Rho and Rho-kinase mediate thrombin-induced phosphatidylinositol 4-phosphate 5-kinase trafficking in platelets.

Phosphatidylinositol 4-phosphate 5-kinase (PIP5K) catalyzes the rate-limiting step in the production of phosphatidylinositol 4,5-bisphosphate (PIP(2)), a signaling phospholipid that contributes to actin dynamics. We have shown in transfected tissue culture cells that PIP5K translocates from the cytosol to the plasma membrane following agonist-induced stimulation of Rho family GTPases. Nonetheless, it is unclear whether Rho GTPases induce PIP5K relocalization in platelets. We used PIP5K isoform-specific immunoblotting and lipid kinase assays to examine the intracellular localization of PIP5K in resting and activated platelets. Using differential centrifugation to separate the membrane skeleton, actin filaments and associated proteins, and cytoplasmic fractions, we found that PIP5K isoforms were translocated from cytosol to actin-rich fractions following stimulation of the thrombin receptor. PIP5K translocation was detectable within 30 s of stimulation and was complete by 2-5 min. This agonist-induced relocalization and activation of PIP5K was inhibited by 8-(4-parachlorophenylthio)-cAMP, a cAMP analogue that inhibits Rho and Rac. In contrast, 8-(4-parachlorophenylthio)-cGMP, a cGMP analogue that inhibits Rac but not Rho, did not affect PIP5K translocation and activation. This suggests that Rho GTPase may be an essential regulator of PIP5K in platelets. Consistent with this hypothesis, we found that C3 exotoxin (a Rho-specific inhibitor) and HA1077 (an inhibitor of the Rho effector, Rho-kinase) also eliminated PIP5K activation and trafficking into the membrane cytoskeleton. Thus, these data indicate that Rho GTPase and its effector Rho-kinase have an intimate relationship with the trafficking and activation of platelet PIP5K. Moreover, these data suggest that relocalization of platelet PIP5K following agonist stimulation may play an important role in regulating the assembly of the platelet cytoskeleton.

Actin assembly and vesicle trafficking are controlled by PIP 2 . Some of these actin signaling pathways depend on intact PIP 2 rather than on the products of its hydrolysis (11,12). This is thought to be due to the binding and displacement of various actin regulatory proteins from actin filaments by PIP 2 , thus allowing the polymerization of these filaments (13). Consistent with an important role for PIP 2 in the organization of the actin cytoskeleton, overexpression of PIP5K has been shown to modulate actin cytoskeletal dynamics and to induce stress fibers (14) (15), membrane ruffles (16), microvilli (17), and motile actin comets (18).
The relationship between PIP5K isoforms and actin is complex but appears to be regulated by small GTPases of the Rho family. A physical association between RhoA and PIP5K has been reported (19), and the addition of recombinant RhoA stimulates PIP5K activity in lysates of mouse fibroblasts (20,21). Consistent with the role of Rho in PIP5K activation, inactivation of Rho GTPases by Clostridium difficile toxin B reduces cellular PIP 2 levels by 50 -90% in several tissue culture cell lines (22,23). This results in an inhibition of receptormediated inositol phosphate formation by phospholipase C and PIP 2 -sensitive phospholipase D and also induces profound changes in cell morphology.
In seeming contrast to these results, Tolias et al. (24) demonstrated a link between PIP5K and Rac1 but not with RhoA or Cdc42. In permeabilized platelets, a constitutively active Rac mutant induced PIP 2 synthesis and uncapped actin filaments by a PIP 2 -dependent mechanism (18,25). In addition, Rozelle et al. (18) have demonstrated that overexpression of PIP5K in tissue culture cells generates actin comets, decreases stress fibers, and suppresses membrane ruffling through pathways that may be mediated by Cdc42. Recently, Oude Weernink et al. (26) showed that all three PIP5K isoforms are positively regulated by expression of RhoA, Rac1, and Cdc42 in HEK-293 cells, and this results in enhanced cellular PIP 2 levels. Lastly, Honda et al. (16) have suggested that activation of another small GTPase, Arf6, is actually the critical step in PIP5K activation. Thus, there is evidence that PIP5K activity may be regulated by Rho, Rac, or Arf6.
Although reports about the association of PIP5K with small GTPases have been controversial, there is agreement that small GTPases regulate the activity of PIP5K isoforms, and this contributes to actin dynamics (14,27,28). The apparent discrepancies about which GTPase is critical for PIP5K activation might be attributable to cell type-specific differences. In the present study, we used intact human platelets to investigate the mechanism by which small GTPases regulate platelet PIP 2 formation by PIP5K. We demonstrate that platelet PIP5K is mostly regulated by Rho and its effector Rho-kinase. We also show that the activated PIP5K translocates to the platelet actin cytoskeleton, where it is available to initiate actin assembly.

EXPERIMENTAL PROCEDURES
Materials-The [␥-32 P]phosphate was obtained from Amersham Biosciences. The Rho-kinase inhibitor, HA-1077, was from Upstate Biotechnology (Lake Placid, NY). The hirudin (2000 ATU) was obtained from Roche Diagnostics. The anti-RhoA and anti-Rac antibodies and GST fusion proteins were purchased from Cytoskeleton, Inc. (Denver, CO). Triton X-100 (Surfact-Amps X-100) and Brij 58 (Surfact-Amps 58) were from Pierce. The two rabbit polyclonal antisera directed against PIP5K isoforms were further purified using a protein A affinity column (29). All other reagents were obtained from Sigma-Aldrich.
Recombinant botulinum C3 exoenzyme (C3) was expressed in bacteria as a glutathione S-transferase fusion protein. Following lysis of the bacteria, the C3 protein was affinity purified with glutathione-Sepharose 4B beads. The C3 exoenzyme was cleaved from GST by thrombin (80 units/liter culture) overnight at 4°C. The thrombin was subsequently removed with benzamidine-Sepharose 6B. The purified C3 protein was resuspended in 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM MgCl 2 , and 0.1 mM dithiothreitol.
Platelet Preparation-Blood was drawn from healthy volunteers and collected into 0.1 volume of 0.15 M sodium citrate. Platelet-rich plasma was obtained by centrifugation at 200 ϫ g for 10 min and adjusted to pH 6.5 with 0.15 M citric acid to prevent platelet activation during further isolation. After a second centrifugation at 800 ϫ g for 10 min, the pellet was washed twice with ACD (74.8 mM sodium citrate, 41.6 mM citric acid, 138.8 mM dextrose, and 145 mM NaCl). The washed platelet pellet was then resuspended in HEPES-Tyrode's buffer (20 mM HEPES, 138 mM NaCl, 0.36 mM NaH 2 PO 4 , 2.9 mM KCl, 12 mM NaHCO 3 , 5.5 mM glucose, 0.4 mM MgCl 2 , pH 7.4) at 1 ϫ 10 9 platelets/ml.
Platelet Activation-Washed platelets (0.5 ml) were incubated with stirring at 37°C in a lumi-aggregometer (Frenius, Bad Homburg, Germany). CaCl 2 (1 mM) was added prior to the addition of 50 -60 M TRAP. As indicated, in some experiments the platelets were preincubated, prior to addition of TRAP, with HA1077 for 1 h, C3 exotoxin for 2 h, or analogues of either cAMP or cGMP for 10 min.
Cell Fractionation and Translocation Assay-Platelets were lysed by the addition of an equal volume of ice-cold 2ϫ Triton lysis buffer containing 100 mM Tris-HCl, pH 7.4, 10 mM EGTA, 2 g/ml leupeptin and pepstatin, 1 g/ml aprotinin, 2 mM phenylmethylsulfonyl fluoride, 2 mM Na3VO4, and 2% Triton X-100 and kept on ice for 30 min. Platelet lysates were fractionated by established methods described previously (34,35) including differential centrifugation to obtain the 15,600 ϫ g fraction (low speed cytoskeleton), the 100,000 ϫ g fraction (membrane skeleton), and the 100,000 ϫ g supernatant fraction containing soluble cytosolic proteins.
Immunoblotting-Samples were reduced in Laemmli sample buffer. After boiling for 3 min, proteins were resolved on SDS polyacrylamide gel and then transferred to polyvinylidene difluoride. The membrane was blocked overnight at 4°C in Tris-buffered saline containing 5% (w/v) bovine serum albumin. The proteins were detected by blotting with the appropriate monoclonal or polyclonal antibodies in Tris-buffered saline, 0.02% Tween, 2% bovine serum albumin followed by incubation with either anti-mouse or anti-rabbit IgG antibody coupled to horseradish peroxidase. Detection was achieved using a chemiluminescent probe (Amersham Biosciences). Protein concentrations were determined using the micro BCA protein assay (Pierce) with bovine serum albumin as standard.
Lipid Raft Preparation-Lipid rafts were isolated by density through a step gradient of 0 -40% (w/v) sucrose as described previously (37)(38)(39). Briefly, resting or stimulated platelets (2 ϫ 10 8 , in 100 l of HEPES-Tyrode's buffer) were lysed in 200 l of ice-cold buffer (20 mM Tris, 150 mM NaCl, 10 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 2 mM Na 3 VO 4 , 10 g/ml leupeptin, 10 g/ml aprotinin, 1 g/ml pepstatin A, pH 8.3) containing 1% Brij 58. The same buffer was used for the sucrose gradient. The lysate was mixed with an equal volume of 80% (w/v) sucrose, giving a final concentration of 40% sucrose. The lysate (600 l) was loaded on the bottom of an Ultracentrifuge tube and overlaid with 1 ml of 25% sucrose and 500 l of lysis buffer, both containing the indicated detergent concentrations. The samples were centrifuged at 200,000 ϫ g for 2.5 h at 4°C. Seven fractions, each 300 l, were sequentially removed from the top of the gradient and added to an equal volume of Laemmli buffer. For immunoprecipitations, the visible lightscattering band (corresponding to fractions 2 and 3, the "rafts fractions") was recovered in 300 l. An equal volume of fractions 6 and 7 was used for immunoprecipitation of the "soluble fractions." Rho and Rac Pull-down Assay-Measurement of activation of RhoA and Rac were performed using commercially available kits (BK036 and PAK02) manufactured by Cytoskeleton, Inc. Briefly, cleared platelet lysates were incubated for 1 h at 4°C with GST-PAK or GST-Rhotekin already bound to glutathione-Sepharose beads to precipitate GTPbound Rac and Rho, respectively. Precipitated complexes were washed three times in lysis buffer and boiled in sample buffer. Total lysates and precipitates were analyzed on Western blots using antibodies against Rac and RhoA.

RESULTS
Comparison of Relative Abundance of PIP5K␣, PIP5K␤, and PIP5K␥ in Platelets-The expression of different PIP5K isoforms in platelets has not been described previously. Using polyclonal antibodies that are specific for PIP5K␣ (ϳ60 kDa), or both PIP5K␤ (ϳ60 kDa) and PIP5K␥ (ϳ90 kDa), we examined the presence of each isoform in human platelets (Fig. 1). We found all three isoforms to be present. PIP5K␥ is known to have alternative splice sites that produce an 87-kDa and a 90-kDa (so-called "brain-specific") variant (4). Because the 90-kDa splice variant has been shown to contribute to focal adhesion formation in tissue culture cells, the 90-kDa PIP5K␥ might play an important role in platelet actin dynamics.

Time Course of PIP5K Redistribution in Human Platelets-
Consistent with studies of most cells, total cellular PIP 2 in platelets does not rise dramatically in response to agonist stimulation of PIP5K (31,32). This suggests that local increases in the PIP 2 concentration may occur at sites of recruited lipid kinases (1). We have demonstrated in transfected COS-7SH cells that stimulation of the PAR1 thrombin receptor induces a translocation of PIP5K␣ from the cytosol toward the plasma membrane (33). In these cells, the translocation of PIP5K␣ was dependent on both Rac and Rho, although the activation of Rho was essential for the trafficking process. Because the regulation of PIP5K appears to vary in different types of cells, we examined the distribution of the PIP5K isoforms in platelets.
To determine the subcellular location of PIP5K in resting and activated platelets, these cells were fractionated by established Triton X-100 lysis protocols (34,35) followed by differential centrifugation. This allowed us to obtain the 15,600 ϫ g fraction ("low speed cytoskeleton fraction" containing actin filaments and associated proteins), 100,000 ϫ g fraction (containing the membrane skeleton), and the 100,000 ϫ g supernatant fraction (containing soluble cytosolic proteins). In resting platelets, PIP5K␤ (Fig. 2) was predominantly present in the cytoplasmic fraction. Gently stirred platelets were stimulated with the thrombin receptor-activating peptide (TRAP). TRAP stimulation led to rapid translocation of PIP5K␤ (Fig. 2) to the low speed cytoskeleton fraction. A similar result was found for the other platelet PIP5K isoforms, PIP5K␥ and PIP5K␣ (not shown). This trafficking event was seen within 30 s and reached a peak within 5 min. This suggests that similar to our observation in tissue culture cells, PIP5K isoforms relocalize within platelets following agonist stimulation.
PIP5K Does Not Localize within Platelet Lipid Rafts-Cholesterol-sphingolipid-rich membrane microdomains (GEMs or lipid rafts) are sites of active phosphoinositide and tyrosine kinase signaling. Lipid rafts are found in platelets (36,37), although their functional role is just beginning to be elucidated (18,38). Rozelle et al. (18) have shown that lipid rafts in Ref52 cells are the preferred platforms for membrane-linked actin polymerization mediated by in situ PIP 2 synthesis. Therefore, we investigated whether the PIP 2 synthesis by PIP5K occurred in platelet lipid rafts and whether the agonist-induced relocalization of PIP5K involved these microdomains.
To define conditions for isolation of platelet rafts, we used CD36, Lyn, and GM1 ganglioside as markers of this microdomain (Fig. 3A) (36,39). The soluble fractions containing nonraft-associated membrane proteins and cytosolic proteins were identified by blotting for cytosolic p42/44 ERK (extracellular signal-regulated kinase) (not shown) (38). In the cell fractions derived from either resting or TRAP-stimulated platelets, we tested for PIP5K. As shown in Fig. 3B (left panel), we were unable to detect any PIP5K␤ protein by immunoblotting with our polyclonal antibody. Similarly, we were also unable to demonstrate any PIP5K enzymatic activity in raft fractions (Fig. 3B, right panel). Although PIP5K does relocalize to dif-ferent subcellular domains following stimulation of the platelet thrombin receptor, our data indicate that platelet PIP5K does not traffic into lipid rafts.
Rac Is Not Required for Trafficking or Activation of PIP5K in Human Platelets-Reports from various cell types have suggested that PIP5K is a downstream target of Rho proteins (19,26,40). We have previously shown that these GTPases regulate the intracellular localization and catalytic activity of PIP5K␣ in Cos7SH cells (33). Offermanns and colleagues (41) demonstrated that cyclic nucleotide analogues of cAMP and cGMP differentially interfere with thromboxane A 2 -induced Rac and Rho activation. Thus, we examined whether cAMP and cGMP, via their effect on Rac and Rho, modulated the trafficking of PIP5K in intact platelets.
For the initial experiments, we tested whether analogues of cAMP and cGMP would modulate the thrombin receptor-mediated activation of Rac and Rho in addition to their published effect on thromboxane A 2 receptor signaling. Washed platelets were incubated with TRAP for 5 min, and activation of Rac and Rho was analyzed using pull-down assays with fusion proteins composed of GST with either a Rac-binding domain or a Rhobinding domain. Incubation of human platelets with TRAP caused a rapid activation of both RhoA and Rac (Fig. 4). Because our capture assay recognizes both Rac1 and Rac2, we could not discern whether there is preferential activation of either Rac isoform. In the presence of the cAMP analogue, there was little activation of either Rac or Rho after stimulation of the thrombin receptor. However, in the presence of cGMP, TRAP still induced activation of RhoA, but activation of Rac was almost completely inhibited.
We next tested whether the analogues of cAMP and cGMP affected the redistribution of PIP5K. Stimulation of platelets with TRAP induced the translocation of PIP5K, RhoA, and Rac into the low speed actin filament-rich fraction (Fig. 5). A dose of cAMP that simultaneously inhibited the translocation of both Rac and Rho also completely eliminated the trafficking and activation of PIP5K␤ (Fig. 5). In contrast, preincubation of platelets with the cGMP analogue inhibited Rac but had minimal effects on the translocation or activation of either Rho or PIP5K. Recent   FIG. 1. Expression of PIP5K isoforms within platelets. Equal quantities of total cell lysates of human platelets were fractionated by 10% SDS-PAGE. Immunoblotting was performed using polyclonal antibodies against either PIP5K␣ or against both PIP5K␤ and PIP5K␥. reports have demonstrated nonspecific effects of cAMP and cGMP analogues (42,43). However, these reagents do demonstrate that the activation or translocation of Rac is not required for the thrombin receptor-mediated translocation or activation of PIP5K. These data suggested that activation of RhoA might be essential for the PIP5K␤ translocation in platelets.
Rho and Rho-kinase Are Required for the Relocalization and Activation of PIP5K-Clostridium toxin 3 (C3 exotoxin) has been shown to specifically inactivate RhoA by ADP ribosylation (44,45). We tested whether ADP ribosylation of platelet Rho by C3 exotoxin affected PIP5K trafficking or activation. Incubation of intact platelets for 2 h with C3 exotoxin resulted in ADP ribosylation of greater than 90% of RhoA, as determined by the limited ability of additional C3 exotoxin to [ 32 P]ADP-ribosylate Rho in lysates of these platelets (Fig. 6A).
Next, to determine whether C3 exotoxin affected the trafficking of PIP5K, platelets were incubated with C3 exotoxin before the stimulation, and the low speed actin filament-rich fraction was isolated. This fraction was analyzed for PIP5K␤ protein (Fig. 6B). Preincubation of intact platelets with C3 exotoxin caused an ϳ60% reduction of agonist-stimulated translocation of PIP5K␤ (Fig. 6B) into the actin filament-rich fraction. This observation supports our data using cyclic nucleotide analogues and demonstrates that the regulation of PIP5K is closely linked to RhoA.
We also tested whether the activation of PIP5K␤ in total (unfractionated) platelet lysates was affected by pharmacologic inhibition of RhoA. Lysates derived from platelets stimulated by their thrombin receptor had 2.5-fold higher PIP5K activity compared with lysates derived from resting platelets (Fig. 7). However, this PIP5K activation was completely eliminated by cAMP or C3 exotoxin. This demonstrates that activation of RhoA is critical for the regulation of PIP5K activity in platelets.
Many of the effects of RhoA are mediated by Rho-kinase (also designated as ROCK or Rho-associated kinase). In addition, the Rho-mediated regulation of PIP5K has been speculated to require Rho-kinase (15,28,46). Therefore, we tested whether the Rho-kinase inhibitor, HA1077, affected PIP5K trafficking or activity. This Rho-kinase inhibitor reduced PIP5K translocation and activation (Figs. 6B and 7). This suggests that the translocation of active PIP5K toward the actin cytoskeleton is regulated by both Rho and its effector, Rho-kinase. DISCUSSION The goal of this study was to investigate the signaling pathway within primary platelets initiated by stimulation of the thrombin receptor and leading to the activation of PIP5K and production of PIP 2 . These observations extend our previous data on the subcellular localization of PIP5K␣ expressed in tissue culture cell lines (33). In that report, stimulation of the thrombin receptor induced the translocation of PIP5K␣ near the plasma cell membrane through a pathway dependent on Rho. Our current findings demonstrate a pathway for activation of PIP5K in platelets that is dependent on both Rho and Rho-kinase. This signaling pathway simultaneously results in the trafficking and enzymatic activation of PIP5K. We hypothesized that both relocalization and biochemical activation of PIP5K are critical for its ability to regulate actin dynamics within platelets, a process critical for adhesion under shear conditions of the arterial system.
Publications by Divencha and colleagues (47,48) previously showed that the "Type C" isoforms of PIP5K migrated to the platelet membrane cytoskeleton after thrombin stimulation. Since that report, it is now accepted that PIP5K C is actually a phosphatidylinositol 5-phosphate 4-kinase (PIP4K) (1). This enzyme, which phosphorylates phosphatidylinositol 5-phosphate, may be involved in an alternative PIP 2 synthesis pathway. It is intriguing that PIP5K and PIP4K, both enzymes  2 and 3 in A) and non-lipid raft fractions (lanes 6 and 7 in A) were analyzed for PIP5K␤ by immunoblotting (left panel) or by an in vitro lipid kinase assay using PI4P as the exogenous substrate (right panel). This demonstrates that PIP5K␤ protein and PIP5K activity do not redistribute into lipid rafts following platelet activation.

FIG. 4. Cyclic nucleotides affect thrombin receptor-induced activation of Rac and Rho GTPases.
Washed human platelets (1 ϫ 10 9 /ml) were preincubated for 10 min at 37°C with or without preincubation of 1 mM 8-pCPT-cAMP or 8-pCPT-cGMP before the addition of 60 M for 5 min. The platelets were lysed, and GTP-loaded Rho or Rac were affinity purified from a portion of the lysates using a "pull-down" assay with either a GST-Rac-binding protein or GST-Rho-binding protein. Shown are anti-Rac and anti-Rho immunoblots of total cell lysates or affinity purified GTP-loaded proteins. This shows that 8-pCPT-cAMP inhibits TRAP-induced activation of Rac and Rho, whereas 8-pCPT-cGMP inhibits the activation of Rac alone. speculated to generate platelet PIP 2 , may compartmentalize to a similar location within platelets.
Schwartz and colleagues (20) were the first to demonstrate that PIP5K is regulated by small GTP-binding proteins. They found that PIP5K␣ catalytic activity in cell lysates increased in the presence of GTP-bound Rho but not in the presence of either GDP-bound Rho or GTP-bound Rac. Since that publication, various reports have presented evidence that Rac, CDC42, or Arf6 may be the small GTPases responsible for activation of PIP5K in various cell types (16, 18, 24 -26). Our results show that inhibition of Rho completely blocks trafficking and activation of PIP5K in thrombin receptor-stimulated platelets. Although activation of Rac is clearly not required for these two processes in platelets following stimulation of the thrombin receptor, our data do not exclude the possibility that Rac contributes to the regulation of platelet PIP5K under certain circumstances. It is also likely that different cell types utilize different GTPases for the activation and perhaps trafficking of PIP5K.
It is unclear how GTP-bound Rho recruits PIP5K to the platelet membrane cytoskeleton. Several investigators have demonstrated that Rho family members shuttle on and off the membrane; and this membrane localization is regulated by their GTP/GDP-bound state (49 -51). Because Rho is constitutively bound to PIP5K, one could speculate that GTP loading of Rho merely serves to help translocate the complex to the cell membrane, the predominant location of its lipid substrate, PI4P. However, our results demonstrate that Rho regulates the total platelet PIP5K activity in platelets. This demonstrates that Rho does more than chaperone PIP5K and must also regulate the specific activity of PIP5K within platelets.
The association of Rho with PIP5K is guanine nucleotideindependent, yet only the GTP-bound form of this GTPase can activate PIP5K. This suggests a model whereby a Rho effector, once activated by GTP-Rho, is also required for the regulation of PIP5K (28). We found that Rho-kinase may be the direct mediator of this effect in platelets. Consistent with our studies, Rho-kinase also regulates PIP5K in several tissue culture cell lines (15,28,46). It is noteworthy that mutations within the catalytic domain of PIP5K block its intracellular trafficking in FIG. 7. Inhibition of either Rho or Rho-kinase impairs total platelet PIP5K activity. Washed platelets were incubated with 1 mM 8-pCPT-cAMP, 1 mM 8-pCPT-cGMP, HA1077, C3 exotoxin, or carrier control. A portion of the platelets was stimulated with TRAP for 5 min prior to lysis. In vitro kinase activity of the total cell lysates were analyzed using PI4P as the exogenous substrate. The mean Ϯ S.E. are derived from at least three independent experiments.
FIG. 5. Activation of Rac is not required for PIP5K trafficking. Washed platelets were incubated with 1 mM 8-pCPT-cAMP or 8-pCPT-cGMP before 5 min of stimulation by TRAP. The cells were lysed, and the actin filament-rich fraction (low speed pellet) was isolated by differential centrifugation. This fraction was divided into two parts for analysis by immunoblotting with PIP5K, Rho, and Rac antibodies to test for translocation of these proteins (A) or in vitro lipid kinase assay using PI4P as the substrate (B). These results indicate that inhibition of Rac does not prevent PIP5K activation or translocation to the membrane cytoskeleton.
FIG. 6. Effect of Rho or Rho-kinase inhibitors on the trafficking of PIP5K into actin filament-rich fraction of platelet lysates. A, ADP ribosylation of Rho in human platelets by C3 exotoxin. Washed human platelets were preincubated in the absence or presence of 400 g/ml C3 exotoxin before TRAP stimulation. The platelets were lysed and subjected to [␣-32 P]ADP ribosylation assay as described under "Experimental Procedures." This demonstrates that more than 90% of Rho in platelets was ADP-ribosylated. B, effects of inactivation of RhoA and RhoA kinase on TRAP-induced translocation of platelets PIP5K. Washed platelets were incubated with 1 mM 8-pCPT-cAMP or 8-pCPT-cGMP, with 1 mM HA1077, or with 400 g/ml C3 exotoxin at 37°C before TRAP stimulation. Platelets were then lysed and centrifuged to isolate the actin filament-rich fraction (low speed pellet) for translocation assay. The actin filament-rich fraction was analyzed by anti-PIP5K␤ immunoblotting.
Chinese hamster ovary cells. 2 This suggests that, in addition to Rho-kinase, a PIP 2 -regulated protein may also be required for relocalization of PIP5K. The identity of this PIP 2 -dependent protein is currently unknown.
Overexpression of PIP5K does not change total cellular concentrations of PIP 2 . Yet, wild type PIP5K overexpression induces actin reorganization through a pathway that requires a functional lipid kinase domain. This suggests that PIP 2 production within discrete cellular subcompartments contributes to actin dynamics. Our study demonstrates that PIP5K protein and PIP 2 production are not preferentially localized with lipid rafts. This is consistent with the previous report that failed to show PIP 2 production within this platelet microdomain (52). However, the results of these platelet experiments are in contrast with the report demonstrating that PIP 2 signaling may be preferentially localized to rafts within Ref52 cells (18). Recently, it has been shown in B lymphocytes that PIP5K is recruited into a microdomain that contains the tyrosine kinase, Btk (29). It is currently unknown whether localized high concentrations of PIP 2 accumulate within a similar microdomain within platelets. It is tempting to speculate that localized production of PIP 2 on the platelet membrane contributes to actin dynamics by binding and regulating profilin or actin-capping proteins such as gelsolin and capZ.
In conclusion, our results demonstrate that platelet PIP5K is tightly regulated by Rho GTPase. Once Rho and its effector, Rho-kinase, become stimulated, they induce the activation and membrane recruitment of PIP5K to generate PIP 2 . We believe that this leads to the localized production of PIP 2 that regulates actin-binding proteins and contributes to platelet actin dynamics.