Differential Regulation of Rho and Rac through Heterotrimeric G-proteins and Cyclic Nucleotides*

Platelets were used to study the activation of Rho and Rac through G-protein-coupled receptors and its regulation by cyclic nucleotides. The thromboxane A2(TXA2) mimetic U46619 rapidly activated both small GTPases independently of integrin αIIbβ3activation. U46619, which leads to the activation of G12/G13 and Gq did not induce Rac activation in Gαq-deficient platelets but was able to activate Rho, to stimulate actin polymerization and phosphatidylinositol 4,5-bisphosphate formation, and to induce shape change. Rac activation by U46619 in wild-type platelets could be blocked by chelation of intracellular Ca2+ and was partially sensitive to apyrase and AR-C69931MX, an antagonist of the Gi-coupled ADP receptor. Cyclic AMP, which completely blocks platelet function, inhibited the U46619-induced activation of Gq and G12/G13 as well as of Rac and Rho. In contrast, cGMP, which has no effect on platelet shape change blocked only activation of Gq and Rac. These data demonstrate that Rho and Rac are differentially regulated through heterotrimeric G-proteins. The G12/G13-mediated Rho activation is involved in the shape change response, whereas Rac is activated through Gq and is not required for shape change. Cyclic AMP and cGMP differentially interfere with U46619-induced Rho and Rac activation at least in part by selective effects on the regulation of individual G-proteins through the TXA2receptor.

The small GTPases Rho and Rac are central regulators of various cellular processes such as actin cytoskeleton dynamics, transcriptional regulation, cell cycle progression, and contractile processes (1). They are activated by a variety of receptors, including those coupled to heterotrimeric G-proteins (2). Various heterotrimeric G-proteins have been involved in linking receptors to the regulation of Rho and Rac. The ␣-subunits of the G 12 family of heterotrimeric G-proteins, G␣ 12 and G␣ 13 , are able to activate Rho (3,4). Rho activation may be mediated by a group of Rho-specific guanine nucleotide exchange factors, which are able to interact with G␣ 12 /G␣ 13 (5)(6)(7). However, in some cells G␣ q or ␤␥ complexes of heterotrimeric G-proteins have been suggested to induce Rho activation (8 -10). G-pro-tein-coupled receptor-mediated activation of Rac has been shown to be mediated by G i -type G-proteins via a mechanism, which in some cases appears to involve G-protein ␤␥ subunits and activation of phosphoinositide 3-kinase (11)(12)(13)(14).
Both Rho and Rac have been involved in early signaling processes underlying platelet activation. Platelets respond to various stimuli, which function through G-protein-coupled receptors with secretion of granule contents, aggregation, and a rapid change of their shape. Rac, which is rapidly activated after activation of the thrombin receptor PAR-1 (15), has been suggested to mediate thrombin receptor-induced actin assembly via stimulation of phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P 2 ) 1 production during platelets activation (16,17), and evidence has been provided that Rho is involved in early processes underlying platelet activation by linking receptors to Rho-kinase and subsequent regulation of myosin light-chain phosphorylation (18 -20). The mechanisms by which Rho and Rac become activated through G-protein-coupled receptors in platelets are unclear. Receptors activated by thrombin or thromboxane A 2 (TXA 2 ) couple to G q and G 12 /G 13 . In addition, thrombin but not TXA 2 is able to induce activation of G i in platelet membrane fractions (18,21). Platelets from G␣ q -deficient mice have been instructive in delineating the initial mechanisms of platelet activation. TXA 2 and thrombin are unable to induce phospholipase C activation as well as platelet aggregation and secretion in the absence of G␣ q , whereas induction of platelet shape change through the activation of G 12 /G 13 appears to be basically unaffected (18,22).
In this report we demonstrate by direct determination of Rho and Rac activation that both small GTPases become rapidly activated by TXA 2 in an integrin ␣ IIb ␤ 3 -independent manner. We used platelets from G␣ q -deficient mice to show that Rho and Rac are differentially regulated. Rho activation occurs through G 12 /G 13 and is involved in processes underlying platelet shape change, including PtdIns(4,5)P 2 formation. In contrast, Rac activation is mediated by G q and is not required for shape change. Cyclic nucleotides cAMP and cGMP differentially interfere with Rac and Rho activation in platelets at least in part by affecting receptor-mediated activation of G q and G 12 /G 13 .
Platelet Preparation and Aggregation-Whole blood was collected from normal and G␣ q -deficient mice (129/Sv ϫ C57BL/6) anesthetized with pentobarbital by puncturing the inferior vena cava with syringes containing acid citrate dextrose (1/9 volume). The blood from 3-4 G␣ qdeficient mice and wild-type mice was pooled for each platelet experiment. Blood was diluted with half the volume of Hepes-Tyrode's buffer (137 mM NaCl, 2 mM KCl, 12 mM NaHCO 3 , 0.3 mM NaH 2 PO 4 , 2 mM CaCl 2 , 1 mM MgCl 2 , 5.5 mM glucose, 5 mM Hepes, pH 7.3) containing 0.35% human serum albumin, and platelet-rich plasma was obtained by centrifugation for 8 min at 250 ϫ g at 37°C. Thereafter, prostacyclin at a final concentration of 500 nM was added to the platelet-rich plasma, and platelets were pelleted twice by centrifugation at 1000 ϫ g for 5 min at 37°C. The platelet pellet was resuspended in Hepes-Tyrode's buffer at a density of 1 ϫ 10 9 platelets per milliliter in the presence of 0.02 unit/ml of the ADP scavenger apyrase (adenosine-5Ј-triphosphate diphosphohydrolase) and incubated for 30 min at 37°C.
For inositol lipid analysis, platelets were labeled with 0.5 mCi/ml [ 32 P]orthophosphate during 50 min in a phosphate-free washing buffer (pH 6.5) at 37°C. 32 P-Labeled platelets were then washed once in the same buffer and finally suspended at a final concentration of 1 ϫ 10 9 cells/ml (pH 7.3). Optical aggregation experiments were conducted in a two-channel aggregometer (Chronolog).
Determination of Activated Cellular Rho and Rac-The amount of activated cellular Rho and Rac was determined by precipitation with a fusion protein consisting of GST and the Rho-binding domain of Rhotekin (GST-RBD) or the Rac-binding domain of PAK1 (PBD) as described (11,24). Platelets were lysed in RIPA buffer (50 mM Tris, pH 7.2, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 500 mM NaCl, 10 mM MgCl 2 , 10 g/ml each of leupeptin and aprotinin, and 1 mM phenylmethylsulfonyl fluoride), and clarified cell lysates were incubated with GST-RBD or GST-PBD (20 g of beads) at 4°C for 45 min. The beads were washed four times with 50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 150 mM NaCl, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 5 mM EGTA, and 5 g/ml each of leupeptin and aprotinin. The bead pellet was finally suspended in 15 l of Laemmli sample buffer. RhoA and Rac were separated by 12% SDS-PAGE and transferred to nitrocellulose membrane, and GTPases were detected using a specific monoclonal antibody against RhoA and Rac.

SDS-PAGE and
Immunoblotting-SDS-PAGE of photolabeled proteins was performed on 12% (w/v) polyacrylamide gels. Photolabeled membrane proteins were visualized by autoradiography of the dried gels. Blotting of membrane proteins separated by SDS-PAGE, processing of immunoblots, and detection of immunoreactive proteins by chemiluminescence procedure (Amersham Pharmacia Biotech, Braunschweig, Germany) have been described previously (23).
Isolation of Actin Cytoskeleton-Reactions were stopped, and the cytoskeleton was immediately extracted by adding one volume of icecold twice-concentrated cytoskeleton buffer containing 100 mM Tris-HCl, pH 7.4, 20 mM EGTA, 2 mM Na 3 VO 4 , 4 g/ml each of aprotinin and leupeptin, 2 mM phenylmethylsulfonyl fluoride, and 2% (v/v) Triton X-100 as described previously (25,26). After 10 min at 4°C, the cytoskeleton was pelleted by centrifugation (12,000 ϫ g for 10 min at 4°C), and a fraction of the supernatant was taken as a control. Pellets were washed once in cytoskeleton buffer with 0.5% Triton X-100 and once with the same buffer without Triton X-100. Cytoskeleton and a  fraction of the post-spin supernatants as well as of the initial lysate were then immediately prepared for SDS-PAGE.
Lipid Extraction and Analysis-Reactions were stopped by addition of chloroform/methanol (1/1, v/v) containing 0.4 N HCl, and lipids were immediately extracted following the modified procedure of Bligh and Dyer (27,28). PtdIns(4,5)P 2 lipids were immediately deacylated by 20% methylamine and analyzed by high performance liquid chromatography on a Whatman Partisphere 5 SAX column (Whatman International Ltd., UK) as described previously (28).
Scanning Electron Microscopy-Isolated platelets were preincubated under the indicated conditions. Thereafter, platelets were incubated in the absence or presence of U46619 (5 M) for 5 s at 37°C and then fixed for 10 min with 3% paraformaldehyde, 3.75% glutaraldehyde, 0.06 mM cacodylate buffer, and 3.4 mM CaCl 2 . The fixed platelets were suctionfiltered onto Nucleopore polycarbonate filters (0.45 m), which had been preincubated with 10 g/ml polylysine. Filters were washed three times with 0.9% NaCl and dehydrated stepwise in aqueous ethanol. After exchange of ethanol for hexadimethyldisilazane, samples were air-dried and sputtered with gold. Scanning electron microscopy was carried out on a Zeiss-Gemini instrument using a beam voltage of 5 kV.

RESULTS
Human as well as wild-type and G␣ q -deficient mouse platelets were incubated with the TXA 2 mimetic U46619, and activation of Rho and Rac was investigated after different incubation times using the pull-down assays based on GST-Rho-binding domain or GST-Rac-binding domain fusion proteins (11,24). Stimulation of human and mouse platelets by U46619 led to a very rapid and sustained activation of Rho and Rac, which reached a maximum a few seconds after addition of the stimulus (Fig. 1, A and B). In contrast, no activation of Rac was observed in G␣ q -deficient platelets stimulated by U46619 (Fig. 1B). However, in the absence of G␣ q the TXA 2 analogue still induced a Rho activation with a time course indistinguishable from that observed in wildtype platelets (Fig. 1A). Quantification of precipitated Rho and of Rho in the lysates by densitometric analysis of immunoblots demonstrated that the amount of activated Rho in G␣ q -deficient platelets exposed to U46619 was the same as that observed in wild-type platelets and amounted to about 10% of the total Rho (data not shown). These data indicate that the G q -mediated pathway is not required for Rho activation in platelets. After 10 s, the reaction was stopped by ice-cold 2ϫ RIPA buffer, and Rac activation was evaluated as described under "Experimental Procedures." B, platelets were preincubated for 15 min with the indicated concentrations of BAPTA-AM before incubation in the absence or presence of 1 M U46619 for 10 s was started. Shown is a densitometric evaluation of a blot representing the precipitated activated Rac. C, platelets were incubated with the indicated concentrations of the Ca 2ϩ ionophore A23187, and Rac activation was determined as described. D, effect of ADP on Rac activation induced by U46619. Mouse platelet suspensions were incubated with apyrase (2 units/ml) or AR-C69931MX (1 M) and stimulated with ADP (5 M) or U46619 (1 M) as indicated. For the Rac pull-down assay, aggregation was assessed at 37°C with stirring, and after 10 s of stimulation the reaction was stopped and Rac activation was measured. E, platelets from G␣ q -deficient mice were stimulated for 10 s with U46619 (1 M), U46619 in the presence of ADP (10 M), ADP alone, or with thrombin (5 IU/ml). Rac activation was assessed as described under "Experimental Procedures." Data are representative of three independent experiments. G␣ q -deficient platelets do not show any aggregation and secretion in response to U46619 (22). Thus, the sustained Rho activation observed in G␣ q -deficient platelets is obviously independent of integrin ␣ IIb ␤ 3 activation and feedback effects of secreted stimuli. To test whether Rac activation was dependent on integrin ␣ IIb ␤ 3 -mediated aggregation, we pretreated platelets with the GP IIb/IIIa antagonist SR 121566A (30) as well as with the anti-GP IIb/IIIa antibody JON/A (31), which completely inhibited aggregation induced by U46619 ( Fig. 2A). Rac activation by U46619 in mouse platelets was not affected by pretreatment of platelets with GP IIb/IIIa blocking agents (Fig.  2B). Similarly, blockade of integrin ␣ IIb ␤ 3 in human platelets by the peptide Arg-Gly-Asp-Ser (RGDS) had no effect on both Rho and Rac activation (Fig. 2C). These data demonstrate that the rapid activation of Rho and Rac after platelet activation occurs independently of integrin outside-in-signaling.
The lack of Rac activation in the absence of G␣ q indicates that the G q -mediated pathway is upstream of Rac. Rac activation by U46619 required an increase in the free cytosolic Ca 2ϩ concentration, because it was strongly inhibited by pretreatment of platelets with the Ca 2ϩ chelator BAPTA-AM and the phospholipase C inhibitor U-73122 but not by the protein ki-nase C inhibitor RO-31-8220 or the extracellular Ca 2ϩ chelator EGTA (Fig. 3, A and B). To test whether an increase in the cytosolic calcium concentration would induce Rac activation, we incubated cells with A23187. Ionophore treatment resulted in a robust activation of Rac (Fig. 3C). The dependence of Rac activation on G q , PLC, and Ca 2ϩ suggests that a direct pathway mediated by phospholipase C␤ is involved in Rac activation or that Rac activation follows the G q -mediated release of mediators, which in turn may activate G i -coupled receptors. ADP, which is an important mediator of some TXA 2 effects, can activate various receptors, including the G i -coupled P2Y 12 receptor (32). To exclude a possible contribution of G i activated through U46619-induced ADP release, we degraded ADP with apyrase or blocked the G i -coupled P2Y 12 receptor with AR-C69931MX. Apyrase and AR-C69931MX completely blocked ADP-induced aggregation (Fig. 3D). Both apyrase and AR-C69931MX partially inhibited the U46619-induced Rac activation in wild-type platelets (Fig. 3D) indicating that ADP and the G i -coupled P2Y 12 receptor are involved in the effect of U46619 on Rac activation. To test whether G i activation alone is sufficient to lead to activation of Rac in platelets, we tested the effect of thrombin and ADP on GTP loading of Rac in FIG. 4. U46619-mediated actin polymerization and PtdIns(4,5)P 2 synthesis in wild-type and G␣ q -deficient platelets. A, wild-type and G␣ q -deficient platelets were preincubated for 3 min in the absence or presence of JON/A antibody (30 g/ml). Cytoskeleton was precipitated by addition of Triton X-100 and collected by centrifugation as described under "Experimental Procedures." Cytoskeletal proteins and 3% of the total post-spin supernatant as well as of the total lysate were then separated by SDS-PAGE (7.5%) and stained with Coomassie Blue. Data are representative of three independent experiments. B, densitometric evaluation of the F-actin band in cytoskeletal fraction of platelets treated in three independent experiments as described. C, time course of changes in the radioactivity of PtdIns(4,5)P 2 after U46619 stimulation in 32 P-labeled platelets. Platelets were stimulated by U46619 (1 M) for the indicated time periods, and the radioactivity of PtdIns(4,5)P 2 was determined as described under "Experimental Procedures." Results are means of two independent experiments. D, wild-type and G␣ q -deficient platelets were preincubated for 30 min in the absence or presence of 10 M Y-27632. Five seconds after addition of 1 M U46619, PtdIns(4,5)P 2 content was determined as described under "Experimental Procedures." Data (-fold increase over control) are means Ϯ S.E. of three independent experiments. Stars indicate significant differences (p Ͻ 0.05; Student's t test). G␣ q -deficient platelets in which both agonists induce activation of G 12 /G 13 and G i (18). Although thrombin and ADP and ADPϩU46619 led to Rac activation in wild-type platelets (data not shown), no Rac activation could be observed in response to these stimuli in G␣ q -deficient platelets (Fig. 3E). This indicates that G i activation alone is not sufficient for Rac activation and that a direct G q /PLC␤-mediated mechanism as well as the Ca 2ϩ -dependent release of ADP and possibly other mediators are required for full Rac activation in platelets.
Platelet activation is accompanied by rapid actin polymerization, a process that is regulated through the formation of phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P 2 ) (33, 34). The major PtdIns(4,5)P 2 -forming enzyme, phosphatidylinositol-4-phosphate 5-kinase, has been shown to be regulated by Rho and Rac (35)(36)(37). Especially in platelets, evidence has been provided that Rac links receptors to PtdIns(4,5)P 2 formation and actin polymerization via stimulation of phosphatidylinositol-4-phosphate 5-kinase (16,17). Although a considerable portion of actin polymerization during platelet activation is mediated in an integrin ␣ IIb ␤ 3 -dependent manner, some actin polymerization accompanies the protrusion of filopodia and lamellipodia during platelet shape change (18,38,39). Accordingly, in G␣ q -deficient platelets, which do not aggregate in response to TXA 2 receptor activation, U46619 induced an increase in the Triton X-100-insoluble actin fraction (Fig. 4, A  and B). The latter was comparable to the effect observed in wild-type platelets treated with a GP IIb/IIIa blocking antibody (Fig. 4, A and B). To test whether there was still some receptordependent PtdIns(4,5)P 2 formation in the absence of G q -mediated Rac activation, we determined PtdIns(4,5)P 2 levels in response to U46619 in wild-type and G␣ q -deficient platelets. Although there was no Rac activation in G␣ q -deficient platelets in response to U46619, a significant increase in PtdIns(4,5)P 2 levels by about 30% could be observed shortly after addition of the TXA 2 analogue, which was about half of that seen in wildtype mouse platelets (Fig. 4C). At longer times of stimulation (3 min), the PtdIns(4,5)P 2 levels decreased in wild-type platelets probably due to the activation of phospholipase C, whereas in G␣ q -deficient platelets, which do not show any phospholipase C activation, PtdIns(4,5)P 2 levels remained elevated. These data show that platelets can form PtdIns(4,5)P 2 and undergo shape change in response to U46619 by a mechanism independent of Ca 2ϩ and Rac activation. To test whether the Rho/Rho-kinasemediated pathway is involved in the formation of PtdIns(4,5)P 2 found in G␣ q -deficient platelets, we preincubated wild-type platelets and platelets lacking G␣ q with 10 M of the Rho-kinase inhibitor Y-27632. Although PtdIns(4,5)P 2 formation in wild-type platelets was partially reduced, the U46619-dependent formation of PtdIns(4,5)P 2 in G␣ q -deficient platelets was completely blocked by Y27632 (Fig. 4D). This clearly suggests that PtdIns(4,5)P 2 formation in platelets is under dual control through a pathway mediated by G q /Rac and G 12/13 /Rho/Rho-kinase.
The two main intracellular mediators of platelet inhibition, cAMP and cGMP strongly inhibit platelet aggregation but have been shown to differentially affect platelet shape change. While cAMP blocks the platelet shape change response, cGMP has no effect on shape change of human or mouse platelets (Fig. 5, Ref.  18). Preincubation of mouse platelets with prostacyclin which induces cAMP formation through the prostacyclin receptor or with the cAMP analogue Sp-5,6-DCl-cBIMPS strongly suppressed U46619-induced Rho activation in wild-type (Fig. 6A) and G␣ q -deficient platelets (data not shown). In contrast, the cGMP analogue 8-pCPT-cGMP was without effect (Fig. 6A). Activation of Rac by U46619 was, however, strongly inhibited by the cGMP analogue as well as by prostacyclin and the cAMP analogue. Very similar effects of cyclic nucleotides on Rho and Rac activation were observed in human platelets (Fig. 6B). Thus, cAMP inhibits both Rac and Rho activation while cGMP blocks only receptor-mediated Rac activation.
Because Rho activation by U46619 appears to be mediated by G 12 /G 13 , whereas Rac involves G q , we tested whether cAMP and cGMP interfere with TXA 2 receptor-mediated activation of G q and/or G 12 /G 13 . Photolabeling of receptor-activated G-proteins in mouse platelet membranes and subsequent immunoprecipitation of individual G-protein ␣-subunits showed that, in wild-type mouse platelets, activated TXA 2 receptors couple to G q , G 12 , and G 13 (18). Preincubation of human platelet membranes with the cGMP analogue 8-pCPT-cGMP almost completely blocked U46619-induced activation of G q but had no effect on the activation of G 12 and G 13 (Fig. 7). In contrast, preincubation with the cAMP analogue Sp-5,6-DCl-cBIMPS resulted in the inhibition of U46619-induced activation of both G q and G 12 /G 13 (Fig. 7). The effects of cAMP were, however, less pronounced than those of cGMP. These data indicate that cAMP and cGMP differentially interfere with coupling of TXA 2 receptors to G q and G 12 /G 13 . The selective effects of cGMP and cAMP on receptor-G-protein coupling may contribute to inhibition of G-protein-mediated Rho and Rac activation in platelets. DISCUSSION Platelets have been extensively used as a model system to study rapid cell regulation through G-protein-coupled receptors. TXA 2 induces platelet shape change, aggregation, and granule secretion acting through receptors that couple to G q and G 12 /G 13 but not to G i -type G-proteins (18, 21). Lack of G q -mediated phospholipase C activation in platelets from G␣ qdeficient mice blocks TXA 2 -induced platelet aggregation and secretion but does not interfere with the ability of the TXA 2 mimetic U46619 to induce platelet shape change, which appears to be mediated by G 12 /G 13 (18,22). Here we used human and mouse platelets to study the regulation of the small GTPases Rho and Rac.
Rho became rapidly activated in response to U46619 in wildtype platelets as well as in platelets lacking G␣ q . This indicates that G-proteins of the G 12 family are involved in the activation of Rho. In contrast, Rac activation in response to U46619 could only be observed in wild-type platelets and was absent in G␣ q -deficient platelets, suggesting that G q mediates this effect and that G 12 /G 13 are not involved. In neuronal cells and cardiomyocytes, it has been shown that Rac activation can be induced by stimuli that are believed to function through G q / G 11 -coupled receptors like bradykinin, endothelin-1, or phenylephrine (40,41). On the other hand, in neutrophils, activation of Rac through the G-protein-coupled receptor of the chemoattractant N-formyl-L-methionyl-L-leucyl-L-phenylalanine has been shown to occur in a pertussis toxin-sensitive manner indicating that it is mediated by G i -type G-proteins (11,13,42). TXA 2 effects in platelets have been shown to be at least partially mediated through the release of mediators like ADP, which exert their effects on platelets in part through G i -coupled receptors (43)(44)(45)(46). Because G␣ q -deficient platelets do not secrete their granule contents in response to U46619, lack of Rac activation in the absence of G␣ q could be due to the defective mediator release. We therefore tested the effect of the ADPdegrading enzyme apyrase and antagonists of the G i -coupled P2Y 12 receptor on U46619-induced Rac activation. Both agents reduced the effect of U46619 on Rac activation. However, activation of G i alone was not sufficient to induce Rac activation as demonstrated by the fact that neither thrombin nor ADP and ADPϩU46619 were able to activate Rac in G␣ q -deficient platelets. Thus, full Rac activation appears to require both a direct G q -dependent signal as well as the G q -mediated release of mediators acting through G i -coupled receptors.
Induction of Rho and Rac activation by U46619 occurred very rapidly. The fact that activation of Rho could also be observed in G␣ q -deficient platelets, which do not secrete their granule contents and aggregate in response to U46619, shows that Rho activation was independent of the G q /PLC␤-mediated pathway resulting in degranulation and integrin ␣ IIb ␤ 3 activation. Blockade of integrin ␣ IIb ␤ 3 in wild-type platelets did not interfere with U46619-induced Rac activation, indicating that Rac activation also occurred independently of integrin signaling. Although Rho and Rac become activated in an integrin ␣ IIb ␤ 3independent manner early during platelet activation, integrin activation may contribute to Rho and Rac activation at later stages of platelet activation in a manner similar to that occurring during integrin-mediated cell adhesion (23,(47)(48)(49).
G␣ q -deficient platelets still undergo shape changes, including actin polymerization and phosphorylation of the myosin light chain (16,50). Rho has been suggested to link receptoractivated G-proteins of the G 12 family to the stimulation of Rho-kinase, which via inhibition of myosin phosphatase results in an increased phosphorylation of the myosin light chain (18 -20). The fact that U46619-induced Rho activation was not affected by the loss of G␣ q supports the concept that Rho is involved in the induction of platelet shape change by transducing the G 12 /G 13 -mediated signal. Whether Rho is involved in regulation of integrin ␣ IIb ␤ 3 in platelets is not clear. Although partial inactivation of the RhoA pool in human platelets by C3 exoenzyme has been shown to inhibit platelet activation (51,52), ADP-ribosylation of about 90% of Rho in human platelets did not affect inside-out signaling of integrin ␣ IIb ␤ 3 , ligandinduced aggregation, and F-actin content (53).
The lack of Rac activation in the absence of G␣ q demonstrates that activation of Rac is not required for the platelet shape change response in G␣ q -deficient platelets that still show increased F-actin formation in response to U46619 (see Fig. 4, A and B (18)). Although Rac has been involved in PtdIns(4,5)P 2 -mediated actin polymerization (16), there appear to be alternative Rac-independent mechanisms that lead to actin polymerization in the absence of Rac activation. In G␣ qdeficient platelets U46619 still increased the formation of PtdIns(4,5)P 2 . However, stimulation of PtdIns(4,5)P 2 formation was about 50% lower than in wild-type platelets. This residual formation of PtdIns(4,5)P 2 in the absence of G␣ q could be completely blocked by the Rho-kinase inhibitor Y-27632 indicating that this G␣ q -and Rac-independent formation is mediated through a Rho/Rho-kinase-dependent pathway. Both Rho and Rho-kinase have been shown to induce activation of phosphatidylinositol-4-phosphate 5-kinase in other cellular systems (35,54). Collectively, these data show that two independent pathways involving G q /Rac and G 12/13 /Rho/Rho-kinase appear to mediate the receptor-dependent formation of PtdIns(4,5)P 2 . This regulation is different from a recently reported model based on a cotransfected Cos-7 cell system in which PtdIns(4,5)P 2 formation through the G-protein-coupled receptor PAR1 was shown to involve the sequential activation of G q , Rac, Rho, and phosphatidylinositol-4-phosphate 5-kinase I␣ (55).
The cyclic nucleotides cAMP and cGMP are physiological mediators of platelet inhibition through activation of cAMPand cGMP-dependent kinases. Although analogues of both cyclic nucleotides can block platelet aggregation, only cAMP analogues inhibit platelet shape change (18,56,57). Similarly, we observed that the cAMP analogue Sp-5,6-DCl-cBIMPS but not the cGMP analogue 8-pCPT-cGMP inhibited TXA 2 receptormediated shape change in wild-type and G␣ q -deficient platelets (Fig. 5). The inhibition of Rho activation by Sp-5,6-DCl-cBIMPS but not by 8-pCPT-cGMP in human and mouse platelets as well as in G␣ q -deficient platelets suggests that the Rho/Rho-kinase-mediated signaling cascade is inhibited by the cAMP-dependent pathway (Fig. 6). A similar role of cAMP was suggested for the inhibition of Rho/Rho-kinase-mediated neurite remodeling and morphology change in epithelial-like cells (58,59). In contrast to platelets, cGMP appears to inhibit the Rho/Rho-kinase-mediated pathway in smooth muscle cells probably at a site downstream of Rho (60,61). Both cGMP and cAMP inhibit Ca 2ϩ mobilization through TXA 2 receptors in platelets (62), which is mediated by the G q /phospholipase C␤ pathway (22). Consistent with that, we found that both cyclic nucleotides blocked U46619-induced Rac activation in mouse and human platelets (Fig. 6). Because cGMP in contrast to cAMP does not interfere with platelet shape change, these data provide additional evidence that Rac activation is not required for platelet shape change. These findings also indicate that cAMP and cGMP exert their inhibitory effects on platelet activation at least in part by interfering with upstream mechanisms of receptor-induced signaling cascades. By studying the effect of cGMP and cAMP analogues on coupling of TXA 2 receptors to G q and G 12 /G 13 , we found that cGMP strongly inhibited TXA 2 receptor-mediated G q activation, whereas cAMP partially inhibited activation of all three G-proteins through the activated receptors (Fig. 7). Thus, cGMP appears to selectively interfere with activation of G q but not of G 12 /G 13 , which may explain the lack of cGMP effects on U46619-induced Rho activation. The TXA 2 receptor has been shown to be a substrate for cAMP-dependent kinase (63, 64) as well as for the cGMP-dependent kinase (65), and cGMP inhibits TXA 2 receptor-mediated stimulation of high affinity GTPase in platelet membranes (65). In addition, cAMP leads to the phosphorylation of TXA 2 receptor-coupled G␣ 13 (66). These cyclic nucleotide-dependent kinase effects on the level of TXA 2 receptors and their interaction with G-proteins may contribute to the observed inhibition of G q -mediated Rac activation by cGMP and cAMP as well as of the G 12 /G 13 -mediated Rho activation by cAMP. Other mechanisms mediated by cyclic nucleotide-dependent kinases are likely to be also involved in cGMP/cAMP effects on Rho and Rac activation (67)(68)(69)(70).
Taken together, our data show that Rho and Rac are differentially regulated by G 12 /G 13 and through G q -mediated mechanisms. Both processes occur rapidly upon receptor activation and are not dependent on integrin activation. Activation of Rac is not required for platelet shape change involving actin polymerization and myosin light-chain phosphorylation. The differential upstream regulation of Rho and Rac activity explains their different sensitivity toward the effects of cyclic nucleotides. Cyclic AMP partially inhibits receptor-mediated activation of G 12 /G 13 and G q whereas cGMP only inhibits activation of G q . Analogously, Rho and Rac activation via heterotrimeric G-proteins is subject to inhibition by cAMP whereas only the Rac activation can be blocked by cGMP.  ) and solubilized, and G-protein ␣-subunits (G␣ 12 , G␣ 13 , and G q ) were immunoprecipitated as described under "Experimental Procedures." Anti-G␣ 12 , anti-G␣ 13 , and anti-G␣ q/11 antisera were used. Precipitated proteins were subjected to SDS-PAGE. Shown are autoradiograms of dried SDS gels. Data are representative of four independent experiments.