p38 mitogen-activated protein kinase phosphorylates cytosolic phospholipase A2 (cPLA2) in thrombin-stimulated platelets. Evidence that proline-directed phosphorylation is not required for mobilization of arachidonic acid by cPLA2.

The Ca2+-sensitive 85-kDa cytosolic phospholipase A2 (cPLA2) is responsible for thrombin-stimulated mobilization of arachidonic acid for the synthesis of thromboxane A2 in human platelets. We have previously shown that thrombin activates p38 kinase, a recently discovered new member of the mitogen-activated protein kinase family (Kramer, R. M., Roberts, E. F., Strifler, B. A., and Johnstone, E. M. (1995) J. Biol. Chem. 270, 27395-27398) and also induces phosphorylation of cPLA2, thereby increasing its intrinsic catalytic activity. In the present study we have examined the role of p38 kinase in the phosphorylation and activation of cPLA2 in stimulated platelets. We have observed that activation of p38 kinase accompanies receptor-mediated events in platelets and coincides with cPLA2 phosphorylation. Furthermore, in the presence of inhibitors of p38 kinase, the proline-directed phosphorylation of cPLA2 was completely blocked in platelets stimulated with the thrombin receptor agonist peptide SFLLRN and was suppressed during the early (up to 2 min) phase of platelet stimulation caused by thrombin. Unexpectedly, we found that prevention of proline-directed phosphorylation of cPLA2 in stimulated platelets did not attenuate its ability to release arachidonic acid from platelet phospholipids. We conclude that: 1) cPLA2 is a physiological target of p38 kinase; 2) p38 kinase is involved in the early phosphorylation of cPLA2 in stimulated platelets; and 3) proline-directed phosphorylation of cPLA2 is not required for its receptor-mediated activation.

On activation of platelets with physiological agonists such as thrombin, significant amounts of arachidonic acid are rapidly liberated for transformation to thromboxane A 2 via the cyclooxygenase-thromboxane synthase pathway. There is substantial evidence to indicate that this efficient receptor-mediated mobilization of arachidonic acid is mediated by a phospholipase A 2 pathway (1,2) and that the involved phospholipase A 2 is the Ca 2ϩ -sensitive cytosolic phospholipase A 2 (cPLA 2 ) 1 (3,4).
Many studies with different cellular systems, including platelets, have documented that phosphorylation of cPLA 2 by receptor-mediated events accompanies the stimulated release of arachidonic acid from cellular phospholipids (5). Lin et al. (6) established that this phosphorylation is "activating" (i.e. it increases the catalytic activity of cPLA 2 severalfold) and occurs at Ser 505 residing within a MAP kinase consensus sequence (Pro-Leu-Ser 505 -Pro). In fact, cPLA 2 was phosphorylated and activated by the MAP kinase ERK2 in vitro and in vivo (i.e. in cultured cells overexpressing cPLA 2 and ERK2; Ref. 6), and accordingly, the ERKs were taken to be responsible for the proline-directed phosphorylation of cPLA 2 observed in various cellular systems. Surprisingly, we have noted that phosphorylation of cPLA 2 occurred in the absence of ERK activation in human platelets stimulated with the thrombin receptor agonist peptide SFLLRN (7). Furthermore, under conditions in which ERK activation was completely suppressed by protein kinase C inhibitors, cPLA 2 phosphorylation induced by thrombin or collagen was unaffected (8). Last, PD 098059, a specific inhibitor of the activation of ERKs, did not block thrombin-induced cPLA 2 phosphorylation (9). Taken together, these findings suggested that kinases other than the ERKs may be involved in receptor-mediated phosphorylation of cPLA 2 .
We have recently shown that in addition to the ERKs, platelets contain p38 kinase (10), a recently discovered MAP kinase typically activated by inflammatory cytokines and environmental stress. On the other hand, we were unable to detect kinases belonging to the Jun nuclear kinase subfamily of the stressactivated MAP kinases (11). Thrombin induces a rapid and robust activation of p38, suggesting that this kinase may play a role in platelet function (10). It was recently reported that p38 kinase provides a signal crucial for platelet aggregation at low agonist concentrations (12). The present study was undertaken to further elucidate the functional role of p38 in cPLA 2 activation. Using inhibitors of p38 kinase we found that cPLA 2 is a physiological target of p38 kinase in platelets stimulated via the thrombin receptor. Our results further indicate that prevention of cPLA 2 phosphorylation by such p38 kinase inhibitors does not attenuate the receptor-mediated liberation of arachidonic acid, suggesting that proline-directed phosphorylation of cPLA 2 is not a prerequisite for its activation in stimulated platelets. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 1 The abbreviations used are: cPLA 2 , cytosolic phospholipase A 2 ; MAP, mitogen-activated protein; ERK, extracellular signal-regulated kinase; CAPS, 3-(cyclohexylamino)-1-propanesulfonic acid; MAFP, methylarachidonyl fluorophosphonate; PAGE, polyacrylamide gel electrophoresis; SB 203580, 4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4-pyridyl)1H-imidazole; SB 202190, 4-(4-fluorophenyl)-2-(4hydroxyphenyl)-5(4-pyridyl)-1H-imidazole; U46619, 911-dideoxy-11␣, 9␣-epoxymethano-prostaglandin F 2␣ .
Labeling of Platelets with [ 32 P]Orthophosphoric Acid and Immunoprecipitation of 32 P-labeled cPLA 2 -To measure thrombin-induced incorporation of phosphate into cPLA 2 , washed platelets were prelabeled with 0.5 mCi/ml [ 32 P]orthophosphate for 2 h and stimulated with thrombin, and cPLA 2 was immunoprecipitated with anti-cPLA 2 antiserum as detailed previously (8). Immunoprecipitates were subjected to SDS-PAGE and transferred to polyvinylidene difluoride membranes, and 32 P-labeled cPLA 2 was visualized by autoradiography.
Other Procedures-Since MAFP targets the active site serine of cPLA 2 , we examined its effect on thrombin, a serine protease, using a thrombin-fibrinogen clotting assay (17). We found that in this assay 50 M MAFP increased the clotting time only minimally (i.e. 1.3-fold). By comparison, PPACK, a potent serine-directed active site inhibitor of thrombin, increased the clotting time by greater than 25-fold. Thromboxane A 2 production was determined by radioimmunoassay using nonaspirinized, gel-filtered human platelets as described previously (18). PLA 2 activity was assayed as described previously (14).

Receptor-mediated Activation of p38 MAP Kinase
Correlates with the Phosphorylation of cPLA 2 -We previously determined the kinase activity of the different platelet MAP kinases resolved by MonoQ chromatography and observed that thrombininduced stimulation of p38 kinase is more rapid and more potent than that of ERK1 and -2 (10). Due to the small shift in electrophoretic mobility and high reactivity with antiphosphotyrosine antibodies, activated p38 can be readily detected in platelet extracts despite its molecular mass being almost the same as that of ERK2 (41.3 versus 41.4 kDa, respectively). Thus, as shown in Fig. 1, p38 kinase activated by thrombin displays only a small decrease in electrophoretic mobility compared with activated ERK2 (and ERK1) (lanes 4 -6 versus 10 -12). However, activated p38 can be readily visualized by antiphosphotyrosine antibodies (lanes 2 and 3), whereas activated ERKs due to comigration with the very abundant actin at 45 kDa react only poorly (lane 3 versus lane 12).
the thromboxane A 2 -mimetic U46619 activating the platelet thromboxane A 2 receptor (20) caused a weak activation of both p38 kinase and the ERKs. As further demonstrated in Fig. 2, the Ca 2ϩ ionophore A23187 solely stimulated p38 kinase, and conversely, the protein kinase C activator phorbol-dibutyrate activated only the ERK kinases. Collectively, these results show that p38 kinase is activated by all receptor agonists tested and some, but not all, agents that selectively target intracellular signaling events (i.e. cytosolic free [Ca 2ϩ ] or protein kinase C). To examine the temporal relationship between receptor-mediated activation of p38 kinase and cPLA 2 phosphorylation, we stimulated platelets with thrombin and the thrombin receptor agonist peptide SFLLRN for 0 -5 min. As shown in Fig. 3, in platelets exposed to thrombin, activation of p38 MAP kinase preceded the time-dependent decrease in the electrophoretic mobility of cPLA 2 characteristic of its phosphorylation on Ser 505 . By comparison, the stimulation of the ERKs was delayed relative to the appearance of phosphorylated cPLA 2 . Fig. 3 further demonstrates that SFLLRN induced less pronounced activation of p38 and failed to activate the ERKs but nonetheless caused phosphorylation of cPLA 2 . Taken together, the above results show that activation of p38 kinase is common to receptor-mediated events in platelets and that the profile of cPLA 2 phosphorylation follows that of p38 activation, suggesting that receptor-induced phosphorylation of cPLA 2 may be caused by p38 kinase.
Inhibitors of p38 MAP Kinase Suppress Phosphorylation of cPLA 2 -To examine whether cPLA 2 is the physiological substrate of p38 kinase in stimulated platelets, we exploited the fact that Lee and Young (21) have recently described pyridinyl imidazoles that are potent and specific inhibitors of p38 kinase. We synthesized two compounds, SB 202190 and SB 203580, and found that they potently inhibit platelet p38 kinase in vitro (IC 50 , 8 and 40 nM, respectively; Fig. 4A). We noted that although SB 202190 is the more potent compound, it is less soluble than SB 203580 at concentrations greater than 10 M. To directly probe the involvement of p38 in the phosphorylation of cPLA 2 in stimulated platelets, we preincubated platelets for 5 min with increasing concentrations of the p38 inhibitors and looked for an effect on thrombin-induced phosphorylation of cPLA 2 . As shown in Fig. 4B, both SB 202190 and SB 203580 suppressed the thrombin-induced shift in electrophoretic mobility of cPLA 2 , indicative of the proline-directed phosphorylation, in a dose-dependent fashion. We then studied the time dependence of the inhibition of cPLA 2 phosphorylation by 10 M SB 202190 in platelets that were stimulated for 0 -5 min with a high dose of either thrombin or SFLLRN. As demonstrated in Fig. 5, in SFLLRN-stimulated platelets, cPLA 2 phosphorylation was completely suppressed in the presence of the p38 inhibitor. Likewise, thrombin-mediated phosphorylation of cPLA 2 was inhibited after 1 min of stimulation and attenuated after 2 min compared with control incubations with vehicle. After 5 min of stimulation with thrombin, however, cPLA 2 became phosphorylated in platelets despite the presence of the p38 inhibitor. These results provide the first direct evidence that p38 kinase is responsible for cPLA 2 phosphorylation in SFLLRN-stimulated platelets and is involved in the early phosphorylation of cPLA 2 in thrombin-stimulated platelets. To further analyze the two temporal phases of cPLA 2 phosphorylation, we examined the effect of p38 inhibition on thrombin-

FIG. 2. Differential activation of platelet MAP kinases by various agonists.
Platelets at 1 ϫ 10 9 /ml were incubated for 0 -5 min with thrombin (5 units/ml), collagen (100 g/ml), U46619 (10 M), calcium ionophore A23187 (2 M), or phorbol-dibutyrate (1 M); reactions were stopped, and solubilized extracts subjected to SDS-PAGE and immunoblotting as detailed under "Experimental Procedures." The blotting antibodies indicated to the right included anti-phosphotyrosine antibodies (anti-PY) and anti-ERK antibodies (anti-ERK1/2) as described in the legend of Fig. 1. The data shown are representative of three independent experiments yielding similar results. mediated cPLA 2 phosphorylation after preincubation of platelets with 30 M SB 203580 followed by stimulation for 2 or 5 min with increasing concentrations of thrombin. As shown in Fig. 6, after 2 min of thrombin treatment, as expected, the phosphorylation of cPLA 2 was completely suppressed in the presence of SB 203580. In contrast, following the 5-min stimulation, cPLA 2 phosphorylation was prevented by the p38 inhibitor at low doses of thrombin but only partially suppressed at higher doses. Altogether these results are consistent with the notion that proline-directed phosphorylation of cPLA 2 is initially mediated by p38 kinase and may subsequently occur via p38 kinase-independent mechanisms. However, we noted that under conditions in which the gel shift is completely prevented by p38 inhibitors (i.e. stimulation for 2 min), the enhancement of cPLA 2 activity in lysates from platelets stimulated with thrombin or SFLLRN is suppressed by only 50% (Table I). Likewise, thrombin-induced phosphorylation of cPLA 2 monitored by 32 P incorporation is only partially blocked. Thus, as demonstrated in Fig. 7, after preincubation of 32 Plabeled platelets with SB 203580 and exposure to thrombin for 2 min, the 32 P labeling of cPLA 2 was reduced to 50% of the control. Collectively, these results suggest that a kinase that phosphorylates cPLA 2 at a site distinct from Ser 505 and therefore does not induce a gel shift of cPLA 2 also participates in the early receptor-mediated phosphorylation and activation of cPLA 2 .
Effect of p38 Inhibitor on Arachidonic Acid Mobilization and Thromboxane A 2 Generation in Stimulated Platelets-As shown above, preincubation with p38 inhibitors resulted in complete inhibition of proline-directed phosphorylation of cPLA 2 in platelets exposed to thrombin up to 2 min or SFLLRN up to 5 min. We therefore questioned whether suppression of this cPLA 2 phosphorylation may affect its ability to hydrolyze platelet phospholipids in response to these agonists. Aspirinized platelets were prelabeled with [ 3 H]arachidonic acid, incubated with and without p38 inhibitor, and then exposed to either SFLLRN or thrombin. The released arachidonic acid was then measured as described under "Experimental Procedures." Unexpectedly, as shown with a dose dependence study for thrombin (Fig. 8A) and a time course experiment for thrombin (Fig. 8B) and SFLLRN (Fig. 8C), the stimulus-induced arachidonic acid mobilization was not affected at any dose or time in the presence of p38 inhibitors. Importantly, as depicted in Fig.  9, the thrombin-and SFLLRN-mediated liberation of arachidonic acid was effectively inhibited by MAFP, a recently described irreversible inhibitor of cPLA 2 (22), but not LY311727, a potent inhibitor of secretory group II phospholipase A 2 (23). Likewise, bromoenol lactone, a known inhibitor of Ca 2ϩ -independent PLA 2 activities (24), did not significantly affect stimulus-induced mobilization of arachidonic acid (data not shown). Taken together, these data demonstrate that effective inhibition of early agonist-induced, proline-directed phosphorylation of cPLA 2 does not attenuate its ability to release arachidonic acid from platelet phospholipids, suggesting that this phosphorylation is not required for activation of cPLA 2 .
While this study was in progress, Saklatvala et al. (12) reported that platelet aggregation caused by minimal concentrations of collagen and U46619 was inhibited by SB 203580. Since in platelets these processes are dependent on the production of thromboxane A 2 , this inhibition might be due to im-   SO), platelets (at 0.5 ϫ 10 9 /ml) were incubated for 2 min in the presence of buffer, thrombin (2.5 units/ml), or SFLLRN (25 M) and lysed by sonication. Lysates (20 g of protein) were assayed for cPLA 2 activity as detailed under "Experimental Procedures." Values in parentheses indicate the fold increase in enzymatic activity of cPLA 2 in lysates from stimulated compared with nonstimulated platelets. The gel shift of cPLA 2 was completely prevented by SB 203580, as verified by SDS-PAGE and immunoblotting analyses. The data are representative of three different experiments; values shown for cPLA 2 activity are means Ϯ SD, assaying three separate platelet incubations in duplicate.  5. Time course of inhibition of cPLA 2 phosphorylation by p38 inhibitors. Platelets at 1 ϫ 10 9 /ml were preincubated for 5 min in the presence of 10 M SB 202190 or vehicle (1% Me 2 SO) and then stimulated with thrombin (10 units/ml) or SFLLRN (100 M) for 0 -5 min. After addition of Triton X-100 stopping mixture and SDS sample buffer, solubilized extracts were subjected to SDS-PAGE and immunoblotting, probing with anti-cPLA 2 IgG. The data shown are representative of two independent experiments yielding similar results. paired thromboxane A 2 formation. In fact, we observed that nonaspirinized platelets preincubated for 5 min with 10 M SB 203580 generated significantly less thromboxane A 2 in response to SFLLRN than control platelets (Fig. 10A). Since our studies described above clearly showed that prevention of p38 kinase-mediated cPLA 2 phosphorylation did not affect its ability to release arachidonic acid from platelet phospholipids, we reasoned that SB 203580 could affect the metabolism of newly released arachidonic acid rather than its liberation. To bypass cPLA 2 , we therefore used arachidonic acid to stimulate platelets. As depicted in Fig. 10B, on preincubation of platelets with 10 M SB 203580, the amount of thromboxane A 2 generated in response to arachidonic acid was indeed markedly decreased compared with control platelets. Hence, in SFLLRN-stimulated platelets (Fig. 10A) the p38 inhibitor SB 203580 attenuated thromboxane A 2 formation downstream of cPLA 2 via inhibition of cyclooxygenase (or thromboxane synthase). Our findings thus differ from those of Saklatvala et al. (12), who stated that SB 203580 does not act as a cyclooxygenase inhibitor. DISCUSSION We believe that p38 kinase is involved in the receptor-mediated phosphorylation of cPLA 2 in stimulated platelets for the following reasons. First, the early (within 2 min) onset of proline-directed cPLA 2 phosphorylation correlates with the rapid activation of p38 kinase by thrombin, consistent with cPLA 2 being a target of p38 kinase. Second, p38 inhibitors effectively suppress the thrombin-mediated phosphorylation of cPLA 2 during this early phase. Third, in SFLLRN-stimulated platelets, in which p38 is the only activated MAP kinase, prolinedirected phosphorylation of cPLA 2 is prevented by p38 inhibitors at all times. Last, we have previously demonstrated that cPLA 2 , but not the S505A mutant cPLA 2 , is a substrate for p38 kinase in vitro, suggesting that p38 mediates the "activating" phosphorylation at Ser 505 of cPLA 2 (7).
Previous studies have convincingly documented that the majority of the released arachidonic acid in stimulated platelets is provided by the action of a phospholipase A 2 and not via a phospholipase C-diacylglycerol hydrolase pathway (1, 2). Platelets, in addition to cPLA 2 , also possess secretory group II phospholipase A 2 that is rapidly secreted on platelet activation (25). In the present study we demonstrate that MAFP, the active site-directed irreversible inhibitor of cPLA 2 (22), is able to effectively suppress thrombin receptor-mediated liberation of arachidonic acid. LY311727, a potent active site inhibitor of secretory group II phospholipase A 2 (23), on the other hand, does not affect stimulated arachidonic acid release. These results are in agreement with earlier observations documenting the critical involvement of cPLA 2 in the stimulated mobilization of arachidonic acid from platelet phospholipids (3,4). To date there is no evidence for the presence of a Ca 2ϩ -independent PLA 2 activity in platelets, and accordingly, bromoenol lactone inhibitors do not significantly alter arachidonic acid mobilization in stimulated platelets.
On activation of platelets with thrombin or collagen, arachidonic acid is rapidly liberated from platelet phospholipids, with the bulk of release occurring within 1-2 min of stimulation (7, 26 -28). In fact, hydrolysis of phosphatidylcholine and phosphatidylethanolamine, in which the majority of the esterified arachidonate resides, could be detected as early as 20 s after stimulation of platelets with thrombin, and this was accompanied by the concomitant accumulation of lyso-phosphatidylcholine and lyso-phosphatidylethanolamine (29,30). At these early time points cPLA 2 is partially phosphorylated, and its catalytic activity is increased 1.5-2-fold (14). The p38 kinase is maximally stimulated by thrombin or collagen within 1 min. Hence, the temporal relationship between p38 kinase activation and the cPLA 2 gel shift strongly suggests that the kinase responsible for this early phosphorylation is p38. The identity of the kinase responsible for the "late" proline-directed phosphorylation of cPLA 2 remains to be elucidated. Although ERKs have been implicated in the phosphorylation of cPLA 2 in numerous cellular systems (5), they are not likely to be involved in the phosphorylation of cPLA 2 in platelets stimulated with physiological agonists (7)(8)(9).
The MAP kinase family consists of at least three subgroups, including the ERKs, Jun nuclear kinases, and p38 protein kinases, that are regulated by distinct signal transduction pathways (11). It was originally found that p38 kinase is activated in response to inflammatory cytokines, bacterial endotoxin, and environmental stress (31)(32)(33)(34)(35). However, recent studies with human platelets revealed that thrombin and collagen also promote activation of p38 kinase (10,12). The functional consequences of p38 activation have not been clear. The p38 inhibitor SB 203580 was used to demonstrate that MAPKAP kinase 2 is a physiological target of p38 kinase in cells exposed to interleukin 1, cellular stresses, and bacterial endotoxin (36). Furthermore, activation of p38 kinase by its activators MKK3 and MKK6 in cultured cells confirmed that the transcription factors ATF2 and Elk-1 are substrates of p38 kinase (37). We observed previously that p38 kinase partially purified from thrombin-stimulated platelets readily phosphorylates cPLA 2 in vitro (7). In this present study we demonstrate that receptormediated phosphorylation of cPLA 2 is prevented by inhibitors of p38 kinase and thereby establish that cPLA 2 is also a target for p38 kinase in stimulated platelets. It was previously reported that bacterially expressed p38 kinase does not act on cPLA 2 (35). However, it should be noted that the enzymatic activity of bacterially expressed MAP kinases is orders of magnitude smaller than that of cellular MAP kinases physiologically activated by phosphorylation on both Thr 180 and Tyr 182 (38). Furthermore, the substrate specificity of recombinant p38 kinase activated due to some autophosphorylation may differ from that of physiologically activated platelet p38 kinase.
Lin et al. (6) observed that in cultured cells overexpressing cPLA 2 the stimulus-induced release of arachidonic acid was markedly enhanced compared with control cells. In sharp contrast, when S505A mutant cPLA 2 was similarly overexpressed, no such increase was observed. This finding suggested that phosphorylation at Ser 505 is required for receptor-mediated activation of cPLA 2 to release arachidonic acid from cellular phospholipids. Many subsequent reports have proposed that such phosphorylation of cPLA 2 is a critical step in the sequence of events leading to the mobilization of arachidonic acid in stimulated cells, but direct evidence to support this hypothesis has not been provided (5). Surprisingly, we observed that prevention of the proline-directed phosphorylation of cPLA 2 in stimulated platelets only partially suppressed the enhancement of the enzymatic activity of cPLA 2 and did not affect its ability to rapidly mobilize arachidonic acid in stimulated platelets. Our findings thus indicate that, at least in platelets, proline-directed phosphorylation is not a prerequisite for receptor-mediated activation of cPLA 2 .
We conclude that the early thrombin-induced, proline-directed phosphorylation of cPLA 2 that contributes to its increased catalytic activity is mediated by p38 MAP kinase. However, in platelets this phosphorylation of cPLA 2 is not a critical factor for the rapid mobilization of arachidonic acid by cPLA 2 , suggesting that other as yet unidentified kinases may be involved. In fact, recently, novel phosphorylation sites of cPLA 2 have been identified (39) that may be targeted by such kinases, and it will be of great interest to study their involvement in the regulation of cPLA 2 .