The a-Adrenergic Stimulation of Atrial Natriuretic Factor Expression in Cardiac Myocytes Requires Calcium Influx, Protein Kinase C, and Calmodulin-regulated Pathways*

It has been shown recently that a-adrenergic ago- nists can stimulate atrial natriuretic factor (ANF) expression in ventricular cardiac myocytes; however, little is known about the intracellular signals media- ting this activation. The present study focused on the potential roles of calcium-regulated kinases and cal- cium influx in the a-adrenergic stimulation of ANF gene expression in ventricular myocardial cell cul- tures. Myocardial cells maintained for 48 h in serum-free medium supplemented with phenylephrine (PE) possessed up to 15-fold higher levels of ANF peptide and ANF mRNA than control cells. The removal of PE, or the addition of nifedipine, resulted in a rapid decline in ANF expression, suggesting that the sustained ele- vation of some intracellular messenger (e.g. calcium and/or phospholipid hydrolysis products) was required for the adrenergic response. The calcium channel agonist BAY K 8644 was capable of increasing ANF expression in a nifedipine-sensitive manner; however, unlike PE, it did not stimulate phosphoinositide hydrolysis. The protein kinase C inhibitor, H7, caused an approximate 75% reduction in PE-stimulated ANF expression, but had no effect on BAY K-stimulated expression.

It has been shown recently that a-adrenergic agonists can stimulate atrial natriuretic factor (ANF) expression in ventricular cardiac myocytes; however, little is known about the intracellular signals mediating this activation. The present study focused on the potential roles of calcium-regulated kinases and calcium influx in the a-adrenergic stimulation of ANF gene expression in ventricular myocardial cell cultures. Myocardial cells maintained for 48 h in serumfree medium supplemented with phenylephrine (PE) possessed up to 15-fold higher levels of ANF peptide and ANF mRNA than control cells. The removal of PE, or the addition of nifedipine, resulted in a rapid decline in ANF expression, suggesting that the sustained elevation of some intracellular messenger (e.g. calcium and/or phospholipid hydrolysis products) was required for the adrenergic response. The calcium channel agonist BAY K 8644 was capable of increasing ANF expression in a nifedipine-sensitive manner; however, unlike PE, it did not stimulate phosphoinositide hydrolysis. The protein kinase C inhibitor, H7, caused an approximate 75% reduction in PE-stimulated ANF expression, but had no effect on BAY K-stimulated expression. W7, a calcium/calmodulin inhibitor, completely blocked the effects of both PE and BAY K 8644. The addition of either H7 or W 7 24 h after the PE addition resulted in a decline of ANF expression.
These results indicate that a-adrenergic agonists augment ANF gene expression through at least two pathways, one that is H7-sensitive, perhaps involving the sustained activation of protein kinase C, and the other that is W7-sensitive, perhaps involving the sustained activation of calmodulin-regulated kinases. Further, it appears that BAY K 8644-mediated increases in ANF expression are independent of protein kinase C activation and dependent on calmodulinregulated events. ). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Supported by Postdoctoral Fellowship Grant 89-99 from the American Heart Association, CA Affiliate.
1 Established Investigator of the American Heart Association (Grant 86-173). To whom correspondence and reprint requests should be addressed.
Although several studies have shown that a variety of hormones can alter ANF' gene expression (Gardner et al., 1987;Argentin et al., 1987;Matsubara et al., 1987a, 198713;LaPointe et al., 1988;Seidman et al., 1988;Fukuda et al., 1989;Wu et al., 1989;Shubeita et al., 1990), very little is known about the intracellular signaling mechanisms through which such hormonal regulation occurs. a-Adrenergic agonists, which are among the most effective stimulators of ANF gene expression (Knowlton et al., 1991), have been shown to activate phosphoinositide (PI) hydrolysis leading to the IP3mediated mobilization of intracellular calcium and DAG/ Ca2+-mediated activation of protein kinase C (Nishizuka, 1986). Although al-adrenergic agonists stimulate PI hydrolysis in cardiac myocytes (Brown et al., Steinberg et al., 1989), the functions of IPS and DAG in regulating myocyte function have never been clarified. Indeed, the classic role of IP3 in mediating the mobilization of intracellular calcium over relatively long periods of time (e.g. seconds to minutes) is not compatible with the rapid calcium transients (e.g. 0.2 s) that normally drive contractions in cardiac myocytes (Bals et al., 1990). Moreover, calcium transients in cardiac myocytes are not known to be dependent on the formation of inositol phosphates.

ANF Expression in
Myocardial Cells 15911 could contribute to regulating contractility (Mochly-Rosen et al., 1990). If PKC translocation, and presumed activation, is all that is needed for the al-adrenergic induction of the ANF gene, then a short term treatment with the agonist phenylephrine (10 to 15 min) should be sufficient to lead to the maximum level of ANF expression. However, we found that the aladrenergic agonist must be continually present in order for continued induction of the gene, indicating that the sustained elevation of intracellular messengers, along with the continual activation of effector enzymes, such as PKC, must be required. In pursuing this hypothesis we have shown in the present study that a,-adrenergic agonists influence ANF expression through at least two pathways; one requires PI hydrolysis with the apparent sustained activation of PKC, and the other involves a calmodulin-regulated step, such as CaM kinase I1 activation.

MATERIALS AND METHODS
Myocardial Cell Cultures-Ventricular myocardial cell cultures were prepared from 1-4-day-old neonatal rat pups and maintained as previously described (Shields et al., 1988). For most experiments, the dissociated cells were plated into fibronectin-coated 16-mm culture dishes at approximately 5 X IO5 cells/culture and maintained for 4 h in Dulbecco's modified Eagle's medium (DMEM)/F-12 supplemented with 10% fetal bovine serum. The cultures were then switched to serum-free DMEM/F-12 containing 1 mg/ml BSA, and, after 14 h, various test compounds were added. Cultures were then maintained for up to 5 days after the initial dissociation, although most experiments were carried out within 48 h after plating. For ANF message analysis, the cells were plated as described above, except into 35-mm dishes at a density of about 2.2 X lo6 cells/culture.
Measurements ofANF Message and Peptide-Total RNA, prepared by acid/guanidinium thiocyanate/phenol/chloroform extraction (Chomczynski and Sacchi, 1987), was fractionated on agarose gels (2 pg per lane) followed by transfer to a Nytran membrane. The membranes were probed with [a-32P]dCTP-labeled rat ANF cDNA (Glembotski et al., 1986), or with [c~-~*P]dCTP-labeled mouse 28 S rRNA cDNA (Iwaki et al., 1990). After washing at high stringency (0.1 X SSPE, 0.1% SDS, 65 "C), the extent of hybridization was determined by scanning laser densitometry (LKB Ultroscan XL) of the resultant autoradiographs, and the ANF signal was normalized to the 28 S rRNA signal to correct for loading and/or transfer efficiencies. The rate of ANF secretion from the cells was assessed at various times after switching to media containing test compounds. The maintenance medium was removed from the cultures and replaced by fresh medium, but otherwise was identical with the maintenance medium for a particular culture set. The quantity of immunoreactive ANF that had accumulated in the medium over a 2-h incubation period was then assessed by radioimmunoassay as described (Shields et al., 1988).
Protein Kinase C Enzyme Assays-Soluble adult rat ventricular PKC enzyme activity was assayed essentially as described (Fearon and Tashjian, 1985). Briefly, ventricular tissue was homogenized in Buffer A (20 mM Tris, pH 7.4, 5 mM EDTA, 5 mM EGTA, 10 pg/ml leupeptin, 0.3 mg/ml phenylmethylsulfonyl fluoride, 30 mM 2-mercaptoethanol), and a 100,000 X g supernatant was applied to a DE52 column pre-equilibrated with Buffer A. Protein kinase C was eluted with Buffer A containing 200 mM NaCl. Enzyme activity in the NaCl eluant was measured in a final volume of 130 pl with 2 pmol of Tris-HCI, 20 mM M e acetate, 100 p~ calcium, 1 mg/ml Histone 111-S, 96 pg/ml phosphatidylserine, 9.6 pg/ml diolein, 25 p~ [y-"'P]ATP (40-60 cpm/pmol), and 30 pl of enzyme with the indicated levels of potential inhibitors, H7, HA-1004, and W7. The reaction was started by the addition of ATP, allowed to proceed for 7 min at 30 "C, and quenched with 3 ml of 25% trichloroacetic acid. Following the addition of 100 pg of BSA, the samples were placed on ice for 20 min to facilitate precipitation. The precipitated proteins were collected by vacuum filtration over glass-fiber filters and washed with 5% trichloroacetic acid, 1 M NaH,PO, and counted. Protein kinase C activity is defined as the difference in histone phosphorylation in the presence and absence of added lipid, as previously described (Nishizuka, 1986). The percentage of remaining activity in samples containing either H7, W7, or HA-1004 were compared to controls that were described as "maximal PKC activity." Phosphoinositide Hydrolysis and Diacylglycerol Measurements-Membrane phosphoinositides and DAG measurements were performed essentially as previously described . Briefly, ventricular myocytes were plated in 10% fetal bovine serumcontaining medium that was supplemented with [3H]inositol (10-20 Ci/mmol) at 15 pCi/ml of medium. The label-containing medium was removed approximately 18 h later, the cultures were washed, and serum-free DMEM/F-12 (1:l with 1 mg/ml BSA) containing various test compounds and 10 mM LiCl was added to each culture. After an incubation of 40 min, the medium was removed, and 10% trichloroacetic acid was added to extract the cells. After scraping the cell debris from the dishes, samples were centrifuged, and labeled inositol phosphate, which was isolated from the supernatant following ether extraction, was analyzed by anion exchange chromatography as described (Jones et al., 1988). For DAG determinations, the precipitate was extracted further with methanol as described (Bligh and Dyer, 1959). The lipid-containing fractions were then assayed for total DAG as previously described (Preiss et al., 1987).

RESULTS
Medium ANF as a Measure of ANF Gene Expression-Primary neonatal ventricular myocytes release ANF constitutively such that little of the peptide is stored within the cells (Bloch et al., 1986). Thus, changes in the level of ANF in the medium of ventricular myocyte cultures provide a reasonable estimate of changes in ANF mRNA levels. In many experiments discussed in this paper, the effects of a treatment on ANF expression were assessed initially by measuring changes in the levels of medium ANF. This was then followed with studies evaluating the effects of specific treatments on ANF mRNA and, in some cases, cellular levels of ANF. It was consistently found that a treatment that caused a change in medium ANF levels coordinately changed cellular ANF as well as ANF mRNA.
Time Course of PE Effects of ANF Production-Maintaining ventricular myocyte cultures in serum-free medium containing phenylephrine (PE) resulted in increased rates of ANF accumulation in the medium (Fig. lA). Within 6 h of exposure to PE, the appearance of ANF in the medium increased by about 2to &fold over control cultures. This differential increased steadily to a final value of about 15-fold over control cultures after 48 h of PE exposure. Northern analyses showed that ANF message levels had increased maximally (3-to 4fold over control) after about 12 h and remained at these levels throughout 48 h of PE exposure (Fig. 1B). The aagonist-mediated changes in message levels were shown to have a specific effect on ANF expression, since all ANF message values were normalized to levels of 28 S ribosomal RNA levels, as previously described . When cultures maintained in PE for 24 h were washed free of the agonist medium, ANF began to drop so that within 6 h it was only about half that observed in cultures remaining in PE (Fig. lA). This trend continued so that 12 h after PE removal, medium ANF levels had declined to nearly control values. Analyses of culture extracts indicated that cellular ANF levels had decreased by 66 and 80% within 6 h and 12 h of PE removal, respectively, a time course similar to the decline in ANF mRNA and medium ANF. The finding that medium ANF levels reflected cellular levels indicated that PE removal affected total ANF production, and not simply secretion. This is consistent with previous findings that ventricular myocytes release ANF constitutively through a nonregulated pathway (Bloch et al., 1986). In accord with the decline in both cellular and medium ANF levels, the removal of PE also resulted in a decline in ANF mRNA (Fig. 1B). Interestingly, the kinetics of this decrease appeared to be slower than that for the peptide, which is perhaps consistent with a transcript half-life of about 17-24 h, as previously determined (Gardner et al., 1987). Although it is clear that both peptide and mRNA levels decrease upon agonist removal, the more rapid decline in peptide levels suggests that ANF production may be regulated at both the translational and transcriptional levels.
These results indicated that PE-mediated elevation of an intracellular messenger, such as IP3 and/or calcium, was important for maintaining ANF expression. The effects of hormones on intracellular calcium in cardiac myocytes have been studied very little, perhaps because of the rapid calcium transients that continually take place in these cells during the contractile cycle. Since it has been inferred from two previous studies that a-adrenergic agonists can affect slight increases in intracellular calcium in cardiac myocytes (Endoh andBlinks, 1988 Fedida et al., 1989), and since it has recently been shown that ANF mRNA and peptide levels in primary myocardial cultures decrease when extracellular calcium is lowered (LaPointe et al., 1990), it was of interest to evaluate whether the mechanism of PE-stimulated ANF expression, at least in part, was dependent on calcium influx.
Effects of BAY K 8644-Treatment of cultures with the calcium channel agonist BAY K 8644 alone resulted in an increase in ANF production of approximately 10-fold over control cultures (Fig. 2 A , see 0  response experiment demonstrated a half-maximal dose of about 0.1 pM with the maximum achieved at 1 pM and maintaining at that level to 10 p~ (not shown). This is in accord with the effects of BAY K 8644 on other peptide systems and its affinity for L-type calcium channels (Enyeart and Hinkle, 1984;Heisler, 1985;Stojilkovic et al., 1988). The maximum response produced by BAY K 8644 amounted to about 20-25% of the response obtained with maximal levels of PE. Interestingly, BAY K 8644 augmented PE-stimulated ANF production at all levels of the a-adrenergic agonist (Fig.

A ) .
Since either PE or BAY K 8644 caused increases in ANF production, and PE has been shown to increase PI hydrolysis in myocardial cultures (Brown et al., 1985;Steinberg et al., 1989), it was of interest to determine whether BAY K 8644 could stimulate PI hydrolysis. As expected, PE increased PI hydrolysis by 8-to 10-fold, with a maximal effect occurring at 10 p~ PE (Fig. 2 B ) . In contrast, BAY K 8644 had no effect on PI hydrolysis, nor did it augment the PI hydrolysis observed in response to PE. In fact, a slight, although statistically insignificant, decrease in PI hydrolysis was observed when BAY K 8644 was added to PE. Thus, the increases in ANF levels in response to BAY K 8644 occurred in the absence of any detectable changes in the rate of PI hydrolysis or DAG formation (not shown) over control. This suggested that BAY K 8644 could augment ANF expression through increases in calcium influx. However, the lower efficacy of BAY K 8644 compared to PE indicated that an a-agonist-mediated event, perhaps related to increased PI hydrolysis, was required for maximal augmentation of ANF expression.
Effects of Nifedipine-Due to the known effects of BAY K 8644 on calcium channel activity, it was of interest to evaluate the effects of nifedipine, a calcium channel antagonist, on ANF expression. As expected, nifedipine inhibited BAY K 8644-dependent increases in ANF peptide and mRNA levels in the cultures (Fig. 3, A and C). Nifedipine also inhibited PE-induced increases in ANF peptide (Fig. 3A) and mRNA (Fig. 3C); however, it did not affect a-adrenergic agoniststimulated PI hydrolysis (Fig. 3B). Dose-response experiments showed that the effects of nifedipine on each of these functions was half-maximal at about 0.05 p~ (not shown), in accord with its effects on other functions regulated by the activity of L-type calcium channels (Wolfe and Brostrom, 1986;Stojilkovic et al., 1988). Thus, it appeared as though the ability of PE to influence ANF expression was primarily dependent on calcium influx.
Effects of H7, W7, and HA-1004-In order to begin to address the potential functions of protein kinases, studies utilizing membrane-permeable enzyme inhibitors were per-  formed. Three inhibitors were chosen based on their enzyme selectivities. The membrane-permeable isoquinoline derivative, H7, which inhibits PKC and protein kinase A with similar potencies, and HA-1004, which is about 20-fold more selective for protein kinase A than for PKC inhibition, were used (Hidaka et al., 1984). In addition to inhibitors of PKC and protein kinase A, a compound that would inhibit calmodulin-regulated kinases was also employed. W7, a relatively selective inhibitor of calmodulin-regulated reactions such as myosin light chain kinase and CaM/kinase I1 (Ki = 30-60 p~) is ineffective as an inhibitor of PKC or protein kinase A (e.g. KipKC > 100 p~; Tanaka and Hidaka, 1980;Ito et al., 1986). Each of these compounds was tested for the ability to inhibit rat ventricular tissue PKC enzyme activity in vitro. In these analyses, H7 produced half-maximal inhibition at 25 pM and up to 80% inhibition at 100 pM, while HA-1004 at up to 100 pM caused only a 29% decrease in PKC enzyme activity (Table I). Like HA-1004, W7 was a poor PKC inhibitor, resulting in a loss of only 3% of enzyme activity at 100 p M . However, based on previous studies (Ito et al., 1986;Davis and Wilson, 1989;Nguyen et aL, 1990), it was expected that W7 could completely inhibit calmodulin-dependent reactions when used at approximately 10 to 30 pM.
In dose-response studies, it was shown that HA-1004 inhibited PE-mediated ANF expression by only 38% at levels of 50 p~ or more (Fig. a), as expected if protein kinase A activation has little involvement in a-adrenergic-stimulated ANF expression. However, H7 inhibited PE-stimulated ANF expression at a half-maximal concentration of between 10 and 20 p~, with a maximal effect of about 80% inhibition at 50 p~ (Fig. 4B). The dose-response range for H7 inhibition of PE-stimulated ANF expression correlated well with that observed for inhibition of PKC (see Table I). Although H7 inhibited PE-stimulated ANF expression, it had no inhibitory effect on PE-stimulated PI hydrolysis or DAG formation (Table 11). Thus, H7-mediated inhibition of the PE effect appeared to be distal to a-adrenergic receptor activation of phospholipase C, most likely at the level of a protein kinase such as PKC. In contrast to H7, W7 was more potent at inhibiting PE-stimulated ANF expression, displaying a halfmaximal concentration of about 2 pM and a maximal effect of greater than 95% inhibition at 10 p~ (Fig. 4C). H7 (100 PM) and W7 (10 p~) also diminished ANF message levels (ANF mRNA/28 S ratios for PE = 4.75 f 0.35; PE + H7 = Homogenates of adult rat ventricular tissue were prepared, and the soluble PKC was partially purified from each fraction by ion exchange chromatography, as described under "Materials and Methods." Enzyme assays were performed in the presence of the indicated concentrations of either H7, HA-1004, or W7. The phospholipid-stimulatable enzyme activity determined in the absence of any inhibitor was set at 100% activity to which all other values were compared. This experiment was carried out using three different preparations of soluble rat ventricular PKC. In early experiments, it was shown that the findings with the soluble enzyme were essentially identical with those of the Triton-extracted membrane-associated enzyme. The average S.E. for each activity measurement was 15%.   Fig. 2B. At this time, the cultures were washed and the medium was replaced with serum-free medium containing either no further additions (Min), 100 p~ phenylephrine (PE), or 100 PM PE and 100 p~ H7 (PE/H7). Following a 60-min incubation, the cells were harvested, and the extracts were analyzed for DAG and labeled inositol phosphates, as described under "Materials and Methods." Each point for each treatment was analyzed using four identical cultures; errors represent the S.E. Statistical analyses were performed by one-way analysis of variance with Newman-Keulspost hoc analysis of variance. 2.10 f 0.07; PE + W7 = 0.34 f 0.05). In order to evaluate whether the sustained activation of kinases was involved in PE-stimulated ANF expression, H7 or W7 was added 24 h after the addition of PE. The late addition of H7 or W7 resulted in a 40% and 55% inhibition of PE-stimulated ANF expression, respectively (not shown), thus indicating the importance of sustained kinase activation.
To determine whether the H7-insensitive component of PE-stimulated ANF expression was due to calcium influx, further experiments were carried out testing the efficacy of H7 against BAY K 8644. Interestingly, BAY K 8644-mediated ANF expression was insensitive to H7; however, both PEand BAY K 8644-mediated responses were inhibited completely by W7 (Fig. 5). This indicated that two signaling pathways mediate the effects of PE, one being PKC-dependent and the other calmodulin-dependent, while BAY K 8644 operates through only the calmodulin-dependent pathway.

DISCUSSION
In the present study, we have investigated the potential roles of calcium influx, protein kinases, and calmodulin in aadrenergic-mediated increases of ANF expression in primary neonatal rat ventricular myocardial cell cultures. We found that the continual occupation of a-adrenergic receptors with an agonist is required for maximal activation of ANF expression in ventricular myocardial cultures, indicating that the presence or activities of some intracellular signal(s) (either messengers and/or kinases) must be sustained in order to achieve the maximal response. It appears as though the stimulation of ANF expression requires calcium influx and a CaMregulated step(s) (Fig. 6, Required Pathway), while PKC activation may play a secondary role by augmenting the calcium influx/CaM-dependent pathway (Fig. 6, Enhancement Puthway). In support of this model is the finding that PE produced an increase in the rate of PIPz hydrolysis and DAG formation in myocardial cells in the presence or absence of calcium influx; however, blocking calcium influx abolished any effects of a-adrenergic agonists on ANF production. Moreover, when calcium influx was stimulated with BAY K 8644, there was no increase in PIPz hydrolysis, but there was a significant increase in ANF expression. Interestingly, PE was consistently more efficacious than BAY K at stimulating ANF expression. Therefore, it seems probable that the enhanced  Sukhatme et al., 1988 Henrich andSimpson, 1988;LaPointe et al., 1990;Mochly-Rosen et al., 1990) as they pertain to the mechanisms through which a-adrenergic agonists and extracellular calcium could influence ANF message and peptide levels in primary neonatal rat ventricular cultures. The central and Required Pathway involved in the regulation of ANF message and, thus, peptide production, is comprised of several pathways that contribute to the regulation of cytosolic free calcium levels (Ca",). Accordingly, Caz+i could arise from extracellular (Ca2+o) and/or sarcoplasmic reticulum ( S R ) pools of calcium. In cardiac myocytes, the influx of calcium through voltage-operated calcium channels (VOCC) is thought to serve as a trigger for the release of sarcoplasmic reticulum-derived calcium during the contractile cycle (Fabiato, 1985;Carafoli, 1985). Thus, it is possible that a major portion of the intracellular free calcium is derived from SR sources and is resequestered within the SR at the end of the contractile cycle. Due to dependence on calcium influx, the release of SR calcium would he inhibited by nifedipine and stimulated by BAY K 8644. The Enhancement Pathway involves PKC activation and may not be required for increases in ANF production mediated by calcium alone.
efficacy of PE on ANF production is related to PI hydrolysis and likely involves the activation of PKC. In contrast to PE, BAY K-mediated ANF expression was not inhibited by H7; however, in comparison to PE, its effects could be completely blocked with W7 at levels known not to affect PKC in bovine heart (see Table I1 and Schatzman et al., 1983).
The precise intracellular mechanism through which aadrenergic agonists and calcium channel activators stimulate ANF expression remains to be elucidated. Although such a mechanism could involve regulation at the transcriptional and the translational levels, it seems certain that at least transcriptional regulation is involved. Perhaps the stimulation of calcium-activated kinases, such as PKC and/or CaM kinase 11, regulates ANF gene transcriptional activity by altering the abilities of specific transcriptional factors to bind to relevant sites on the ANF promotor. An analysis of the 5'flanking sequence of the rat ANF gene has indicated the presence of consensus sequences for the binding of several well characterized transcriptional regulatory proteins. For example, in the region from -495 to -487 bp in the rat ANF promotor (Seidman et al., 1988), there is a consensus sequence for AP-1 (fos/jun) binding, while the region from -461 to 452 bp contains a consensus sequence for the binding of another, more recently discovered transcriptional regulatory protein, Egr-1 (Sukhatme et al., 1988). Among the various hormonal stimuli that activate fosljun heterodimer formation and, thus, the binding to and regulation of certain AP-1-containing genes, are those that stimulate PKC (Chiu et al,, 1988). Recent studies have indicated that in primary ventricular myocardial cells, a-adrenergic stimulation results in increases in fos, jun, and Egr-1 message levels (Iwaki et al., 1990), substantiating the hypothesis that PKC-regulated transcriptional factors are likely involved in a-adrenergic-stimulated ANF gene expression. Furthermore, the expression of constitutively activated forms of PKC in cardiac myocytes leads to the transcriptional activation of the ANF gene.' It is also possible that calcium and calmodulin augment ANF expression through mechanisms not requiring PKC. Recent studies have indicated that the transcription of a number of other genes is regulated by intracellular calcium in a calmodulin-dependent fashion. These include genes for the heat shock protein family (Resendez et al., 1988), prolactin (Davis, 1990), plasminogen activator (Ziegler et al., 1990), enkephalin (Nguyen et al., 1990), proopiomelanocortin (Loeffler et al., 1989), and several immediate-early genes, such as those for fos and jun (Morgan and Curran, 1988). In several of these cases, regions of the relevant 5'-flanking sequences which are responsible for calcium/calmodulin responsiveness have been isolated and have been shown to be distinct from other known transcriptional regulatory regions. For example, the calcium-response element in the prolactin gene contains all the information required for a full response to calcium (induced with BAY K 8644) and is considered to be distinct from, yet near a phorbol ester/epidermal growth factor responsive element (Jackson and Bancroft, 1988). Interestingly, similar to ANF expression, prolactin expression in GH3 cells can be enhanced by the activation of a phospholipase Ccoupled hormone receptor (thyrotropin-releasing hormone) or through BAY K 8644-mediated increases in calcium influx (Hinkle et al., 1988). Increases in prolactin expression in response to either compound are sensitive to nifedipine and calmodulin inhibitors such as W7 (Davis and Wilson, 1989). In PC12 cells, it has been shown that c-fos can be induced by calcium influx through L-type calcium channels in response to either depolarization or BAY K 8644 (Morgan and Curran, 1986). As with prolactin and proenkephalin gene transcription, it has been proposed that since the calcium-dependent augmentation of c-fos expression occurs in a W7-sensitive manner, it is also likely to involve the calmodulin-mediated activation of a kinase which phosphorylates a transcription factor and then stimulates expression (Morgan and Curran, 1988).
The calcium response element in the c-fos promotor lies about 60 bases upstream of the start site and is distinct from the regulatory elements for polypeptide growth factors, serum, and phorbol esters (Sheng et al., 1988). Since only a few calcium-response elements have been characterized, a consensus sequence has not yet been established. However, in several cases, it has been shown that the calcium-response elements are indistinguishable from cyclic AMP response elements (CRE) (Sheng et al., 1988;Nguyen et al., 1990). Further, it has been shown that the CRE binding protein, CREB, is phosphorylated in response to either increases in calcium or cAMP in such a way that it activates transcription (Sheng et al., 1990(Sheng et al., , 1991. Interestingly, between -603 and -596 bp in the rat ANF promotor is a TGACTTCA sequence (Seidman et al., 1988) that is similar to the CRE consensus (TGACGTCA, Montminy et al., 1986). Although extensive studies have not yet been carried out, preliminary experiments in our laboratory have indicated that in primary cardiac myocytes cAMP does not stimulate ANF expression. Perhaps the CRE sequence in the ANF promotor serves as a calciumresponse element.
In summary, calcium is an important determinant of ANF expression in ventricular myocytes. The nifedipine sensitivity H. E. Shubeita, E. Martinson, K. R. Chien, and J. H. Brown, manuscript in preparation.
of PE-and BAY K 8644-stimulated ANF expression implies that the optimal effects of these compounds depend on calcium influx through L-type calcium channels, indicating that calcium-sensitive kinases may ultimately regulate transcriptional activity. The partial H7 sensitivity of PE-stimulated ANF expression suggests that PKC activation plays an important signaling role. Additionally, since BAY K 8644-enhanced ANF expression is insensitive to H7, but sensitive to W7, it is apparent that calmodulin is also involved. Preliminary studies from our laboratory have shown that the CaM kinase 11-specific inhibitor, KN-62 (a gift from H. Hidaka, Nagoya University School of Medicine, Showa-ku Nagoya, Japan), is a potent blocker of PE-stimulated ANF expression, thus describing at least one role for calmodulin in this process. Future studies evaluating further the potential roles of PKC and calmodulin-regulated kinases, as well as the regulation of nuclear transcription factor levels by calcium influx and the mapping of calcium-sensitive promotor regions, will be of interest in unraveling the mechanisms of hormone-regulated ANF production in cardiac myocytes.