Identification of promoter elements involved in cell-specific regulation of rat smooth muscle myosin heavy chain gene transcription.

In order to identify cis-acting regulatory elements involved in smooth muscle cell-specific gene regulation, we have cloned a 4. 7-kilobase pair fragment of the promoter for the rat smooth muscle myosin heavy chain, a protein expressed in differentiated smooth muscle cells. Sequence analysis of a 1.7-kilobase pair portion of this clone reveals potential binding sites for known transcription factors. A comparison of the primary sequence between the rat and rabbit smooth muscle myosin heavy chain promoters reveals numerous conserved consensus binding sites. Transient transfection analysis of promoter deletion constructs in rat aorta and tracheal smooth muscle cells, L8 myoblast cells, and rat pulmonary aorta endothelial cells suggests that a region of the promoter located between -1,249 and -1,317 base pairs is important for the restriction of gene expression to smooth muscle cells. Electrophoretic mobility shift analysis of a highly conserved region located between -1,317 and -1, 085 base pairs reveals specific DNA-protein complexes formed in smooth muscle cell extracts, which can be competed with an oligonucleotide containing a nuclear factor 1 binding site.

In order to identify cis-acting regulatory elements involved in smooth muscle cell-specific gene regulation, we have cloned a 4.7-kilobase pair fragment of the promoter for the rat smooth muscle myosin heavy chain, a protein expressed in differentiated smooth muscle cells. Sequence analysis of a 1.7-kilobase pair portion of this clone reveals potential binding sites for known transcription factors. A comparison of the primary sequence between the rat and rabbit smooth muscle myosin heavy chain promoters reveals numerous conserved consensus binding sites. Transient transfection analysis of promoter deletion constructs in rat aorta and tracheal smooth muscle cells, L8 myoblast cells, and rat pulmonary aorta endothelial cells suggests that a region of the promoter located between ؊1,249 and ؊1,317 base pairs is important for the restriction of gene expression to smooth muscle cells. Electrophoretic mobility shift analysis of a highly conserved region located between ؊1,317 and ؊1,085 base pairs reveals specific DNA-protein complexes formed in smooth muscle cell extracts, which can be competed with an oligonucleotide containing a nuclear factor 1 binding site.
Smooth muscle cells (SMCs) 1 express a repertoire of specialized proteins important for the cell's contractile function. Smooth muscle myosin is the primary type-II myosin present in fully differentiated SMCs. Four isoforms of smooth muscle myosin heavy chain (smMHC) exist, produced by alternative splicing from a single gene (1)(2)(3)(4). smMHC isoform composition is tissue-restricted and developmentally restricted and is also modulated in response to changes in SMC differentiation state (2,5,6). The physiological function of the different isoforms remains unclear; however, it is probable that shifts in isoform composition allow cellular adaptation to physiological stresses, as is the case in skeletal muscle (7)(8)(9).
Vascular diseases such as atherosclerosis, hypertension, and restenosis following angioplasty involve smooth muscle cell dedifferentiation and proliferation (5, 10 -13). smMHC is an extensively studied marker, which is specific for differentiated SMCs (2,5,14,15). The expression of smMHC is decreased or absent in proliferating, dedifferentiated SMCs (6, 14, 16 -20). Once SMCs in the neointima of atherosclerotic vessels cease proliferating, they reexpress the SM1 smMHC isoform (6,14). Therefore, the smMHC gene provides a potentially useful promoter for studying the regulation of SMC differentiation state and the factors involved in specifying the proliferative or quiescent/differentiated SMC phenotype.
Relatively little is known regarding the transcription factors involved in the coordinate regulation of smooth muscle-specific genes in differentiated SMCs. Recent studies of the most extensively studied promoter, that for smooth muscle ␣-actin, have shown that there is a complex regulation involving both positive and negative cis-acting elements, as well as important CArG box elements required for smooth muscle-specific expression (21)(22)(23). However, the regulation of this promoter may be somewhat complicated by elements involved in regulating its expression in non-smooth muscle cells such as myofibroblasts and developing cardiac and skeletal muscle (24 -26). Furthermore, the expression of ␣-actin during development and in cell culture suggests that it is an earlier marker on the smooth muscle differentiation pathway than smMHC, which appears to be one of the final markers expressed in fully differentiated smooth muscle cells (5). Studies of additional SMC-specific genes are therefore essential, as it is the coordinate regulation of these genes that is important for establishing the fully differentiated SMC phenotype.
Recently, characterizations of the rabbit smMHC promoter have been reported. Katoh et al. reported a region between Ϫ1,223 and Ϫ1,548 bp, which appears to be important for cell-specific promoter activity in transiently transfected rat SMCs (27). Kallmeir et al. reported the presence of an enhancer element located between Ϫ1,225 and Ϫ1,332 bp, which appears to contribute to SMC-specific expression in rabbit vascular SMCs (42).
The goals of the current study were to clone, sequence, and identify key regulatory elements in the rat smMHC promoter; determine important cis-acting DNA elements governing SMCspecific transcriptional expression; and examine conserved promoter regions for evidence of SMC DNA-binding proteins. The results of this study provide important new evidence for: 1) conservation of smMHC promoter elements in rat to rabbit sequence comparison, including a set of three CArG-like elements; 2) differential gene regulation in vascular versus airway SMCs; 3) the presence of a DNA element located between Ϫ1,249 and Ϫ1,317 bp, which is involved in repressing promoter activity in non-smooth muscle cell types; and 4) the existence of a DNA-binding protein, which interacts with a nuclear factor 1-like site that is present in SMC extracts. * 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) U55179.

EXPERIMENTAL PROCEDURES
Screening of the Rat Genomic Library-A rat genomic DASH II library (Stratagene) containing 2.0 ϫ 10 6 independent clones was screened using a 32 P-radiolabeled fragment containing the 5Ј-untranslated region of a rat smMHC cDNA. Approximately 5.0 ϫ 10 6 clones were screened under stringent hybridization and wash conditions, as described previously (3). Eleven positive clones were obtained and further examined by Southern blot analysis of the phage DNA with the 5Ј-untranslated fragment of the rat smMHC cDNA as probe. A positive clone, RtG-8, was then selected for further analysis. Restriction endonuclease digestion of this clone with NotI, EcoRI, BamHI, and HindIII followed by electrophoresis of the digested fragments on a 0.6% agarose gel indicated that this clone is approximately 16 kb in length. A 4.2-kb BglII fragment and a 4.0-kb EcoRI (RtG-EcoRI) fragment were determined by Southern analysis to contain the 5Ј-untranslated exon and upstream region of the promoter. These clones were used for subsequent mapping, sequencing, and generation of deletion constructs.
Construction of Promoter-Luciferase Expression Plasmids-A 1.2-kb EcoRI-SacI fragment was subcloned from the 4.0-kb RtG-EcoRI clone. This clone extends from a SacI site located at ϩ44 in the 5Ј-untranslated exon to an EcoRI site located at position Ϫ1,259 from the start site. This clone was fully sequenced in both directions using complimentary oligonucleotide primers by the method of Sanger (28) with a Sequenase sequencing kit (U. S. Biochemical Corp.). Deletion constructs were derived via polymerase chain reaction (PCR) using the 1.2-kb EcoRI-SacI promoter fragment as a template. Complementary oligonucleotide primers were designed with SacI sites and commercially synthesized (Operon). Promoter segments were amplified using Taq polymerase (Perkin-Elmer) on a Thermolyne thermocycler. Following a phenol:chloroform extraction and ethanol precipitation, PCR products were digested with SacI and purified on a low melt agarose gel. The isolated promoter segments were then ligated into a SacI-digested pGL2-Basic plasmid (Promega) with T4 ligase (New England Biolabs). Ligated plasmids were then used to transform JM109 cells for preparation of large quantities of plasmid DNA. The orientation and sequence fidelity of the plasmids was determined by Sanger dideoxy sequencing, as well as by restriction enzyme digestion and electrophoresis. The 4.2-kb BglII promoter fragment was directly cloned into the BglII site of pGL2-Basic. Two deletion constructs (p1621LUC and p1317LUC) were generated via PCR using this fragment as a template. All other promoter constructs used the 1.2-kb EcoRI-SacI fragment as a PCR template. Due to the presence of SacI sites in the promoter upstream of Ϫ1.2 kb, the p1621LUC and p1317LUC constructs were cloned into the pGL2 vector using oligonucleotides with HindIII restriction sites, and these clones extend to ϩ88 in the 5Ј-untranslated exon.
Promoter-luciferase expression construct plasmid DNAs used for transient transfections were prepared by alkaline lysis and affinity column-purified (Wizard Mega-prep, Promega) followed by phenol:chloroform extraction, ethanol precipitation, and CsCl density gradient centrifugation. Plasmid preparations were examined by gel electrophoresis for purity and the presence of linearized DNA prior to transfection.
Sequence Comparisons and Analysis-The MacVector DNA analysis program (IBI) was used to compare rat and rabbit smMHC promoter sequences and to identify transcription factor binding sites.
Cell Culture, DNA Transfections, and Reporter Assay-Smooth muscle cells were obtained from adult rat aorta and tracheal tissues by a modification of procedures described by Bochaton-Piallat et al. (29). Adult Sprague-Dawley rats were anesthetized with sodium pentobarbital, and the aorta and trachea were aseptically removed. Following a 15-min digestion in collagenase solution (400 units/ml collagenase, 0.5 mg/ml elastase, 0.5 mg/ml trypsin inhibitor in Hank's balanced salt FIG. 1. A, structure of the rat genomic clone RtG8 with partial restriction map. The BglII sites (*) and EcoRI-SacI sites (**) used for sequencing and subcloning are indicated. B, structure of the rat smooth muscle myosin heavy chain promoter-luciferase fusion constructs. The relative positions of the three CArG-like elements, the TATA box, and a MEF2-like site are indicated. The p4.2, p1621, and p1317 constructs extend to ϩ88 bp in the 5Ј-untranslated exon. All other constructs end at ϩ45 bp in the 5Ј-untranslated exon. solution containing 300 units/ml penicillin, 300 g/ml streptomycin sulfate, and 0.75 g/ml amphotericin B), the trachea were scraped and minced for further digestion. The adventitial layer of the aorta was removed and endothelial cells were removed by scraping with blunt forceps, and the aorta was minced prior to enzymatic digestion. The tissues were digested in fresh collagenase solution for 1-3 h at 37°C, until the cells were completely dissociated. The cells were collected by centrifugation at 3,000 rpm for 5 min, resuspended in Dulbecco's modified Eagle's medium (DMEM, Life Technologies, Inc.), and then recentrifuged to remove all collagenase. The cells were then resuspended in DMEM containing 100 units/ml penicillin, 100 g/ml streptomycin sulfate, and 0.25 g/ml amphotericin B, and seeded onto 100-mm 2 tissue culture dishes. Fetal bovine serum (FBS) was added to 10% after allowing the cells to attach for 1-2 h. These cell cultures typically reached confluence after 3-7 days and were then passaged onto 12-well plates for transfection. At the time of transfection, at approximately 60 -80% confluence, many cells expressed detectable levels of smMHC and smooth muscle ␣-actin as determined by indirect immunofluorescence analysis (data not shown).
L8 myoblast cells were obtained from the American Tissue Culture Collection (ATCC 1769-CRL). These cells were cultured in DMEM containing 10% FBS. The rat pulmonary endothelial cells used were isolated by Dr. Una S. Ryan, according to a previously described method (43). These cells were originally identified as endothelial cells by the presence of angiotensin-converting enzyme activity and reactivity to factor VIII antibodies (44). Endothelial cells were also grown in DMEM with 10% FBS. Cells were cultured to 60 -80% confluence in 12-well plates for transfections.
Transfections were performed in triplicate using Lipofectamine rea-gent (Life Technologies, Inc.) with each well receiving 1.5 g of test plasmid and 5 l of Lipofectamine in a total volume of 500 l of Opti-MEM reduced serum media (Life Technologies, Inc.). Plasmid DNA and Lipofectamine were incubated for 0.5-1 h prior to transfection to allow the formation of DNA-liposome complexes. Cells were washed once with Opti-MEM, and then the DNA-liposome mix was layered over the cells. At 16 h post-transfection, the DNA-liposome mix was removed, the cells were washed with 1 ml of DMEM, 10% FBS, and then fresh medium (DMEM, 10% FBS) was added to each well and the cells grown for 24 h, during which time they usually reached confluence. Cell extracts were obtained by washing the cells three times with phosphate-buffered saline (137 mM NaCl, 27 mM KCl, 8 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 ), followed by incubation for 15 min in 100 l of 1 ϫ reporter lysis buffer (Promega). Cells were scraped into microcentrifuge tubes and centrifuged at 14,000 rpm for 5 min to remove cell debris. A 20-l aliquot of the supernatant was removed for the determination of luciferase activity. The luciferase activity was determined using a luciferase reporter assay kit (Promega) with signal detected via luminometer (Berthold).
Early transfection experiments utilized a secreted alkaline phosphatase plasmid (p-SEAP, Tropix) as a co-transfection plasmid to measure transfection efficiency. However, the secreted alkaline phosphatase measurements did not result in qualitative changes in the data and also had no effect on experimental variability. The use of a co-transfection plasmid was discontinued, primarily to allow the use of lower levels of test DNA in the transfection, as higher levels of DNA per well (Ͼ5 g) were observed to have toxic effects on the cells, and because the p-SEAP plasmid might compete for transcription factors with the smMHC promoter plasmids. Instead, for each experiment, pGL2-Basic, a promot- erless vector was transfected in triplicate to serve as the base-line indicator of luciferase activity. An SV40 enhancer/promoter-luciferase construct (pGL2-SV40, Promega) was transfected in triplicate to serve as a positive control of transfection and the luciferase assay.
All luciferase activity values represent at least three independent experiments in which each plasmid construct was tested in triplicate in each experiment. At least two separate plasmid preparations for each construct were tested. Relative luciferase activity data are expressed as the means Ϯ standard error.
Preparation of Cell Extracts and Electrophoretic Mobility Shift Analysis-Whole cell extracts of cultured cells were prepared by freezethaw, essentially by the method of Ladias et al. (30). Tissue extracts were prepared in a similar manner, with the exception that the tissues were frozen in liquid nitrogen upon dissection from the animal, then crushed under liquid nitrogen and homogenized in a Dounce homogenizer in cell scraping buffer (40 mM Tris, pH 7.4, 1 mM EDTA, and 150 mM sodium chloride) before proceeding with the extraction protocol. Cell extracts were aliquoted and stored at Ϫ70°C until use.
The promoter segment used in the electrophoretic mobility shift assay (EMSA) analysis was generated by PCR using complimentary primers and the 4.2-kb BglII promoter construct (p4.2LUC) as a template. Following PCR, the promoter segment was purified by gel purification on a low melt agarose gel. This fragment was end-labeled using [␥-32 P]ATP and T 4 kinase, followed by removal of unincorporated nucleotides on a NucTrap push column (Stratagene).
Binding reactions were loaded and electrophoresed on a 5% polyacrylamide gel, which had been pre-run at 150 V for 1 h. Electrophoresis was performed at 150 V in Tris-glycine buffer (200 mM glycine, 25 mM Tris, 1 mM EDTA, pH 8.3). Gels were dried and exposed to film for 24 -72 h with an intensifying screen at Ϫ70°C.

Isolation of the Rat smMHC Gene Promoter and Sequence
Analysis-The rat smMHC gene promoter was isolated from a rat genomic DASH II library by screening with a 1.2-kb EcoRI fragment of a smMHC cDNA containing 80 bp of the 5Ј-untranslated region. The strongest positive of 16 positive clones examined by Southern blot, RtG8, was further characterized by restriction enzyme and Southern blot analysis and determined to contain the rat smMHC gene promoter. This clone was found to contain approximately 4.7 kb of upstream promoter (Fig. 1).
The sequence of the promoter extending to Ϫ1,699 bp is shown in Fig. 2. A TATA box (TATAA; Ref. 31) was located at position Ϫ24 bp; however, no typical CAAT box motif could be found in this promoter. The transcription start site, as well as the 5Ј-untranslated exon's intron/exon boundaries, were determined by sequencing several cDNA clones obtained by a rapid amplification of cDNA ends procedure conducted on rat aorta mRNA (data not shown). An examination of the sequence for DNA-binding protein/transcription factor consensus binding sites revealed numerous sites for known factors. Three CArG box-like motifs (32,33) are located at positions Ϫ1297, Ϫ1223, and Ϫ1106 bp upstream of the transcription initiation site. Eight E boxes, as well as numerous Sp1 and Ap1 sites are found throughout the promoter (32). Several NF1-like sites (34) are located in the promoter, most notably at Ϫ1276 and Ϫ1143 bp, adjacent to the CArG box cluster.
A comparison of the rat smMHC promoter sequence to the rabbit smMHC promoter sequence revealed that the two sequences are remarkably dissimilar over the Ϫ1.7 kb to ϩ75 bp region. Two main regions of relatively high similarity were identified. The Ϫ1323 to Ϫ1098 bp region contains the CArG box-like motifs, which are completely conserved (Fig. 3A). This region also contains two NF1-like sites, which are not precisely conserved between the rat and rabbit primary sequences. The Ϫ118 to ϩ60 bp region contains the TATA box and flanking sequences (Fig. 3B). The TATA box is completely conserved, as are several regions adjacent to it. The E box present in the rat

Negative-acting cis-Elements Are Involved in Cell-type Specific Regulation of the Smooth Muscle Myosin Heavy Chain
Promoter-As a first step toward defining the regulatory elements required for promoter activity in smooth muscle cells, SMCs derived from adult rat aortas and trachea were transiently transfected with fusion plasmids containing portions of the rat smMHC promoter fused to a luciferase reporter gene (Fig. 1B). The promoter construct p1249 was found to be the most active in adult aorta cells with a relative luciferase activity approximately 30-fold over background (Fig. 4A). Addition of upstream sequences led to decreases in relative luciferase activity, as did deletion of sequences downstream of Ϫ1249 bp. The p48 construct was found to be completely inactive, indicating that elements located between Ϫ138 and Ϫ48 bp are required for basal promoter activity. The two larger constructs, p4.2 and p1621, were found to have relatively low levels of luciferase activity, suggesting the presence of negative regulatory elements within these regions.
Transient transfection of promoter constructs into adult rat tracheal SMCs produced a somewhat different pattern of luciferase activity (Fig. 4B). The p1249 was again one of the most active plasmids; however, deletions of downstream elements appear to have less of an effect on luciferase activity in tracheal SMCs, compared to aorta cells. As observed for the adult aorta SMCs, the p48 construct is inactive. Addition of regions upstream of Ϫ1249 bp also leads to decreased luciferase activities, as observed for the rat aorta SMCs.
In order to identify cis-elements involved in cell-specific gene regulation, the smMHC promoter deletion constructs were transiently transfected into non-SMC types (Fig. 5). Neither rat pulmonary aorta endothelial cells (rPAECs) nor rat L8 myoblast cells express endogenous smMHC protein, based upon immunofluorescence staining with smMHC-specific antibodies (data not shown). In the rPAECs, the p48 construct was inactive, as it was in SMCs. The three shortest constructs had relatively high luciferase activities, ranging from 18-to 50-fold over background. The luciferase activity was somewhat depressed in p825 and p1249 constructs, but was still 8 -15-fold higher than background. Addition of regions upstream of Ϫ1249 bp led to abrupt decreases in luciferase activity. The activity of the p4.2, p1621, and p1317 constructs was reduced to background or near background levels.
Transient transfections into L8 myoblasts revealed a pattern of luciferase activity similar in some respects to the rPAEC pattern. The p48 construct again showed minimal levels of activity, while the p825, p602, p291, and p138 constructs all showed high levels of activity. However, unlike the rPAECs, L8 cells transfected with p1249 demonstrated very high activity levels, approximately 50 -100-fold over background. Addition of sequences upstream of Ϫ1249 bp, however, led to greatly decreased levels of luciferase activity similar to that observed in the rPAECs. The composite data strongly suggest the presence of a negative-acting cis element, located between Ϫ1317 and Ϫ1249 bp, which functions to restrict expression of the smMHC promoter to SMCs.
The p4.2, p1621, and p1317 constructs contain an extra 43 bp of 5Ј-untranslated exon, due to the use of the BglII site for cloning rather than the adjacent SacI site. In order to eliminate the possibility that this 43-bp region was responsible for the cell-specific expression of the p4.2, p1621, and p1317 con-structs, two additional constructs (Ϫ1249 to ϩ88 and Ϫ1249 to ϩ600 bp) were made and transiently transfected into rat aorta and tracheal SMCs, L8 myoblasts, and rat pulmonary endothelial cells. The pattern of luciferase activity observed for these constructs did not differ from that of the p1249 construct (data not shown). Therefore, the additional 43 bp of 5Ј-untranslated exon present in the p4.2, p1621, and p1317 constructs is not sufficient to cause the cell-specific expression observed for these promoter constructs.
Factors Present in Smooth and Non-smooth Muscle Cell Types Bind to the Ϫ1317 to Ϫ1085 bp Fragment of the smMHC Promoter-To investigate the relevance of the observed conservation of sequence between the rat and rabbit promoters in the Ϫ1.3 to Ϫ1.1 kb region, EMSAs were performed on cell extracts using a probe which extended from Ϫ1,317 to Ϫ1,085 bp. This region contains the three CArG box-like elements, as well as two NF1-like elements.
EMSA using whole cell extracts from primary rat tracheal and aortic SMCs, rat pulmonary aorta endothelial cells, and rat L8 myoblast cells reveals a large shifted complex, which appears similar in all cell extracts (Fig. 6). This DNA-protein complex can be completely abolished by addition of excess unlabeled probe or by the addition of 100-fold excess unlabeled oligonucleotide containing an NF1 binding site, when tested with rat tracheal SMC extract (Fig. 7).
Whole cell extracts made from intact aorta and tracheal tissues form a pattern of shifted complexes which is different from that observed in cultured cells (Fig. 8). These shifted bands can also be abolished by addition of excess unlabeled probe or NF1 oligonucleotide to a binding reaction containing rat aorta extract (Fig. 9). In order to determine which of the NF1-like sites located in the Ϫ1,317 to Ϫ1,085 probe is involved in the formation of this DNA-protein complex, two oligonucleotides, corresponding to the two NF1-like sites were used in a competitive EMSA (Fig. 10). Addition of cold double-stranded oligonucleotide spanning from Ϫ1147 to Ϫ1124 was found to abolish formation of the DNA-protein complexes in extracts derived from cultured primary SMCs and in tracheal and aorta tissue cell extracts (Fig. 10A). Competitive EMSA using an oligonucleotide spanning the Ϫ1275 NF1 site did not affect formation of the DNA-protein complex (Fig. 10B). However, an oligonucleotide containing the NF1 site from an adenovirus sequence is capable of effectively competing for the proteins present in this DNA-protein complex. These data suggest that factors present in cultured cell extracts or extracts derived from smooth muscle tissues bind specifically to the NF1-like element located at Ϫ1143 bp.

DISCUSSION
The goal of the present study was to isolate the rat smMHC promoter and begin the characterization of cis-acting elements and trans-acting factors that regulate the promoter. A comparison of the rat to the rabbit promoter sequence revealed two regions of relatively high identity, with highly divergent intervening sequence. The TATA box and flanking sequences are somewhat conserved, as is a region located over 1 kb from the start of transcription. The region, Ϫ1323 to Ϫ1098 bp, contains three CArG box-like elements, which are completely conserved, as well as two NF1-like elements that are not completely conserved between rat and rabbit sequences. The conservation of regions of the promoter between the two species suggests that they may be important for regulation of gene expression. constructs shown in Fig. 1B. Luciferase activity is expressed relative to the base-line luciferase activity of a promoterless-luciferase construct (pGL-Basic) set to equal 1. An SV40 enhancer/promoter-luciferase construct (pGL-SV40) was transfected as a positive control. Data shown are from 5 independent experiments performed in triplicate. Data are presented Ϯ S.E. Sequence analysis of the promoter for known transcription factor binding sites revealed numerous binding motifs. The three CArG box-like motifs clustered at Ϫ1297, Ϫ1223, and Ϫ1106 bp were most striking. Serum response factor (SRF), a member of the MADS box transcription factor family, recognizes and binds to CArG box elements (45). CArG box elements may confer tissue specificity to gene expression by competitions for the site between different factors, or by utilization of tissuerestricted SRFs or SRF-associated proteins (46). The triplet set of CArG boxes may potentially play a role in the coordinate regulation of SMC contractile proteins, as three CArG box motifs have also been located in the smooth muscle ␣-actin promoter (21)(22)(23). These CArG box elements have been shown to be important for tissue-specific positive activation of the smooth muscle ␣-actin promoter, which may involve cell-specific CArG box-binding proteins (21).
Two nuclear factor 1-like sites were identified, located at Ϫ1275 and Ϫ1143 bp. NF1 exists as a family of proteins, which were originally isolated as adenovirus replication factors (38). NF1, which is ubiquitously expressed, has been shown to act as either a positive or a negative regulator of transcription (39,40). The importance of these elements in smMHC gene regulation is currently unknown.
Eight E box motifs were found scattered throughout the smMHC promoter. The smooth muscle ␣-actin promoter has also been shown to contain E boxes, known to be important for muscle-specific transcription (41), as do many promoters that are active in striated muscle. E box elements are bound by heterodimers of E proteins and helix-loop-helix proteins. Their importance with regard to smooth muscle-specific transcription is unknown; however, the presence of helix-loop-helix proteins in SMCs suggests that they may potentially play some role in   SMC-specific gene regulation (41).
The transient transfections clearly demonstrate the complexity of the smMHC promoter. Multiple positive and negative regulatory regions exist, and it appears that they are differentially utilized in different smooth muscle cell types. In both the rat aorta and tracheal SMCs, the Ϫ1,249 bp promoter construct produced high luciferase activity. However, whereas shorter constructs produced decreased luciferase activity in aorta SMCs, this decrease was not observed in tracheal SMCs. The Ϫ1249 bp construct appears to contain a positive regulatory element important for gene expression in aorta SMCs, but this element appears to be less important for promoter activity in tracheal SMCs. This result also further emphasizes the importance of cell context with respect to studies of the smMHC promoter, as these two SMC types appear to be able to utilize different regulatory strategies.
A surprising result of the transient transfection experiments was the high activity of smMHC promoter constructs p1249, p825, p602, p291, and p138 in rat L8 myoblast and rat pulmonary aorta endothelial cells. This activity was, however, nearly abolished by the addition of 68 bp of 5Ј-flanking sequence to the Ϫ1249 bp construct. This result strongly suggests the presence of a negative regulatory element, which represses promoter activity in non-SMC types. Addition of this region also reduced luciferase activity in the rat aorta and tracheal SMCs. This result may be a reflection of the heterogeneous population of cells present in primary smooth muscle cell cultures representing varying degrees of differentiation, some of which are no longer able to express smMHC (16 -21). For example, an ele-  10. A, competitive electrophoretic mobility shift analysis of proteins binding to the Ϫ1317 to Ϫ1085 bp promoter fragment competed with DNA containing NF1 consensus binding sites. Lane 1 shows the probe alone. Lanes 2 and 3 show the shifted complexes observed in cultured rat tracheal SMCs (rTSMC) and tracheal tissue extract (Trach.). Unlabeled probe, an oligonucleotide containing the NF1-like site present at Ϫ1143 bp, and a CTF/NF1 oligonucleotide were added as indicated in the reactions in lanes 4 -9. Lanes 10 and 11 show the shifted complexes observed in rat aorta SMCs (rASMC) and aorta tissue extracts. Lanes 12 and 13 show the effect of addition of the oligonucleotide containing the NF1-like site at Ϫ1143 bp (SMHC 1143) to the reactions. B, competitive electrophoretic mobility shift analysis of proteins binding to the Ϫ1317 to Ϫ1085 bp promoter fragment competed with DNA containing NF1-like binding motifs. Lane 1 shows the probe ment contained in this 68-bp region may be important for restricting gene expression in non-SMCs and in silencing SMCspecific gene expression in SMCs that have modulated their phenotype and are no longer able to express differentiated SMC-specific genes.
Studies of the rabbit smMHC promoter also demonstrate that a region of the promoter located between Ϫ1332 bp and Ϫ1225 bp is important for high activity specific to SMCs (42). In another study, transfections of the rabbit promoter fragments into SMCs, fibroblasts (NIH3T3), and myoblasts (C2C12) also showed a dramatic SMC-specific increase in activity when a Ϫ1548 bp construct was used, whereas a Ϫ1223 bp construct's CAT activity was not significantly different from that observed for the fibroblasts and myoblast cells (27). These results are consistent with our findings, and we are further able to refine the location of an important regulatory element to a region lying between Ϫ1249 and Ϫ1317 bp. However, unlike the enhancer element recently described in the rabbit promoter, this study has revealed an element that acts as a repressor of gene activity in non-SMCs. This combination of positive and negative regulatory elements contribute to the cell specificity of smMHC expression.
Electrophoretic mobility shift assays of a region that is highly conserved between the rat and the rabbit smMHC promoter sequence (Ϫ1317 to Ϫ1085 bp) demonstrated a specific DNA-protein complex formed in extracts from cultured smooth muscle cells. The DNA-binding proteins were found to interact with an NF1-like site in the promoter located at Ϫ1143 bp. Cell extracts derived from aorta or tracheal smooth muscle tissues show two smaller complexes formed in EMSA analysis using the Ϫ1317 to Ϫ1085 bp probe. Both complexes are also due to binding to the NF1-like site at Ϫ1143 bp. The importance of this NF1-like site in terms of promoter regulation and the precise identity of the DNA-binding proteins that interact with it remains to be determined, as does the relevance of the different complexes formed in extracts from cultured SMCs versus SMCs from smooth muscle tissue.
In summary, we have presented new evidence for a regulatory element that acts to repress activity of the smMHC gene in non-SMCs. We have also identified a nuclear factor 1-like site that is recognized and bound by proteins present in SMC and non-SMC extracts. Transfection analysis of vascular and airway SMCs provides the first evidence for cell-specific regulation of the smMHC gene. Future studies of this promoter will be required to further define the repressor element, to identify the factors that bind to this region and to ascertain the role of the nuclear factor 1-like site in the regulation of gene activity.