Differential translocation of phospholipase C isozymes to integrin-mediated cytoskeletal complexes in thrombin-stimulated human platelets.

To investigate a role of phospholipase C (PLC) isozymes in the integrin alphaIIbbeta3-mediated signaling, their location was examined in thrombin-activated human platelets, revealing different regulation of their translocation to the cytoskeleton (CSK). In resting platelets, the major PLCs such as PLCbeta2, PLCbeta3a (155 kDa), and PLCgamma2 and the minor PLCs (PLCbeta1 and PLCgamma1) were located in the Triton X-100-soluble (Tx.Sol) fraction and the membrane skeleton, whereas PLCbeta3b (140 kDa) was present only in Tx.Sol fraction when examined by Western immunoblotting. Thrombin stimulation caused a rapid and transient translocation of PLCbeta3a and PLCbeta3b and a slower accumulation of PLCbeta2 and PLCgamma2 in the reorganized CSK. The translocation to CSK of both PLCbeta3a and PLCbeta3b, but not PLCbeta2, was dependent on integrin alphaIIbbeta3-mediated aggregation. Furthermore, an actin polymerization inhibitor, cytochalasin D, or a protein tyrosine kinase inhibitor, genistein, abolished the CSK association of alphaIIbbeta3, PLCbeta3a, and PLCbeta3b. In the genistein-pretreated platelets, pp60(c-)src, Gq, and protein kinase Calpha were no longer able to associate with CSK. In contrast, these agents had no or marginal inhibitory effects on the CSK association of PLCbeta2 and Gi2. The late diacylglycerol generation induced by thrombin stimulation was significantly reduced by the genistein treatment. These results suggest that the integrin alphaIIbbeta3-mediated cytoskeletal association of PLCbeta3 is regulated by protein tyrosine kinase and also that the activation of the relocated PLC may play a role in the late platelet-to-platelet aggregation in thrombin-stimulated human platelets.

Integrins have recently been known to play important roles in cellular functions, e.g. adhesion, migration, cell proliferation, and differentiation (1). Several lines of evidence indicate that the platelet integrin ␣ IIb ␤ 3 is involved in bi-directional signaling across the plasma membrane (2)(3)(4)(5). When platelets are stimulated by agonists such as thrombin, intracellular signal transduction (intracellular calcium increase and protein kinase C activation) can be activated, leading to a conformational change in the extracellular domains of integrin ␣ IIb ␤ 3 that increases its affinity for adhesive ligands, for example fibrinogen, resulting in platelet aggregation. Conversely, ligand binding to integrin ␣ IIb ␤ 3 is involved in outside-in signals; its association with cytoskeleton causes biochemical changes including tyrosine phosphorylation of several proteins, increase in intracellular calcium level, and activation of the calcium-dependent neutral protease (calpain) (6 -11). Although the mechanisms underlying integrin-dependent increases in intracellular calcium have not been elucidated, in some instances they appear to involve either PLC 1 -medicated, inositol 1,4,5-trisphosphateinduced calcium mobilization from the endoplasmic reticulum or the transport of extracellular calcium through plasma membrane channels.
Various protein tyrosine kinases including pp60 c-src become associated with integrins through interactions with cytoskeletal complexes induced by integrin-dependent cytoskeletal reorganization (11)(12)(13)(14)(15). Other signaling enzymes, PI 3-kinase, PI 4-kinase, PI 4,5-kinase, protein kinase C, phospholipase C, and small GTP-binding proteins, are also known to relocate to the integrin-mediated reorganized CSK in thrombin-stimulated platelets (2, 8, 16 -19). The role of PI 3-kinase in integrininduced changes in cell behavior is not fully understood. However, the finding that wortmannin, an inhibitor of PI 3-kinase, blocks thrombin-induced actin rearrangement suggests a possible involvement of the kinase in a signaling event that occurs upon ligand engagement of integrin and platelet aggregation evoked by thrombin (20). PI 4-kinase and PLC activities are increased in the reorganized cytoskeletal complex produced by thrombin stimulation (2,8). The increased amount of phosphatidylinositol 4,5-bisphosphate caused by integrin engagement would be important for the polymerization of actin, because besides the principal role as the preferred substrate for PLC, phosphatidylinositol 4,5-bisphosphate can regulate actin polymerization by binding to actin-binding proteins such as profilin and gelsolin.
Our previous study indicated evidence suggesting that thrombin activation of human platelets causes cleavage of PLC␤3 by calpain leading to its enhanced activity by G␤␥ subunits (21). However, it remains unclear if there would be functional consequences of association of PLC␤3 with CSK. The co-localization of signaling molecules with the CSK raises the intriguing possibility that CSK may be intimately involved in signal transduction induced by integrin-ligand interaction. Therefore, this study was undertaken to examine the CSK association of different PLC isozymes and also the regulatory mechanisms of the cytoskeletal reorganization that occur as a consequence of integrin-ligand interaction in thrombin-stimulated human platelets.
Platelet Activation and Aggregation-Washed platelet suspension (0.5 ml) was incubated with stirring at 37°C in an aggregometer (Frenius, Bad Homburg, Germany). CaCl 2 (1 mM) was added 30 s prior to addition of 1 unit/ml thrombin. Aggregation was measured by percentage change in light transmission. In some experiments, as indicated, platelets were preincubated for 15 min with synthetic peptide consisting of the sequence RGDS (0.5 mM) or with mAb-IN␤ 3 (50 g/ml). In some experiments, platelets were pretreated with calpain inhibitor calpeptin (30 M), cytochalasin D (20 M), or genistein (100 -150 M), or Me 2 SO (0.2%, control) for 5-20 min at 37°C prior to addition of 1 unit/ml thrombin.
Determinations of Diacylglycerol Content-Washed platelets (1 ϫ 10 9 /ml) were pretreated with Me 2 SO (0.3%, control) or genistein (150 M) at 37°C for 20 min and stimulated with thrombin (1 unit/ml) for 2 min and terminated by adding 2 ml of chloroform/methanol (1:2, v/v). Lipids were extracted by the method of Bligh and Dyer (22). Diacylglycerol (DAG) mass content was by the conversion of DAG into [ 32 P]phosphatidic acid by Escherichia coli DAG kinase in the presence of [␥-32 P]ATP, according to the method as described previously (23).
Proteins were separated on SDS-polyacrylamide gels containing 6% polyacrylamide, transferred to the polyvinylidene difluoride membrane, and then incubated with antibodies against PLC isozymes. Antigenantibody complexes were detected with the chemiluminescence (ECL) method (Amersham Corp.). The intensities of the bands were quantitated by densitometry (ATTO Densitograph AE-6900M) and quantitations were performed within the linear range. To know actin amounts in the CSK fraction, proteins were separated by SDS-polyacrylamide gel electrophoresis (10%), followed by staining of the gel with 0.2% Coomassie Brilliant Blue, and the bands corresponding to actin were measured by the densitometry.
Protein concentrations were determined using the Bio-Rad protein assay with bovine serum albumin as standard.

Distribution of PLC Isozymes in Human Platelets with or
without Thrombin Stimulation-We have previously demonstrated that human platelets contained two PLC␤3 forms that are different in molecular mass (PLC␤3a, 155 kDa and PLC␤3b, 140 kDa) when analyzed by using the Ab-PLC␤3C against the C-terminal (amino acids 1202-1217) of PLC␤3 and Ab-PLC␤3M against the 550 -561 amino acid residue of PLC␤3 (21). The PLC␤3b was only detected with Ab-PLC␤3M but not Ab-PLC␤3C, suggesting that there is C-terminal proteolysis. As indicated in Fig. 1, the distribution of these two PLC␤3 forms was different in the resting platelets; PLC␤3a was present in the Tx⅐Sol and MSK, whereas PLC␤3b was only present in Tx⅐Sol fraction. PLC␤2, PLC␤3a, and PLC␥2 were present in both Tx⅐Sol and MSK fractions but absent in CSK fraction. A small amount of PLC␥1 was present and PLC␤1 was hardly detectable. PLC␤4, PLC␦1, and PLC␦2 were unable to be detected under the conditions used here. The same volume (20 l) from each fraction (Tx⅐Sol, 1.0 ml; MSK, 0.15 ml; and CSK, 0.2 ml) was subjected to densitometric analysis of immunoblots (n ϭ 4). It has been shown that 90% of PLC␤2, 87% of PLC␤3a, and 90% of PLC␥2 were located in the Tx⅐Sol fraction, whereas 9, 10, and 8% of each PLC isozyme were in the MSK fraction. In contrast, PLC␤3b was entirely distributed in the Tx⅐Sol fraction.
Upon platelet stimulation with 1 unit/ml thrombin for 30 s, all PLC isozymes appeared in the CSK at different levels. The extents of CSK association of PLC␤3a and PLC␤3b were increased 24-and 80-fold, respectively, while those of PLC␤2 and PLC␥2 were 3-and 10-fold. At 30 s after stimulation with thrombin, approximately one-fourth of total PLC␤3b was translocated to CSK. The time course of translocations of PLC isozymes in thrombin-stimulated platelets displayed decreases of PLC␤3(a, b) in the Tx⅐Sol fraction (Fig. 2, left panels) and increases in the CSK fraction, reaching a maximum at 30 s after stimulation (Fig. 2, right panels). No significant changes of PLCs were seen in the MSK fraction (Fig. 2, middle panels). These results suggested that the PLCs translocated to CSK were originated from Tx⅐Sol but not from MSK. The marked decreases of PLC␤3(a, b) in the CSK after 30 s stimulation were compatible with our previous observations that PLC␤3(a, b) were proteolytically degraded to 100 kDa by the activation of calpain in thrombin-stimulated platelets (21). This truncated form (100 kDa) was detected in CSK (data not shown).
Although PLC␤2 and PLC␥2 were also translocated to CSK from the Tx⅐Sol fraction by thrombin stimulation, the extents were much smaller compared with those of PLC␤3. Interestingly, unlike PLC␤3, both PLC␤2 and -␥2 underwent progres- sive accumulation in CSK (Fig. 2, right panels) The total amounts of these two PLCs (␤2, ␥2) in the Triton X-100-soluble and -insoluble fractions were consistent, indicating that they are resistant to proteolytic attack by calpain.
Regulation of Cytoskeletal Association of PLC Isozymes-Aggregation requires fibrinogen binding to the activated integrin ␣ IIb ␤ 3 in thrombin-stimulated platelets (2)(3)(4)(5). To examine involvement of the fibrinogen binding to the integrin in the CSK association of PLCs, platelets were pretreated with the fibrinogen antagonist peptide RGDS or with the inhibitory monoclonal integrin ␤ 3 antibody (mAb-IN␤ 3 ) (25). It was demonstrated that the thrombin stimulation caused an increase of protein content in CSK in a time-dependent manner and reached a plateau at 30 -60 s after stimulation (Fig. 3). The increased proteins contain mainly polymerized actin as well as MSK proteins (3). The enhancement of actin polymerization was prevented by RGDS or mAb-IN␤ 3 , thereby leading to considerable inhibition of thrombin-induced aggregation. Under these conditions we have examined the CSK association of different PLC isozymes. It has been known that integrin ␣ IIb ␤ 3 is located in MSK in resting platelets and then translocated to the reorganized actin-rich CSK by platelet aggregation (3,24). As shown in Fig. 4, the integrin ␣ IIb ␤ 3 appeared in CSK at 15 s after thrombin stimulation and reached the maximal level at 120 s when platelets were nearly fully aggregated. The reloca-tion of ␣ IIb ␤ 3 to the actin-rich CSK was inhibited by RGDS or mAb-IN␤ 3 . The time course of PLC translocation to CSK showed a good correlation with that of the integrin ␣ IIb ␤ 3 translocation. Interestingly, the pretreatment of platelets with RGDS or mAb-IN␤ 3 caused differential inhibition in CSK association of PLC isozymes (Table I)

FIG. 3. Effects of RGDS and mAb-IN␤ 3 on the protein content in CSK and aggregation in thrombin-stimulated platelets.
Washed platelets were pretreated with RGDS (0.5 mg/ml) or mAb-IN␤ 3 (50 g/ml) for 5 min at 37°C and then stimulated with thrombin (1 unit/ml) for the indicated times. Platelet aggregation was measured as described under "Experimental Procedures." A, the protein content in the CSK fraction was isolated from 10 9 platelets and expressed in mean Ϯ S.D. from three different experiments. B, aggregation was measured in changes of light transmission and the value at 5 min was designated as 100%.

FIG. 4. Effects of RGDS and mAb-IN␤ 3 on CSK association of PLC isozymes and integrin ␣ IIb ␤ 3 in thrombin-stimulated platelets.
Washed platelets were pretreated with RGDS (0.5 mg/ml) or mAb-IN␤ 3 (50 g/ml) and then stimulated by thrombin (1 unit/ml) for 30 s. Control platelets without the pretreatments were stimulated with thrombin for the indicated times. The CSK fractions were prepared, lysed, and subjected to SDS-PAGE and immunoblotted as described in Fig. 1. Data are representative of three different experiments. PLC␤3(a, b) was almost completely prevented by either RGDS or mAb-IN␤ 3 , but the association of PLC␥2 and ␤2 was much less affected. The PLC isozymes, PLC␤2, PLC␤3a, and PLC␥2, in MSK of resting platelets were unaffected with these treatments (data not shown). These results suggested the differential regulation in the integrin ␣ IIb ␤ 3 -mediated cytoskeletal association of PLC isozymes.
As seen in Figs. 2 and 4, during a concomitant increase in aggregation, there were decreases in CSK association of PLC␤3(a, b) after the peak at 30 s after stimulation. In our previous study, similar decreases of these PLC␤3(a, b) enzymes have been observed in intact platelets stimulated with thrombin, which were thought to be induced by calpain activation (21). Then we examined whether the translocation of PLCs to CSK are associated with calpain activation. As shown in Fig. 5, A and B, calpeptin, a specific calpain inhibitor, did not interfere with relocation of all PLCs to CSK, thus indicating that once translocated to the reorganized CSK, PLC␤3(a, b) were proteolytically modified by calpain.
It has been demonstrated that the extracellular triggering of integrin response involves hierarchies of protein interactions upon tyrosine phosphorylation and actin polymerization (26). To examine whether the CSK association of PLCs is dependent on actin integrity or protein tyrosine phosphorylation, platelets were pretreated with cytochalasin D, an inhibitor of actin polymerization, or genistein, a protein tyrosine kinase inhibitor. As indicated in Fig. 5B, an increase of polymerized actin in CSK fraction was reduced to 27% by cytochalasin D pretreatment. The same treatment caused the decreased CSK association of PLC␤3(a, b) to 23 and 14% of the untreated control, respectively (Fig. 5, A and B). In contrast, no inhibition by cytochalasin D was observed in the CSK association of PLC␤2. These results indicated that actin polymerization is involved in the CSK association of PLC␤3(a, b).
Experiments with tyrosine kinase inhibitor, herbimycin A, has indicated that tyrosine phosphorylation of various proteins is required for the cytoskeletal attachment of integrin ␣ IIb ␤ 3 (15). In the present study, we have examined the effect of genistein on the cytoskeletal association of PLC enzymes in thrombin-stimulated platelets. Genistein (100 M) pretreatment of platelets caused a marked reduction (80%) in the thrombin-induced CSK association of ␣ IIb ␤ 3 (Fig. 5, A and B). Similar inhibitory effects were observed in PLC␤3(a, b), although genistein had no effect on actin polymerization. There was no inhibition for PLC␤2 translocation. Moreover, it was shown that pp60 c-src and protein kinase C␣ were translocated to the CSK fraction in thrombin-stimulated platelets (Fig. 6). However, this event was blocked by pretreatment of genistein. Heterotrimeric GTP-binding proteins, G i2 and G q , as detected by specific antibodies, were also translocated to the reorganized CSK in thrombin-stimulated platelets. The translocation to CSK was regulated differently; the CSK association of G q was inhibited by treatment of genistein but that of G i2 was not. These findings that genistein selectively inhibited the ␣ IIb ␤ 3mediated cytoskeletal association of signaling molecules raise the possibility that tyrosine phosphorylation is involved in their differential translocation to CSK.
Possible Roles of Cytoskeletal Association of PLC␤3 in Platelet Aggregation-Our previous study indicated that thrombin stimulation of human platelets induced a biphasic accumulation of DAG by PLC activation; the early phase showed a sharp  rapid peak at 10 s after stimulation, and the subsequent second phase DAG formation is slow and sustained, reaching a maximum at 2 min, where platelets are fully aggregated (23). Since genistein prevented the translocation to CSK of PLC␤3(a, b) (Fig. 5), effect of genistein on DAG formation was examined at 2 min after thrombin stimulation. As shown in Fig. 7, a significant reduction of DAG formation was observed in genisteinpretreated platelets. Similar inhibition of DAG formation was also caused by another tyrosine kinase inhibitor, herbimycin A (data not shown). These results suggested that the late DAG formation could be partly due to PI hydrolysis via the PLC␤3(a, b) translocated to the reorganized CSK.
At the present time, we do not have relevant interpretation for the role of PLC␤3 activation at the reorganized CSK in the late stage of platelet aggregation (platelet-to-platelet aggregation). However, a recent report has shown that the ␣ IIb ␤ 3mediated signaling regulated by protein tyrosine phosphorylation participates in formation of fibrin polymers (15). In fact, we also have observed the inhibition of the thrombin-induced clot retraction in genistein-treated platelets (data not shown).

DISCUSSION
Several lines of evidence indicate that integrins are involved in bi-directional signaling across the plasma membrane (1). In platelets, the agonist stimulation causes activation of ␣ IIb ␤ 3 integrin by unique signal transduction (inside-out signaling), and the subsequent fibrinogen binding to the ␣ IIb ␤ 3 integrin, leading to cytoskeletal rearrangement and platelet aggregation, causes tyrosine kinase activation including pp125 FAK and pp60 c-src (outside-in signaling) (1,3,4,27). The ability of ␣ IIb ␤ 3 to undergo cytoskeletal rearrangement is blocked by the monoclonal antibody to integrin ␤ 3 or the fibrinogen antagonist, RGDS (27). Furthermore, recent study has shown that protein tyrosine kinase inhibitors, genistein or herbimycin A, and actin polymerization inhibitor, cytochalasin D, prevent the association of various signaling molecules with integrin (26). These observations suggest that the integrin ␣ IIb ␤ 3 is anchored to underlying cytoskeletal molecules and that the association of ␣ IIb ␤ 3 with cytoskeletal proteins is regulated by activation of protein tyrosine kinases.
In the present study, we have demonstrated that in thrombin-stimulated human platelets, the PLC␤3(a, b) were preferentially relocated to the detergent-insoluble actin-rich CSK (Figs. 1 and 2). This translocation was dependent on the integrin ␣ IIb ␤ 3 activation (Fig. 4 and Table I). Furthermore, it is inhibited by either cytochalasin D or genistein (Figs. 5 and 6), suggesting that the translocation of PLC␤3(a, b) to CSK is mediated by formation of microfilament-dependent complexes containing ␣ IIb ␤ 3 protein tyrosine kinase and their substrates in aggregated platelets.
The decreases in the PLC␤3(a, b) level were observed in CSK after the peak at 30 s of thrombin stimulation (Fig. 2, right  panels). In our earlier report it was indicated that PLC␤3 is one of the substrates for calpain in thrombin-stimulated human platelets (21). Thus it was conceivable that the decreased PLC␤3 level reflects its cleavage by calpain. In fact, calpeptin, a calpain inhibitor, blocked the cleavage of PLC␤3(a, b), although the inhibitor did not inhibit their translocation to CSK mediated by ␣ IIb ␤ 3 activation. In addition to the PLC␤3, some other proteins also are cleaved by calpain in the late aggregation phase, and their products are located in CSK (28,29). The calpain activation has been known to be dependent on ␣ IIb ␤ 3ligand interaction (7). The fibrinogen antagonist RGDS or antibody mAb-IN␤ 3 could prevent the cleavage of PLC␤3 (data not shown). Therefore, it can be assumed that the cleavage by calpain of PLC␤3(a, b) may be associated with the integrin activation.
We have previously shown in in vitro experiments that the cleavage of purified PLC␤3a by calpain produced a 100 kDa PLC and that this truncated PLC was, to a greater extent, activated by G␤␥ subunits than its original forms (30). The present study using intact platelets indicated that heterotrimeric G proteins (G i2 and G q ) involving PLC␤3 activation were translocated to the reorganized CSK (Fig. 6). The location of the heterotrimeric G protein in the cytoskeleton was described in a previous report that demonstrated that G␥ 5 was co-localized with vinculin in association with focal adhesions (31). Moreover, it was recently shown that G q ␣/G 11 ␣ was recovered in the actin-rich fraction in agonist-stimulated rat mammary tumor (WRK) cells and that cytochalasin D inhibited G proteininduced PLC activation, suggesting an important role of the actin-rich cytoskeleton in the G protein-regulated PLC activation (32). Accordingly, PLC␤3(a, b) translocated to CSK was assumed to be activated by the G proteins in thrombin-stimulated platelets. There was DAG production with a peak at 2 min after stimulation. This late DAG formation was reduced by genistein pretreatment which blocked the translocation of PLC␤3 and G q (Figs. 6 and 7). Our previous study has suggested that thrombin stimulates a biphasic accumulation of DAG, with an early phase reaching a peak at 10 s and a later phase at 2-3 min, and that the second phase of DAG accumulation may be derived from PIs in human platelets (23). These results lead us to speculate that the integrin-ligand interaction induces translocation to the actin-rich CSK of PLC␤3(a, b) and G q , resulting in DAG accumulation by activated PLC␤3 in the late aggregation step.
Several lines of evidence suggest that protein tyrosine kinases such as pp60 c-src are required for the cytoskeletal attachment of integrin ␣ IIb ␤ 3 (9,26,33) and also for the retraction of fibrin polymers in thrombin-stimulated platelets (15). We have demonstrated here that the pretreatment of platelets with genistein prevented the CSK association of various signaling molecules, such as PLC␤3(a, b), pp60 c-src , G q , PKC␣, and ␣ IIb ␤ 3 , thus leading to inhibition of the PLC activation. These findings have led us to postulate that the translocation of PLC␤3 to CSK which was controlled by protein tyrosine kinase may play a role in the late aggregation (formation of large platelet-to-platelet aggregation).
It has been known that platelets have at least two adhesive glycoprotein receptors, integrin ␣ 2 ␤ 1 (GPIa-IIa) and ␣ IIb ␤ 3 (GPIIb-IIIa), which mediate transmembrane signaling (3-5, 35). Platelet ␣ 2 ␤ 1 acting as a receptor for collagen mediates activation of tyrosine kinases including pp125 FAK (36,37). The ␤1 or ␤2 family of integrins induces activation of the tyrosine kinase PLC␥1 pathway (37,38). Recently, PLC␥2, but not ␥1, was shown to be activated via tyrosine phosphorylation in platelets stimulated by collagen but not by thrombin (39,40). Our data obtained in this study demonstrated that in the cytoskeletal association of thrombin-stimulated platelets, PLC␥2 is distinct from PLC␤3 in that PLC␤3 is entirely integrin ␣ IIb ␤ 3 -dependent and inhibited by cytochalasin D. Thus, PLC␥2 appears to play a role in the transmembrane signaling initiated through ␣ 2 ␤ 1 rather than ␣ IIb ␤ 3 in platelet aggregation. In addition, PLC␤2, which is most abundant in platelets, also becomes associated with CSK by aggregation with thrombin, but its association was independent of ␣ IIb ␤ 3 and not inhibited by cytochalasin D and genistein, suggesting that PLC␤2 participates in the early stage of agonist-induced signaling but not in the integrin-mediated signaling pathway in human platelets. However, the activation mechanisms and exact roles of different PLC isozymes in these two pathways remain to be disclosed.