Cytosolic phospholipase A2 is phosphorylated in collagen- and thrombin-stimulated human platelets independent of protein kinase C and mitogen-activated protein kinase.

Human platelets pretreated with indomethacin release arachidonic acid predominantly through the activity of cytosolic phospholipase A2 (cPLA2), an 85-kDa protein. This enzyme is regulated by an increase in intracellular Ca2+, a necessary condition for arachidonic acid liberation, and by phosphorylation. Phosphorylation of cPLA2 enhanced the Ca2+-induced arachidonic acid release in platelets stimulated by the ionophore A23187 and phorbol ester (phorbol 12,13-dibutyrate (PDBu)). In thrombin-stimulated platelets, however, phosphorylation appeared not to be necessary for arachidonic acid release since the latter response was not impaired in the presence of staurosporine, which inhibited phosphorylation. Collagen, thrombin, and PDBu induced phosphorylation of platelet cPLA2 as well as activation of mitogen-activated protein kinase (MAPK; p42mapk and p44mapk). cPLA2 activation was not dependent on protein kinase C (PKC) in thrombin- and collagen-stimulated platelets, as preincubation with the PKC inhibitor Ro 31-8220 neither interfered with cPLA2 phosphorylation nor reduced arachidonic acid release. However, collagen- and thrombin-induced activation of MAPK was inhibited by Ro 31-8220, indicating that PKC is necessary for MAPK stimulation in platelets. Although MAPK may underlie phosphorylation of cPLA2 in PDBu-activated human platelets, our results provide evidence for PKC- and MAPK-independent phosphorylation of cPLA2 in platelets stimulated by the physiological activators collagen and thrombin.

Upon stimulation, human platelets release eicosanoids as mediators in blood clotting events. Arachidonic acid is cleaved from the sn-2 position of phospholipids through the activity of phospholipase A 2 (PLA 2 ) 1 and metabolized by cyclooxygenase and lipoxygenase enzymes. Two different forms of PLA 2 exist in platelets: a 14-kDa PLA 2 , which is secreted into the plasma and depends on millimolar Ca 2ϩ for activation (1), and the recently discovered cytosolic PLA 2 (cPLA 2 ), which is an 85-kDa enzyme requiring submicromolar Ca 2ϩ for activation (2). cPLA 2 has been purified (3,4) and cloned (5,6). The enzyme is regulated by intracellular Ca 2ϩ , which induces translocation to membranes (5) through a Ca 2ϩ -dependent lipid-binding motif in its N terminus (7). Phosphorylation on serine residues leads to an increase in enzymatic activity (2,8,9). A third structural feature of cPLA 2 , serine 228, is involved in the catalytic mechanism (10).
The signaling events leading to phosphorylation and activation of cPLA 2 are not established. Phorbol 12-myristate 13acetate potentiates release of arachidonic acid by the Ca 2ϩ ionophore A23187 in platelets suggesting a role for protein kinase C (PKC; Ref. 11). Nemenoff et al. (12) reported that purified cPLA 2 can be phosphorylated by either purified PKC or mitogen-activated protein kinase (MAPK). The site of phosphorylation of cPLA 2 by MAPK has been identified as serine 505 in transfected Chinese hamster ovary cells and lies in a consensus sequence for this kinase (9). MAPK activity is regulated through MAPK kinase (13)(14)(15), which is itself activated downstream of PKC-dependent and -independent pathways (16,17).
In human platelets, cPLA 2 is phosphorylated after thrombin stimulation, and this is associated with an increase in its specific activity (2). Both p42 mapk and p44 mapk are present in platelets (18 -20), but only p42 mapk has been reported to undergo activation in thrombin-stimulated human platelets (19).
In this study, we have investigated the regulation of cPLA 2 in human platelets stimulated by collagen and thrombin. Collagen binds to glycoprotein receptors on the platelet surface, which results in an increase in phospholipase C (PLC) activity through phosphorylation of PLC␥-2 on tyrosine residues (21,22). In contrast, thrombin activates PLC␤ isoforms via a G protein-dependent pathway (23). To investigate the importance of PKC and MAPK in the regulation of cPLA 2 , we used kinase inhibitors that block the intracellular signaling pathways elicited by collagen and thrombin. Staurosporine is a strong inhibitor of tyrosine and serine/threonine kinases, including PKC, but its lack of specificity limits its use for studying the role of PKC (24,25). We therefore included the staurosporine analogue Ro 31-8220 in our study, which is a potent and more selective inhibitor of PKC (26,27). The role of Ca 2ϩ has also been investigated using the divalent cation ionophore A23187 and the intracellular Ca 2ϩ chelator BAPTA-AM. The results demonstrate that cPLA 2 undergoes phosphorylation in collagen-and thrombin-stimulated platelets independent of PKC and MAPK and that Ca 2ϩ has a major role in its regulation independent of phosphorylation.
Preparation of Platelets-Blood was drawn on the day of the experiment from healthy volunteers by venepuncture using acidic citrate dextrose (ACD: 120 mM sodium citrate, 110 mM glucose, 80 mM citric acid) as anticoagulant. All solutions were prewarmed to 30°C. The blood was centrifuged at 200 ϫ g for 20 min to obtain platelet-rich plasma. Platelets were collected by centrifugation at 1000 ϫ g for 10 min in the presence of 0.1 g/ml prostacyclin. The pellet was gently resuspended in 1 ml of Tyrode's buffer (20 mM HEPES, 135 mM NaCl, 3 mM KCl, 0.35 mM Na 2 HPO 4 , 12 mM NaHCO 3 , and 1 mM MgCl 2 , pH 7.3) and 150 l of ACD. For the measurement of [ 3 H]arachidonic acid release, this suspension was incubated with 5 Ci of [ 3 H]arachidonic acid for 2 h. Platelets were washed with a mixture of 25 ml of Tyrode's and 3 ml of ACD and centrifuged at 1000 ϫ g for 10 min in the presence of prostacyclin. The pellet was resuspended in 1 ml of Tyrode's buffer. For cPLA 2 and pleckstrin phosphorylation, platelets were labeled with 0.5 mCi/ml 32 P i for 2 h. The volume of the platelet suspension was adjusted to give 4 ϫ 10 8 cells/ml for the [ 3 H]arachidonic acid release studies and 2 ϫ 10 9 cells/ml for immunoprecipitation. Indomethacin (10 M) was added to the platelet suspension in all experiments in order to block cyclooxygenase. For the study of [ 3 H]arachidonic acid release, BW4AC (3 M) was added to block lipoxygenase, and for immunoprecipitation procedures, EGTA (1 mM) was present to prevent platelet aggregation. The platelets were rested for 30 min at 30°C prior to experimentation. Platelets were preincubated with BAPTA-AM for 15 min or with other substances for 5 min at 37°C. The incubation mixture was gently shaken for 2 min during stimulation with thrombin, A23187, and PDBu or stirred in an aggregometer at 1200 rpm for 5 min during stimulation with collagen.

Measurement of [ 3 H]Arachidonic
Acid Release-Stimulation of platelets (200 l) was stopped with an equal volume of 6% glutaraldehyde. Samples were centrifuged at 13,000 ϫ g for 15 min at 4°C. The radioactivity of an aliquot of the supernatant (300 l) was determined by liquid scintillation spectrometry in a Beckman scintillation counter to a 5% level of significance.
Measurement of Intracellular Ca 2ϩ -Platelet-rich plasma was incubated with fura2-AM (3 M) at 30°C for 45 min, and platelets were prepared as descibed above. The volume of the platelet suspension was adjusted to give 3 ϫ 10 8 cells/ml, and indomethacin (10 M) was added. Aliquots (450 l) were incubated with either Me 2 SO (1%), Ro 31-8220 (10 M), or staurosporine (10 M) for 5 min or with BAPTA-AM (10 M) for 15 min. Platelets were stimulated with thrombin (1 unit/ml) or collagen (100 g/ml) at room temperature in a stirred solution. Fluorescence was measured at excitation wavelengths 340 and 380 nm with emission at 510 nm, using a Perkin-Elmer LS50B spectrofluorimeter. Data are presented as the excitation fluorescence ratio (340:380 nm).
Measurement of cPLA 2 Phosphorylation-The stimulation of 32 P i radiolabeled platelets (500 l incubation volume) was stopped with an equal volume of lysis buffer (final concentrations: 1% Triton X-100, 0.5% SDS, 0.75% deoxycholate, 10 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 100 g/ml aprotinin, 10 M pepstatin A, 10 M staurosporine, 10 mM Na 4 P 2 O 7 , 50 mM NaF, 200 M Na 3 VO 4 ) and placed on ice for 30 min. The cell lysates were separated from insoluble material by centrifugation at 13,000 ϫ g for 15 min at 4°C in a microcentrifuge, precleared with 40 l of protein A-Sepharose CL-4B (1:2 in H 2 O (v/v)) and incubated with 2 l of polyclonal rabbit cPLA 2 antiserum at 4°C for 2 h and with 40 l of protein A-Sepharose CL-4B for another 30 min. The immunoprecipitates were collected and washed 3 times with 1 ml of wash buffer (0.5% Triton X-100, 150 mM NaCl, 10 mM Tris-HCl, pH 7.4) followed by two washes with the wash buffer containing 750 mM NaCl and finally 2 times with the initial wash buffer. The immunoprecipitates were dried with strips of filter paper, extracted with Laemmli sample buffer, heated for 15 min at 60°C, and resolved on 10% SDS-PAGE. Gels were transferred to polyvinylidene difluoride membranes, which were blocked with 10% BSA. Membranes were exposed to Amersham Hyperfilm at Ϫ70°C using intensifying screens for 1-7 days. Phosphorylation was quantified by densitometry.
Electrophoretic mobility analyses to distinguish phosphorylated from nonphosphorylated cPLA 2 was performed on 10% SDS-PAGE as described previously (2).
Phosphorylation of Pleckstrin (p47)-A small sample of 32 P i -labeled platelet lysate was removed prior to cPLA 2 immunoprecipitation, placed in Laemmli sample buffer, boiled for 5 min, and analyzed on 10% SDS-PAGE. Gels were stained with Coomassie Blue, dried, and autoradiographed. Regions of the gels corresponding to pleckstrin were cut and scintillation counted for radioactivity.
Measurement of MAPK Activation-The stimulation of platelets (500-l incubation volume) was stopped with an equal volume of denaturing lysis buffer (final concentrations: 2% SDS, 5 mM EDTA, 10 mM Tris, 0.5 mM phenylmethylsulfonyl fluoride, pH 7.3). Samples were boiled for 10 min and diluted 40-fold with Tris-buffered saline (TBS-T 20 mM Tris, 137 mM NaCl, 0.1% Tween 20, pH 7.6) containing 2 mg/ml BSA, 1 mM phenylmethylsulfonyl fluoride, and 1 mM EDTA. After preclearing the platelet extracts for 1 h, they were incubated overnight with either 1 g of polyclonal anti-p42/p44 mapk antibody or with 1 g of monoclonal anti-p42 mapk antibody and 0.76 g of rabbit anti-mouse IgG and with 40 l of protein A-Sepharose CL-4B (1:2 suspension) followed by centrifugation at 3000 ϫ g for 10 min. Immunoprecipitates were washed 3 times with TBS-T containing 1 mM EDTA, dried, and boiled for 10 min in 20 l of Laemmli sample buffer (during the last wash, one-tenth of the immunoprecipitate was removed for Western blotting controls). The immunoprecipitates were resolved on 10% SDS-PAGE, with MBP (0.5 mg/ml) copolymerized in the polyacrylamide gel. The in-gel renaturation was carried out as described by Papkoff et al. (19). Briefly, gels were washed twice in a mixture of 50 mM Tris, pH 8.0, and 20% isopropanol and then twice in 50 mM Tris, pH 8.0, containing 5 mM ␤-mercaptoethanol. Proteins were denatured with 6 M guanidine HCl and then renatured for 16 h at 4°C with several changes of 50 mM Tris, pH 8.0, containing 5 mM ␤-mercaptoethanol and 0.04% Tween 20. For the kinase assay, gels were incubated at 37°C for 1 h with 10 ml of kinase assay buffer (50 mM Tris, pH 8.0, 5 mM MgCl 2 , 1 mM EGTA, 5 mM dithiothreitol, 50 M ATP, and 200 Ci of [␥-32 P]ATP) and then extensively washed in a mixture of 5% trichloroacetic acid and 1% Na 4 P 2 O 7 . Gels were dried, and autoradiographs were exposed for up to 2 days at Ϫ70°C with intensifying screens. The region of MAPK was cut from the gel and scintillation counted for radioactivity.
Analysis of Results-For [ 3 H]arachidonic acid release, experiments were performed in quadruplicate. Results are expressed as mean Ϯ S.E. Statistical significance was assessed using Student's t test with p Ͻ 0.05.

Release of [ 3 H]Arachidonic Acid from Platelets Is Dependent
on Ca 2ϩ -The release of [ 3 H]arachidonic acid was measured as an index of cPLA 2 activity in human platelets. We compared a procedure adapted from a 5-hydroxytryptamine release assay (24) with the separation of a lipophilic platelet extract on thinlayer chromotography prior to scintillation counting (11) and found that both methods gave similar results. The cyclooxygenase inhibitor indomethacin (10 M) and the lipoxygenase inhibitor BW4AC (3 M) were included to prevent arachidonic acid metabolism (28) and secondary stimulation by thromboxane A 2 (29). During a 2-min stimulation, the Ca 2ϩ ionophore A23187 released 9.8 Ϯ 0.74% (n ϭ 14) of total tissue levels of [ 3 H]arachidonic acid, and thrombin released 4.4 Ϯ 0.49% (n ϭ 14; Table I). Collagen stimulated the release of 3.9 Ϯ 0.54% (n ϭ 5) of total tissue levels during 5 min.
The Ca 2ϩ dependence of arachidonic acid formation in collagen-stimulated platelets has previously been demonstrated by Smith et al. (28) using the Ca 2ϩ chelator BAPTA-AM (100 M), which permeates the plasma membrane and becomes trapped as a tetraanion inside the cell after cleavage by nonspecific cytosolic esterases. Since BAPTA-AM markedly suppressed PLC activation at concentrations higher than 15 M (30), we used BAPTA-AM at a concentration of 10 M. This blocked the thrombin-and collagen-stimulated Ca 2ϩ rise in platelets ( Fig.  1) and inhibited over 90% of the [ 3 H]arachidonic acid release after stimulation by collagen, thrombin, or A23187 (Table I). To check that BAPTA-AM did not inhibit PLC/PKC activation, we monitored phosphorylation of pleckstrin, the major PKC substrate in platelets (31,32). BAPTA-AM had no effect on [ 32 P]phosphorylation of pleckstrin induced by collagen (106.8 Ϯ 5.1%, n ϭ 3, relative to collagen, 100%) or by thrombin (108.7 Ϯ 6.4%, n ϭ 4, relative to thrombin, 100%), indicating that the observed inhibition of [ 3 H]arachidonic acid release is due to the Ca 2ϩ -chelating properties of BAPTA-AM. A rise in intracellular Ca 2ϩ is therefore a necessary condition for the formation of arachidonic acid in platelets.

Release of [ 3 H]Arachidonic Acid from Thrombin-and Collagen-stimulated Platelets Is Not Dependent on PKC Activity-
Halenda et al. (11) reported that phorbol 12-myristate 13acetate enhanced the formation of arachidonic acid in ionophore-stimulated platelets without having any effect on its own. This synergy was reversed by the kinase inhibitor staurosporine, although the response to the ionophore in the absence of phorbol 12-myristate 13-acetate was not altered by staurosporine (33). In the present study, we used a different phorbol ester, PDBu, and observed potentiation of [ 3 H]arachidonic acid release over the length of the concentration-response curve to A23187 (Fig. 2). Preincubation with the specific PKC inhibitor Ro 31-8220 (26,34) reversed the effect of PDBu but had no significant effect on A23187-stimulated [ 3 H]arachidonic acid release (Fig. 2). Staurosporine did not significantly alter the [ 3 H]arachidonic acid release from A23187-stimulated platelets (not shown). These data demonstrate that PKC activity does not contribute to Ca 2ϩ ionophore-induced release of [ 3 H]arachidonic acid in platelets but that additional stimulation of PKC can enhance this response.
Further experiments were therefore designed to investigate whether PKC is involved in the formation of [ 3 H]arachidonic acid after activation by physiological stimuli, such as collagen and thrombin. Both Ro 31-8220 and staurosporine inhibited phosphorylation of the PKC substrate pleckstrin in collagenand thrombin-treated platelets (Fig. 3). Ro 31-8220 did not alter the collagen-induced [ 3 H]arachidonic acid release significantly, whereas staurosporine blocked it (Fig. 4A). This is in agreement with results reported by Daniel et al. (21) and parallels the effect of each inhibitor on PLC activation (35). The inhibitory effect of staurosporine is consistent with a role for tyrosine phosphorylation in early signaling events in collagenstimulated platelets (21,22,36). Since Ro 31-8220 did not decrease the formation of [ 3 H]arachidonic acid after collagen stimulation, PKC activity does not appear to be necessary for arachidonic acid liberation.
[ 3 H]Arachidonic acid was generated in a concentration-dependent manner following stimulation of platelets with thrombin (Fig. 4B). At submaximal concentrations of thrombin (0.2-1 units/ml), both staurosporine and Ro 31-8220 enhanced [ 3 H]arachidonic acid release without having any significant effect on their own (for example, compare release of [ 3 H]arachidonic acid at 0.01 units/ml thrombin). Activation of PLC by thrombin is not altered in the presence of staurosporine but is potentiated significantly by Ro 31-8220 (24,37); the effect of Ro 31-8220 can be explained by inhibition of the negative feedback action of PKC on PLC (38). This is in good agreement with the effect of Ro 31-8220 on [ 3 H]arachidonic acid release, H]Arachidonic acid-labeled platelets (4 ϫ 10 8 /ml) were incubated with Me 2 SO (1%) or BAPTA-AM (10 M) for 15 min and then stimulated with A23187 (2 M) or thrombin (1 unit/ml) for 2 min or with collagen (100 g/ml, stirred at 1200 rpm) for 5 min at 37°C. The data are presented as the percentage of the total 3 H radioactivity incorporated into the platelet after subtraction of the radioactivity released from unstimulated platelets (basal). Typical values for basal release of 3 H radioactivity and stimulation with A23187 were 2200 dpm and 8000 dpm, respectively. Experiments were carried out in quadruplicate; results are given as mean Ϯ S.E. (n ϭ number of independent experiments). suggesting that thrombin-induced formation of [ 3 H]arachidonic acid is an event downstream of PLC and Ca 2ϩ .
To further investigate this hypothesis, intracellular Ca 2ϩ elevation was measured in the presence of Ro 31-8220 and staurosporine. As shown in Fig. 4C, Ro 31-8220 had no effect on the intracellular Ca 2ϩ rise in collagen-stimulated platelets, whereas staurosporine blocked it. Both inhibitors increased and prolonged the intracellular Ca 2ϩ elevation after thrombin stimulation. These data correlate with the release of [ 3 H]arachidonic acid (Fig. 4, A and B) and provide further evidence that arachidonic acid is generated downstream of Ca 2ϩ in human platelets.
Phosphorylation of cPLA 2 in Platelets-Kramer et al. (2) have recently shown that thrombin stimulates phosphorylation of cPLA 2 in human platelets. Collagen also induced [ 32 P]phosphorylation of cPLA 2 , immunoprecipitated from 32 P i -labeled platelets (Fig. 5A, upper panels). In both collagen-and thrombin-stimulated platelets, phosphorylation of cPLA 2 was not altered in the presence of Ro 31-8220, whereas staurosporine inhibited it completely as illustrated in Fig. 5A (for densitometry analysis see the legend). Western blotting for cPLA 2 demonstrated that a comparable amount of protein was immunoprecipitated in all samples (Fig. 5A, lower panels). To investigate the time-course of cPLA 2 phosphorylation and to confirm the results obtained with the inhibitors, we carried out Western blots for cPLA 2 on total tissue lysates and monitored the electrophoretic mobility shift of phosphorylated cPLA 2 (2,8,9). Collagen induced detectable cPLA 2 phosphorylation only after 5 min of stimulation with less than half of protein in the retarded, phosphorylated form; thrombin stimulated cPLA 2 phosphorylation within 2 min (Fig. 5B). In confirmation of the results described above, the phosphorylated form of cPLA 2 was observed in platelets stimulated in the presence of Ro 31-8220, whereas staurosporine blocked cPLA 2 phosphorylation (no mobility shift; Fig. 5B). Our results demonstrate that phosphorylation of cPLA 2 is independent of PKC but is mediated downstream of kinases sensitive to staurosporine.
Activation of MAPK in Collagen-and Thrombin-stimulated Platelets-Lin et al. (9) and Nemenoff et al. (12) have recently proposed a role for MAPK in the phosphorylation of cPLA 2 . Platelets contain both p42 mapk and p44 mapk (18 -20), but little is known about the regulation of platelet MAPK. We therefore investigated the activation of MAPK and its dependence on PKC in platelets stimulated by collagen and thrombin. Kinase activity of immunoprecipitated p42/p44 mapk was determined using an in-gel renaturation kinase assay. The substrate MBP was copolymerized with polyacrylamide in the gels.
Both isoforms of MAPK were activated in collagen-and thrombin-stimulated platelets, whereas there was little p42 mapk activity under basal conditions (Fig. 6A). p42 mapk was the predominantly active form, but in contrast to results obtained by Papkoff et al. (19), p44 mapk was also activated. The collagen-induced MAPK activation peaked after 2 min of stimulation (Fig. 6A) and declined slightly by 5 min to 13.4 Ϯ 3.0-fold (n ϭ 5) over basal levels. Thrombin stimulated maximal MAPK activity between 1 and 2 min (at 2 min, 94.2 Ϯ 31.9-fold, n ϭ 5), which declined to half-maximal levels after 5 min. Preincubation of platelets with Ro 31-8220 or staurospo- FIG. 3. Phosphorylation of pleckstrin (p47). Platelets were labeled with 32 P i , pretreated with indomethacin and were incubated with Me 2 SO (D, 1%), Ro 31-8220 (Ro, 10 M), or staurosporine (st, 10 M) for 5 min at 37°C. They were stimulated with collagen (100 g/ml) for 5 min in a stirred solution (1200 rpm) or with thrombin (1 unit/ml) for 2 min. Total tissue samples were resolved on 10% SDS-PAGE, and the dried gel was autoradiographed. rine blocked MAPK activation by either stimulus at all time points measured (Fig. 6A). All samples contained similar levels of p42/p44 mapk as analyzed by Western blotting of a small portion of each immunoprecipitate (Fig. 6B). We have confirmed these results by using a monoclonal antibody to immunoprecipitate p42 mapk prior to the in-gel renaturation kinase assay (not shown), and by Western blotting MAPK immunoprecipitates for phosphotyrosine residues. Thrombin-stimulated p42 mapk tyrosine phosphorylation peaked at 2 min (Fig.  6C); faint p44 mapk tyrosine phosphorylation was detectable at longer exposure times (not shown). Tyrosine phosphorylation was blocked in the presence of Ro 31-8220 or staurosporine (Fig. 6C). Collagen-induced p42 mapk tyrosine phosphorylation was much weaker but could be detected after 2 and 5 min of stimulation (not shown); it was also inhibited by Ro 31-8220 and staurosporine. These results suggest that p42/p44 mapk is regulated downstream of PKC in platelets stimulated by collagen or thrombin.
FIG. 6. Activation of p42 mapk and p44 mapk is inhibited by staurosporine and Ro 31-8220. Indomethacin-treated platelets (1 ϫ 10 9 / sample) were incubated with Me 2 SO (1%), Ro 31-8220 (10 M), or staurosporine (10 M) for 5 min at 37°C. A, platelets were stimulated with collagen (100 g/ml) for 5 min in a stirred solution (1200 rpm) or with thrombin (1 unit/ml) for 2 min. p42 mapk and p44 mapk were immunoprecipitated under denaturing conditions using a polyclonal antibody recognizing both isoforms and were resolved on SDS-PAGE (10%, copolymerized with 0.5 mg/ml MBP). Gels were renatured as described under "Materials and Methods" and were incubated with 50 M ATP and 20 Ci/ml [␥-32 P]ATP in kinase buffer. Autoradiographs were taken from dried gels. Under the conditions used during the kinase assay, no autophosphorylation was detectable (not shown). B, small portions of the immunoprecipitates were resolved on 10% SDS-PAGE, transferred to polyvinylidene difluoride and Western blotted for MAPK using a polyclonal anti-p42/p44 mapk antibody. C, membranes obtained as in B were Western blotted for phosphotyrosine residues using the monoclonal anti-phosphotyrosine antibody 4G10. Bands around 50 kDa (B and C) represent IgG heavy chains and were detected by secondary antibody alone. The results are representative of three similar experiments.
Phosphorylation in Platelets-Since an increase in PKC activity could enhance ionophore-stimulated [ 3 H]arachidonic acid release, we investigated the effect of PDBu and A23187 with regard to activation of MAPK and phosphorylation of cPLA 2 . A23187 stimulated weak phosphorylation of pleckstrin but did not alter the marked phosphorylation induced by PDBu (Fig.  7A). A23187 stimulated neither p42 mapk (investigated using a monoclonal anti-p42 mapk antibody, Fig. 7B) nor p44 mapk (not shown), whereas PDBu activated both p42 mapk (Fig. 7B) and p44 mapk (not shown). PDBu also caused phosphorylation of cPLA 2 (Fig. 7C and cPLA 2 mobility shift experiments, not shown). cPLA 2 phosphorylation stimulated by A23187 was weaker (Fig. 7C), as previously observed by Lin et al. in Chinese hamster ovary cells overexpressing cPLA 2 (8), and the cPLA 2 mobility shift could only be detected after 5 min of stimulation with A23187 (not shown). Both stimuli together had an additive effect (Fig. 7C, for densitometry analysis see the legend). PDBu-induced pleckstrin phosphorylation, MAPK activation and cPLA 2 phosphorylation were inhibited by Ro 31-8220 (not shown). The synergy between PKC and Ca 2ϩ on [ 3 H]arachidonic acid release (Fig. 2) may be explained by PKCstimulated cPLA 2 phosphorylation, possibly through the activity of MAPK, leading to an increase in the specific activity of cPLA 2 (2,8). DISCUSSION The abundance of cPLA 2 in human platelets over secretory PLA 2 (2) and the millimolar Ca 2ϩ requirement of the latter (1) suggest that the bulk of arachidonic acid is generated through the activity of cPLA 2 . Consistent with this, inhibitors of secretory PLA 2 do not reduce the release of arachidonic acid from human platelets (39), whereas arachidonyl trifluoromethyl ketone, an inhibitor of cPLA 2 (40) and Ca 2ϩ -independent PLA 2 (41), induces substantial inhibition after stimulation with thrombin (39) and A23187 (42).
Evidence against a major role for diacylglycerol lipase in the liberation of arachidonic acid is provided by Halenda et al. (11,33), demonstrating that Ca 2ϩ ionophores induce substantial arachidonic acid release in the absence of PLC activation, and by the present work with BAPTA-AM. BAPTA-AM markedly inhibited thrombin-and collagen-stimulated [ 3 H]arachidonic acid release (Table I) but had little effect on pleckstrin phosphorylation, an indicator of PKC activation in human platelets (32,43,44). This shows indirectly that PLC activation (and therefore formation of diacylglycerol, the substrate for diacylglycerol lipase) was not impaired (see also Ref. 30). Our results support the model of cPLA 2 activation whereby a Ca 2ϩ -dependent lipid-binding domain is responsible for the translocation of cPLA 2 to phospholipid membranes (5,7).
To address the question as to whether PKC and kinases regulated downstream of PKC are involved in the activation of cPLA 2 , experiments were carried out using agents that either increase (PDBu) or inhibit (Ro 31-8220, staurosporine) PKC activity. Human platelets contain the PKC isoenzymes ␣, ␤, ␦, and of which PKC-␣, -␤ and -translocate to the membrane fraction after stimulation by thrombin, which correlated with the first phase of diacylglycerol formation and pleckstrin phosphorylation (45). Ro 31-8220 inhibits purified PKC-␣, -␤, -␥, and -⑀ isoforms as well as the PKC activity of a partially purified rat brain PKC preparation (which presumably also contains PKC-␦ and -isoforms) with an IC 50 of less than 30 nM (27). We would therefore expect Ro 31-8220 (10 M) to completely inhibit all PKC activity in platelets, and consistent with this, Ro 31-8220 inhibited pleckstrin phosphorylation (Fig. 3). Ro 31-8220 also blocks PKC-dependent processes such as secretion of 5-hydroxytryptamine and slows down the rate of aggregation, whereas it does not affect Ca 2ϩ -dependent shape change in platelets, which is mediated via myosin light chain kinase (34), or tyrosine phosphorylation (35).
When applied on their own, phorbol esters do not stimulate arachidonic acid release from platelets or elevate intracellular Ca 2ϩ (46). However, the synergy of phorbol esters and A23187 on arachidonic acid formation, and the inhibition of this by staurosporine and Ro 31-8220, provides evidence for a role of PKC in the regulation of PLA 2 activity (present results and Ref. 33). The absence of any effect of Ro 31-8220 on A23187-induced release of [ 3 H]arachidonic acid (Fig. 2) is consistent with the observation that A23187 does not stimulate significant PKC activity (Fig. 7A). Similar results have been described by Qiu and Leslie (47) in macrophages where inhibition of PKC either by GF109203X (25) or through down-regulation did not interfere with A23187-stimulated release of [ 3 H]arachidonic acid. The molecular basis of the synergy between phorbol esters and Ca 2ϩ -ionophore is probably mediated through PKC/MAPK dependent phosphorylation of cPLA 2 ( Fig. 7B and C) which, despite increasing the specific activity of cPLA 2 , does not induce activation of cPLA 2 in vivo in the absence of Ca 2ϩ -depend- FIG. 7. Effect of PDBu and A23187 on PKC and p42 mapk activities and on cPLA 2 phosphorylation. Platelets were treated with indomethacin and stimulated with Me 2 SO (1%, basal), PDBu (1 M), A23187 (2 M), or both drugs together at 37°C for 2 min. A, phosphorylation of pleckstrin as indicator of PKC activity. Total tissue samples from 32 P i -labeled platelets were resolved on 10% SDS-PAGE, and the dried gel was autoradiographed. The position of the PKC substrate pleckstrin is indicated. B, activity of p42 mapk (in-gel renaturation kinase assay). p42 mapk was immunoprecipitated using a monoclonal anti-p42 mapk antibody and renatured in MBP (0.5 mg/ml) containing 10% SDS-PAGE gels. The autoradiograph shows bands of renatured p42 mapk activity. C, autoradiograph of immunoprecipitated cPLA 2 from 32 P ilabeled platelets. After stimulation, platelets were lysed, and cPLA 2 was immunoprecipitated as described under "Materials and Methods." As analyzed by densitometry, cPLA 2 phosphorylation induced by A23187 corresponded to 40% of the response to PDBu 100%, and both agents together caused 130% of the PDBu response. An equal amount of cPLA 2 was immunoprecipitated in each sample as controlled by Western blotting (not shown). ent translocation (7).
In contrast to the results observed with PDBu and ionophore, PKC appears not to directly regulate collagen-or thrombininduced arachidonic acid formation. Ro 31-8220 did not block [ 3 H]arachidonic acid release, nor did it inhibit formation of inositol phosphates (35,37) and the intracellular Ca 2ϩ rise by collagen and thrombin. Moreover, staurosporine, which neither inhibited PLC activity in thrombin-stimulated platelets (24) nor intracellular Ca 2ϩ elevation, did not inhibit formation of [ 3 H]arachidonic acid (Fig. 4, also mentioned in Ref. 33). Additional evidence for the regulation of arachidonic acid formation downstream of PLC but not downstream of PKC comes from the effect of PDBu on thrombin-stimulated platelets; PDBu decreased [ 3 H]arachidonic acid release by 50% (not shown) in parallel with its inhibiting effect on PLC (24). These results are in contrast with reports on zymosan-stimulated macrophages where inhibition of PKC partially decreased the release of [ 3 H]arachidonic acid (47), and on thrombin-and ATP-stimulated Chinese hamster ovary cells where phorbol esters enhanced [ 3 H]arachidonic acid release (48). The present results indicate a major role of PLC and Ca 2ϩ in the stimulation of arachidonic acid formation by both collagen and thrombin but provide evidence against a direct role for PKC in this response.
Activation of cPLA 2 in intact cells is associated with phosphorylation of the enzyme (2, 49 -53), which increases cPLA 2 activity in a reconstitution assay (2,9,12). Experiments using phorbol ester indicate that phosphorylation is not sufficient for cPLA 2 activation (present results and Ref. 49). In contrast to results by Lin et al. (8), who observed an inhibitory effect of staurosporine on A23187-and ATP-induced arachidonic acid release in Chinese hamster ovary cells overexpressing cPLA 2 , we show that phosphorylation of cPLA 2 is not necessary for arachidonic acid formation. In platelets, staurosporine abolished cPLA 2 phosphorylation after thrombin stimulation but did not reduce formation of [ 3 H]arachidonic acid (Figs. 4B and 5). It might be possible that a decrease in cPLA 2 activity due to the inhibition of phosphorylation by staurosporine was counterbalanced by the enhancing effect of the drug on Ca 2ϩ . However, it is interesting to see that despite the different effects of staurosporine and Ro 31-8220 regarding cPLA 2 phosphorylation, both agents increased thrombin-induced Ca 2ϩ elevation and [ 3 H]arachidonic acid release in a similar manner. Furthermore, A23187 was a powerful stimulus of [ 3 H]arachidonic acid release within 2 min but caused minimal phosphorylation of cPLA 2 during this stimulation time.
Both p42 mapk and p44 mapk are present in platelets (18 -20), and we found that not only p42 mapk but also p44 mapk is activated upon stimulation with collagen, thrombin, and PDBu. In agreement with a report by Qiu and Leslie (47), where A23187 did not activate MAPK in macrophages, a rise in intracellular Ca 2ϩ was not sufficient to activate this enzyme in platelets. However, Ca 2ϩ -dependent activation of MAPK has been described previously (54,55).
Lin et al. (9) demonstrated a causal link between MAPK activation and cPLA 2 phosphorylation by cotransfection experiments and mutation of Ser-505, the MAPK phosphorylation site of cPLA 2 . Subsequent studies in several cell types showed strong correlation between MAPK activation and cPLA 2 phosphorylation under physiological conditions (12,47,51,56,57). Since the correlation between MAPK activation and cPLA 2 phosphorylation does not on its own establish a causal link between these two events, we investigated the PKC dependence of the two signaling pathways. PKC stimulates MAPK in a number of cell types through a pathway that may involve direct activation of Raf (58,59). In agreement with this, we observed that activation of PKC by phorbol ester was sufficient to stimulate platelet MAPK. Moreover, PKC was necessary for MAPK activation since Ro 31-8220 inhibited p42/p44 mapk when platelets were stimulated with the agonists tested in this study (Fig. 6). In contrast, phosphorylation of cPLA 2 was maintained in the presence of Ro 31-8220, which is consistent with the absence of an inhibitory action of Ro 31-8220 on collagenand thrombin-stimulated release of [ 3 H]arachidonic acid. cPLA 2 phosphorylation does therefore not appear to be mediated downstream of PKC, which contrasts with the regulatory pathway of MAPK in platelets stimulated by collagen and thrombin. This suggests that platelets must contain enzymes different from PKC and MAPK that are capable of phosphorylating cPLA 2 . In agreement with this, two reports dissociating MAPK activation and cPLA 2 phosphorylation were recently published (60,61).
In conclusion, the present study demonstrates that activation of MAPK and phosphorylation of cPLA 2 are regulated by distinct pathways in collagen-and thrombin-stimulated platelets. p42 mapk and p44 mapk seem to be regulated downstream of PKC but do not contribute to the physiological regulation of cPLA 2 . An unidentified kinase therefore mediates phosphorylation of cPLA 2 in collagen-and thrombin-stimulated human platelets.