Identification and characterization of the cell type-specific and developmentally regulated alpha7 integrin gene promoter.

Expression of alpha7 is mainly confined to skeletal and cardiac muscle in which it appears to be the major laminin-binding integrin. When myoblasts differentiate to myotubes, alpha7 mRNA and protein expression is up-regulated. To explore the mechanisms involved in the tissue-specific and developmentally regulated expression of alpha7, we isolated and characterized a genomic clone containing approximately 2.8 kilobase pairs (kb) of the 5'-flanking region of the murine alpha7 gene. The 5'-flanking region lacks both TATA and CCAAT boxes but contains five putative Sp1 binding sites located in a CpG island. Two transcription start sites, located near an initiator-like sequence, are 176 and 170 base pairs upstream of the translation start site. There are numerous binding sites for developmental and cell type-specific transcription factors, including AP-1, AP-2, GATA, and several AT-rich sites. There are also eight consensus E-boxes that bind the basic helix-loop-helix family of muscle-specific transcription factors. The approximately 2.8-kb 5'-flanking region was an active promoter in C2C12 skeletal myoblasts and exhibited increased expression upon conversion to myotubes but was inactive in HtLM2 cells, a mouse breast carcinoma epithelial cell line that does not express alpha7. Deletion analysis identified both positive and negative regulatory elements within the approximately 2.8-kb fragment. In 10T1/2 fibroblasts the approximately 2.8-kb alpha7 promoter was trans-activated by the myogenic basic helix-loop-helix proteins myogenin and MyoD but not by MRF4 and myf5. These results suggest that muscle-specific transcription factors play a role in regulating the cell-type expression of the alpha7 gene during development.

expression of the ␣7 integrin receptor is elevated in the nonmetastatic phenotype but is lacking in metastatic cells, suggesting that absence of this laminin adhesion receptor can play a role in melanoma tumor cell dissemination. 1 Expression of ␣7, as detected by reverse transcriptase-PCR, 2 has also been reported to be present in a number of tissues including for example stomach and uterus (7). However, it has not yet been determined if any of the ␣7 protein isoforms are expressed in these tissues. In myoblasts and cardiomyocytes, ␣7 appears to be the predominant laminin-binding integrin expressed (8,9). In vitro, basal levels of ␣7 expression are seen in replicating rat and mouse myoblasts (7,10,11). Expression of ␣7 is increased upon differentiation of myoblasts to myotubes. This up-regulation is paralleled by increased expression of myogenin (11), a member of the muscle-specific basic helixloop-helix (bHLH) transcription factor family, suggesting that these factors play a regulatory role in ␣7 expression during muscle development.
At embryonic day 7.5, ␣7 can be detected in the ectoplacental cone (differentiating murine trophoblast), where it is believed to play a role in trophoblast adhesion and/or differentiation. The use of laminin fragments suggested that ␣7␤1 may be one of the major trophoblast laminin-binding integrins involved in embryo implantation (12). Recently, the cDNAs for rat and mouse ␣7 have been isolated and shown to be alternatively spliced in both the extracellular (giving rise to two isoforms, X1 and X2) and cytoplasmic (producing three isoforms, A, B, and C) domains. Subsequent work has shown that these alternatively spliced forms are also developmentally regulated in muscle (7,10,13).
Taken together, these results indicate that expression of ␣7 plays a role in embryo implantation and muscle development. Furthermore, the developmental and restricted tissue expression (in skeletal muscle and cardiac muscle) of the ␣7 subunit suggests that the elements governing its expression are complex. As an approach to understanding the mechanisms regulating cell type-and differentiation-specific expression of the ␣7 subunit, we isolated and characterized the promoter region for the murine ␣7 integrin gene.

EXPERIMENTAL PROCEDURES
Isolation of a Mouse ␣7 Genomic Clone, Restriction Enzyme Analysis, and Sequencing-Approximately 1 ϫ 10 6 plaque-forming units of a mouse spleen genomic library constructed in a Fix II phage vector (Stratagene Cloning Systems, La Jolla, CA) were screened as described previously (10). An ϳ17-kb positive clone, G9, was identified and subcloned into pBluescript (Stratagene Cloning Systems). A partial restriction enzyme map was determined by standard techniques using single * This work was supported by Grants from the National Institutes of Health CA51884, DE10564, and DE10306 (to R. H. K.) and by National Cancer Institute Research Fellowship Award CA66272 (to B. L. Z.). 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  and double restriction enzyme digestion and Southern blot analysis (14). The probes used in the Southern analysis were labeled with [␣-32 P]dCTP using a Multiprime DNA labeling kit (Amersham Corp.). Hybridized nylon filters (Nytranϩ, Schleicher & Schuell) were stringently washed and exposed to x-ray film at Ϫ80°C as described previously (10). A subclone of G9 designated as G2 that contains the ␣7 promoter region was sequenced and analyzed further.
All nucleotide sequences were determined by the dideoxy chain termination method for double-stranded DNA as described previously (10). Primers for sequencing and polymerase chain reaction (PCR) were synthesized using the 391 DNA Synthesizer PCR-Mate from Applied Biosystems (Foster City, CA).
Plasmid Constructs-To create the ϳ2.8-kb ␣7 chloramphenicol acetyltransferase (CAT) construct, the ␣7 genomic clone G2 was digested with NcoI and blunt-ended with mung bean nuclease to remove the ATG start site (14). A second digest with SalI released the ϳ2.8-kb fragment. The XbaI site of pCAT-Basic (Promega, Madison, WI) was digested and filled in with Klenow, and pCAT-Basic was then digested with SalI and ligated to the blunt-ended/SalI ϳ2.8-kb ␣7 promoter fragment. The orientation of the insert was confirmed by sequencing. pCAT-Basic plasmids containing deletions of the ␣7 gene 5Ј-flanking region were created using PCR and the ϳ2.8-kb construct as the template for each reaction. A common 3Ј-oligonucleotide, promSalI, 5Јd-(A-AGGTCGACGGATCAACGCTCTC-CCAGCTAGTGC)-3Ј, corresponding to the sequence between positions ϩ151 and ϩ175 and containing a SalI restriction enzyme site at the 5Ј end (underlined) was used in the PCR reactions to create all deletion constructs. The oligonucleotides used as 5Ј PCR primers each contained a HindIII restriction enzyme site (underlined) and are as follows: All PCR reactions were performed in 1 ϫ Vent buffer (New England Bio-Labs, Beverly, MA) with 200 M of each deoxynucleotide, 40 pmol of each primer, and 2 units of Vent DNA polymerase (New England Bio-Labs) with proofreading activity. Typically, the thermocycling program consisted of an initial denaturation step at 94°C followed by 20 -25 cycles consisting of 1 min denaturation at 94°C, 30 s annealing at 58°C, and 30 s elongation at 72°C. A final 10-min elongation at 72°C followed. The PCR fragments were digested with HindIII and SalI and ligated into the corresponding sites in pCAT-Basic. Each construct was verified by sequencing.
Cell Culture-The mouse C2C12 myoblast cell line was provided by Dr. H. Blau (Stanford University). C2C12 cells were maintained in Dulbecco's minimal essential medium (DME-H-21, Cell Culture Facility, UCSF) supplemented with 20% fetal bovine serum. To induce their differentiation into myotubes, myoblasts were switched to differentiation medium (DM) containing DME-H21 with 2% horse serum for the indicated amount of time. Murine breast carcinoma cells, HtLM2 (provided by Dr. C. Damsky, UCSF), were grown in DME-H-16 with 10% fetal bovine serum. 10T1/2 fibroblasts were maintained in growth medium containing DME-H-16 supplemented with 10% fetal bovine serum. The mouse melanoma cell line K1735 cl19 was maintained as described previously (6).
Transfection Assays-Mouse C2C12 cells were maintained in DME-H-21 containing 20% fetal bovine serum. Cells were plated at 1.35 ϫ 10 5 cells per 35-mm dish and incubated for 18 -24 h. Two g of each chimeric CAT reporter gene construct and 0.5 g of a reference SV40luciferase plasmid (pGL2, Promega) were combined with 6 l of Lipo-fectAMINE (Life Technologies, Inc.) in 200 l of Opti-MEM (Life Technologies, Inc.). Cells were washed with 2 ml of Opti-MEM and overlaid with the LipofectAMINE-DNA complexes diluted to 1 ml with Opti-MEM. After 5 h, one-half of the cultures were switched to DME-H-21 with 20% fetal bovine serum. The remaining cultures were switched to DM to stimulate the formation of myotubes during growth for an additional 72 h. HtLM2 cells were transiently transfected in the same way as C2C12 myoblasts except that 10 l of LipofectAMINE was used. After 5 h the LipofectAMINE-DNA complex was replaced with growth medium, and the cells were incubated for 72 h.
CAT Assays-Cells were rinsed twice in phosphate-buffered saline and manually harvested using 1 ml of phosphate-buffered saline. Cell pellets were resuspended in 100 l of phosphate-buffered saline containing 1 mM phenylmethylsulfonyl fluoride. Fifty l of this cell suspension was subjected to three freeze-thaw cycles, spun at 500 ϫ g for 5 min to remove cell debris, and assayed for CAT activity. CAT activity was detected and quantitated by assaying 20 l of either the C2C12 or the HtLM2 cell lysate, using an enzyme-linked immunosorbent assay (ELISA). This colorimetric CAT-ELISA (5 Prime-3 Prime, Boulder, CO) detects the fully native, enzymatically active (trimeric) form as well as the denatured, enzymatically inactive (monomeric) form of the CAT protein. Each ELISA was performed according to the manufacturer's instructions, and CAT activity in cell lysates was determined from a standard curve run for each experiment.
To normalize the CAT activity between transfections, we assayed the luciferase activity transcribed by cotransfecting plasmid pGL2. The remaining 50 l of the cell suspension (see above) was lysed using 1 ϫ Reporter lysis buffer (Promega) and spun briefly to pellet cell debris. Three l of the cell extract was mixed with 100 l of luciferase assay reagent (Promega), and luciferase activity was measured using a Turner luminometer (Promega). The luciferase values were used to normalize CAT activity readings for transfection efficiency.
For trans-activation studies, 1.35 ϫ 10 5 10T1/2 fibroblasts plated in 35-mm dishes were transfected with LipofectAMINE-DNA complexes containing 2 g of the p2.8kb-CAT construct (full-length promoter) and 0.5 g of one of the expression vectors for muscle-specific bHLH proteins, MyoD (15), myogenin (16), MRF4 (17), and myf5 (18). The cDNA for each of the bHLH proteins was contained in the expression vector pEMSVcribe (15). Following transfection, 10T1/2 cells were grown for 72 h in DM prior to harvesting. S1 Nuclease Protection and Primer Extension Analysis-S1 nuclease protection assays were performed on total RNA isolated by the guanidinium isothiocyanate method as described (10). A single-stranded ϳ290-bp 5Ј genomic fragment of ␣7 extending from the initiation methionine to Ϫ115 was used as the probe for S1 analysis. This probe contained 175 bp of the untranslated region and 115 bp of the 5Јflanking sequences. The probe was end-labeled with T4 polynucleotide kinase, and ϳ3 ϫ 10 5 cpm of the probe was hybridized with 40 g of RNA from K1735 cells and from C2C12 myoblasts and myotubes for 16 h at 30°C in an S1 hybridization mixture containing 40 mM Pipes, pH 6.4, 400 mM NaCl, 1 mM EDTA, pH 8, and 80% deionized formamide. The hybridization mixture was digested with 1200 units/ml S1 nuclease (Boehringer Mannheim) in S1 nuclease buffer containing 0.28 M NaCl, 50 mM sodium acetate, pH 4.5, 4.5 mM ZnSO 4 , and 20 mg/ml salmon sperm DNA for 1 h at 30°C. The reaction products were ethanolprecipitated, and the protected fragments were analyzed on a denaturing polyacrylamide gel. A sequencing reaction primed with the same oligonucleotide (PE1) used to make the single-stranded S1 nuclease probe was run as size and site standard. DNA fragments labeled with 32 P or 35 S were run at the same position and as such should not affect length determinations.
Primer extension was performed using antisense oligonucleotide PE1 (GGATCAACGCTCTCCCAGCTAGTGC) (Fig. 2). PE1 was end-labeled with T4 polynucleotide kinase, and 10 5 -10 6 cpm was hybridized to 10 -15 g of total RNA for 16 h at 30°C in S1 hybridization buffer. The complementary DNA was extended in 1 ϫ PCR buffer (see below) containing 5 mM dithiothreitol, 100 g of actinomycin, 0.5 units of RNasin, 0.4 mM dNTPs, and Superscript II (Life Technologies, Inc.) for 2 h at 42°C (19). The reaction products were run on a denaturing polyacrylamide gel along with a sequencing reaction using the ϳ2.8-kb promoter clone as template primed with PE1 to determine length of the extended products.
RNA Isolation, Northern Blot-Total RNA was isolated by the guanidinium-isothiocyanate/phenol method (20). RNA samples (15 g) were electrophoresed in a 1.2% agarose gel containing formaldehyde. After electrophoresis, the RNA was transferred to nylon membranes (Nytranϩ, Schleicher & Schuell) by capillary blotting and fixed to the filter by exposure to UV light. The RNA was hybridized with a ϳ900-bp ␣7 cDNA fragment labeled with 32 P using a Multiprime DNA labeling kit (Amersham Corp.). To control for differences in RNA loading, blots were re-probed with a 300-bp 32 P-labeled mouse glyceraldehyde-3-phosphate dehydrogenase cDNA fragment. Hybridizations were carried out at 42°C in 50% formamide, 5 ϫ SSC, 5 ϫ Denhardt's, 0.1% SDS, and 300 g/ml salmon sperm DNA. Filters were washed twice in 1 ϫ SSC, 0.1% SDS at room temperature and once at 65°C in 0.1 ϫ SSC, 0.1% SDS. Filters were exposed to x-ray film at Ϫ80°C with intensifying screens.

RESULTS AND DISCUSSION
Isolation of a 5Ј ␣7 Genomic Clone-The 5Ј-flanking sequences of genes, which contain promoter and enhancer elements, regulate developmental and tissue-specific gene transcription. In order to study the elements governing the tissuespecific and developmental regulation of the ␣7 integrin subunit in more detail, we screened a mouse genomic library in Fix II for the 5Ј-untranslated and 5Ј-flanking region of the ␣7 gene. Using a cDNA fragment that contained sequences near the start of translation, we isolated one positive clone and plaque-purified it. A subclone (designated clone G2) that contained the 5Ј-untranslated region and 5Ј-flanking region of the ␣7 gene was examined further.
Mapping the Site of Transcription Initiation-Partial restriction enzyme mapping and Southern blotting indicated that clone G2 contains ϳ7 kb (Fig. 1) of sequence. To map the site of transcription initiation, we used a combination of primer extension and S1 nuclease analysis. S1 nuclease analysis was performed by hybridizing a 5Ј end-labeled 290-bp genomic fragment (extending from Ϫ115 to ϩ175) with total mRNA from mouse melanoma K1735 cells, C2C12 myoblasts, and C2C12 myotubes (Fig. 2). Two protected fragments were identified in all cell lines indicating that there are two major transcriptional start sites, located 170 and 176 bp, respectively, upstream of the translation initiation methionine (Figs. 2 and 4). This result also indicates that there is no change in the transcription start positions in myoblast versus myotubes. To confirm the S1 analysis result, we performed primer extension with antisense oligonucleotide PE1 (Figs. 3 and 4). Primer extension identified two major extension products whose sizes matched the two transcriptional initiation sites determined by S1 nuclease analysis. Previously, we reported the isolation and characterization of the mouse ␣7 cDNA (10). The start of the 5Ј-untranslated region, as determined by 5Ј rapid amplification of cDNA ends, of the full-length ␣7 cDNA terminated at the same position as the 170-bp fragment identified in S1 nuclease and primer extension analysis. Thus, by S1 analysis and primer extension, we have confirmed the presence of two primary transcription initiation sites at 176 and 170 bp upstream of the initiation methionine. Multiple transcriptional start sites are not uncommon for promoter regions lacking a TATA box.
Sequence analysis indicated that the genomic clone G2 contains ϳ2.8 kb of the 5Ј-flanking sequence, the first exon, all of intron 1, and exon 2. Designating the Ϫ176 initiation site as the start of transcription and performing sequence analysis indicated that exon 1 is comprised of 383 bp, of which 212 bp encodes the signal sequence and translation start site (Figs. 1  and 2). Exon 1 is followed by ϳ3.5 kb that makes up the first intron. Adjacent to this intron is an open reading frame of ϳ124 bp that encodes exon 2. Similar genomic organization is seen for other integrin genes (21,22). Each of the intron/exon boundaries analyzed (Fig. 2) conformed to the consensus 5Ј and 3Ј splice sequences, G(T/A)G (23).
The 5Ј-Flanking Region of the a7 Gene-The promoter regions for a number of the integrin receptor genes (␣II␤, ␣2, ␣4, ␣5, ␣6, 3 ␣v, ␤1, ␤2, ␤3, ␤7, CD11a, CD11b, CD11c and CD18) have recently been identified and characterized (24 -37). The FIG. 2. S1 nuclease protection analysis to identify the transcription initiation site. A single-stranded ϳ290-bp 5Ј genomic fragment of ␣7 genomic clone G2, extending from the initiation methionine to Ϫ115, was used as the probe for S1 analysis. This probe contained 175 bp of the untranslated region and 115 bp of the 5Ј-flanking sequences. The probe was end-labeled with T4 polynucleotide kinase and hybridized with 40 g of total RNA from K1735 cells (lane 3) and from C2C12 myoblasts (lane 1) and myotubes (lane 2). The RNA-DNA hybrid was subjected to S1 digestion, and the protected products were analyzed on a denaturing 6% polyacrylamide gel. Digested probe only was run in lane 4. The sequence (order of sequencing reactions is TCGA) of the two transcription start sites is denoted by ** and *. Oligonucleotide PE1 was used as the primer for sequencing analysis. human ␣4 promoter) (26), lacks both a TATA and a CCAAT box (Fig. 4). Transcription from promoters without TATA and CCAAT boxes can initiate at a consensus sequence defined as the initiator sequence (Inr) (38). The original consensus Inr sequences, as designated by the terminal deoxynucleotide transferase gene (38), closely resemble the initiator sequences found in the ␣II␤, ␣5, CD11a, and CD11b 5Ј-flanking regions (24,27,32,34). The sequence at the initiation site for ␣7, like that for the ␣2 integrin gene, diverges somewhat from the consensus Inr sequence (consensus, CTCANTCT; ␣7, CT-GGGTCC) ( Fig. 2; 25). In addition, like the promoters for ␣5 and ␣6, 3 the transcription start site for the ␣7 gene appears to start not in the Inr-like sequences but in close proximity to it (27). In fact, the start site at 170 bp 5Ј of the translational start site is only 2 bp upstream of the Inr-like sequence (Fig. 4). Typically, the Inr, in combination with nearby Sp1/GC boxes (38,39), directs correct transcription initiation. Located within 90 bp of the transcription start sites of the ␣7 gene are three tandem sequences for Sp1 (Fig. 4). Further upstream, at positions Ϫ165 and Ϫ194, are two additional Sp1 sites. All five Sp1 sites are located in the 5Ј-flanking region between ϩ1 and Ϫ238. This region is composed of 70% guanine (G) and cytosine (C) residues with three HpaII restriction enzyme sites (CCGG). Similar G/C rich regions, termed CpG islands, have been described for the ␣2, ␣5, and ␣6 3 promoters (25,27). When methylated, G/C islands can affect chromatin structure and thus influence the regulation and tissue specificity of gene expression (40,41).
In addition to the Inr sequence and the five putative Sp1 binding sites, the ϳ2.8-kb sequence of the ␣7 5Ј-flanking region contains numerous binding sites for ubiquitous, developmental, and cell type-specific transcription factors. Binding sites for ubiquitous transcription factors include three sites for AP-1 at Ϫ2100, Ϫ838, and Ϫ109 and four AP-2/AP-2-like sites at Ϫ955, Ϫ147, Ϫ127, and ϩ20 (Fig. 4). Transcription is mediated from AP-1 and AP-2 sites via the phorbol ester/diacylglycerol-activated protein kinase C pathway, whereas the cAMP pathway induces gene transcription from AP-2 (25,42). Whether these two pathways play a role in ␣7 transcription remains to be determined. AP-1 and AP-2 sites are present in a number of the integrin genes including ␣2, ␣4, ␣5, and CD11c and a number of muscle-specific genes such as muscle creatine kinase and desmin (25)(26)(27)36). In addition to sequence recognition sites for AP-1 and AP-2, two GATA sites are present, at position Ϫ327 (GATA-like) and Ϫ756 (consensus site) (Fig. 4). GATA sites bind the heart progenitor cell-specific transcription factor, GATA-4, which has recently been shown to be necessary for cardiac muscle cell development (43). This factor may be important in regulating expression of ␣7 in late fetal and adult heart. A second putative developmental binding site, for the ets transcription factor, is located at position Ϫ752 (Fig. 4). Recently, the ets DNA-binding domain was shown to be required for high level ␤-enolase promoter activity in myoblasts and to be important for repression of expression in fibroblasts (44). Further analysis of the 5Ј-flanking region of the ␣7 gene revealed a number of potential cell-specific and differentiationrelated regulatory sequence elements. There are eight consensus E-boxes (CANNTG), located at Ϫ2539, Ϫ2210, Ϫ2146, Ϫ2127, Ϫ2059, Ϫ2032, Ϫ1449, and Ϫ571 (Fig. 4). E-boxes are the DNA-binding sites for a family of muscle-specific bHLH proteins (45,46). The members of this family, comprised of MyoD, myf5, MRF4, and myogenin, are able to convert nonmuscle cells into myogenic lineages, can direct expression of many muscle-specific genes including desmin, troponin I, Mcreatine kinase, are expressed only in skeletal (not cardiac) muscle, and are believed to play a major role in muscle differentiation (47)(48)(49). Previously, it has been shown that the ␣7 mRNA level in rat L8 myoblast is increased 3-4-fold upon conversion of myoblasts to myotubes (11). This increase in ␣7 mRNA is paralleled by increases in the bHLH transcription factor myogenin. In addition, ␣7 expression is seen in the developing embryo during the time the MRFs, MyoD and myogenin are expressed (50 -52). These results suggest that the E-boxes in the promoter region of ␣7 may play an important role in ␣7 expression during muscle development and differentiation. Finally, it is interesting to note that the E-box at Ϫ2127 contains the symmetrical core consensus sequences, CAGCTG, that have been shown by in vitro studies to be the preferred binding sites for MyoD (53).
A/T-rich sites, termed the MADS box, CArG box, and the serum response element, have recently been identified in a number of muscle-specific genes (54 -56). Several factors that are important in embryogenesis and skeletal and heart muscle development (Mhox, Oct 1) and are required for transcriptional activity in skeletal and cardiac myocytes (MEFS, Mef2a-d, serum response factor) bind to these sequence elements (56,57). There are five A/T-rich sites in the ␣7 promoter (Fig. 4) that may function as binding sites for these factors and possibly regulate ␣7 expression in skeletal and/or cardiac tissue. Finally, at position Ϫ585 is an MCAT-like sequence. The MCAT motif is the binding site for transcriptional enhancer factor-1 (TEF-1) or a TEF-1-like protein (58,59) and is believed to be critical in directing cardiac-specific expression (60,61).
The 5Ј-Flanking Region of the ␣7 Gene Acts as a Promoter and Directs Cell-specific Transcription in Transfection Assays-To determine whether the 5Ј-flanking region of the mouse ␣7 gene could act as a promoter and direct transcription in a cell-specific and differentiation-specific manner, we ligated the ϳ2.8-kb genomic fragment upstream of the CAT gene in pCAT-Basic (see Fig. 6 and "Experimental Procedures"). The CAT-promoting activity of this fragment was determined by transiently transfecting the ϳ2.8-kb/CAT (p2.8kb-CAT) construct into HtLM2 mouse breast carcinoma cells and C2C12 myoblasts. Replicating C2C12 myoblasts express moderate levels of ␣7 mRNA (Fig. 5). In contrast, the HtLM2 mouse breast carcinoma cell line is of epithelial origin and does not express ␣7, as determined by Western blotting (not shown). The ␣7 gene 5Ј-flanking region directs CAT enzymatic activity in C2C12 myoblasts. When normalized for transfection efficiency, the ϳ2.8-kb/CAT construct generated ϳ14-fold greater activity than the CAT-structural sequences alone in C2C12 myoblasts (Fig. 6). In contrast, the same construct showed no activity in HtLM2 mouse breast carcinoma cells. Thus, the ϳ2.8-kb 5Јflanking sequence of the ␣7 gene can function as a promoter and demonstrates cell-specific activity.
To characterize the 5Ј-flanking region required for either myoblast-or myotube-specific promoter activity, 5Ј deletion mutants of p2.8kb-CAT (Ϫ2669 to ϩ175), p1.22kb-CAT (Ϫ982 to ϩ175), p600bp-CAT (Ϫ439 to ϩ175), p400bp-CAT (Ϫ223 to ϩ175), p300bp-CAT (Ϫ114 to ϩ175), and p200bp-CAT (Ϫ23 to ϩ175) were constructed (Fig. 6). We assayed the promoter deletion constructs in C2C12 myoblasts since they display basal levels of ␣7 expression as myoblasts and an increasing level of ␣7 mRNA when induced to form myotubes (Fig. 5). All deletion constructs directed CAT expression in both C2C12 myoblasts and myotubes. The p200bp-CAT construct, which contains only 23 bp 5Ј of the transcription start site, displayed the lowest promoter activity in both myoblast and myotubes. In myotubes this construct directed 2-fold higher enzymatic activity than in myoblasts. In contrast, the same construct was unable to direct CAT activity in HtLM2 mouse breast carcinoma cells. The addition of ϳ96 bp further 5Ј in the p300bp-CAT construct resulted in a ϳ14-fold increase in enzymatic activity in both myoblasts and myotubes. This construct directed the greatest CAT activity of all promoter deletion constructs. The p300bp-CAT construct displayed the highest activity in the HtLM2 mouse breast carcinoma cells as well. Comparison of the activity of p300bp-CAT in myoblasts and myotubes indicated that it was capable of directing ϳ2-fold higher CAT expression in myotubes than in myoblasts. A number of possible factors may be important for the increase in promoter activity displayed by this construct. For example, located within p300bp-CAT are three tandem Sp1/GC boxes, an AP-1 site, and an ets-like site (Fig. 2). These sites, either alone or in combination, may play a role in this increased activity.
The p400bp-CAT construct appears to contain a negative element, as indicated by the 3.5-and 12-fold reduction in CAT activity in myoblast and myotubes, respectively (Fig. 6). No consensus sequences for known negative regulatory elements or silencers are located in this region. Either an unidentified element is present in this construct or other regulatory elements located further upstream are required for efficient expression of these sequences. In HtLM2 cells, this construct also directed decreased enzymatic activity. In contrast, the addition of 200 bp 5Ј in p600bp-CAT only slightly elevated the CAT activity in myoblasts but increased the enzyme level 5-fold in C2C12 myotubes. Comparison of the CAT activity stimulated by this construct in myoblasts and myotubes showed only a modest increase in myotubes. In the mouse breast epithelial cell line, little to no activity was elicited by this construct.
Both the p1.22kb-CAT and p2.8kb-CAT deletion constructs were able to efficiently direct transcription in both myoblasts and myotubes (Fig. 6). The p1.22kb-CAT was able to promote ϳ13and ϳ10-fold enzymatic activity in myoblasts and myotubes, respectively, in comparison to p200bp-CAT. p1.22kb-CAT showed only slightly less activity than the highest-promoting deletion construct, p300bp-CAT. p1.22kb-CAT contains an E-box and an A/T-rich region, which may explain the higher CAT activity of this construct compared with p600bp-CAT. p2.8kb-CAT was able to induce ϳ10and ϳ7-fold higher CAT activity in myoblasts and myotubes, respectively, as compared with p200bp-CAT. The increased expression elicited by both constructs suggests that a positive regulatory element or an enhancer-like element may be located 5Ј of the p400bp-CAT construct. Neither construct was able to direct enzymatic activity in the murine breast carcinoma cell line, indicating that both constructs contain sequence elements required for tissuespecific expression. p1.22-CAT could direct nearly 2-fold more transcription in myotubes than in myoblasts, which may be due to the E-box. The p2.8kb-CAT that contains all E-boxes was only able to promote ϳ1.3-fold more activity in myotubes than myoblast. The 2-and 1.3-fold induction elicited by these constructs is of borderline relevance given the complexity of the  Fig. 2 ligated to the CAT structural gene. Constructs p1.22kb-CAT (Ϫ982 to ϩ175), p600bp-CAT (Ϫ439 to ϩ175), p400bp-CAT (Ϫ223 to ϩ175), p300bp-CAT (Ϫ114 to ϩ175), and p200bp-CAT (Ϫ23 to ϩ175) were deletion mutants derived from p2.8kb-CAT. B, promoter activity in C2C12 myoblasts. All constructs in A were transfected using Lipofectamine in parallel with pCAT-control (Promega), which contains the SV-40 promoter and enhancer, into the mouse myoblast C2C12 cell line. Cotransfection of pGL2, which contains the luciferase gene, was used to control for transfection efficiency. After 72 h in growth medium cell extracts were assayed using a colorimetric CAT-ELISA that detects the fully native, enzymatically active (trimeric) form as well as the denatured, enzymatically inactive (monomeric) form of the CAT protein. Relative CAT activity of the constructs, which is the average of three to five independent experiments, is plotted. Bars represent range of experimental values. C, promoter activity in C2C12 myotubes. All deletion constructs were transfected as described in B. Cultures were switched to DM in order to stimulate the formation of myotubes. After 72 h of incubation, cell extracts were isolated and assayed as in B. D, promoter activity in the mouse epithelial cell line HtLM2. All constructs in A were transfected into the murine breast carcinoma cell line HtLM2 and after 72 h cell extracts were assayed as described in B. The results presented are the average of two independent transfections. CAT reporter assay system, suggesting other factors may account for the large increase in ␣7 mRNA during myotube differentiation. It is possible that this increase may be a due to a combination of factors, for example elevated promoter activity and mRNA stability. Taken together, these results indicate that the 5Ј-flanking region of the ␣7 gene functions as a promoter and shows the expected tissue-specific and developmentally regulated activity. Studies are ongoing to determine which sequence elements and post-transcriptional events are involved in this regulation.
As a start to determine the elements involved in the regulation of ␣7 in myoblasts and myotubes, we undertook transactivation assays (see "Experimental Procedures"). trans-Activation assays have been routinely used to determine if the family of muscle-specific bHLH factors can activate transcription from muscle-specific gene promoters (47,(62)(63)(64)(65). Vectors expressing MyoD, myogenin, MRF4, and myf5 were transiently cotransfected into 10T1/2 fibroblast cells with the ␣7 promoter-CAT fusion construct, p2.8kb-CAT. As a negative control the p2.8kb-CAT construct was cotransfected with pGEM-11fZ(ϩ). A number of reports have previously demonstrated that bHLH transcription factors can differentially trans-activate musclespecific genes. For example, the ␥ subunit gene of the acetylcholine receptor is efficiently trans-activated by myogenin and myf5 whereas MRF4 preferentially trans-activates the ⑀ subunit gene promoter (62,66). Only myogenin and MyoD were able to efficiently trans-activate the ␣7 promoter-CAT construct (Fig. 7). Myogenin trans-activated the promoter by ϳ2fold whereas MyoD was able to trans-activate by nearly 4-fold, indicating that both of these factors may play a role in ␣7 gene expression during muscle development. Co-transfection of MyoD, myogenin, and p2.8kb-CAT at the same time did not show any additive or synergistic effects (data not shown). It is possible that other factors, like the ubiquitously expressed E-proteins, form heterodimers with these factors during transactivation of the ␣7 promoter. p2.8kb-CAT alone produced very little activity in 10T1/2 cells. These results are supported by recent work with myoblasts showing that ␣7 expression levels increase during myotube formation in a pattern that is similar to increases in myogenin expression (11). In addition, when MyoD was transfected into 10T1/2 cells, it significantly increased the level of ␣7 mRNA expression. In contrast, no ␣7 expression was seen in the untransfected 10T1/2 cells (Fig. 8). Taken together, these results demonstrate that the bHLH factor MyoD (and myogenin) activates ␣7 gene transcription and that this activation appears to be mediated by the interaction of MyoD (and myogenin) with the ␣7 gene promoter. Finally, it has recently been shown that transfection of MRF4 into 10T1/2 cells caused a slight induction in ␣7 mRNA expression (11). This is in contrast to our trans-activation assays that indicated that MRF4 could not directly activate the ␣7 promoter. This apparent discrepancy between the trans-activation of the native gene as compared with the reporter gene may be the result of a single trans-activating factor generating a new or novel transcription start site. It is also possible that MRF4 may activate ␣7 expression through an indirect pathway (67).
In summary, we have identified the murine ␣7 gene promoter and its activity in C2C12 myoblasts and myotubes. Characterization of the ␣7 5Ј-flanking region of the murine gene revealed that the ␣7 integrin subunit's expression is modulated by a number of positive and negative regulatory elements that confer cell type-specific and development-specific expression. We have located regions that contain possible enhancer and silencer elements and have identified potential transcription factors that may regulate ␣7 expression during muscle development. Further work will be required to define how these elements as well as the regulatory sequences and binding factors are important in controlling the tissue-specific and developmentally regulated expression of the ␣7 integrin subunit. FIG. 7. trans-Activation of the ␣7 promoter by bHLH transcription factors. Expression plasmids (pEMSV) encoding myogenin (myoG), myf5, MyoD, or MRF4 were cotransfected into 10T1/2 cells along with p2.8kb-CAT. The data are expressed as fold activation of the cotransfection of the muscle-specific transcription factors over the p2.8kb-CAT cotransfected with pGEM-11Zf(ϩ) alone. CAT activity in cell extracts was determined by a CAT-ELISA assay. The results presented are the average of two independent co-transfections.