies to a Myc epitope tag located at the NH 2 terminus of the proteins (Fig. 6A). In contrast, neither MyoD nor myogenin bound to the E1 probe (data not shown).
The E1 and E2 probes were also tested for binding to nuclear extracts of C2C12 cells in gel mobility shift assays. Nuclear extracts were prepared from both proliferating C2C12 myoblasts and C2C12 cells cultured for 4 days in media containing 2% horse serum to induce myotube formation. The E2 probe bound to both myoblast and myotube C2C12 nuclear extracts; mutation of the CANNTG motif within the E2 probe abolished this binding (Fig. 6B). In contrast, the E1 probe showed negligible binding to C2C12 nuclear extracts (data not shown).
As shown in Fig. 6B, the complexes formed between the E2 probe and the myotube nuclear extract appeared more intense than those formed with the myoblast nuclear extract. To quantify the difference in binding, we measured the intensities of the probe-protein complexes that formed with the E2 probe using equivalent amounts of nuclear protein prepared from myoblast or myotube cultures in four separate gel mobility shift experiments (see "Experimental Procedures"). On average, the intensity of the E2 probe-protein complex formed with myotube nuclear extract was 1.9-fold greater than that formed with myoblast nuclear extract (Fig. 6B). These data indicate that the E2 element binds to proteins that are enriched during myogenic differentiation of C2C12 cells, and that these proteins are likely to include MyoD and myogenin.
Binding of Serum Response Factor (SRF) to the Barx2 Promoter-The proximal Barx2 promoter contains a noncanonical CArG box/SRE immediately adjacent to the MyoD/myogenin binding site. This element is a recognition motif for SRF, which has been shown in previous studies to interact with MyoD and myogenin at the promoter of particular muscle-specific genes (38,39). To determine whether SRF binds to the Barx2 promoter, we prepared a probe corresponding to the Barx2 CArG box/SRE and tested its ability to bind to nuclear extracts from C2C12 cells. The CArG box/SRE probe formed complexes with nuclear extracts from both C2C12 myoblasts and myotubes; this binding was eliminated by mutation of the CArG box consensus motif (Fig. 6C). Because the sequence of the Barx2 CArG box/SRE (CCCAAAAAGG) diverges from the canonical recognition motif for SRF (CC(A/T) 6 GG), we examined whether FIG. 5. Analysis of the Barx2 promoter and 5-untranslated region. A, sequence comparisons show extensive homology between the mouse and human Barx2 genes within the 1-kb region immediately upstream of the first exon. Nucleotides that are identical in the two genes are highlighted in boldface type. The region between Ϫ0.44 and Ϫ0.08 kb contains putative binding sites for Sp1 and a cluster of regulatory elements that are involved in muscle-specific transcription, including binding motifs for basic helix-loop-helix proteins (E-box), MEF2, Ets proteins, and serum response factor (CArG box/SRE) (indicated by boxes). The arrow indicates the predicted transcription start site. This sequence corresponds to GenBank TM accession number AY188085, nucleotides 1953-3014. B, six Barx2 promoter/luciferase reporter plasmids were constructed in the luciferase reporter plasmid pGL3basic by deletion of a 1.7-kb promoter fragment that contains the transcription start site. C, C3H10T1/2 cells were transfected with each promoter construct, together with the lacZ reporter plasmid CMV␤gal to provide an internal reference for transfection efficiency. Luciferase and ␤-galactosidase activities were measured 48 h post-transfection. The data are the average of four experiments, performed in duplicate. D, schematic representation of the Barx2 promoter depicting the core promoter, enhancer, and repressor activities that were identified in transfection experiments. Regulatory elements within the core promoter region are indicated by polygons: Sp1 (circle), E-box (rectangle), CArG box/SRE (triangle), and MEF2 (square).
antibodies to SRF could supershift or block the formation of complexes between the Barx2 CArG box probe and C2C12 nuclear extracts. The predominant probe-protein complex was supershifted by antibodies to SRF, suggesting that the complex contains SRF (see lane 5 in Fig. 6C).
In C2C12 cells, SRF regulates both mitogenic genes such as c-fos, as well as muscle-specific genes (40,41). However, SRF has been reported to bind with higher affinity to the CArG box/SRE from the c-fos gene promoter than to the noncanonical CArG boxes/SREs present in many muscle-specific genes (40). To examine the relative strength of binding of SRF to the Barx2 CArG box/SRE, we compared binding of C2C12 nuclear extracts to the Barx2 CArG box/SRE probe as well as to a control probe corresponding to the c-fos CArG box/SRE (40). The Barx2 CArG box/SRE probe formed much less intense complexes than the c-fos CArG box/SRE probe when incubated with equivalent amounts of nuclear protein from C2C12 cells. Collectively these data suggest that both E-box binding proteins and SRF may regulate expression of Barx2 in skeletal muscle cells.
MyoD and Myogenin Activate the Proximal Barx2 Promoter in C2C12 Cells-The results of the gel mobility shift analyses described above indicated that the proximal Barx2 promoter region is bound by factors involved in muscle-specific gene regulation. To examine the role of these elements in the regulation of the Barx2 gene, C2C12 myoblasts were transfected with various Barx2 promoter constructs, together with expression plasmids that encode MyoD and myogenin (Fig. 7). Constructs 1.0 and 0.44 showed lower levels of basal activity in C2C12 cells than in C3H10T1/2 cells (see Fig. 5). However, co-transfection of a MyoD expression plasmid activated these constructs approximately 13-and 11-fold, respectively (Fig. 7). Expression of myogenin activated these two constructs 6-and 5-fold, respectively (Fig. 7). Deletion of a 220-bp SmaI restriction fragment from construct 0.44 (0.44⌬Sma) prevented its activation by both MyoD and myogenin. The 220-bp SmaI fragment alone (construct Sma) also functioned as a promoter in C2C12 myoblasts and was activated by MyoD and myogenin 7-and 4-fold, respectively (Fig. 7). These results indicate that the SmaI restriction fragment within the Barx2 proximal promoter, which contains Sp1, E-box, and CArG box/SRE motifs, is necessary and sufficient for promoter activation by MyoD, and to a lesser extent, by myogenin.
The E2 and CArG Box/SRE Motifs Are Required for Activation of the Barx2 Promoter by MyoD-To examine the role of the MyoD/myogenin binding site (E2) in activation of the Barx2 promoter by MyoD and myogenin, this element was mutated within the 0.44-kb Barx2 promoter construct. Because SRF has been shown to act cooperatively with MyoD (38), the SRF binding site (CArG box/SRE) was also mutated within the 0.44-kb construct. C2C12 cells were co-transfected with pGL3basic, the 0.44-kb Barx2 promoter construct, the 0.44⌬E2 construct, or the 0.44⌬CArG construct, together with pcDNA3, MyoD, or myogenin expression plasmids. As shown in Fig. 8A, mutation of either the E2 or CArG box/SRE element prevented activation of the 0.44-kb Barx2 promoter construct by MyoD and myogenin.
To determine whether these two elements are sufficient for activation by MyoD, we generated a synthetic construct containing a segment of the Barx2 promoter that spans the E2 and CArG box/SRE motifs. This 58-bp segment (hereafter referred to as the myogenic regulatory region, or MRR) was synthesized (lanes 1 and 2, arrow at left). This complex was supershifted by antibodies to SRF (lanes 3-5, see arrowhead at right). The Barx2 CArG probe formed a much less intense complex with the C2C12 extract than the c-fos CArG box/SRE, suggesting that the Barx2 CArG box/SRE has a lower affinity for SRF (compare lanes 3 and 6).  1-3, arrow at left). Antibodies to the Myc tag supershifted these complexes (lanes 4 -6, arrows at right). B, the E2 probe was also tested for binding to nuclear extracts prepared from C2C12 myoblast (MB) or myotube (MT) cultures. The probe formed more intense complexes with myotube extracts than with myoblast extracts (lanes 1 and 2) whereas a mutated version of the probe in which the E-box probe was eliminated (mE2) did not bind to either extract (lanes 3 and 4). The average increase in binding was determined by measuring the relative intensities of the complexes formed between the E2 probe and the myoblast or myotube extracts in four independent experiments. C, a probe corresponding to the CArG/SRE within the Barx2 promoter (CArG), and a mutant version in which the core motif was eliminated (mCArG), were tested for binding to C2C12 extracts. As a control, a probe containing the consensus CArG/SRE from the c-fos promoter was also tested for binding. The Barx2 CArG probe formed a complex with the C2C12 myoblast extract (MB) that was not formed with the mCArG probe and inserted upstream of a minimal TATA box promoter driving the luciferase reporter gene (see Fig. 8B). When transfected into C2C12 cells, this synthetic promoter construct (MRR) showed ϳ7-fold greater activity than that of the minimal promoter alone (construct pLuc). Co-transfection with a MyoD expression plasmid activated the MRR construct ϳ6-fold, relative to transfection of the empty pcDNA3 plasmid (Fig. 8B). In contrast, co-transfection of a myogenin expression plasmid did not significantly activate the MRR construct (data not shown), suggesting that activation of the Barx2 promoter by myogenin (see Fig. 7) involves additional sequences outside of the MRR.
To examine whether the E2 or CArG box/SRE elements are required for activation of the MRR by MyoD, we prepared constructs in which either the E2 or CArG box/SRE motif was mutated, and tested their ability to be activated by MyoD in cellular transfection experiments. Mutation of either the E2 or CArG box/SRE motif reduced the activity of the MRR promoter construct to the level of the minimal promoter alone. Similar to the results obtained when the E2 and CArG box/SRE elements were mutated in the native promoter construct, mutation of either element with the MRR was sufficient to prevent activation by MyoD. These results indicate that activation of the Barx2 promoter by MyoD requires an accessory factor that binds to the CArG box/SRE element, most likely SRF. DISCUSSION The homeobox protein Barx2 is expressed in several phases of mesenchymal tissue morphogenesis, including the formation of pre-chondrogenic and pre-muscle mesenchymal condensations and the differentiation of skeletal and smooth muscle (16 -18). Barx2 is also expressed in adult skeletal and smooth muscle, and its expression in C2C12 myoblasts is increased during differentiation into myotubes (18). In this study we showed that expression of Barx2 is required for the differentiation of MyoD-positive limb bud cells into myotubes, and that overexpression of Barx2 accelerates myotube formation by both primary myoblasts and C2C12 cells. However, overexpression of Barx2 did not increase the number of MyoD-positive cells in limb bud cultures. This result, together with the observation that overexpression of Barx2 does not convert MyoD-negative C3H10T1/2 cells into skeletal myoblasts, strongly suggests that Barx2 does not activate MyoD but rather acts downstream of MyoD. Importantly, the latter observation also indicates that Barx2 cannot substitute for MyoD activity.
Taken together, our results suggest that the role of Barx2 in myoblast differentiation may be distinct from that of other homeobox genes previously shown to influence muscle development, such as Pax3, Pax7, Lbx1, and Msx1. Each of these factors has been shown to act upstream of MRFs in myoblast determination. For example, Pax3 is expressed in myogenic progenitor cells and proliferating myoblasts and is required for the activation of MyoD and thus myogenesis (10). Lbx1 also acts upstream of MRFs to specify myoblasts and both induces, and is induced by, Pax3 (11). Msx1 antagonizes the activity of Pax3 by direct interaction leading to repression of myogenesis (12,42) and ectopic expression of Msx1 in C2C12 myotubes reduces the expression of MRFs leading to de-differentiation of myotubes (43). In contrast to the activities of these factors, our study indicates that Barx2 promotes myotube formation without inducing the expression of MyoD; thus Barx2 is the first homeobox gene shown to have a significant role in myoblast differentiation downstream of MyoD expression.
The mechanism by which Barx2 promotes myotube formation is yet to be determined. However, because the portion of the Barx2 protein that contains the DNA binding and activation domains was sufficient to promote myotube formation, it is likely that Barx2 activates particular genes that are required for the differentiation of skeletal myoblasts. For example, in this study we found that Barx2 up-regulated SM-␣-actin expression in both primary limb bud cultures as well as in the C2C12 myoblast cell line. Moreover, the observation that Barx2 increased SM-␣-actin expression in proliferating myoblasts suggests that Barx2 induces cellular changes that prime skeletal myoblasts for differentiation. Other downstream targets of Barx2 that are likely to influence myoblast fusion include CAMs. Barx2 was previously shown to regulate the N-CAM promoter and to induce cadherin-6 expression (19,44). Previous studies have shown that both N-CAM and particular cadherins, such as N-cadherin, can promote myoblast fusion (27,45). In addition, two new Ig family CAMs, CDO and BOC, were shown recently to play an essential role in myoblast fusion via FIG. 7. MyoD and myogenin activate the mouse Barx2 promoter in C2C12 myoblasts. A, Barx2 promoter constructs used to examine activation by MyoD and myogenin. The 0.44⌬Sma plasmid was generated by deleting a 220-bp internal SmaI fragment that contains Sp1, E-box, and CArG box elements. The Sma plasmid contains the 220-bp SmaI fragment upstream of the luciferase gene in pGL3basic. B, C2C12 cells were co-transfected with each of the Barx2 promoter reporter plasmids and either empty pcDNA plasmid or a pcDNA3 expression plasmid encoding MyoD or myogenin. The CMV␤gal plasmid provided an internal reference for transfection efficiency. Luciferase and ␤-galactosidase activities were measured 48 h post-transfection. The data are the average of six experiments, performed in duplicate. their heterophilic interaction (46,47). In future studies, it will be important to determine whether these and other CAMs are direct targets of Barx2 in myoblasts.
Our analysis of the Barx2 gene promoter provides a molecular mechanism for the control of Barx2 expression during myogenic differentiation. We identified a proximal region of the Barx2 promoter containing an E-box (E2) that binds to both MyoD and myogenin, and additional binding sites for SRF and MEF2 proteins. Binding of nuclear proteins to the E2 element was found to increase during differentiation of C2C12 myoblasts, suggesting that the binding of MRFs to this promoter element mediates the previously reported induction of Barx2 expression during myotube formation (18). Both the E2 element and an adjacent CArG box/SRE, which binds to the SRF are required for activation of the Barx2 promoter by MyoD. This result is similar to previous observations that both E-box and CArG/SRE motifs are involved in activation of the SM-␣actin and dystrophin genes by MyoD (38,48), and suggests that MyoD and SRF cooperatively activate Barx2 expression in C2C12 cells.
SRF also plays a central role in the differentiation of smooth muscle (41,49,50) and has been shown to interact with various homeodomain proteins including variants of both Barx1 and Barx2 (18,(51)(52)(53)(54). In particular, Barx2 can stimulate the binding of SRF to its cognate CArG box/SRE motif (18), and interaction of chicken Barx1b with SRF regulates smooth muscle cell-specific expression of the ␤-tropomyosin gene (54). Our studies also suggest a role for Barx2 in SMC differentiation. Overexpression of Barx2 in C3H10T1/2 cells increased SM-␣-actin expression and promoted cell spreading and formation of stress fibers. These morphological changes are consistent with differentiation of SMCs or myofibroblasts, cells that have contractile properties like smooth muscle, but retain characteristics of fibroblasts (33,55,56). Because the SM-␣-actin gene also contains a CArG box that binds to SRF (41), it is possible that induction of SM-␣-actin after overexpression of Barx2 is mediated by a complex of Barx2 and SRF. Thus both the regulation of Barx2 by SRF, and the interaction of Barx2 with SRF, may be part of the muscle differentiation program.
Overall, our studies suggest that Barx2 regulates multiple muscle differentiation pathways. In committed myoblasts Barx2 expression can be induced by MRFs such as MyoD and myogenin and increased expression of Barx2 then promotes myotube formation. In uncommitted mesenchymal cells that do not express MRFs, Barx2 may promote differentiation into other contractile cell types such as SMCs. The mechanisms by which Barx2 influences the differentiation of both skeletal and smooth muscle could be elucidated by chromatin immunoprecipitation or microarray analysis to identify targets of Barx2 that are specifically regulated in each of these contexts. FIG. 8. The E2 and CArG/SRE elements are required for activation of the Barx2 promoter by MyoD and myogenin. A, constructs used to examine the role of the E2 and CArG elements in the Barx2 promoter. Each of these elements was eliminated from the 0.44 Barx2 promoter luciferase construct by site-directed mutagenesis. B, C2C12 cells were transfected with each of the Barx2 promoter constructs shown in A, together with the expression plasmids, pcDNA3, pcMyoD, or pcMyogenin. The CMV␤gal plasmid provided an internal reference for transfection efficiency. C, a synthetic promoter construct was prepared containing the E2 and CArG elements upstream of a TATA box promoter, together with two mutated versions of this construct in which each motif was eliminated. D, C2C12 cells were co-transfected with each of the promoter constructs shown in C, together with pcDNA3 or pcMyoD and the CMV␤gal plasmid. For all experiments, luciferase and ␤-galactosidase activities were measured 48 h post-transfection and the data shown are the average of four experiments, performed in duplicate.