Thrombin induces activation of p38 MAP kinase in human platelets.

In human platelets a proline-directed kinase distinct from the ERK MAP kinases is stimulated by both thrombin and the thrombin receptor agonist peptide SFLLRN and may be involved in the activation of Ca(2+)-dependent cytosolic phospholipase A2 (Kramer, R. M., Roberts, E. F., Hyslop, P. A., Utterback, B. G., Hui, K. Y., and Jakubowski, J.A. (1995) J. Biol. Chem. 270, 14816-14823). Here we show that this kinase is identical with or closely related to p38 (the mammalian homolog of HOG1 from yeast), a recently discovered protein kinase typically activated by inflammatory cytokines and environmental stress. Further, we demonstrate that activation of this kinase by thrombin is transient (with maximal stimulation at 1 min), is accompanied by tyrosine phosphorylation, and precedes the activation of the ERK kinases. This is the first report to show that p38 kinase is activated by thrombin and to suggest a role for this MAP kinase in the thrombin-mediated signaling events during platelet activation.


In human platelets a proline-directed kinase distinct from the ERK MAP kinases is stimulated by both thrombin and the thrombin receptor agonist peptide SFLLRN and may be involved in the activation of Ca
Here we show that this kinase is identical with or closely related to p38 (the mammalian homolog of HOG1 from yeast), a recently discovered protein kinase typically activated by inflammatory cytokines and environmental stress. Further, we demonstrate that activation of this kinase by thrombin is transient (with maximal stimulation at 1 min), is accompanied by tyrosine phosphorylation, and precedes the activation of the ERK kinases. This is the first report to show that p38 kinase is activated by thrombin and to suggest a role for this MAP kinase in the thrombin-mediated signaling events during platelet activation.
We have recently shown that thrombin stimulates the activity of the MAP 1 kinases ERK1 and ERK2 but also activates another proline-directed kinase that is distinguishable from ERK1/2 based on its strong binding to anion exchange resin and the lack of reactivity with anti-ERK1/2 antibodies (1). We further noted that this kinase readily phosphorylates cPLA 2 but not the S505A mutant of cPLA 2 . This observation indicated that the serine residing within the MAP kinase consensus sequence (i.e. Pro-Leu-Ser 505 -Pro) is the target phosphorylation site for the kinase. Significantly, the thrombin receptor agonist peptide SFLLRN also activated this proline-directed kinase but completely failed to stimulate ERK1/2. Nonetheless SFLLRN, like thrombin, mediated activation of cPLA 2 by phosphorylation, and we reasoned that this unidentified kinase could play a role in the signal transduction pathways activated through the thrombin receptor. We therefore further characterized the kinase with the goal to determine its identity and define its role in the thrombin-induced signaling events during platelet activation.
Partial Purification of p38 and ERK Kinases by MonoQ Chromatography-High speed supernatants were subjected to chromatography on a MonoQ HR 5/5 column (Pharmacia Biotech Inc.) at a flow rate of 1.5 ml/min collecting 0.5-ml fractions using two different procedures. In order to partially purify the p38 kinase, the column was first equilibrated in buffer A (1 mM EGTA, 1 mM DTT, 100 M Na 3 VO 4 , 50 mM ␤-glycerophosphate, pH 7.5) containing 150 mM NaCl and subjected to a 30-ml linear salt gradient from 150 to 530 mM NaCl. For rapid partial purification of the p38 and the ERK kinases, the MonoQ column was equilibrated in buffer A containing 50 mM NaCl and then subjected to a step gradient from 50 to 250 mM NaCl (for elution of ERK1/2), followed by a step gradient from 250 to 450 mM NaCl (for elution of p38). MonoQ fractions (0.5 ml) were collected into 10 l of a mixture providing (final concentrations) 100 nM microcystin, 100 M leupeptin, 0.1 mg/ml aprotinin, and 1 mM Pefabloc.
Assay for Proline-directed Kinases-Kinase assays were performed as described before (1) using the Thr 669 peptide substrate (KRELVE-PLTPSGEAPNQALLR from Macromolecular Resources, Colorado State University) and the SpinZyme system (Pierce).

RESULTS AND DISCUSSION
In human platelets thrombin activates several kinases that readily phosphorylate the Thr 669 peptide derived from the epidermal growth factor receptor (1). Based upon the distinct chromatographic and immunological characteristics, these kinases could be distinguished and found to consist of the MAP kinases ERK1/2, as well as another unidentified proline-directed kinase. When extracts from control and thrombin-stimulated platelets were applied to MonoQ in buffer containing 150 mM NaCl, ERK1/2 flowed through the column. The unknown thrombin-activated Thr 669 kinase, on the other hand, bound tightly to the column and eluted at Ն350 mM NaCl ( 1A). Consequently, the isoelectric point (IEP) of this kinase had to be significantly lower than that of ERK1/2 (ϳ6.8). For example, the IEP of cPLA 2 that binds similarly to MonoQ eluting with ϳ400 mM NaCl is 5.1 (4). Only three recently identified proline-directed kinases exhibit calculated IEPs in agreement with the observed chromatographic behavior of the platelet Thr 669 kinase. These include JNK2 (p54␣ 1 ) (5), p38 kinase (6), and an ERK3 homolog (referred to as p97 MAPK ) (7) with IEPs of 5.7, 5.6, and 4.8, respectively.
We therefore subjected extracts from thrombin-stimulated platelets and the most active Thr 669 kinase MonoQ fractions 34 and 35 (see Fig. 1A) to SDS-PAGE/immunoblotting, probing with antibodies against JNK2, p38, and ERK3. As shown in Fig. 2A anti-p38 antibodies specifically recognized a protein of ϳ40 kDa that was enriched in the MonoQ fractions (lane 2) compared with the loaded platelet extract (lane 1). The same ϳ40-kDa protein strongly reacted with anti-phosphotyrosine antibodies (lane 4). Although ERK1 and ERK2 could be readily detected in extracts (lane 5), they were absent in the MonoQ fractions containing the Thr 669 kinase activity (lane 6). As demonstrated in Fig. 2B, no immunoreactivity could be detected in extracts or active MonoQ fractions when probing with anti-JNK antibodies (lanes 8 and 9) or anti-ERK3 antibodies (lanes 11 and 12). By comparison both kinases could be readily seen when standard proteins were tested (lane 7 and lane 10). We confirmed the specificity of two anti-p38 antibodies (anti-p38N and anti-p38C) for the ϳ40-kDa protein by competition experiments with the respective C-and N-terminal peptides of p38 used for immunization. As demonstrated in Fig. 2C, the reactivity with the ϳ40-kDa protein was significantly decreased when the immunizing peptides were present during immunoblotting, including the anti-p38N (lane 14 versus lane 13) and the anti-p38C (lane 16 versus lane 15) antibodies. We further examined whether the kinase activity eluting from the MonoQ column (Fig. 1A) paralleled the immunoreactivity of the eluting kinase protein probing with both anti-p38 and antiphosphotyrosine antibodies. As shown in Fig. 1B (upper right  panel), the MonoQ elution of the ϳ40-kDa protein recognized by the anti-p38 antibodies correlated with the thrombin-induced Thr 669 kinase activity. Likewise, coincident with the peak of kinase activity we detected thrombin-induced tyrosine  1 and 2), anti-phosphotyrosine (␣P-Y) monoclonal antibodies 4G10 (Upstate Biotechnology) plus PY20 (ICN) at 1 g/ml each (lanes 3 and 4), and anti-ERK1/2 polyclonal antibody erk1-CT (Upstate Biotechnology) at 1 g/ml (lanes 5 and 6). B, anti-JNK2 polyclonal antibody (Santa-Cruz Biotechnology) at 0.1 g/ml (lanes 7-9) and anti-ERK3 antibody (Transduction Laboratories) at 1 g/ml (lanes 10 -12). The ability of anti-JNK2 and anti-ERK3 antibodies to recognize JNK2 and ERK3, respectively, was verified with puri- p38 Kinase Activation in Thrombin-stimulated Platelets 27396 phosphorylation of the same ϳ40-kDa protein (Fig. 1B, lower  panel). Taken together, these data indicate that the prolinedirected kinase activated by thrombin is identical with, or closely related to, the p38 MAP kinase.
We determined the kinetics of thrombin-mediated activation of the p38 kinase, resolving it from ERK1/2 by MonoQ chromatography. As shown in Fig. 3A, thrombin induced a transient stimulation of p38 kinase activity that reached a maximum at 1 min and was still detected at the latest time point measured (5 min). The amount of p38 protein purified by MonoQ chromatography was the same for all time points examined, as verified by SDS-PAGE immunoblotting (Fig. 3B). The appearance and disappearance of p38 kinase activity in thrombinstimulated platelets temporally coincided with the tyrosine phosphorylation of p38 (Fig. 3C). By comparison, activation of the ERKs was delayed with maximal stimulation at 2 min following thrombin stimulation, as shown by the ability of ERK1/2 to phosphorylate the Thr 669 peptide substrate (Fig. 4A) and the decreased electrophoretic mobility of the ERK proteins (Figs. 3C and 4B), indicative of activation (8). The data in Fig.  4 also reveal that activation by thrombin of p38 is more prominent than that of ERK1/2. A similar robust activation of p38 and delayed stimulation of ERK1/2 were observed in aspirinized platelets (where the synthesis of endogenous thromboxane A 2 is inhibited), demonstrating that p38 kinase is the target of thrombin and not of the secondary agonist thromboxane A 2 that is released from activated platelets. The p38 kinase belongs to a new subfamily of stress-activated MAP kinases related to the HOG1 gene product, a kinase required for adaptation to osmotic stress in Saccharomyces cerevisiae (9), and has only recently been identified in mammalian cells. Thus, Han et al. (10) first described this novel kinase FIG. 3. Time course of thrombin-mediated activation of p38 kinase. Platelets (0.6 ml at 1.25 ϫ 10 9 /ml) were incubated at 37°C in the absence (ϪTHR) and presence of 5 units/ml of thrombin (ϩTHR) as described under "Experimental Procedures." After addition of 150 l of Triton X-100 stopping mixture and ultracentrifugation, the extracts were diluted with 150 mM NaCl, 1 mM EGTA, 1 mM DTT, 100 M Na 3 VO 4 , and 50 mM ␤-glycerophosphate, pH 7.5 and subjected to MonoQ chromatography as described under "Experimental Procedures." A, determination of kinase activity in pooled MonoQ fractions containing the p38 kinase using 2 mM Thr 669 peptide substrate as detailed under "Experimental Procedures." B, SDS-PAGE/immunoblotting verifying that equal amounts of p38 kinase protein were present in the pooled active MonoQ fractions derived from platelets incubated without thrombin (ϪTHR) for 30 s to 5 min (lanes 1-4) and with thrombin (ϩTHR) for 0 -5 min (lanes 5-10). C, SDS-PAGE/immunoblotting of solubilized lysates (10 l) from platelets incubated with thrombin (5 units/ml) for 0 -5 min as indicated, probing for the presence of tyrosine-phosphorylated proteins (␣P-Y, lanes 1-5), p38 kinase (lanes 6 -10), and ERK1/2 (lanes 11-14) using the antibodies as described in Fig. 2. The data shown are representative of two independent experiments yielding similar results, and the values shown in A are means Ϯ range of duplicate incubations.

FIG. 4. Differential activation of p38 and ERK kinases by thrombin.
Platelets (0.6 ml at 1.25 ϫ 10 9 /ml) were incubated at 37°C with 5 units/ml thrombin, and the reaction was quenched as described under "Experimental Procedures." The solubilized lysates were cleared by ultracentrifugation, diluted with 1 mM EGTA, 1 mM DTT, 100 M Na 3 VO 4 , and 50 mM ␤-glycerophosphate, pH 7.5 (buffer A), and passed over a MonoQ column equilibrated in buffer A containing 50 mM NaCl. ERK1/2 and p38 were eluted with a step salt gradient and recovered in 1.5 ml of buffer A containing 250 mM NaCl and 1.5 ml of buffer A containing 450 mM NaCl, respectively, as described under "Experimental Procedures." A, determination of kinase activity in 8.3 l of the 1.5-ml MonoQ pools using 2 mM Thr 669 peptide substrate as detailed under "Experimental Procedures." B, SDS-PAGE/immunoblotting demonstrating the presence of ERK1/2 (lanes 5-8) and p38 (lanes 1-4) in the respective 1.5-ml MonoQ pools probing with the antibodies described in Fig. 2. The data shown are representative of two independent experiments yielding similar results, and the values shown in A are means Ϯ range of duplicate incubations, each assayed in duplicate.
p38 Kinase Activation in Thrombin-stimulated Platelets 27397 of apparent molecular mass of 38 kDa (therefore referred to as p38) in cells of monocytic lineage, observing that it is rapidly phosphorylated on tyrosine residues in response to endotoxin. Cloning of the p38 kinase revealed that its predicted sequence is 52% identical to the yeast kinase HOG1 (6) and shares with HOG1 the unique sequence TGY comprising the dual phosphorylation site typical of MAP kinases. Lee et al. (11) identified the new kinase CSBP, a target of cytokine synthesis inhibitors, that was found to be identical with p38 kinase. Furthermore, Rouse et al. (12) discovered a stress-activated kinase recognized by antibodies against the Xenopus kinase Mpk2, a kinase closely related to HOG1 from yeast, and Freshney et al. (13) purified an interleukin-1-stimulated kinase from human epidermal carcinoma cells whose biochemical properties closely resembled those of the p38 kinase. Recent studies by Raingeaud et al. (14) showed that p38 kinase is activated not only by osmotic stress and endotoxin but is also stimulated by inflammatory cytokines, particularly tumor necrosis factor, and exposure to UV radiation. In contrast, the p38 kinase was only poorly activated by growth factors, interferon-␥ and phorbol ester (6,10,14). Here, we report that the serine protease thrombin known to activate a heterotrimeric G protein-coupled receptor causes a marked activation of the p38 kinase. The stimulation of p38 kinase by thrombin not only precedes that of the ERKs but is also more pronounced than that of the ERKs. This suggests that p38, rather than the ERKs, may be involved in early proline-directed phosphorylation events during thrombin-mediated platelet activation. The difference in the temporal pattern of activation is consistent with the notion that the p38 and ERK MAP kinases are independently regulated by distinct signaling pathways (15). The sequential kinase cascade leading to the activation of the ERK MAP kinases lies downstream of Ras and consists of two protein kinases (Raf and MAP kinase kinase) acting sequentially to activate the ERKs (15). In contrast, the upstream regulatory mechanisms and protein kinases involved in the activation of p38 are not yet fully elucidated. Activation of p38 kinase requires dual phosphorylation on Thr 180 and Tyr 182 (14). Recently, a MAP kinase kinase referred to as JNKK (16) or MKK4 (17) was identified that activates the p38 and also the JNK kinases. Another kinase, called MKK3, cloned as the human homolog of the yeast MAP kinase kinase PBS2, activated only the p38 kinase and not the ERK and JNK MAP kinases (17). Thus, MKK3 may be a component of an independent signaling pathway that specifically activates the p38 kinase. Recently, the small GTP-binding proteins Rac1 and Cdc42 were shown to be efficient activators of the signaling cascade affecting the p38 kinase (18).
Our studies show that the p38 kinase is present in human platelets, where it is transiently and potently stimulated by thrombin. Although the nature of the signaling pathways that participate in stimulation of platelets upon activation of the thrombin receptor has not been completely defined, it is known that the activated thrombin receptor couples to phosphatidylinositol biphosphate metabolism and inhibition of adenylate cyclase via the G proteins G q and G i , respectively (19). Receptors coupled to heterotrimeric G proteins are thought to activate the ERK MAP kinase pathway via activated ␣ and ␤␥ G protein subunits (20). The thrombin receptor agonist peptide SFLLRN also stimulates the proline-directed kinase, now identified as p38, but, unlike thrombin, does not activate ERK1/2 (1). These findings suggest the existence of parallel pathways leading to the activation of either p38 or the ERKs. While both pathways can be stimulated by thrombin, SFLLRN activates solely the signaling pathway causing p38 stimulation. It thus appears that, at least in SFLLRN-stimulated platelets, the p38 kinase pathway is responsible for regulation of cPLA 2 . While other physiological functions of the p38 kinase remain to be elucidated, p38 is likely to be an integral part of a signaling pathway utilized by the thrombin receptor in platelets, and it will be of great interest to further investigate the role of p38 in platelet function. CONCLUSION Thrombin rapidly and potently stimulates the p38 kinase in human platelets, demonstrating that extracellular stimuli other than stress-related events and proinflammatory cytokines can activate this proline-directed kinase. Taken together with our previous findings (1) these observations suggest that in platelets (i) thrombin activates two distinct signaling pathways that result in the activation of either the p38 or the ERK MAP kinases, (ii) the thrombin receptor agonist peptide SFLLRN exclusively signals through the p38 pathway, and (iii) cPLA 2 appears to be one of the downstream targets of p38 kinase.