INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE SARCOLEMMAL Na⁺/Ca²⁺ exchanger is the primary Ca²⁺ efflux mechanism involved in regulating cardiac intracellular Ca²⁺ and beat-to-beat contractility (3, 34). Three Na⁺/Ca²⁺ exchanger genes (NCX1, NCX2, NCX3) have been identified (21, 32, 33). Of these, only NCX1 is expressed in cardiac myocytes. NCX1 uses alternative promoters that are important in controlling the level of expression of the gene in various tissues (2, 19, 31). In particular, tissue-specific NCX1 transcripts with distinct 5'-untranslated regions have been identified in high concentrations in the heart, kidney, and brain of the rat (19) and cat (2). In each organ, the exchanger is localized to specific cell types. In the heart, NCX1 is expressed mainly in cardiac myocytes, whereas in the kidney, NCX1 appears to be most abundant in the distal tubule. In the brain, NCX1 is most prevalent in neurons. Whereas some of the factors that determine tissue specificity of multiple genes, such as skeletal muscle genes, have been described, little is currently known about the control of cardiac-specific genes and, in particular, the cardiac-specific regulation of NCX1.

The Myo-D basic helix-loop-helix (bHLH) family and other related transcriptional factors (myogenin, Myf-5, and MRF4) are important in skeletal muscle for the restricted expression of genes involved in skeletal myogenesis (6, 9, 23). Recently, other bHLH family members (eHAND/dHAND) (37), previously identified homeobox transcription factors (Nk proteins) (11), and the zinc finger GATA proteins (particularly GATA-4, -5, and -6) were shown to be involved in cardiac development (16). Additionally, some of these proteins, such as GATA-4 and Nkx2.5, may interact with each other to fine-tune regulation and to synergize the transcription of certain cardiac-specific genes (36).

In general, the bHLH family of transcription factors binds to a consensus DNA binding site CANNTG (where N can be any nucleotide) called the "E-box" (30). eHAND and dHAND were the first bHLH family members found in the heart (7, 37). On the other hand, the murine cardiac-specific homeobox gene Nkx2.5/Csx (a vertebrate homolog of the Drosophila tinman gene) recognizes the DNA binding element TNAAGTG (5). Both of these proteins demonstrate differential expression during cardiac development, and mutational studies indicate that disruption of these proteins results in cessation of chick cardiogenesis at the looping tube stage (22, 37).

The GATA family of transcription factors, including GATA-1-6, have two adjacent zinc fingers that bind to a WGATAR (where W is A or T and R is A or G) consensus sequence with a highly conserved basic DNA binding domain. It is not known how each member discriminates its target site, but it has been suggested (16) that this is accomplished by different binding specificities to flanking DNA sequences. The COOH terminus of GATA factors contains a translocation domain. The NH₂ terminus is typically the site of activation at which interaction with other proteins, such as cofactors or the general transcription factors, occurs (28). Whereas GATA-1, -2, and -3 are involved in nonredundant functions of hematopoietic-restricted gene expression, GATA-4, -5, and -6 are expressed predominantly in heart and gut and appear to have a role in regulating embryonic heart formation and gut differentiation in various species (1, 17, 18, 25). In Xenopus, GATA-4, -5, and -6 genes are associated with cardiac specification and regulation of cardiac-specific actin and -myosin heavy chain gene transcription during embryogenesis (15). Whereas no study clearly establishes the presence of GATA-5 or GATA-6 expression at the neonatal stage of development, it is unequivocal that regulation of multiple cardiac-specific genes by GATA-4 at this stage is necessary (10, 14, 24). In postnatal development, gene expression of GATA-4 and GATA-6 is evident; however, GATA-5 gene expression is no longer observed in the heart but is restricted to the gut, bladder, and lung (26, 27).

Immunofluorescence studies indicate that GATA-4 is present in rat neonatal cardiac myocytes. GATA-4 can regulate the cardiac-specific expression of both the troponin I gene and the troponin C enhancer in nonmuscle cells (29). Durocher et al. (8) and Lee et al. (20) have demonstrated that both GATA-4 and GATA-5 act synergistically with Nkx2.5 to activate the rat cardiac atrial natriuretic factor and the brain natriuretic peptide (BNP) promoters in noncardiac cells. Sepulveda et al. (36) have further shown that the cardiac -actin gene functions similarly, requiring the combinatorial interaction of three transcription factors, including Nkx2.5 and GATA-4 as well as the serum response factor.

Recently, it has been shown that the proximal cardiac NCX1 promoter in rat and cat both contain E-box elements and multiple binding sites for GATA transcription factors (2, 31). However, it is yet to be determined whether these sites are functional or involved in cardiac-specific expression of NCX1. In this study, supershift experiments indicate that GATA-4, and not GATA-5 or GATA-6, is present in neonatal myocyte nuclear extract and interacts with the NCX1 promoter. Site-directed mutagenesis of GATA DNA elements (particularly the proximal GATA-4 element) significantly reduces activity of the rat cardiac NCX1 promoter, but mutation of the E-box binding site does not alter its activity. This is the first study to establish that GATA-4 is responsible for cardiac tissue specificity of the cardiac NCX1 minimum promoter.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Plasmid constructs. NCX1-luciferase wild-type promoter constructs containing the rat cardiac full-length promoter (at 336 bp) and the minimum promoter (at 137) were cloned in the pGL2 basic (pGL2B; Invitrogen) luciferase reporter vector background as previously described to create pGL2-336 and pGL2-137 (31). Mutant promoter constructs were created by the QuikChange site-directed mutagenesis method (Stratagene) using PCR primers (Retrogen, San Diego, CA) and the wild-type plasmid pGL2-336. The E-box at 175 was mutated from CATGTG to ACGTGT, and each of two GATA DNA elements (at 75 and 145) was mutated to TCGC. PCR primer pairs to the first GATA binding site at 75 (sense: 5'-GAAAGGCACATCGCAGCTGAGAGC-3' and antisense: 5'-GCTCTCAGCTGCGATGTGCCTTTC-3') were used to construct pGL2-336mG1 and pGL2-137mG1. PCR primer pairs to the second GATA binding site at 145 (sense: 5'-TGAGAGCTGCCATCGCGCTTCTCGACAG-3' and antisense: 5'-CTGTCGAGAAGCGCGATGGCAGCTCTCA-3') were used to construct pGL2-336mG2. A double mutant, pGL2mG1,2, was similarly created containing mutations in both GATA sites. PCR primer pairs (sense: 5'-ATTTTTATCACCACGTGTTTGGATGAGGCT-3' and antisense: 5'-AGCCTCATCCAAACACGGGTGGTGATAAAAAT-3') were used to construct the E-box mutant promoter pGL2mE. All plasmids were sequenced in both directions using Sequenase (version 2.0; USB, Cleveland, OH).

Cell culture and transfections. Primary neonatal cardiac myocytes from 24- to 48-h-old Sprague-Dawley rats were isolated as previously described (31) and plated at 6 × 10⁶ cells/35-mm culture dish. The cells were maintained in DMEM-F-12 with 10% horse serum and 5% fetal bovine serum at 37°C in 5% CO₂ for 18-24 h before transfection.

Enzyme assays. Luciferase assays were performed using 20 µl of cell lysate and 100 µl of assay buffer (100 mM Tricine, 10 mM MgSO₄, and 2 mM EDTA, pH 7.80), containing 2 mM ATP and 10 mM luciferin in a Monolight 1500 luminometer (Analytical Luminescence). The relative light units (RLU) were automatically recorded.

Nuclear extract preparation. Nuclear extract from primary cardiac myocytes was prepared from culture dishes by the following method. The cells were pelleted at 2,000 rpm in a bench-top centrifuge for 4 min at 4°C and then resuspended in 400 µl of cold buffer A [10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 1 mM dithiothreitol (DTT), and 0.5 mM phenylmethylsulfonyl fluoride (PMSF)] and incubated on ice for 15 min. A total of 10 µl of 10% NP-40 was added, and the suspension was vortexed at 14,000 rpm for 10 s. The pellet was collected by centrifugation at 6,000 rpm for 4 min and resuspended in 50 µl of cold buffer B (20 mM HEPES, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM DTT, and 1 mM PMSF). The tube was intermittently shaken vigorously at 4°C for ~2 h. The mixture was centrifuged at 14,000 rpm for 5 min at 4°C, and the supernatant containing the nuclear extract was collected. The protein concentration was measured by spectrophotometry. The sample was then aliquoted and stored at 70°C for use within 6 mo.

Electrophoretic mobility shift assay. Double-stranded oligonucleotides were labeled with [-³²P]ATP using T4 polynucleotide kinase (Life Technologies). The labeled DNA was purified over a G-50 Sephadex Quick Spin column (Boehringer Mannheim, Indianapolis, IN), and the counts per minute (cpm) were measured in a scintillation counter. Each electrophoretic mobility shift assay (EMSA) binding reaction consisted of a total of 20 µl with 4 µl of assay buffer (40 mM KCl, 15 mM HEPES, pH 7.9, 1 mM EDTA, 0.5 mM DTT, 6 mM MgCl₂, and 10% glycerol), 10-12.5 µg of nuclear extract protein, 8 µg of BSA, and 10 µg of poly(dI-dC) to reduce nonspecific protein-DNA binding. The binding reaction mixture was placed on ice for 15-20 min, followed by the addition of 20,000-40,000 cpm of labeled probe. GATA consensus (sense: 5'-CACTTGATAACAGAAAGTGATAACTCT-3' and antisense 5'-AGAGTTATCACTTTCTGTTATCAAGTG-3') and mutant (sense: 5'-CACTTCTTAACAGAAAGTCTTAACTCT-3' and antisense: 5'-AGAGTTAAGACTTTCTGTTAAGAAGTG-3') oligonucleotides obtained from Santa Cruz Biotechnology were used. Competition studies were done with varying concentrations of unlabeled competitor DNA before the addition of labeled probe. Antibody (40 µg) for supershift experiments was added for 15-20 min after the addition of probe. Antibody to GATA-4 was obtained from Dr. D. Wilson (Washington University, St. Louis, MO), and the antibodies to GATA-5 and GATA-6 were obtained from Santa Cruz Biotechnology. The reaction was run on a 4% acrylamide gel, prerun for 1 h in 1× Tris borate-EDTA, for 3.5 h at ~200 V. The gel was dried and exposed to Kodak X-Omat film at 70°C with intensifying screens for the indicated time.

Statistical analysis. The raw data were collected and analyzed using standard statistical analyses and Student's t-test.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

GATA-4 binds to the rat cardiac NCX1 minimum promoter. As demonstrated by Nicholas et al. (31) and Barnes et al. (2), the activity of the minimum promoter of the rat (137) and cat (250) NCX1 genes are both cardiac tissue specific. Sequence comparison of the promoters indicates a common GATA element that may account for this characteristic. Analogous GATA elements have proven to be the binding sites (particularly for GATA-4 transcription factor) responsible for directing cardiac specificity for other cardiac genes, such as -myosin heavy chain (-MHC) and BNP (10, 24). Possibly, the proximal GATA element common to both the rat and cat minimum promoter is responsible for tissue specificity of the NCX1 gene, and the upstream GATA element and E-box binding sites may play some accessory role. Figure 1 shows a limited comparison of the genomic sequence of the rat and cat cardiac NCX1 promoters. There is >90% sequence identity between the E-box site and the transcription start site but <20% identity in the 5'-flanking DNA sequence upstream of the E-box. Thus these GATA and E-box elements are reasonable targets for investigation of their roles in cardiac-specific expression of NCX1.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Comparison of rat and cat cardiac Na+/Ca2+ exchanger NCX1 promoters. A: schematic shows relative location of proximal and distal GATA elements and E-box binding site to transcription start site in rat and cat cardiac NCX1 full-length promoters. B: comparison between rat and cat cardiac NCX1 full-length promoter DNA sequences indicates >90% sequence identity between E-box binding site and transcription start site. Region upstream from E-box has <20% sequence identity. Nonidentical nucleotide bases are represented in lowercase type.

View larger version (135K): [in this window] [in a new window]   Fig. 2.   GATA-4 transcription factor is present in neonatal primary cardiomyocyte nuclear extract. A: In this electrophoretic mobility shift assay, a commercially available oligonucleotide (oligo) containing a GATA consensus sequence (SC-GATA) was labeled and incubated with nuclear extract (N.E.) obtained from neonatal primary cardiac myocytes. All lanes contain labeled probe. Lane 2: band shift in presence of nuclear extract. Lane 3: positive competition in presence of excess unlabeled probe. Lane 4: lack of competition with unlabeled probe containing a mutation in GATA sequence. Lane 5: no evidence of nonspecific binding between probe and GATA-4 antibody (Ab). Lane 6: a supershifted band (arrow) appears when antibody to GATA-4 is added. B: labeled SC-GATA was again incubated with nuclear extract obtained from neonatal primary cardiac myocytes. All lanes contain labeled probe. Lane 2: band shift in presence of nuclear extract. Lane 3: supershifted band (arrow) similar to that described in A (lane 6) with antibody to GATA-4. Lanes 5 and 7: lack of a supershifted band when antibody to either GATA-5 or GATA-6 is added. Lanes 4 and 6: lack of nonspecific DNA binding with antisera.

View larger version (80K): [in this window] [in a new window]   Fig. 3.   GATA-4 binds to proximal GATA site (at 75 bp) of cardiac NCX1 promoter. This competition and supershift assay was performed with labeled NCX1 oligonucleotide containing GATA binding site at 75. All lanes contain labeled probe. Lane 2: band shift in presence of neonatal primary cardiac myocyte nuclear extract. Lane 3: result of positive competition in presence of excess unlabeled probe. Lane 4: lack of competition with unlabeled probe containing a mutation in GATA sequence. Lane 5: supershifted band when antibody to GATA-4 is added. Lanes 6 and 7: no supershift indicated with antibody to GATA-5 or GATA-6. Parallel studies with probe and antisera showed no evidence of nonspecific binding.

Mutation of the GATA site in the cardiac NCX1 minimum promoter abolishes its activity. We next determined whether a mutation in the single GATA site at 75 of the cardiac NCX1 minimum promoter would alter promoter activity. A cardiac NCX1 minimum promoter with the GATA sequence mutated to TCGC was generated by site-directed mutagenesis to create the plasmid pGL2-137mG1. To investigate the contribution of the GATA elements in the context of the full-length NCX1 promoter (336), three other mutant promoters were similarly constructed. The plasmid pGL2-336mG1 contained a mutation in the proximal GATA site at 75, pGL2-336mG2 contained a mutation in the distal GATA binding site at 145, and pGL2-336mG1,2 contained mutations in both GATA binding sites. The plasmid pGL2-336mE, containing a mutation in the E-box at 175, was also created. Each of the five plasmids (pGL2-137mG1, pGL2-336mG1, pGL2-336mG2, pGL2-336mG1,2, and pGL2-336mE) was transfected into primary neonatal cardiac myocytes, and mutant promoter activity was compared with that of the corresponding wild-type promoter.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Disruption of GATA binding, but not E-box binding, significantly reduces activity of cardiac NCX1 minimum (137) and full-length (336) promoter by >90%. A: neonatal primary cardiac myocytes were transfected with promoterless vector pGL2B or plasmid containing either wild-type cardiac NCX1 minimum promoter pGL2-137 or mutant minimum promoter pGL2-mG1, which has a mutation in GATA site at 75. Minimum promoter activity was reduced by >90% by mutation. Ordinate represents activity compared with that of RSV promoter, which provided maximum non-tissue-specific promoter activity (Max). * P < 0.05. Data represent average of at least N = 3 experiments done in duplicate. B: neonatal primary cardiac myocytes were transfected with promoterless vector pGL2B or either cardiac NCX1 wild-type full-length promoter pGL2-336 or mutated promoters pGL2-336mG1, containing a mutation in GATA site at 75; pGL2-336mG2, containing a mutation in GATA site at 145; or pGL2-336mG1,2, containing mutations in both GATA sites. Ordinate represents activity compared with that of RSV promoter. Data represent average of at least N = 4 experiments done in duplicate. * P <0.05. C: transfections were performed with wild-type cardiac NCX1 full-length promoter pGL2-336 and pGL2-336mE, containing a mutation in E-box binding site at 175. Data represent average of at least N = 4 experiments done in duplicate.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Identification of skeletal muscle-specific transcription factors has allowed a better understanding of the mechanism of muscle differentiation and the expression of skeletal muscle-specific genes. In the same way, the discovery of cardiac muscle-specific transcription factors now enhances our knowledge of the role of alternative promoters of particular genes and how the promoters may be regulated. We and others (2, 31) have previously shown that the NCX1 gene has alternative promoters that direct expression of tissue-specific transcripts in the heart, kidney, and brain. In particular, the cardiac NCX1 minimum promoter contains all the information required for expression of the cardiac-specific transcript. This study has focused on understanding how the cardiac NCX1 promoter accomplishes this task.

The results of electrophoretic mobility shift and supershift assays indicate that the transcription factor GATA-4, and not GATA-5 or GATA-6, is present in rat neonatal (1-2 days old) primary cardiac nuclear extract and that it interacts with a GATA DNA element within the cardiac NCX1 minimum promoter (Fig. 3). Disruption of this binding site by mutagenesis not only eliminates protein interaction with the DNA but also results in >90% loss of activity of the minimum promoter (137) and ~60% reduction in activity of the full-length promoter (336) (Fig. 4). These data correlate well with those observed for other cardiac-specific genes (10, 14, 24).

Although the minimum promoter (pGL2-137) possesses cardiac-specific activity, the level of that activity (~1.4% maximum; Fig. 3A) is approximately less than one-half that of the full-length promoter (pGL2-336, ~3.4% maximum; Fig. 4B). Furthermore, loss of the single GATA-4 binding site (at 75), common to both promoters, results in almost complete loss of activity of the minimum promoter but only a partial loss of activity of the full-length promoter. This suggests that, whereas the proximal GATA-4 site may be responsible for complete activity of the minimum promoter, the distal GATA-4 site functionally cooperates to enhance activity of the full-length promoter. These experiments do not exclude the possibility that other enhancer elements of the NCX1 promoter may exist.

Transcription factors, such as Myo-D, that bind to the E-box have a distinct role in expression of skeletal muscle-specific genes (6, 9, 23). The E-box also appears to be important in cardiac morphogenesis (37), but, so far, E-box-dependent transcription has not emerged as an important mechanism in controlling the activity of cardiac muscle-specific genes, such as NCX1 (Fig. 4C).

Early experiments suggested that GATA-4 played an important role in cardiac development (1). However, it is only recently that investigation with GATA-4-deficient null mice by Kuo et al. (17) and Molkentin et al. (25) clearly defined the role for this transcription factor as necessary for the control of rostral-to-caudal and lateral-to-ventral folding morphogenesis essential during normal cardiogenesis. We show that the cardiac promoter of the rat NCX1 gene is also GATA-4 dependent and can therefore be added to the list of other GATA-4-dependent, cardiac-specific genes. With this information, an intriguing question arises. Why is GATA-4 an important and common regulator of these gene promoters?

The literature indicates that some of these GATA-4-dependent genes (particularly -MHC, cardiac troponin C, and skeletal -actin, as well as NCX1) undergo orchestrated upregulation during early cardiac morphogenesis and become transcriptionally reactivated during cardiac hypertrophy. Various in vivo and in vitro models of cardiac hypertrophy indicate that numerous autocrine and paracrine systems participate to effect the hypertrophic signaling response (40). A role for the mitogen-activating protein kinase system in the reexpression of embryonic muscle genes has been examined (38, 39). Recent work indicates that GATA-4 may also be involved in the hypertrophic signaling response. Hasegawa et al. (12) have shown that GATA-4 is important in upregulation of -MHC in the aortic banding model of cardiac hypertrophy, whereas Herzig et al. (13) have demonstrated GATA-4 involvement in the pressure-overload induction of the angiotensin II type Ia receptor. It may not be surprising to find that GATA-4 may also participate in the inducible expression of NCX1 during cardiac hypertrophy.