GATA-5 Is Involved in Leukemia Inhibitory Factor-responsive Transcription of the β-Myosin Heavy Chain Gene in Cardiac Myocytes*

Leukemia inhibitory factor is a member of a family of structurally related cytokines sharing the receptor component gp130. Activation of gp130 by leukemia inhibitory factor is sufficient to induce myocardial cell hypertrophy accompanied by specific changes in the pattern of gene expression. However, the molecular mechanisms that link gp130 activation to these changes have not been clarified. The present study investigated the transcriptional pathways by which leukemia inhibitory factor activates β-myosin heavy chain expression during myocardial cell hypertrophy. Mutation of the GATA motif in the β-myosin heavy chain promoter totally abolished leukemia inhibitory factor-responsive transcription without changing basal transcriptional activity. In contrast, endothelin-1 responsiveness was unaffected by the GATA mutation. Among members of the cardiac GATA transcription factor subfamily (GATA-4, -5, and -6), GATA-5 was the sole and potent transactivator for the β-myosin heavy chain promoter. This transactivation was dependent on sequence-specific binding of GATA-5 to the β-myosin heavy chain GATA element. Cardiac nuclear factors that bind to to the β-myosin heavy chain GATA element were induced by leukemia inhibitory factor stimulation. Last, leukemia inhibitory factor stimulation markedly increased transcripts of cardiac GATA-5, the expression of which is normally restricted to the early embryo. Thus, GATA-5 may be involved in gp130 signaling in cardiac myocytes.

Cardiac muscle cells exit the proliferative cell cycle soon after birth, with little or no capacity for subsequent cell division. Hence, the adult myocardium responds to hemodynamic stimuli through an adaptive hypertrophic response that is characterized by an increase in myocardial cell size without a concomitant increase in myocyte number (for review, see Refs. 1 and 2). During chronic exposure to hemodynamic stress, however, the myocardium ultimately develops an irreversible loss of function and ensuing cardiac muscle failure (3). As such, the identification of the signaling pathways that mediate cardiac muscle hypertrophy is critical to the ultimate elucidation of the molecular basis of cardiac muscle failure.
Cardiac myocyte hypertrophy is associated with specific changes in the pattern of gene expression, exemplified by the induction of ␤-myosin heavy chain (MHC) 1 and atrial natriuretic factor in rodents (4 -6). Although the human ventricular myocardium predominately expresses ␤-MHC under basal conditions, the induction of this gene occurs in atrial myocardium in response to hemodynamic overload (7,8). The regulated expression of cardiac genes has been studied using primary cultures of neonatal rat cardiac myocytes (9 -16). In this in vitro system, a number of growth factors signaling through G-protein-coupled receptors, including ␣ 1 -adrenergic agonists, angiotensin II, and endothelin-1 (ET-1), stimulate increases in myocyte volume and reproduce many of the changes in cardiac gene expression characteristic of the hypertrophic program in vivo. Transcriptional regulation of cardiac genes by these stimuli has been extensively studied using transient transfection assays. DNA binding factors that might mediate the nuclear response to ␣ 1 -adrenergic stimulation include the transcription enhancer factor-1 family, serum-responsive factor, and Sp1 (11)(12)(13).
Recent work has demonstrated that members of a family of structurally related cytokines including leukemia inhibitory factor (LIF) and cardiotrophin-1 induce an increase in cell size in cardiomyocyte culture (17,18). The receptors of this cytokine family are multimeric and share the class-specific transmembrane signal-transducing component gp130 (19 -23). Signaling is triggered through the homodimerization of gp130 (24) or the heterodimerization of gp130 with a related transmembrane signal transducer, the LIF receptor subunit ␤ (25,26). Overexpression of both interleukin-6 and its receptor results in constitutive tyrosine phosphorylation of gp130 (i.e. activation) in the myocardium and left ventricular hypertrophy in vivo (27). Thus,theinductionofcardiomyocytehypertrophythroughgp130dependent signaling pathways is not confined to the in vitro hypertrophy assay but is also observed in vivo. Activation of gp130 by LIF or cardiotrophin-1 is also associated with specific changes in cardiac gene expression (18). The molecular mechanisms that link gp130 activation to these changes have not been clarified.
Members of the interleukin-6-LIF cytokine family have been shown to activate the Janus kinase/signal transducer and activator of transcription (STAT) pathway and phosphorylate STAT3 (23, 28 -30). It is also clear that this family of cytokines can activate Ras and mitogen-activated protein kinase cascades (23,28,31). An activated Ras gene, targeted to myocardium in transgenic mice, elicits ventricular enlargement, atrial natriuretic factor expression, myofibrillar disarray, and impaired relaxation in diastole (32). Conversely, microinjection of dominant-negative Ras protein was reported to block ␣ 1 -adrenergic induction of both morphological changes in myofibrillar structure and expression of atrial natriuretic factor (33), demonstrating a requirement of Ras-dependent pathways for Gprotein-coupled signaling in myocardial cell hypertrophy. The relative contributions of the Ras/mitogen-activated protein kinase and Janus kinase/STAT pathways to gp130-induced cardiac hypertrophy are presently unclear, however, because selective pharmacological inhibition of mitogen-activated protein kinase activation does not block hypertrophy (34). Recently, we and others have shown that zinc finger GATA transcription factors are required for transcriptional activation of the genes for angiotensin II type 1a receptor and ␤-MHC during pressure overload hypertrophy in vivo (35,36). However, pressure overload is a complex stimulus consisting of multiple factors. A specific stimulus linked to GATA factors has not been clarified. In addition, although so far six members of the GATA transcription factor family have been cloned, it is unclear which member of this family plays the most important role. Thus, the present study analyzed cis-acting elements and trans-acting factors required for LIF-responsive ␤-MHC transcription during myocardial cell hypertrophy.

EXPERIMENTAL PROCEDURES
Measurement of Protein Synthesis Rate-Primary ventricular cardiac myocytes were prepared from hearts of 1-2-day-old Sprague Dawley rats as described previously (37). Twenty-four hours after plating, the cells were washed twice with serum-free media and then incubated with 5 Ci/ml [ 3 H]phenylalanine (120 Ci/mmol) and unlabeled phenylalanine (0.36 mmol/liter) in serum-free medium for 48 h in the presence of 2.5 ϫ 10 Ϫ9 M LIF (AMRAD, Melbourne, Australia), 10 Ϫ7 M ET-1 (Peptide Institute, Osaka, Japan), or saline as a control. The cells were washed twice with phosphate-buffered saline, and 10% trichloroacetic acid was added at 4°C for 60 min to precipitate protein. The precipitate was washed three times with 95% ethanol and then resuspended in 0.15 N NaOH. Aliquots were measured by a scintillation counter.
Plasmids were purified by anion exchange chromatography (Qiagen, Hilden, Germany), quantified by measurement of A 260 , and examined on agarose gels stained with ethidium bromide before use.
Transfection and Luciferase/CAT Assays-Twenty-four hours after plating, cells were washed twice with serum-free media and then cotransfected with 4 g of the luciferase construct of interest and 1 g of pRSVCAT using LipofectAMINE (Life Technologies, Inc.) according to the manufacturer's recommendation. After a 2-h incubation with DNA-LipofectAMINE complex, the cells were washed twice with serum-free media and further incubated for 48 h in serum-free media in the presence of 2.5 ϫ 10 Ϫ9 M LIF, 10 Ϫ7 M ET-1, or saline as a control. The cells were then washed twice with ice-cold phosphate-buffered saline and lysed as described (36,37,40,43,47). Luciferase activities were determined in duplicate samples from each plate using a Monolight LB 9501 luminometer (EG&G, Berthold) (36,37,40,43,47). Chloramphenicol acetyltransferase (CAT) activities were determined in the same cell lysate as that used for the luciferase assay (36,37,40,43,47).
Reverse Transcriptase-Polymerase Chain Reaction (PCR)-To detect GATA-5 transcripts in cardiac myocytes, a reverse transcriptase-polymerase chain reaction was carried out as described previously (51). For this particular experiment, we used ventricular myocytes isolated from 1-2-day-old DDY mice, because the rat GATA-5 sequence has not been published. One litter (8 -12 pups) yielded ϳ4 ϫ 10 5 cells. Total RNA was isolated as described (36,37) from these cells and subjected to reverse transcription (8 g of total RNA/sample) with a first-strand cDNA synthesis kit (Amersham Pharmacia Biotech) according to the manufacturer's recommendation.
The PCR primers were designed on the basis of published mouse cDNA sequences for GATA-5 (44) and GAPDH (39) as follows; sense for GATA-5, TCCCACTCTCCTCAACTCT; antisense for GATA-5, ACAC-CAGGTCTCCTGACGTA; sense for GAPDH, TTGCCATCAACGAC-CCCTTC; and antisense for GAPDH, TTGTCATGGATGACCTTGGC. To define the optimal amplification conditions, a series of pilot studies were performed using various amounts of reverse transcription products and various cycle numbers of PCR amplification as described (51). On the basis of these initial experiments, the linear portion of the amplification was determined for both genes. The following conditions were therefore chosen as standard for the PCR reactions in a volume of 50 l: reverse transcription products from 300 ng of RNA for GATA-5 or 150 ng of RNA for GAPDH, 2.5 units of TaqAmpli polymerase (Perkin-Elmer), and 35 cycles of amplification for GATA-5 or 30 cycles for GAPDH and 100 ng of each sense and antisense primers. The amplification was carried out as follows: denaturation, 45 s at 94°C; annealing, 45 s at 54°C; and extension, 90 s at 72°C. The PCR products (10 l/lane) were electrophoresed on a 1.5% agarose gel and stained with ethidium bromide.
Statistical Analysis-All data are expressed as means Ϯ S.E. The significance of differences between mean values was evaluated by the two-tailed Student's t test, and differences were considered significant at the p Ͻ 0.05 level.

LIF Increases Protein Synthesis Rate and ␤-MHC Transcription in Cardiac
Myocytes-Neonatal rat ventricular cardiac myocytes respond to various hypertrophic stimuli by increasing protein synthesis and by specifically changing their patterns of gene expression, e.g. induction of ␤-MHC (12,15). To assess whether LIF induces a hypertrophic response to an extent similar to that of other previously well defined hypertrophic stimuli, such as ET-1, cardiomyocytes were treated with 2.5 ϫ 10 Ϫ9 M LIF or 10 Ϫ7 M ET-1. As reported previously, these two stimuli elicit distinct forms of hypertrophy (width versus length; Ref. 18). Therefore, we have used bulk protein synthesis as a measure of hypertrophy. Stimulation with LIF and ET-1 resulted in a 40 Ϯ 7 and 46 Ϯ 5% increase in the protein synthesis rate, respectively (Fig. 1A). Then we examined the expression of ␤-MHC by Northern blot using an oligonucleotide probe specific for the ␤-isoform of MHC mRNA. We performed these experiments using three independent preparations of cardiac myocytes. Stimulation with LIF and ET-1 increased the expression of ␤-MHC gene in cardiac myocytes by 3.8 Ϯ 0.4and 3.0 Ϯ 0.5-fold, respectively, compared with the saline-stimulated states (Fig. 1B). However, neither LIF nor ET-1 activated the expression of a ubiquitously and constitutively expressed GAPDH gene. Thus, both LIF and ET-1 increased the protein synthesis rate and specifically activated ␤-MHC gene expression to a similar extent in cardiac myocytes.
To determine whether the increase in ␤-MHC gene expression during LIF and ET-1 stimulation of neonatal rat ventricular cells is mediated at the transcriptional level by elements within the 5Ј-flanking region of the ␤-MHC gene, cardiomyocytes were transfected with a ␤-MHC-luciferase reporter construct containing 3542-bp rat ␤-MHC upstream sequence (p-3542␤-MHCluc). To control for transfection efficiency, the cells were co-transfected with a small quantity of pRSVCAT. After 48 h of stimulation with LIF, ET-1, or saline as a control, cardiomyocytes were harvested for luciferase and CAT assays. The 3542-bp ␤-MHC promoter fragment conferred LIF-and ET-1-inducible expression to the luciferase reporter gene (1.9 Ϯ 0.3-and 2.4 Ϯ 0.1-fold, respectively). In contrast, neither LIF nor ET-1 stimulation induced the expression of a transfected luciferase gene driven by the 2936-bp ␣-MHC promoter (0.9 Ϯ 0.2-and 1.2 Ϯ 0.3-fold, respectively). These findings suggest that the upreguated expression of ␤-MHC gene by LIF or ET-1 is mediated, at least in part, through a transcriptional mechanism and that the proximal 3542-bp ␤-MHC promoter sequences are sufficient to mediate LIF-and ET-1-responsive transcription.
LIF-responsive ␤-MHC Transcription Requires an Intact GATA Element-To more precisely determine the downstream molecular events during LIF-induced cardiac hypertrophy, we examined cis-acting elements that mediate LIF-responsive ␤-MHC transcription. A previous study demonstrated that the proximal 333 bp of the rat ␤-MHC promoter are sufficient to mediate muscle-specific transcription in cultured neonatal cardiac myocytes and in sol8 cells (52). As shown in Fig. 2B, in cultured neonatal cardiac myocytes, the transfected Ϫ333/ ϩ34-bp ␤-MHC promoter responded to LIF and ET-1 stimulation, increasing the expression 2.0-and 3.0-fold, respectively. These data demonstrate that important elements exist within the rat ␤-MHC promoter sequences Ϫ333/ϩ34, although they do not rule out possible elements outside these sequences.
The rat ␤-MHC promoter sequences Ϫ333/ϩ34 contain distal and proximal M-CAT elements, previously demonstrated to mediate muscle-specific and ␣ 1 -adrenergic-stimulated transcription of the ␤-MHC gene (12). These also contain a GATA element, shown to mediate cardiac-specific transcription of other genes (50,53,55). Thus, we mutated these elements in the context of the 333-bp ␤-MHC promoter. Mutations were designed to abolish the binding of cardiac nuclear factors (12,53,56). As shown in Fig. 2A, basal transcriptional activity of the transfected 333-bp ␤-MHC promoter was attenuated by simultaneous mutations in both distal and proximal M-CAT elements (66% decrease versus wild type), compatible with a role for the M-CAT element in muscle-specific transcription. Basal activity was unaffected, however, by mutating the GATA motif. LIF-or ET-1-responsive ␤-MHC transcription is shown in Fig. 2B. In contrast to the basal activity, mutating the M-CAT elements affected neither ET-1 nor LIF responsiveness. Notably, however, LIF-but not ET-1-responsive transcription was totally abolished by mutating the GATA element (wild type, 2.0-fold, versus GATA mutant, 0.9-fold; p Ͻ 0.001). Thus, an intact GATA element is required for LIF-responsive ␤-MHC transcription, suggesting a role for this element in LIF-induced cardiac hypertrophy.
GATA whether expression of GATA-4, -5, and -6 can transactivate the LIF-responsive Ϫ333/ϩ34 bp ␤-MHC promoter sequences, we performed transient transfection experiments. We co-trans-fected a luc expression vector driven by the Ϫ333/ϩ34-bp ␤-MHC promoter with a eukaryotic expression plasmid encoding one of GATA-4, -5, or -6 or ␤-galactosidase as a control. Transfection efficiency was monitored by co-transfected pRSV-CAT activity. GATA-4, -5, or -6 could not transactivate the Ϫ333/ϩ34-bp ␤-MHC promoter in cultured neonatal cardiac myocytes, possibly because of the competition for co-factors with endogenous GATA factors. To circumvent this problem, we performed these experiments in NIH3T3 cells, which do not express GATA-4, -5, or -6. As shown in Fig. 3A, expression of GATA-5 resulted in marked (12-fold) activation of the Ϫ333/ ϩ34 bp ␤-MHC promoter. In contrast, a promoter derived from the ubiquitously expressed ␤-actin gene was activated only mildly (3.2-fold). The extent of the ␤-MHC promoter transactivation by GATA-4 or -6 was Ͻ3-fold and did not differ significantly from that of ␤-actin promoter activation. Compatible with a previous report (53), a 2936-bp ␣-MHC promoter was not transactivated by GATA-4, -5, and -6 (Ͻ3-fold). We showed that GATA-4, -5, and -6, to a comparable degree (Ͼ10-fold), transactivated the smooth muscle myosin heavy chain promoter, which contains two GATA motifs. 2 Thus, among members of the cardiac GATA transcription factor subfamily (GATA-4, -5, and -6), GATA-5 is the sole potent activator of the ␤-MHC promoter. In addition, this marked activation occurs in the ␤-MHC promoter but not in the ␣-MHC promoter, compatible with LIF responsiveness in cardiac myocytes.
Next we addressed whether GATA-5 stimulation of the ␤-MHC promoter activity occurred in a sequence-specific manner. Point mutations were introduced into the GATA site of the Ϫ333/ϩ34-bp ␤-MHC promoter to ablate LIF responsiveness (Fig. 2B) as above. The resulting promoter construct (p-333GATA␤-MHCluc) was co-transfected with an expression plasmid, pcDNA-GATA-5, and subsequently assayed for the relative luciferase activity. As seen in Fig. 3B, transactivation of the ␤-MHC promoter was reduced by the GATA site mutation to levels only slightly greater than those exhibited by the ␤-actin promoter. These findings demonstrate that the transactivation of the ␤-MHC promoter by GATA-5 is dependent on an intact GATA sequence.
GATA-5 Strongly Binds to the ␤-MHC GATA Element-To determine whether the GATA motif in the ␤-MHC promoter can interact with GATA-5, EMSAs were performed. In vitro-translated GATA-5 was probed with a radiolabeled oligonucleotide containing the ␤-MHC GATA site (Fig. 4, lanes 2-6) in the presence or absence of competitor DNAs. Competition EMSAs revealed that a retarded band represented specific binding (Fig. 4, lane 2), as evidenced by the fact that it was competed by a 100-fold molar excess of unlabeled ␤-MHC GATA oligonucleotide (Fig. 4, lane 3). The retarded band represents an interaction of the probe with GATA-5, because it was absent in unprogrammed lysate (Fig. 4, lane 1). The retarded band was also competed by an unlabeled CEF-1 oligonucleotide (Fig. 4, lane  4), which contains the GATA motif in the cTnC promoter (48) previously demonstrated to be a binding site of GATA-5. In contrast to the wild-type ␤-MHC GATA site, the gel shift could not be competed by an excess of an oligonucleotide containing the ␤-MHC GATA site into which point mutations (Fig. 2B) that ablate LIF responsiveness had been introduced (Fig. 4,  lane 5) or by the same amount of a nonspecific oligonucleotide (Fig. 4, lane 6), confirming the sequence-specific nature of the interaction. These findings demonstrate that GATA-5 can bind the ␤-MHC GATA site in a sequence-specific manner.
Although cTnC promoter has been shown to be efficiently transactivated by both GATA-4 and -5 (44,50), the present study demonstrated that the ␤-MHC promoter was activated by GATA-5 but not by GATA-4. We have investigated whether this difference in the transactivation intensity is attributable to the ability of cTnC and ␤-MHC GATA sites to bind GATA-5 relative to GATA-4. In vitro-translated GATA-4 or GATA-5 was probed with a radiolabeled ␤-MHC GATA oligonucleotide (Fig.  5A, lanes 3-6). Although a retarded band showing the interaction of the ␤-MHC GATA site with GATA-5 was strong in its intensity (Fig 5A, lane 5), a band showing the interaction with GATA-4 was very weak (Fig. 5A, lane 3). In vitro-translated GATA-4 or GATA-5 derived from the same lysates with those used for the ␤-MHC GATA site was also probed with CEF-1 oligonucleotide containing the GATA motif in the cTnC promoter (Fig. 5B, lanes 3-6). In contrast to the ␤-MHC GATA site, the intensity of the band showing the interaction of the CEF-1 with GATA-5 (Fig. 5B, lane 5) was similar to that showing the interaction with GATA-4 (Fig. 5B, lane 3). Thus, the ability of GATA elements to bind GATA-5 relative to GATA-4 differs between ␤-MHC and cTnC promoters.
LIF Induces the Expression of GATA-5 in Neonatal Cardiac Myocytes-To determine whether LIF modulates the ␤-MHC GATA binding activity in cardiac myocytes, EMSAs were performed with nuclear extracts from saline-and LIF-stimulated neonatal cardiac myocytes. Nuclear extracts were probed with a radiolabeled ␤-MHC GATA oligonucleotide in the presence or absence of competitor oligonucleotides (Fig. 6, lanes 1-5). Competition EMSAs revealed that one retarded band (Fig. 6, lanes  1 and 2, arrow) represented GATA sequence-specific binding, as evidenced by the fact that it was competed by an unlabeled ␤-MHC GATA oligonucleotide (Fig. 6, lane 3) or by an oligonucleotide containing a previously demonstrated GATA site in the cTnC promoter (CEF-1; Fig. 6, lane 4), but not by an excess of the ␤-MHC GATA site containing point mutations that ablate LIF responsiveness (Fig. 6, lane 5) or by a nonspecific oligonucleotide (data not shown). Notably, the activity of the specific band was increased in nuclear extracts from LIF-stimulated myocytes (Fig. 6, lane 2) compared with those from salinestimulated cells (Fig. 6, lane 1). This experiment was repeated three times using three independent preparations of cells and found to be reproducible. Thus, LIF up-regulated the ␤-MHC GATA binding activity in nuclear extracts from cardiac myocytes.
To investigate whether the up-regulated ␤-MHC GATA binding activity represents increased GATA-5 transcripts in cardiac myocytes, we examined GATA-5 mRNA levels in saline-and LIF-stimulated cells. Previous studies demonstrated that the GATA-5 expression in the heart is restricted to the early embryonic stage and not detectable in the late embryo or in the adult by Northern blots. Using highly sensitive reverse transcriptase-PCR, a faint band indicating GATA-5 was detectable in saline-stimulated neonatal rat cardiac myocytes (Fig. 7). Notably, the band intensity markedly increased in LIF-stimulated cardiac myocytes. We confirmed by sequencing that this band represents a specific PCR product derived from GATA-5 cDNA. In contrast, the intensity of the band indicating GAPDH was almost comparable between saline-and LIF-stimulated cardiac myocytes. We repeated these experiments with three independent preparations of cardiac myocytes. With the use of a previously described semiquantitative reverse transcriptase-PCR (51), the cumulative results indicated that GATA-5 mRNA relative to GAPDH mRNA was 6.2 Ϯ 0.5-fold higher in the LIF-stimulated cardiac myocytes than in the saline-stimulated cells.

DISCUSSION
Cardiac myocyte hypertrophy is a central feature of all types of cardiac muscle disease and is an interesting example of the response of a terminally differentiated cell type to growth stimulation. Current insights into the mechanisms controlling cardiomyocyte hypertrophy have been obtained primarily from a cell culture model, in which growth factors signaling through G-protein-coupled receptors induce hypertrophy (9 -16). Growing evidence suggests that gp130 activation is also coupled to myocardial cell hypertrophy (17,18,27). Using neonatal cardiac myocytes in culture, the present study demonstrates that mutation of the GATA motif in the ␤-MHC promoter totally abolished LIF-responsive transcription. Among members of the cardiac GATA transcription factor subfamily, GATA-5 alone was able to potently transactivate the ␤-MHC promoter. This transactivation was dependent on sequence-specific binding of GATA-5 to the ␤-MHC GATA element. Last, LIF stimulation markedly increased levels of GATA-5 transcripts in cardiac myocytes. These findings demonstrate that GATA-5 is important in the LIF-mediated up-regulation of ␤-MHC expression in cardiac myocytes and may represent the mechanism underlying the cardiac hypertrophy induced by the gp130 ligand family.
Role of M-CAT and GATA Elements in ␤-MHC Transcrip-tion-Once the hypertrophy signal is transduced from the membrane to the nucleus, a fundamental reprogramming occurs within cardiac myocytes that results in the reexpression of genes encoding fetal protein isoforms. Genes such as skeletal ␣-actin, ␤-MHC, and atrial natriuretic factor become highly expressed within ventricular myocytes (4 -8). Studies focused on elucidating the transcriptional regulation of these genes have identified a group of cis-acting regulatory elements that might mediate the nuclear response to hypertrophic stimuli.  6) were probed with a radiolabeled oligonucleotide containing the ␤-MHC GATA site (A) or that containing the cTnT GATA site (CEF-1) (B). Unlabeled competitor DNAs are the same oligonucleotides as those used as probes.
Analysis of the ␤-MHC promoter has demonstrated that the M-CAT element, a binding site of the transcription enhancer factor-1 family, may play a role in both basal and hypertrophicresponsive transcription. For example, this element mediates both ␣ 1 -adrenergic and ␤-protein kinase C-stimulated ␤-MHC transcription (11,12). M-CAT elements have also been implicated in ␣ 1 -adrenergic-stimulated expression of the skeletal ␣-actin and ␤-type natriuretic peptide promoters (13,57). In addition, the M-CAT motif is present in the promoters of several striated muscle-specific genes, where it functions to positively regulate basal transcription (57)(58)(59)(60)(61)(62)(63). Our observation that simultaneous disruption of both M-CAT elements in the setting of a 333-bp ␤-MHC promoter decreases transcriptional activity is consistent with previous reports demonstrating that M-CAT elements play an important role in basal ␤-MHC transcription. Unexpectedly, however, simultaneous mutations in both M-CAT elements, adequate to abrogate binding of nuclear proteins and to destroy enhancer function, had no effect on the ET-1 or LIF responsiveness of the 333-bp rat ␤-MHC promoter. Although this finding does not rule out the possibility that M-CAT elements contribute to ET-1 or LIF responsiveness, it demonstrates that other elements within these sequences suffice to mediate this transcriptional response.
Sequences Ϫ333/ϩ34 of the rat ␤-MHC promoter also contain a GATA element (36). GATA elements have been shown to be important for cardiac-specific transcription in many cardiac genes, including ␣-MHC, ␤-type natriuretic peptide, myosin light chain 1/3 and cTnC (50,(53)(54)(55). We show here that mutation of the GATA element in the 333-bp ␤-MHC promoter totally abolished LIF-responsive transcription without changing basal transcriptional activity. Thus, this GATA element plays a critical role in LIF-responsive ␤-MHC transcription. In contrast, ET-1 responsiveness was unaffected by the GATA mutation. These findings suggest that LIF and ET-1 activate ␤-MHC gene transcription through distinct cis-acting elements. Previous work has shown that G-protein and gp130 pathways elicit morphologically distinct forms of myocardial cell hypertrophy (18). Thus, it appears that these two stimuli induce distinct hypertrophic processes through different pathways. The elucidation of the differences in these signaling pathways and the pathophysiological significance of these two forms of hypertrophy would be of particular interest.
GATA Factors Mediate gp130 Signaling in Cardiac Myocytes-To date, six related zinc finger-containing proteins have been described, which recognize and bind the GATA motif (44,45,64). The proteins fall into two subgroups: those containing GATA-1, -2, and -3, and those with GATA-4, -5, and -6. The subgroups are defined by both sequence homology and expression pattern, with GATA-1, -2, and -3 predominating in blood and ectodermal derivatives and GATA-4, -5, and -6 in heart and endodermal derivatives. Interestingly, the genes encoding GATA-4 and -6 are expressed in the heart throughout embryonic and postnatal development, whereas the murine GATA-5 gene is normally expressed in a temporally and spatially restricted pattern within the embryonic heart (44,45). These findings raise the possibility that GATA-4, -5, and -6 play differential roles during LIF-induced hypertrophy. The present study demonstrated that neither GATA-4 nor -6 significantly activated the LIF-responsive 333-bp ␤-MHC promoter. In contrast, GATA-5 markedly stimulated this promoter. This activation required an intact GATA element, suggesting a direct effect. Consistent with this model, GATA-5 bound the ␤-MHC GATA element in a sequence-specific manner. Importantly, LIF stimulation increased ␤-MHC GATA binding activity in cardiac nuclear extracts. Although GATA-5-specific antisera for supershift experiments is not available at present, a complex formed with the ␤-MHC GATA element is clearly GATA sequence specific. In addition, LIF stimulation increased expression of GATA-5 in neonatal cardiac myocytes. Taken together, these findings demonstrate that GATA factors are involved in LIFresponsive ␤-MHC transcription and that GATA-5 is the factor that is primarily involved.
The signal transduction pathways by which members of the gp130 ligand family activate target genes have been well studied in several cell types (28 -31). Typically, the phosphorylated STAT proteins dimerize, translocate into the nucleus, and bind to the promoter of target genes. The DNA binding targets of STATs include the interferon-␥ activation site-like elements (TTC/ANNNG/TAA) and the interferon-␥-stimulated response elements (AGTTTCNNTTTCNC/T) (31). Our data demonstrate that GATA-5 markedly activated the ␤-MHC promoter through specific binding to the GATA element and that ␤-MHC GATA binding activity in cardiac myocytes is induced by LIF stimulation. At present, the molecular events that may link the Janus kinase/STAT pathway to this augmentation are unknown. LIF activates STATs within 15 min after LIF stimulation, whereas the induction of the ␤-MHC expression occurs much later (48 h after LIF stimulation). Therefore, it is unlikely that STATs directly associate with GATA factors in the activation of the ␤-MHC promoter. We demonstrate here that LIF stimulation increased GATA-5 transcripts in neonatal cardiac myocytes. Thus, one model is that the Janus kinase/STAT pathway is linked to the regulation of GATA-5 gene expression. Another model is that LIF-induced intracellular signaling cascades activate GATA factors or GATA co-activators post-translationally by phosphorylation or other mechanisms. In any event, because gp130 activation is one component of the hemodynamic overload stimulus (65), additional studies to delineate the precise mechanisms by which gp130 activation induces ␤-MHC transcription are likely to provide significant insight into the pathways that mediate hemodynamic overload-induced hypertrophy in vivo. FIG. 7. The effect of LIF on the expression of GATA-5 mRNA by cultured ventricular cardiomyocytes. Cardiomyocytes were challenged with saline or LIF for 48 h. Representative photographs of PCR products after reverse transcriptase-PCR for GATA-5 and GAPDH mRNAs are shown. Each lane represents RNA from a separate culture plate. Similar results were obtained from three independent experiments. M, molecular marker.