Identification of a Contractile-responsive Element in the Cardiac (cid:97) -Myosin Heavy Chain Gene*

The mechanisms by which the cardiac-specific (cid:97) -my-osin heavy chain ( (cid:97) -MHC) gene responds to contractile activity was studied in cultured cardiomyocytes and in vivo . Deletion analysis of the (cid:97) -MHC promoter transiently transfected into neonatal rat cardiomyocytes lo- calized the contractile-responsive element within (cid:50) 80 to (cid:50) 40 base pairs of the transcriptional start site. Mutational analysis of an E-box motif at position (cid:50) 47 showed that it was necessary for the contractile response both in cultured cardiomyocytes and in the intact heart. Competition gel mobility shift experiments indicated that the protein-DNA complex formed within the (cid:50) 39 to (cid:50) 59 base pair region could be competed by the E-box element at (cid:50) 309 of the (cid:97) -MHC gene and that base substitutions within an E-box motif at (cid:50) 47 eliminated the protein-DNA complex. To identify the contractile-re- sponsive nuclear protein, antibodies specific for E12/ E47, an E-box binding basic-helix-loop-helix (bHLH) pro- tein, and antibodies recognizing upstream stimulatory factor (USF), a widely expressed bHLH-leucine zipper transcription factor, were studied for their ability to inhibit cardiomyocyte nuclear protein binding to the E-box motif at (cid:50) 47. Anti-USF antibody abolished formation of the protein-DNA complex, thus identifying the protein as antigenically related to USF and demonstrat-ing ligated into a promoterless firefly luciferase expression plasmid (pLUC) The sequence identity of all verified by DNA sequence analysis. A second series of promoter constructs generated that terminated at the downstream position (cid:49) 32 and extended upstream to 388, 163, (cid:50) 40. These sequences generated using polymer- ase chain reaction to amplify the regions from the plasmid pSVOMCAT. primers Bgl II site and antisense primers Nco I site to facilitate subcloning into the multiple cloning region of pLUC.Two (cid:97) -MHC constructs tire

The mechanisms by which the cardiac-specific ␣-myosin heavy chain (␣-MHC) gene responds to contractile activity was studied in cultured cardiomyocytes and in vivo. Deletion analysis of the ␣-MHC promoter transiently transfected into neonatal rat cardiomyocytes localized the contractile-responsive element within ؊80 to ؊40 base pairs of the transcriptional start site. Mutational analysis of an E-box motif at position ؊47 showed that it was necessary for the contractile response both in cultured cardiomyocytes and in the intact heart. Competition gel mobility shift experiments indicated that the protein-DNA complex formed within the ؊39 to ؊59 base pair region could be competed by the E-box element at ؊309 of the ␣-MHC gene and that base substitutions within an E-box motif at ؊47 eliminated the protein-DNA complex. To identify the contractile-responsive nuclear protein, antibodies specific for E12/ E47, an E-box binding basic-helix-loop-helix (bHLH) protein, and antibodies recognizing upstream stimulatory factor (USF), a widely expressed bHLH-leucine zipper transcription factor, were studied for their ability to inhibit cardiomyocyte nuclear protein binding to the E-box motif at ؊47. Anti-USF antibody abolished formation of the protein-DNA complex, thus identifying the protein as antigenically related to USF and demonstrating that bHLH-leucine zipper proteins are involved in the contractile-induced expression of the cardiac ␣-MHC gene.
The cardiac myocyte responds directly to mechanical stimuli such as load, stretch, or contractility by changes in cell mass and by alterations in specific gene expression resulting in changes in contractile function (1)(2)(3)(4)(5). In the pressure-overloaded rat heart, expression of the myosin heavy chain (MHC) 1 genes are altered such that the ␤-MHC isoform appears de novo while the ␣-MHC isoform is decreased (6,7). Hemodynamic unloading of the heart as occurs in the heterotopically transplanted isograft results in a decrease in ␣-MHC expression which is mediated by a decrease in promoter activity of this gene (3,8). The mechanism by which the hemodynamic or contractile stimulus is transduced to the nucleus remains unclear. Stretch-induced alterations in plasma membrane-associated ion channels, phospholipases, G proteins, and their associated cytoplasmic second messengers, including cAMP, inositol phosphates, calcium, and diacylglycerol leading to the induction of protein kinase cascades are potential signaling pathways (9 -11). Alternatively, stretch-activated release of various autocrine factors may promote immediate-early gene expression and/or growth factor gene induction which in turn may induce cardiac-specific gene transcription (9,11,12).
Based on consensus sequence, mutational analysis and nuclear protein binding activity, the identities of several regulatory elements have been delineated in the ␣-MHC gene. Elements sufficient for both high levels of expression and cardiac myocyte-restricted expression have been localized to the proximal 5Ј-flanking region of the gene from Ϫ380 to Ϫ40 bp of the transcriptional start site. These DNA elements include a site that binds the myocyte-specific enhancer-binding factor-2 at position Ϫ327/Ϫ337 (13,14), an E-box sequence at position Ϫ308/Ϫ313 (15), and M-CAT and A-rich motifs at positions Ϫ236/Ϫ242 and Ϫ217/Ϫ223, respectively (16). Recently, two sites located at Ϫ258/Ϫ269 that interact with the transcription factor, GATA-4, have been shown to be necessary and sufficient to impart cardiac myocyte-restricted expression of the ␣-MHC gene (17). Responsiveness of the ␣-MHC promoter to activators of the cyclic-AMP/protein kinase A signaling pathway has been mapped to an M-CAT/E-box hybrid element (18).
We have shown that the contractile activity per se of cultured neonatal rat cardiac myocytes modulates the expression of the MHC genes (1,5,19). In the present study we used deletion analysis of the 5Ј-flanking region of the ␣-MHC gene to identify DNA elements that determine responsiveness of the promoter to contractile activity, and have identified putative transcription factors that mediate this response.
For the preparation of nuclear extracts, cardiomyocytes were plated onto collagen-coated 75-cm 2 tissue culture flasks and maintained in DMEM/F-12 medium containing ITS plus T3 for 48 h prior to harvest.
Transfection Studies-Plasmid DNA was introduced into the cultured cells 18 h after plating onto six-well dishes and immediately following removal of the nonadherent cells. Plasmids containing 5Јflanking regions of the ␣-MHC gene were co-transfected with a constitutively active RSV (Rous sarcoma virus long terminal repeat) ␤-galac-tosidase plasmid (pRSVZ, ATCC, Rockville, MD) using the lipofection method of transfer. Cells in each 35-mm well were exposed for 6 h to DMEM/F-12:PC-1 (2:1) medium containing Lipofectin reagent (Life Technologies, Inc.), 2.5 g of ␣-MHC promoter/luciferase reporter plasmid, and 0.25 g of pRSVZ. After lipofection, the cells were washed twice in Hanks' buffered salt solution and maintained in DMEM/F-12 supplemented with ITS plus other reagents as indicated with daily medium changes. Luciferase and ␤-galactosidase were analyzed after 48 -72 h by lysis of the cells in 300 l of lysis buffer (Promega, Madison, WI)/well. Luciferase activity was determined by the addition of 10 l of lysate to 100 l of luciferin reagent and measured as light production using a Turner Designs model 20 luminometer. ␤-Galactosidase activity was measured in 50 -150 l of lysate and compared with a standard curve of purified ␤-galactosidase (Promega) from 0.1 to 0.5 milliunits. Luciferase activity is expressed as a function of ␤-galactosidase activity in the same volume of cell lysate (luciferase luminescence units/␤galactosidase). To determine efficiency of transfection, myocyte cultures were stained for ␤-galactosidase activity 48 h after lipofection with pRSVZ. Cells were fixed in solution containing 2% paraformaldehyde, 5 mM EGTA, 2 mM MgCl 2 , 0.1 M Pipes, pH 7.3, and stained by standard methods.
Plasmid Constructs-The ␣-MHC promoter constructs were generated from plasmid pSVOMCAT (20) generously provided by Dr. B. E. Markham (Ann Arbor, MI). The 5Ј-flanking region of the ␣-MHC gene was restricted using NheI, BglII, EcoRI, and BglI to generate fragments terminating at positions Ϫ2560, Ϫ1660, Ϫ612, and Ϫ195, respectively, and at the same HindIII site at ϩ421 of the transcriptional start site. Fragments were ligated into a promoterless firefly luciferase expression plasmid (pLUC) that has been described previously (8). The sequence identity of all clones was verified by DNA sequence analysis. A second series of promoter constructs was generated that terminated at the downstream position ϩ32 and extended upstream to positions Ϫ388, Ϫ163, Ϫ80, and Ϫ40. These sequences were generated using polymerase chain reaction to amplify the regions from the plasmid pSVOMCAT. Sense primers contained a BglII site and antisense primers contained a NcoI site to facilitate subcloning into the multiple cloning region of pLUC.
Two ␣-MHC promoter constructs were made that contained the entire genomic region between transcriptional and translational start sites and included the sequences from Ϫ2560 to ϩ1036 and from Ϫ1660 and ϩ1036. Polymerase chain reaction methodology was used to amplify the region from ϩ421 to ϩ1036 from rat genomic DNA (Clontech, Palo Alto, CA). The sense primer contained sequences at ϩ421 and antisense primer contained the first ATG of the ␣-MHC coding sequence. This fragment was ligated to previously constructed plasmids containing sequences from Ϫ2560 to ϩ421 and Ϫ1660 to ϩ421. Proper orientation and ligation were ascertained by automated DNA sequence analysis (Applied Biosystems 373A sequencer). All plasmids were purified by ion exchange chromatography (QIAGEN, Inc., Chatsworth, CA) and examined by agarose gel electrophoresis with ethidium bromide staining.
Mutagenesis of the ␣-MHC Promoter-The Ϫ388␦HME mutant plasmid contains a C to A transversion at position Ϫ50 within the context of the Ϫ388 to ϩ32-bp region of the gene. An antisense primer containing the base substitution at position Ϫ50 and the Eco47III site at Ϫ21 and a sense primer containing sequences from Ϫ388 to Ϫ368 and a BglII site were synthesized and used for polymerase chain reaction amplification of the Ϫ388 to Ϫ21 sequence. The wild type Ϫ388 to Ϫ21 sequence in the Ϫ388/ϩ32 luciferase expression plasmid was excised using BglII and Eco47III and replaced by the amplified fragment containing the mutation to generate the Ϫ388␦HME plasmid. The sequence was verified by automated DNA sequence analysis.
Electrophoretic Gel Shift Assay-The sense strands of each deoxyoligonucleotide used for EMSA is shown in Table I. Both sense and antisense oligonucloetides contained an AGCT sequence at their 5Ј ends and were synthesized by Appligene, Inc. (Pleasanton, CA). The sense strands were end-labeled using T4 polynucleotide kinase and [␥-32 P]ATP (3000 Ci/mmol) (DuPont NEN) to a specific radioactivity of 1.5-2 ϫ 10 6 dpm/pmol, separated from free label over Sephadex G-25 spin columns, and annealed to a 50-fold molar excess of nonradiolabeled complementary oligonucleotide. Reactions (20 l) contained 5-10 g of nuclear extract protein, 1 g of poly(dI-dC), in 12 mM Hepes-NaOH pH 7.9, 60 mM KCl, 4 mM Tris-HCl, 0.6 mM EDTA, 0.6 mM dithiothreitol, 12% glycerol, 0.6 mM phenylmethylsulfonyl fluoride. Reactions were preincubated at room temperature for 15 min prior to incubation with 20 -40 fmol of labeled DNA (40,000 dpm) for 30 min. In competition experiments, competitor oligonucleotides were added at 100-fold excess of the labeled DNA. In gel shift assays with antibodies to E12/E47 and USF (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), the binding reaction was completed prior to incubation with the antibody for 60 min on ice. The complexed products were resolved on 6% native polyacrylamide gels which were subsequently dried and exposed to x-ray film.
Nuclear Extract Preparation-Cardiomyocytes from approximately 15 neonatal rat hearts plated on two 75-cm 2 flasks (T75) were lysed and pooled for a single sample of nuclear extract. Myocytes (Ͼ90% of total cells) were cultured in DMEM/F-12 plus ITS and L-triiodothyronine (10 Ϫ7 M) for 48 h prior to harvest. Typical yields from two T75 flasks were 100 g of nuclear protein. Myocytes were Dounce homogenized in ice-cold buffer containing 0.3 M sucrose, 10 mM Hepes, pH 7.9, 10 mM KCl, 1.5 mM MgCl 2 , 0.1 mM EGTA, 0.5 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and 50 g/ml each of antipain, leupeptin, benzamidine, and aprotinin. Nonidet P-40 was added to 0.4% final concentration. Nuclei were recovered by centrifugation at 3000 ϫ g for 10 min at 4°C and washed once with homogenization buffer without detergent. Nuclei were extracted as described by Sierra et al. (21) except that protease inhibitors were added at 50 g/ml. Protein concentration in the final extract was determined by Lowry assay, aliquoted, and stored at Ϫ80°C for single use.
Direct Gene Transfer-The procedure for directly injecting plasmid DNA into the intact heart has been described previously (8,22). Five g of either wild type Ϫ388/ϩ32 ␣-MHC promoter construct (Ϫ388wt) or E-box Ϫ47 mutant construct (Ϫ388␦HME) were co-injected with 0.5 g of pRSVCAT in 20 l of saline into the ventricular myocardium. The animals were sacrificed after 3 days, and the ventricular tissue that included the area of injection (150 -200 mg) was homogenized in low salt buffer. The supernatant resulting from centrifugation at 9500 ϫ g was used for both chloramphenicol acetyltransferase (CAT) and luciferase (LUC) analyses as described previously (8,22). ␣-MHC promoter activity is expressed as luminescence units normalized to percent CAT conversion (LUC/CAT) in the same volume of homogenate.

DNA Transfection into Cultured Myocytes-
The efficiency of DNA transfer by lipofection into the neonatal cardiomyocytes was assessed using ␤-galactosidase reporter plasmid (pRSVZ) and by staining the cells 48 h after transfection. From 16 separate experiments, the transfection efficiency was estimated at 6 Ϯ 1% of the myocytes.
Identification of a Contractile-responsive Region by Deletion Analysis of the ␣-MHC Promoter-To identify the region of the ␣-MHC gene that is both necessary and sufficient for contractile responsiveness, the activities of deletion mutants of the 5Ј-flanking sequences were determined by transient transfection of neonatal cardiac myocytes. Fig. 1A compares the activities of promoter constructs terminating at positions Ϫ2560, Ϫ1660, Ϫ612, and Ϫ195 with a common downstream site at ϩ421 in spontaneously contracting cardiomyocytes and in noncontracting verapamil-treated cultures. All promoter deletions were contractile-responsive with 2.5-4-fold greater activity in contracting myocytes compared with noncontracting cells.
To determine contractile responsiveness of the proximal region of the promoter, a series of promoter constructs were made that terminated at position ϩ32 and extended upstream to positions Ϫ388, Ϫ163, Ϫ80, and Ϫ40. Fig. 1B shows that con-

TRE135
Ϫ159 GCTGTCCTCCTGTCACCTCC 140 structs containing sequences upstream of Ϫ80 bp of the transcriptional start site were contractile-responsive, while sequences downstream of position Ϫ40 were not responsive. These data suggest that the DNA region between Ϫ40 and Ϫ80 bp is necessary to mediate the contractile response of the ␣-MHC promoter.
To determine whether contractile-responsive elements exist downstream of the transcriptional start site, two constructs were made that contained the entire genomic region between the transcriptional and translational start sites and extending to Ϫ2560 nd Ϫ1660 bp upstream of the transcriptional start site. Transient transfection analysis of these two constructs was compared with promoter constructs containing the same sequences upstream but terminating at position ϩ421 within the first intron. Activities of all promoter constructs were induced 2.5-3-fold in response to contractile activity (data not shown). These results indicate the absence of contractile-responsive enhancers downstream of the transcriptional start site.
Effects of Contractile-arresting Agents on ␣-MHC Promoter Activity-The spontaneous contractions characteristic of the neonatal cardiac myocytes plated at high density can be prevented by culturing in medium containing 10 M verapamil, an inhibitor of calcium influx across the sarcolemma (23). Studies with the fluorescent indicator, Fura-2, indicated that verapamil treatment of cardiomyocytes eliminated phasic levels of intracellular [Ca 2ϩ ] but maintained diastolic [Ca 2ϩ ] i at approximately 200 nM. 2 An alternative agent that reduces contractile activity in cardiomyocytes is BDM (24). At 5 mM concentration, BDM prevented tension development in myocytes by inhibiting actomyosin cross-bridge formation while having no effect on intracellular Ca 2ϩ transients. 2 ␣-MHC promoter activities were measured by transient transfection of cardiomyocyte cultures treated either with 10 M verapamil or 5 mM BDM and compared with untreated spontaneously contracting cultures. Activities of three promoter constructs containing sequences from positions Ϫ2560 to ϩ421(p2560), Ϫ612 to ϩ421 (p612), and Ϫ388 to ϩ32 (p388) are expressed as a percent of their activity in contracting cultures (Table II). Verapamil and BDM had similar effects on ␣-MHC promoter activity, significantly decreasing activity by 2.5-4.5-fold compared with spontaneously contracting cultures (p Ͻ 0.01).
To ascertain that culture conditions used in these experiments did not alter the expression of the control plasmid, ␤-galactosidase activity was measured in both contracting and noncontracting myocytes in defined medium in the absence or presence of the agents used in this study. ␤-Galactosidase activity measured per mg of myocyte protein was not significantly altered by myocyte contractile activity. Activity of the promoterless luciferase reporter plasmid, pLUC, was measured in every culture condition. Its activity was significantly lower than any of the ␣-MHC promoter constructs studied and was not altered by contractile activity or by addition of any agent studied.
Identification of a Contractile-responsive Cis-regulatory Element-Sequence analysis of the Ϫ80 to Ϫ40-bp region revealed a consensus E-box motif at position Ϫ74/Ϫ80, a CArG element at Ϫ60/Ϫ70 (18), an E-box at Ϫ47/Ϫ52 and an M-CAT site on the opposing DNA strand at position Ϫ42/Ϫ48 (18). Table I shows the contractile-responsive region from Ϫ83 to Ϫ40 to indicate the arrangement of these DNA elements. Also listed are the oligonucleotide sequences used for electrophoretic mobility shift analysis to determine cardiomyocyte nuclear protein binding to these cis-acting elements. The HME oligomer contains the E-box and overlapping M-CAT elements at Ϫ42/Ϫ52;  E74 oligomer contains the E-box motif at Ϫ74. Probe E318 contains the E-box at Ϫ309/Ϫ314 (15); probe MCT241 contains the M-CAT element at position Ϫ237/Ϫ243 (16), and the TRE135 oligomer contains the thyroid hormone-responsive element located at position Ϫ142/Ϫ157 (25) of the ␣-MHC promoter. Fig. 2 shows EMSA using contracting myocyte nuclear extracts with the ds oligonucleotide probes listed in Table I. The HME probe that contains both consensus E-box and M-CAT elements shows one prominent retarded band (B1) and a minor broad band (B2). The single protein-DNA complex formed with the E74 probe containing a consensus E-box motif corresponds to the higher molecular weight band, B1, seen with the HME probe. Binding to the E-box sequence at position Ϫ309 (probe E318) retards a band corresponding to B1 and a second band in the region of B2. Two protein-DNA complexes are observed binding to the M-CAT element at Ϫ237 of ␣-MHC (probe MCT241) as has been described previously (26,27).
Competition experiments were used to identify the nuclear proteins bound to the HME and E74 DNA probes. Fifty-and 100-fold molar excess unlabeled competitor DNA sequences were used to determine specificity of competition. A ds oligonucleotide containing the ␣-MHC thyroid hormone response element (TRE135) was used as nonspecific competitor DNA. Fig. 3 (lane 1) shows the retarded bands bound to the ds HME oligonucleotide probe. Competition with 100-fold excess of the E318 ds oligonucleotide probe (lane 2) completely eliminated the B1 complex. To determine whether nuclear protein bound to the consensus M-CAT sequence within the HME probe, 100-fold excess of a known M-CAT sequence (MCT241) was added to the binding reaction. MCT241 did not compete for binding to either the prominent B1 protein-DNA complex or the minor B2 complex (lane 3). The TRE135 sequence did not compete for any of the HME-protein complexes (lane 4). Since the HME B1 complex is retarded similarly to the complex retarded by the E-box at Ϫ309 (E318) and is competed by the E318 ds oligonucleotide but not by the M-CAT sequence (MCT241), the proteins contained within the B1 complex appear to be E-box-and not M-CAT-binding proteins.
The protein complex bound to the E-box motif at Ϫ74 was retarded similarly to the B1 protein-DNA complex of the HME probe (Fig. 3, lane 5) and was competed by 100-fold excess of the HME oligomer (lane 6). Similarly, 100-fold excess of the E74 oligomer effectively competed B1 binding to the labeled HME probe (lanes 7 and 8). The relative quantity of B1 complex bound to E-box Ϫ74 was consistently less than the protein-DNA complex bound to E-box at position Ϫ47 (lane 5 versus 7), suggesting differences in binding affinities, protein oligomerization, or that the E-box motifs bind distinct nuclear proteins.
Mutational Analysis of the HME Sequence-To determine which nucleotides within the HME sequence were essential for protein-DNA complex formation, four separate transversion mutations were introduced into the consensus E-box and M-CAT elements (Fig. 4). Gel shifts using ds oligonucleotides containing these base substitutions are shown in Fig. 4. Lane 1 shows the standard gel shift analysis using the wild type HME sequence. Lane 2, which corresponds to mutation (2), shows that substitution of the conserved T in the E-box motif did not inhibit formation of the B1 protein-DNA complex, whereas the C to A transversion at position Ϫ50 (mutation 3) abolished B1 binding. The G to T base substitution at position Ϫ53 (mutation 4) which is outside the consensus E-box sequence CANNTG had no effect on B1 complex formation. To determine whether nuclear protein binding to the M-CAT consensus sequence at Ϫ42/Ϫ48 was involved in B1 complex formation, a ds oligomer was synthesized containing three base substitutions at positions Ϫ43 to Ϫ45 (mutation 5). Gel shift analysis showed that the mutant M-CAT oligomer did not alter B1 protein-DNA complex formation (lane 5), suggesting that binding to this M-CAT sequence did not occur in these cardiomyocyte nuclear extracts.
E-box-47 Is Necessary for Contractile Responsiveness of the ␣-MHC Promoter-To determine if the E-box element at position Ϫ47/Ϫ52 was necessary for the contractile-mediated activation of the ␣-MHC promoter, the T to C transversion at Ϫ50 that abolished B1 complex formation (Fig. 4) was introduced into the 5Ј-flanking region of the gene for analysis in cultured cardiomyocytes and in the intact animal. The base substitution at position Ϫ50 was introduced into the context of a larger promoter region that included Ϫ388 to ϩ32 bp of the transcriptional start site, generating the ␣-MHC promoter/luciferase reporter plasmid, Ϫ388␦HME. Transient transfection analysis of the mutant construct showed that the E-box element was necessary for the contractile-mediated activation of the promoter in neonatal rat cardiomyocytes. Activity of the mutant  Table I. End-labeled probes are indicated above each lane. B1 and B2 indicate the location of slower and faster mobility protein-DNA complexes for the HME probe, respectively.
FIG. 3. Identification of nuclear factors that bind to the HME and E74 sequences of the ␣-MHC gene. Competition gel shift experiments, described under "Experimental Procedures," were performed to determine the identity of the protein complexes formed with the HME probe (lanes 1-4) and to determine whether E74 and HME bound similar nuclear proteins (lanes 5-8). End-labeled ds HME or E74 oligomer probes were incubated with cardiomyocyte nuclear extract (5 g of protein), without and with 100-fold molar excess of competitor probes as indicated above each lane. Oligomer sequences are listed in Table I. B1 and B2 are protein-DNA complexes retarded by the HME probe.
Ϫ388␦HME luciferase reporter plasmid in contracting cardiomyocytes was 37 Ϯ 4% of the activity of the wild type promoter (Ϫ388wt) in contracting myocytes and the level of activity was the same as the wild type promoter in noncontracting myocyte cultures (48 Ϯ 7%) ( Table III). Activities of the wild type and mutant promoter constructs in noncontracting myocytes were the same (48 Ϯ 3 versus 33 Ϯ 7%), suggesting that the E-box element is necessary for the contractile-mediated activation of the promoter and not the basal activity of the promoter.
To determine whether the E-box element at Ϫ47 is necessary for promoter activity in the contracting myocardium in vivo, the same wild type (Ϫ388wt) and mutant (Ϫ388␦HME) promoter/luciferase reporter plasmids used in cultured myocytes were injected directly into the ventricular myocardium of normal animals. Three days after DNA injection, luciferase analysis of the injected ventricular tissue showed that the activity of the mutant promoter was 38 Ϯ 6% of the wild type promoter (p Ͻ 0.01) (Table III). These data in cell culture and in the intact working heart suggest that the E-box element at position Ϫ47 is required to enhance ␣-MHC promoter activity in the contracting cardiac myocyte. HME-binding Protein Is Antigenically Related to USF-Several distinct families of nuclear proteins bind to DNA sequences containing the core CANNTG element or E-box motif, including those comprising the basic-helix-loop-helix (bHLH) proteins such as the myogenic factors and E12/E47 (28,29) and the bHLH-leucine zipper (bHLHZ) proteins such as upstream stimulatory factor (USF) (30). The E12/E47 nuclear proteins are widely expressed heterodimerization partners of cell-spe-cific bHLH proteins such as MyoD (28), allowing for genespecific regulation. Similarly, USF is a ubiquitously expressed protein but has been shown to be important in cell-specific gene regulation (31,32). To identify the proteins which form the Ϫ47 E-box B1 complex, antibodies produced against E12/E47 and USF were used in gel shift experiments (Santa Cruz Biotechnology).
The protein-DNA complex B1 that forms in the presence of cardiomyocyte nuclear extracts and the HME probe is shown in Fig. 5, lane 1. Formation of the B1 complex was not altered by including 2 g of rabbit IgG (lane 2) or 1 g E12/E47 antibody (lane 3) in the binding reaction. However, USF antibody blocked the formation of the B1 complex (lane 4). This blocking ability of the USF antibody has been observed previously with other E-box sequences (31)(32)(33). To ascertain that the blocking effect of the USF antibody was specific to the B1 complex, myocyte nuclear extracts were incubated with the MCT241 probe containing the ␣-MHC M-CAT element at position Ϫ237 (lane 5). Addition of USF antibodies to the DNA/nuclear extract binding reaction did not affect protein binding to the M-CAT motif (lane 6). These data indicate that the protein(s) that form the B1 complex is USF or is antigenically related to USF and that it is distinct from the E12/E47 proteins. DISCUSSION The importance of cardiomyocyte contractile activity in maintaining protein synthesis and cardiac mass has been well documented in the intact heart and in cultured myocytes (1)(2)(3)19). Hemodynamic load and mechanical stimuli have also been shown to modulate the expression of cardiac-specific proteins (3, 5, 34 -36). We reported that the effect of contractile activity on myosin heavy chain mRNA expression in cultured cardiomyocytes was independent of serum effects (19), suggesting that the contractile stimulus per se influenced phenotype. Studies of cultured cardiac myocytes subjected to metabolic inhibition and subsequently allowed to recover showed that only when contractile function was re-established did the expression of the cardiac myosin light chain-2 gene return to control levels and the myofilaments reorganize (37).
To study the mechanisms by which a contractile/mechanical stimulus could regulate specific gene transcription, we used primary cultures of neonatal rat cardiomyocytes in which spontaneous contractile activity could be prevented by verapamil or BDM (23,24). When myocyte contractile activity was prevented using either agent, ␣-MHC promoter activity was significantly decreased. Deletional analysis of the ␣-MHC promoter showed that the contractile-responsive region resided within Ϫ80 to Ϫ40 bp of the transcriptional start site. Sequence analysis of this region identified four consensus elements, in-  cluding a novel E-box motif at position Ϫ74, a CArG sequence at Ϫ60, and overlapping M-CAT and E-box elements at Ϫ42/ Ϫ52 (18). E-box binding proteins including the basic helix-loophelix (bHLH) family of proteins are involved in muscle-specific gene expression (15,29) and in establishing diverse cell lineages (38,39). E-box-binding proteins have been shown to modulate expression of several cardiac genes, including ␣and ␤-MHC and cardiac ␣-actin (15,29), and the bHLH2 protein, USF, has been shown to regulate expression of the cardiac ventricular myosin light chain-2 gene (40). In the present study, gel mobility shift analysis showed that the nuclear protein complexed to the contractile-responsive region from Ϫ60 to Ϫ40 bp was an E-box binding protein(s). Although a consensus M-CAT sequence overlaps with this E-box motif (18), we could not detect TEF-1 binding to this region of the promoter. Anti-USF antibodies prevented protein binding to the Ϫ47 E-box element, suggesting that the contractile-responsive nuclear protein is antigenically related to the bHLH/leucine zipper family of transcription factors. Members of this family include Myc, Mad, Mxil, and AP-4 which are able to bind to DNA as oligomers (38 -42). Although many of these transcription factors are ubiquitous, the ␣-MHC proximal E-box may potentially bind a unique combination of protein oligomers which function in transcription preinitiation complex formation (43). Since protein oligomerization and DNA binding affinity have been shown to be altered by phosphorylation of bHLH proteins (44), it is attractive to speculate that phosphorylation of E-box binding factors modulate ␣-MHC promoter activity via a protein kinase signaling pathway (5,18,19). Such a mechanism may explain the process by which changes in myocyte contractile activity regulates specific gene transcription. The importance of this putative contractile response element in ␣-MHC promoter activity in the intact rat heart was evaluated by DNA transfer directly into the ventricular myocardium. In the normal hemodynamically contracting rat heart, mutation of the Ϫ47 E-box element significantly reduced ␣-MHC promoter activity compared with the wild type promoter. These data are the first to identify a contractile-respon-sive element in a gene coding for a sarcomeric protein that is independent of the activation of immediate-early genes (11,12). FIG. 5. The HME E-box-binding protein is antigenically related to USF. Cardiomyocyte nuclear extracts were incubated with the HME probe without USF antibody (lane 1) or in the presence of normal rabbit IgG (lane 2), anti-E12/E47 antibody, or anti-USF antibody (lane 4). The protein-DNA complex B1 that formed in the presence of wild type HME is shown. As a control for the ability of the USF antibodies to interfere specifically with B1 binding to the HME probe, cardiomyocyte nuclear extracts were incubated with the MCT241 oligomer containing an M-CAT motif without (lane 5) and with anti-USF antibody (lane 6).