Stimulus-specific Differences in Protein Kinase C (cid:1) Localization and Activation Mechanisms in Cardiomyocytes*

Protein kinase C (PKC) isoforms play key roles in the regulation of cardiac contraction, ischemic precondi-tioning, and hypertrophy/failure. Models of PKC activation generally focus on lipid cofactor-induced PKC translocation to membranes. This study identifies tyrosine phosphorylation as an additional mechanism that regulates PKC (cid:1) actions in cardiomyocytes. Using immunoblot analysis with antibodies to total PKC (cid:1) and PKC (cid:1) pY 311 , we demonstrate that PKC (cid:1) partitions between soluble and particulate fractions (with little Tyr 311 phosphorylation) in resting cardiomyocytes. Phorbol 12-my-ristate 13-acetate (PMA) promotes PKC (cid:1) translocation to membranes and phosphorylation at Tyr 311 . H 2 O 2 also increases PKC (cid:1) -pY 311 in association with its release from membranes. Both PMA- and H 2 O 2 -dependent in- creases in PKC (cid:1) -pY 311 are mediated by Src family ki-nases,buttheyoccurviadifferentmechanisms.TheH 2 O 2 - dependent increase in PKC (cid:1) -pY 311 results from Src activation and increased Src-PKC

Protein kinase C (PKC) 1 comprises a multigene family of at least 10 structurally distinct phospholipid-dependent serine-threonine kinases that regulate cardiac contraction, play a role in ischemic preconditioning, and contribute to the pathogenesis of cardiac hypertrophy and heart failure (1,2). PKC isoforms are single polypeptide chains with structurally homologous C-terminal catalytic domains and more variable N-terminal regulatory domains. This diverse group of enzymes is subdivided into three distinct subfamilies based upon structural differences in their N-terminal regulatory domain that confer distinct patterns of cofactor activation. Conventional PKC isoforms (cPKCs; ␣, ␤I, ␤II, ␥) contain an autoinhibitory pseudosubstrate domain followed by membrane-targeting C1 and C2 domains that are regulated by diacylglycerol (DAG) and calcium, respectively. Novel PKCs (nPKCs; ␦, ⑀, , and ) lack a calcium-binding C2 domain and are maximally activated by DAG and phorbol ester, in the absence of calcium. Atypical PKCs (aPKCs; and /) are regulated by lipids, but are not activated by second messengers such as calcium and DAG. Current models of PKC isoform activation in the heart have focused largely on the conformational changes induced by cofactor interactions with N-terminal membrane-targeting modules that anchor the enzyme to membranes, expel the autoinhibitory pseudosubstrate domain from the substrate-binding pocket, and thereby relieve autoinhibition. According to this model, individual PKC isoforms elicit distinct (and occasionally functionally opposing) cellular responses as a result of cofactorinduced compartmentation to distinct membrane subdomains, in close proximity to their unique sets of target protein substrates (1).
Recent studies identify an additional mechanism for PKC regulation via sequential phosphorylations on a conserved threonine in the activation loop and two conserved serine/threonines in turn and hydrophobic motifs in the C terminus (3). For cPKCs, these phosphorylation events are completed during enzyme maturation (and are critical for the generation of an enzymatically active and conformationally stable protein in the cytosol). In contrast, our recent studies identify dynamically regulated nPKC isoform phosphorylation events that accompany agonist-induced nPKC activation/translocation (and contribute to the regulation of nPKC function) in cardiomyocytes (4). Agonist-induced phosphorylation of PKC␦ at Thr 505 in the activation loop regulates its kinase activity, whereas agonist-induced phosphorylation of PKC⑀ at its C-terminal hydrophobic motif influences the kinetics of trafficking/down-regulation.
PKC isoforms (in particular PKC␦) also serve as targets for regulatory phosphorylations on tyrosine residues in cells transformed with Src and Ras or acutely stimulated with PMA, epidermal growth factor, platelet-derived growth factor, and H 2 O 2 (5)(6)(7)(8)(9). However, the precise biological consequences of PKC␦ tyrosine phosphorylation have been difficult to decipher, at least in part because of the presence of multiple sites for independently regulated tyrosine phosphorylations in PKC␦'s regulatory domain (Tyr 52 , Tyr 155 , and Tyr 187 ), catalytic domain (Tyr 512 and Tyr 523 ), and hinge region (Tyr 311 and Tyr 332 ). Current literature suggests that there is no uniform pattern or consequence of PKC␦ tyrosine phosphorylation. Rather, the precise configuration of tyrosine residues phosphorylated on PKC␦ depends upon the nature of the inciting stimulus and dictates the functional properties of the enzyme in cells. Most studies have relied on in vitro kinase assays to resolve tyrosine phosphorylation-dependent changes in PKC␦ function and variably describe the catalytic activity of tyrosine phosphorylated PKC␦ as decreased, increased or even altered its substrate specificity and cofactor requirements (5)(6)(7)(8)(9). However, a singular focus on phosphorylation-driven effects for PKC␦ may be myopic (or even misplaced), given recent evidence that phosphotyrosines on PKC␦ can also serve as docking sites for other signaling proteins (10,11) and that the PKC␦ role as a signalregulated scaffold (rather than as an active serine/threonine kinase) may underlie certain of its effector functions (12).
Oxidative stress is a common feature of many cardiovascular risks (including hypertension, diabetes, and smoking) and contributes to the pathogenesis of heart failure syndromes (13). There is ample evidence that intracellularly generated reactive oxygen species (ROS) and exogenously administered hydrogen peroxide (H 2 O 2 ) activate both mitogenic and apoptotic signaling pathways in cardiomyocytes (14,15). PKC␦ is one of several signaling molecules typically activated in this context. However, the precise mechanisms whereby oxidant stress activates PKC␦ and the role of PKC␦ in ROS triggered structural and functional remodeling of the ventricle has not been examined. This study demonstrates that H 2 O 2 treatment releases PKC␦ from membranes and generates a tyrosine-phosphorylated form of PKC␦ that exhibits lipid-independent catalytic function (and is poised to phosphorylate distinct PKC␦ target proteins throughout the cell, not just on lipid membranes). These results suggest that there are two independent signaling modes for PKC␦, with PKC␦ actions/phosphorylations resulting from GPCR activation and the generation of DAG in membranes functionally distinct from the events induced by the tyrosinephosphorylated form of PKC␦ in the soluble fraction of cells exposed to oxidant stress.
Cardiomyocyte Culture and Transfection-Cardiomyocytes were isolated from the hearts of 2-day-old Wistar rats by a trypsin dispersion procedure using a differential attachment procedure to enrich for cardiomyocytes followed by irradiation as described previously (4). The yield of cardiomyocytes typically is 2.5-3 ϫ 10 6 cells per neonatal ventricle. 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 MEM (Invitrogen, Life Technologies, Inc.) supplemented with 10% fetal calf serum and then serum-deprived for the subsequent 24 h.
Immunoblot Analysis-Immunoblot analysis was performed on whole cell extracts or soluble and particulate fractions prepared accord-ing to methods described previously (16,17). Briefly, cells were washed with phosphate buffered saline and then immediately transferred to ice-cold homogenization buffer (20 mM Tris-HCl, pH 7.5, 2 mM EDTA, 2 mM EGTA, 6 mM ␤-mercaptoethanol, 50 g/ml aprotinin, 48 g/ml leupeptin, 5 M pepstatin A, 1 mM phenylmethylsulfonyl fluoride, 0.1 mM sodium vanadate, and 50 mM NaF). Cells were lysed by sonication, centrifuged at 100,000 ϫ g for 1 h, and the supernatant was saved as the soluble fraction. PKC isoforms were extracted from the pellets by incubation on ice for 10 min in homogenization buffer containing 1% Triton X-100 followed by centrifugation at 100,000 ϫ g for 30 min at 4°C. Electrophoretic separation was performed using 8% SDS-polyacrylamide gels, followed by transfer to nitrocellulose for immunoblotting with a panel of antibodies that recognize total PKC␣, PKC␦, or PKC⑀ protein expression (previously identified as the phorbol estersensitive PKC isoforms expressed by cardiomyocyte cultures (16,18)), PKC␦ phosphorylated at the activation loop (anti-PKC␦-pT 505 ), or PKC␦ tyrosine-phosphorylated at the hinge region (anti-PKC␦-pY 311 ). The specificity of the anti-PKC antibodies was established previously (4,19); the specificity of all anti-phospho-PKC antibodies was validated in control experiments showing that anti-phospho-PKC immunoreactivity is stripped by treatment of samples with alkaline phosphatase (data not shown) and is down-regulated along with PMA-dependent down-regulation of the cognate protein (4). Bands were detected by enhanced chemiluminescence, with each panel in each figure from a single gel exposed for a uniform duration.
Immunoprecipitates (equivalent to the amount of protein extracted from 0.3 ϫ 100-mm dish per assay) were used in immunocomplex kinase assays carried out in 200 l of a reaction mixture containing buffer (26 mM Tris, pH 7.5, 5 mM MgCl 2 , 0.6 mM EGTA, 0.6 mM EDTA, 0.5 M PKI, 10 M PP1, and 0.25 mM DTT) in the absence and presence of 80 g/ml phosphatidylserine (PS) plus 160 ng/ml PMA with histone III-S (0.5 mg/ml), PKC-⑀-peptide (50 M) or PKC-␦-peptide (50 M) as substrate. PKC-⑀-peptide is a synthetic peptide that corresponds to the pseudosubstrate domain of PKC⑀, with a phosphorylatable serine for alanine substitution (ERMRPRKRQGSVRRRV); this peptide can serve as a good substrate for all PKCs except PKC (20). PKC-␦-peptide (which corresponds to amino acids 422-443 of murine eEF-1␣, RFAVRDMRQTVAVGVIKAVDKK) is a relatively specific substrate for PKC␦ (and not other PKC isoforms, Ref. 20). Of note, we previously established that the phosphorylation of ␦and ⑀-peptide (which do not contain tyrosines) or histone is mediated by PKC (rather than another co-immunoprecipitating kinase); all phosphorylations are completely inhibited by the addition of 5 M GF109203X to the in vitro kinase assay buffer (4). Reactions were initiated by the addition [␥-32 P]ATP (13 Ci, 66 M) and were performed in quadruplicate at 30°C for 16 min. Assays were terminated by placing samples on ice followed by centrifugation at 1,500 ϫ g for 10 min at 4°C, and 40 l of each supernatant was spotted onto phosphocellulose filter papers (P-81). Each P-81 disc was immediately dropped into water, washed (five times for 5 min), and counted for radioactivity. Pellets were subjected to SDS-PAGE and immunoblotting for PKC␦, to normalize for minor differences in the amount of immunoprecipitated enzyme.
In Vitro Phosphorylation of PKC␦ by Src-0.1 g of recombinant human PKC␦ (rPKC␦) was preincubated in the absence or presence of Src kinase (0.66 units) in 160 l of a reaction buffer containing 43 mM Tris-Cl, pH 7.5, 6.25 mM MgCl 2 , 10 mM MnCl 2 , 0.75 mM EDTA, 0.77 mM EGTA, 0.3 mM DTT, 125 mM NaCl, 5% glycerol, 0.006% Brij-35, 0.04 mM phenylmethylsulfonyl fluoride, 0.2 mM benzamidine, and [␥-32 P]ATP (13 Ci, 83 M). Incubations were for 15 min at 30°C and were stopped by placing samples on ice. PKC substrates were then added, and kinase assays were carried out (in a final volume of 200 l) in the absence or presence of PS/PMA as described above. Triplicate 10-l aliquots of each sample were spotted onto phosphocellulose filter paper (P-81), dropped into water, washed (5ϫ for 5 min), and counted for radioactivity. The remainder of the sample (0.07 g of PKC␦) was subjected to SDS-PAGE and immunoblotting for P-Y or PKC␦-pY 311 .

Distinct Effects of H 2 O 2 , PMA, and Norepinephrine on PKC␦ Translocation and Phosphorylation in
Cardiomyocytes-Immunoprecipitation of PKC␦ followed by immunoblot analysis with anti-YP was used as a screen for PKC␦ tyrosine phosphorylation by the ␣ 1 -adrenergic receptor agonist norepinephrine (NE), PMA, and H 2 O 2 . Fig. 1A shows that there is a low level of PKC␦ tyrosine phosphorylation in quiescent cultures and that PKC␦ tyrosine phosphorylation is increased in cultures treated with NE, PMA, and H 2 O 2 . The magnitude of the NE-dependent increase in PKC␦ tyrosine phosphorylation is relatively modest. In contrast, PMA and H 2 O 2 induce substantial increases in PKC␦ tyrosine phosphorylation, with the effect of H 2 O 2 considerably greater than the effect of PMA. To begin to resolve an agonist-dependent difference in the level of PKC␦ tyrosine phosphorylation at a single site from stimulus-specific differences in the number/identity of sites on PKC␦ that undergo tyrosine phosphorylation, immunoblot analysis was performed with an antibody that selectively recognizes PKC␦ phosphorylation at Tyr 311 . We focused on this tyrosine (in the hinge region, between the cysteine-rich and kinase domains) because it has been identified as a major site for tyrosine phosphorylation in cells treated with H 2 O 2 , it is flanked by sequence that conforms to an optimal Src substrate (and is reported to be a target for Src-dependent phosphorylation), and it represents a modification that reportedly alters PKC␦ kinase activity and accelerates its down-regulation kinetics (6,22,23). Fig. 1A shows that PKC␦-Y 311 phosphorylation is increased by PMA and (to a somewhat greater extent) by H 2 O 2 , but not by NE.
Cardiomyocyte cultures were partitioned into soluble and particulate fractions prior to immunoblot analysis to determine whether PKC isoforms (including the Tyr 311 -phosphorylated form of PKC␦) reside in distinct subcellular locations following stimulation with PMA and H 2 O 2 . Fig. 1B shows that PKC␦ protein partitions between the soluble and particulate fractions of resting cardiomyocytes. Stimulation with PMA leads to the translocation of PKC␦ to the particulate fraction and a decline in its mobility in SDS-PAGE, characteristic of a phosphorylation-induced change in electrophoretic mobility. In contrast, H 2 O 2 promotes a marked redistribution of PKC␦ from the particulate to the soluble fraction without changing its electrophoretic mobility in SDS-PAGE. PKC⑀ also partitions between the soluble and particulate fractions of resting cardiomyocytes and translocates to the particulate fraction upon activation with PMA. Whereas H 2 O 2 promotes the redistribution of PKC⑀ to the soluble fraction, the magnitude of this effect is relatively modest compared with the H 2 O 2 effect on PKC␦. In contrast, PKC␣ resides mainly in the soluble fraction of resting cardiomyocytes and is recovered in the particulate fraction of PMAtreated cardiomyocytes; PKC␣ partitioning between soluble and particulate fractions is not appreciably altered by H 2 O 2 .
PKC␦-Y 311 phosphorylation is detected at very low levels (and only in the particulate fraction) in resting cardiomyocytes (Fig. 1B). Stimulation with PMA leads to a large increase in particulate PKC␦-pY 311 (in association with the PMA-induced FIG. 1. Agonist-induced PKC␦ tyrosine phosphorylation and translocation in cardiomyocytes. Cardiomyocyte cultures were serum-starved for 24 h followed by stimulation for 20 min with vehicle, NE (10 M), PMA (100 nM), or H 2 O 2 (5 mM) as indicated. Panel A, extracts (1 mg/lane) were subjected to immunoprecipitation with anti-PKC␦ followed by immunoblot analysis for anti-YP, anti-PKC␦-pY 311 and total PKC␦ immunoreactivity (to control for immunoprecipitation and protein loading). The results are representative of data obtained in three experiments on separate culture preparations. It should be noted that these experiments used 1 mg/lane starting cell extract, because the NE-dependent increase in overall PKC␦ tyrosine phosphorylation was not detected (and the PMA response was at the limits of detection) in two preliminary experiments that used lesser amounts (0.5 mg/lane, although the much more vigorous H 2 O 2 -dependent increase in PKC␦ tyrosine phosphorylation was readily detected under these more stringent conditions). Panel B, soluble and particulate fractions were subjected to SDS-PAGE and immunoblot analysis with antibodies that recognize total PKC␦, PKC␣, and PKC⑀, PKC␦-pY 311 or PKC␦-pT 505 . Results are representative of data obtained on 4 -5 separate culture preparations. Panel C, whole cell lysates from cardiomyocytes stimulated with vehicle, norepinephrine (NE, 10 Ϫ5 M for the indicated intervals), or PMA (100 nM for 10 min, as a positive control) were subjected to SDS-PAGE and immunoblot analysis for PKC␦-pY 311 , PKC␦-pT 505 , and total PKC␦ protein. Similar results were obtained in three separate experiments on whole cell lysates (as well as two experiments that partitioned extracts into soluble and particulate fractions).
translocation of PKC␦ to membranes). PKC␦-Y 311 phosphorylation also is markedly increased in cells treated with H 2 O 2 . However, under this condition PKC␦-pY 311 is detected in both soluble and particulate fractions. This and subsequent experiments in this study report the cellular response to 5 mM H 2 O 2 ; increased amounts of Tyr 311 -phosphorylated PKC␦ also are recovered from cardiomyocytes subjected to lower levels of oxidant stress (100 -500 M, data not shown). Interestingly, both H 2 O 2 and PMA promote PKC␦-Y 311 phosphorylation, but only PMA slows the mobility of PKC␦ in SDS-PAGE ( Fig. 1B; also see Figs. 2 and 4). This result suggests that the electrophoretic mobility of PKC␦ is not influenced by phosphorylation at Tyr 311 ; rather the PMA-induced mobility shift must derive from a phosphorylation at a site distinct from Tyr 311 . We recently demonstrated that PKC␦ is phosphorylated at the turn motif (Ser 643 ) and hydrophobic motif (Ser 662 ) sites, but not at the activation loop (Thr 505 ) site, in resting cultures and that phosphorylation at Thr 505 (but not Ser 643 or Ser 662 ) increases during stimulation with PMA (4). Similarly, Fig. 1B  To better understand agonist-induced compartmentation of PKC␦, we studied the kinetics of agonist-induced PKC␦ translocation and Tyr 311 phosphorylation in greater detail (Fig. 2). The H 2 O 2 -induced release of PKC␦ from membranes (and translocation to the soluble fraction) is quite rapid (detectable within the first min, and maximal by 2 min); PKC␦ recovery in the soluble fractions remains increased for at least 30 min of H 2 O 2 stimulation. PKC␦-Y 311 phosphorylation accompanies the translocation to the soluble fraction. PKC␦-Y 311 phosphorylation is detected within the first 1-2 min and increases progressively (in both the soluble and particulate fractions) during the first 30 min of H 2 O 2 treatment. Of note, PKC␦-Y 311 phosphorylation increases progressively in the particulate fraction, despite a dwindling amount of PKC␦ protein recovery. PMA-induced PKC␦-Y 311 phosphorylation also follows rapid kinetics (Fig. 2B), with both PKC␦ translocation to the particulate fraction and PKC␦-Y 311 phosphorylation being maximal at 5 min (the first time point sampled). PMA-induced PKC␦-Y 311 phosphorylation is sustained for up to 5 h, declining in parallel with the PMA-induced down-regulation of the protein.
Most cardiac disorders are characterized by the simultaneous activation of hormone/neurotransmitter receptors coupled to lipid signaling (and the generation of cofactors that anchor PKC to membranes) as well as variable amounts of oxidant stress. To determine whether the H 2 O 2 -dependent release of PKC␦ from membranes competes with the PMA-dependent translocation of PKC␦ to membranes, PKC␦ localization and phosphorylation were compared in cardiomyocytes exposed to H 2 O 2 or PMA alone, versus PMA in the presence of H 2 O 2 . Fig.  3 shows that H 2 O 2 alone promotes the release of PKC␦ from membranes, whereas PMA alone effectively clears PKC␦ from the soluble fraction. However, PMA-dependent translocation of PKC␦ to the particulate fraction is attenuated by H 2 O 2 , illus- trating the dynamic convergence of these signals. Although PMA-dependent phosphorylation of PKC␦ at Thr 505 proceeds unhindered in the membrane fraction, no PMA-dependent PKC␦-T 505 phosphorylation is detected in the pool of PKC␦ that remains in the soluble fraction of PMA ϩ H 2 O 2 treated cardiomyocytes. This result is consistent with the notion that PKC␦-T 505 phosphorylation requires PKC␦ localization (in an open conformation, to expose the activation loop Thr 505 ) in membranes. The effects of H 2 O 2 to modify PMA (and presumably receptor-dependent) activation of PKC␦ appear to be specific (as opposed to a nonspecific effect of H 2 O 2 to inactivate PMA), because PMA-dependent translocation of PKC␣ is not influenced by H 2 O 2 . Fig. 4A shows that H 2 O 2and PMA-dependent PKC␦-Y 311 phosphorylations are markedly inhibited by the relatively specific Src family kinase inhibitor PP1. In contrast, H 2 O 2 -and PMA-dependent PKC␦-Y 311 phosphorylations are not blocked by AG1478 (an inhibitor of the EGF receptor tyrosine kinase), GF109203X (a general inhibitor of cPKCs, PKC␦, and PKC⑀, under conditions that were previously validated to completely inhibit cardiomyocyte PKC isoforms (4)) or PP3 (a structurally similar, but functionally inactive, PP1 analogue that serves as a negative control; Fig.  4B and data not shown). Neither PP1 nor GF109203X influence the PMA-dependent translocation of PKC␦ to the particulate fraction or the H 2 O 2 -dependent release of PKC␦ from membranes.

PMA-and H 2 O 2 -dependent Tyrosine Phosphorylation of PKC␦ Requires Src Kinase Activity-
Because H 2 O 2 -dependent tyrosine phosphorylation of PKC␦ is not necessarily limited to Tyr 311 , and there is evidence that PKC␦ phosphorylation at certain regulatory domain tyrosines need not be limited to the actions of Src family kinases (24), the effect of PP1 to more generally inhibit H 2 O 2 -dependent PKC␦ tyrosine phosphorylation also was considered. PKC␦ was im-munoprecipitated from cardiomyocytes exposed to vehicle or H 2 O 2 (in the presence or absence of PP1) and analyzed using an anti-YP antibody. In this context, H 2 O 2 -dependent PKC␦ tyrosine phosphorylation (at all potential sites recognized by the anti-YP antibody) is completely blocked by PP1 (Fig. 4C). The observation that PP1 prevents H 2 O 2 -dependent PKC␦ tyrosine phosphorylation but it does not prevent the H 2 O 2 -dependent release of PKC␦ from membranes argues that the H 2 O 2 -dependent release of PKC␦ from membranes is via a mechanism that does not involve PKC␦ tyrosine phosphorylation.
Src Kinase Activation and Complex Formation with PKC␦-Neonatal rat cardiomyocyte cultures express multiple Src family kinases that might mediate PKC␦ tyrosine phosphorylation (25). To begin to explore the mechanism(s) for PKC␦ tyrosine phosphorylation in cardiomyocytes treated with H 2 O 2 or PMA, we examined Src family kinase expression, partitioning between soluble and particulate fractions, and activity in resting and activated cardiomyocytes. Fig. 5A shows that Src, Fyn, Yes, and Lyn are readily detected, and recovered largely in the particulate fraction, in resting cardiomyocytes. Whereas the partitioning of each of these proteins between soluble and particulate fractions is not altered by PMA, treatment with H 2 O 2 consistently increased the recovery of Lyn (and more variably Src) in the soluble fraction. Yes and Fyn partitioning between soluble and particulate fractions were not influenced by H 2 O 2 .
Immunoblot analysis with anti-phospho-Src-Y 416 was used as a screen to detect H 2 O 2 -or PMA-dependent changes in Src kinase activity. This antibody specifically recognizes the activated (activation loop phosphorylated) form of Src; anti-phospho-Src-Y 416 also cross-reacts with activated forms of Fyn, Lyn, and Yes (which are phosphorylated at equivalent activation loop residues). Fig. 5B shows that anti-Src-pY 416 immunoreactivity is at the limits of detection in resting cultures. Anti-Src- pY 416 immunoreactivity increases rapidly in the particulate fraction (maximal activation detected at 2 min, the first time point assayed) when cardiomyocytes are treated with H 2 O 2 . Src-pY 416 immunoreactivity also increases in the soluble fraction, but with a slight delay relative to the increase in membranes. Of note, Src-pY 416  To determine whether there are potential differences in the activation kinetics for individual Src kinase family members, Src family kinases were immunoprecipitated individually from vehicle-or H 2 O 2 -treated cardiomyocyte cultures and Src-pY 416 immunoreactivity was monitored as a measure of their activity. These experiments used whole cell lysates and track Src activation loop phosphorylation throughout the cell (without resolving potential differences in soluble and particulate fractions). Fig. 5C shows that H 2 O 2 markedly increases Src, Lyn, and Fyn activation loop phosphorylation. In each case, the H 2 O 2 -dependent response is maximal by 2 min and sustained for an additional 20 min. In contrast, Src family kinase activation loop phosphorylation is not increased by PMA. Similar results were obtained in separate experiments that used immune complex kinase assays with enolase as substrate (data not shown).
Interactions between PKC␦ and several Src family kinases have been reported. To determine whether Src kinases interact (either in a constitutive or regulated fashion) with PKC␦ in cardiomyocytes, lysates from cardiomyocyte cultures treated with vehicle, PMA, or H 2 O 2 were subjected to immunoprecipitation with antibodies to Src, Fyn, Yes, or Lyn followed by immunoblot analysis to probe for PKC␦. Fig. 6 shows that Src, and to a lesser extent other Src family kinase members, are recovered constitutively in complexes with PKC␦. H 2 O 2 induces a dramatic increase in complex formation between PKC␦ and Fyn, Yes, and Lyn. PKC␦ complexes with Src, Fyn, and Yes are not influenced by PMA. Of note, PMA consistently decreased complex formation between Lyn and PKC␦. A similar effect of PMA to disrupt PKC␦-Lyn complexes was recently identified in mast cells (where it was attributed to PKC-dependent phosphorylation of Lyn, (11)). Src family kinase interactions with PKC␦ are isoform specific; PKC⑀ or PKC␣ were not detected in these complexes.
The marked H 2 O 2 -dependent increase in Src-pY 416 immunoreactivity (and Src kinase family member complex formation with PKC␦) provides a plausible mechanism to explain H 2 O 2dependent tyrosine phosphorylation of PKC␦. However, a mechanism for PKC␦-Y 311 phosphorylation in cells treated with PMA (where Src kinases are not detectably activated) is less obvious. To explore the mechanism for PMA-dependent PKC␦-Y 311 phosphorylation, we performed reciprocal immunoprecipitation/immunoblotting experiments. PKC␦ was immunoprecipitated from quiescent cultures and subjected to kinase assays with [␥-32 P]ATP in the absence or presence of sonicated vesicles containing the PKC-activating lipid cofactors phosphatidylserine [PS] plus PMA (PS/PMA). This protocol allows for PKC␦ autophosphorylation as well as PKC␦ phosphorylation in trans by associated kinases that co-purify during immunoprecipitation. Samples were subjected to SDS-PAGE followed by autoradiography (as a general measure of phosphorylation at serine, threonine, or tyrosine residues) as well as immunoblot analysis with anti-YP and anti-PKC␦-pY 311 antibodies (to resolve total and Tyr 311 -specific tyrosine phosphorylation). Fig. 7 shows that PKC␦ is recovered from quiescent cardiomyocytes with only a low level of tyrosine phosphorylation (including at Tyr 311 ). In vitro kinase assays without added lipid resulted in a low level of 32 P incorporation into PKC␦ (detected by autoradiography) and PKC␦ tyrosine phosphorylation (detected by the anti-YP antibody). PKC␦ radiolabeling was enhanced further by added PS/PMA. Of note, the PS/PMAdependent increase in 32 P incorporation into PKC␦ was blocked by GF109203X (but not PP1); this result indicates that 32 Pincorporation largely provides a measure of PKC-dependent autophosphorylation. In contrast, PS/PMA increases PKC␦ ty-rosine phosphorylation (including at Tyr 311 ); the PS/PMA-induced tyrosine phosphorylation of PKC␦ is blocked by PP1 (but not GF109203X). Collectively, these results identify distinct mechanisms for in vitro PKC␦ phosphorylation at serine/threonine versus tyrosine residues. The phosphorylation of PKC␦ at Tyr 311 (and perhaps other tyrosine residues) during the in vitro kinase assays validates the conclusion that PKC␦ is phosphorylated by one (or more) Src family kinase(s) that constitutively associates (and co-precipitates) with PKC␦ from resting cardiomyocyte cultures. In fact, both Src and Lyn were identified by immunoblot analyses in these anti-PKC␦ immune complexes (data not shown). Because PS/PMA allosterically activates PKC (but does not activate Src kinases), the mechanism for the PS/PMA-induced increase in PKC␦ tyrosine phosphorylation is presumed to involve a PS/PMA-induced conformational change that renders PKC␦ a better substrate for phosphorylation by one or more precomplexed Src kinase.  32 P incorporation into PKC␦ was detected by autoradiography; tyrosine phosphorylation in general, phosphorylation at PKC␦-Y 311 , and total PKC␦ protein recovery were detected by immunoblot analysis as indicated. Arrowheads denote the mobility of PKC␦. Results were replicated in experiments on two separate culture preparations.

FIG. 6. Src family kinase interactions with PKC␦ in cardiomyocytes.
Cardiomyocyte cultures were serumstarved for 24 h followed by stimulation with vehicle, 100 nM PMA, or 5 mM H 2 O 2 for 15 min. Whole cell lysates were subjected to immunoprecipitation with the indicated antibodies for individual Src kinases followed by immunoblot analysis for PKC␦, PKC⑀, or PKC␣ (with immunoblot analysis for cognate Src family kinase proteins as loading controls). All results were replicated in 4 -5 separate culture preparations.
Tyrosine Phosphorylation Regulates the Kinase Activity (but not the PMA-dependent Down-regulation Kinetics) for PKC␦-PKC␦-pY 311 has been viewed as a modification that destabilizes the enzyme and promotes down-regulation (23). However, Fig. 8 shows similar kinetics for PKC␦ down-regulation in cardiomyocytes exposed to PMA alone and PMA plus PP1 (to prevent Tyr 311 phosphorylation); the down-regulation kinetics for PKC␣ (examined in parallel) also are not influenced by PP1. These results argue that Src-dependent PKC␦ phosphorylation at Tyr 311 does not appreciably influence PMA-dependent PKC trafficking/down-regulation mechanism(s) in cardiomyocytes.
Tyrosine phosphorylation also has been implicated as a mechanism to regulate PKC kinase activity. Therefore, we compared the catalytic activity of PKC␦ recovered from quiescent cultures and cultures treated with PMA or H 2 O 2 ; assays were performed under basal conditions and with lipid (PS/ PMA) added to the in vitro incubations. To determine whether PKC␦ activation/phosphorylation leads to a change in substrate specificity, kinase assays measured the incorporation of 32 P i from [␥-32 P]ATP into ␦-peptide and ⑀-peptide (model substrates for PKC␦ and PKC⑀, respectively) and histone (generally considered to be a better substrate for cPKC isoforms, than for nPKC isoforms). Fig. 9A shows that PKC␦ immunopurified from quiescent cardiomyocyte cultures exhibits little-to-no lipid-independent kinase activity toward any substrate; histone-, ␦-peptide-, and ⑀-peptide-kinase activities increase markedly upon addition of PS/PMA to the in vitro kinase assays (although as recently reported, 32 P i incorporation into ␦-peptide is surprisingly modest, compared with the activity measured with ⑀-peptide or histone as substrates (4)). PKC␦ is recovered from PMA-and H 2 O 2 -treated cultures with markedly different cofactor requirements, substrate specificity, and activity. Here, lipid-independent kinase activity is markedly increased. With ␦-peptide or ⑀-peptide as substrates, the lipid-independent kinase activity of the in vivo stimulated enzyme (recovered from PMA-or H 2 O 2 -stimulated cardiomyocytes) approaches the in vitro PS/PMA-stimulated (maximally activated) PKC␦ kinase activity recovered from resting cardiomyocytes. Although lipidindependent histone kinase activity also is significantly increased by in vivo treatment with PMA or H 2 O 2 , addition of lipid cofactors leads to little further increase in histone phosphorylation. Consequently, the absolute level of histone kinase activity (even in the presence of PS/PMA) for PKC␦ recovered from H 2 O 2 -activated cardiomyocytes is quite low (Ͻ10% of the PS/PMA-stimulated histone kinase activity of PKC␦ recovered from quiescent cultures). Fig. 9B shows that H 2 O 2 increases PKC␦ kinase activity in both soluble and particulate fractions. H 2 O 2 -dependent changes in PKC␦ kinase activity are illustrated in this figure, without correcting for the H 2 O 2 -dependent translocation of PKC␦ from the particulate to the soluble fraction. This serves to underestimate the H 2 O 2 -dependent increase in particulate PKC␦ kinase activity. However, the H 2 O 2 -dependent increase in PKC␦ kinase activity in the soluble fraction could result from an increase in PKC␦ recovery in the soluble fraction, activation of the soluble pool of PKC␦, or (more likely) both, because the 3.5-and 2.8-fold increment in ␦and ⑀-peptide kinase activity in the soluble fraction exceeds the ϳ2-fold increased in soluble PKC␦ protein recovery.
H 2 O 2 can modulate PKC␦ kinase activity indirectly by stimulating Src kinase-dependent tyrosine phosphorylation or directly through oxidative modification of cysteines in the regulatory and catalytic domains. Oxidation of reactive cysteines in the zinc fingers of the regulatory C1 domain induces a conformational change that relieves autoinhibition and leads to a constitutively active enzyme. Oxidation of critical cysteines in the catalytic domain has the opposite effect and disrupts catalytic activity (26). Fig. 9C shows that H 2 O 2 , directly added to the in vitro assay buffer according to conditions used for the intact cell experiments (in an attempt to mimic the amount of oxidant stress achieved in vivo in cardiomyocytes), induces a significant increase in both lipid-independent and PS/PMA-dependent PKC␦ kinase activity; under these in vitro assay conditions, the stimulatory effects of PMA and H 2 O 2 are additive. This result is surprising, because activating oxidative modifications in the regulatory domain of PKC␦ are predicted to produce a form of PKC␦ that no longer binds phorbol esters (i.e. is not further activated by PMA). These results suggest that additional types of redox modifications may exist for PKC␦ and that oxidative modifications may contribute to PKC␦ activation in H 2 O 2 -treated cardiomyocytes.
Finally, in vitro kinase assays were performed with recombinant PKC␦ (rPKC␦) and active Src to resolve tyrosine phosphorylation-dependent changes in PKC␦ kinase activity. PKC␦ was preincubated with Src (in the presence of [␥-32 P]ATP, to detect PKC␦ and Src phosphorylation) and then kinase assays were initiated by the addition of substrate (histone, ␦-peptide, or ⑀-peptide). In assays performed with PS/ PMA, lipid cofactors were present throughout the preincubation and kinase assay. Aliquots from each assay were counted for radioactivity, as a measure of PKC␦ kinase activity. The remainder of the sample was then subjected to SDS-PAGE followed by autoradiography (to detect 32 P incorporation into PKC␦ and Src) and immunoblot analysis with anti-YP and anti-PKC␦-pY 311 (to detect tyrosine phosphorylation of Src and to resolve total and Tyr 311 -specific tyrosine phosphorylation of PKC␦). The ␦-peptide and ⑀-peptide sequences lack tyrosine residues and are not substrates for Src. The autoradiogram and Western blots from incubations with histone are presented to address the potential concern that a Src-induced increment in 32 P incorporation into histone results from tyrosine phosphorylation (rather than a Src-induced increase in PKC␦ kinase activity). Fig. 10A (left) shows that PS/PMA markedly increases 32 P incorporation into histone (detected as major and minor bands, denoted by asterisks) and PKC␦; 32 P incorporation into Src is similar in incubations without and with PS/ PMA. Differences in 32 P incorporation into histone are attributable to changes in PKC␦ kinase activity (and not histone phosphorylation by Src) because no tyrosine phosphorylation of histone is detected (Fig. 10A, right). 32 P incorporation into PKC␦ also largely provides a measure of autophosphorylation, because it is detected in incubations without Src. However, immunoblot analysis with anti-YP and anti-PKC␦-pY 311 shows that Src also promotes PKC␦-tyrosine phosphorylation, including at Tyr 311 . Fig. 10B shows that Src-dependent tyrosine phosphorylation influences PKC␦ catalytic activity (with the magnitude of the effect varying for different substrates). Tyrosine phosphorylation leads to a ϳ50% increase in lipid-independent catalytic activity toward ⑀-peptide and histone. Maximal PS/PMA-stimulated PKC␦ kinase activity toward histone also is increased, whereas maximal PS/PMA-dependent phosphorylation of ⑀-peptide is not. The robust PMA-dependent histone kinase activity displayed by Src-phosphorylated rPKC␦ contrasts with the very low level of histone phosphorylation (even in the presence of lipid cofactors) by PKC␦ recovered from H 2 O 2treated cardiomyocytes (Fig. 9A). These results suggest that PKC␦-Y 311 phosphorylation alone is not sufficient to impair histone kinase activity, and that treatment with H 2 O 2 leads to FIG. 9. In vitro PKC␦ kinase assays; PKC␦ is recovered from PMA-and H 2 O 2 -activated cardiomyocytes with altered activity, cofactor requirements, and substrate-specificity. Panel A, cultures were treated with vehicle, PMA (300 nM) or H 2 O 2 (5 mM, each for 10 min); whole cell lysates were subjected to immunoprecipitation and immune complex PKC␦ kinase assays with histone, ␦-peptide, or ⑀-peptide as substrate. Assays were in the absence or presence of PS/PMA as described under "Experimental Procedures." Panel B, cultures were treated for 10 min with vehicle or H 2 O 2 . Lysates were partitioned into soluble and particulate fractions and subjected to immunoprecipitation and immune complex PKC␦ kinase assays (without added PS/PMA) with ␦-peptide or ⑀-peptide as substrate. Samples also were subjected to immunoblot analysis for total PKC␦ and PKC␦-pY 311 . Panel C, cultures were treated for 10 min with vehicle or 5 mM H 2 O 2 . Whole cell lysates were subjected to immunoprecipitation and immune complex PKC␦ kinase assays were performed in the absence or presence of PS/PMA or 1 mM H 2 O 2 with ␦-peptide as substrate; similar results were obtained in separate assays that used ⑀-peptide as substrate, (data not shown). Results for kinase assays are mean ϩ S.E. of quadruplicate measurements from a representative experiment, with similar result obtained in three (panels A and B) or two (panel C) separate experiments on separate culture preparations. defective histone phosphorylation via a different modification on PKC␦ (that could include tyrosine phosphorylation of PKC␦ at other tyrosine residues, perhaps by other non-receptor tyrosine kinases). Finally, Fig. 10B shows that tyrosine phosphorylation of PKC␦ markedly augments lipid-independent and PS/PMA-dependent PKC␦ kinase activity toward ␦-peptide (34and 5-fold, respectively). Collectively, these results indicate that PKC␦ is phosphorylated at Tyr 311 by Src, that Src-dependent phosphorylation of PKC␦ at Tyr 311 (and potentially other tyrosine residues) leads to an increase in the catalytic activity of PKC␦, and that the magnitude of the increment is skewed by the choice of substrate used in the assay; the effect is most striking in assays with a substrate that conforms to an optimal PKC␦ phosphorylation motif. DISCUSSION PKC regulates a wide range of functions in cardiomyocytes. Signaling specificity for PKC generally has been attributed to the co-expression of multiple isoforms with distinct modes of activation, subcellular localizations, and/or target substrates. However, this study suggests that (at least for PKC␦) this model must be broadened to consider distinct modes of activation and functions for a single PKC isoform. In addition to the traditional mode of allosteric activation by lipid cofactors, PKC␦ is regulated by oxidative stress and tyrosine phosphorylation. This mechanism holds particular interest for cardiomyocyte physiology, where the traditional model of receptor-driven lipid cofactor-dependent PKC␦ activation readily accounts for substrate phosphorylation in lipid membranes, but it does not adequately explain the well known effects of PKC to modulate contractile function by phosphorylating proteins in the sarcomere (27). This study provides a viable solution to this dilemma by identifying conditions that allow for lipid-independent PKC␦ kinase activity in the soluble fraction of cardiomyocytes.
Three distinct stimulus-specific patterns for PKC␦ phosphorylation in cardiomyocytes treated with NE, PMA, and H 2 O 2 were identified. All three stimuli increase PKC␦ tyrosine phosphorylation, but only PMA and H 2 O 2 promote PKC␦-Y 311 phosphorylation. Because NE does not increase PKC␦-pY 311 , it must promote a tyrosine phosphorylation event elsewhere in the protein. Similarly, H 2 O 2 and PMA promote quantitatively similar increases in PKC␦-Y 311 phosphorylation, but overall PKC␦ tyrosine phosphorylation (detected with anti-YP) is substantially higher in cardiomyocytes treated with H 2 O 2 . This cumstances (such as ischemia), which release PKC␦ from membranes (and free it of its lipid requirement). Substrate phosphorylations that occur in the soluble fraction of H 2 O 2 -treated cells would not be blocked by peptide inhibitors that prevent RACK-driven PKC isoform compartmentation. In view of the accumulating evidence that oxidant stress is an important feature of heart failure syndromes and many cardiovascular risks (13), it will be important to consider the distinct PKC␦ signaling mode triggered by oxidative stress (and its consequences, which are predicted to differ from the actions of PKC␦ during growth factor signaling) in future efforts to design and evaluate PKC␦-targeted therapeutics.