Differential effects of protein kinase C, Ras, and Raf-1 kinase on the induction of the cardiac B-type natriuretic peptide gene through a critical promoter-proximal M-CAT element.

The cardiac genes for the A- and B-type natriuretic peptides (ANP and BNP) are coordinately induced by growth promoters, such as α1-adrenergic receptor agonists (e.g. phenylephrine (PE)). Although inducible elements in the ANP gene have been identified, responsible elements in the BNP gene are unknown. In this study, reporter constructs transfected into neonatal rat ventricular myocytes showed that in the context of 2.5 kilobase pairs of native BNP 5′-flanking sequences, a 2-base pair mutation in a promoter-proximal M-CAT site (CATTCT) disrupted basal and PE-inducible transcription by more than 98%. Expression of constitutively active forms of Ras, Raf-1 kinase, and protein kinase C, all of which are activated by PE in cardiac myocytes, strongly stimulated BNP reporter expression. Isolated M-CAT elements conferred PE, protein kinase C, and Ras inducibility to a minimal BNP promoter, however, they did not confer Raf-1 inducibility. These results show that M-CAT elements can serve as targets for Ras-dependent, Raf-1-independent pathways, implying the involvement of c-Jun N-terminal kinase and/or p38 mitogen-activated protein kinases, but not extracellular signal-regulated protein kinase/mitogen-activated protein kinase. Moreover, the essential M-CAT element distinguishes the BNP gene from the ANP gene, which utilizes serum response elements and an Sp1-like sequence.

The A-, B-, and C-type natriuretic peptides (NPs) 1 are structurally related cardiac-derived peptides with vasorelaxant, diuretic, and natriuretic effects (1)(2)(3)(4)(5)(6)(7). Upon treatment with stimuli that can eventually lead to the hypertrophic growth of ventricular myocytes, several embryonic cardiac genes are reactivated, including those for ANP, BNP, ␤-myosin heavy chain, and skeletal ␣-actin (6 -18). While the precise function of this recapitulation of embryonic cardiac gene expression is unclear, it can be speculated that increased NP production represents a compensatory endocrine response to stimuli that often increased blood pressure. Accordingly, a knowledge of the mechanisms by which the NPs are induced during the hypertrophic growth program will provide a better framework upon which to understand how the expression levels of the hormones are regulated under less severe, but nonetheless hemodynamically challenging physiological conditions.
A variety of studies have addressed the mechanisms responsible for ANP induction in primary neonatal rat cardiac myocytes (11,12,15,17). Recent studies have demonstrated the importance of serum response elements (SREs) (18) as well as SP-1-like elements (16) in the transcriptional activation of ANP in response to ␣ 1 -adrenergic agonists. Earlier reports have also indicated the probable involvement of AP-1-binding cis-sequences, also known as 12-O-tetradecanoylphorbol-13-acetate response elements (TREs), in regulating transcription of human ANP (19). Accordingly, it is believed that ␣ 1 -adrenergic agonists, and perhaps other ANP inducers, stimulate myocardial cell signaling pathways, which eventually lead to the activation of serum response factor, an Sp1-like protein, and perhaps AP-1, each of which converge on the transcriptional enhancement of ANP.
Relatively little is known about the regulation of BNP expression. Phorbol esters and diacylglycerol increase BNP mRNA and peptide levels (20,21), suggesting the involvement of PKC. Additionally, BNP promoter activity, measured using constructs containing approximately 2.5 kb of the rat BNP 5Ј-FS upstream of a luciferase reporter, is inducible by phorbol esters, serum, or the ␣-adrenergic agonist, phenylephrine (PE) (22). Thus, it is possible that in part the coordinated induction of ANP and BNP involves the convergence of intracellular signaling mechanisms upon cis-elements that are conserved between the genes. The present study was undertaken to test this hypothesis by mapping and identifying regions of the BNP 5Ј-FS that are critical for basal and inducible transcription in rat cardiac myocytes.

Mutagenesis
Preparation of Truncated BNP/Luciferase Constructs-A 2.5-kb portion of the rat BNP 5Ј-FS was inserted into a luciferase reporter construct (pGL2; Promega, Madison, WI), as described previously (22). Truncated versions of BNP/luciferase were created either by using native restriction sites, by synthesizing specific PCR primers, or by unidirectional deletion (Erase-a-Base, Promega), starting with BNP-2501GL, as described previously (18).
Preparation of Cluster and Point Mutated BNP/Luciferase Constructs-A series of 6-bp cluster mutations covering the BNP 5Ј-FS between Ϫ103 bp and Ϫ42 bp was created in full-length BNP-2501 by site-directed mutagenesis using Altered Sites (Promega), as described previously for ANP (18). Briefly, a fragment of the BNP 5Ј-FS from Ϫ116 to ϩ80 bp was inserted into the pAlter vector and used as a template. Oligonucleotides containing the desired 6-bp mutation (mutants A-H in Fig. 1) flanked by 12 nucleotides of native BNP sequence on either side, were synthesized and used to prepare the mutant constructs using methods described by the manufacturer. Point mutations in the M-CAT site (⌬M-CAT) were prepared beginning with oligonu-cleotides containing the changed nucleotides flanked on either side by native BNP sequences, extending on the 5Ј side to BNP-116, where there is a SacI site. Using these oligonucleotides as sense primers, and an oligonucleotide complementary to sequences in the 5Ј-region of the luciferase gene as a common antisense primer, PCR was carried out using BNP-116GL as the template. PCR products were then digested with SacI (BNP-116) and BamHI (BNPϩ80), and then cloned into pGL2 to create BNP-116GL possessing cluster mutations A-H and the ⌬M-CAT mutation shown in Fig. 1.
To create multiple mutations (e.g. ⌬M-CAT/GATA(Ϫ95)), a BNP-116GL construct possessing the appropriate GATA-directed point mutation(s), prepared previously (22), was used as the template and the sense primer, possessing the 2-nucleotide M-CAT mutation (see above), was coupled with the luciferase primer to prepare the appropriate PCR-generated product. As above, PCR products were then digested with SacI (BNP-116) and BamHI (BNPϩ80), and then cloned into pGL2 to create BNP-116GL possessing various combinations of mutations in the M-CAT and GATA sites (see Fig. 3, top, for M-CAT/GATA mutations).
To prepare the cluster and point mutations in the 2.5-kb BNP 5Ј-FS, we utilized the SacI sites located 5Ј of Ϫ2501 in the pGL2 multiple cloning site and at Ϫ116 in the BNP 5Ј-FS. Wild type BNP-2501GL was digested with SacI, and the fragment from Ϫ2501 to Ϫ116 was purified. This fragment was then cloned into SacI-digested BNP-116GL constructs possessing the mutations described above to restore the fulllength BNP-2501.
Preparation of M-CAT/BNP-81GL-Two synthetic oligonucleotides (see below) were designed so that after hybridization there is a tandem repeat of two canonical BNP M-CAT core elements (boxed) separated by 5 bp (see Sequence 1).
Some flanking BNP sequences were included such that the repeats represent BNP(Ϫ112 to Ϫ97)/BNP(Ϫ109 to Ϫ97). The lowercase nucleotides at the ends are not native to the rat BNP 5Ј-FS and were added to provide PspA1 sites. The double-stranded synthetic oligonucleotide was ligated into the PspA1 site on the 5Ј boundary of the rat BNP sequences in BNP-81GL. Positive clones were sequenced and some clones contained one insert (i.e. 2XM-CAT/BNP-81GL), and others contained two inserts (i.e. 4XM-CAT/BNP-81GL). All plasmid constructions were verified by dideoxy sequencing.

Cell Culture and Transfections
Myocardial cells were prepared as described (18,22). For transfections, freshly dissociated cells were resuspended at a density of 30 million cells/ml of minimal medium (Dulbecco's modified Eagle's medium/F-12 medium (Life Technologies, Inc.) containing 1 mg/ml bovine serum albumin). For each transfection, 300 l, or 9 million cells, were mixed with 15 or 30 g of BNP/luciferase (test reporter), and 9 g of CMV-␤-galactosidase (normalization reporter), and in some experiments, 45 g of a PKC, Ras, or Raf-1 expression construct (see below). Each 300-l aliquot was then electroporated in a Bio-Rad Gene Pulser at 700 V, 25 microfarads, 100 ohms in a 0.2-cm gap cuvette. This procedure results in an approximate 30% viability (18); accordingly, the 3 million viable cells were plated into fibronectin-coated 35-mm wells, at 1 ϫ 10 6 cells/well, or into 24-mm wells at 0.5 ϫ 10 6 cells/well. Thus, the plasmid concentrations per 10 6 viable cells were 5 g of BNP/ luciferase (10 g in some experiments), 3 g of CMV-␤-galactosidase, and in some experiments, 15 g of PKC, Ras, or Raf-1 test construct.
Transfected cells were maintained in Dulbecco's modified Eagle's medium/F-12 medium supplemented with 10% fetal bovine serum for approximately 14 h after electroporation. The cells were then washed thoroughly, and the medium was replaced with minimal medium. Unless otherwise stated, 24 h later, the medium was again replaced with minimal medium Ϯ 50 M phenylephrine with 1 M propranolol added to block ␤-adrenergic receptors. Luciferase and ␤-galactosidase assays were performed as described (18,22). Luciferase activity was measured for 30 s on a Bio Orbit 1251 Luminometer (Pharmacia Biotech Inc.). Data are expressed as "relative luciferase" ϭ arbitrary integrated luciferase units/␤-gal units, representative of at least three independent experiments performed with two different plasmid preparations, and represent the mean and S.E. of triplicate 35-or 24-mm wells.
To assess the effects of PKC on BNP/luciferase reporter expression, PKC␤⌬OP, which codes for the production of a catalytically inactive form of PKC (23) was used as a control, and PKAC, which codes for the expression of a constitutively active form of PKC-␤ (24), was used as described previously (25). To assess the effects of Ha-Ras on BNP/ luciferase reporter expression, pDCR Ha-Ras V12 , which codes for the production of a constitutively active form of Ha-Ras (26,27) was used with pCEP4 as the empty vector control, as described (28). To assess the effects of Raf-1 kinase on BNP/luciferase reporter expression, pCEP4 ⌬Raf-1:ER, an expression plasmid coding for an estrogen-activated form of Raf-1 kinase (29), was used and pCEP4 was used as the empty control vector. As a second test of the effects of Raf-1 kinase, Raf BXB, which encodes a constitutively active form of c-Raf-1 (30,31), was used.
In all transfection experiments, three identically treated cultures were used for each treatment. Each experiment was replicated at least three times, and the average of three experiments is shown.

Electrophoretic Mobility Shift Assay (EMSA)
EMSA was carried out using nuclei from neonatal rat ventricular tissue obtained as described (18,32). Briefly, probes were prepared by Klenow fragment-mediated filling of the sticky ends of double-stranded oligonucleotides. A typical binding assay contained 20,000 cpm doublestranded probe and 10 g of nuclear extract protein in 1 ϫ binding buffer (10 mM Hepes, pH 7. 0.1 mM EGTA, 5% glycerol, 0.5 mM dithiothreitol). After a 10-min preincubation of extract with 0.1 g of nonspecific competitor (poly(dI-dC), Pharmacia) Ϯ competitor, the probe was added. Binding was allowed to proceed at room temperature for 30 min prior to separation of bound and free probe on a 4% native polyacrylamide gel (29:1 bis/ acrylamide) in 0.5 ϫ Tris borate-EDTA buffer at 4°C at 150 V. DNAprotein complexes were detected by autoradiography. The autoradiograms of some gels in this report were scanned using a Molecular Dynamics Personal Densitometer, and the resulting image was imported to Adobe Photoshop and Claris MacDraw Pro II for final figure preparation.

A Promoter-proximal M-CAT Element Is Critical for BNP
Transcription-In a previous study we cloned and sequenced approximately 2.5 kilobase pairs of the rat BNP 5Ј-FS (22). A search of this sequence revealed the presence of various putative regulatory cis-elements in this region of the rat BNP gene (Fig. 1, top). To begin mapping areas of the gene involved in regulating BNP transcription, a series of reporter constructs was prepared with various lengths of the BNP 5Ј-FS driving luciferase expression. When primary ventricular cardiac myocytes were transfected with these constructs, it was apparent that the removal of about 1.5 kb of 5Ј-FS, down to BNP-535GL, had very little effect on basal or PE-inducible reporter activity ( Fig. 2A). Interestingly, the removal of 137 bp between Ϫ535 and Ϫ398 appeared to result in greater basal and inducible promoter activity, suggesting the presence of repressor elements in this region of the gene. While the physiological role of such repressor elements is unknown, this result is similar to that recently observed in similar truncation analyses of the human BNP gene (33). Further truncation resulted in a gradual decline of promoter activity such that BNP-140GL displayed about 75% of the basal and inducible activities as the full-length construct. Truncation beyond this point seemed to have a more severe effect, such that BNP-116GL possessed only about 15% of the promoter activity observed with BNP-2501GL, while BNP-81GL and BNP-58GL possessed only 8% and 1% of original activity.
A series of 6-bp cluster mutations (see Fig. 1) targeted at promoter proximal regions within the full-length, BNP-2501GL was prepared. Mutants B, C, G, and H each decreased basal and PE-inducible promoter activity by about 30 -50%, mutant D had no effect, while mutants E and F increased PE-inducible promoter activity by as much as 25% (Fig. 2B). Strikingly, however, mutant A, which spans the BNP 5Ј-FS between Ϫ103 to Ϫ98, resulted in a drastic, 98% reduction of reporter activity, implying the presence of a critical regulatory sequence in this region.
A CATTCT, or M-CAT, consensus sequence lies between Ϫ109 and Ϫ102 nucleotides in the rat BNP gene (Fig. 1). CATTCT elements bind a family of skeletal-and cardiac muscle-derived proteins originally named M (muscle)-CAT-binding proteins (see Fig. 1; Ref. 34). M-CAT-binding proteins, which are present in high levels in cardiac myocytes, are related to a family of proteins called transcription enhancement factors (e.g. TEF-1), also known as enhancers of SV40 transcriptional activity, GT-IIC (34). Additionally, M-CAT-binding proteins have been implicated in the transcription of other cardiac muscle genes, such as those for cardiac troponin C, ␣-skeletal actin and ␣-myosin heavy chain, and ␤-myosin heavy chain (34 -38). Accordingly, in the context of 2.5 kb of the BNP 5Ј-FS, a 2-bp double-point mutation was prepared (⌬M-CAT), which was predicted from previous studies to specifically disrupt the binding of M-CAT-binding protein (39). Basal and PE-inducible promoter activity from ⌬M-CAT were reduced by at least 98%, as seen for mutation A (Fig. 2C), emphasizing the absolute requirement for an intact M-CAT element in the promoter-proximal region of the BNP gene.
The GATA family of transcription factors, which have been implicated as regulators of myocardial cell BNP gene expression (22, 40 -42), bind to DNA possessing the consensus, WGTAR sequence (43,44). Since the cluster mutations B and C displayed somewhat reduced basal and PE-inducible promoter activity (Fig. 2B), and since these mutations were located over GATA-binding protein consensus sequences (Fig.  1, lower), point mutations known to disrupt GATA binding were prepared in the context of BNP-2501GL. In agreement with our previous study, which tested similar mutations in reporter constructs possessing only 116 bp of BNP 5Ј-FS (i.e. BNP-116GL) (22), the mutation at GATA(Ϫ95) was of little consequence, while the mutation at GATA(Ϫ84) resulted in an approximate 20% decline in PE-inducibility (Fig. 3). However, when the mutations were combined (i.e. GATA(Ϫ95,Ϫ84)), there was a much greater than expected, 60% reduction in basal and PE-inducible reporter activity (Fig.  3). One explanation for this unexpected reduction is the possibility that these two GATA sites, and perhaps the protein(s) that bind there, are interactive, either with each other, or with other elements. For example, perhaps a GATA-binding protein (BP) must bind to the BNP gene somewhere between about Ϫ100 and Ϫ80 to confer optimal promoter activity. If a GATA-BP binds to the native BNP gene primarily at GATA(Ϫ84), which would leave enough room for an M-CATbinding protein to bind at Ϫ106, one would expect the Ϫ95 mutation to be of little consequence. However, the Ϫ84 mutation might be expected to disrupt GATA-BP binding there, perhaps promoting GATA-BP binding to the alternate, and apparently less effective GATA site at Ϫ95. Then, mutating both Ϫ84 and Ϫ95 would completely disrupt all GATA binding, resulting in the unexpectedly low activity observed in the GATA(Ϫ95,Ϫ84) double mutant.
To test for possible interactions between the GATA and M-CAT elements, the various GATA-directed mutations were combined with the ⌬M-CAT point mutation. As expected, overall promoter activities of the constructs harboring the ⌬M-CAT mutation were severely decreased, probably due to the disruption of the consensus M-CAT site. Additionally, the GATA(Ϫ95) mutation had no additional effect when combined with the ⌬M-CAT mutation; however, the GATA(Ϫ84) and GATA(Ϫ95,Ϫ84) mutations further decreased promoter activity. The abilities of the GATA-directed mutations to decrease promoter activity in the ⌬M-CAT constructs, as well as those possessing the native M-CAT site, suggests that the GATA sites behave independently of the M-CAT site.
A TEF-1-like Protein Binds to the BNP M-CAT Element-Due to its absolute requirement for basal and inducible BNP promoter activity, the properties of the M-CAT element in the BNP 5Ј-FS were studied further. To evaluate whether cardiac nuclear proteins could bind to the promoter-proximal M-CAT site in a manner consistent with the functional consequences of the mutations in the reporter genes, EMSA were carried out. A variety of oligonucleotides were prepared either synthetically or by restriction digestion of the appropriate BNP/luciferase reporters (Fig. 4); the BNP(Ϫ113/(Ϫ95) oligomer was used as a labeled probe and a competitor, while the others were used as competitors.
In the absence of any competitor, cardiac nuclear proteins and the BNP probe formed a single, major complex observed by EMSA (Fig. 5, lane 2). While this complex was disrupted effectively using the unlabeled probe (Fig. 5, lane 3), or an oligonucleotide containing a canonical M-CAT element modeled after that in the chicken troponin gene (Fig. 5, lane 4), the BNP(Ϫ113/(Ϫ95) mutant, which mimics the double point mutation, ⌬M-CAT, was an ineffective competitor (Fig. 5, lane 5). Oligonucleotides mimicking larger stretches of the promoterproximal region of the BNP 5Ј-FS and containing the M-CAT site (e.g. BNP(Ϫ116/(Ϫ52) and BNP(Ϫ116/(Ϫ71)), were also effective competitors (Fig. 5, lanes 6 and 8). However, when  (Fig. 5, lanes 7 and 9), they served as ineffective competitors, consistent with a requirement for the binding of an M-CAT/ TEF-1-related protein to this region of the gene. Oligonucleotides mimicking mutants B and D, which harbored 6-bp clustered changes in regions outside the putative M-CAT region, were effective competitors (Fig. 5, lanes 10 and 11), consistent with the relatively minor roles of these mutations on BNP promoter function. As further controls, it was shown that neither the human c-fos serum response element (SRE), the col-lagense 12-O-tetradecanoylphorbol-13-acetate response element, or the promoter-proximal rat ANF SRE (18) acted as competitors of the shifted complex (Fig. 5, lanes 12-14). Thus, the EMSA analyses are consistent with the required binding of a protein possessing the characteristics of M-CAT-binding protein, or TEF-1, to the putative M-CAT element in the BNP 5Ј-FS for optimal PE-inducible promoter activity.
Roles of PKC, Ras, and Raf-1 in BNP Inducibility-Further experiments were carried out to investigate whether the promoter-proximal sequences, such as the M-CAT element, might participate in BNP promoter activation by PE and by intracellular signals activated by PE. In cardiac myocytes, ␣ 1 -adrener- gic receptors couple through G q to the activation of PKC and to stimulation of the Ras/Raf-1 kinase/extracellular signal-regulated protein kinase (ERK)/MAPK 2 pathways (45)(46)(47)(48)(49)(50)(51)(52). In part, these pathways are believed to participate in mediating the gene induction that results from myocardial cell ␣ 1 -adrenergic receptor activation by PE (15,25,48,(53)(54)(55), although activation of ERK/MAPK alone appears to be insufficient to activate ANP expression (56). Accordingly, we assessed and compared the effects of PE to overexpressed PKC, Ha-Ras, or Raf-1 kinase on various forms of BNP/luciferase reporter genes.
Consistent with the earlier truncation studies, BNP-81GL was virtually inactive, regardless of the stimulus (Fig. 6). PE activated reporter expression from BNP-140GL about 50 -60% as well from BNP-2501GL (Fig. 6A), further supporting the view that sequences between Ϫ140 and Ϫ81 bp could confer significant ␣ 1 -adrenergic agonist inducibility. Basal expression of luciferase was similar for BNP-140GL and BNP-2501GL, but much less for BNP-81GL; the decrease between Ϫ140 and Ϫ81 is most likely due to the absence of the M-CAT and GATA elements from the latter construct (see Fig. 1). Constitutively active PKC strongly and similarly activated both BNP-2501GL and BNP-140GL (Fig. 6B), supporting the view that sequences lying between Ϫ140 and Ϫ81 bp could confer PKC inducibility. Ha-Ras also served as a potent activator of BNP-2501GL; however, it stimulated BNP-140GL only about 40% as well as BNP-2501GL (Fig. 6C), suggesting that in contrast to PKC, sequences residing both distal and proximal to Ϫ140 are re-quired for optimal Ras inducibility. In further support of a role for Ha-Ras was the finding that Raf-1 kinase served as a potent activator of BNP-2501 (Fig. 6D); surprisingly, however, Raf-1 stimulated reporter expression from BNP-140GL relatively poorly, by only about 25% as well as from BNP-2501GL (Fig. 6D).
These results indicated that sequences lying proximal to Ϫ140 bp were responsible for significant levels of inducibility, although they displayed somewhat differential responsiveness to the stimuli, conferring relatively strong induction in response to PE, PKC, and Ha-Ras, but very weak induction in response to Raf-1. To explore further the importance of the promoter-proximal M-CAT element in basal and inducible reporter activity, the BNP-related M-CAT sequences were cloned upstream of position Ϫ81 in BNP-81GL, a construct that normally expresses very low reporter activity. In this context the M-CAT sequences were found to confer a significant recovery of basal as well as PE, PKC, and Ras inducibility to BNP-81GL, with the extent of basal and inducible reporter expression being approximately proportional to the number of M-CAT sites (Fig. 7).
These results supported the hypothesis that the promoterproximal M-CAT element can contribute to BNP inducibility in response to PE, and two of its major effectors, PKC and Ha-Ras. This view is corroborated further by the lack of PE inducibility of ⌬M-CAT/BNP-2501GL (Fig. 2C) and a decrease in Ha-Ras inducibility of this same construct by over 80% (data not shown). However, since Ras is a known activator of Raf-1 in the ERK/MAPK pathway, the low Raf-1 responsiveness of BNP-140GL was inconsistent with this hypothesis, implying that while they were responsive to Ha-Ras, sequences in BNP-140GL were poorly responsive to Raf-1. Accordingly, the ability of Raf-1 kinase to activate the M-CAT/BNP-81GL constructs was evaluated using the ⌬Raf:ER expression construct. Consistent with its poor ability to activate BNP-140GL, Raf-1 did not enhance reporter expression significantly from either 2Xor 4XM-CAT/BNP-81GL (Fig. 8A). As expected, BNP-2501GL was induced significantly by Raf-1, while BNP-140GL was induced by less than 15% compared to BNP-2501GL. The lack of Raf-1 responsiveness supports the view that the PE-, PKCand Ras-mediated increases in luciferase expression from 2Xor 4XM-CAT/BNP-81GL, shown in Fig. 7, represent induction above basal expression; thus, the M-CAT element is apparently required for basal transcription and it can mediate inducible transcription. To confirm the unexpected lack of responsiveness of the M-CAT element to Raf-1, a different Raf-1 expression construct, Raf BXB, was used in a similar experiment. Again, Raf-1 kinase was ineffective as an enhancer of reporter expression from 4XM-CAT/BNP-81GL (Fig. 8B), but like ⌬Raf: ER, it served as a strong inducer of reporter expression from BNP-2501GL and a relatively weak inducer of BNP-140GL. DISCUSSION The results of this study indicate that BNP transcription requires a promoter-proximal M-CAT element that can mediate transcriptional stimulation in response to PE, as well as PKC and Ras, both of which are activated by ␣ 1 -adrenergic agonists. Interestingly, however, the M-CAT element does not appear to contribute to BNP induction in response to Raf-1 kinase, which is also activated by ␣ 1 -adrenergic agonists. Instead, the Raf-1-inducible elements reside distally, between Ϫ2501 and Ϫ140 of the BNP 5Ј-FS, and are yet to be identified. Accordingly, this is the first report to suggest that in addition to conferring inducibility in response to PKC, which was previously shown for the ␤-MHC gene (37), M-CAT elements in muscle-specific genes might also be responsive to Ras-activated signals, but not those involving Raf-1 kinase.

FIG. 5. Electrophoretic mobility shift assay.
A double-stranded oligonucleotide consisting of BNP(Ϫ113 to Ϫ95) was the probe; binding followed by polyacrylamide gel electrophoresis were carried out as described under "Experimental Procedures." Optimal levels of dI-dC were determined over a concentration range of 0 -8 g (not shown). In some reactions unlabeled, double-stranded oligonucleotides were added at a 100-fold molar excess prior to addition of the probe to test their efficacies as competitors. Refer to Fig. 4 for oligonucleotide sequences.
Although M-CAT elements have not been previously shown to mediate Ras inducibility, a large body of information supports a role for M-CAT sequences as determinants of cardiac and skeletal muscle-specific gene expression (35,57,58). Cardiac and striated muscle tissues are particularly enriched in M-CAT binding factors; however, they are also found in many other cell-and tissue types (39, 59 -61). Nonetheless, it is  Fig. 2. Panel B, in addition to the BNP/luciferase and CMV-␤-galactosidase constructs, myocardial cells were also transfected with 15 g of a control plasmid, p⌬OP (Con), which codes for the production of catalytically inactive PKC, or 15 g of pPKAC (PKC), which codes for the production of a constitutively active form of PKC-␤ (24). After 48 h in minimal media, the cultures were extracted and assayed for reporter enzyme activities. Panel C, in addition to the BNP/luciferase and CMV-␤-galactosidase constructs, myocardial cells were also transfected with 15 g of a control plasmid, pCEP4 (Con), which contains no insert, or 15 g of the plasmid Ha-Ras V12 (H-Ras), which codes for the production of a constitutively active form of Ha-Ras (26,27). After 48 h the cultures were extracted and assayed for reporter enzyme activities. Panel D, in addition to the BNP/luciferase and CMV-␤-galactosidase constructs, myocardial cells were also transfected with 15 g of pCEP4/⌬Raf-1:ER (Raf-1), which codes for the production of an believed that in combination with nearby regulatory sequences, such as E-boxes, M-CAT elements contribute to the musclespecific gene expression (62).
In addition to conferring tissue specificity to certain muscle genes, the hormonal inducibility of several cardiac genes is also thought to involve M-CAT elements. For example, a promoterproximal (Ϫ210 bp) M-CAT element in the rat ␤-MHC gene is required for induction by ␣ 1 -adrenergic agonists and PKC (37). A similar element, also located near the promoter of the rat skeletal ␣-actin gene, appears to mediate induction by transforming growth factor-␤ (36), as well as ␣ 1 -adrenergic agonists and PKC (63). Indeed, the abilities of PKC expression constructs to activate M-CAT-containing reporter genes for ␤-MHC (37) and for BNP (present study) lead to the conclusion that PKC may serve as an important regulator of TEF-1-enhanced transcription.
The precise mechanism by which TEFs might enhance BNP transcription in response to PKC activation remains unknown. However, it has been postulated that PKC could directly regulate the transcriptional enhancement effects of TEF (64). For example, TEF-1 possesses known, PKC phosphorylation sites that could alter the ability of the protein to bind to M-CAT elements and thus confer transcriptional activation. However, it is also possible, if not probable, that in comparison to PKC's abilities to activate transcription through SRF or AP-1, it may indirectly activate TEF, perhaps by way of altering the activities of other kinases or phosphatases which would ultimately alter the phosphorylation state of TEF, or closely associated proteins. Another possibility is that PKC and Ras might converge on a single pathway to ultimately effect enhancement of transcription through TEF-1.
Since PKC and Ras can both activate Raf-1 and, thus, ERK/ MAPK, it is tempting to speculate that ERK/MAPK might serve as a downstream effector through which PE can stimulate transcription via TEF-1. However, the finding in this study that the M-CAT element appears to participate in Ras-but not Raf-1-inducible BNP promoter activation suggests a role for Ras-dependent, ERK/MAPK-independent pathways. Consistent with this are recent results demonstrating that PD 098059, a specific MEK inhibitor (65), blocks PE-inducible ERK/MAPK in cardiac myocytes by 80%, but does not block PE-inducible BNP promoter activity, as measured with BNP-2501GL (66). Among the Ras-dependent, ERK/MAPK-independent pathways of possible interest are those converging on the activation of the JNK and p38 members of the MAPK family (67). It is possible, therefore, that p38 MAPK and/or JNK might ultimately affect transcriptional enhancement through TEF-1, or accessory factors. Further, it is reasonable to hypothesize that even though all three MAPK family members might be activated in cardiac myocytes in response to PE, they could differentially effect changes in transcription. Indeed, the MAPKs display some substrate selectivity amongst certain transcription factors. For example, while JNK appears to preferentially phosphorylate c-Jun, ERK/MAPK preferentially phosphorylates c-Myc, p38 MAPK preferentially phosphorylates ATF-2 and ERK/MAPK and p38 MAPK phosphorylate Elk-1 to similar extents (68). Thus, to further dissect the mechanism of Rasinduction through TEF-1, it will be of interest to evaluate the abilities of all three MAPK family members to enhance transcription through BNP promoter-proximal M-CAT element.
In summary, the results from this study add new information to our understanding of the cis-elements and the signaling mechanisms responsible for ␣ 1 -adrenergic agonist-mediated induction of BNP and other cardiac genes. We have found that in comparison to the ␣-skeletal actin and ␤-MHC genes, a promoter-proximal M-CAT element is important for ␣ 1 -adrenergic inducibility of the BNP gene, and that this induction could be mediated at least partly by PKC. The present study has added further to our knowledge of how TEF-1 might mediate cardiac gene expression, demonstrating that the M-CAT element confers Ras-dependent induction, but in a Raf-independent manner; this finding implies the involvement of p38/and/or JNK/ MAPK-mediated events, and/or other Ras-activated pathways not yet clearly identified. The mechanisms by which TEF-1 responds to these, and perhaps other signaling pathways, remain unknown. And while the recent findings that there are multiple forms of the TEFs (69, 70) might also add potential complication to the mechanism, it is also possible that such multiple pathways provides the potential for somewhat independent induction of TEF-responsive genes to match a variety of physiological requirements. Future studies of how the various forms of the TEFs interact with other accessory proteins such as E-box-binding proteins (62) or nearby transcription factors such as GATA-binding proteins (22, 40 -42) will be required to further our understanding of this complex gene induction mechanism.
Acknowledgments-We thank Nichole Arnold for expert technical assistance, and Deanna Hanford, Patrick McDonough, and Dietmar Zechner for critical reading of the manuscript. PKC␤⌬OP and PKAC were gifts from M. Muramatsu (DNAX Research Institute, Palo Alto, CA), pDCR Ha-Ras V12 was a gift from D. Bar-Sagi (SUNY, Stony Brook, NY), pCEP4 ⌬Raf-1:ER was a gift from A. Thorburn (University of Utah, Salt Lake City, UT), and Raf BXB was a gift from U. Rapp (Universitä t Wurzburg, Wurzburg, Germany).