Identification of Functional Promoter Elements in the Rabbit Smooth Muscle Myosin Heavy Chain Gene*

Despite the importance of smooth muscle cell proliferation in vascular pathophysiological states, the mechanisms regulating smooth muscle cell growth and differentiation are poorly understood. Previous studies have shown that adult rabbit smooth muscles express two types of myosin heavy chain (MHC) isoforms, SM1 and SM2, which are generated through alternative RNA splicing from a single smooth muscle MHC (SMHC) gene. In the present study, we isolated and characterized the rabbit SlVWC gene promoter. DNA sequence analysis of the upstream region of the SMHC gene revealed several putative cis-DNA regulatory elements proximal to the transcription start site. Most notably, cis-acting regulatory elements that closely resemble CC(A/T),GG (CArG box) and myocyte enhancer binding factor 2 (MEF-2)- type sequence motifs were found in the SMHC 5"flank-ing region. In addition, six E-box motifs were found in the 5"flanking region of the SMHC gene between -374 and -2109 base pairs from the transcription start site. A series of transient transfection assays using SMHC promoter deletion constructs indicated 'that a promoter fragment extending to 2266 base pairs upstream of the transcription start site has the highest reporter activity in cultured rat aortic smooth muscle cells. Gel mobility shift analyses using the MEF-2-like sequence located at -1540 revealed a specific DNAprotein complex, whereas the CArG-like element located at -1275 did not show protein binding. The SMHC promoter construct, p509-CAT, which included neither the CArG-

Despite the importance of smooth muscle cell proliferation in vascular pathophysiological states, the mechanisms regulating smooth muscle cell growth and differentiation are poorly understood. Previous studies have shown that adult rabbit smooth muscles express two types of myosin heavy chain (MHC) isoforms, SM1 and SM2, which are generated through alternative RNA splicing from a single smooth muscle MHC (SMHC) gene. In the present study, we isolated and characterized the rabbit SlVWC gene promoter. DNA sequence analysis of the upstream region of the SMHC gene revealed several putative cis-DNA regulatory elements proximal to the transcription start site. Most notably, cis-acting regulatory elements that closely resemble CC(A/T),GG (CArG box) and myocyte enhancer binding factor 2 (MEF-2)type sequence motifs were found in the SMHC 5"flanking region. In addition, six E-box motifs were found in the 5"flanking region of the SMHC gene between -374 and -2109 base pairs from the transcription start site. A series of transient transfection assays using SMHC promoter deletion constructs indicated 'that a promoter fragment extending to 2266 base pairs upstream of the transcription start site has the highest reporter activity in cultured rat aortic smooth muscle cells. Gel mobility shift analyses using the MEF-2-like sequence located at -1540 revealed a specific DNAprotein complex, whereas the CArG-like element located at -1275 did not show protein binding. The SMHC promoter construct, p509-CAT, which included neither the CArGnor MEF-2-type motifs, conferred 32% of chloramphenicol acetyltransferase activity in the same cells, whereas the construct pl88-CAT, which contained the minimal promoter elements (TATA box), was significantly less active (7%; 2.0fold over background). This is the first report describing the promoter elements of a gene whose expression is restricted to smooth muscle cells.
Myosin, the major protein constituent of the contractile apparatus in all muscle tissues, is composed of two heavy chains and four light chains. The myosin heavy chain (MHCY mole-cule contains an ATPase that acts as the chemo-mechanical energy transducer during muscle contraction. The MHC isoforms are encoded by a large multigene family, the expression of which is regulated in a tissue-specific and developmental stage-specific manner (Mahdavi et al., 1986;Gulick et al., 1987;Kropp et al., 1987). In smooth muscle cells, there are at least four types of MHC, two of which are smooth muscle specific (SM1,204 kDa,and SM2,200 kDa) and two of which are expressed in both smooth muscle and nonmuscle tissues 196 kDa,200 kDa). We have shown that the SM1 and SM2 myosin isoforms are products of the same MHC gene generated by alternative RNA splicing Periasamy, 1989, Nagai et al., 1989). During early stages of smooth muscle development, SM1 myosin is predominantly expressed, and SM2 myosin appears only during the postnatal period . Recent studies have shown that the nonmuscle myosin NMHC-B/SMemb is also expressed in smooth muscle cells during early stages of development (Kuro-o et al., 1991). In the adult stage, however, SM1 and SM2 are the only two myosin isoforms detectable in smooth muscle tissues of rat and rabbit (Kuro-o et al., 1989;Borrione et al., 1989). Interestingly, SM1 but not SM2 myosin is expressed in proliferating smooth muscle cells of arteriosclerotic neointimas (Kuro-o et al., 1991). In addition, we have shown that primary cultured smooth muscle cells express only the SM1type MHC, and down-regulate the SM2 isoform (Babij et al., 1992). Although the physiological significance of these isoforms is not understood at this time, it is likely that differential expression of the SM1 and SM2 isoforms may have important roles in smooth muscle development and function.
Abnormal neointimal smooth muscle cell proliferation appears to be a common cause of vascular disease such as atherosclerosis or restenosis after coronary angioplasty (Ross, 1986). During this process, smooth muscle cells become more secretory and less contractile and reach a dedifferentiated state. The expression level of myosins is altered, involving a complete loss of SM2 myosin. A number of protein markers have been identified in smooth muscle cells, including a-smooth muscle actin, SMHCs, heavy caldesmon, calponin, and others (Skalli et al., 1986;Nagai et al., 1989;Glukhova et al., 1988;Frid et al., 1992). Several of these markers have been found to be expressed in other tissues, however. For example, a-smooth muscle actin is found in myofibroblasts, skeletal, and cardiac muscle; in cells of the eye lens and hair follicles; and in bone marrow stroma (Sappino et al., Schmitt-Graff et al., 1990;Jahoda et al., 1991;Peled et al., 1991). On the other hand, the SMHC gene is expressed highly specific in smooth muscle tissues, and its expression pattern correlates well with smooth muscle growth and differentiation. Therefore, the SMHC gene is a good candidate for studying and characterizing factors that influence growth and differentiation specific to smooth muscle cells.
The purpose of this study was to isolate and characterize the rabbit SMHC gene promoter, including its 5'-upstream regions required for transcriptional activation in smooth muscle cells. In this study, a genomic clone containing the SMHC gene promoter and 8 kb of 5"flanking region was isolated. DNA sequence analyses revealed several putative cis-regulatory elements including MEF-2-like and CArG box elements found in other muscle-specific genes. Gel mobility shift analysis using the MEF-2-like sequence located at -1540 revealed a specific DNA protein compex, whereas the CArG-like element located at -1275 did not show protein binding in our study. Studies using a series of SMHC promoter deletion constructs transiently transfected into primary smooth muscle cells demonstrate that a 2266-bp promoter fragment is capable of high level expression in smooth muscle cells. On the other hand, the same SMHC promoter constructs were unable to drive high levels of chloramphenicol acetyltransferase (CAT) expression in other cell types such as NIH3T3 and C,C,, cells. These studies demonstrate that the SMHC gene contains promoter elements that are highly tissue specific.

EXPERIMENTAL PROCEDURES
Screening of the Rabbit Genomic Library-A rabbit genomic library constructed in the cloning vector EMBL-3 SP6pT7 (Clontec Laboratories Inc.) containing 1.4 x IO6 independent clones and with an insert size ranging from 8 to 22 kb was used. Screening of the genomic library was carried out as described previously under conditions of high stringency (Zarain-Herzberg et al., 1990) using a -1.3-kb fragment from the SMHC genomic clone ARG4 as a probe (Babij et al., 1991). The clones were digested with restriction endonucleases, and selected genomic DNA fragments were subcloned into p-Bluescript M13+ and pJRCATX vectors (Robbins et al., 1989). For Southern blot analysis, restriction endonuclease digests of the genomic clones were separated by electrophoresis in 0.8% agarose gel and blotted onto nitrocellulose (Maniatis et al., 1989). The filters were hybridized with DNAfragments labeled with 32P by random primer extension. The SMHC genomic DNA probes that were used to determine the gene map (see Fig. 1) were as follows: a -2-kb EcoRI -Hind111 fragment containing the TATA box and the first exon and a -1.3-kb SalI-EcoRI fragment from the 5'-upstream region of the SMHC genomic DNA (Babij et al., 1991). DNA sequencing was performed by the procedure of Sanger et al. (1977) using the Sequenase kit (U. S. Biochemical Corp.).
Construction of the SMHC Gene Promoter-CAT Chimeric Constructs-A -3.3-kb DNA fragment containing 2266 bp of the 5'flanking region of the SMHC gene was excised from the clone hRG4 using SalI-Hind111 endonucleases. The -3.3-kb fragment was bluntended with Klenow enzyme and ligated with HindIII linkers using standard methods (Maniatis et al., 1989). The resulting fragment was ligated in a 5' to 3' orientation into the unique HindIII site of the CAT expression vector pJRCATX (Robbins, 1989). The resulting vector, p2266-CAT (including -2266 bp of 5"flanking region), was used as a template for 5' to 3' exonuclease 111 digestion (Erase-a-Base kit, Promega). This produced a series of unidirectionally deleted promoter constructs (see Fig. 31, which were used in transfection studies. The 5'-ends of the deleted constructs were confirmed by dideoxynucleotide sequencing (Sanger et al., 1977). In addition, the longest construct, p4.2-CAT, was created using a KpnI-Hind111 fragment from the genomic clone ARG8.
Cell Culture-Smooth muscle cells from rat thoracic aorta were isolated and cultured by a modification of the procedures described by Owens et al. (1986). 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 from 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,, 95% air atmosphere in Hanks'balanced salt solution in the presence of 1 mg/ml collagenase (219 unitdmg, 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-mm2 pieces, and incubated (37 "C, 5% CO,, 95% air) for an additional 80-110 min. The dissociated cells were separated from the undigested tissue by filtration through an 85 micrometer 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,, 95% air with medium changes 3 times weekly. Confluent cultures were usually obtained after 4-8 days. The confluent cultures were washed with 1 x phosphatebuffered saline, harvested with 0.05% trypsin, 0.53 m M EDTA, and replated at a 1 5 or 1:6 split area ratio into 90-mm tissue culture plates. Since it has been reported that SM1 and SM2 myosin disappeared in cultured rat aortic smooth muscle cells after several passages and was replaced by a 196-kDa nonmuscle MHC (NMCH-A) (Kawamoto et al., 1987), we chose to use rat aortic smooth muscle cells after the first passage for our in vitro transfection studies and nuclear extraction.
DNA Dansfections-Transient transfections of the SMHC promoter constructs were performed using the calcium phosphate co-precipitation method (Brent et al., 1987). Briefly, duplicate dishes of cells (5 x cellddish) were transfected with 15 pg of each of the SMHC-CAT constructs plus 5 pg of MSVpGAL. DNA used for transfections was purified by two successive CsCI, density gradient centrifugations. After 8-14 h, cells were washed with PBS buffer (137 m M NaCl, 27 m M KC], 8 m M Na,HPO,, 1.5 m~ KH,PO,) and glycerol shocked for 1-2 min. Cells were then washed twice with PBS, and fresh growth medium (10% fetal calf serum) was added. For cultured rat aortic smooth muscle cells and NIH3T3 fibroblasts, cells were harvested 48 h posttransfection. For C,C,, myotubes, myoblasts were allowed to propagate in growth medium for 8-10 h and then were replaced with differentiation medium to induce myotube formation. C,C,, myotubes were harvested 36-48 h following medium change.
CAT Assays-The transfected cells were rinsed twice with cold Hanks' solution and harvested in release buffer (40 m M Tris-HC1, pH 7.4, 1 m M EDTA, 150 m M NaCl). The cells were lysed in 250 m~ Tris-HCI, pH 7.5, by three cycles of freeze-thawing, and the cell lysate was used for the enzyme assays. p-Galactosidase assays (Herbomel et al., 1984) 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 p1 of the extract in a 150-pl assay containing 4 m~ acetyl coenzyme A (Sigma) and 0.05 mCi of ['4Clchloramphenicol(54 mCi/mmol, Amersham Corp.) in 250 m M Tris-HC1, pH 7.8. The reaction was stopped and extracted with 1 ml of ethyl acetate. The organic phase was dried, suspended in 20 pl of ethyl acetate, and spotted onto a silica gel thin layer plate. The chromatogram was developed in a chlorofodmethanol (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. All constructs were tested in at least three separate transfection experiments with at least two different plasmid preparations. The construct pSV,CAT was used as a positive control (Gorman et al., 1982). Experiments in which the transfection efficiencies varied more than 10% between the cultures were discarded.
Preparation of Nuclear Extracts and Gel Mobility Shif? Assays-Nuclear extracts from rat aortic cultured smooth muscle cells were prepared essentially as described by Dignam et al. (1983). The protein concentration was determined by the Bradford assay (Bradford, 1976). Gel mobility shift assays were performed as described previously (Sukovich et al., 1993). Oligonucleotide was end-labeled with [y32P1ATP, using T4 polynucleotide kinase. Synthetic oligonucleotide top strands were annealed with a 3-fold molar excess of the corresponding bottom strand unlabeled oligonucleotide. Briefly, 2-10 pg of nuclear extracts was incubated in a total volume of 30 p1 for 30 min at room temperature in the presence of 1 ng of radiolabeled double-stranded oligonucleotides and 2 pg of poly(dI-dC) and analyzed by electrophoresis in 6% polyacrylamide gels. Competition experiments were performed with 100-500fold molar excess of specific or nonspecific unlabeled DNA fragments. The oligonucleotides used as probes and competitors are described (see Fig. 6B). In "super shift" assays, 1.5-3.0 pl of the MEF-2 (C-21) Trans-CruzTM Gel Supershift antibody (Santa Cruz Biotechnology, Inc.) was added per 30 pl of reaction volume subsequent to addition of 32P-labeled oligonucleotide probe, and it incubated at 4 "C for 1 h. Following electrophoresis, the gels were stained in 10% acetic acid, 30% methanol; dried; and autoradiographed.

RESULTS
Isolation of the Rabbit SMHC Gene Promoter and Its 5 ' -Upstream Region-To initiate studies on the control elements regulating SMHC gene expression, the rabbit SMHC gene promoter and its 5"flanking region were isolated. Arabbit genomic library constructed in EMBL-3 SP6/"7 was screened with a -1.3-kb SalI-Hind111 fragment from the 5'-end of clone ARG4 (Babij et al., 1991) under conditions of high stringency. This fragment was previously identified to contain the TATA box and transcription initiation site. Screening of 2 x lo6 independent plaques yielded five positive recombinant phages. These genomic clones were subjected to further analysis by restriction enzyme mapping and Southern blotting. A single clone of interest (hRG8), estimated to contain a n 8-kb upstream region of the SMHC gene promoter, was chosen for detailed analysis (Fig. 1).

Sequence Analysis of the SMHC Gene Promoter and Its 5'-
Flanking Region-To map precisely the promoter and regulatory sequences in the genomic clone, ARG8 restriction fragments were subcloned into the M13mp19 vector, and DNA sequence analysis was performed. The nucleotide sequence of the SMHC promoter region extending to -2266 bp upstream of the transcription initiation site is shown in Fig. 2. Detailed analysis of the nucleotide sequence upstream of the transcription initiation site revealed a canonical TATA box (5"TATAAA-3') sequence at position -26 bp (Breathnach and Chambon, 1981). The CAAT box motif usually found around -80 bp was not found. The consensus sequence for SP1 binding site 5'-GGGCGG-3' (present in constitutively expressed genes (Dynan and Tjian, 1985)) is represented 3 times at positions -2225 to -2220, -1956 to -1951, and -1502 to -1497; and the complementary sequence 5'-CCGCCC-3' is found 3 times at positions -1177 to -1172, -579 to -574, and -523 to -518. Several DNA sequences previously reported to bind specific trans-acting factors are also present in this gene. Two AP2 binding sites are found at positions -48 t o -41 (5'-CCGCGGGC-3') and -1248 to -1241 (5"CCCCGGGC-3') (Imagawa et al., 1987); and two CF1 recognition sites are found at positions -847 to -842 and -1852 to -1847 (5"ANATGG-3') (Khoury-Christianson et al., 1992) (Fig. 2). Two CArG box-like motifs, previously identified in skeletal and cardiac muscle actin genes (Minty and Kedes, 1986), are present at -1275 to -1266 (5'-CCATATTTAG-3') and -1194 to -1185 (5'-CCTTTTTGGG-3') upstream of the transcription initiation site. Interestingly, a myocyte enhancer binding factor consensus (MEF-2) (Gossett et al., 1989) like sequence was found at -1540 to -1530 (5'-TATTAATATAA-3'), and a DNA sequence resembling the MEF-2 site is located at cloned gene carried a functional promoter and the appropriate regulatory sequences, we subcloned two restriction endonuclease fragments containing upstream regions -2266 bp and -1223 bp into the CAT reporter vector, pJRCATX (Fig. 3). Transient transfection analyses of these two promoter constructs into smooth muscle cells showed that the construct p2266 produced high levels of CAT activity, whereas the p1223 construct produced low levels of CAT activity (Fig. 4). To delineate further the boundaries of the functional promoter and regulatory elements within the 5"flanking region of the SMHC gene, we prepared a series of deletion constructs using the p2266-CAT construct and exonuclease I11 (Fig. 3). The resulting constructs were verified by DNA sequencing analysis before transfection. Each construct contained different lengths of the 5"flanking region, the first exon (78 bp, 5'-untranslated), and 787 bp of the first intron linked to the CAT reporter gene. The following deletion constructs were chosen for further analysis: p1947-CAT, and p188-CAT (Fig. 3). In addition, we created a larger promoter construct p4.2-CAT, which contained 4.2 kb of the upstream region (Fig. 3).
To define promoter elements, the above described constructs were transfected into primary cultures of rat aortic smooth muscle cells. We have previously shown that these cultures are myosin positive and express high levels of SMHC transcripts (Babij et al., 1992). As shown in Fig. 4 the highest CAT activity was observed with the p2266 SMHCEAT construct, and therefore this was treated as 100% for comparison with other constructs. The larger promoter construct, p4.2-CAT, produced only 14% of the maximal CAT activity.
The construct pl88-CAT, which contained the minimal promoter elements including the TATA box, produced low levels of CAT activity (7% of maximal activity), whereas the promoter construct p509-CAT produced 32% of maximal CAT activity showing a 5-fold increase over p188-CAT. Therefore, the region between -188 and -509 may contain important positive regulatory elements promoting transcription of the SMHC gene.
Inclusion of additional upstream sequences from -509 up to -1223 produced a significant decrease in reporter activity. The CAT constructs p753-CAT, p1042-CAT, and p1223-CAT, gave only 12, 11, and 10% of maximal activity, respectively, suggesting that the region between -509 and -1223 acts as a negative regulator on this promoter. Interestingly, the addition of further upstream sequences located between -1223 and -2266 produced significant increases in CAT activity, as seen with promoter constructs p1392-CAT (30%), p1548-CAT (62%), and p1947-CAT (65%). The CAT activity reached the maximum with p2266-CAT (100%). Therefore, a second positively acting upstream region is located between -1392 and -2266. In contrast, the largest construct (p4.2-CAT) produced only 15% of ' p4.2 shown in Fig. 5, these constructs did not yield significant CAT activity in NIH3T3 or C,C,, cells (maximal CAT activity was %fold over background; whereas the CAT activity produced by pSV,CAT was more than 60-fold over background, data not shown) as compared with smooth muscle cells. These results are in agreement with our in uiuo observations, where SMHC gene expression is limited to smooth muscle cells .
Gel Mobility Shift Assay with a MEF-2-like (A/T-rich) Element-Our transient transfection analysis revealed that deletion of promoter sequences between -1548 and -1392 and between -1392 and -1223 decreased CAT activity from 62 to 32%, and from 32 to lo%, respectively. To define the DNA regulatory elements present in these areas, we performed gel mobility shift assays on these two putative regulatory elements. A MEF-2-like sequence (5"TATTAATATATAA-3') located at -1540 to -1530 and a CArG-like sequence (5'-CCATA'M-TAG-3') located at -1275 to -1266 oligomers were used as probes. As shown in Fig. 6 A , the gel shift analyses using the MEF-2-like sequence revealed that this element binds to a specific protein complex in the nuclear extracts from vascular smooth muscle cells (lanes 13). A 100-fold molar excess of cold homologous oligonucleotide abolished the protein binding (lane 4 ) , whereas a 500-fold molar excess of a mutated form of MEF-2-like element did not eliminate protein binding (lane 5). Interestingly, the muscle creatine kinase MEF-2 consensus element also failed to compete with protein binding to this element. Furthermore, the addition of the MEF-2 (C-21) antibody did not interfere with this protein binding, suggesting that a protein other than MEF-2 may bind to this APT-rich sequence. On the other hand, the CArG-like element found at -1275 t o -1266 did not reveal any specific protein binding, whereas many bands were revealed by SP-1 oligomer as a positive control (data not shown).

NIH3T3 c2c12
box usually located around -80 bp is not found. The proximal 5"flanking region of the SMHC gene is highly GC-rich. The SP-1 binding site (5"GGGCGG-3') described in promoters of several genes (Dynan and Tijan, 1985) is present 4 times. The rabbit SMHC gene also contains two CArG box-like motifs. The CArG motif at -1194 position differs from the CC(A/T),GG configuration since only 5 of the 6 core APT residues are present. The second CArG motif that is found at position -1275 also differs from the consensus in that it ends in AG nucleotides instead of GG. The CArG motif, which was first described in the 5"regulatory region of striated muscle a-actin genes, is also found in several other muscle-specific genes (Zakut et aZ., 1982;Miwa and Kedes, 1987;Sternberg et al., 1988). The importance of CArG elements in the regulation of the smooth muscle a-actin promoter has been established (Carroll et al., 1986). Another important regulatory element, namely the MEF-2-like site, is situated at positions -1540 and -789 bp in the SMHC upstream region. MEF-2 is a trans-acting factor and is expressed abundantly in cardiac, skeletal, and smooth muscle tissues (Yu et al., 1992;Gossett et al., 1989). MEF-2 recently has been shown to be important for the transcriptional regulation of a number of skeletal and cardiac muscle-specific genes and to act in concert with myogenic regulatory proteins such as myogenin in myogenesis (Edmondson et al., 1992). In addition, six E-box motifs are found in the 5'flanking region between -374 and -2109. E-box motifs have been shown to bind muscle-specific and basic helix-loop-helix trans-activating factors (Olson, 1990) and are known to be important for muscle-specific gene expression (Blackwell and Weintraub, 1990). The exact role of these sequences in the SMHC gene regulation is yet to be defined.
Our transient transfection analyses have identified multiple positive and negative regulatory elements in the 5"flanking region of the SMHC gene. Most notably, the promoter element located between -1392 and -2266 produced the highest reporter CAT activity upon transfection into smooth muscle cells. Another region of importance is located between -188 and -509, since inclusion of this region produced a large increase in CAT activity over that of the minimal promoter. These two FIG. 6. Gel mobility shift assay of MEF-2-like motif located at -1640 to -1560 with nuclear extracts from rat aortic smooth muscle cells. A, lunes I 3 included 2, 5, and 10 pg of nuclear extract from smooth muscle cells, respectively; lune 4 included a 100-fold molar excess of cold homologous oligonucleotides; lune 5 includes a 500-fold molar of mutant oligos for the MEF-2-like element; lune 6 includes a 500-fold molar excess of cold muscle creatine kinase (MCK) MEF-2 oligos; lunes 7 and 8 include a MEF-2 antibody (C-21, Santa Cruz), 1.5 and 3 p1, respectively. Similar results were obtained from three additional experiments. B, sequences of oligonucleotides used in the gel mobility shift assay. Mutated bases in the MEF-2-like sequences are underlined.
regions act as positive regulatory elements. Interestingly, progressive deletions in the upstream region from -2266 bp to -1392 resulted in a gradual loss of CAT activity from 100 to 30%. This finding suggests that there are multiple positive regulatory elements in this region, and that maximal expression depends upon cooperative interactions between these elements. Actually, four out of six E-boxes are found within the upstream region from -2266 to -1392, which exhibits the highest promoter activity.
Our deletion analysis also revealed negative elements in the SMHC promoter. In particular, when the DNA sequence between -509 and -753 bp is included in the promoter construct, there is a significant decrease (-2.7 fold) in activity in comparison with the promoter construct that includes only 509 bp of upstream DNA.
Interestingly, the deletion of a MEF-2-like site present at -1540 bp results in a significant loss of CAT activity (from 62 to 32%). However, deletion of the second MEF-2-like site present a t -789 does not modify CAT activity. The deletion of the CArGlike motif, present at -1275 results in a loss of CAT activity (from 32 to lo%), whereas the deletion of the CArG motif at -1194 did not affect CAT activity.
Gel mobility shift analyses using the MEF-2-like (Afl"rich) sequence (5"TATTAATATAA-3') located at -1540 to -1530 showed that this element binds to a specific protein complex in nuclear extracts from vascular smooth muscle cells. However, this binding was not competed out by the muscle creatine kinase MEF-2 consensus element. Also, the protein binding to this element was not modified by a MEF-2 antibody, suggesting that the nuclear factor binding to this A5"rich sequence is not a MEF-2 protein. Further analyses will be needed to define the dissimilarity between this AA"rich element binding protein and other already known proteins that bind to some "rich elements such as related to serum response factor (RSRF) (Pollock and Treisman, 1991) or Antp-type homeoproteins (Beachy et al., 1988;Muller et al., 1988). On the other hand, the CArGlike element (5'-CCATATTTAG-3') found at -1275 to -1266 did not demonstrate any specific protein binding. To establish firmly the functional role of these and other elements in the SMHC gene, it will be necessary to perform additional experiments.
In the smooth muscle a-actin gene, a proximal E-box has been reported to be important for expression of the gene (Carroll et al., 1988). Although E-box motifs are present both in the proximal and distal 5'-upstream regions, the exact function of these E-box elements in the SMHC promoter is unclear. However, the presence of constitutively expressed basic helixloop-helix transcription factors in smooth muscle cells (Blackwell and Weintraub, 1990) suggests that E-boxes may have important roles in SMHC gene regulation.
In this study, we show that the SMHC promoter constructs, when transfected into NIH3T3 fibroblasts and C2C,, muscle cells, did not produce significant CAT activity. This further demonstrates that the SMHC gene contains promoter elements that restrict its expression to smooth muscle cells. This is in contrast to the a-smooth muscle actin promoter, which has been shown to be expressed not only in smooth muscle cells but also in fibroblasts and myoblasts (Carroll et al., 1988).
In summary, the results of this study demonstrate that the 5'4anking region of the rabbit SMHC gene contains multiple positive and negative regulatory elements extending to -2266 bp upstream of the transcription start site. Furthermore, we demonstrate that the SMHC promoter directs high levels of CAT expression only in cultured rat smooth muscle cells but not in NIH3T3 fibroblasts or C2C,, myotubes. Therefore, our in vitro transfection studies confirm our earlier observations that the SMHC gene expression is highly restricted to smooth muscle cells. This is also the first study describing functional elements of a smooth muscle-specific promoter. Of particular interest in future studies will be the definition of discrete DNA elements controlling SMHC gene expression in smooth muscle cells. The promoter provides an important tool toward understanding mechanisms regulating smooth muscle cell growth and differentiation and may serve to target expression of specific gene products into smooth muscle cells. script, Dr. Masahiko Kurabayashi for helpful suggestions and continuous encouragement, and C a r p Miller for excellent secretarial assistance.