MEF2B Is a Component of a Smooth Muscle-specific Complex That Binds an A/T-rich Element Important for Smooth Muscle Myosin Heavy Chain Gene Expression*

To understand smooth muscle-specific gene expression, we have focused our studies on the smooth muscle myosin heavy chain (SMHC) gene, a smooth muscle-specific marker. In this study, we demonstrate that the SMHC promoter region (−1594 to −1462 base pairs) containing the A/T-rich element can activate the heterologous thymidine kinase promoter in smooth muscle cells, but not in fibroblasts. Mutations of this A/T-rich element decreased SMHC promoter activity significantly. Both gel mobility shift assays and DNase I footprinting revealed that this region binds to specific protein complexes from smooth muscle nuclear extracts, whereas nuclear extracts from skeletal muscle and fibroblasts produced a different binding pattern. We also demonstrate that the protein complex obtained from smooth muscle nuclear extract reacts with MEF2B-specific antibody, but not with antibodies specific to MEF2A, MEF2C, or MEF2D, suggesting that only MEF2B protein binds to the A/T-rich element. Furthermore, MEF2B overexpression in smooth muscle cells up-regulated the SMHC promoter, suggesting that MEF2B is important for SMHC gene regulation. This is the first report demonstrating a role for MEF2 factors in smooth muscle-specific gene expression.

Smooth muscle cells have been the subject of intense study because abnormal growth and proliferation of smooth muscle cells are involved in the pathogenesis of both atherosclerosis (1) and restenosis following percutaneous transluminal coronary angioplasty or atherectomy. The hallmark of restenosis following percutaneous transluminal coronary angioplasty or coronary atherectomy is vascular smooth muscle cell proliferation and migration causing obstructive lesions (2). Numerous observations suggest that smooth muscle cells in vascular lesions undergo phenotypic modulations from a contractile to a synthetic phenotype. These proliferating smooth muscle cells secrete increased extracellular matrix and express several embryonic markers, but lose SM-2 myosin present in mature smooth muscle cells, more closely resembling immature fetal smooth muscle cells. These studies suggest that upon injury, adult smooth muscle cells recapitulate many aspects of embry-onic development during their remodeling. Therefore, a better understanding of smooth muscle development and differentiation is important to define the vascular disease process. From this standpoint, a central question in vascular biology relates to understanding the transcriptional control mechanisms underlying smooth muscle growth and differentiation.
To date, transcription factors that are responsible for smooth muscle commitment and differentiation or modulation of the smooth muscle cell phenotype have not been identified (3). The transcription factors that regulate skeletal muscle development and differentiation are better understood and may provide useful paradigms for elucidating control of gene expression in smooth muscle. In recent years, skeletal muscle has become the paradigm for understanding tissue-specific gene activation and cell differentiation due to the discovery of a set of master regulatory genes, namely the MyoD family, which includes MyoD, myogenin, myf-5, and MRF-4/herculin/myf-6 (4). The myogenic factors are unique in their abilities to orchestrate an entire program of skeletal muscle-specific gene activation when introduced into diverse cell types. However, the members of the MyoD gene family are not expressed in smooth muscle, and related helix-loop-helix transcription factors controlling smooth muscle cell differentiation have not yet been identified. This raises the possibility that there are other types of transcription factors involved in smooth muscle myogenesis.
A second class of transcription factors, namely MEF2 (myocyte-specific enhancer factor 2)/RSRF (related to serum response factor), has been implicated in striated muscle differentiation and transcriptional control (5). The MEF2 family includes four genes, MEF2A (6), MEF2B (7), MEF2C (8), and MEF2D (9 -11). Each of the MEF2 proteins cloned thus far appears to be subject to complex forms of regulation at different levels. Members of this gene family are transcribed in a wide range of cell types including skeletal, cardiac, and smooth muscles as well as brain and spleen. MEF2 was originally described as a DNA-binding activity present in differentiating myotubes (12). The MEF2-binding site, CTA(A/T) 4 TA(G/A), has been shown to be important for transcriptional regulation of many cardiac and skeletal muscle-specific genes (12)(13)(14)(15)(16). It has also been demonstrated that MEF2 transcription factors can interact with myogenic bHLH 1 proteins (17) to synergistically activate muscle-specific genes. Recent studies in the fruit fly Drosophila melanogaster showed that MEF2 is indispensable for muscle development since disruption of the single MEF2 gene, D-MEF2, causes lethality with abnormalities in cardiac, somatic, and visceral muscle development. This demonstrates that MEF2 plays an important role in all three muscle lineages (18,19). MEF2A, MEF2B, and MEF2D are also expressed (mRNA and protein) in cultured smooth muscle cells and in the adult rat aorta, and their expression is increased in neointimal smooth muscle cells during vascular remodeling (20). However, the exact role of MEF2 transcription factors in regulating smooth muscle gene expression remains to be determined.
Toward understanding smooth muscle-specific gene expression, we have recently isolated and characterized the rabbit smooth muscle myosin heavy chain (SMHC) gene promoter (21), whose expression is highly tissue-specific. We have shown that the promoter region extending to Ϫ2266 bp is highly active in cultured rat aortic smooth muscle cells, but not in other cell types. Promoter deletion analyses identified a region between Ϫ1548 and Ϫ1392 bp as important for high level promoter activity. This region includes a MEF2-like A/T-rich element located at Ϫ1540 bp that binds to a specific protein complex in nuclear extracts from vascular smooth muscle cells (21). The goal of this study was to examine the nature of protein binding to this element and its functional relevance to SMHC promoter activity. In this study, we demonstrate that an antibody specific to MEF2B protein supershifted the protein complexes formed on the A/T-rich region from smooth muscle nuclear extracts. However, MEF2B does not appear to bind this A/Trich element directly. We further show that overexpression of MEF2B protein in smooth muscle cells up-regulates SMHC gene promoter activity. This is the first report demonstrating a role for MEF2 in smooth muscle-specific gene expression.

MATERIALS AND METHODS
Cell Culture-Smooth muscle cells from rat thoracic aorta were isolated and cultured by a modification of the procedures described by Owens et al. (22). Male Sprague-Dawley rats (4 -6 weeks old) were anesthetized with an intraperitoneal injection of sodium pentobarbital, and the descending thoracic aorta was excised aseptically. Two vessels were placed in Hanks' balanced salt solution (Life Technologies, Inc.) with 1% antibiotic/antimycotic (Life Technologies, Inc.), cleaned free of adhering fat and connective tissue by blunt dissection, and opened longitudinally. After preincubation of the vessels for 15-25 min at 37°C in a 5% CO 2 and 95% air atmosphere in Hank's balanced salt solution in the presence of 1 mg/ml collagenase (2219 units/mg; Worthington) and 0.16 -0.2 mg/ml elastase (4.2 units/mg; Worthington) with 1% antibiotic/antimycotic, the adventitia was carefully stripped off under a dissecting microscope, and the luminal surface was scraped with the convex side of curved forceps to remove endothelial cells. The resulting aorta pieces were placed into fresh collagenase/elastase solution, minced into 1-2-mm 2 pieces, and incubated (37°C, 5% CO 2 and 95% air) for an additional 80 -110 min. The dissociated cells were separated from the undigested tissue by filtration through an 85-m stainless steel screen, and fetal calf serum (Life Technologies, Inc.) was added to a final concentration of 30%. The isolated cells were collected by sedimentation at 1500 rpm for 6 min and resuspended in Medium 199 (Life Technologies, Inc.) containing 10% fetal calf serum and the above antibiotics. Cells were seeded into plastic tissue culture dishes (Falcon) and cultured at 37°C in a humidified atmosphere of 5% CO 2 and 95% air with medium changes three times weekly. The confluent cultures were usually obtained after 4 -8 days. The confluent cultures were washed with 1 ϫ phosphate-buffered saline, harvested with 0.05% trypsin and 0.53 mM EDTA, and replated at a 1:5 or 1:6 split area ratio into 90-mm tissue culture plates. Rat aortic smooth muscle cells after the first passage were utilized for in vitro transfection studies and for isolation of nuclear extracts. For C 2 C 12 and Sol-8 myotubes, cells were allowed to propagate in growth medium for 8 -10 h, and then growth medium was removed and replaced with differentiation medium. C 2 C 12 and Sol-8 myotubes were harvested 36 -48 h after the fusion medium change.
Plasmid Construction-Unidirectional deletions of the SMHC promoter were produced using the Erase-a-base system (Promega) in the pJRCATX vector (21). Heterologous SMHC-TK promoter constructs were made by ligating a 133-bp fragment generated by PCR corresponding to the Ϫ1594 to Ϫ1462 bp region to pBLCAT5. This produced constructs containing both sense and antisense orientations. Orienta-tion of the SMHC fragment was confirmed by SnaBI digestion, which is unique to this fragment. The MEF2 expression vectors for MEF2C (pCDNA1-MEF2C) (11) and MEF2B (pCGNR2-MEF2B) were described earlier (23). Constructs pJRCATX, pBLCAT5, pBLCAT6, and pSV 2 CAT were used as positive and negative controls in transfection analyses (24).
Site-directed Mutagenesis of the A/T-rich (MEF2-like) Element in the SMHC Promoter-A single base substitution of the A/T-rich element was performed by a two-step symmetrical PCR procedure using a pair of complementary template-mismatched central primers containing the desired sequence (25). Briefly, the first PCR was done with a primer (5Ј-CCCGGGATCCCCTGTAGGGA-3Ј) that hybridizes to a region encompassing the BamHI site at Ϫ1584 bp and a primer (5Ј-GTAT-TCATATAAGACACC-3Ј) that includes the mutation of the A/T-rich sequence. A second reaction was performed with a primer (5Ј-TTCCT-GCAGGAATTCCCAGC-3Ј) that hybridizes to a region encompassing the EcoRI site at Ϫ1224 bp and a primer (5Ј-GGTGTCTTATAT-GAATAC-3Ј) that is the complementary sequence of the above-mentioned MEF2-like (A/T-rich) mutated sequence. The products of both reactions were mixed together, and a third reaction was carried out with the underlined primers. PCR products corresponding to the fulllength product was cloned once into plasmid pCRII (Invitrogen) and digested with BamHI and EcoRI, and then the BamHI-EcoRI and EcoRI-HindIII (which contains the proximal promoter region including the TATA box) fragments were subcloned into the pJRCATX vector. The presence of the correct mutations was verified by sequence analysis before cloning into pJRCATX.
Western Blot Analyses-For Western blots, equivalent quantities of nuclear extracts from (second passage) primary rat aortic smooth muscle cells were separated through a 10% SDS-polyacrylamide gel as described previously (20). Proteins were then transferred to nitrocellulose, incubated with polyclonal antibodies at 1:1000 dilutions, and detected with an enhanced chemiluminescence kit (ECL, Amersham Corp.). Antibodies to MEF2A (GST-MEF2A-(129 -263)) and MEF2B (GST-MEF2B-(88 -365)) were obtained from Dr. Yie-Teh Yu. MEF2A and MEF2C were from the laboratory of Olson and co-workers (20). The rabbit antisera to MEF2B and MEF2D were provided by Dr. Ron Prywes and have been described previously (23).
DNA Transfections-Transient transfections of the SMHC promoter constructs were performed using the calcium phosphate coprecipitation method (21). Briefly, duplicate dishes of cells (5 ϫ 10 Ϫ5 cells/dish) were transfected with 15 g of the SMHC constructs plus 5 g of pMSV-␤gal into cultures of rat aortic smooth muscle cells and NIH3T3 fibroblasts.
Chloramphenicol Acetyltransferase (CAT) Assays-The transfected cells were rinsed twice with cold Hanks' balanced salt solution and harvested in release buffer (40 mM Tris-HCl, pH 7.4, 1 mM EDTA, and 150 mM NaCl). The cells were lysed in 250 mM Tris-HCl, pH 7.5, by three cycles of freeze-thawing, and the cell lysate was used for the enzyme assays. ␤-Galactosidase assays were performed on 30% of the cell extract to provide values of relative transfection efficiency. The resulting values were used to normalize the amount of extract added to the subsequent CAT assays. The reaction mixture contained 20 l of the extract in a 150-l assay containing 4 mM acetyl coenzyme A (Sigma) and 0.05 mCi of [ 14 C]chloramphenicol (54 mCi/mmol; Amersham Corp.) in 250 mM Tris-HCl, pH 7.8. The reaction was stopped and extracted with 1 ml of ethyl acetate. The organic phase was dried, suspended in 20 l of ethyl acetate, and spotted onto a silica gel thin-layer plate. The chromatogram was developed in a chloroform/methanol (95:5) system, dried, and autoradiographed. For quantitating the conversion of chloramphenicol to its acetylated forms, the spots were excised, and the radioactivity was measured in a liquid scintillation counter. CAT activity for each promoter construct was determined by at least three separate transfection experiments using two different plasmid preparations. The construct pSV 2 CAT was used as a positive control (26).
Preparation of In Vitro Translated MEF2 Proteins, Vascular Smooth Muscle Nuclear Extracts, and Gel Mobility Shift Assays-In vitro transcriptions and translations were performed together in a coupled rabbit reticulocyte lysate system (Promega) according to the manufacturer's recommended conditions. The T7 promoter was used in plasmids pCDNA1-MEF2A, pCDNA1-MEF2C, pCDNA1-MEF2D1a or pCDNA1-MEF2D1b, and pCI-MEF2B (described in Refs. 8, 11, and 27) to give transcription products followed by translation products of MEF2A, MEF2C, MEF2D, and MEF2B, respectively. Nuclear extracts from cultured rat aortic smooth muscle cells were prepared essentially as described by Dignam et al. (28). The protein concentration was determined by the Bradford assay (29). Gel mobility shift assays were performed as described previously (21). A 133-bp fragment that contains the A/T-rich element (see Fig. 1B) was 3Ј-end-labeled with [␣-32 P]dATP using DNA polymerase (Klenow fragment, New England Biolabs Inc). Briefly, 10 -20 g of nuclear extract was incubated in a total volume of 30 l for 30 min at room temperature in the presence of 1 ng of radiolabeled fragment or double-stranded oligonucleotides and 2 g of poly(dI-dC) and analyzed by electrophoresis on 4 and 5% polyacrylamide gels. Competition experiments were performed with a 100 -500-fold molar excess of specific or nonspecific unlabeled DNA oligomers. The oligonucleotides used as probes or competitors in the gel mobility shift assays were as follows: SMHC-A/T-rich, CTGGGGGTATTAATATAAGACACCC; A/T-rich (which contains three nucleotide substitutions), CTGGGGGCATTCC-TATAAGACACCC; ⌬A/T-rich (which contains a point mutation), CTG-GGGGTATTCATATAAGACACCC; E-box (E3), CCACGCACACGTG-GTTGCAG; MCK-MEF2, GCTCTAAAAATAACCCTG; and ␣MHC-MEF2, CTTTCAGATTAAAAATAACTAAGG. To determine whether MEF2 protein families bind to the A/T-rich element in the 133-bp fragment, antibodies that react specifically to the MEF2A, MEF2B, MEF2C, and MEF2D isoforms were incubated together with the DNA and nuclear extracts in the gel shift reaction. MEF2A, MEF2B, MEF2C, or MEF2D antibodies (1-2 l) were added to a 30-l gel shift reaction and incubated at 4°C for 2 h. Following electrophoresis, the gels were stained in 10% acetic acid and 30% methanol, dried, and autoradiographed.
DNase I Footprinting-DNase I footprinting was performed on the Ϫ1594 to Ϫ1462 bp promoter region using nuclear extracts from cultured rat aortic smooth muscle cells. Briefly, a 133-bp probe labeled on only one strand was prepared by PCR using an end-labeled sequence primer. 120 g of smooth muscle nuclear extract and 4 ng of the labeled probe were incubated in the same buffer as that used in the gel mobility shift assays for 10 min at room temperature and for an additional 10 min on ice. 50 l of DNase I (Worthington) at a concentration of 10 g/ml in 10 mM Tris-HCl, pH 8.0, 10 mM MgCl 2 , and 1 mM CaCl 2 was added to the binding mixture and incubated for 90 s at room temperature. This reaction was terminated by the addition of 100 l of stop solution (200 mM NaCl, 30 mM EDTA, and 1% SDS), phenolized, and ethanol-precipitated. The samples were heat-denatured and loaded onto a 6% sequencing gel. The A ϩ G ladders were generated by the Maxam-Gilbert chemical sequencing method (30).

The Rabbit SMHC Promoter Region (Ϫ1594 to Ϫ1462 bp) Can Activate the Heterologous TK Promoter in Smooth Muscle
Cells-Using promoter deletion mapping analyses, we have previously demonstrated that the region between Ϫ1594 and Ϫ1392 bp is essential for high level SMHC promoter activity in cultured smooth muscle cells (21). Deletion of this region resulted in a 52% decrease in CAT activity in smooth muscle cells. This region includes an A/T-rich element (5Ј-TATTA-ATATA-3Ј) located at Ϫ1540 bp (Fig. 1A). To determine whether this region contained positive regulatory (enhancerlike) elements, the DNA from the Ϫ1594 to Ϫ1462 bp region was cloned upstream of the heterologous TK promoter-CAT plasmid (pBLCAT5) both in the sense and antisense orientations. These heterologous promoter constructs together with control vectors were transiently transfected into primary cultures of rat aortic smooth muscle cells and NIH3T3 fibroblast cells. The basal TK promoter pBLCAT5 produced low levels of CAT activity in smooth muscle and NIH3T3 fibroblasts (Fig.  2B). On the other hand, inclusion of the SMHC promoter region (Ϫ1594 to Ϫ1462 bp) increased TK promoter activity significantly in smooth muscle cells. The sense orientation produced an 8 -10-fold increase in TK-CAT expression, whereas the antisense orientation produced an increase of 3-4-fold compared with the basal TK promoter pBLCAT5 (Fig. 2). However, the same heterologous promoter constructs did not produce elevated CAT expression in NIH3T3 fibroblasts (Fig. 2). Thus, the SMHC promoter region (Ϫ1594 to Ϫ1462 bp) appears to function as a positive regulatory element in smooth muscle cells.

DNase I Footprinting Reveals That the A/T-rich Element Located at Ϫ1540 bp Binds Nuclear Proteins from Smooth
Muscle Cells-The SMHC promoter region (Ϫ1594 to Ϫ1462 bp) described above includes an A/T-rich element (5Ј-TATTA-ATATA-3Ј) located at Ϫ1540 bp (Fig. 1A). To determine the nature of protein-binding sites within the Ϫ1594 to Ϫ1462 bp DNA region, in vitro DNase I footprint analysis was carried out using nuclear extracts from cultured rat aortic smooth muscle cells. A 133-bp probe (Ϫ1594 to Ϫ1462 bp) was generated in which one of the amplification primers was 5Ј-end-labeled with [␥-32 P]ATP and T4 polynucleotide kinase. The same DNA (endlabeled in the sense strand) was used to generate a Maxam-Gilbert chemical-degradation sequencing ladder (Fig. 3, first and fourth lanes and fifth and eighth lanes). DNase I footprinting on both sense and antisense strands showed strong protection of the A/T-rich region (Fig. 3, third and seventh lanes). The footprint spans a large region (GGGGGTATTAATATAA) that includes the A/T-rich element and several flanking nucleotides, suggesting that more than one protein is binding to this region. Footprinting analyses also showed a weak protection of the E-box located at Ϫ1516 bp in the antisense strand, but absent in the sense strand (Fig. 3).

Mutations of the A/T-rich Element Produced a Significant Decrease in SMHC Promoter Activity in Vascular Smooth
Muscle Cells-To determine whether the A/T-rich element (located at Ϫ1540 bp) is critical for SMHC promoter activity, the A/Trich element was modified in the p1548-CAT plasmid using site-directed mutagenesis. A point mutation was introduced into the A/T-rich element (GTATTAAT ATA 3 GTAT-TCATATA) where a single nucleotide was modified from adenine to cytidine. The introduction of C in the place of A abolished MEF2-binding activity in the MCK promoter (31). The CAT reporter constructs containing the wild-type SMHC promoter (p1548-CAT) and the mutated A/T-rich element (p1548-CATA/T) were transfected into primary cultures of rat aortic smooth muscle cells, and CAT reporter activity was determined. The wild-type SMHC promoter produced a 20-fold increase in CAT activity over the promoterless CAT vector (pJR-CATX) (Fig. 4). Conversely, mutation of the A/T-rich element produced a 6.5-fold reduction (68% decrease) in reporter activity (Fig. 4). These results indicate that the A/T-rich element is critical for maximal SMHC promoter activity.

MEF2B-specific Antibody Supershifts the Protein Complexes That Bind to the A/T-rich Element-
To determine precisely what protein factors bind to the A/T-rich element, we performed gel shift analyses using nuclear extracts from smooth, skeletal, and nonmuscle cells. The 133-bp DNA fragment (Ϫ1594 to Ϫ1462 bp) that contained the A/T-rich element (5Ј-TATTAATATA-3Ј) was used as a probe. As shown in Fig. 5A, the gel shift analyses revealed three closely migrating protein complexes with the nuclear extracts from vascular smooth muscle cells (lanes 1 and 2). However, the same probe produced a different protein binding pattern with nuclear extracts from Sol-8 myotubes, C 2 C 12 myotubes, and NIH3T3 fibroblasts, where only a single protein complex was observed in C 2 C 12 and NIH3T3 fibroblast extracts (Fig. 5B). Interestingly, a 100-fold molar excess of an oligonucleotide containing the A/T-rich element (5Ј-CTGGGGGTATTAATATAAGACACCC-3Ј) abolished all three protein complexes (Fig. 5A, lane 3). However, an A/T-rich oligomer containing three nucleotide substitutions, A/T rich (5Ј-CTGGGGGCATTCCTATAAGACACCC-3Ј), did not compete protein binding at a 100-fold excess (lane 4). In addition, we performed competition assays using an oligonucleotide containing a point mutation in the A/T-rich core sequence (which was shown to decrease SMHC promoter activity in the transient transfection analyses) (Fig. 4). This oligonucleotide also failed to compete protein binding at 10-, 20-, and 100-fold molar excesses (data not shown), suggesting that any modification to the A/T-rich core region abolishes protein binding. An E-box oligonucleotide corresponding to the E-box sequence located at Ϫ1516 bp did not act as a competitor (Fig. 5B,  lane 2). To our surprise, a MEF2 consensus oligonucleotide derived from the MCK promoter (5Ј-GATCCTCGCTCTA-AAAATAACCCTGTC-3Ј) did not compete protein binding (lanes 5 and 6).
To determine whether any members of the MEF2 family bind to this A/T-rich region in nuclear extracts from vascular smooth muscle cells, antibodies that react specifically to the MEF2A, MEF2B, MEF2C, and MEF2D isoforms were incubated together with the DNA and proteins in the gel shift reaction. As shown in Fig. 5A, MEF2B antibody alone supershifted the three protein complexes (lane 9 and 10), whereas the antibodies to MEF2A, MEF2C, and MEF2D did not interfere with protein binding (lanes 7, 8, and 11-14). MEF2B-specific antibody from two different sources (obtained from Drs. Yu and Prywes) gave the same result, supporting that MEF2B is present in this complex. The entire complex is shifted by MEF2B antibody, suggesting that other MEF2 proteins (MEF2A, MEF2C, and MEF2D) are not part of the protein complexes that bind to the A/T-rich element. These results unambiguously demonstrate that the MEF2B protein isoform is contained in the A/T-rich binding complex in smooth muscle cells, but not in skeletal muscle cells. The pretreatment with MEF2B antibody also produced the same supershift pattern (Fig. 5B, lane 3), suggesting that MEF2B antibody does not interfere with protein binding to the A/T-rich element. In addition, an oligonucleotide corresponding to the E-box element located at Ϫ1581 bp did not compete protein binding (Fig. 5B, lane 2).
The from Cultured Smooth Muscle Cells-To determine whether MEF2 transcription factors are present in smooth muscle nuclear extracts, we performed Western blot analyses using MEF2 isoform-specific antibodies. As shown in Fig. 6, all four MEF2 isoforms (A, B, C, and D) were present in the nuclear extract from early passage rat aortic smooth muscle cells. Therefore, it appears that primary rat aortic smooth muscle cells express all of the MEF2 isoforms.

The MEF2 Family Members (MEF2A, MEF2B, MEF2C, and MEF2D) Can Bind to the A/T-rich Element in the SMHC Gene
Promoter-To determine whether the A/T-rich element in the SMHC promoter can bind directly to each member of the MEF2 family, we performed gel mobility shift assays using in vitro translated proteins for MEF2A, MEF2B, MEF2C, and MEF2D with a 25-bp oligomer containing the A/T-rich element. In parallel, gel mobility shift assays were performed with the MCK-MEF2 and ␣MHC-MEF2 site oligomers as high and low affinity controls, respectively. As shown in Fig. 7, each of the MEF2 family members can bind to the A/T-rich oligomers from the SMHC promoter. The binding of the MEF2B in vitro translated product to each of the probes was somewhat more diffuse when compared with the other MEF2 translation products, consistent with its previous description (27). However, it is important to note that the ratio of each of the different MEF2 translation products is similar between the three different probes. This suggests that the minimal A/T-rich site itself does not direct the preferential binding of MEF2B over the other MEF2 proteins.
Cotransfection of the MEF2-B Expression Vector Together with the SMHC Promoter Increases Reporter Activity in Vascular Smooth Muscle Cells, but Not in NIH3T3 Fibroblasts-To determine whether MEF2B protein can positively regulate the SMHC promoter, we cotransfected the SMHC promoter constructs (p1548-CAT, p1392-CAT, and pJRCATX) together with the MEF2B expression vector pCGNR2 (23) into cultured rat aortic smooth muscle cells and NIH3T3 fibroblasts. In response to MEF2B overexpression, the construct p1548-CAT, which includes the MEF2B-binding site, showed a 61% increase in promoter activity, whereas the promoter construct p1392-CAT, which lacks the A/T-rich element, failed to show an increase in promoter activity in cultured rat aortic smooth muscle cells (Fig. 8). On the other hand, MEF2B cotransfected with p1548-CAT and p1392-CAT into NIH3T3 fibroblasts did not significantly increase SMHC promoter activity, which is very low in fibroblasts. Cotransfection of the MEF2C expression vector (pCDNA1-MEF2C) failed to increase SMHC promoter activity significantly both in smooth muscle cells and in NIH3T3 cells (data not shown). These results demonstrate that MEF2B can up-regulate SMHC promoter activity in smooth muscle cells, but not in NIH3T3 fibroblast cells. These data further suggest that MEF2B function is dependent upon other tissue-specific cofactors.

DISCUSSION
The goal of this study was to characterize cis-regulatory elements important for SMHC gene expression. In this study, we demonstrated that the SMHC promoter region (Ϫ1594 to Ϫ1462 bp), which contains the A/T-rich element, is capable of activating a heterologous (TK) promoter in smooth muscle cells, but not in fibroblasts. Point mutation of this A/T-rich element reduces SMHC promoter activity significantly in vascular smooth muscle cells. Gel mobility shift analyses demonstrated that the probe (133 bp) containing the A/T-rich element binds to three closely migrating protein complexes from smooth muscle cell nuclear extracts, whereas nuclear extracts from skeletal muscle showed a different protein binding pattern. We further demonstrated that the protein complexes were supershifted by an antibody specific to MEF2B protein, indicating that MEF2B is a component of a smooth muscle-specific complex that binds the A/T-rich element in the SMHC promoter. By Western blot analyses, we demonstrated that all four MEF2 isoforms (A, B, C, and D) were present in nuclear extracts from rat aortic smooth muscle cells. Therefore, the preferential binding of MEF2B to the A/T-rich element is not simply a reflection of its abundance relative to the other MEF2 factors. Furthermore, our data suggest that MEF2B overexpression can activate the SMHC promoter in smooth muscle cells, but not in NIH3T3 fibroblast cells. These findings taken together suggest that MEF2B protein may play an important role in regulating SMHC gene expression in smooth muscle cells.
In this study, we have shown that the antibody specific to MEF2B protein can recognize the protein complexes formed on the A/T-rich element. This is a novel finding and suggests that MEF2 transcription factors might be important in the regulation of smooth muscle-specific gene expression. MEF2B mRNA is expressed in brain, heart, and skeletal muscle (7) and in smooth muscle (20). However, the precise role of MEF2B protein in transcriptional regulation is not well understood. Initial studies showed that MEF2B protein did not bind DNA efficiently (6,7). Nevertheless, forced expression of MEF2B resulted in expression of a reporter gene containing multimers of the MEF2 recognition sequence. These studies suggested that MEF2B may be heterodimerizing with other MEF2 isoforms to FIG. 4. Mutation of the A/T-rich element decreases SMHC promoter activity in rat aortic smooth muscle cells. 15 g of the SMHC-CAT reporter plasmids (schematically represented on the left) and 5 g of the pMSV-␤gal reference plasmid were transfected into cultures of primary vascular smooth muscle cells, and CAT and ␤-galactosidase activities were determined. CAT activities, corrected for differences in transfection efficiencies, were normalized to the CAT activity observed following transfection of the promoterless control plasmid, pJRCATX. A representative CAT assay is shown in the middle, and relative CAT activities Ϯ S.E. are shown on the right. In this study, we observed that the probe containing the A/T-rich region revealed three closely migrating protein complexes only in smooth muscle nuclear extracts. Interestingly, all three protein complexes were supershifted by MEF2B anti-body. This raises the possibility that MEF2B protein may be complexing with other proteins to produce these complexes. Alternatively, we should also consider the possibility that these multiple protein complexes are due to phosphorylated forms of MEF2B protein (27). Although MEF2B protein is known to be expressed in Sol-8 and C 2 C 12 skeletal muscle cells, the nuclear extracts from these cells did not produce the same protein binding pattern. This finding also supports the idea that MEF2B binding may depend upon cell type-specific interacting proteins. DNase I footprinting with smooth muscle nuclear extracts revealed extended protection including the A/T-rich element and its flanking nucleotides, further supporting multiple protein binding in the A/T-rich region.
A double-stranded oligonucleotide containing the A/T-rich element effectively competed all three protein complexes, whereas the oligonucleotides containing mutations of the A/Trich core sequence (TATTAAT) failed to act as competitors. These results suggest that direct protein interaction with the A/T-rich element is critical for the formation of the three protein complexes. In this study, we have shown that in vitro translated MEF2 family members (MEF2A, MEF2B, MEF2C, and MEF2D) can directly bind to the A/T-rich element. However, the binding affinity for the SMHC-A/T-rich element is much lower compared with the MCK-MEF2 consensus element, but comparable to the ␣MHC-MEF2 site. Although other MEF2 family members (MEF2A, MEF2C, and MEF2D) are also expressed in smooth muscle cells, they do not bind to this A/T-rich element in smooth muscle nuclear extracts, which suggests that MEF2B binding specificity may depend upon the interacting proteins present in smooth muscle cells.
MADS box proteins have been shown to be highly interactive molecules in both yeast and higher organisms (reviewed in Ref. 32). Recently, it has been shown that the MADS box protein MEF2C physically interacts with the myogenic bHLH proteins as a heterodimer with E-box proteins (17). The MADS box protein SRF has also been shown to physically interact with this same heterodimer formed between the myogenic bHLH proteins and E-box proteins (33). Furthermore, MEF2 proteins have also been shown to interact with the bHLH protein Twist (34) as well as with the bHLH protein MASH1 (35). Together, these results suggest that MADS box proteins are interactive molecules that complex with other cell type-specific transcrip-FIG. 6. Detection of MEF2 proteins in nuclear extracts from primary cultured smooth muscle cells. Nuclear extracts were prepared from primary cultured rat aortic smooth muscle cells as described under "Materials and Methods," and proteins were resolved by SDS-10% polyacrylamide gel electrophoresis and transferred to nitrocellulose. Blots were then probed with antibodies specific to MEF2 isoforms A, B, C, and D. Molecular mass markers (in kilodaltons) are shown on the sides, and the position of each MEF2 protein is indicated by brackets. tion factors to direct tissue-restricted gene expression.
To date, transcription factors that are unique to smooth muscle cells have not been identified, and smooth musclespecific gene expression remains underexplored. However, the recent isolation and characterization of the smooth musclespecific gene promoters for smooth muscle ␣-actin (36 -38), smooth muscle ␥-actin (39), SM22␣ (40,41), calponin (42,43), and SMHC provide unique opportunities for dissecting the various cis-and trans-regulatory factors involved in smooth muscle cell-specific gene expression. Studies using the abovementioned promoters have identified a number of known ciselements including the CArG box, E-box, GATA-binding site, AP-2, SP1, and A/T-rich element. In particular, CArG elements that bind SRF have been shown to be important for smooth muscle ␣and ␥-actin gene expression (36,37,44). Similarly, CArG box elements are found in the SM22␣ proximal promoter and may be important for its regulation in myogenic lineages (41). On the other hand, there are no CArG elements in the human calponin gene, which is expressed exclusively in smooth muscle-containing tissues in adult stages.
The SMHC promoter region (Ϫ1281 to Ϫ1038 bp) contains three CArG box-like elements (Fig. 1) and is conserved between mouse, rat, and rabbit promoters. This region is found to be necessary for maximal SMHC promoter activity (45,46). 2 The CArG box element is known to bind to SRF or related factors; however, the precise role of SRF transcription factors in smooth muscle-specific gene expression remains to be explored. Future experiments will be focused toward identifying cis/trans-factors responsible for smooth muscle cell-specific gene expression.
In summary, the data presented here suggest that MEF2B protein may play an important role in regulating SMHC gene expression in smooth muscle cells. We propose that MEF2B protein may interact with other transcription factors in controlling the smooth muscle-specific expression of this gene. Future experiments will attempt to identify and characterize the protein factors that interact with MEF2B protein.