L E T T E R S

1418 VOLUME 9 | NUMBER 11 | NOVEMBER 2003 NATURE MEDICINE

Platelet activation at sites of vascular injury is essential forprimary hemostasis, but also underlies arterial thrombosisleading to myocardial infarction or stroke1,2. Platelet activatorssuch as adenosine diphosphate, thrombin or thromboxane A2 (TXA2) activate receptors that are coupled to heterotrimericG proteins1,3. Activation of platelets through these receptorsinvolves signaling through Gq, Gi and Gz (refs. 4–6). However,the role and relative importance of G12 and G13, which areactivated by various platelet stimuli7–9, are unclear. Here weshow that lack of Gα13, but not Gα12, severely reduced thepotency of thrombin, TXA2 and collagen to induce plateletshape changes and aggregation in vitro. These defects wereaccompanied by reduced activation of RhoA and inability toform stable platelet thrombi under high shear stress ex vivo.Gα13 deficiency in platelets resulted in a severe defect inprimary hemostasis and complete protection against arterialthrombosis in vivo. We conclude that G13-mediated signalingprocesses are required for normal hemostasis and thrombosisand may serve as a new target for antiplatelet drugs.

G12 and G13 constitute a subfamily of heterotrimeric G proteinsthat are activated through various receptors, including thoseinvolved in many platelet stimuli. Mice lacking Gα13, the α-subunitof G13, die in utero because of a defect in angiogenesis, whereasGα12-deficient mice are phenotypically normal10,11. To circumventembryonic lethality of mice deficient in Gα13 and to study the roleof Gα12- and Gα13-mediated signaling in platelet activation, weused Cre/loxP-mediated recombination to conditionally inactivateGna13, the gene encoding Gα13, alone or in a Gα12-deficient(Gna12–/–) background. We generated a Gna13 allele, Gna13ta, con-taining three loxP sites and a cassette containing the neomycinresistance (neor) and the thymidine kinase gene (tk), by gene target-ing of embryonic stem cells (Fig. 1a). The Gna13ta allele was con-verted into a conditional allele, in which exon 2 is flanked by loxPsites (Gna13flox), or into a deleted allele (Gna13−) by crossing theseanimals with the Cre-deleter mouse strain EIIa-Cre (ref. 12;Fig. 1a–c). Gna13flox/flox mice were normal, whereas full recombi-

nation of both Gna13flox alleles results in a phenotype characteris-tic of the Gna13– null allele10 (data not shown).

To generate mice lacking Gα12 and Gα13 in platelets, we used amouse line expressing the Cre gene under the control of the inter-feron-inducible Mx promoter (Mx-Cre)13. Deletion of the Gna13gene in mice carrying Mx-Cre was induced by intraperitonealinjections of polyinosinic-polycytidylic acid (PIPC). Treatmentwith PIPC resulted in almost complete recombination in liver, bonemarrow and spleen, as assessed by Southern blotting (data notshown). Four weeks after the last PIPC injection, there was no Gα13protein detectable in platelets, liver or spleen, whereas levels ofother G-protein α-subunits were unchanged (Fig. 1d and data notshown). Platelet counts were not affected by Gα13 deficiency, andsurface expression of prominent membrane receptors, includingglycoprotein Ib-Ix (GPIb-Ix), GPV, GPVI, β1 and β3 integrins andCD9, was normal (data not shown).

Because previous studies suggested a role for G12 and G13 in thereceptor-mediated platelet shape change response7, we exposedplatelets deficient in both Gα12 and Gα13 (Gα12/Gα13-deficient) toincreasing concentrations of various platelet activators. In wild-type platelets, thrombin, collagen and the TXA2 mimetic U46619led to platelet shape change at concentrations one to two orders ofmagnitude lower than those required to induce aggregation. Incontrast, there was no shape change in Gα12/Gα13-deficientplatelets in response to low agonist concentrations (Fig. 2a,b andSupplementary Fig. 1 online). However, at high agonist concentra-tions, Gα12/Gα13-deficient platelets underwent shape change andall stimuli induced aggregation, although with considerably lowerpotency than in wild-type platelets. Surprisingly, platelets lackingonly Gα13 showed defects indistinguishable from those observed inGα12/Gα13-deficient platelets, whereas platelets lacking only Gα12behaved like wild-type. Thus, Gα13, but not Gα12, is required forthe induction of shape change and aggregation in response to lowand intermediate concentrations of TXA2, thrombin and collagen.

To test whether the aggregation defect observed in Gα13 andGα12/Gα13-deficient platelets was caused by defective inside-outactivation of αIIbβ3 integrin, we directly assessed this process by

1Institute of Pharmacology, University of Heidelberg, Im Neuenheimer Feld 366, 69120 Heidelberg, Germany. 2Vascular Biology, Rudolf Virchow Center forExperimental Biomedicine, University of Würzburg, 97078 Würzburg, Germany. 3Deutsches Herzzentrum und 1. Medizinische Klinik, Technische UniversitätMünchen, D-80636 München, Germany. 4Division of Biology 147-75, California Institute of Technology, Pasadena, California 91125, USA. 5Present address:INSERM, Unité 326, Hôpital Purpan, 31059 Toulouse Cedex, France. Correspondence should be addressed to S.O. (Stefan.Offermanns@urz.uni-heidelberg.de).

Published online 5 October 2003; doi:10.1038/nm943

G13 is an essential mediator of platelet activation inhemostasis and thrombosisAlexandra Moers1, Bernhard Nieswandt2, Steffen Massberg3, Nina Wettschureck1, Sabine Grüner2,Ildiko Konrad3, Valerie Schulte2, Barsom Aktas2, Marie-Pierre Gratacap1,5, Melvin I Simon4,Meinrad Gawaz3 & Stefan Offermanns1

©20

03 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

rem

edic

ine

L E T T E R S

NATURE MEDICINE VOLUME 9 | NUMBER 11 | NOVEMBER 2003 1419

flow cytometry. Because this analysis was conducted in dilutedplatelet suspensions, which largely excludes the accumulation ofreleased mediators, U46619 does not lead to significant αIIbβ3 acti-vation and degranulation unless Gi-type G proteins are activated bycoadministration of adrenaline or adenosine diphosphate14.Thrombin-induced αIIbβ3 activation was indistinguishablebetween wild-type and Gα12/Gα13-deficient platelets, and therewas only a very small decrease in Gα13 and Gα12/Gα13-deficientplatelets, as compared with wild-type and Gα12-deficient platelets,when exposed to U46619 in the presence of adrenaline (data notshown). Thus, the defective aggregation response of Gα13- andGα12/Gα13-deficient platelets is not caused by impaired inside-outactivation of αIIbβ3.

We next examined whether the absence of Gα12, Gα13 or bothhad any effect on platelet degranulation. Whereas U46619- oradrenaline-induced P-selectin exposure and serotonin secretionwas markedly reduced in Gα13 and Gα12/Gα13-deficient platelets

a b c

d

Figure 1 Generation of Gna13flox alleleand Cre-mediated recombination inmice. (a) Targeting strategy (seeSupplementary Methods online fordetails). �, loxP sites; neortk, neomycinresistance/thymidine kinase cassette;Gna13+, wild-type allele; Gna13ta,targeted allele; Gna13–, deleted allele;Gna13flox, floxed allele. S, SpeI; Sc,SacI; K, KpnI; H, HindIII; N, NheI; A, probe A; B, probe B. (b) HindIII-digested genomic DNA of indicatedembryonic stem cell clones hybridizedwith probe B. (c) Identification of Gna13−

and Gna13flox alleles after Cre-mediatedrecombination in SpeI-digested tail-tipDNA hybridized with probe A. (d) Westernblots of platelet lysates from wild-type,Gna12–/–, PIPC-induced Mx-Cre Gna13flox/flox or Mx-Cre Gna12–/–Gna13flox/flox mice probed withantibodies to G-protein α-subunits (anti-Gα). fl, floxed allele. Arrows indicate 42-kDa marker protein.

Figure 2 Ex vivo analysis of wild-type andGα12/Gα13-deficient platelets. (a,b) Platelets werestimulated with increasing concentrations ofthrombin (a) or U46619 (b). Data are shown asaggregometric traces; addition of stimuli is indicatedby arrows. (c,d) Secretion induced by U46619 andadrenaline (c) or PAR-4 peptide (d) was analyzed byflow cytometry using FITC-labeled antibodies againstP-selectin. Stimuli were given at the indicatedconcentrations. WT, wild-type; 12–/–, Gna12–/–;13–/–, Gna13–/–; 12/13–/–, Gna12–/–Gna13–/–.

(Fig. 2c and data not shown), thrombin-induced secretion was notaffected by Gα12 and/or Gα13 deficiency (data not shown).Thrombin exerts its effects on mouse platelets primarily throughthe protease-activated receptor (PAR)-4 (ref. 15). P-selectin expo-sure and serotonin secretion in response to the PAR-4-activatingpeptide AYPGKF was clearly reduced in platelets lacking Gα13(Fig. 2d). This suggests that thrombin induces additional, PAR-4-independent signaling pathways that compensate for the lack ofGα12 and/or Gα13 in the degranulation response, but not in aggre-gation and shape change. GPIb, which has a role in thrombin-induced activation processes16, does not seem to be involved, asproteolytic removal of the 45-kDa N-terminal ligand-bindingdomain of GPIb-α did not affect the ability of thrombin to inducesecretion in Gα13-deficient platelets (data not shown).

Because the Rho- and Rho kinase–mediated signaling pathwayleading to myosin light chain (MLC) phosphorylation is activatedthrough G12 and G13 together (ref. 17), and because Rho and Rho

kinase are involved in the induction of plateletshape change7,18,19, we tested whether activa-tion of RhoA and phosphorylation of MLCwere affected in Gα12/Gα13-deficient platelets.In wild-type platelets, U46619 induced RhoAactivation and MLC phosphorylation, with amaximal effect at 10 nM. Again, Gα12-deficient platelets were indistinguishable fromwild-type platelets in this regard. InGα12/Gα13- or Gα13-deficient platelets, how-ever, no RhoA activation or MLC phosphory-lation was observed in response to 10 nMU46619 (Fig. 3a), indicating that RhoA activa-

a b c

d

©20

03 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

rem

edic

ine

L E T T E R S

1420 VOLUME 9 | NUMBER 11 | NOVEMBER 2003 NATURE MEDICINE

tion and subsequent MLC phosphorylation is primarily inducedthrough G13 in platelets.

We next studied the potential role of the G13-mediated signalingpathway in thrombus formation during perfusion of whole bloodover collagen under high shear conditions ex vivo (Fig. 3b). Wild-type and Gα12-deficient platelets adhered to collagen fibers, andadherent platelets initiated the formation of platelet aggregateswithin 2 min. Initial adhesion of Gα13- or Gα12/Gα13-deficientplatelets was indistinguishable from that observed with wild-typeplatelets. However, platelet thrombus formation was severelyimpaired, and virtually no thrombi could be observed after rinsingthe chamber (Fig. 3b–d). These findings suggest that the formationand stabilization of platelet thrombi under high shear flow condi-tions requires intact signaling through the G13-mediated pathway.The defective activation of RhoA in the absence of G13 may con-tribute to the observed defect in the formation of stable plateletaggregates, as RhoA may be required for platelet aggregation underhigh shear conditions and for irreversible aggregation of plateletsin suspension20,21.

To test whether defects observed in Gα13- and Gα12/Gα13-defi-cient platelets in vitro would have any effect under in vivo condi-tions, we determined tail-bleeding times as a measure for primaryhemostasis (Fig. 4a). Mice with Gα13- or Gα12/Gα13-deficientinterferon-responsive tissues had massively increased bleedingtimes compared with wild-type or Gα12-deficient mice (Fig. 4a). Torule out the possibility that Gα13 deficiency in endothelial cells,hepatocytes or other interferon-responsive organs contributed tothe increased bleeding time, we restricted the Mx-Cre-induceddeletion of Gna13 to the hematopoietic system by transferring

bone marrow cells derived from PIPC-induced Mx-CreGna12–/–Gna13flox/flox or Mx-Cre Gna13flox/flox mice into irradiatedwild-type recipient animals. Four weeks after the transfer, all trans-planted animals had normal platelet counts, but bleeding times ofmice transplanted with Gα13- or Gα12/Gα13-deficient bone mar-

a

b

c d

Figure 3 Ex vivo analysis of wild-type and Gα12/Gα13-deficient platelets. (a) Detection of activated Rho in platelets stimulated with U46619, asdescribed in Methods. Arrows indicate 20-kDa marker protein. WT, wildtype; IB, immunoblot; anti-RhoA, anti-P-MLC, anti-MLC, antibodies toRhoA, phosphorylated MLC and MLC, respectively. (b) Phase-contrastmicroscopy of whole blood from wild-type, Gα12-deficient (Gna12–/–), PIPC-induced Mx-Cre Gna12–/–Gna13flox/flox (Gna12/13–/–) or PIPC-induced Mx-Cre Gna12+/+Gna13flox/flox (Gna13–/–) mice, perfused over a collagen-coatedsurface. (c,d) Analysis of flow chamber experiments showing mean ± s.d. ofpercentage of surface area covered by thrombi (c) and numbers of singleattached platelets per microscopic field (d; n = 5–8).

a b c d e

f

Tim

e (m

in)

Figure 4 In vivo analysis of hemostasis and thrombosis. (a) Bleeding times of PIPC-induced mice with theindicated genotypes, as described in Methods. Each point represents one individual mouse. fl, floxed allele. (b–d) Bleeding time experiments (b), platelet adhesion (c) and thrombus formation (d) in wild-type mice (Recip.)transplanted with bone marrow from the following animals (Donor): noninduced Mx-Cre Gna12+/+Gna13flox/flox

(WT), PIPC-induced Mx-Cre Gna12–/–Gna13flox/flox (12/13–/–) or PIPC-induced Mx-Cre Gna12+/+Gna13flox/flox

(13–/–). n = 6 for c and d. (e,f) Representative cross-sections of carotid arteries of wild-type mice (e) and micewith Gα13-deficient platelets (f).

©20

03 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

rem

edic

ine

L E T T E R S

NATURE MEDICINE VOLUME 9 | NUMBER 11 | NOVEMBER 2003 1421

row were massively prolonged (>20 min) compared with micetransplanted with bone marrow from noninduced Mx-CreGna12+/+Gna13flox/flox mice (Fig. 4b). This indicates that G13-medi-ated signaling is required for normal hemostasis.

We next assessed the role of Gα13 in arterial thrombus formationby in vivo fluorescence microscopy of the injured carotid artery.The experiment was conducted using wild-type animals trans-planted with wild-type or Gα13-deficient bone marrow. Afterinduction of injury by ligation of the carotid artery, wild-typeplatelets adhered to the subendothelial surface and massivethrombi subsequently developed in the carotid artery (Fig. 4c–e).Platelets lacking Gα13 also adhered rapidly to the injured vesselwall (Fig. 4c). However, the total number of platelets that firmlyadhered was reduced by ∼ 50% compared with wild-type platelets.In contrast to wild-type animals, mice with Gα13-deficient plateletsdid not show any thrombi after vascular injury, indicating thatGα13-deficient platelets were not able to form stable thrombi invivo (Fig. 4d,f). This is consistent with the data obtained ex vivo andshows that the initial adhesion of platelets to the injured vessel wallwas not substantially altered, whereas the subsequent formationand stabilization of platelet thrombi was altered by the absence ofG13-mediated signaling.

Platelet activation during hemostasis and thrombosis is a com-plex process involving many stimuli that act on platelets at sites ofvascular injury. Most of these stimuli directly or indirectly use G-protein-mediated signaling pathways to induce platelet activation.Defining the relative importance of signaling through differentheterotrimeric G proteins is useful in understanding the molecularmechanisms of platelet activation. Our data clearly indicate thatG13-mediated signaling is necessary for platelet activation in hemo-stasis and thrombosis. Whereas the phenotypic changes in plateletsfrom Gα i2- or Gαz-deficient mice seem to be less severe6,22,23, thedefects observed in mice with Gα13-deficient platelets are compa-rable with the in vivo defects in mice lacking Gαq (ref. 4). Thus,platelet activation seems to be an integrated process that requiresdifferent G-protein signaling pathways involving Gq, Gi, Gz andG13. Although signaling through two of these pathways can be suf-ficient to induce some platelet activation5,14,24, all three pathwaysseem to be required for efficient platelet activation under physio-logical and pathological conditions. The severe defects observed inmice with Gα13-deficient platelets may be caused by the decreasedresponsiveness of Gα13-deficient platelets to various stimuli.Reduced potency of platelet activators may become limiting, espe-cially under high flow conditions that lead to rapid clearance of sol-uble stimuli. We also observed impaired stabilization of plateletthrombi under high shear stress, a process that may require acuteand continuous activation of RhoA-mediated signaling pathways.The fact that Gα13 deficiency protected animals against thrombosissuggests that inhibition of the G13-mediated signaling pathway inplatelets is a promising strategy to prevent or treat platelet activa-tion in thrombosis.

METHODSConditional inactivation of the Gna13 gene. See Supplementary Methodsonline.

Western blot analysis. Platelet lysates were subjected to SDS-PAGE. Afterblotting, nitrocellulose membranes were probed with antibodies to Gα12

(ref. 25) and to Gα13, Gα i and Gαq/11 (Santa Cruz Biotechnology).

Platelet preparation and aggregation. All animal experiments and care wereapproved by the local Animal Care & Use Committee. Whole blood was

collected from the retro-orbital plexus in acid citrate dextrose (150 µl perml blood) from anesthetized mice. Blood was diluted with half the volumeof HEPES–Tyrode buffer [137 mM NaCl, 2 mM KCl, 12 mM NaHCO3, 0.3mM NaH2PO4, 2 mM CaCl2, 1 mM MgCl2, 5.5 mM glucose and 5 mMHEPES (pH 7.3)] containing 0.35% human serum albumin. Platelet-richplasma was washed once, and optical aggregation experiments were con-ducted in a two-channel aggregometer (Chrono-Log).

Flow cytometry. Washed platelets (2 × 106) were incubated with the indi-cated stimuli for 2 min at 37 °C, stained with fluorophore-conjugated mon-oclonal antibodies (emfret Analytics) at saturating concentrations for 15 min at room temperature and directly analyzed on a FACSCalibur(Becton Dickinson). Platelets were gated by forward and side scatter14.

Secretion of 5-hydroxy[14C]tryptamine. Platelets (5.8 × 108 /ml) wereloaded with 5-hydroxy[14C]tryptamine (200 nCi/ml) and stimulated withU46619 for 45 s. After fixation and centrifugation, 5-hydroxy[14C]trypta-mine release was determined by liquid-scintillation counting.

Determination of activated cellular Rho and MLC phosphorylation. Theamount of activated cellular Rho was determined by precipitation with afusion protein consisting of glutathione-S-transferase (GST) and the Rho-binding domain of rhotekin (GST-RBD)26. Platelets were stimulated with10 nM of U46619 for 10 s, lysed and subjected to GST-rhotekin pull-down.Precipitated GTP-bound RhoA and lysates were immunoblotted with anti-bodies against RhoA, MLC or the phosphorylated form of MLC (all fromSanta Cruz Biotechnology).

Flow chamber experiments. Experiments were conducted as described27.Briefly, transparent flow chambers with a slit depth of 50 µm, equipped withHorm-type collagen-coated cover slips, were connected to a syringe filledwith anticoagulated blood. Perfusion was done using a pulse-free pumpunder high shear stress equivalent to a wall shear rate of 1000 s–1 (4 min). Thechambers were then rinsed with a 4 min perfusion of HEPES buffer contain-ing 2 mM Ca2+, at the same shear stress, and phase-contrast images wererecorded from at least five different microscope fields (×40 objectives).Shown are experiments representative of five to eight experiments per group.

Determination of bleeding times. Adult mice were anesthetized withintraperitoneal pentobarbital (50 mg per kg body weight) and intramuscu-lar xylazin (3 mg/kg), and the tail was cut 5–6 mm from the tip andimmersed in saline at 37 °C. Bleeding time was defined as the time at whichall visible signs of bleeding from the incision had stopped. The experimentwas stopped after 20 min.

Carotid artery ligation and assessment of platelet adhesion and aggrega-tion. Experiments were conducted as described28. Platelets were isolated andlabeled with 5-carboxyfluorescein diacetate succinimidyl ester as described.Labeled platelets (200 × 106 platelets per 250 µl) were infused intravenouslyinto anesthetized mice of the same genotype as those used to prepare theplatelets. The right common carotid artery was dissected free and ligated vig-orously near the carotid bifurcation for 5 min to induce vascular injury.Adhesion of autogeneous platelets was assessed before and after carotidinjury by in vivo video microscopy28. Videotaped images were evaluatedusing a computer-assisted image analysis program28 (Cap Image 7.4, H.Zeintl). The number of adherent platelets was assessed by counting the cellsthat did not move or detach from the endothelial surface within 10 s. Thearea covered by thrombi was also quantified and is presented in µm2.

Histology. Carotid arteries were perfusion-fixed in situ with 4%paraformaldehyde (pH 7.0) 15 min after induction of injury, excised, fixed in0.1 M cacodylate-buffered Karnovsky solution (2.5% glutaraldehyde and 1%paraformaldehyde) overnight at room temperature and postfixed in 1%osmium tetroxide for 2 h at pH 7.3. The samples were dehydrated in gradedethanols and embedded in EmBed-812 epoxy resin (all reagents from ScienceServices). After 48 h of heat polymerization at 60 °C, semithin (0.8 µm) sec-tions were cut with a diamond knife (Diatome) and double stained withaqueous solutions of 1% toluidine blue and basic fuchsin (60 °C, 1 min).

©20

03 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

rem

edic

ine

L E T T E R S

1422 VOLUME 9 | NUMBER 11 | NOVEMBER 2003 NATURE MEDICINE

Bone marrow transplantation. Bone marrow cells from mice (10–12 weeksold) were removed aseptically from femurs and tibias. Cells (5 × 106) wereresuspended in DMEM and transplanted by tail-vein infusion into lethallyirradiated (10 Gy) recipients (C57BL/6; Charles River) 1 d after irradiation.Gα13 deficiency was verified by western blotting.

Note: Supplementary information is available on the Nature Medicine website.

ACKNOWLEDGEMENTSWe thank A. Rogatzki and A. Rippberger for expert technical help and B. Arnoldfor kind help with bone marrow transplantation experiments. This work wassupported by the Deutsche Forschungsgemeinschaft and the VolkswagenFoundation.

COMPETING INTERESTS STATEMENTThe authors declare that they have no competing financial interests.

Received 24 July; accepted 13 September 2003Published online at http://www.nature.com/naturemedicine/

1. Ruggeri, Z.M. Platelets in atherothrombosis. Nat. Med. 8, 1227–1234 (2002).2. Bhatt, D.L. & Topol, E.J. Scientific and therapeutic advances in antiplatelet therapy.

Nat. Rev. Drug Discov. 2, 15–28 (2003).3. Abrams, C.S. & Brass, L.F. Platelet signal transduction. in Hemostasis and Thrombosis

(eds. Colman, R.W., Hirsh, J., Marder, V.J., Clowes, A.W. & George, J.N.) 541–559(Lippincott William & Wilkins, Philadelphia, 2001).

4. Offermanns, S., Toombs, C.F., Hu, Y.-H. & Simon, M.I. Defective platelet activation inGαq deficient mice. Nature 389, 183–186 (1997).

5. Jin, J. & Kunapuli, S.P. Coactivation of two different G protein-coupled receptors isessential for ADP-induced platelet aggregation. Proc. Natl. Acad. Sci. USA 95,8070–8074 (1998).

6. Yang, J. et al. Signaling through Gi family members in platelets. Redundancy and speci-ficity in the regulation of adenylyl cyclase and other effectors. J. Biol. Chem. 277,46035–46042 (2002).

7. Klages, B., Brandt, U., Simon, M.I., Schultz, G. & Offermanns, S. Activation of G12/G13results in shape change and Rho/Rho-kinase-mediated myosin light chain phosphoryla-tion in mouse platelets. J. Cell Biol. 144, 745–754 (1999).

8. Offermanns, S., Laugwitz, K.-L., Spicher, K. & Schultz, G. G proteins of the G12 familyare activated via thromboxane A2 and thrombin receptors in human platelets. Proc.Natl. Acad. Sci. USA 91, 504–508 (1994).

9. Manganello, J.M., Djellas, Y., Borg, C., Antonakis, K. & Le Breton, G.C. Cyclic AMP-dependent phosphorylation of thromboxane A(2) receptor-associated Galpha(13). J. Biol. Chem. 274, 28003–28010 (1999).

10. Offermanns, S., Mancino, V., Revel, J.-P. & Simon, M.I. Vascular system defects and

impaired cell chemokinesis as a result of Gα13 deficiency. Science 275, 533–536(1997).

11. Gu, J.L., Müller, S., Mancino, V., Offermanns, S. & Simon, M.I. Interaction of Gα12 withGα13 and Gαq signaling pathways. Proc. Natl. Acad. Sci. USA 99, 9352–9357 (2002).

12. Lakso, M. et al. Efficient in vivo manipulation of mouse genomic sequences at thezygote stage. Proc. Natl. Acad. Sci. USA 93, 5860–5865 (1996).

13. Kühn, R., Schwenk, F., Aguet, M. & Rajewsky, K. Inducible gene targeting in mice.Science 269, 1427–1429 (1995).

14. Nieswandt, B., Schulte, V., Zywietz, A., Gratacap, M.-P. & Offermanns, S. Costimulationof Gi- and G12/G13-mediated signaling pathways induces integrin alpha IIbbeta 3 acti-vation in platelets. J. Biol. Chem. 277, 39493–39498 (2002).

15. Sambrano, G.R., Weiss, E.J., Zheng, Y.W., Huang, W. & Coughlin, S.R. Role of thrombinsignaling in platelets in hemostasis and thrombosis. Nature 413, 74–82 (2001).

16. Savage, B., Cattaneo, M. & Ruggeri, Z.M. Mechanisms of platelet aggregation. Curr.Opin. Hematol. 8, 270–276 (2001).

17. Sah, V.P., Seasholtz, T.M., Sagi, S.A. & Brown, J.H. The role of Rho in G protein-coupledreceptor signal transduction. Annu. Rev. Pharmacol. Toxicol. 40, 459–489 (2000).

18. Paul, B.Z., Daniel, J.L. & Kunapuli, S.P. Platelet shape change is mediated by both cal-cium-dependent and -independent signaling pathways. Role of p160 Rho-associatedcoiled-coil-containing protein kinase in platelet shape change. J. Biol. Chem. 274,28293–28300 (1999).

19. Bauer, M. et al. Dichotomous regulation of myosin phosphorylation and shape changeby Rho-kinase and calcium in intact human platelets. Blood 94,1665–1672 (1999).

20. Missy, K. et al. Rho-kinase is involved in the sustained phosphorylation of myosin andthe irreversible platelet aggregation induced by PAR1 activating peptide. Thromb.Haemost. 85, 514–520 (2001).

21. Schoenwaelder, S.M. et al. RhoA sustains integrin alpha IIbbeta 3 adhesion contactsunder high shear. J. Biol. Chem. 277, 14738–14746 (2002).

22. Jantzen, H.M., Milstone, D.S., Gousset, L., Conley, P.B. & Mortensen, R.M. Impairedactivation of murine platelets lacking G alpha(i2). J. Clin. Invest. 108, 477–483(2001).

23. Yang, J. et al. Loss of signaling through the G protein, Gz, results in abnormal plateletactivation and altered responses to psychoactive drugs. Proc. Natl. Acad. Sci. USA 97,9984–9989 (2000).

24. Dorsam, R.T., Kim, S., Jin, J. & Kunapuli, S.P. Coordinated signaling through bothG12/13 and G(i) pathways is sufficient to activate GPIIb/IIIa in human platelets. J. Biol.Chem. 277, 47588–47595 (2002).

25. Kuner, R., Swiercz, J., Zywietz, A., Tappe, A. & Offermanns, S. Characterization of theexpression of PDZ-RhoGEF, LARG and Gα12/Gα13 proteins in the murine nervous sys-tem. Eur. J. Neurosci. 16, 2333–2341 (2002).

26. Ren, X.D. & Schwartz, M.A. Determination of GTP loading on Rho. Methods Enzymol.325, 264–272 (2000).

27. Kuijpers, M.J. et al. Complementary roles of platelet glycoprotein VI and integrinalpha2beta1 in collagen-induced thrombus formation in flowing whole blood ex vivo.FASEB J. 17, 685–687 (2003).

28. Massberg, S. et al. A crucial role of glycoprotein VI for platelet recruitment to the injuredarterial wall in vivo. J. Exp. Med. 197, 41–49 (2003).

©20

03 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

rem

edic

ine