Transcriptional regulation of the chicken caldesmon gene. Activation of gizzard-type caldesmon promoter requires a CArG box-like motif.

Caldesmon, which plays a vital role in the actomyosin system, is distributed in smooth muscle and non-muscle cells, and its isoformal interconversion between a high M(r) form and low M(r) form is a favorable molecular event for studying phenotypic modulation of smooth muscle cells. Genomic analysis reveals two promoters, of which the gizzard-type promoter displays much higher activity than the brain-type promoter. Here, we have characterized transcriptional regulation of the gizzard-type promoter. Transient transfection assays in chick gizzard smooth muscle cells, chick embryo fibroblasts, mouse skeletal muscle cell line (C2C12), and HeLa cells revealed that the promoter activity was high in smooth muscle cells and fibroblasts, but was extremely low in other cells. Cell type-specific promoter activity depended on an element, CArG1, containing a unique CArG box-like motif (CCAAAAAAGG) at -315, while multiple E boxes were not directly involved in this event. Gel shift assays showed the specific interaction between the CArG1 and nuclear protein factors in smooth muscle cells and fibroblasts. These results suggest that the CArG1 is an essential cis-element for cell type-specific expression of caldesmon and that the function of CArG1 might be controlled under phenotypic modulation of smooth muscle cells.

Caldesmon, which plays a vital role in the actomyosin system, is distributed in smooth muscle and non-muscle cells, and its isoformal interconversion between a high M r form and low M r form is a favorable molecular event for studying phenotypic modulation of smooth muscle cells. Genomic analysis reveals two promoters, of which the gizzard-type promoter displays much higher activity than the brain-type promoter. Here, we have characterized transcriptional regulation of the gizzard-type promoter. Transient transfection assays in chick gizzard smooth muscle cells, chick embryo fibroblasts, mouse skeletal muscle cell line (C2C12), and HeLa cells revealed that the promoter activity was high in smooth muscle cells and fibroblasts, but was extremely low in other cells. Cell type-specific promoter activity depended on an element, CArG1, containing a unique CArG box-like motif (CCAAAAAAGG) at ؊315, while multiple E boxes were not directly involved in this event. Gel shift assays showed the specific interaction between the CArG1 and nuclear protein factors in smooth muscle cells and fibroblasts. These results suggest that the CArG1 is an essential cis-element for cell type-specific expression of caldesmon and that the function of CArG1 might be controlled under phenotypic modulation of smooth muscle cells.
Smooth muscle cells (SMCs) 1 undergo remarkable phenotypic modulation during embryogenesis. A converse transition of arterial SMCs from a differentiated to dedifferentiated phenotype is one of major events in the onset of atherosclerosis (1,2). Although molecular approaches of such phenotypic modulation are important for understanding vascular pathogenesis, only limited information is available. Of these, ␣-smooth muscle actin (␣-SM actin) was considered to be a suitable molecular marker for differentiation of SMCs (3). Recent studies have led to suspect the significance of this protein because its expression has been found in skeletal muscle cell line (4) and certain stromal cells (5).
Caldesmon (CaD) plays a vital role in the Ca 2ϩ -dependent regulation of smooth muscle and non-muscle contraction (6,7).
The two CaD isoforms have been identified. h-CaD (high M r form) is dominantly expressed in differentiated SMCs, while l-CaD (low M r form) is widely distributed in non-muscle tissues and cells (8 -10). In particular, the isoformal interconversion of CaD is tightly associated with phenotypic modulation of SMCs, in which the CaD isoform converts from the l-to h-form during differentiation and vice versa (11)(12)(13). CaD is therefore a favorable molecular marker for studying phenotypic modulation of SMCs. Genomic analysis has revealed that the expression of h-or l-CaDs depends on a unique selection of two 5Ј-splice sites within the exon 3 (14,15). Another important molecular event is the up-regulation of CaD expression during SMC differentiation (11). Several cytoskeletal proteins such as myosin heavy and light chains (16,17), ␣-SM actin (3), tropomyosin (18), vinculin (12), metavinculin (12), calponin (13,19), and SM22 (20) are also up-regulated in association with SMC differentiation. Contrarily, their expressions are down-regulated during dedifferentiation. The expressional changes of these cytoskeletal proteins in their amounts might be controlled at a transcriptional level. However, their transcriptional regulations have been scarcely investigated. The ␣-SM actin and vinculin genes have been partially characterized (4,(21)(22)(23). In our previous report (24), we have identified two CaD promoters, gizzard-type and brain-type promoters, in which the gizzard-type promoter shows much higher activity than the brain-type promoter. Here, we have characterized the transcriptional regulation of the gizzard-type promoter, which actively functions in SMCs and chick embryo fibroblasts (CEFs) but is unable to promote high levels of transcriptional activity in other cell types such as C2C12 and HeLa cells. The promoter activity in differentiated SMCs was higher than that in dedifferentiated SMCs. This result coincided with the high expression of h-CaD in differentiated SMCs compared with the low expression of l-CaD in dedifferentiated cells. We have also demonstrated that the cell type-specific expression of the CaD gene is regulated by a single cis-element, CArG box-like motif (CCAAAAAAGG), located between Ϫ309 to Ϫ300, whereas multiple E boxes located in the 5Ј-upstream region are not directly involved in this event.
dishes. It has been reported that laminin retards the oneset of dedifferentiation of cultured arterial SMCs (26). In this experiment, we have also found that the h-CaD expression in chick gizzard SMCs was maintained for several days under our culture conditions (see "Results"). Therefore, we chose to use this culture system as differentiated SMCs for transient transfection of CAT constructs. The isolated SMCs were cultured in DMEM supplemented with 10% fetal calf serum for more than 1 week to promote dedifferentiation. In these cells, the CaD expression was observed to convert from the h-to l-form (see "Results"). CEFs were cultured as described elsewhere (24). A clonal cell line from mouse skeletal muscle cell, C2C12 myoblast, was cultured as described previously by Blau et al. (27). Mononucleated myoblasts were cultured in growth medium (DMEM supplemented with 20% fetal calf serum) at low density, while differentiation was induced by switching confluent cultures to low serum medium (DMEM supplemented with 2% horse serum). HeLa cells were cultured in DMEM supplemented with 10% fetal calf serum. Transfections and CAT assay (28) were carried out as follows. Calcium phosphate-DNA precipitates containing 8 g of CAT construct plus 1 g of control plasmid carrying the luciferase cDNA under chicken ␤-actin or Rous sarcoma virus (RSV) promoter were added to the cultured cells. Transfection into differentiated or dedifferentiated SMCs was performed at 12 h postseeding the cultured SMCs on laminin-coated dishes in DMEM supplemented with 0.2% bovine serum albumin for differentiated SMCs or in DMEM supplemented with 10% fetal calf serum for dedifferentiated SMCs, and the cells were harvested at 24 h after transfection. Both SMCs and CEFs were exposed to glycerol shock for 30 s after 4 h of transfection, which enhanced the transfection efficiencies. CEFs were harvested at 48 h after transfection. Transfection into C2C12 myoblasts and HeLa cells was performed according to the same procedure for CEFs except that the cells were left in the presence of calcium precipitates for 15-20 h without glycerol shock. For C2C12 myotubes, myoblasts were transfected as the same procedure described above, and then the medium was changed to low serum medium to induce myotube formation. C2C12 myotubes were harvested at 72 h after the medium change. Standardization of transfection efficiency using luciferase activity was carried out according to the method described elsewhere (29). The appropriate volume of the cell extracts after heating to inactivate endogenous deacetylases was incubated at 37°C with 1 mM acetyl coenzyme A and 3.7 kBq of [ 14 C]chloramphenicol (Amersham Corp.) and analyzed by thin layer chromatography. pUC0CAT and pUC2CAT (30) were used as negative and positive controls, respectively. The transfection experiments were repeated on multiple sets of cultures with two or three different plasmid preparations. CAT activity was quantified by scanning Imager (Molecular Dynamics), and the average values are shown.
In Vivo Competition Assay and Serum Effect on CaD Promoter Activity-For in vivo competition assay, GP1Db-21CAT and either competitor plasmid, pUCGE100, or control plasmid, pUC18, were cotransfected into CEFs. Serum effect was analyzed as follows. After transfection of the indicated CAT construct and RSV-luciferase plasmid, CEFs were either serum starved (in DMEM without fetal calf serum) for 50 h, or serum starved for 42 h and then restimulated in the growth medium containing 10% fetal calf serum for 8 h. To analyze the effect of serum on the gizzard-type CaD expression in CEFs, we performed Northern blotting. Total cellular RNAs isolated from CEFs cultured under the same conditions as described above were hybridized with the gizzard-type CaD-specific probe as described elsewhere (24). We also carried out the positive control of serum inducibility in CEFs as follows. CEFs were transfected with ␤-actin luciferase plus pUC2CAT. The luciferase activities from ␤-actin promoter under serum-starved or stimulated conditions were assayed using CAT activity of cotransfected pUC2CAT as the standardization of transfection efficiency.
Analysis of DNA-Protein Interaction by Gel Shift Assay-Probes are described under "Results." GE100 and GE80(⌬CArG) were isolated from GP1Db-21CAT and GP1(⌬CArG)aCAT by EcoRI/SphI-digestion, respectively. CArG1, CArG2, CArG3, and CArG1M were prepared by annealing respective sense and antisense synthesized oligonucleotides to form duplex DNA. Nuclear extracts from SMCs and CEFs were prepared according to the procedures described elsewhere (31). For characterization of DNA-protein binding, samples of nuclear extracts were mixed with 0.1-0.2 ng of 32 P-labeled probe and 3.5 g of heatdenatured herring sperm DNA in the presence or absence of nonradiolabeled competitor at room temperature for 30 min in 20 l containing 5 mM HEPES, pH 7.8, 5 mM ␤-mercaptoethanol, 1 mM EDTA, 60 mM NaCl, 5 mM spermidine, and 10% glycerol. Samples for gel shift assay were analyzed on 4% polyacrlyamide gels in 0.5 ϫ Tris/borate/EDTA buffer. Fig. 1 shows the expression of CaD isoforms in several cell types. High levels of the expression were detected in SMCs and CEFs (lanes 1-3). It has been demonstrated that the expression of h-CaD is specific in differentiated SMCs (8 -10), and that the CaD isoforms convert from the h-to l-form during dedifferentiation of SMCs (11-13) (Fig. 1, lanes 1 and 2). In addition, the h-CaD expression at the protein level in differentiated SMCs was higher than the l-CaD expression in dedifferentiated SMCs (Fig. 1, lanes 1 and 2). Suprisingly, CEFs expressed high levels of both h-and l-CaDs (24) (Fig. 1, lane 3). Skeletal muscle cell line (C2C12 cells) and carcinoma cell line (HeLa cells) expressed low levels of CaD (lanes 4 -7). Among subtypes of CaD, the gizzard-type CaD was dominant in in SMCs (data not shown) and CEFs (24). Although two distinct transcriptional machineries (gizzard-and brain-type promoters) have been identified in the chicken CaD gene, the gizzardtype promoter displays much higher activity than the braintype promoter (24). We further characterized the transcriptional regulation of the gizzard-type promoter. Fig. 2A shows the schematic diagram of the gizzard-type promoter from Ϫ3041 to ϩ60 (24) and its CAT constructs. The relative activities of CAT constructs are shown in Fig. 2B. A series of deletions from GP3CAT to GP1Db-21CAT showed equally high activities in differentiated SMCs and CEFs, whereas the activity of GP1(SphI)CAT was dramatically decreased. These results indicate that the sequence from Ϫ217 to ϩ1 is the basal promoter region consisting of TATA box, Sp1 binding site-like sequence, and CCAAT box, and that the upstream from Ϫ315 to Ϫ218, GE100, is essential for the positive promoter activity, whereas E boxes (reviewed in Refs. 32-34) located in the upstream region would not be directly involved in the transcriptional activities in SMCs and CEFs. The promoter activities of the deletions from GP3CAT to GP1Db-21CAT in dedifferentiated SMCs were 40 -70% of those in differentiated SMCs. In this case, the CaD expression at the protein level (Fig. 1, lanes  1 and 2) were in good agreement with the gizzard-type promoter activities. In C2C12 myoblasts and myotubes and HeLa cells, the activities of CAT constructs were low. These results were coincided with immunoblotting data, in which the CaD contents were very low in those cells (Fig. 1). However, the proximal E box (Ϫ560 to Ϫ565) might be slightly involved in the up-regulation of the promoter in C2C12 myoblasts and myotubes, because the CAT activity was decreased between GP1CAT and GP1Db-21CAT. GE100 was also able to increase the brain-type promoter activity; BP1CAT showed only weak promoter activity (24), while the chimeric construct, GE100/ BP1CAT, carrying GE100 at the upstream of the brain-type promoter, increased in the promoter activity of BP1CAT in SMCs and CEFs. In addition, the activity of GP1Db-21CAT in CEFs was suppressed by cotransfection of pUCGE100 carrying only GE100 (Fig. 3A). These results indicate that cell typespecific high expression of gizzard-type CaD depends on an enhancer element in GE100. Specific DNA-protein complex was demonstrated by gel shift assay using nuclear extracts from CEFs and 32 P-labeled GE100 as a probe (Fig. 3B, lanes 1  and 2). Since GE100 consists of a CArG box-like motif (Fig. 4), we investigated to identify the protein binding region in GE100. The complex formation was suppressed by the addition of unlabeled CArG1 containing a CArG box-like motif and its 5Ј-and 3Ј-flanking 6-nucleotide sequences, but not by GE80(⌬CArG) (Fig. 3B, lanes 3 and 4, and Fig. 4). The DNAprotein complexes showing identical migration in gels were also found using respective nuclear extracts from both differentiated and dedifferentiated SMCs, and the quantities of such complexes were especially high using the differentiated SMC extracts (Fig. 3C).

Specific Expression of the Gizzard-type CaD Depends on GE100 -
A Unique CArG Box-like Motif, CCAAAAAAGG, Is a Key cis-Element in GE100 -To search for a key cis-element in GE100 which is directly involved in activation of the gizzard-FIG. 3. In vivo competition assay and specific DNA-protein interaction using GE100. A, In vivo competition assay was carried out by cotransfection with GP1Db-21CAT (4 g) and competitor (20 g), pUCGE100, or control, pUC18, into CEFs. Relative CAT activity was based on the activity of the cells cotransfected with GP1Db-21CAT and pUC18. B and C, specific DNA-protein interaction was analyzed by gel shift assay using 32 (24). Schematic structures of deleted or chimeric CAT constructs are shown under the map. Thick and thin lines indicate the sequences of the gizzard-and brain-type CaD promoters, respectively. Open boxes indicate the GE100 and the CArG1. B, relative CAT activities of respective CAT constructs in differentiated SMCs, dedifferentiated SMCs, CEFs, C2C12 myoblasts, C2C12 myotubes, and HeLa cells are shown. They were normalized to the activity of pUC2CAT in respective cells as 100%. BP1CAT and GE100/BP1CAT were transfected into only both types of SMCs and CEFs. To account for differences in transfection efficiencies, the level of luciferase activities from control plasmid carrying RSV promoter and luciferase cDNA were assayed. type promoter in SMCs and CEFs, we constructed deletions from GP1Db-21CAT (Fig. 5). The activities of GP1 (⌬CArG)aCAT and GP1(⌬CArG)bCAT, in which a CArG boxlike motif was deleted, were decreased to 25% of GP1Db-21CAT. In contrast, 3Ј⌬GE100/GP1(SphI)CATc14, in which 3Ј-region of GE100 (Ϫ262 to Ϫ218) was deleted, retained high activities. CArG1M/GP1(⌬CArG)aCAT carrying a mutated CArG box-like motif, CArG1M (Fig. 4), was unable to enhance the promoter activity. CArG1AS/GP1(⌬CArG)aCAT, in which the CArG1 was inserted at Ϫ299 in the antisense orientation, displayed 3-4-fold higher activity than GP1(⌬CArG)aCAT. The CArG1 also enhanced the promoter activity at the same level as GP1Db-21CAT when it was inserted at Ϫ248 (CArG1/ GP1(⌬CArG)bCAT). On the contrary, the CArG2 and CArG3 lacking the 5Ј-or 3Ј-flanking sequence (Fig. 4) were unable to enhance the promoter activity as strongly as was the CArG1 (data not shown). The activities of CArG1-0/GP1(SphI)CAT, in which the CArG1 was inserted at the SphI site (Ϫ217 to Ϫ212), were 4 -7-fold higher than those of the GP1(SphI)CAT. The present results indicate that the CArG1 is a key cis-element in GE100 for activation of the basal promoter in a cell typespecific manner.
Gel shift assay using CEF and differentiated SMC nuclear extracts revealed the specific CArG1-protein complex formation with identical migration in gels, because such complexes were suppressed by unlabeled CArG1, but not CArG2, CArG3, and CArG1M (Figs. 4 and 6, A and B). Conversely, radiolabeled deletion and/or mutation probes did not form the DNA-protein complex (data not shown). The amounts of such complex were also high in the differentiated SMC extracts (Fig. 6B). The results of transient transfection assay (Fig. 5) and gel shift assay (Fig. 6) indicate that the interaction between the CArG1 and nuclear protein factors is essential for enhancement of the basal promoter activity. Since the promoter activities in differentiated SMCs and CEFs were nearly equal in spite of the difference in the amounts of the CArG1-protein complex, the quantities of the complex were not directly correlated to the enhancement. The CArG1-protein complex was resistant to high salt (120 mM NaCl) and did not require Mg 2ϩ , but was sensitive to orthophenanthroline, a Zn 2ϩ chelator; 5 mM orthophenanthroline suppressed the complex formation (data not shown).

Serum Effect on CaD Promoter Activity-The present results
in which the gizzard-type promoter in differentiated SMCs was able to promote high levels of transcriptional activity under serum-free conditions suggest that the promoter would not be affected by serum stimulation. Since the serum stimulation promotes modulation of cultured SMCs from a differentiated into a dedifferentiated phoenotype, it is impossible to study the serum effect on the promoter activity in differentiated SMCs. In this case, we chose CEFs to examine the serum effect on the promoter activity. CEFs transfected with GP1CAT and GP1Db-21CAT were cultured under either serum-starved or -stimulated conditions. The transcription from both constructs was not activated by serum (Fig. 7A). Similarly, the endogenous CaD expression was not affected when examined by Northern blotting (Fig. 7B) and immunoblotting (data not shown). In contrast, ␤-actin promoter, which shows serum inducibility (35), was activated (3.6-fold) by serum (Fig. 7C). We further confirmed the serum inducibility of vinculin expression (3-fold) by immunoblotting (data not shown). Therefore, the activation by the CArG1 would not depend on serum stimulation. DISCUSSION The expressional changes of CaD isoforms and in their contents are closely associated with phenotypic modulation of SMCs (11)(12)(13). CaD is therefore considered to be a favorable molecular marker for studying such phenotypic modulation. The CaD expression is regulated by two means, splicing and transcription, within a single gene (14,15,24). Maturation of mRNAs for CaD isoforms is determined by a unique splicing; the expression of h-or l-CaDs depends on a selection of two 5Ј-splice sites within exon 3 (14,15). The change in CaD content is determined by the regulation of promoter activities. Characterization of the factors involved in the expressional regulation of the key genes is important for phenotypic modulation of SMCs. Although little is known about the transcriptional regulation of the ␣-SM actin gene in differentiated SMCs, the CArG boxes are reported to be key cis-elements in the regulation of ␣-SM actin promoter in dedifferentiated SMCs and skeletal muscle cell lines (4,21,22). Recent studies have expressed doubt that the ␣-SM actin is suitable for a molecular marker of SMC phenotypic modulation because its expression is not restricted within SMCs (4,5). Recently, the promoter region of the human vinculin gene has been partially analyzed (23). Although it contains a CArG box and shows serum inducibility, the involvement of cis-elements in the activation of the vinculin promoter is unknown. At present, the expressional regulations of SMC-specific molecular markers have not been well characterized, and cis-elements and transacting factors involving in SMC-specific transcription also remain unclear.
We have previously demonstrated the cloning of the gizzardand brain-type CaD promoters, in which the gizzard-type promoter displays much higher activity than the brain-type promoter (24). In the present study, we have characterized the transcriptional regulation of the gizzard-type promoter in SMCs, CEFs, C2C12, and HeLa cells. The gizzard-type promoter displays SMC-specific high expression except for CEFs (Fig. 2). In addition, both the promoter activity and the protein level of CaD in differentiated SMCs were higher than in dedifferentiated SMCs (Figs. 1 and 2). At present, it is unknown why the promoter activity is high in CEFs and the h-CaD is expressed in these cells. At any rate, CEFs are one of suitable subjects for the present purpose. It has been further clarified that only a limited region expanding from Ϫ315 to Ϫ218, GE100, enhances the basal promoter (Ϫ217 to ϩ1) activity in SMCs and CEFs, while the upstream region from Ϫ3041 to Ϫ316 containing multiple E boxes is not directly involved in this event (Fig. 2). In vivo competition and gel shift assays suggest the presence of trans-acting factors bound to cis-element in GE100 (Fig. 3A). Detailed analyses indicate that the target element is a unique CArG box-like motif, located at Ϫ309 to Ϫ300 and that the CArG1 composition of this motif in addition to its 5Ј-and 3Ј-flanking sequences is essential for binding of trans-acting factors (Figs. 5 and 6). CArG boxes have been found in several actin genes as well as the c-fos gene (36,37). They interact with multiple nuclear protein factors and are required for skeletal or cardiac muscle-specific expression of the actin genes, for basal constitutive expression of the nonmuscle actin genes, and for rapid and transient activation of the c-fos gene in response to serum growth factors. The sequence of the CArG1 is specific because both the binding and transcriptional activities were decreased by deletion or mutation in CArG1 (Figs. 5 and 6). Therefore, the inner core of the CArG1 in the CaD gene, CCAAAAAAGG, is unique compared with other CArG boxes. The CArG1 was also able to activate the promoter activity in spite of its position and orientation. Based on these findings, we conclude that the CArG1 plays a role as a cell type-specific enhancer. The interaction between the CArG1 and nuclear protein factors was essential for activation of the gizzard-type promoter, while the amounts of the CArG1-protein complex were variable in differentiated and dedifferentiated SMCs and CEFs (Figs. 3, A and B, and 6). Therefore, the quantities of CArG1-binding protein factors would not be directly related to the promoter activity. These variations suggest the multiple interactions between the CArG1-protein complex and basal promoter units including CCAAT box, Sp1 site, and TATA box in respective cell types. Compared with the factors interacting with CArG boxes in the skeletal ␣-actin gene (38,39), the CArG1-binding factors were resistant against high salt concentrations. Our preliminary studies by UV cross-linking using the CArG1 or the c-fos serum response element as probes suggest that distinct proteins with different M r values bind to the each probe. 2 Based on our A, GP1CAT and GP1Db-21CAT were independently transfected into CEFs. Cells were cultured either serum-starved for 50 h (minus fetal calf serum) or serum-starved for 42 h and then restimulated in the growth medium for 8 h (plus fetal calf serum). Relative CAT activity was based on the activity of each CAT construct under conditions of serum starved. B, total RNAs (5 g) prepared from CEFs under serum-starved (Ϫ) or serum-stimulated after starvation (ϩ) were analyzed by Northern blotting using a oligonucleotide probe specific to the gizzard-type CaD. Ethidium bromide staining of the gels (at the bottom) is also shown. C, relative luciferase activities from the ␤-actin promoter under respective conditions of serum starved and stimulated are shown. results, we speculate that the CArG1-binding factors would be distinct from such CArG box-binding factors which have already been characterized. Further studies will be necessary to establish the CArG1-binding protein factors in the cell typespecific expression of the CaD gene.
The CArG1 fails to function as a serum-responsive element because the gizzard-type promoter was not affected by serum (Fig. 7A). This result coincided with the expression of endogenous CaD gene in CEFs (Fig. 7B) and high levels of promoter activity in differentiated SMCs cultured under serum-free condition (Figs. 2 and 5). Considering the serum responsiveness of vinculin and ␣and ␤-actin genes (23,35,38,40,41), serum inducibility might depend on cell type and might require another factor to mediate between the CArG box-binding factor and basal transcription initiation factors.
In summary, the present studies demonstrated that the gizzard-type CaD promoter exhibits high levels of transcriptional activity in SMCs and CEFs, but extremely low levels in other cell types such as C2C12 and HeLa cells, and that the promoter activity in differentiated SMCs is higher than that in dedifferentiated SMCs. The protein levels of CaD in differentiated and dedifferentiated SMCs were in good agreement with the promoter activities in the respective cells. These results suggest that the gizzard-type CaD promoter activity might be controlled under phenotypic modulation of SMCs. In addition, we have identified that the CArG1, located at Ϫ309 to Ϫ300 upstream of the transcriptional starting site of the gizzard-type CaD promoter is an essential cis-element for the SMC-specific expression, and that specific DNA-protein complex formation is found between the CArG1 and nuclear extracts from SMCs and CEFs. Further studies regarding SMC-specific gene expression are required for understanding the molecular events of phenotypic modulation of SMCs.
Addendum-During submission of this paper, promoter elements of the smooth muscle myosin heavy chain gene have been identified (43). E boxes, myocyte enhancer binding factor 2 (MEF2)-like motifs, and CArG box-like motifs are found in the myosin heavy chain gene and are involved in the SMC-specific expression. Based on their study, the protein binding to the MEF2-like motifs is revealed to be different from a MEF2 protein, while the CArG box-like motif does not show protein binding. In our present studies, a MEF2-site is absent in the gizzardtype CaD promoter, and E boxes are not important in the CaD expression, whereas only CArG1 is the essential cis-element for activation of the promoter.