A Shear-restricted Pathway of Platelet Procoagulant Activity Is Regulated by IQGAP1*

Circulating blood platelets regulate the initial phase of the hemostatic response through adhesive and aggregatory events and by providing the necessary procoagulant surface for prothrombinase complex assembly and thrombin generation. The signaling pathway(s) that regulate platelet procoagulant activity are largely unknown, although they are distinct from platelet aggregatory signals linked to fibrinogen ligation to the conformationally active αIIBβ3 integrin. We describe a novel intracellular signaling mechanism involving platelet IQGAP1 that specifically regulates the development of platelet procoagulant activity under conditions of mechanical shear stress. Murine platelets that are deficient in IQGAP1 demonstrate increased prothrombinase activity compared with wild-type littermate controls when activated by a physiological shear stress of 16 dynes/cm2 (shear rates of 1600 s-1) (p < 0.0001), corresponding to ∼2.5 times the normal shear stress, or ∼40% degree of stenosis in coronary arteries. The exaggerated prothrombinase activity is not associated with enhanced platelet microvesiculation (cytoskeletal proteolysis) and occurs independently of the intracellular calcium release, [Ca2+]i, but it is specifically coupled to the α-granule exocytic pathway without concomitant effects on aminophospholipid exposure. These observations identify platelet IQGAP1 as an important modulator of normal hemostasis and as an appropriate pharmacological target for control of platelet procoagulant function.

The hemostatically neutral surface of quiescent platelets is preserved by an aminophospholipid translocase that maintains phosphatidylserine (PS) 1 and phosphatidylethanolamine within the inner platelet leaflet and phosphatidylcholine on the outer leaflet. This asymmetry is abolished by thrombin, collagen, and ionophores, presumably related to changes in the intracytosolic calcium concentration, [Ca 2ϩ ] i . The rapid loss of membrane asymmetry is mediated by a calcium-dependent phospholipid scramblase (with inhibition of the translocase), resulting in rapid bidirectional movement of phospholipids, the appearance of procoagulant PS, and an ϳ6-log acceleration of ␣-thrombin generation via the assembly of cell surface prothrombinase (1,2). The importance of the sudden reversal of this membrane phospholipid in the maintenance of normal hemostasis is best exemplified by Scott's syndrome. This syndrome is a rare bleeding disorder caused by an uncharacterized molecular defect associated with a platelet procoagulant deficiency, which is related to defective membrane PS movement (3). Nonetheless, the aggregatory potential of platelet agonists (mediated by fibrinogen ligation to the conformationally active ␣ IIb ␤ 3 integrin) is distinct from that linked to exposure of the aminophospholipids PS and phosphatidylethanolamine. In thrombin-activated human platelets, for example, PS exposure is mediated by protease-activated receptor-1 (PAR1), with little or no contribution from PAR4 (2,4).
Platelet microvesiculation (generation of platelet microparticles (PMP)) is closely associated with the exposure of anionic PS on the outer surface of the platelet, thereby considerably enhancing the surface area for activation of the coagulation cascade. PMPs can be generated from high shear stress and are able to cross-activate platelets and endothelial cells, and they are elevated in patients with increased thromboembolic risk and in the subsets of those patients undergoing coronary artery bypass grafting (1,2). A causal role for PMPs in thrombosis, however, has been difficult to establish because platelet microvesiculation is intimately associated with platelet activation and aggregation. Although the molecular basis of microvesiculation remains poorly characterized, it appears to be calciumdependent, requires cytoskeletal reorganization, is associated with calpain and caspase-3 activation, and is morphologically similar to the membrane blebbing and phospholipid exposure phase of nucleated cell apoptosis (5). The molecular mechanism of PMP generation remains unknown, although the molecular pathway appears to be distinct from that regulating PS exposure or platelet exocytosis and is associated with cytoskeletal proteolysis (1)(2)(3).
Mammalian IQGAPs are conserved homologues of an extended family of proteins; their counterparts in yeast (Saccharomyces cerevisiae Iqg1p/Cyk1p) and amoebae (Dictyostelium discoideum DdGAP1) have key roles in actomyosin ring formation and cytokinesis. Two homologous IQGAPs have been identified and characterized, although expression of IQGAP1 appears broad compared with the previously described hepatocyte-restricted distribution of IQGAP2 (6,7). IQGAPs are multidomain scaffolding proteins that presumably integrate intracellular signals with actin polymerization by interactions with the Rho GTPases rac1 and cdc42 (through Ras-GAP-related homology domains, or GRD), Ca 2ϩ /calmodulin (through IQ motifs), and F-actin (through calponin homology domains). Inter-actions with E-cadherin (8), ␤-catenin (9), and the microtubular protein CLIP-170 (10) have been described previously, although the in vivo physiological significance of these interactions and the regulatory mechanisms leading to IQGAP activation remain incompletely characterized. Our data now link IQGAP1 to a shear stress-restricted (mechanoreceptor) signaling pathway of platelet activation, specifically regulating the procoagulant properties of the platelet membrane.

MATERIALS AND METHODS
Murine Functional Studies-All animal studies were completed in accordance with approved Institutional Animal Care and Use Committee protocols at the State University of New York, Stony Brook. mIQGAP1 Ϫ/Ϫ mice and inbred wild-type littermate controls were expanded and maintained on an SV129 background and genotyped by PCR (11). Blood was removed by retro-orbital puncture into 0.1 volume of acid/citrate/dextrose supplemented with 0.1 M prostaglandin E 1 , and gel-filtered platelets (GFP) were isolated over a Sepharose 2B column as described previously (12). Protein-solubilized lysates were generated from liquid nitrogen snap-frozen livers or GFP (1 ϫ 10 9 platelets), and immunoblot analysis was completed by 4 -15% SDS-PAGE using anti-IQGAP1 (1:500 dilution) (Transduction Laboratories) or anti-IQGAP2 (1:1000) (Upstate Biotechnologies) antibodies, which were detected by enhanced chemiluminescence (Amersham Biosciences). Complete hemograms were determined using automated Coulter counters, and the F-actin content of GFP was determined by flow cytometry using fixed and permeabilized platelets incubated with 30 units/ml FITC-phalloidin (13). Transmission electron microscopy was completed using 1 ϫ 10 9 GFP fixed in 2.5% (v/v) glutaraldehyde and 2% (w/v) paraformaldehyde. After postfixation in a freshly prepared solution of 1% (w/v) osmic acid, alcohol-dehydrated samples were pelleted and embedded in an Epon-812 resin mixture, and thin sections were examined in a blinded fashion.
Platelet calmodulin activity assays were determined from plateletrich plasma prepared in 0.5 M prostaglandin E 1 as described previously (14). In brief, platelets were resuspended at 3 ϫ 10 8 platelets/ml in 25 mM HEPES, pH 7.5, 150 mM NaCl, and 1 mM CaCl 2 and freezethawed twice, and supernatants were isolated after centrifugation. Assays were completed in duplicate using calmodulin-dependent chicken gizzard phosphodiesterase for cyclic AMP generation, and determinations were made from the linear portion of the standard curve using pure porcine testis calmodulin as the standard.
Flow Cytometric Analyses-FITC-conjugated anti-mouse CD41 monoclonal antibody (integrin ␣ IIb ␤ 3 chain) and FITC-conjugated anti-CD62 (P-selectin) were purchased from Pharmingen. CD41 was used for delineating platelet gates, and debris and dead cells were excluded using scatter gates. PS exposure of intact and microvesiculated platelets was quantified by annexin-V binding. Briefly, appropriately treated platelets were incubated with 0.0125 g of phycoerythrin-conjugated annexin V in the presence of 2 mM CaCl 2 for 30 min at 25°C in the dark followed by FACScan analysis using logarithmic gain settings for light scatter and fluorescence (BD Biosciences). For all studies, 10,000 gated events were acquired; forward and side scatter and fluorochrome binding were used for delineation and quantification of platelets and platelet-derived microparticles.
Prothrombinase Determinations and Shear Activation-Quantitative assessment of phosphatidylserine exposure was completed using both annexin-V binding (see above) and a modified prothrombinase assay using acetylated prothrombin (15,16). Acetylated prothrombin retains Ͻ0.1% clotting activity (and is thereby defective in platelet activation) but is activable by factor Xa in the presence of platelet factor Va to produce thrombin activity measured by cleavage of the chromogenic substrate tosyl-Gly-Pro-Arg-p-nitroanilide (Chromozym-TH; Roche Applied Science). Briefly, 0.1-ml samples of GFP (4 ϫ 10 4 /l, subjected to various agonists or exposure to shear stress) in Ca 2ϩ -free HBMT were incubated with 200 nM acetylated prothrombin, 5 mM Ca 2ϩ , 100 pM factor Xa, and individual agonists for 10 min at 37°C. Then, duplicate 20-l aliquots were assayed, using 0.3 mM Chromozyme-TH for thrombin generation, by reading in a microplate reader. All samples were individually normalized to parallel determinations using the same GFP (agonist or shear-exposed) activated with 5 M ionomycin.
Flow-dependent (shear stress) prothrombinase activities were determined using polytetrafluoroethylene tubing in a circulating flow loop as described previously (16). GFP (2 ϫ 10 4 /l) were resuspended in 2.3 ml of Ca 2ϩ -free HBMT supplemented with 0.1% bovine serum albumin and activated for various times and shear rates at 37°C, and duplicate 20-l aliquots were then removed for assays. The result of a single shear loop represents the platelet activation rate, determined by linear regression. The statistical significance of mean platelet activation rates between wild type and mIQGAP1 Ϫ/Ϫ was calculated using a paired Student's t test. For calcium chelation, GFPs were preloaded with 50 M BAPTA-AM (cell-permeant 5,5Ј-dimethylacetoxymethyl ester) (Molecular Probes, Eugene, OR) for 45 min at 37°C followed by a second gel filtration step prior to shear activation (or microspectrofluorimetry using 1 nM ␣-thrombin). For some experiments, determining prothrombinase activity of the microparticle fractions was carried out in parallel using the supernatants of samples centrifuged at 10,000 ϫ g (flow cytometry using FITC-conjugated anti-CD41 confirmed the presence of Ͻ2% intact platelets in microparticulate fractions). Control experiments on uncirculated platelets showed no detectable prothrombinase activity.

IQGAP Characterization in Murine Platelets-
We have previously identified both IQGAP1 and IQGAP2 in human platelets and demonstrated that IQGAP2 activation is restricted to platelet stimulation by thrombin (13). Targeted disruption of the mIQGAP1 gene has been described previously, with no evident phenotypic consequences except for gastric hyperplasia (11). Immunoblot analysis using either murine liver or platelet lysates established that murine platelets specifically express mIQGAP1, with no evidence for mIQGAP2 expression (Fig. 1). These expression patterns are distinct from those of human platelets, which express both IQGAP1 and IQGAP2. Furthermore, there are no compensatory changes in mIQGAP2 expression levels in mIQGAP1 Ϫ/Ϫ platelets. Complete blood counts from 30 mice (wild type (n ϭ 10)), mIQGAP1 ϩ/Ϫ heterozygotes (n ϭ 9), and mIQGAP1 Ϫ/Ϫ (n ϭ 11) demonstrated no differences in hemoglobin, hematocrit, white blood cell counts, platelet counts, mean platelet volumes, platelet F-actin content, or morpholoy as evaluated by Wright-Giemsa staining or transmission electron microscopy. Similarly, in a comparison of heterozygote, wild-type controls (n ϭ 3 in each arm), and mIQ- GAP1 Ϫ/Ϫ platelets, platelet aggregometry using gel-filtered platelets demonstrated small but nonreproducible differences in agonist-induced hyperresponsiveness. These results were most evident using strong receptor agonists such as thrombin or fibrillar collagen (known to activate platelets via members of the PAR family or the glycoprotein VI/Fc␥R-chain receptor tyrosine kinase complex, respectively) and, when evident, were manifest by a more rapid initial slope, with slightly exaggerated amplitudes (data not shown).
Exaggerated Calcium Mobilization in mIQGAP1 Ϫ/Ϫ Platelets-Because IQGAP1 is known to bind Ca 2ϩ /calmodulin via four homologous IQ motifs (IQXXXRGXXXR) (17), microspectrofluorimetry was completed to establish the effect of IQGAP1 deficiency on intracellular calcium responses ([Ca 2ϩ ] i ) in agonist-stimulated platelets. Initial studies using gel-filtered platelets from mIQGAP1 Ϫ/Ϫ mice demonstrated exaggerated elevations of intracytosolic calcium compared with wild-type controls at both concentrations of thrombin (1 and 10 nM). Subsequent experiments using both male and female mice (n ϭ 3 in each arm) confirmed these results, extending these observations to platelet stimulation using the PAR4-activating peptide AYPGKF (Fig. 2). The exaggerated responses were readily evident in the rate of cytosolic calcium release and were concentration-dependent, using both thrombin and the PAR4activating peptide. Because IQGAP1 is a major calmodulinbinding protein, we sought to exclude the possibility that these exaggerated [Ca 2ϩ ] i responses were caused by secondary changes in cytosolic calmodulin concentrations. As seen in Fig.  2, calmodulin concentrations were essentially identical between mIQGAP1 Ϫ/Ϫ and wild-type platelets and comparable with those previously determined in human platelets (19).
Shear Stress-restricted Defect in mIQGAP1 Ϫ/Ϫ Platelet Procoagulant Activity-Platelet procoagulant activity is intimately associated with the calcium-induced microvesiculation that parallels the generation of the prothrombinase complex, although the two endpoints are independently regulated molecular events (20,21). Thus, whereas microvesiculation requires cytoskeletal proteolysis, assembly of the prothrombinase complex is contingent on two processes: (i) the release of activated platelet factor V from the storage pool found within ␣-granules (see below), and (ii) exposure of aminophospholipids, including negatively charged PS. Different platelet agonists cause the extent of PS exposure to vary. For example, ionophores typi-cally effect maximum exposure followed by fibrillar collagen and thrombin; PS exposure is also known to result from mechanical shear stress (22). To further dissect the relationship of the exaggerated [Ca 2ϩ ] i release in response to PS exposure, platelets were activated by distinct agonists or by mechanical shear, and procoagulant activity was determined using a modified prothrombinase assay that closely parallels PS exposure as quantified by annexin-V binding (15). Consistent with prior observations (2, 4), thrombin and collagen (but not ADP or PAR4 agonist peptide) caused dose-dependent increases in prothrombinase activity, results that were essentially the same for both mIQGAP1 Ϫ/Ϫ and wild-type platelets (Fig. 3a). In sharp contrast, however, mIQGAP1 Ϫ/Ϫ platelets activated by shear stress demonstrated enhanced prothrombinase activity compared with littermate controls (either at 12 or 16 dynes/cm 2 , equivalent to shear rates of 1200 and 1600 s Ϫ1 , respectively). This exaggerated procoagulant activity was unrelated to inherent differences in membrane function or phospholipid content because the maximal prothrombinase activities between wild type and mIQGAP1 Ϫ/Ϫ platelets were essentially identical (Fig.  3a). Similarly, there was no evidence that the enhanced procoagulant activity could be explained by inherent differences in microvesiculation as determined by flow cytometric analyses.
More exhaustive experiments were completed using platelets activated by shear stress of 16 dynes/cm 2 (corresponding to ϳ2.5 times normal shear stress in coronary arteries, or ϳ40% degree of stenosis). Because the prothrombinase measurement of platelet procoagulant activity reflects the composite sum generated from microvesicular and intact platelets (23), parallel determinations were made of the proportion of the total activity derived from platelet microparticles. As shown in Fig.  4, by using larger numbers of mouse platelets activated by moderate shear stress (n ϭ 12 in each arm, with male:female ratios of 1:1), mIQGAP1 Ϫ/Ϫ platelets demonstrated an enhanced ability for prothrombinase generation, differences that were highly significant compared with wild-type controls (p Ͻ 0.0001). Although microvesiculation is known to accompany high shear activation, minimal (and essentially comparable) microparticle formation was evident in both mIQGAP1 Ϫ/Ϫ and wild-type platelets activated by low-shear (Fig. 5), consistent with prior observations (22). Moreover, the prothrombinase activity was primarily found in the intact platelets from both groups of mice, with an identical fraction (ϳ26%) accounted for  (18). a, agonist incubation (downward arrow) using 1 nm or 10 nM ␣-thrombin. b, concentration-dependent rate responses using ␣-thrombin or AYPGKF (PAR4 peptide). All determinations represent the mean Ϯ S.E. from quadruplicate wells from one complete set of studies (*, p Ͻ 0.05; **, p Ͻ 0.01; error bars were omitted from a to improve clarity). c, calmodulin activity determinations were completed from platelet-rich plasma as described under "Materials and Methods." by platelet microparticles. Thus, although IQGAP1 is known to have actin-polymerizing functions that could be involved in the cytoskeletal actin reorganization known to accompany microvesiculation, mIQGAP1 deficiency is uniquely associated with enhanced shear-associated platelet prothrombinase activity, without a concomitant defect in microvesiculation.  3. Quantitative prothrombinase determinations of mIQGAP1 Ϫ/Ϫ and wild-type platelets. a, aliquots of gel-filtered platelets (4 ϫ 10 6 /well) in HBMT were loaded into individual wells and activated by agonists (or shear) for 10 min followed by prothrombinase determinations as described under "Materials and Methods." Results are expressed as the mean Ϯ S.E. from duplicate wells from one complete set of representative studies (repeated once), except for shear-determinations, which are from one well. The inset shows the maximal attainable prothrombinase activity using 5 M ionomycin (n ϭ 10/genotype). b, flow cytometric scattergrams (lower panels) demonstrate no differences in annexin-V binding or PMP formation between wild-type (solid lines) and mQGAP1 Ϫ/Ϫ (light brown) platelets maximally stimulated using 5 M ionomycin (representative dot plot is shown in upper right panel). The upper left panel dot plot demonstrates unstimulated platelets. PE, phycoerythrin. ther delineate the intracellular signaling pathways regulating the procoagulant phenotype, we studied the role of the aberrant [Ca 2ϩ ] i and dissected the two critical components of platelet prothrombinase assembly (negatively charged phospholipid exposure and factor V surface expression caused by ␣-granule exocytosis). Initially, platelets were incubated with the intracellular calcium chelator BAPTA-AM followed by activation using the identical shear rate of 16 dynes/cm 2 . Unexpectedly, BAPTA-treated mIQGAP1 Ϫ/Ϫ platelets demonstrated persistently exaggerated procoagulant activity (Fig. 4), despite evidence for effective chelation as established by microspectrofluorimetry after shear activation (data not shown). These results were in sharp contrast to wild-type control platelets, which demonstrated inhibitory effects on the generation of platelet procoagulant activity with calcium chelation.
To more specifically dissect the mechanism for the enhanced prothrombinase generation data, simultaneous determinations quantified annexin-V binding (as a marker for PS exposure) and P-selectin cell surface expression (as a marker for ␣-granule exocytosis). For both wild-type and mIQGAP1 Ϫ/Ϫ platelets, the distribution of intact platelets, microparticles, and annexin-V binding were essentially identical when platelets were activated by shear rates of 16 dynes/cm 2 (Fig. 5). In contrast, mIQGAP1 Ϫ/Ϫ platelets demonstrated a shorter lag phase of P-selectin expression compared with wild-type platelets; these differences were limited to intact platelets with no differences evident in the generation of P-selectin microparticles between both platelet phenotypes. Indeed, the time to half-maximal response (t1 ⁄2 ) was Յ5 min for mIQGAP1 Ϫ/Ϫ platelets compared with t1 ⁄2 ϳ12 min for wild-type controls. This abbreviated lag time in mIQGAP1 Ϫ/Ϫ platelets was not associated with any changes in either maximal P-selectin expression or its expression in platelet microparticles. These observations strongly support a model linking platelet IQGAP1 with a shear-acti-vated, exocytic pathway that specifically regulates platelet prothrombinase complex assembly, which is independent of platelet microvesiculation. Furthermore, despite the fact that maximal ␣-granule release is essentially identical between wild-type and mIQGAP1 Ϫ/Ϫ platelets by 15 min, mIQGAP1 Ϫ/Ϫ platelets display persistently enhanced procoagulant activity at all subsequent time points (see Fig. 4). These data did not directly quantify cell surface factor V expression as a marker of prothrombinase assembly. Nonetheless, if ␣-granule-derived factor V secretion parallels that of P-selectin, an extrapolation of these observations suggests that the initial rate of prothrombinase complex assembly represents a major determinant of subsequent thrombin generation over time. DISCUSSION What is the mechanism by which IQGAP1 modulates platelet procoagulant activity, and how is this related to the exaggerated [Ca 2ϩ ] i transients? Previous observations have demonstrated complex relationships among Ca 2ϩ , calmodulin, and structural domains of IQGAP1. Indeed, Ca 2ϩ can directly associate with IQGAP1 (24), and Ca 2ϩ /calmodulin binding modulates the functional interactions of IQGAP1 with actin and cdc42 (25,26). Furthermore, the IQ motifs display differential binding to apocalmodulin compared with Ca 2ϩ /calmodulin (27). This observation implies that IQGAP1 may facilitate Ca 2ϩ association to calmodulin-bound IQGAP1, a model that may explain the exaggerated [Ca 2ϩ ] i release in IQGAP1-null platelets. Irrespective of the mechanism, this aberrant [Ca 2ϩ ] i response appears unrelated to the enhanced procoagulant phenotype evident in shear-induced mIQGAP1 Ϫ/Ϫ platelets, data that are entirely consistent with previous observations disjoining the intracytosolic calcium release from the procoagulant capacity of the activated platelet membrane (2,4,20,21). Rather, our data establish that IQGAP1 functions in a calcium- FIG. 4. Platelet shear-activation studies. a, gel-filtered platelets (4 ϫ 10 4 /l) from wild-type and mIQGAP1 Ϫ/Ϫ mice were subjected to shear stress of 16 dynes/cm 2 at 37°C for 20 min, and 100-l aliquots were removed at various time intervals for prothrombinase assays using total or PMP fractions. Results represent the means Ϯ S.E., and time points are connected using linear best-fit determinations for the determination of platelet activation rate (p Ͻ 0.0001; mIQGAP1 Ϫ/Ϫ versus wild-type totals). Clear square represent prothrombinase activities of wild-type or mIQGAP1 Ϫ/Ϫ (solid square) platelets circulated without shear. b, intracellular calcium chelation and platelet procoagulant activity. GFP were preloaded with 50 M BAPTA-AM for 45 min at 37°C (n ϭ 5 in each arm) followed by shear activation and prothrombinase determinations as in a (p values are shown). Note that the higher prothrombinase activity in wild-type platelets (no BAPTA) in these experiments is explained by the additional time (60 -75 min) between platelet isolation and shear activation (16), although the prothrombinase activity is still statistically lower than that of mIQGAP1 Ϫ/Ϫ platelets. independent secretory pathway of ␣-granule release, specifically activated by shear stress.
Regulated platelet exocytosis proceeds through a sequence involving granule docking and fusion, utilizing core SNARE (soluble NSF (N-ethylmaleimide-sensitive factor) attachment protein receptor)-related molecules that are functional in neuronal secretion and conserved in yeast (28). To date, there is no evidence that IQGAP1 functions within this complex, although the S. cerevisiae Iqg1p homologue is known to interface proteins regulating polarity-dependence with those involved in exocytosis/secretion and cytokinesis (29). Alternatively, IQGAP1 could modulate ␣-granule secretion by interactions with microtubules or microtubule-associated proteins (known to be important for platelet secretion (28), a prime candidate being CLIP-170 (8)). Recent evidence would suggest that an IQGAP1/CLIP-170 scaffold complexed with [GTP]-bound rac1/ cdc42 functions as a link between microtubules and the cortical actin meshwork, thereby providing a calcium-independent, integrated signaling complex that could normally function to negatively regulate the early stages of granule exocytosis.
What is the mechanoreceptor linked to platelet IQGAP1 activation, and how does IQGAP1 become activated during shear stress? Although various studies have delineated key IQGAP1 structural domains mediating biochemical interactions with regulatory proteins in vitro, our data provide the first evidence for a physiological signal (i.e. shear) directly linked to IQGAP1 activation. Previous data have established a role for the von Willebrand factor/GPIb␣ mechanoreceptor axis in shear-induced conformational activation of ␣ IIb ␤ 3 , although the role of this agonist-receptor complex (and its signaling pathway) in the generation and assembly of the prothrombinase complex remains incompletely delineated (30). Our shear loop did not include exogenous von Willebrand factor, suggesting that the intracellular signaling pathway occurs independently of von Willebrand factor/ GPIb␣ ligation, although we cannot definitively exclude the possibility of GPIb␣ activation from endogenous plateletderived von Willebrand factor. Interestingly, the physical properties of the phospholipid bilayer may be sufficient for shear stress-induced activation of membrane-bound G proteins in the absence of cell surface receptors (31), thereby providing a direct link to IQGAP1 via GTP-charging of rac1 and/or cdc42. Finally, although IQGAP1 specifically modulates the kinetics of prothrombinase assembly, our data also suggest that IQGAP1 deficiency results in progressive procoagulant activity over time, imputing a regulatory role for IQGAP1 in sustaining cell surface thrombin generation. These data implicate IQGAP1 in a fundamental signaling mechanism that regulates platelet procoagulant activity and presumably normal hemostasis; further characterization of this signaling pathway should identify novel targets capable of modulating platelet procoagulant function.