Functional Interaction of Protein Kinase Cα with the Tyrosine Kinases Syk and Src in Human Platelets*

There is a high degree of cross-talk between tyrosine phosphorylation and the serine/threonine phosphorylation signaling pathways. Here we show a physical and functional interaction between the classical protein kinase C isoform (cPKC), PKCα, and two major nonreceptor tyrosine kinases in platelets, Syk and Src. In the presence of the cPKC-selective inhibitor Gö6976, platelet 5-hydroxytryptamine release was abolished in response to co-activation of glycoproteins VI and Ib-IX-V by the snake venom alboaggregin A, whereas platelet aggregation was substantially inhibited. Of the two platelet cPKCs, PKCα but not PKCβ was activated, occurring in an Syk- and phospholipase C-dependent manner. Syk and PKCα associate in a stimulation-dependent manner, requiring Syk but not PKC activity. PKCα and Syk also co-translocate from the cytosol to the plasma membrane upon platelet activation, in a manner dependent upon the activities of both kinases. Although PKCα is phosphorylated on tyrosine downstream of Syk, we provide evidence against phosphorylation of Syk by PKCα, consistent with a lack of effect of PKCα inhibition on Syk activity. PKCα also associates with Src; although in contrast to interaction with Syk, PKCα activity is required for the association of these kinases but not the stimulation-induced translocation of Src to the cell membrane. Finally, the activity of Src is negatively regulated by PKC, as shown by potentiation of Src activity in the presence of the PKC inhibitors GF109203X or Gö6976. Therefore, there is a complex interplay between PKCα, Syk, and Src involving physical interaction, phosphorylation, translocation within the cell, and functional activity regulation.

The protein kinase C (PKC) 1 family comprises 10 isozymes grouped into the following three classes: conventional (␣, ␥, ␤I, and ␤II), novel (␦, ⑀, /L, and ), and atypical ( and /). In addition, PKC is considered as a fourth class, now generally referred to as a separate family termed protein kinase D (1). The PKC family has long been known to be involved in a number of platelet processes, most importantly aggregation and secretion, where stimulation of platelets with diacylglycerol (DAG) or phorbol ester can induce aggregation, and agonist-induced secretion can be prevented by pharmacological inhibition of a broad range of PKC isoforms (2)(3)(4)(5).
At least seven PKC isoforms (␣, ␤, ␦, , ⑀, , and ) are expressed in platelets (6 -12), and it is becoming clear that each isoform may play different roles in platelet function and may have different modes of activation and downstream targets. PKC␣ has been identified recently as an essential factor in positively regulating ␣and dense granule secretion in platelets (13) as well as platelet aggregation (14). We were therefore interested to determine how PKC␣ activity may be modulated and its signaling role in platelets.
There are now known to be a variety of different mechanisms by which PKC activity and localization may be regulated, important among which is the phosphorylation of serine, threonine, and more recently tyrosine residues (15). Regulation of novel PKC isoforms by tyrosine phosphorylation has been well characterized in a variety of cell types and in response to a variety of stimuli (7, 16 -30). We have shown recently (6,31) that the novel isoforms PKC and PKC␦ may be phosphorylated on tyrosine through physical and functional associations with the nonreceptor tyrosine kinases Btk and Fyn, respectively. Some reports have also documented tyrosine phosphorylation of PKC␣ in several other cell types and in response to a variety of stimuli including insulin (32), ␤ 1 integrin ligation (33), and oxidative stress induced by H 2 O 2 in COS-7 cells (34). This last report concludes that tyrosine phosphorylation of PKC␣ recovered from these cells was catalytically active independent of DAG binding, indicating a possible pathway for early activation of PKC not dependent on PLC hydrolysis of inositol phospholipid. It has also been shown that PKC␣ may be positively regulated by Src (35), and therefore there is a precedent in the literature for regulation of this classical isoform by tyrosine phosphorylation.
Here we were interested to investigate the mutual regulation of PKC␣, Syk, and Src kinases in human platelets. We show that the two tyrosine kinases Syk and Src physically interact with PKC␣ leading to distinct functional consequences. Syk-PKC␣ interaction does not depend upon PKC␣ activity in contrast to Src-PKC␣, interaction of which does depend upon the activity of PKC␣. Although PKC␣ activity was dependent upon Syk, as may be expected because of the proximal role played by Syk in GP VI-dependent signaling (36), there was no reciprocation because Syk activity was not regulated by PKC␣. However, Src activity was negatively regulated by PKC␣. The results suggest that PKC␣, Syk, and Src interact to regulate each other and cellular activities in platelets EXPERIMENTAL PROCEDURES Materials-Trimeresurus albolabris venom was a kind gift from Professor R. G. D. Theakston (Liverpool, UK). Alboaggregin A was pre-pared from venom by ion-exchange chromatography as described previously (37). Protein A-agarose and protein G-Sepharose were obtained from Oncogene Research Products (San Diego, CA); complete mini protease inhibitor tablets were from Roche Applied Science, and protein phosphatase cocktails I and II were from Sigma. Anti-PKC␣, anti-PKC␤, anti-PKC␥, and mouse anti-Syk antibodies were purchased from BD Biosciences; and anti-phospho-Ser PKC substrate, anti-phosphothreonine, anti-phospho-Syk (Tyr-525/Tyr-526), and anti-phospho Src family (Tyr-416) antibodies were from Cell Signaling Technology (New England Biolabs, Hitchin, UK). Anti-phosphotyrosine (4G10) monoclonal antibody, anti-Src antibody, and dephosphorylated MBP were obtained from Upstate Biotechnology Inc. (Lake Placid, NY); and anti-Fyn, anti-Lyn, rabbit anti-Syk, and rabbit anti-Src were from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). TRITC-conjugated antimouse IgG and FITC-conjugated anti-rabbit IgG antibodies were from Molecular Probes (Eugene, OR). Gö6976, GF109203X, and U73122 were obtained from Tocris (Avonmouth, UK); piceatannol and PP1 were from Alexis Corp. (Nottingham, UK); and myo-[ 3 H]inositol, [ 3 H]5-HT, [␥-32 P]ATP, and enhanced chemiluminescence (ECL) kits were from Amersham Biosciences. Raytide TM peptide was from Calbiochem, and AG 1-X8 resin columns were from Bio-Rad. All other reagents were of analytical grade.
Preparation of Human Platelets-Human blood was drawn from healthy, drug-free volunteers on the day of the experiment. Acid citrate dextrose (120 mM sodium citrate, 110 mM glucose, 80 mM citric acid, used at 1:7 v/v) was used as anticoagulant. Platelet-rich plasma was prepared by centrifugation at 200 ϫ g for 20 min, and platelets were then isolated by centrifugation for 10 min at 400 ϫ g in the presence of prostaglandin E 1 (40 ng/ml). The pellet was resuspended to a density of 4 ϫ 10 8 platelets/ml in a modified Tyrode's-HEPES buffer (145 mM NaCl, 2.9 mM KCl, 10 mM HEPES, 1 mM MgCl 2 , 5 mM glucose, pH 7.3). 10 M indomethacin was added to this platelet suspension, and a 30-min rest period was allowed before stimulation.
Measurements of Platelet Aggregation-Platelets were prepared as described above, preincubated with different inhibitors or vehicle solution (0.1% Me 2 SO final concentration) for 10 min at 37°C, and stimulated in an aggregometer (Chrono-Log Corp., Labmedics Ltd., Manchester, UK) at 37°C, with continuous stirring at 800 rpm. Alboaggregin A was used at 1 g/ml, a concentration determined previously (37) for the induction of 5-HT release. Platelet aggregation was followed by turbidimetric aggregometry over a 1-min period.
Measurement of 5-HT Release-Platelets were loaded by incubation of platelet-rich plasma with 0.5 Ci/ml [ 3 H]5-HT for 1 h at 37°C. Platelets were then pelletted and resuspended in Tyrode's-HEPES buffer. After incubation with different inhibitors or vehicle solution (0.1% Me 2 SO final concentration) for 10 min at 37°C, platelets were activated and rapidly centrifuged, and the [ 3 H]5-HT released into the supernatant was determined by scintillation counting and expressed as a percentage of the total cell content, as described previously (38).
Protein Immunoprecipitation-Following activation, platelet suspensions were generally lysed with an equal volume of 2ϫ radioimmunoprecipitation assay buffer (2% Triton X-100, 2% sodium deoxycholate, 0.2% SDS, 300 mM NaCl, 20 mM Tris, 1 mM phenylmethylsulfonyl fluoride, 10 mM EDTA). Alternatively, for in vitro kinase assays, 2ϫ Nonidet P-40 (1% Nonidet P-40, 300 mM NaCl, 20 mM Tris, 1 mM phenylmethylsulfonyl fluoride, 10 mM EDTA) was used. In both lysis . Platelets were stimulated with alboaggregin A (1 g/ml), and aggregation was assessed by turbidimetric aggregometry over a period of 3 min. Data shown are from one experiment representative of at least three independent experiments. D, concentration-response curve for Gö6976-mediated inhibition of platelet aggregation. Platelets were pretreated for 10 min with various concentrations of Gö6976 and stimulated with alboaggregin A (1 g/ml; 3 min). Inhibition of platelet aggregation by Gö6976 from four independent experiments is expressed as mean Ϯ S.E. (IC 50 ϭ 390 nM). buffers, protease activity was inhibited with complete mini protease inhibitor tablets, whereas phosphatase activity was inhibited with protein phosphatase cocktails I and II. Lysates were precleared with protein A-agarose or protein G-Sepharose for 3 h at 4°C, before overnight incubation with antibody at 4°C. Antibody-protein complexes were precipitated by incubation with protein A-agarose or protein G-Sepharose for 2 h at 4°C. Beads were then washed with Tris-buffered saline plus Tween (0.1%) before addition of 2ϫ Laemmli sample solvent and boiling for 5 min.
SDS-PAGE and Immunoblot-Proteins were resolved by electrophoresis in 8 -10% SDS-polyacrylamide gels. Samples were then transferred to polyvinylidene difluoride membranes, blocked with 5-10% bovine serum albumin, and incubated for 1 h or overnight at room temperature with the appropriate primary antibody. Membranes were then washed before incubation with the appropriate secondary antibody followed by thorough washing. Bound peroxidase activity was detected using enhanced chemiluminescence (ECL).
In Vitro Kinase Assay for Protein Kinase C-PKC isoforms were immunoprecipitated from platelets lysed into Nonidet P-40 buffer using 1 g/ml of selective antibody and resuspended in 20 l of kinase assay (KA) buffer (5 mM MgCl 2 , 5 mM MnCl 2 , 100 mM NaCl, 1 mM ATP, 2 mM Na 3 VO 4 , 20 mM HEPES, pH 7.2). After incubation for 10 min at 37°C, the reaction was terminated by addition of 0.5 ml of ice-cold EDTA (100 mM). Immunoprecipitated proteins were then washed in radioimmunoprecipitation assay buffer before separation by SDS-PAGE and immunoblot with anti-phosphoserine PKC substrate or anti-phosphothreonine antibodies (1:1000). In this way the assay is a direct, nonradioactive in vitro kinase assay for PKC.
Inositol Phosphate Accumulation Assay-Platelets were loaded by incubation of platelet-rich plasma with 50 Ci/ml myo-[ 3 H]inositol for 3 h at 30°C. After addition of LiCl to a final concentration of 20 mM, platelets were separated from the supernatant by centrifugation and resuspended in Tyrode's-HEPES buffer containing 20 mM LiCl. After incubation with different inhibitors or vehicle solution (0.1% Me 2 SO final concentration) for 10 min at 37°C, platelets were activated and reactions stopped by addition of an equal volume of 0.8 M perchloric acid. 350 l of 0.72 N KOH, 0.6 M KHCO 3 solution was added to neutralize the sample solution, and samples were centrifuged for 5 min at 20,000 ϫ g and inositol phosphates were separated on AG 1-X8 resin columns. Total labeled inositol phosphates were then determined by liquid scintillation counting, and data are presented as counts/min.
Immunofluorescence Confocal Imaging-Platelets were prepared as described, before being pretreated with antagonists and stimulated. Reactions were terminated by addition of 4% paraformaldehyde in phosphate-buffered saline (PBS: 137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 ). Platelets were pelleted by centrifugation at 4000 rpm for 2 min in a microcentrifuge and washed twice in PBS. Platelets were immobilized on poly-L-lysine-coated coverslips overnight, permeabilized by incubation of coverslips with 0.05% Triton X/PBS at room temperature for 10 min, and incubated for 30 min at room temperature with 1% bovine serum albumin (BSA) in PBS to block nonspecific antibody binding. Samples were then incubated for 3 h in 1% BSA/PBS at room temperature with primary antibodies; coverslips were washed in PBS, 0.05% Triton X-100 and incubated for 30 min at room temperature with 1% BSA in PBS. Fluorescein isothiocyanate (FITC)-labeled anti-rabbit IgG secondary antibody and tetramethylrhodamine isothiocyanate (TRITC)-labeled anti-mouse IgG secondary an-

FIG. 2. PKC␣ but not PKC␤ activation is induced upon platelet activation by alboaggregin A.
A, platelets were lysed into Laemmli sample buffer (WCL) or 1% Nonidet P-40 lysis buffer, and PKC␣ or PKC␤ was immunoprecipitated (IP) from lysates. Proteins were separated by SDS-PAGE, transferred to polyvinylidene difluoride membrane, and immunoblotted for either PKC␣ or PKC␤ as indicated. Data shown are representative of three independent experiments. WB, Western blot. B and C, platelets were treated for 1 min with either 1 g/ml alboaggregin A (Albo. A) or vehicle (water), and the reaction was stopped by lysis into 1% Nonidet P-40 buffer as described under "Experimental Procedures." PKC␣ or PKC␤ was immunoprecipitated from lysates and resuspended in kinase assay (KA) buffer containing 5 g of dephosphorylated MBP and incubated for 10 min at 37°C. MBP samples were immunoblotted with 1:1000 anti-phospho-Ser PKC substrate antibody (B, p-(Ser)PKC) or 1:1000 anti-phospho-Thr antibody (C, p-(Thr)). Data shown are representative of three independent experiments. D, time dependence of PKC␣ activation by alboaggregin A. Platelets were treated for 1 min with vehicle solution (water) or for 5, 15, 30, and 60 s with 1 g/ml alboaggregin A, and the reaction stopped by lysis into ice-cold 1% Nonidet P-40 buffer. PKC␣ was immunoprecipitated from platelet lysates, and in vitro kinase assay was performed as described for B and C. Ser phosphorylation of MBP was assessed by immunoblotting with anti-phospho-Ser PKC substrate antibody. Data shown are representative of three independent experiments. E, platelets were pretreated for 10 min with Gö6976 (1 M) or vehicle solution (0.1% Me 2 SO final concentration). Samples were then stimulated for 1 min with either 1 g/ml alboaggregin A or vehicle (water) and lysed into ice-cold 1% Nonidet P-40 buffer, and in vitro kinase assay for immunoprecipitated PKC␣ was performed as described for B and C. Ser phosphorylation of MBP was assessed by immunoblotting with anti-phospho-Ser PKC substrate antibody. Data shown are representative of three separate experiments.

FIG. 3. PKC␣ activity is regulated by Syk and PLC. A and B, PKC␣ activation is abolished by Syk inhibition and reduced by PLC inhibition.
Platelets were pretreated for 10 min with piceatannol (10 g/ml), U73122 (10 M), or vehicle solution (0.1% Me 2 SO final concentration). Platelets were then stimulated with alboaggregin A (Albo. A) (1 g/ml) for 1 min, or control platelets were treated with vehicle solution (water). Reactions were stopped by lysis into ice-cold 1% Nonidet P-40 buffer, and immunoprecipitated (IP) PKC␣ was resuspended in kinase buffer containing 5 g of dephosphorylated MBP and incubated for 10 min at 37°C. Samples were separated by SDS-PAGE and immunoblotted with 1:1000 antiphospho-Ser PKC substrate antibody. Data shown in A are representative of four independent experiments, quantified by densitometry, and represented in B. WB, Western blot. B, shows data as mean Ϯ S.E. (* ϭ p Ͻ 0.01 and ** ϭ p Ͻ 0.001, compared with platelets pretreated with vehicle solution and stimulated with alboaggregin A). C, effect of piceatannol and U73122 on inositol phosphate (IP) accumulation. Platelets were labeled with myo-[ 3 H]inositol, pretreated for 10 min with piceatannol (10 g/ml), U73122 (10 M), piceatannol (10 g/ml) plus U73122 (10 M), or vehicle solution (0.1% Me 2 SO final concentration) and stimulated with alboaggregin A (1 g/ml) for 1 min, whereas control platelets were treated with vehicle solution (water). Total inositol phosphates were determined as described under "Experimental Procedures," and data are expressed as mean Ϯ S.E. from three independent experiments (* ϭ p Ͻ 0.01, compared with platelets pretreated with vehicle solution and stimulated with alboaggregin A). D, platelets were stimulated with alboaggregin A (1 g/ml) for 1 min, or control platelets were treated with vehicle solution (water). Reactions were stopped by lysis into ice-cold 1% Nonidet P-40 buffer, and immunoprecipitated PKC␣ was resuspended in kinase buffer containing 5 g of dephosphorylated MBP in the presence of either piceatannol (10 g/ml), GF109203X (10 M), or vehicle solution (0.1% Me 2 SO final concentration) and incubated for 10 min at 37°C. Samples were separated by SDS-PAGE and immunoblotted with 1:1000 anti-phospho-Ser PKC substrate antibody. Data shown are representative of three independent experiments. E, effect of piceatannol and U73122 on Syk activity. Platelets were pretreated for 10 min with piceatannol (10 g/ml), U73122 (10 M), or vehicle solution (0.1% Me 2 SO final concentration) and stimulated with alboaggregin A (1 g/ml) for 1 min, or control platelets were treated with vehicle solution (water). Immunoprecipitated Syk was immunoblotted with anti-Syk Tyr(P)-525/526 antibody and reprobed with anti-Syk antibody as shown. Data shown are representative of three separate experiments. F and G, effect of piceatannol on Fyn and Lyn activity. Platelets were pretreated for 10 min with piceatannol (10 g/ml), PP1 (20 M), or vehicle solution (0.1% Me 2 SO final concentration) and stimulated with alboaggregin A (1 g/ml) for 1 min, or control platelets were treated with vehicle solution (water). Fyn (F) or Lyn (G) was immunoprecipitated, and their autophosphorylation was analyzed by immunoblotting with anti-phospho-Src family Tyr(P)-416 antibody. Data shown are representative of three independent experiments. tibody were then added at a concentration of 4 g/ml in PBS, 1% BSA for 1 h at room temperature. Subsequently, coverslips were washed four times in PBS, 0.05% Triton X-100 and mounted onto slides using a 13.5% Mowiol solution containing 2.5% 1,4-diazobicyclo-(2,2,2)octane to prevent bleaching of fluorescence. Platelets were imaged using a Leica TCS-NT confocal laser scanning microscope equipped with Kr/Ar laser attached to a Leica DM IRBE inverted epifluorescence microscope with phase contrast.
In Vitro Raytide Phosphorylation Assay-Src activity was assayed using Raytide peptide as an exogenous substrate. Immunoprecipitated kinase was resuspended in 20 l of KA buffer, and 10 g of Raytide peptide was added to each sample. The reaction was started by the addition of 10 l of ATP buffer (0.15 mM ATP, 30 mM MgCl 2 , and 200 Ci/ml [␥-32 P]ATP in KA buffer). After incubation at 30°C for 30 min, the reaction was terminated by addition of 10% phosphoric acid. Samples were applied to 2 ϫ 2-cm squares of P81 ion-exchange chromatography paper, extensively washed in 0.5% phosphoric acid followed by a wash in acetone. Papers were then dried, and labeled Raytide was quantified by liquid scintillation counting.
Statistics and Data Presentation-Bar graphs were obtained by using GraphPAD Prism (GraphPAD software), and where appropriate, statistical significance was assessed using one-way analysis of variance with Bonferroni's multiple comparison post-test. Log concentrationeffect curves were fitted to logistic expressions for sigmoidal concentration-effect curves with variable slope using GraphPAD Prism.

Dependence of Platelet Responses upon Classical PKC Isoforms-It is clear that PKC isoforms play important roles in regulating both secretion and aggregation events in platelets,
and there is evidence that classical isoforms may be involved in these processes (13,14). We chose to verify this involvement by using the PKC inhibitors Gö6976, which has selectivity for the classical isoforms, and GF109203X, which nonselectively inhibits PKC isoforms. Platelets were stimulated with 1 g/ml alboaggregin A, a purified lectin-type snake venom capable of binding and activating both GP Ib-IX-V and GP VI adhesion receptors for von Willebrand factor and collagen, respectively (39). As shown in Fig. 1A, preincubation with maximally effective concentrations of either Gö6976 (1 M) or GF109203X (20 M) abolished 5-HT release induced by platelet activation. The concentration dependence of Gö6976 for inhibition of 5-HT release shows an IC 50 of 180 nM (Fig. 1B). Although platelet aggregation was abolished by 20 M GF109203X, 1 M Gö6976 only partially inhibited this response (Fig. 1C). Again, the concentration dependence of Gö6976 for inhibition of this response was determined, and the IC 50 was shown to be 390 nM (Fig. 1D).

Selective Activation of PKC␣ by Alboaggregin A in Platelets-
The classical isoforms of PKC predominantly expressed in platelets are PKC␣ and -␤ (6, 7). By having determined the dependence of platelet responses upon classical isoforms of PKC, it was important to determine whether individual isoforms became activated upon platelet activation. Initially, it was important to confirm expression of both PKC␣ and PKC␤ in human platelets and to demonstrate that the antibodies used were isoform-selective. In Fig. 2A, we immunoprecipitated either PKC␣ or PKC␤ and blotted each immunoprecipitate with the antibodies for each isoform. The data show expression of each isoform in platelets and confirm that each antibody is only able to detect the single isoform it was raised against. PKC␣ or PKC␤ were then immunoprecipitated using these isoform-specific antibodies, and the kinase activity was assessed in vitro by using myelin basic protein (MBP) as a substrate. PKC-dependent phosphorylation of MBP was detected using anti-phospho-Ser PKC substrate antibody (Fig. 2B) or anti-phospho-Thr antibody (Fig. 2C). Stimulation of platelets induced phosphorylation of MBP by PKC␣ but not by PKC␤, as detected by anti-phospho-Ser PKC substrate antibody, suggesting that PKC␣ is selectively activated by alboaggregin A in platelets. The time dependence of PKC␣ activation by alboaggregin A was investigated by stimulating the platelets for different times, from 5 to 60 s. Kinase activity was measured as described by using anti-phospho-Ser PKC substrate antibody (Fig. 2D) and was shown to increase during the 60-s stimulation period. As a control, and in order to demonstrate that phosphorylation of MBP detected during in vitro kinase assay was mediated by classical PKCs, the assay was repeated in the absence or presence of Gö6976 (1 M). Fig. 2E shows that phosphorylation of MBP was markedly diminished in the presence of inhibitor, showing that it is mediated by classical PKC activity.
Regulation of PKC␣ Activity by PLC and Syk-The molecular mechanism underlying PKC␣ activation was next investigated. As signaling downstream of GP VI is mediated by activation of Syk kinase, it was important to determine the activation dependence of PKC␣ by Syk and to determine whether this dependence was mediated solely through activation of PLC␥ 2 or by additional Syk-dependent signaling. PKC␣ activity, measured in vitro as described above, was shown to be completely dependent upon the activity of Syk kinase, because it is fully blocked in the presence of the Syk inhibitor piceatannol (Fig. 3,  A and B). In the presence of the PLC inhibitor U73122, the activity was substantially diminished, but a PLC-independent, Syk-dependent component remained. This suggested that Syk- dependent signals, other than through activation of PLC␥ 2 by Syk, may play a role in regulating PKC␣ activity. In order to confirm that U73122 was able to fully block PLC under the conditions used in the experiment, and to show the dependence of PLC activity upon Syk activation in our system, we assessed inositol phosphate turnover directly in the absence or presence of the inhibitors. Fig. 3C shows that both U73122 and piceatannol abolished stimulation of PLC upon platelet activation.
By way of an additional control, Fig. 3D shows the effect on PKC␣ activity of piceatannol directly added to the in vitro kinase buffer. The inhibitor had no significant effect upon PKC␣ activity directly, suggesting no direct inhibition of PKC␣ by piceatannol. Fig. 3E shows that Syk activity, as assessed by autophosphorylation of Tyr-525/526 in the activation loop of the kinase (40), is also abolished by piceatannol, although the PLC inhibitor U73122 does not affect its activity. These data are consistent with the established signaling pathway from GP VI where Syk lies essentially upstream of PLC␥ 2 (36) . Fig. 3, F and G, shows that under our conditions piceatannol is also selective for Syk over members of the Src family kinases, because it does not inhibit either Fyn (Fig. 3F) or Lyn (Fig. 3G) activity, as assessed by autophosphorylation of each kinase detected by an antibody specific for phosphorylated tyrosine in the activation loop of the Src family kinase domain (anti-Srcphospho-Tyr-416) (41).
Activation-dependent Association between PKC␣ and Syk-As we had shown PKC␣ to be activated by Syk-dependent signals, both PLC-dependent and -independent, we were interested to determine whether the functional interaction of these two proteins also involves their association in an heteromeric complex. The cellular distribution of these two proteins was first analyzed by confocal microscopy. Fig. 4 shows that, in resting platelets, Syk and PKC␣ were broadly expressed in the platelet cytoplasm. Upon platelet activation,  /ml). A-C, reactions were stopped by lysis into ice-cold 1% Nonidet P-40 buffer, and PKC␣ was immunoprecipitated, separated by SDS-PAGE, and immunoblotted with 1:1000 anti-Syk antibody. The position of Syk on the blot is indicated. Both immunoprecipitating antibody (anti-PKC␣) and blotting antibody (anti-Syk) were derived from mice, generating a prominent IgG heavy chain band as indicated (IgG HC). A and B, the presence of PKC␤ was assessed by reblotting with 1 g/ml anti-PKC␤ antibody. C, Tyr phosphorylation of PKC␣ and associated Syk was assessed by reblotting with 1 g/ml anti-phospho-Tyr 4G10 antibody (center panel). D, PKC␣ is phosphorylated on tyrosine in an Syk-dependent manner. Platelets were pretreated for 10 min with piceatannol (10 g/ml) or vehicle solution (0.1% Me 2 SO final concentration) and stimulated with alboaggregin A (1 g/ml) for 1 min, or control platelets were treated with vehicle solution (water). Reactions were stopped by lysis into ice-cold 1% Nonidet P-40 buffer, and PKC␣ immunoprecipitates were immunoblotted with 1 g/ml anti-phospho-Tyr 4G10 antibody. A-D, immunoprecipitation of PKC␣ was confirmed by reblotting membranes with 1 g/ml anti-PKC␣ antibody. All data shown are representative of three independent experiments. both Syk and PKC␣ translocated from the cytoplasm to a sub-plasma membrane localization, where the two proteins were shown to co-localize. Pretreatment with either piceatannol or Gö6976 abolished the translocation of both Syk and PKC␣, suggesting that the simultaneous activity of both proteins is necessary for translocation of either protein in response to platelet activation.
The association between Syk and PKC␣ was also investigated by co-immunoprecipitation. The association between the two kinases depends upon the activity of Syk but not upon the activity of PKC␣, because piceatannol, but not U73122 or inhibitors of PKC, abolishes the interaction as assessed by coimmunoprecipitation (Fig. 5, A and B). As shown in Fig. 5C (top  panel), Syk is abundantly co-precipitated with PKC␣ after 5 s of alboaggregin A stimulation. Longer stimulations (up to 180 s) do not further increase the amount of Syk co-precipitated with PKC␣. Most interestingly, both PKC␣ and associating Syk are also rapidly tyrosine-phosphorylated in response to platelet activation (Fig. 5C (middle panel)), although phosphorylation of PKC␣ is weak by comparison with Syk. Fig. 5D shows that tyrosine phosphorylation of PKC␣ is dependent upon Syk activity, because piceatannol substantially inhibited tyrosine phosphorylation of PKC␣.
By having shown that Syk may feed forward to induce phosphorylation of PKC␣ on tyrosine and positively regulate the activity of PKC␣, it was also important to assess whether Syk was phosphorylated and/or reciprocally regulated by PKC␣. In  Fig. 6A, we have used the anti-phospho-Ser PKC substrate antibody, raised against a consensus sequence for phosphorylation by PKC, to show that although phosphorylation of Syk may be detected upon stimulation, this is not downstream of PKC␣ because there is no inhibition of phosphorylation by Gö6976. On the other hand, phosphorylation is abolished by the PLC inhibitor U73122 and partially inhibited by the broad spectrum PKC inhibitor GF109203X, suggesting that at least part of the phosphorylation event may be downstream of a nonclassical PKC isoform. Phosphorylation data correlated with assay of Syk activity, because Fig. 6B shows autophosphorylation of Tyr-525/526 is induced upon activation and is not affected by inhibition of classical PKC isoforms using Gö6976. There was a small increase in Syk activity upon broad spectrum inhibition of PKCs using GF109203X, suggesting a possible negative feedback on Syk activity by nonclassical PKCs.
Platelet Activation-dependent Association of PKC␣ and c-Src-As it had been shown previously that PKC␣ may also be positively regulated by another nonreceptor tyrosine kinase, Src (35), we were interested to address whether PKC␣ would associate with Src in platelets. Fig. 7A shows that PKC␣ does not co-immunoprecipitate with Src in resting platelets, but upon activation with alboaggregin A, the two kinases associate. This association may be seen in both Src and PKC␣ immunoprecipitates (shown in Fig. 7A, i and ii). Unlike for the PKC␣-Syk association, however, the interaction was dependent upon PKC activity, as pretreatment with GF109203X or Gö6976 abolished co-immunoprecipitation, possibly suggesting that PKC␣ needs to be in its active state to associate with Src. Consistent with this and with data from Fig. 3 showing that PKC␣ activity depends upon Syk activity, inhibition of Syk by piceatannol also substantially inhibited the association between PKC␣ and Src (Fig. 7B). In order to assess the selectivity of interaction between PKC␣ and Src, the presence of PKC␤ in Src immunoprecipitates was investigated and shown not to be present in either basal or stimulated conditions (Fig. 7C). Additionally, PKC␣ was also shown not to associate with other members of the Src family kinases, Fyn and Lyn (Fig. 7, D and   E), thereby demonstrating a degree of specificity in the association between PKC␣ and Src.
It was therefore important to determine whether associating Src was phosphorylated by PKC␣. We reprobed Src immunoprecipitates in Fig. 7A with anti-phospho-Ser PKC substrate antibody or an anti-phosphothreonine antibody, but as shown in the figure we could not detect phosphorylation of Src on these residues. This suggests that Src is not phosphorylated by PKC␣ either in resting platelets or upon platelet activation. Nevertheless, we cannot completely exclude the possibility that Src is phosphorylated by PKC␣ on a site that is not recognized by the antibodies used.
We also studied the translocation of PKC␣ and Src in the cell by confocal immunofluorescence microscopy. As shown in Fig.  8, these two kinases are sparsely distributed in the cytoplasm of resting platelets, but following platelet activation both PKC␣ and Src translocate to a sub-plasma membrane location where they appear co-localized. Pretreatment of platelets with PP1, an Src family kinase inhibitor, abolishes the translocation of both PKC␣ and Src. On the other hand, the classical PKCselective inhibitor Gö6976 abolishes only PKC␣ translocation without inhibiting Src translocation. This is consistent with our finding that the association between PKC␣ and Src depends upon PKC activity (Fig. 7A) and suggests that although the two kinases physically associate, neither this association nor the activation of PKC␣ is necessary for translocation of Src.
Functional Regulation of Src by PKC-By having shown the association of Src and PKC␣, it was important to determine whether PKC activity could regulate the activity of associating Src kinase. We chose to assay Src activity by the following two methods: analysis of autophosphorylation of Tyr-416, and in vitro phosphorylation of exogenous substrate peptide Raytide. Under basal conditions Src is not phosphorylated on Tyr-416 (Fig. 9, A and B), and upon activation this residue becomes phosphorylated. Pretreatment of platelets with PKC inhibitors, either Gö6976 or GF109203X, induced an increase in the stimulated phosphorylation of Tyr416, suggesting that PKC␣ may negatively regulate Src activity. Pretreatment of platelets with the Src inhibitor PP1 completely abolished activation-induced autophosphorylation of Src. When assayed in vitro by 32 P labeling of Raytide peptide by immunoprecipitated Src, inhibition of PKC isoforms by GF109203X also potentiated the activity of Src (Fig. 9C), confirming that PKCs negatively regulate the activity of Src in this system. Finally, we assessed the effect of the Syk inhibitor piceatannol on Src activity (Fig. 9D), and we show that the activity is not diminished in the presence of this inhibitor, demonstrating selectivity of piceatannol for Syk over Src in our assay. This is consistent with gene knock-out studies, where absence of Syk had no effect upon adhesionmediated activation of Src (42). There is a small but reproducible increase in Src activity in fact, which is consistent with the working model we present in Fig. 10, as Syk positively regulates PKC␣ activity which in turn negatively regulates Src activity. It would be predicted that inhibition of Syk would therefore lead to an increase in Src activity. DISCUSSION It is widely accepted that tyrosine phosphorylation is critical for early activation of platelets by the adhesion receptors GP VI and GP Ib-IX-V. However, in addition to tyrosine kinases, the PKC family of serine/threonine kinases has been shown to be essential in aggregation and secretion (13,14,(43)(44)(45)(46). This study has defined a critical role for PKC␣ in complex with Syk and Src tyrosine kinases in human platelets activated downstream of GPVI and GPIb-IX-V, and this study also demonstrates an essential role for this PKC isoform in regulating secretion of 5-HT and a major contributory role in regulating platelet aggregation (illustrated diagrammatically in Fig. 10).
In this study we have analyzed the association, phosphorylation, activation, and translocation of PKC␣, Syk, and Src in human platelets activated downstream of GP VI and GP Ib-IX-V. Here we have used the snake venom component alboaggregin A as an agonist to stimulate platelets through combined activation of GP Ib-V-IX and GP VI, receptors for von Willebrand factor and collagen, respectively (37,39,47). Co-activation of these receptors is likely to occur in vivo, and they may act synergistically with each other to induce substantial platelet activation (48). Murugappan et al. (7) have recently reported selective inhibition of platelet responses to collagen, rather than thrombin, by the classical PKC inhibitor Gö6976, suggesting a primary role for classical PKCs in signaling mediated by adhesion. Of the classical PKC family, platelets express PKC␣ and PKC␤ isoforms predominantly, with little or no expression of PKC␥ (6,7). In our study, co-activation of GP VI and GP Ib-IX-V leads to selective activation of PKC␣ rather than the PKC␤ isoforms, allowing us to infer that responses blocked by the classical isoform-selective inhibitor Gö6976 are likely to be primarily mediated by PKC␣. We therefore conclude from data presented in Fig. 1 that platelet 5-HT secretion induced by alboaggregin A is absolutely dependent upon PKC␣ activity. Platelet aggregation is also completely dependent upon PKC activity, being blocked by GF109203X, but although a major component is PKC␣-dependent, there is also a small component that is not blocked by Gö6976 and is therefore PKC␣-independent, and may therefore be downstream of a nonclassical isoform of PKC. We have recently shown that PKC is activated by alboaggregin A (6), and this may be one of the nonclassical isoforms that play a role in mediating platelet aggregation. Indeed, deficiency of PKC has been shown recently to lead to signaling defects in the activation of integrin ␣ IIb ␤ 3 in human platelets (49).
Syk is a major early tyrosine kinase component of the GP VI signaling pathway (36,50) and is also activated by GP Ib-IX-V (37,51). It has been shown recently to regulate classical PKC isoforms by direct tyrosine phosphorylation of PKCs (52,53), in particular on Tyr-658 of PKC␣, leading to recruitment of Grb-2/Sos and activation of the Ras/ERK pathway in mast cells. We were therefore interested to determine whether PKC␣ was activated in an Syk-dependent manner, whether it was associated with Syk, and whether it was phosphorylated by Syk in human platelets.
Our evidence, from Fig. 3, suggests that PKC␣ is activated in a manner absolutely dependent upon Syk activity, because inhibition of Syk using the inhibitor piceatannol abolishes PKC␣ activity. Although piceatannol has a narrow concentration range over which it is selective for Syk over members of the Src family (54,55), our control experiments shown in Fig. 3 demonstrate that under our conditions piceatannol is at least selective for Syk inhibition relative to inhibition of the two major upstream Src kinases Fyn and Lyn. Also, piceatannol does not inhibit PKC␣ directly, as addition of this Syk inhibitor to the in vitro kinase assay for PKC␣ does not affect its kinase activity. It is important to state here that genetic approaches to selective modulation of Syk activity would also be valuable to support our findings, although the usefulness of the mouse Syk gene knock-out would be limited, because analysis of interaction of PKC␣ with Syk would not be possible. Future studies may however involve generation of a point mutant kinase-dead Syk gene knock-in, where full analysis of the role of Syk kinase activity would be made possible. PKC␣ is known to be regulated by products of phospholipase C, diacylglycerol, and inositol 1,4,5-trisphosphate-mediated calcium rise. Syk is established as lying upstream of PLC␥ 2 in platelets, in both GP VI and GP Ib-IX-V-mediated pathways (36,37), and it was therefore possible that the Syk dependence of PKC␣ activity was solely mediated by its activation of PLC␥ 2 . However, this was not the case, because Fig. 3 also shows that a maximally effective concentration of the PLC inhibitor is not able to abolish PKC␣ activity completely, in contrast to piceatannol that does abolish PKC␣ activity. We therefore conclude that PKC␣ is activated in an Syk-dependent manner through PLC-dependent and PLC-independent pathways. The latter may involve direct tyrosine phosphorylation of PKC␣, and we show in Fig. 5 that although minor by comparison with tyrosine phosphorylation of Syk, there is a clearly discernible tyrosine phosphorylation of PKC␣ in platelets that occurs over a similar time course to that of PKC␣ activation. This phosphorylation is downstream of Syk kinase, because it is blocked by piceatannol, as shown in Fig. 5D. This would be consistent with other reports where tyrosine phosphorylation of PKC␣ has been shown in other cell types, for example in insulin-stimulated CHO cells expressing the insulin receptor (32) and in mast cells downstream of Fc⑀RI activation (52,53). Indeed, Konishi et al. (34) showed that in response to H 2 O 2 stimulation of COS-7 cells, PKC␣ becomes phosphorylated on tyrosine and is catalytically activated (34). Evidence is provided by these authors that phosphorylation of conserved tyrosine residues within the catalytic domain of PKC can activate the kinase in a manner unrelated to hydrolysis of inositol phospholipids and generation of diacylglycerol. In the present work, we show that PKC␣ is both tyrosine-phosphorylated and activated in an Syk-dependent manner. This is not a definitive proof that direct phosphorylation by Syk induces the activation of PKC␣, rather that PKC␣ activity lies downstream of Syk activity. Most importantly, however, there is not likely to be a reciprocation by PKC␣ on Syk activity, because there was no change in activity seen in the presence of Gö6976 (Fig. 6B). This was consistent with the lack of evidence for direct phosphorylation of Syk by PKC␣, as shown in Fig. 6A using the antibody directed against PKC substrates in the consensus phosphorylation sequence Platelets were then stimulated with alboaggregin A (Albo. A) (1 g/ml) for 1 min, or control platelets were treated with vehicle solution (water). A, B, and D, Src was immunoprecipitated (IP) from basal and stimulated platelets, and its phosphorylation at Tyr-416 was measured by SDS-PAGE and immunoblotting. WB, Western blot. Immunoblot shown in A is representative of three independent experiments. Three independent experiments were analyzed by densitometry for B, which shows means Ϯ S.E. from these experiments (* ϭ p Ͻ 0.01, compared with nonstimulated platelets; ** ϭ p Ͻ 0.01, compared with platelets stimulated with alboaggregin A). There was no significant difference between the vehicle alone condition and that stimulated in the presence of PP1 (p Ͼ 0.05). C, Src was immunoprecipitated from basal and stimulated platelets with 1 g/ml anti-Src antibody, and in vitro phosphorylation of Raytide peptide was measured as described under "Experimental Procedures." Incorporation of 32 P into Raytide was assayed by liquid scintillation counting, and data shown are means Ϯ S.E. from three independent experiments (* ϭ p Ͻ 0.01, compared with nonstimulated platelets; ** ϭ p Ͻ 0.01, compared with platelets stimulated with alboaggregin A). The immunoblot shown in D was reprobed with anti-Src antibody, and the data shown are representative of three independent experiments.
(K/R)XpS(K/R) (where is a hydrophobic residue, pS is phosphoserine, and X is any amino acid).
It is becoming increasingly clear that interaction of PKC isoforms with signaling partners, including tyrosine kinases, is a major mechanism for the selective regulation of the activity of PKC isoforms and for directing PKCs to their substrates and functional effectors (reviewed in Refs. 56 and 57). We had shown previously that individual novel isoforms of PKC associate selectively with different nonreceptor tyrosine kinases in platelets upon activation (6,31), and we have shown recently that PKC␣ itself is constitutively associated with the tyrosine phosphatase SHP-1 through the SH2 domains of SHP-1 (58). Evidence has been provided previously that PKC␣ may associate with Src kinase in other cell types (59), and in this study we show that PKC␣ associates with both Syk and Src upon platelet activation. Most interestingly, the dependence of these interactions upon the activity of PKC␣ differs between the two tyrosine kinases. Although Src-PKC␣ association depends upon PKC␣ activity, the association of Syk with PKC␣ is independent of PKC␣ activity. Most intriguingly, the interaction between Syk and PKC␣ is dependent upon Syk activity (Fig. 5,  A and B), although the amount of Syk that associates with PKC␣ remains maximal by 180 s despite a fall off in Syk activity, as assessed by the tyrosine phosphorylation status of Syk (Fig. 5C). We are not sure of the underlying reason for this, but it is possible that Syk activity is merely required for initiating the interaction between the two proteins, which is then sustained by another Syk kinase-independent mechanism. The molecular basis for the physical associations and the difference between the two tyrosine kinases and their interactions with PKC␣ are not clear at present, and indeed it is possible that the interactions we see here are mediated by third party proteins and are not direct. This is an important area for future study.
Given a physical association between PKC␣ and Src, we were interested to investigate whether PKC␣ would phosphorylate Src and whether the activity of Src could be regulated by PKC␣. Liebenhoff et al. (60) had shown previously that in thrombinstimulated platelets, Src becomes phosphorylated by PKC at Ser-12. We were not able to detect phosphorylation of Src however in alboaggregin A-stimulated platelets, using the antiphospho-Thr and Ser (PKC substrate) antibodies detailed above (data not shown). However, we cannot rule out the possibility that Src is phosphorylated by PKC␣, but on a site that would not be detected by these antibodies. Despite the lack of evidence for phosphorylation of Src by PKC, however, Src activity is markedly elevated by inhibition of PKC␣ during platelet stimulation (Fig. 9). This is consistent with previous literature in other cell types showing negative regulation of Src by PKC isoforms. Song et al. (61) showed that overexpression of PKC␦ resulted in a dramatic inhibition of Src activity, indicating that PKC isoforms may have a negative effect on Src. In addition, fibroblast growth factor treatment of porcine aortic endothelial cells induces a decrease in Src activity, and this inhibition of Src is mediated by PKC (59).
Finally, we examined the subcellular localization of Src, Syk, and PKC␣ by confocal immunofluorescence microscopy. All three kinases undergo a stimulation-dependent translocation from cytosol to plasma membrane, or sub-plasma membrane localization. The translocation from the cytoplasm to the cell membrane upon activation is an accepted pattern for PKC function in various cell types (62,63) and platelets (31,64,65). Syk and Src have also been shown to translocate to the cell membrane upon platelet stimulation (66,67), and the two have been shown recently to associate in living cells by an elegant BRET approach (68). Because of the early role played by Src family members in the signaling pathways activated by alboaggregin A, it was not surprising that inhibition of Src by PP1 led to inhibition of translocation of both Src and PKC␣. Similarly, because we had shown that PKC␣ is activated in an Syk-dependent manner in our system, its translocation was also dependent upon Syk activity. Most interestingly, however, PKC␣ regulated Syk membrane translocation, which was accordingly blocked by Gö6976. The reason for the PKC dependence of translocation of Syk is not clear. It is, however, possible to hypothesize that the translocation of the heteromeric complex containing PKC␣ and Syk is driven by PKC␣ translocation. As described previously, the C1 and C2 domains appear to promote the membrane translocation of PKC by providing high affinity membrane-binding sites that interact with membranebound DAG and phosphatidylserine (69). Most interestingly, we have described previously the regulation by PKC␦ of the activation-induced membrane translocation of Fyn in human platelets (31). Finally, our data with Src are in contrast with Liebenhoff et al. (60), who showed that in addition to phosphorylation of Src by PKC in thrombin-stimulated platelets, inhibition of PKC prevented thrombin-induced Src translocation to the cytoskeleton. This discrepancy between our findings and those of Liebenhoff et al. (60) may be as a result of the different agonists used in each of the studies.
In conclusion, we have shown that PKC␣ plays a critical role in regulating both 5-HT secretion and aggregation, as shown in Fig. 10. Early activation of Syk by co-stimulation of the adhesion receptors GP VI and GP Ib-IX-V regulates PKC␣ by two pathways, PLC-dependent and PLC-independent. The PLCindependent pathway may be mediated by direct phosphorylation of PKC␣ by Syk. In turn, PKC␣ negatively regulates Src kinase, although this may not be by direct phosphorylation of FIG. 10. Model depicting the functional interaction of Syk and Src with PKC␣. The model shown depicts the activation pathway of key signaling molecules investigated in the present study. An early activation of Syk by co-stimulation of adhesion receptors GP VI and GP Ib-IX-V leads to activation of PKC␣ by two pathways, PLC-dependent and PLC-independent. The PLC-independent pathway may be mediated by direct phosphorylation of PKC␣ by Syk. PKC␣ does not feedback to phosphorylate or regulate the activity of Syk, although PKC␣ does negatively regulate Src kinase. However, this regulation of Src may not be by direct phosphorylation of Src. Functionally, PKC␣ plays a critical role in regulating both 5-HT secretion and aggregation, although there are additional PKC␣-independent pathways to regulation of aggregation. In addition to control of activities, PKC␣ physically associates with both Syk and Src in a stimulation-dependent manner, and the complexes translocate to a sub-plasma membrane location upon activation of the cell. The association of Syk-PKC␣ depends upon activation of Syk, but not PKC␣, whereas the translocation of each kinase is dependent upon both their activities. The association of Src-PKC␣ depends upon PKC␣ activity, in contrast to Syk-PKC␣ association, and although the translocation of PKC␣ depends upon PKC␣ activity, translocation of Src is independent of PKC␣.
Src. In addition to control of activities, PKC␣ physically associates with both Syk and Src in a stimulation-dependent manner. The complex interplay between these signaling partners now requires further detailed analysis to elucidate how the associations may take place, in order to develop tools to specifically disrupt interactions.