Ser1928 is a common site for Cav1.2 phosphorylation by protein kinase C isoforms.

Voltage-dependent Ca(2+) channel (Ca(v)1.2, L-type Ca(2+) channel) function is highly regulated by hormones and neurotransmitters in large part through the activation of kinases and phosphatases. Regulation of Ca(v)1.2 by protein kinase C (PKC) is of significant physiologic importance, mediating, in part, the cardiac response to hormonal regulation. Although PKC has been reported to mediate activation and/or inhibition of Ca(v)1.2 function, the molecular mechanisms mediating the response have not been definitively elucidated. We show that PKC forms a macromolecular complex with the alpha(1c) subunit of Ca(v)1.2 through direct interaction with the C terminus. This interaction leads to phosphorylation of the channel in response to activators of PKC. We identify Ser(1928) as the residue that is phosphorylated by PKC in vitro and in vivo. Ser(1928) has been identified previously as the site mediating, in part, the protein kinase A up-regulation of channel activity. Thus, the protein kinase A and PKC signaling pathways converge on the Ca(v)1.2 complex at Ser(1928) to increase channel activity. Our results identify two mechanisms leading to regulation of Ca(v)1.2 activity by PKC: pre-association of the channel with PKC isoforms and phosphorylation of specific sites within the alpha(1c) subunit.

The influx of Ca 2ϩ through Ca v 1.2 is essential for activation of excitation-contraction coupling in the heart (1). Ca 2ϩ influx also contributes to the plateau phase of the cardiac action potential, pacemaker activity in nodal cells, and the modulation of gene expression (2). A variety of diseases such as atrial fibrillation, heart failure, and ischemic heart disease have been associated with alterations in Ca v 1.2 density and/or function (3,4). The activation of members of the protein kinase C (PKC) 1 family plays an important role in the signal transduction pathways within the cardiac myocyte leading to modulation of con-traction and cell phenotype (5). Regulation of Ca v 1.2 activity by PKC signaling pathways has been well established, with several studies demonstrating that activators of PKC increase Ca 2ϩ channel currents in cardiac and smooth muscle (6 -11). PKC is a serine/threonine kinase consisting of an N-terminal regulatory region and a C-terminal catalytic region and is activated by diacylglycerol (DAG) produced in response to hormones such as ␣-adrenergic agonists, angiotensin II, and endothelin-1 (12, 13). Three PKC subgroups have been identified and functionally distinguished: conventional PKC (cPKC), novel PKC (nPKC) and atypical PKC (aPKC) (14,15).
The molecular basis for Ca v 1.2 regulation by PKC has been widely studied, but remains incompletely elucidated and controversial. The cardiac Ca v 1.2 ␣ 1c and ␤ 2a subunits are phosphorylated by PKC on unidentified sites in vitro (16). Electrophysiologic studies of recombinant channels have attempted to identify specific mechanisms that mediate PKC regulation, but have yielded conflicting information, potentially due to differences in experimental conditions and/or sequences of ␣ 1c /␤ subunits (17)(18)(19)(20).
Several studies have suggested that the N terminus of ␣ 1c is important for PKC up-regulation of channel function. Deletion of the first 46 or 139 amino acid residues of rabbit heart ␣ 1c causes a 5-10-fold increase in whole cell current in oocytes (21,22), without affecting channel density. These findings suggest that the N terminus provides an inhibitory control, which can be relieved after PKC phosphorylation, potentially of a remote phosphorylation site(s) (23). To elucidate the mechanism(s) through which PKC isoforms regulate channel activity, we have utilized glutathione S-transferase (GST) fusion proteins to screen potential phosphorylation sites. Here, we demonstrate that the protein kinase A (PKA) phosphorylation site (Ser 1928 ) on the ␣ 1c subunit (24), which mediates, in part, Ca v 1.2 activation (25), is a substrate for PKC in vitro and in vivo. The phosphorylation of Ser 1928 is mediated, in part, through a direct association between the ␣ 1c subunit and PKC isoforms. The results suggest that both PKA and PKC converge on Ser 1928 of the ␣ 1c subunit to mediate phosphorylation-dependent regulation of Ca 2ϩ influx.

EXPERIMENTAL PROCEDURES
Preparation of Adult Heart and Brain and Human Embryonic Kidney (HEK) 293 Cell Extracts/Membranes-Membranes were prepared from rat ventricular tissue as described previously (26). All preparative procedures were performed at Ϫ4°C. Rat ventricles were minced and homogenized in 10 mM Tris maleate (pH 6.8) containing 1 mM benzamidine, 5 g/ml pepstatin, 5 g/ml leupeptin, 1 mM calpain I inhibitor, 1 g/ml aprotinin, and 1 mM Pefabloc SC. Homogenates were spun for 20 min at 4000 ϫ g, filtered through gauze, and centrifuged at 8000 ϫ g for 20 min. Supernatants were centrifuged at 40,000 ϫ g for 30 min. Pellets were resuspended in 2 ml of 10 mM Tris maleate containing 0.3 M sucrose and 0.9% NaCl. Membranes were stored at Ϫ80°C. Rat brain extracts were prepared in phosphate-buffered saline containing 1% (v/v) Triton X-100, one Complete mini-tablet (Roche Applied Science), 44 M calpain I inhibitor, 18 M calpain II inhibitor, and 200 M phenylmethylsulfonyl fluoride as described previously (27). Insoluble material was removed by centrifugation (twice at 14,000 rpm for 10 min), and supernatants were collected. HEK cells were washed with 1ϫ phosphate-buffered saline and lysed in 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 1% Triton X-100 containing protease inhibitors. The lysates were centrifuged at 14,000 ϫ g for 10 min to clarify supernatants.
Isolation of Neonatal Cardiac Myocytes-Cardiomyocytes were isolated from 2-day-old Wistar rat hearts by trypsin dispersion using a differential attachment procedure to enrich for cardiomyocytes, followed by irradiation as described previously (28). Cells were plated on protamine sulfate-coated culture dishes at a density of 5 ϫ 10 6 cells/ 100-mm dish. Experiments were performed on cultures grown for 5 days in minimal essential medium (Invitrogen) supplemented with 10% fetal calf serum and then serum-deprived for the subsequent 24 h. Neonatal cardiac myocytes maintained in cultures were washed with 1ϫ phosphate-buffered saline and scraped into 10 mM Tris maleate (pH 6.8) containing a protease and phosphatase inhibitor mixture (Calbiochem). Membranes were prepared according to methods described previously (29). The pellets were suspended in 100 l of 1ϫ sample buffer.
cDNA Clones and Site-directed Mutagenesis-The rabbit ␣ 1c subunit (NCBI accession number X15539) and the rabbit ␤ 2a subunit (NCBI X64297) were ligated into pcDNA3 (Invitrogen) for expression. Sitedirected mutagenesis was done using the QuikChange XL kit (Stratagene). All clones were sequenced on both strands prior to use. Transfections into HEK293 cells were performed using Lipofectamine 2000 (Invitrogen).
Preparation of Phospho-specific Antibody-The phospho-Ser 1928 -specific antibody was prepared at Zymed Laboratories Inc. utilizing the peptide NH 2 -LGRRApSFHLECLK-COOH as described previously (27). The sensitivity and specificity of the phospho-specific antibody were confirmed utilizing GST fusion proteins/enzyme-linked immunosorbent assay.
Immunoprecipitations and Kinase Assays-Immunoprecipitations were performed overnight in modified radioimmune precipitation assay buffer consisting of 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.25% Triton X-100, Complete mini-tablets (one/7 ml), and 200 M phenylmethylsulfonyl fluoride using 2 g of anti-␣ 1c antibody (ACC-003, Alomone Laboratories). Immune complexes were collected using protein A (Amersham Biosciences) for 2 h, followed by extensive washing. All immunoprecipitations included negative controls (peptide-blocked, preimmune serum, antibody alone). Additional antibodies included horseradish peroxidase-conjugated anti-GST antibody (sc-138, Santa Cruz Biotechnology, Inc.) and a panel of PKC isoform-specific antibodies (BD Biosciences). Blots were developed with the use of ECL (Amersham Biosciences) or SuperSignal detection (Pierce). In all cases, data shown are representative of three or more similar experiments.

RESULTS
Ca v 1.2 Forms a Macromolecular Complex with PKC in the Heart and Brain-Although significant evidence supports the concept that PKC signaling pathways can modulate Ca v 1.2 activity in the heart, there is still little to no information on the identity of the PKC isoform(s) and the molecular mechanisms underlying this regulation. In keeping with the emerging role of protein-protein interactions as a mechanism to locally regulate channel activity by regulatory kinases (particularly in differentiated cells such as cardiomyocytes and neurons), the initial studies examined whether PKC isoforms interact with Ca v 1.2 in cardiac or brain tissue. Cardiomyocytes coexpress conventional (cPKC␣), novel (nPKC␦ and nPKC⑀), and atypical (aPKC) isoforms; cPKC␤ has also been variably detected in cardiomyocytes by some investigators (30 -32). Fig. 1 shows that Ca v 1.2 co-immunoprecipitated with cPKC␣, but nPKC␦ and nPKC⑀ were not recovered in the anti-Ca v 1.2 immune complexes (Fig. 1A). Similarly, all three cPKC isoforms expressed in the brain (PKC␣, PKC␤, and PKC␥) were recovered in Ca v 1.2 immune complexes (Fig. 1B). In contrast, nPKC␦ (which is abundant in cardiac and brain tissue) did not coimmunoprecipitate with Ca v 1.2 in either preparation under these conditions (Fig. 1, A and C). nPKC⑀ also did not coimmunoprecipitate with Ca v 1.2 from cardiomyocytes ( Fig. 1A), although some nPKC⑀ was detected in Ca v 1.2 immune complexes from the brain (Fig. 1C), perhaps due to the higher levels of nPKC⑀ expression in this tissue. Other PKC isoforms such as aPKC and aPKC (which are abundant in neuronal tissue and are not expressed at any appreciable level in cardiomyocytes) also were recovered with Ca v 1.2 in the brain (Fig. 1C). In each case, the PKC isoform interactions with Ca v 1.2 were specific, as PKC isoforms were not recovered when immunoprecipitations were performed with irrelevant IgGs. Collectively, these results show a constitutive interaction between Ca v 1.2 and certain PKC isoforms (primarily cPKCs) in cardiac and brain tissue.
To determine whether the PKC isoforms directly interact with the ␣ 1c subunit (and to map the PKC interaction domains on ␣ 1c subunits), a panel of GST fusion proteins containing the ␣ 1c subunit N terminus, individual intracellular loops, and non-overlapping regions of the ␣ 1c C terminus were generated ( Fig. 2A). Fig. 2B shows that cPKC␣ co-sedimented with GST fusion proteins containing domains in the ␣ 1c subunit C terminus. The interactions between cPKC␣ and these non-overlapping regions of the ␣ 1c subunit C terminus were specific, as cPKC␣ did not co-sediment with GST alone or GST fused to other ␣ 1c subunit intracellular domains (including the N terminus or the intracellular loop sequences) (Fig. 2B). To determine whether the cPKC␣-␣ 1c C terminus interaction is direct (i.e. does not require another coprecipitating protein), additional in vitro studies were performed with the ␣ 1c C terminus-GST fusion proteins and recombinant cPKC␣. Fig. 2C shows that both C terminus-GST fusion proteins pull-downed cPKC␣; cPKC␣ was not recovered when the pull-down was performed with GST alone. Collectively, these results identify a direct interaction between cPKC␣ and the ␣ 1c subunit.
We next sought to determine whether the interaction between Ca v 1.2 and cPKC␣ (which we detected in native tissues and in in vitro binding assays) could be reconstituted in a mammalian expression system. To this end, we expressed Ca v 1.2 (␣ 1c and ␤ 2a ) in HEK293 cells, which have nearly undetectable levels of endogenous ␣ 1c upon immunoblotting. HEK293 cells demonstrated relatively high expression of cPKC␣, aPKC, and aPKC and weaker expression of nPKC␦, nPKC⑀, nPKC, and aPKC upon immunoblotting (data not shown). Fig. 3A shows that ␣ 1c specifically co-immunoprecipitated with endogenous cPKC␣ in unstimulated HEK293 cells (Fig. 3A, left panel). No cPKC␣ immunoreactivity could be detected when immunoprecipitation was performed with preimmune serum or with anti-␣ 1c antibody in the presence of blocking (immunogenic) peptide (Fig. 3A, left panel), consistent with the lack of immunoprecipitation of ␣ 1c (right panel).
A constitutive interaction between Ca v 1.2 and cPKC␣ could constitute a mechanism to optimize channel phosphorylation in the context of stimulatory conditions that are associated with the generation of Ca 2ϩ and DAG. To determine whether ␣ 1c subunits are phosphorylated by PKC␣ docked to the channel, we performed in vitro kinase assays. Immune complexes were FIG. 1. Ca v 1.2 forms a macromolecular complex with PKC in the heart and brain. A, Ca v 1.2 ␣ 1c was immunoprecipitated from rat heart membranes, size-fractionated on SDS-polyacrylamide gel, transferred to nitrocellulose, and immunoblotted with PKC isoform-specific antibodies as indicated. Ca v 1.2 associated with cPKC␣ (upper left panel), but not with nPKC␦ (middle panel) or nPKC⑀ (right panel). The immunoblot (lower left panel) of the immunoprecipitation (IP) of the ␣ 1c subunit demonstrates specificity of the antibody; preimmune serum did not immunoprecipitate ␣ 1c (lower left panel) or PKC␣ (upper left panel). ϩ cont represents 10% input. HC, IgG heavy chain. B, shown are cPKC isoform immunoblots of ␣ 1c /preimmune serum precipitates from rat brain extracts. Brain ␣ 1c associated with cPKC␣, cPKC␤, and cPKC␥; preimmune serum did not precipitate the PKC isoforms. The immunoblot (lower left panel) of the immunoprecipitation of the ␣ 1c subunit demonstrates specificity of antibody; preimmune serum did not immunoprecipitate ␣ 1c (lower left panel) or PKC␣ (upper left panel). C, shown are nPKC and aPKC isoform immunoblots of ␣ 1c /preimmune precipitates from rat brain extracts. Brain ␣ 1c weakly associated with nPKC⑀, aPKC, and aPKC, but not with nPKC␦. In all cases, blots are representative of at least three separate experiments. isolated with anti-␣ 1c subunit antibody or preimmune serum and subjected to kinase assays with [␥-32 P]ATP in the presence of sonicated vesicles containing the PKC-activating lipid cofactors and DAG. Although endogenous cPKC␣ co-immunoprecipitated with ␣ 1c subunits under these conditions, exogenous cPKC␣ (alone or with a PKC peptide inhibitor; Promega catalog no. V5691) was added to some incubations as indicated. Fig. 3B shows that ␣ 1c subunits were labeled in the in vitro kinase assays by both exogenous and endogenous PKCs. ␣ 1c subunit phosphorylation increased when excess exogenous PKC␣ was added to the assays (and it was blocked in the presence of a PKC inhibitor). These results indicate that the ␣ 1c subunit is phosphorylated by PKC, rather than another kinase(s) that also might co-immunoprecipitate with the Ca v 1.2 channel complex.
PKC Phosphorylates the ␣ 1c Subunit C Terminus-The sites for PKC phosphorylation of Ca v 1.2 ␣ 1c subunits have not been identified. Based upon electrophysiologic findings, PKC inhibitory modulation of channel activity has been mapped to several residues in the ␣ 1c subunit N terminus (Thr 27 and Thr 31 ) (33), although evidence that this region is not directly phosphorylated in vitro by PKC has also been presented (23), which we have confirmed (data not shown). Because cPKC␣ interacts directly with the C terminus, we hypothesized that regions within the C terminus may contain the authentic PKC phosphorylation sites. To map PKC phosphorylation sites on the ␣ 1c subunit, we utilized GST fusion proteins containing non-overlapping regions of the C terminus as substrates for in vitro PKC phosphorylation. Whereas cPKC␣ phosphorylated both regions 1509 -1905 and 1906 -2170 of the C terminus, nPKC⑀ and aPKC phosphorylated primarily region 1906 -2170 (Fig. 4A). These results suggest that individual PKC isoforms can exert distinct regulatory control on channel function by phosphorylating different regions of the ␣ 1c subunit. Previous studies mapped the PKA phosphorylation site on the ␣ 1c subunit to a PKA consensus motif (RRAS 1928 ) located in the C terminus (24). In vitro ␣ 1c subunit phosphorylation assays are consistent with this conclusion, as PKA induced the radiolabeling of fragment 1906 -2170 (which encompasses the Ser 1928 site) (Fig. 4B).
␣ 1c Ser 1928 Is a PKC Phosphorylation Site-An examination of the amino acid sequence within fragment 1906 -2170 revealed that Ser 1928 represents a consensus PKC phosphorylation site (Fig. 5A) (34). PKA and cPKC␣ induced similar increases in 32 P incorporation into GST-fused wild-type (WT) fragment 1906 -2170; in each case, radiolabeling was nearly completely abrogated by a single substitution (S1928A) (Fig.  5B). Similarly, utilizing a phospho-specific antibody developed to detect Ser 1928 phosphorylation, prominent immunoreactive bands were detected (with a range of mobilities corresponding to the GST-fused full-length protein as well as the proteolytic/ truncated fragments) in both PKA and cPKC␣ phosphorylation assays in GST-fused WT fragment 1906 -2170 (Fig. 5C). In contrast, no anti-phospho-Ser 1928 antibody immunoreactivity was detected when a single substitution (S1928A) was introduced into the substrate sequence. Similar to cPKC␣, nPKC⑀ and aPKC phosphorylated Ser 1928 using GST-fused fragment 1906 -2170 (Fig. 5D).
Because Ser 1928 represents a common target for PKA and PKC in vitro (Fig. 5, B and C) and PKA has been shown to form a macromolecular complex with Ca v 1.2, we sought to determine whether PKC directly phosphorylates Ser 1928 . The PKA inhibitor PKI-(5-24) specifically blocked phosphorylation of Ser 1928 by PKA (but not cPKC␣) of GST-fused fragment 1906 -2170 (Fig. 5E) and full-length recombinant ␣ 1c expressed in HEK293 cells (Fig. 5F). Similar to cPKC␣, nPKC⑀ and aPKC specifically phosphorylated Ser 1928 in vitro in full-length recombinant ␣ 1c (Fig. 5G). These results demonstrate that PKA and PKC can independently phosphorylate ␣ 1c Ser 1928 .
Having validated the specificity of the phospho-Ser 1928 -spe-cific antibody as a reagent to track kinase-dependent phosphorylation of the ␣ 1c C terminus, we used it to examine the phosphorylation status of full-length recombinant WT and S1928A ␣ 1c subunits expressed in HEK293 cells. Fig. 6A shows that WT ␣ 1c subunits (immunoprecipitated from quiescent HEK293 cells and subjected to immune complex kinase assays without lipid cofactors) displayed very low levels of Ser 1928 phosphorylation (second lane); in contrast, Ser 1928 phosphorylation was not detected when studies were performed on HEK293 cells expressing the S1928A mutant (first lane), despite equivalent expression and immunoprecipitation of WT and mutant ␣ 1c subunits (lower panels). WT ␣ 1c Ser 1928 phosphorylation was increased by the addition of lipid cofactors to the in vitro kinase assays (Fig. 6A, fourth lane), consistent with Fig. 3 results showing that ␣ 1c subunits were recovered in complexes with cPKC␣. Exogenously added cPKC␣ or PKA further increased WT ␣ 1c Ser 1928 phosphorylation (Fig. 6A, fifth and sixth lanes). Similar in vitro studies were performed on endogenous ␣ 1c subunits recovered from adult rat hearts. In native tissues (heart, brain, and skeletal muscle), the ␣ 1 subunit displayed a slight increase in mobility upon SDS-PAGE (relative to the mobility of full-length ␣ 1c subunits expressed in HEK293 cells) due to the previously described C-terminal truncation (35)(36)(37). The common PKA and PKC phosphorylation site (Ser 1928 ) is present only in the full-length form of ␣ 1c (37) and is removed by calpain, a Ca 2ϩ -activated protease in the brain (38) and an unidentified protease in the heart (39). Fig. 6B demonstrates that ␣ 1c recovered from adult rat heart was phosphorylated at Ser 1928 by PKC␣ in an in vitro kinase assay. Ser 1928 was partially phosphorylated under basal conditions in adult rat heart by either PKA or PKC (Fig. 6B, second lane).
In HEK cells overexpressing WT Ca v 1.2 channels, incubation with phorbol 12-myristate 13-acetate (PMA; 500 nM) prior to lysis led to phosphorylation of Ser 1928 as assessed using the phospho-Ser 1928 -specific antibody (Fig. 7A). The PMA-induced phosphorylation was inhibited by preincubating the cells with bisindolylmaleimide (500 nM; GF109203X), indicating that the phosphorylation was mediated by PKC. Exposure of the cells to the protein phosphatase (PP) inhibitor calyculin A (50 nM) prior to stimulation with PMA caused increased phosphorylation of Ser 1928 in vivo, which was also significantly inhibited by bisindolylmaleimide (Fig. 7A). Calyculin A-induced phosphorylation of Ser 1928 was partially inhibited by H-89 (500 nM) and bisindolylmaleimide (500 nM), but completely inhibited by both inhibitors (Fig. 7B). This suggests that both PKA and PKC con-  Fig. 2B. Lower molecular mass bands are truncation products. Autoradiographs were exposed for equivalent times (10 min at room temperature). Equivalent kinase units were utilized for each reaction, which was optimized for the given PKC isoform. B, an in vitro kinase reaction was performed with recombinant PKA and demonstrates the phosphorylation of GST-fused ␣ 1c C-terminal fragment 1906 -2170. tribute to the calyculin A-induced phosphorylation of Ser 1928 in HEK293 cells overexpressing Ca v 1.2. Similarly, phosphorylation of Ser 1928 by PKC was also observed in neonatal cardiomyocytes treated with PMA and calyculin A (Fig. 7C). The phosphorylation was inhibited by the PKC inhibitor bisindolylmaleimide. Taken together, these findings demonstrate that Ser 1928 is phosphorylated by PKC in vivo. Phosphorylation of ␣ 1c at Ser 1928 is believed to lead, in part, to increased channel FIG. 6. Ser 1928 is phosphorylated by PKC in full-length recombinant channels and rat heart. A, extracts from HEK cells transfected with WT or S1928A ␣ 1c ϩ ␤ 2a were prepared, followed by ␣ 1c immunoprecipitation and kinase reaction as indicated. Samples were size-fractionated on SDS-polyacrylamide gel, transferred to nitrocellulose membrane, and probed with anti-phospho-Ser 1928 antibody (pS1928). Under basal conditions, the majority of Ser 1928 was not phosphorylated (second lane), compared with maximal phosphorylation with PKA and cPKC␣ (fifth and sixth lanes). cPKC␣ associated with channels phosphorylated Ser 1928 (fourth lane). B, rat heart lysates were prepared in the presence of phosphatase inhibitors, and ␣ 1c was immunoprecipitated (IP) and probed with anti-phospho-Ser 1928 (upper panel) or ␣ 1c (lower panel) antibody. In the last lane, the immunoprecipitates were phosphorylated with PKC␣ in vitro. Native ␣ 1c demonstrated partial C-terminal truncation as demonstrated by the broad band, with a significant portion below 200 kDa (compared with recombinant channels in A). The signal for anti-phospho-Ser 1928 antibody was present only in the full-length channel because truncation was N-terminal to the phosphorylation site. shown is an anti-phospho-Ser 1928 antibody immunoblot of PKC isoform in vitro kinase reaction with GST-fused WT fragment 1906 -2170 and S1928A. cPKC␣, nPKC⑀, and aPKC phosphorylated Ser 1928 equivalently. E: upper panels, immunoblotting using anti-phospho-Ser 1928 antibody demonstrates that PKA phosphorylation of GST-fused fragment 1906 -2170 was inhibited by the PKA inhibitor PKI-(5-24), whereas cPKC␣ phosphorylation was not affected. Lower panels, Ponceau staining of nitrocellulose membrane indicates equivalent loading of GST fusion proteins. A dark exposure was selected to highlight differences between PKI inhibition of PKA and PKC. F: shown are immunoblots with anti-phospho-Ser 1928 (upper panel) and anti-␣ 1c (lower panel) antibodies of in vitro kinase reaction of ␣ 1c with the indicated kinase/inhibitor. Ca v 1.2 was expressed in HEK293 cells, and ␣ 1c was immunoprecipitated, followed by the kinase reaction. Ser 1928 was phosphorylated by PKA and cPKC␣, but PKI inhibited PKA (but not cPKC␣) phosphorylation. G: shown are immunoblots with anti-phospho-Ser 1928 antibody (upper panels) and anti-␣ 1c antibody (lower panel) demonstrating cPKC␣, nPKC⑀, and aPKC phosphorylation of WT (but not S1928A) ␣ 1c . Recombinant ␣ 1c was recovered by immunoprecipitation. In each case, phosphorylation conditions were optimized for the indicated isoform, and equivalent kinase units were utilized. Lower panels, ␣ 1c immunoblot demonstrating equivalent ␣ 1c input into phosphorylation assay. activity (25,40,41), thereby potentially accounting for the activation of the channel observed after stimulation with agents that activate PKC signaling pathways.

DISCUSSION
This study provides the identification of a molecular mechanism that mediates the regulation of Ca v 1.2 in response to activators of PKC. We have mapped specific regions/residues within the Ca v 1.2 channel complex that are sites for PKC-dependent phosphorylation in vitro and in vivo. Observed differences in Ca 2ϩ channel regulation by individual PKC isoforms serve to highlight the complexity that likely exists in vivo, where regulatory controls are through the orchestrated effects of a series of stimulatory and inhibitory kinases and phosphatases that regulate channel activity in a highly specific manner in response to a given physiologic stimulus.
A common theme in signal transduction is the close association of signaling molecules (such as kinases and phosphatases) with effectors (such as ion channels) to achieve specific and localized regulation. Indeed, our findings that specific PKC isoforms can associate with Ca v 1.2 are consistent with this premise. Moreover, the ability to bind different PKC isoforms may lead to variable regulation of channel activity, given the findings that several PKC isoforms demonstrate different abilities to phosphorylate GST-fused C-terminal fragment 1509 -1905. Several PKC-binding proteins have been identified; for instance, RACKs (receptors for activated C kinase) bind activated forms of PKC and enhance its activity (42). In nonstimulated rat heart, RACK1 did not bind Ca v 1.2 (data not shown). Interestingly, AKAP15-PKA also associates with Ca v 1.2 in the distal C terminus of the ␣ 1c subunit (43). Prior studies have established that PKC isoforms can bind directly with ion channels/substrates; for instance, PKC binds directly to the glutamate receptor (44), the ␥-aminobutyric acid receptor (45), and phospholipases D 1 and D 2 (46).
The Ca v 1.2 channel complex is an important PKC target in the heart, as it plays a critical role in excitation-contraction coupling and potentially in myocyte survival and growth (13). Angiotensin II, ␣ 1 -adrenergic receptors, and endothelin-1 re-ceptors are members of a G protein-coupled receptor family that lead to stimulation of PKC isoforms through activation of phospholipase C and production of inositol 1,4,5-trisphosphate and DAG. Angiotensin II and endothelin-1 have been reported to increase Ca 2ϩ entry and thus contractility in mammalian cardiac myocytes (13,47). However, conflicting findings regarding the effects on Ca v 1.2 have been generated for several direct activators of PKC, including phorbol esters (6, 48 -50), 1,2dioctanoyl-sn-glycerol (51), and 1-oleoyl-2-acetyl-sn-glycerol (52). Agonists coupled to PKC signaling pathways are frequently positive inotrophs, whereas cell-permeable PKC activators are often negative inotrophs (13,53). This may be due to PMA effects on PKC localized near the surface membrane as opposed to transverse tubules and/or the loss of specificity (13). Our findings that Ca v 1.2 pre-associates with specific PKC isoforms in the heart and brain are consistent with this hypothesis and provide a potential molecular mechanism for the differences observed between PMA and agonists such as endothelin-1 and angiotensin II.
The identification of Ser 1928 as a PKC phosphorylation site has important implications regarding the study of phosphorylation-dependent modulation of Ca v 1.2. PKA increases channel activity in the heart and brain. The full-length ␣ 1c subunit is phosphorylated by PKA at the single residue (Ser 1928 ) in the C terminus, which is missing in the truncated form of the channel (24). However, the C terminus remains associated with the channel, potentially modulating its activity (39,54). Similar to PKA, our data demonstrate that three representative PKC isoforms (cPKC␣, nPKC⑀, and aPKC) specifically phosphorylate Ser 1928 in vitro and in vivo. Because Ser 1928 is the main and probably only PKA site on the ␣ 1c subunit (24), our findings suggest that Ser 1928 phosphorylation plays an important role in mediating the PKA and PKC modulation of Ca v 1.2. Thus, Ser 1928 phosphorylation may represent a target for the convergence of the two signal transduction pathways, with specificity imparted by the PKC-mediated phosphorylation of other sites within the ␣ 1c and/or ␤ subunit.
Our findings also demonstrate the importance of phospha- FIG. 7. Reconstitution of PMA/PKC-mediated phosphorylation of ␣ 1c in HEK293 cells. A, recombinant WT and S1928A ␣ 1c ϩ ␤ 2a were expressed in HEK293 cells. Cells were exposed to PMA (500 nM) ϩ calyculin A (Cal; 50 nM) for 10 min. In the indicated lanes, cells were pretreated with bisindolylmaleimide (Bis; 500 nM; GF109203X) for 1 h prior to stimulation with PMA ϩ calyculin A. Cells were harvested and lysed in the presence of phosphatase inhibitors. Lysates (100 g) were size-fractionated on SDS-polyacrylamide gel, transferred to nitrocellulose membranes, and blotted with anti-phospho-Ser 1928 (pS1928; upper panel) or anti-␣ 1c (lower panel) antibody. Immunoprecipitation of the ␣ 1c subunit was not required. Activation of PKC by PMA increased phosphorylation of Ser 1928 , which was inhibited by bisindolylmaleimide. Inhibition of phosphatase activity with calyculin A in combination with PMA also increased phosphorylation of Ser 1928 , which was significantly (but not completely) inhibited by bisindolylmaleimide. Basal phosphorylation of recombinant ␣ 1c in HEK293 cells was minimal. The specificity of the antibody and PMA effect is shown using the S1928A mutant. S1928A phosphorylation demonstrated no basal phosphorylation. A small signal was observed in the S1928A mutant exposed to PMA, indicating a low level of endogenous Ca v 1.2 in these cells (signal was also observed in untransfected cells (not shown)). Equal expression of the ␣ 1c protein is demonstrated in the lower panel. B, anti-phospho-Ser 1928 antibody immunoblotting was performed with lysates prepared from calyculin A-treated HEK293 cells expressing WT ␣ 1c ϩ ␤ 2a subunits exposed to a PKC (bisindolylmaleimide) and/or PKA (H-89; 500 nM for 1 h) inhibitor as indicated (upper panel). Exposure of the cells to the cell-permeable PKA or PKC inhibitor partially inhibited calyculin A-induced phosphorylation of Ser 1928 , whereas both inhibitors nearly completely inhibited phosphorylation. Equal expression of the ␣ 1c protein is shown in the lower panel. C, membranes were prepared from rat neonatal cardiac myocytes exposed to PMA, calyculin A, and/or bisindolylmaleimide as indicated. Membranes (100 g) were size-fractionated on SDS-polyacrylamide gel and blotted with anti-phospho-Ser 1928 (upper panel) or anti-␣ 1c (lower panel) antibody. PMA and calyculin A significantly increased Ser 1928 phosphorylation, which was inhibited by the PKC inhibitor bisindolylmaleimide. tase regulation of Ser 1928 phosphorylation. Recent studies have suggested that PPs may regulate, in part, the activity of the Ca v 1.2 channel. PP2A has been shown to associate with Ca v 1.2 in a macromolecular complex in the brain (55). The PP1/PP2A inhibitor calyculin A (125 nM) increases Ca v 1.2 activity by 70% in murine ventricular myocytes (10). The calyculin A effect is believed to be due primarily to its effects on PP1 rather than PP2A activity because okadaic acid (10 -100 nM) or fostriecin (500 nM) fails to increase steady-state Ca v 1.2 current (11). The counteracting kinase for the effect of PP1 on Ca v 1.2 current was found to be PKC, implying that PKC and PP1 activities determine the steady-state level of cardiac Ca v 1.2 activity (11). Ser 1928 phosphorylation may represent the Ca v 1.2 target that mediates this effect because the site is regulated by both PKC and calyculin A (Fig. 7). Our data do not preclude regulation of Ser 1928 phosphorylation by additional kinases such as protein kinase D, which is activated by PMA and is downstream of PKC. The protein kinase D consensus phosphorylation site is LXRXXS (34,56), which corresponds to the amino acid sequence surrounding Ser 1928 .
PKC isoforms are targeted to the Ca v 1.2 complex through direct interaction with the C terminus, thereby enabling specific regulation of ␣ 1c phosphorylation by PKC. We have identified Ser 1928 as a PKC phosphorylation site, which has previously been shown to be phosphorylated by PKA (24) and is responsible, in part, for the PKA-mediated activation of the channel. This has important implications in the study of neuronal and cardiovascular diseases. An increase in Ca 2ϩ influx through Ca v 1.2/Ca v 1.3 is thought to contribute to aging-induced neuronal dysfunction (41). For instance, Ca 2ϩ influx in hippocampal neurons is up-regulated by Ͼ2-fold in old versus adult rats (57). Ser 1928 phosphorylation was recently shown to be increased by Ͼ2-fold, leading to the conclusion that channel activity is up-regulated in the hippocampus of the aging rat (41). However, the molecular basis for the increase in Ca v 1.2 phosphorylation was not found since the levels of cAMP, PP2A, and PP1 inhibitor-1 and -2 and the association of PKA with the channel complex were not altered. Based upon our findings, it is conceivable that the Ser 1928 hyperphosphorylation may be due to elevated PKC rather than PKA activity. Aging has been associated with increased PKC activity in specific regions of the brain, including the hippocampus (58). Further investigation is required to determine the contribution of PKC to mediating Ser 1928 phosphorylation in aging brain.
Taken together, our findings identify PKC as an integral part of the Ca v 1.2 macromolecular complex in the brain and heart. Ser 1928 is phosphorylated by both PKA and PKC. Differential regulation of the channel by specific PKC isoforms through phosphorylation of unique sites within the ␣ 1c subunit may lead to highly specific modulation of channel activity.