A Critical Role for Phospholipase Cγ2 in αIIbβ3-mediated Platelet Spreading*

The interaction of fibrinogen with the integrin αIIbβ3 plays a crucial role in platelet adhesion and platelet activation leading to the generation of intracellular signals that nucleate the reorganization of the cytoskeleton. Presently, we have only a limited understanding of the signaling cascades and effector proteins through which changes in the cytoskeletal architecture are mediated. The present study identifies phospholipase Cγ2 (PLCγ2) as an important target of the Src-dependent signaling cascade regulated by αIIbβ3. Real time phasecontrast microscopy is used to show that formation of filopodia and lamellapodia in murine platelets on a fibrinogen surface is dramatically inhibited in the absence of PLCγ2. Significantly, the formation of these structures is mediated by Ca2+ elevation and activation of protein kinase C, both directly regulated by PLC activity. With the involvement of Syk, SLP-76, and Btk, αIIbβ3-induced PLCγ2 activation partly overlaps with the pathway used by the collagen receptor glycoprotein VI. Important differences, however, exist between the two signaling cascades in that activation of PLCγ2 by αIIbβ3 is unaltered in murine platelets, which lack the FcR γ-chain or the adaptor LAT, but is abolished in the presence of cytochalasin D. Therefore, PLCγ2 plays not only a crucial role in activation of αIIbβ3 by collagen receptors but also in αIIbβ3-mediated responses.

␣ IIb ␤ 3 induces outside-in signals that lead to reorganization of the cytoskeleton and synergize with other agonists to mediate activation. The central role of the integrin ␣ IIb ␤ 3 in thrombosis and hemostasis is highlighted by the severe bleeding disorders in patients with Glanzmann thrombasthenia, which lack functional integrin.
One of the earlier events to occur following ligation of ␣ IIb ␤ 3 is the activation of the tyrosine kinase Syk via one or more Src kinases (1,2). This leads to tyrosine phosphorylation of the adaptor molecule SLP-76, which is constitutively associated with a second adaptor SLAP-130 (3,4), also known as Fynbinding protein or adhesion-and degranulation-promoting adapter protein (5,6). Together with proteins of the Vav GTPase exchange family, the adapter Nck, and the actin-binding protein VASP, this cascade has been shown to lead to activation of phosphoinositol 3-kinase and phosphorylation of FAK, and subsequent reorganization of the cytoskeleton in ␣ IIb ␤ 3 -transfected Chinese hamster ovary cells (7)(8)(9). Evidence that this cascade mediates reorganization of the cytoskeleton in platelets by ␣ IIb ␤ 3 has been provided using kinase inhibitors and murine cells deficient in Src kinases, Syk and SLP-76 (2,3,10).
Recently, we have shown that ␣ IIb ␤ 3 as well as the receptor for vWF, the glycoprotein (GP) Ib-IX-V complex, stimulate tyrosine phosphorylation of PLC␥2 (11,12). PLC␥2 is known to be a major target of signaling by the collagen receptor GPVI in platelets. GPVI exists in a complex with the Fc receptor ␥-chain (FcR ␥-chain), which contains one copy of an immunoreceptortyrosine-based activation motif (13). GPVI activates platelets through tyrosine phosphorylation of the FcR ␥-chain immunoreceptor tyrosine-based activation motif by the Src kinases Lyn and Fyn and recruitment of Syk (14 -16). Syk regulates a cascade that involves the adapters LAT, Gads, and SLP-76, the Tec family kinase Btk, and phosphatidylinositol 3-kinase (17)(18)(19). Functional activation of PLC␥2 downstream of this cascade is crucial and absolutely necessary for platelet activation by collagen (20). In contrast, phosphorylation of PLC␥2 downstream of GPIb-IX-V does not lead to a functional activation of the phospholipase suggesting that PLC␥2 activity is not required for GPIb-IX-V-mediated signals (12).
The role of PLC␥2 downstream of the fibrinogen receptor ␣ IIb ␤ 3 is not known. In the present study, we show that PLC␥2 is activated downstream of the fibrinogen receptor ␣ IIb ␤ 3 and that this plays a critical role in spreading through the mobilization of calcium and activation of protein kinase C. In addition, we also identify a number of additional proteins that are regulated downstream of the integrin-regulated signaling cascade but demonstrate important differences with the cascade used by GPVI. The present study expands the role of the PLC␥2 in platelet activation by demonstrating a central role in remod-eling of the cytoskeleton by immunoreceptor tyrosine-based activation motif and integrin receptors.

EXPERIMENTAL PROCEDURES
Antibodies and Reagents-PLC␥2 and anti-Syk polyclonal antibodies were from previously described sources (16). Polyclonal rabbit anti-FAK antibody was from Santa Cruz Biotechnology Inc., Santa Cruz, CA. Fibrinogen depleted of plasminogen and vWf were from Kordia Laboratory Supplies, Leiden, NL. 2-aminoethoxydiphenylborate (2-ABP) was from Tocris (Ellisville, MO), the PLC inhibitor U71322 and the PKC inhibitor Ro 31-8220 were from Calbiochem. All other reagents were from Sigma or previously named sources (11,19,21).
Animals-C57B1/6 mice deficient in FcR ␥-chain were obtained as previously described (22). Mice deficient in the adapter LAT were obtained as previously described and bred from heterozygotes on a B6 background (23). Mice deficient in PLC␥2 were generated as described (20) and bred from heterozygotes on a B6 background. Wild type littermates were used as controls.
Preparation of Human Platelets-Human blood was taken from drug-free volunteers on the day of the experiment and drawn into sodium citrate. Platelet-rich plasma (PRP) was obtained by centrifugation of the blood samples at 200 ϫ g for 20 min. Platelets were isolated from PRP by centrifugation at 800 ϫ g for 10 min in the presence of prostacyclin (0.1 g/ml). The pellet was resuspended in modified Tyrode's-HEPES buffer (134 mM NaCl, 0,34 mM Na 2 HPO 4 , 2,9 mM KCl, 12 mM NaHCO 3 , 20 mM HEPES, 5 mM glucose, 1 mM MgCl 2 , pH 7.3) containing 0.1 g/ml prostacyclin. The platelets were recentrifuged at 800 ϫ g for 10 min and resuspended at 5 ϫ 10 8 cells/ml in Tyrode's-HEPES buffer.
Preparation of Mouse Platelets-Blood (750 -1000 l) was taken into 100 l of acid citrate dextrose by cardiac puncture under terminal CO 2 narcosis. PRP was obtained by centrifugation of the blood samples at 300 ϫ g for 10 min at room temperature. PRP was centrifuged at 1000 ϫ g in the presence of prostacyclin (0.1 g/ml) for 6 min at room temperature. The pellet was resuspended in a modified Tyrode's-HEPES buffer to the required concentration and left for 30 min at room temperature prior to stimulation.
Adhesion Assays-Surfaces (Petri dishes, 6 well plates, coverslips) were coated with fibrinogen (200 g/ml) or BSA (5 mg/ml) overnight at 4°C. Surfaces were washed twice with PBS, blocked with denatured BSA (5 mg/ml) for 1 h, and washed again twice with PBS before use in the spreading experiments. Platelets did not adhere or become activated to surfaces coated with BSA.
To study tyrosine phosphorylation events in response to ligation to ␣ IIb ␤ 3 , platelets (5 ϫ 10 8 ) were incubated for 30 min in dishes coated with fibrinogen or BSA in the presence of 2 units/ml apyrase and 10 M indomethacin (24). Dishes coated with fibrinogen were washed twice with PBS to remove non-adherent cells. Platelets adherent to fibrinogen or in suspension over BSA were lysed in ice-cold immunoprecipitation buffer and lysates were subjected to immunoprecipitation assays or used directly for SDS-PAGE.
Proteins were separated by SDS-PAGE on 10% gels and electrically transferred onto polyvinylidene difluoride membranes. Membranes were blocked in 10% (w/v) BSA dissolved in TBS-T. Antibodies were diluted in TBS-T containing 2% (w/v) BSA and incubated with polyvinylidene difluoride membranes for 1 h at room temperature. Membranes were washed in TBS-T after each incubation and developed using an enhanced chemiluminescence system (ECL, Amersham Biosciences, Cardiff, United Kingdom).
Measurement of Platelet Cytosolic Ca 2ϩ Concentration-Platelets isolated from PRP were resuspended in modified Tyrode's-HEPES buffer to a concentration of 3 ϫ 10 8 cells/ml and incubated with Fura 2-acetoxymethylester (Fura 2-AM, 3 M, 1 h, 30°C) (Molecular Probes, Eugene, OR). After being washed in Tyrode's-HEPES buffer, platelets were resuspended at 2 ϫ 10 7 cells/ml in the presence of 2 units/ml apyrase and 10 M indomethacin. Platelets were allowed to spread on a fibrinogen-coated coverslip and real time calcium imaging was performed using Openlab software (Improvision, Coventry, UK).
Microscopy-Platelets (1.5 ϫ 10 7 in 0.5 ml of Tyrode's-HEPES buffer) were added to fibrinogen-coated coverslips and incubated for 30 min at 37°C. Non-adherent platelets were washed away and attached platelets were fixed with 3.7% paraformaldehyde for 10 min and permeabilized with 0.2% Triton X-100 in PBS. Platelets were stained for F-actin with rhodamine-conjugated phalloidin and viewed under an inverted fluorescence microscope using Openlab imaging software (Improvision).
Platelet spreading was observed in real time at 37°C and recorded using time-lapse laser scanning phase-contrast microscopy. The system consisted of a Bio-Rad Microradiance laser scanning instrument (Bio-Rad) attached to a Zeiss Axiovert inverted microscope equipped with a Solent Scientific environmental chamber; a Ph3 Plan-apochromat 63 ϫ 1.4 NA objective was used and the images were acquired using Lasersharp 2000 (version 4.2) software.
Determination of Phosphatidic Acid Production-Platelets were suspended in Tyrode's-HEPES without phosphate and were labeled with [ 32 P]orthophosphate (0.5 mCi/ml) for 1 h at 30°C. Platelets were washed twice and resuspended in Tyrode's buffer at a concentration of 5 ϫ 10 8 /ml. Platelets were incubated for 30 min in 6-well plates (5 ϫ 10 8 /well) coated with fibrinogen or over BSA in the presence of 2 units/ml apyrase and 10 M indomethacin. Dishes coated with fibrinogen were washed twice with PBS to remove non-adherent cells. Platelets adherent to fibrinogen or in suspension over BSA were lysed in a buffer containing 100 mM EDTA, 5 N HCl, and 1% Nonidet P-40. Phospholipids were extracted by addition of 400 l of chloroform-methanol-HCl (100/200/1, v/v/v), and [ 32 P]phosphatidic acid was separated by thin-layer chromatography. Thin-layer chromatography plates were exposed to Kodak PhosphorScreen and phosphatidic acid signals were quantified using Molecular Imager FX and Quantity One Software version 4 for Macintosh (Bio-Rad) Statistical analyses were performed using Student's t test.

RESULTS
Differential Effects of Cytochalasin D on ␣ IIb ␤ 3 and GPVIinduced Phosphorylation-It has been shown previously that tyrosine phosphorylation by ␣ IIb ␤ 3 is regulated through activation of a Src family kinase and that it is modulated by disruption of actin polymerization using cytochalasin D (2). Thus ␣ IIb ␤ 3 -mediated tyrosine phosphorylation events can be divided into those that are mediated upstream and downstream of actin polymerization. We have compared the effect of the actin polymerization inhibitor cytochalasin D and the Src kinase inhibitor PP2 on signals mediated by ␣ IIb ␤ 3 and GPVI. Stimulation of ␣ IIb ␤ 3 was achieved by incubation of platelets over a fibrinogen-coated matrix for 30 min. GPVI was activated with the snake toxin convulxin under stirring conditions in suspension using an incubation time of 30 s, a time point known to be the peak for tyrosine phosphorylation induced by this agonist. ␣ IIb ␤ 3 and GPVI induce overlapping patterns of increases in tyrosine phosphorylation as measured using whole cell lysates. Both stimuli induced marked increases in tyrosine-phosphorylated bands of 70 -80 kDa, which co-migrate with Syk, Btk, and SLP-76. The constitutively phosphorylated bands between 50 and 60 kDa co-migrate with Src family kinases. There is also a marked increase in a band of 130 kDa and a number of minor bands around this molecular mass in response to the two agonists. In contrast, a doublet of tyrosine-phosphorylated proteins with a molecular mass between 28 kDa and 32 kDa is regulated downstream of ␣ IIb ␤ 3 but not by GPVI (Fig. 1A, arrows). Convulxin stimulates marked tyrosine phosphorylation of proteins of 36 and 12 kDa, which co-migrate with LAT and the FcR ␥-chain, respectively (Fig. 1A, arrows). Both signaling cascades are strongly dependent on the presence of functionally active Src kinases. Incubation of platelets with 20 M PP2 causes a complete loss of inducible tyrosine phosphorylation following convulxin stimulation and markedly inhibits the response to fibrinogen, although tyrosine phosphorylation of a 130-kDa band is preserved (Fig. 1A, arrow). Cytochalasin D has a distinct effect on the degree of phosphorylation by ␣ IIb ␤ 3 and GPVI. Whereas the response to GPVI is only slightly inhibited by 10 M cytochalasin D, there is a marked reduction in tyrosine phosphorylation of all bands by ␣ IIb ␤ 3 showing that it has a much greater dependence on the reorganization of the cytoskeleton (Fig. 1A).
As a prelude to comparing the signaling cascades regulated Dishes coated with fibrinogen were washed twice with PBS to remove non-adherent cells. Platelets adherent to fibrinogen or in suspension over BSA were lysed in ice-cold immunoprecipitation buffer and subjected to immunoprecipitation for the indicated proteins or used directly for SDS-PAGE and immunoblotted for tyrosine-phosphorylated proteins. Samples were adjusted prior to immunoprecipitations to a similar number of cells in each group. For convulxin (14 nM) platelets in solution were stimulated in the presence of the indicated inhibitors for 30 s and lysates were subjected to immunoprecipitation for the indicated proteins. Results are representative of three experiments. by ␣ IIb ␤ 3 and GPVI, we investigated some of the proteins that undergo increases in tyrosine phosphorylation following incubation of platelets over a fibrinogen-coated matrix for 30 min through immunoprecipitation with specific antibodies and Western blotting with the antiphosphotyrosine antibody, 4G10.
In agreement with studies from other groups, platelet spreading over fibrinogen leads to a substantial phosphorylation of the integrin ␤ 3 subunit (Fig. 1C), tyrosine kinase Syk, focal adhesion kinase (FAK) (Fig. 1B), adapter proteins SLP-76 (Fig.  1C) and SLAP-130 as well as the guanine nucleotide exchange factor Vav1 and the ubiquitin-regulator c-Cbl (not shown). In addition, we identified a number of proteins that have been previously shown to be phosphorylated downstream of GPVI, namely the FcR ␥-chain, the Tec family tyrosine kinase Btk (Fig. 1C), as well as PLC␥2 (Fig. 1B). The adapter molecules Grb2 and Gads were also recruited to signaling complexes downstream of ␣ IIb ␤ 3 and formed associations with unidentified tyrosine-phosphorylated proteins of 32, 55, and 150 kDa, and 50, 76, and 130 kDa, respectively (not shown). There was, however, one important omission in the proteins that undergo tyrosine phosphorylation downstream of ␣ IIb ␤ 3 relative to GPVI, namely the adapter LAT. Whereas LAT is one of the major tyrosine-phosphorylated proteins downstream of GPVI we were not able to detect an increase in LAT phosphorylation after ␣ IIb ␤ 3 stimulation (Fig. 1B). These data demonstrate that platelet adhesion to fibrinogen leads to tyrosine phosphorylation of PLC␥2 and identifies many similarities but also impor-tant differences with the events regulated downstream of GPVI.
The effect of cytochalasin D and PP2 on phosphorylation of Syk, LAT, FAK, and PLC␥2 by ␣ IIb ␤ 3 and GPVI was explored following their immunoprecipitation and Western blotting for phosphotyrosine. Tyrosine phosphorylation of Syk induced by ␣ IIb ␤ 3 and convulxin was reduced in the presence of cytochalasin D (Fig. 1B) and abolished in the presence of PP2 (not shown). In contrast, cytochalasin D had a differential effect on tyrosine phosphorylation of a band of 12 kDa, which co-precipitates with Syk and was shown through immunoprecipitation studies to be the FcR ␥-chain (not shown). Tyrosine phosphorylation of FcR ␥-chain by ␣ IIb ␤ 3 was abolished by treatment with cytochalasin D, whereas it was only slightly inhibited in convulxin-stimulated cells (Fig. 1B, arrows). This demonstrates that tyrosine phosphorylation of FcR ␥-chain is dependent on actin polymerization in response to ␣ IIb ␤ 3 . As already discussed, ␣ IIb ␤ 3 does not cause phosphorylation of LAT. In contrast LAT is strongly phosphorylated downstream of GPVI and this phosphorylation is reduced following cytochalasin D treatment (Fig. 1B) and abolished after inhibition of Src kinases with PP2 (not shown). In contrast ␣ IIb ␤ 3 stimulation induces a robust, cytochalasin D-sensitive phosphorylation of FAK, whereas convulxin stimulation only causes minimal tyrosine phosphorylation of FAK (Fig. 1B) even at times up to 30 min (not shown). Cytochalasin D had a differential effect on the regulation of tyrosine phosphorylation of PLC␥2 by ␣ IIb ␤ 3 and GPVI. Tyrosine phosphorylation of PLC␥2 by ␣ IIb ␤ 3 was almost abolished in the presence of cytochalasin D, whereas the response to convulxin was inhibited by ϳ50% (Fig. 1B). PP2 completely blocked tyrosine phosphorylation of PLC␥2 by both receptors (not shown).
The differential effect of cytochalasin D on tyrosine phosphorylation of FcR ␥-chain and PLC␥2 by ␣ IIb ␤ 3 and GPVI further distinguishes the two signaling cascades. Thus, activation of Syk by ␣ IIb ␤ 3 is independent of phosphorylation of FcR ␥-chain and signals from the kinase are not mediated through tyrosine phosphorylation of LAT. In contrast, Syk is regulated downstream of FcR ␥-chain by GPVI (25) and signals are associated with robust phosphorylation of LAT. The differential role of FcR ␥-chain in these two pathways agrees with previous work from the Shattil group (26) showing that activation of Syk by ␣ IIb ␤ 3 is through a Src kinase-dependent pathway but does not require prior phosphorylation of an immunoreceptor tyrosinebased activation motif-containing protein.
Phosphorylation of PLC␥2 Downstream of ␣ IIb ␤ 3 Is Independent of the Adapter LAT and Fc Receptor ␥-Chain-Previously, we have reported that tyrosine phosphorylation of Syk and PLC␥2 through GPVI is abolished in FcR ␥-chain-deficient murine platelets, whereas tyrosine phosphorylation of PLC␥2 is markedly reduced in LAT-deficient cells (21,25). We have now investigated the role of these two proteins in tyrosine phosphorylation of Syk and PLC␥2 by ␣ IIb ␤ 3 . Adhesion of murine platelets to fibrinogen induced a marked increase in tyrosine phosphorylation of Syk and PLC␥2, similar to that seen in human platelets. This increase in phosphorylation was not altered in platelets deficient in FcR ␥-chain or LAT (Fig. 2). Similarly, tyrosine phosphorylation of the adapter SLP-76 by fibrinogen was unaltered in platelets deficient in FcR ␥-chain or LAT (Fig. 2B). As with human platelets, FcR ␥-chain coprecipitated with Syk upon stimulation by ␣ IIb ␤ 3 . This association was preserved in LAT-deficient platelets ( Fig. 2A, arrow). These results confirm that neither FcR ␥-chain nor LAT are required for tyrosine phosphorylation of Syk and PLC␥2 by ␣ IIb ␤ 3 in contrast to signals from GPVI. ␣ IIb ␤ 3 Stimulates Activation of PLC␥2, Calcium Elevation, and Spreading-The functional consequence of PLC␥2 phosphorylation downstream of ␣ IIb ␤ 3 was investigated by measurement of phosphatidic acid and intracellular calcium, two indirect markers of PLC activity. Platelets spread on a fibrinogen-coated surface had a 2.4 Ϯ 0.1-fold increase in the production of phosphatidic acid relative to cells exposed to a BSAcoated surface (p ϭ 0.005, Fig. 3A). The increase in phosphatidic acid was inhibited strongly in the presence of cytochalasin D (1.5 Ϯ 0.1-fold) and reduced to basal levels using the PLC inhibitor U-73122 (0.96 Ϯ 0.13-fold). PP2 also caused a complete inhibition of phosphatidic acid accumulation, reducing the level below that of cells exposed to BSA (0.57 Ϯ 0.11fold), strongly suggesting that the increase is mediated via PLC␥ rather than PLC␤ isoforms (Fig. 3A).
To further clarify the role of PLC␥2 downstream of ␣ IIb ␤ 3 we determined the production of phosphatidic acid in PLC␥2-deficient murine platelets. As seen in human platelets, spreading of murine platelets on fibrinogen caused an increase in the production of phosphatidic acid (1.5 Ϯ 0.08-fold) relative to cells exposed to a BSA-coated surface (1.0 Ϯ 0.05-fold). This increase was inhibited following cytochalasin D treatment (0.90 Ϯ 0.12-fold) and in PLC␥2-(1.13 Ϯ 0.10-fold) deficient platelets (Fig. 3B). These data suggest that PLC␥2 is the main enzyme responsible for the increase in phosphatidic acid following spreading of platelets on fibrinogen.
The increase in formation of phosphatidic acid was associated with a sustained increase in intracellular calcium, as measured using the calcium reporter Fura 2 (delivered as Fura 2-AM). All of the platelets that had undergone partial or complete spreading on fibrinogen had elevated levels of intracellular calcium. Moreover, the increase in calcium and spreading (measured by monitoring fluorescence at 380 nm) occurred in parallel as illustrated in Fig. 4B suggesting that they are causally related. The calcium increase was observed 2-3 min after the initial contact with the fibrinogen-coated surface and is accompanied by the formation of filopodia and lamellipodia. The increase in intracellular calcium and spreading were substantially inhibited by cytochalasin D and PP2 although they had only minimal effects on attachment of the platelets to the fibrinogen-coated surfaces (Fig. 4A). Calcium elevation and spreading was also strongly inhibited by 2-ABP an inositol 1,4,5-trisphosphate receptor antagonist. The increase in intracellular calcium was specific for ␣ IIb ␤ 3 because it was absent in platelets from a patient diagnosed with Type III Glanzmann thrombasthenia, which express a non-functional ␣ IIb ␤ 3 . Importantly, the platelets from this patient were able to adhere to the fibrinogen-coated surface but spreading and calcium mobilization were almost completely inhibited (Fig. 4C). In contrast, these platelets spread and flux calcium normally on a collagencoated surface (not shown).
The spreading of platelets on fibrinogen was further investigated by staining of the actin cytoskeleton using rhodaminelabeled phalloidin. In agreement with previous reports, human platelets are able to spread on fibrinogen with the formation of filopodia, lamellipodia, and limited stress fibers (Fig. 5A). Spreading is enhanced by the addition of ADP resulting in most of the platelets forming stress fibers (Fig. 5B). Spreading and formation of stress fibers was completely inhibited in the presence of cytochalasin D, BAPTA-AM (which chelates intracellular calcium), U71322, and 2-ABP (Fig. 5, E-G and I). PP2 and the PKC inhibitor Ro 31-8220 also markedly inhibited spreading although a limited formation of filopodia was preserved (Fig. 5, C  and K). The partial spreading in the presence of PP2 does not appear to be because of an incomplete inhibition of Src kinases as Platelets were allowed to adhere and spread to the fibrinogen-coated coverslip and calcium imaging was performed using an inverted fluorescence microscope and Openlab software. B, platelets and coverslips were prepared as described above and real time spreading and calcium mobilization of a Fura 2-AM-loaded platelet in response to fibrinogen was performed using an inverted fluorescence microscope and Openlab software. C, platelets from a patient diagnosed with a type III Glanzman thrombasthenia were loaded with Fura 2-AM and placed onto coverslips coated with fibrinogen in the presence of 2 units/ml apyrase and 10 M indomethacin. Platelets were allowed to adhere to the fibrinogen-coated coverslip and calcium imaging was performed using an inverted fluorescence microscope and Openlab software. Platelets are shown in the bright field image (a) or as a 340/380 nm ratio indicating the concentration of intracellular calcium (b).
higher concentrations of PP2 (60 M) had a similar effect (not shown). The inhibitory action of PP2 and Ro 31-8220 was overcome by the addition of ADP (Fig. 5, D and L). Interestingly, this is not the case for inhibition by the general PLC inhibitor U71322 (Fig. 5H), indicating that activation of at least one PLC isoform is crucial for spreading on fibrinogen. Consistent with our previous studies on ␣ IIb ␤ 3 signaling (11) depletion of cholesterol using methyl-␤-cyclodextrin had no effect on spreading (Fig. 5J).
To investigate the role of PLC␥2 in spreading on fibrinogen we used murine platelets deficient in the phospholipase. Murine platelets, however, undergo a more limited degree of spreading on fibrinogen relative to human platelets. The average surface area of a murine platelet on fibrinogen, as measured with the actin stain phalloidin, increases by 55% from 2.00 Ϯ 0.04 m 2 (n ϭ 318) in the presence of BAPTA-AM to 3.11 Ϯ 0.06 m 2 (n ϭ 378). In contrast human platelets increase their size by 335% from 2.6 Ϯ 0.13 m 2 (n ϭ 83) in the presence of BAPTA-AM to 8.56 Ϯ 0.60 m 2 (n ϭ 103) under the same conditions. Importantly, a larger proportion of the PLC␥2-deficient platelets had not spread on fibrinogen relative to control cells, although the difference in surface area of the population was not statistical significant possibly because of the relative small change in overall size (Fig. 6). Nevertheless, the ability of ADP to induce full spreading of the wild type platelets was reduced by around 10% in the PLC␥2 Ϫ/Ϫ cells relative to the controls (surface area: 7.91 Ϯ 0.55 m 2 ,n ϭ 209; versus 8.77 Ϯ 0.66 m 2 , n ϭ 177; p ϭ 0.008).
Interestingly, during the course of this work, we noticed that the proportion of non-spread platelets in the PLC␥2 Ϫ/Ϫ population relative to the controls increased with time suggesting a kinetic difference between the two populations (not shown). To investigate this further, we measured spreading on fibrinogen using time-lapse laser scanning phase-contrast microscopy. With this method we were able to observe spreading over time and to visualize structures that could not be readily detected with the actin-based stain. Control platelets were seen to rapidly (within 1 min) form elongated filopodia on fibrinogen, which are far greater in length than those observed in phalloidin-stained platelets (Fig. 7). This was accompanied by the partial formation of lamellipodia. The change in morphology is dynamic with a clear synthesis and retraction of filopodia and lamellipodia continuing to occur over a course of 15-30 min before reaching an end stage consisting of a partial spread central core and a wide network of filopodia (Fig. 7). In comparison, the majority of the PLC␥2 Ϫ/Ϫ platelets underwent a far less pronounced change in morphology. A far smaller number of filopodia was seen at early times and this was accompanied by a very limited formation of lamellipodia in ϳ50% of the population. Moreover, in contrast to the wild type platelets, these structures were not maintained and the lamellapodia disappeared within a few minutes. The filopodia were also seen to either disappear within a few minutes or to change to a small network of very thin fibers that could barely be detected at the level of the light microscope (Fig. 7). In comparison, spreading of cells deficient in the adapter LAT on fibrinogen was indistinguishable from that of wild type platelets (not shown). These results demonstrate a critical role for PLC␥2 but not LAT in spreading on fibrinogen. DISCUSSION In this study we show that engagement of ␣ IIb ␤ 3 by fibrinogen causes activation of PLC␥2 and increase in intracellular Ca 2ϩ and that these events are crucial for platelet spreading. We also show that although the regulation of PLC␥2 by ␣ IIb ␤ 3 has many similarities with the signaling events downstream of the collagen receptor GPVI, the two pathways can be distinguished on the role of key signaling proteins. Signaling by ␣ IIb ␤ 3 is dependent on Src and Syk family kinases and leads to cytoskeletal reorganization followed by a marked phosphorylation of FAK (2,27), but is independent of the glycolipidenriched membrane domain-associated adapter LAT and the FcR ␥-chain ( Figs. 1 and 2). In addition, ␣ IIb ␤ 3 can signal via a Src kinase-independent pathway, which is associated with tyrosine phosphorylation of an unidentified protein of about 130 kDa and limited spreading. The present study therefore identifies a novel pathway of activation of PLC␥2 by ␣ IIb ␤ 3 that is dependent on cytoskeletal reorganization and mediates platelet spreading. Dissecting out the full sequence of signaling events downstream of ␣ IIb ␤ 3 will be challenging in that not only does the integrin induce cytoskeletal reorganization, its signaling cascade is also strongly regulated by this event.
GPVI signaling is also dependent on Src and Syk family kinases, but its signaling cascade takes place mainly in glycolipid-enriched membrane domains and is mediated through a pathway that utilizes LAT and the FcR ␥-chain (11,13,16,21). In addition, GPVI signaling is associated with minimal FAK phosphorylation. It is important to note, however, that GPVI signaling is also partly dependent on cytoskeletal rearrangements in that is it inhibited partially in the presence of cytochalasin D. It is possible that cytoskeletal rearrangements contribute to GPVl clustering thereby enhancing the functional outcome or that they reinforce activation independent of clustering.
It has been shown that Src is constitutively associated with ␣ IIb ␤ 3 and plays a key role in the activation of Syk following fibrinogen binding (2). This initiates downstream processes such as phosphorylation of the adapter molecules SLP-76 and SLAP-130, activation of Vav family proteins, as well as phosphorylation of the kinase FAK (1, 3). Distal ␣ IIb ␤ 3 signaling events, e.g. FAK phosphorylation, have been shown to be heavily dependent on reorganization of the cytoskeleton and can be FIG. 7. Real time imaging of platelet spreading on fibrinogen. Platelets were exposed to the fibrinogen-coated surface in Tyrode's buffer and spreading was observed in real time at 37°C using time-lapse laser scanning phase-contrast microscopy using a Zeiss Axiovert inverted microscope with a Ph3 Plan-APOCHROMAT 63/1.4 oil objective and Laser Sharp 2000 version 4.2 for windows software (Bio-Rad). A, representative time course of a single platelet spreading on fibrinogen: a, platelet from a wild type mouse; b, platelet from a PLC␥2-deficient mouse. B, spreading of mouse platelets after 30 min on a fibrinogen-coated surface: a, platelets from a wild type mouse; b, platelets from a PLC␥2-deficient mouse.
prevented by inhibitors of actin polymerization such as cytochalasin D (2). Similar to FAK phosphorylation, PLC␥2 phosphorylation is also sensitive to cytochalasin D, identifying this as a late event in the ␣ IIb ␤ 3 signaling cascade. Interestingly, FAK is implicated in the regulation of PLC␥1 in COS-7 cells and a mouse fibroblast cell line in response to integrin activation. In these cells FAK and PLC␥1 interact directly following autophosphorylation of Tyr-397 of the kinase and it is suggested that this interaction recruits PLC␥1 to the membrane and increases the enzymatic activity of the kinase (28). A similar interaction between FAK and PLC␥2 may take place in platelets following ␣ IIb ␤ 3 activation, and would provide a route of recruitment to the membrane that is independent of LAT phosphorylation as FAK is associated with the ␤ 3 subunit of the integrin in platelets (2).
Although tyrosine phosphorylation of PLC␥2 does regulate activation of the enzyme (29), this does not necessarily lead to a functional activity as shown downstream of the platelet vWF receptor GPIb-IX-V (12). The activity is also dependent on substrate availability and membrane recruitment and it is therefore crucial to identify the functional activity of the phospholipase. Fibrinogen stimulation caused an increase in the production of phosphatidic acid and intracellular calcium through pathways that were inhibited strongly by cytochalasin D and completely by the Src kinase inhibitor PP2 and U71322 (Figs. 3 and 4), an inhibitor of all PLC isoforms (30). The sensitivity to PP2 strongly suggests that the increase in phosphatidic acid is mediated by a PLC␥ isoform rather than another isotype. This was supported by the observation that mice deficient in PLC␥2 also showed decreased production of phosphatidic acid. The lack of complete inhibition of phosphatidic acid formation in these studies could reflect the fact that mouse platelets express a low level of PLC␥1 (31). These data indicate that fibrinogen stimulation triggers functional activation of PLC␥2 downstream of the platelet integrin ␣ IIb ␤ 3 . Activation of PLC isoforms and its importance in integrin-mediated Ca 2ϩ signaling is a well established phenomenon and has been observed in various cells expressing different integrins (32)(33)(34)(35)(36) including platelets activated by the ␣ 2 ␤ 1 -specific peptide GFOGER (37).
Functional activation of isoforms of phospholipase C results in the formation of the second messengers inositol 1,4,5trisphosphate and 1,2-diacylglycerol leading to a mobilization of intracellular calcium and activation of PKC isoforms, respectively. Calcium and PKC have been shown to play major roles in mediating many of the responses associated with platelet activation including secretion and aggregation. Spreading on fibrinogen leads to an elevation of intracellular calcium and is inhibited by treatment with the calcium chelator BAPTA-AM, the inositol 1,4,5-trisphosphate receptor antagonist 2-ABP, and the PKC inhibitor Ro 31-8220 (Figs. 4 and 5). A role for calcium in ␣ IIb ␤ 3 -mediated spreading under static conditions has previously been shown on a vWf-coated surface although it is likely that this reflects a role in integrin activation (38). Several studies have linked ␣ IIb ␤ 3 engagement with calcium elevation at high shear. It is suggested that under flow conditions, the interaction of vWf with GPIb-V-IX induces transient calcium signals promoting ␣ IIb ␤ 3 activation and platelet arrest. Subsequent ␣ IIb ␤ 3 engagement induces intracellular calcium mobilization followed by transmembrane calcium influx. The integrin calcium response is sustained and necessary for irreversible adhesion (39 -41). These results indicate a crucial role of both PLC-regulated second messengers for ␣ IIb ␤ 3 -mediated responses.
Significantly, ADP is able to overcome the effect of the Src family kinase inhibitor PP2 on spreading of human platelets, whereas this is not the case in the presence of the nonspecific inhibitor of PLC isoforms, U71322 or the calcium chelator BAPTA-AM (Fig. 5). This indicates that activation of a PLC isoform is necessary for platelet spreading on fibrinogen but that the increase in lipase activity could be mediated through activation of PLC␥2 by fibrinogen or PLC␤ by the P2Y 1 ADP receptor.
A dramatic reduction in spreading was observed in the PLC␥2Ϫ/Ϫ mice platelets using time-lapse laser scanning phase-contrast microscopy (Fig. 7). These results demonstrate that the activation of PLC␥2 is required for the formation of filopodia and lamellipodia. The observation that the filopodia and the limited formation of lamellipodia is a highly dynamic process and that this leads to continuous changes in morphology in wild type but not PLC␥2Ϫ/Ϫ platelets over a course of 15-30 min suggests that this is mediated by sustained activation of the phospholipase. We speculate that the transient formation of filopodia and spreading in PLC␥2 Ϫ/Ϫ platelets is mediated by a limited and non-sustained mobilization of calcium, possibly mediated by activation of the limited amount of PLC␥1 that is also present in murine platelets (31) as spreading is completely inhibited following chelation of intracellular calcium or by inhibition of protein kinase C. Further work is required to compare the dynamics of calcium elevation and spreading under these conditions. Importantly, and consistent with the phosphorylation results, spreading on fibrinogen in LAT Ϫ/Ϫ platelets was not altered.
There is increasing evidence that PLC␥2 plays a central role in many aspects of platelet activation. It is well established that the phospholipase is the main target of signaling cascades downstream of Fc␥RIIA and the collagen receptor GPVI. PLC␥2 is necessary for the activating function of these receptors and ultimately causes activation of the integrin ␣ IIb ␤ 3 , which is crucial for platelet adhesion and thrombus formation. The present study shows that PLC␥2 also plays a critical role downstream of the integrin ␣ IIb ␤ 3 thereby providing a new concept of regulation of the phospholipase in platelets. This activation is essential for functional responses mediated by the integrin. Spreading of platelets on fibrinogen is dependent on activation of PKC and mobilization of cytosolic calcium demonstrating a critical role for both arms of the lipase-based signaling cascade. Interestingly, despite substantial similarities between the pathways of GPVI and ␣ IIb ␤ 3 stimulated PLC␥2 activation, the two signaling cascades can be distinguished by their differential usage of specific signaling molecules, the requirement of cytoskeletal rearrangements, and the spatial organization. Further investigations will reveal if ␣ IIb ␤ 3 -mediated PLC␥2 activation is involved in other late events in platelet function such as thrombus stability and clot retraction.