The Glycoprotein Ib-IX-V Complex Mediates Localization of Factor XI to Lipid Rafts on the Platelet Membrane*

Factor XI binds to activated platelets where it is efficiently activated by thrombin. The factor XI receptor is the platelet membrane glycoprotein (GP) Ib-IX-V complex (Baglia, F. A., Badellino, K. O., Li, C. Q., Lopez, J. A., and Walsh, P. N. (2002) J. Biol. Chem. 277, 1662–1668), a significant fraction of which exists within lipid rafts on stimulated platelets (Shrimpton, C. N., Borthakur, G., Larrucea, S., Cruz, M. A., Dong, J. F., and Lopez, J. A. (2002) J. Exp. Med. 196, 1057–1066). Lipid rafts are membrane microdomains enriched in cholesterol and sphin-golipids implicated in localizing membrane ligands and in cellular signaling. We now show that factor XI was localized to lipid rafts in activated platelets ( (cid:1) 8% of total bound) but not in resting platelets. Optimal binding of factor XI to membrane rafts required prothrombin (and Ca 2 (cid:1) ) or high molecular weight kininogen (and Zn 2 (cid:1) ), which are required for factor XI binding to platelets.

Factor XI (FXI) 1 is a 160,000-Dalton disulfide-linked homodimeric coagulation protein that exists in plasma in a complex with high molecular weight kininogen (HK) (1)(2)(3)(4)(5)(6)(7). In the presence of HK (and Zn 2ϩ ) or prothrombin (and Ca 2ϩ ), FXI can bind specifically and reversibly to high affinity sites on the surface of stimulated platelets (8,9). Activated platelets can thereby promote FXI activation by thrombin (8,10). We have previously demonstrated that the Apple 3 (A3) domain of FXI mediates the binding of FXI to platelets (11,12). FXI binds to the glycoprotein Ib-IX-V (GPIb-IX-V) complex since 1) Bernard-Soulier platelets lack the GPIb complex and are deficient in binding FXI; 2) a monoclonal antibody against GPIb␣ (SZ-2) inhibits the binding of FXI to the platelet surface; 3) bovine von Willebrand protein, which binds GPIb␣, also inhibits the binding of FXI to activated platelets; and 4) FXI binds to glycocalicin (the soluble extracellular region of GPIb␣) in the presence of Zn 2ϩ ions (13). It has recently been determined by analyzing fractions from sucrose density gradients of Triton X-100 platelet lysates that a significant fraction of the GPIb-IX-V complex is localized within membrane rafts in resting platelets and that this fraction was increased upon platelet activation (14). The precise physiological role of lipid rafts is unclear, although they have been postulated to be involved in localizing membrane ligands and in cellular signaling.
Lipid rafts (also called detergent-insoluble/resistant membranes and glycolipid-enriched membranes) are microdomains of the cell membrane that contain saturated phospholipids, glycosphingolipid-containing assemblies of cholesterol that are highly ordered in array when compared with the rest of the cell membrane (15,16). It is well documented that lipid rafts are enriched in proteins such as the Src family kinases (17), the active Lyn protein tyrosine kinase (18), and the glycosylphosphatidylinositol-linked outer membrane protein (19). These microdomains can be isolated from many hematopoietic cells and are thought to be involved in cell signaling (15,20,21). It has been suggested that clustering of raft components results in a coalescence of lipid rafts, thus connecting raft proteins and accessory molecules into a larger domain (22). However, the role of lipid rafts in receptor signaling in non-immune cells such as platelets is not known.
Since a significant component of the GPIb complex is localized within membrane rafts (14) and FXI binds to the GPIb complex (13), we have asked whether FXI binds to membrane rafts in activated and resting platelets. Our studies show that platelet-bound FXI is associated with membrane rafts and that FXI is bound to the GPIb-IX-V complex in membrane rafts. This interaction may have a physiological role as a platform for the colocalization and activation of FXI by thrombin on the platelet surface.

EXPERIMENTAL PROCEDURES
Materials-All reagents were obtained from Sigma unless stated otherwise. Human FXI, human prothrombin, human FXIa, and human HK were purchased from Hematologic Technologies Inc. (Essex Junction, VT). Human ␣-thrombin ( from William F. Nye Inc. (Fairhaven, MA). Carrier-free Na 125 I was from Amersham Biosciences. The chromogenic substrate for measurement of FXIa activity (S2366) was obtained from Chromogenix (Mölndal, Sweden). The thrombin receptor agonist peptide (TRAP), SFLLRN-amide, was synthesized at the Protein Chemistry Facility of the University of Pennsylvania on the Applied Biosystems 430A Synthesizer, and reverse-phase high performance liquid chromatography was used to purify it to Ͼ99% homogeneity. A monoclonal antibody (SZ-2), which recognizes the NH 2 -terminal extracellular globular domain of GPIb␣, blocks thrombin binding to platelets at low concentration, and inhibits thrombin-induced platelet aggregation, was used in experiments examining FXI binding to platelets. An isotype-specific mouse IgG2A chain control antibody was purchased from Sigma.
Radiolabeling of FXI-Purified FXI was radiolabeled with 125 I by a minor modification (8) of the Iodogen method to a specific activity of 5 ϫ 10 6 cpm/g. The radiolabeled FXI retained Ͼ98% of its biological activity.
Platelet Binding Experiments-Platelets were prewarmed to 37°C and incubated at a concentration of 1 ϫ 10 8 /ml in calcium-free Hepes-Tyrode's buffer, pH 7.3, in a 1.5-ml Eppendorf plastic centrifuge tube with a mixture of radiolabeled FXI, divalent cations, a thrombin receptor (PAR-1) activation peptide (SFLLRN-amide) as a platelet agonist (8,11), and HK or antibodies or other proteins as designated in the figure legends. All incubations were performed at 37°C without stirring after an initial mixing of the reaction mixture. At various added FXI concentrations, aliquots were removed (100 l) and centrifuged through a mixture of silicone oils as described in Refs. 8 -11. Platelet Lysis and Sucrose Density Gradient Fractionation-Gel-filtered platelets (1-2 ϫ 10 8 /ml, containing 125 I-FXI and cofactors) were washed and pelleted in 80% sucrose in MES-buffered saline (MBS: 25 mM MES, pH 6.5, 150 mM NaCl). The washed and pelleted platelets were lysed with 2 ml of ice-cold Triton X-100 containing 10 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, 20 g/ml aprotinin, 20 g/ml leupeptin, 100 g/ml soybean trypsin inhibitor. All subsequent steps were performed at 4°C. The lysate was adjusted to 40% (w/v) sucrose by addition of an equal volume of 80% (w/v) sucrose in MBS. Four milliliters of 30% sucrose followed by 2 ml of 5% sucrose were gently layered over the 40% sucrose fraction in the ultracentrifugation tube. The samples were then centrifuged at 200,000 ϫ g at 4°C for 18 h in a SW40 Rotor (Beckman Instruments). Eleven equal fractions were collected from the top of the gradient. Protein concentrations were determined using the Micro BCA Protein Assay (Pierce) according to the manufacturer's instructions. Dot blots were performed using a Bio-Dot Microfiltration Apparatus (Bio-Rad) according to the manufacturer's instructions and probed for the ganglioside GM1 with horseradish peroxidase-conjugated cholera toxin B-subunit (10 g/ml).
Cholesterol Depletion of Platelets-To specifically deplete the platelet membrane of cholesterol, platelet-rich plasma was incubated with methyl-␤-cyclodextrin (M␤CD) at a final concentration of 20 mM for 30 min at 37°C. Cells were repleted with cholesterol by incubating them in the presence of a cholesterol/M␤CD mixture for 1 h at 37°C. A stock solution of 0.4 mg/ml cholesterol and 10% cyclodextrin was prepared by vortexing at ϳ40°C in 10 ml of 10% M␤CD with 200 l of cholesterol (20 mg/ml in ethanol solution).
Assay of FXI Activation-Activation of FXI (60 nM) by thrombin (1.25 nM) was measured by chromogenic assay as described previously (8,10). Briefly, incubations were carried out at 37°C in 200 l of Tris-buffered saline (50 mM Tris, 150 mM NaCl, pH 7.3) and 1% bovine serum albumin. Gel-filtered platelets were activated by incubation at 37°C for 1 min with thrombin receptor activation peptide (SFLLRN-amide, 25 M). After dilution to a final volume of 1 ml with Tris-buffered saline with 1% bovine serum albumin containing 600 M S2366 (EPR-paranitroanilide, Chromogenix), the amount of free para-nitroanilide was determined by measuring the changes in absorbance at 405 nm (A 405 ). The amount of FXIa generated was assayed by reference to a standard curve constructed using purified FXIa.

Isolation of Platelet Membrane
Rafts-Platelet rafts have been isolated previously (18,23) at low temperatures. They are insoluble in non-ionic detergents and are isolated by continuous gradient centrifugation based on their low density buoyant characteristics. TRAP-activated platelets were lysed in an icecold solution of Triton X-100 (at various concentrations, 0.025-1%), and the lysate was centrifuged in a sucrose gradient for 18 h at 4°C. The results shown in Fig. 1 were those obtained when TRAP-activated platelets were lysed in 0.25% Triton X-100 since this concentration of detergent resulted in maximal raft association of platelet-bound FXI as shown in Fig. 2 and discussed below. Eleven fractions (900 l each) were collected from the top of the gradient. In agreement with previous studies on platelet rafts (18,23), the low density raft-containing fractions 1-4 contained less than 5% of total cellular protein with the higher density fractions 7-11 containing the majority of the protein (Fig. 1A). The raft marker ganglioside GM1 determined the position of lipid rafts within the sucrose gradient. This was identified by ligand blotting with cholera toxin B-subunit, which specifically binds this lipid (24). Dot blot analysis revealed that lipid rafts localize to fractions 1-4 within the gradient. Dot blot staining was evident also in fractions 9, 10, and 11 of the sucrose gradient. This might represent lipid rafts that are solubilized by detergent or unable to float as a result of linkage to the cytoskeleton. In all subsequent experiments (below), GM1 localization was determined to be unaffected by platelet activation, by the concentration of Triton X-100 utilized, by protein (HK or prothrombin) or metal Detergent-sensitive Association of Factor XI with Membrane Rafts-Since a significant proportion of the GPIb-IX-V complex is contained in membrane rafts and FXI binds to GPIb, we designed experiments to determine whether labeled FXI is also incorporated into membrane rafts. Since it has been reported (25,26) that certain ligand-membrane raft associations may be disrupted by concentrations of Triton X-100 Ͼ0.5%, we initially aimed to optimize the detergent concentration required for optimal FXI-membrane raft association. We therefore examined the distribution of 125 I-FXI in sucrose density fractions of lysates (using Triton X-100 concentration of 0.025-1%) from TRAP-activated platelets (in the presence of cofactors that optimize FXI binding to platelets). Fig. 2 indicates that at higher concentrations of Triton X-100 (0.5-1%) little or no FXI is associated with platelet membrane rafts, while at concentrations of Triton X-100 Յ0.25%, ϳ5-8% of the bound FXI is associated with the membrane raft fractions 1-4. The majority of bound FXI (ϳ87-90%) is localized within fractions 6 -11 of the sucrose gradient, and ϳ5% of bound FXI was recovered in the pellet of Triton X-100 lysates. These data indicate that platelet-bound FXI is associated with the membrane raft fraction, and its association with rafts is dependent upon the concentration of Triton X-100 used to solubilize the platelets.
Cofactor Requirements for Binding Factor XI to Lipid Rafts in TRAP-activated and Unactivated Platelets-We have previously determined the optimal cofactor requirements for FXI binding to the platelet surface. It was shown that HK, Ca 2ϩ , and Zn 2ϩ ions or prothrombin and Ca 2ϩ ions are required as cofactors for optimal binding of FXI to the activated platelet surface (8,27). Fig. 3, A and B, shows the distribution of 125 I-FXI in sucrose density fractions of lysates from both stimulated and unactivated platelets, respectively, using the various cofactors necessary to bind FXI to the platelet surface. The data show that the presence of prothrombin (1.2 M) and CaCl 2 (2 mM) or HK (42 nM), ZnCl 2 (25 M), and CaCl 2 (2 mM) are required not only for binding of FXI to the activated platelets but also for FXI association with the membrane raft fraction. The amount of FXI associated with the raft fraction (5-8%) was the same for both cofactors. Platelets that were not activated did not bind FXI, and thus FXI was not associated with membrane rafts. A lesser amount of FXI was bound to membrane rafts (3-4% of total binding) in the presence of ZnCl 2 (25 M) without added HK, but no FXI binding was observed in the presence of CaCl 2 (2 mM) without added prothrombin. Thus, the optimal conditions for binding FXI to the activated platelets are also optimal for FXI association with membrane rafts.
Raft Disruption by Cholesterol Depletion-Membrane cholesterol depletion, using the cholesterol-binding agent M␤CD, is a well established method for disrupting lipid rafts (21). It has been demonstrated in platelets that M␤CD effectively removes cholesterol from platelets as a function of time (14). M␤CD did not induce microparticle formation nor did it alter platelet count. In platelets treated with M␤CD for 30 min, FXI was completely removed from the lipid raft fractions, and this removal was reversed by cholesterol repletion (Fig. 4). GM1 localization was also monitored by ligand blotting with the cholera toxin B-subunit and demonstrated complete elimination of raft formation by M␤CD treatment (i.e. only fractions 8 -11 were positive for GM1), whereas treatment with M␤CD in the presence of cholesterol had no effect on GM1 localization (data not shown). To exclude the possibility that the loss of FXI binding to lipid rafts was due to a loss of receptors for FXI on the platelet surface, we assessed whether M␤CD affected total binding of FXI to the activated platelet. Fig. 5 demonstrates that even after treatment of platelets with concentrations of M␤CD as high as 40 mM there was no effect on the amount of FXI bound to the activated platelet.
The Effects of an Antibody to Glycoprotein Ib and the Recombinant Apple 3 Domain on the Binding of Factor XI to Membrane Rafts-FXI binds activated platelets specifically and reversibly, and this interaction is mediated by amino acids within the A3 domain of FXI (11,12,27). Fig. 6A demonstrates that the recombinant A3 domain abolishes the binding of FXI to membrane rafts as well as the non-raft fractions in the sucrose gradient. In a study demonstrating that the platelet receptor for FXI is the GPIb-IX-V complex, an antibody (SZ-2) that recognizes the NH 2 -terminal extracellular globular domain was shown to block FXI binding to activated platelets (13). Fig.  6B demonstrates that the antibody SZ-2 also abolishes binding of FXI to lipid rafts as well as non-raft fractions in the sucrose gradient. These experiments are consistent with the conclusion that FXI binds through its A3 domain to GPIb␣ and that the FXI-GPIb-IX-V complex is compartmentalized within membrane raft structures.
The Activation of Factor XI by Thrombin Occurs in Lipid Rafts-Activated gel-filtered platelets as well as glycocalicin promote FXI activation by thrombin in the presence of HK or prothrombin at optimal rates, thereby initiating intrinsic coagulation independent of contact proteins (13). Initial rates of FXI activation by thrombin were significantly decreased compared with normal platelets when activated Bernard-Soulier platelets (deficient in GPIb complex) were used as a surface (13). These observations suggest that the GPIb-IX-V complex serves as a receptor for facilitating thrombin-catalyzed FXI activation. To investigate the physiological relevance of FXI-GPIb-IX-V complex-raft association, we disrupted the structural integrity of lipid rafts by cholesterol depletion with M␤CD. Initial rates of FXI activation by thrombin on TRAP-activated platelets were inhibited Ͼ85% (from ϳ17 to 2.4 nM/min) by cholesterol depletion (Fig. 7). These experiments suggest that the localization of FXI to lipid rafts containing the GPIb complex is impor-tant for the activation of FXI by thrombin on the platelet surface.

DISCUSSION
Utilizing a sequence of amino acids (Ser 248 -Val 271 ) in the A3 domain (11,12,27), FXI binds reversibly and specifically to high affinity sites on the activated platelets in the presence of HK (and Zn 2ϩ ions) or prothrombin (and Ca 2ϩ ions) (8, 9) as a consequence of which the rate of FXI activation by thrombin is . 125 I-FXI (22 nM) and gel-filtered platelets (1-2 ϫ 10 8 ml), treated with 0 -40 mM M␤CD, were incubated with the above cofactors for 30 min at 37°C. Aliquots were removed and centrifuged as described under "Experimental Procedures." Each point is an average of three determinations, and the maximum variation of counts/min bound was Ͻ3% of total counts/min bound. One hundred percent binding of FXI represents an average of 110,600 cpm bound. accelerated Ͼ5,000-fold (8,10). FXI interacts with the GPIb-IX-V complex on the platelet surface, and this interaction promotes thrombin-catalyzed FXI activation (13). The GPIb-IX-V complex is a large plasma membrane complex comprising four polypeptide chains, GPIb␣, GPIb␤, GPIX, and GPV, arranged in stoichiometry of 2:2:2:1 (28). Approximately 25,000 copies of the first three peptides are constitutively expressed on the unactivated platelet surface along with half as many copies of GPV (29 -31). The complex also binds thrombin and von Willebrand factor with high affinity (32). However, platelet activation results in a ϳ65% decrease of GPIb from the platelet surface (33) and a redistribution to the surface canalicular system (34). Therefore, how do we account for the fact that platelet activation is required for the binding of FXI to platelet GPIb-IX-V complex and its activation by thrombin? Recently it has been reported that specialized membrane microdomains (lipid rafts) are involved in GPIb-IX-V-mediated signaling (14).
Lipid rafts are dynamic assemblies of cholesterol and sphingolipids that are more ordered in structure than the rest of the membrane (15,21). They are believed to act as "platforms" for signal transduction and ligand localization by selectively attracting certain proteins while excluding others. It was found that a significant fraction of the GPIb-IX-V complex exists within rafts in resting platelets (8%), which increased to 26% upon von Willebrand factor-induced activation of platelets (14). Since FXI binds to 1,500 sites on the platelet surface, i.e. to only a small fraction of the GPIb-IX-V molecules, it is possible that lipid rafts may compartmentalize and colocalize FXI with thrombin in the GPIb-X-V complex on the activated platelet surface.
The present studies support the conclusion that FXI can interact with lipid rafts on the platelet surface and that this interaction is mediated through the GPIb-IX-V complex. The structural integrity of lipid rafts is important for promoting optimal rates of FXI activation by thrombin on the platelet surface. The evidence that supports this conclusion is as follows. 1) ϳ5-8% of bound FXI is associated with membrane rafts in TRAP-activated platelets (Figs. 2 and 3, A and B); 2) FXI-raft association was disrupted by cholesterol depletion (Fig. 4); 3) FXI binds to lipid rafts through the A3 domain (Fig.  6A); 4) the SZ-2 antibody blocks the binding of FXI to lipid rafts suggesting FXI interacts with GPIb in lipid rafts (Fig. 6B); and 5) the initial rate of FXI activation by thrombin on activated platelets was inhibited Ͼ85% by cholesterol depletion leading to raft disruption (Fig. 7).
There have been contradictory results using detergent insolubility to determine whether various signaling proteins are associated with membrane rafts (35,36). This was probably because proteins that are strongly associated with membrane rafts are Triton X-100-insoluble, whereas weakly associated proteins are sensitive to detergent extraction. An example of a ligand weakly associated with rafts is GPVI as evidenced by the fact that its recovery with the low density fraction is only achieved using low concentrations of Triton X-100 (25). Therefore, we first optimized the experimental conditions for FXI association with platelet membrane rafts. At high concentra- tions of Triton X-100 (0.5-1%), FXI was not associated with membrane rafts, whereas at concentrations Յ0.25%, ϳ5-8% of bound FXI is associated with membrane rafts. Since the detergents Brij58 and Brij46 gave the same pattern of distribution of the FXI within the sucrose gradient, apparently the presence of FXI in the lipid fraction is not an artifact of Triton X-100 solubilization (data not shown).
Three different conditions were utilized in the present study to characterize FXI interactions with membrane rafts on platelets: i.e. in the presence of HK and Zn 2ϩ ions, prothrombin and Ca 2ϩ ions, or Zn 2ϩ alone. The interaction of FXI with activated platelets is optimal in the presence of either prothrombin and Ca 2ϩ or HK and Zn 2ϩ with 1,500 sites per platelet and K d(app) ϳ 10 -15 nM (8,9). However, in the presence of Zn 2ϩ ions alone (25 M), FXI interacts with half the number of platelet sites (ϳ800 sites per platelet and K d(app) ϳ 12 nM) (13). We determined that ϳ8% of FXI bound to activated platelets appears in lipid rafts in the presence of HK and Zn 2ϩ ions or prothrombin and Ca 2ϩ ions (Fig. 3, A and B) and ϳ4% of FXI bound to activated platelets in the presence of Zn 2ϩ alone appears in lipid rafts. Thus FXI binding to membrane rafts appears to occur even in the absence of added protein cofactors provided Zn 2ϩ ions are present. A possible mechanism to account for this binding in the absence of added HK or prothrombin is the presence of HK in platelet ␣-granules and its secretion and surface membrane binding after platelet activation (37).
Several reports have implicated platelet membrane lipid rafts in signaling events mediated by von Willebrand factor binding to GPIb-IX-V and convulxin binding to GPVI (14,18,23,25). We examined both TRAP-activated and resting platelets and determined that association of FXI with membrane rafts is activation-dependent. Although the GPIb-IX-V complex is present (ϳ8% of total) in lipid rafts of resting platelets, FXI is not able to bind to resting platelets or to associate with the raft membrane of the unstimulated platelet. The mechanisms accounting for activation-dependent FXI binding to platelets and its association with GPIb-IX-V in lipid rafts are currently unclear. It is possible that the conformation of GPIb-IX-V is altered when platelets are activated, allowing it to bind FXI and associate with lipid microdomains. Alternatively the character of the lipid raft may change with platelet activation (e.g. lipid distribution across different membrane leaflets may be drastically different when platelets are activated). Also it is possible that an unknown cofactor (e.g. platelet HK) is exposed or secreted upon platelet activation that permits optimal binding of FXI to GPIb-IX-V and/or the association of the FXI-GPIb-IX-V complex with lipid rafts.
To corroborate the finding that FXI is recruited to the GPIb-IX-V complex within lipid raft fractions upon TRAP activation, we determined FXI distribution by sucrose density centrifugation after incubating platelets with the SZ-2 antibody (directed against the NH 2 -terminal globular domain of GPIb␣) or the recombinant A3 domain, which contains the platelet binding site for FXI. We previously determined that the SZ-2 antibody blocks the binding of FXI to activated platelets (13) and that the radiolabeled recombinant A3 domain binds to activated platelets with the same K d (ϳ10 nM) and number of sites (ϳ1,500) as FXI, suggesting it is the only binding site on FXI for the platelet surface (11). Both of these probes blocked the binding of FXI to membrane rafts as well as the binding to non-raft fractions, suggesting that FXI can also interact with GPIb-IX-V in the non-raft fractions. Thus, FXI binds the GPIb-IX-V complex both in membrane rafts and in non-raft membranes, and the A3 domain of FXI mediates this interaction. Whether some biochemical modification of FXI may occur (i.e. palmitoylation) to direct it to the lipid rafts is unknown but possible since enzymatic acylation and deacylation of the GPIb-IX-V complex may regulate its association with rafts (14).
To investigate the physiological significance of the partitioning of FXI into lipid rafts, we selectively depleted platelet membrane cholesterol using the cholesterol-binding agent M␤CD. It has been determined that 10 mM M␤CD efficiently removes cholesterol from [ 3 H]cholesterol-labeled platelets as a function of time with ϳ75% of incorporated cholesterol released in 30 min (14). We therefore utilized 20 mM M␤CD in our experiments. Cell viability was not affected by this treatment. In platelets treated with M␤CD for 30 min, FXI was completely removed from the lipid raft fractions, and replenishment of cholesterol to depleted platelets restored the binding of FXI to the lipid raft fractions (Fig. 4). To exclude the possibility that the loss of FXI interaction with rafts following cholesterol depletion was due to a failure to activate cholesterol-depleted platelets, we assessed receptor levels in FXI binding studies (Fig. 5). FXI binding to the activated platelets remained unchanged following a 30-min treatment of up to 40 mM M␤CD when compared with untreated platelets. Thus, it is important to emphasize that cholesterol depletion does not prevent the exposure of FXI receptors by TRAP-mediated PAR-1 activation but does prevent the association of these receptors with membrane rafts.
We also carried out experiments to determine whether the activation of FXI by thrombin may occur within platelet membrane microdomains. Cholesterol-depleted, M␤CD-treated platelets were activated with TRAP and examined for their capacity to support FXI activation by thrombin. Initial rates of FXI activation by thrombin were inhibited Ͼ85% by cholesterol depletion (Fig. 7). That this inhibition by M␤CD of the capacity of activated platelets to facilitate FXI activation was not a consequence of inhibition of platelet activation is evidenced by the fact that M␤CD-treated platelets respond to TRAP normally by exposing FXI binding sites and by aggregation responses to TRAP (data not shown). Thus, FXI association with the GPIb-IX-V complex within membrane microdomains is important for the initiation of intrinsic coagulation independent of contact proteins.
The stoichiometry of interaction between GPIb-IX-V and FXI within membrane lipid rafts remains to be determined. Thus, of the ϳ25,000 copies of GPIb-IX-V per platelet, 8% (ϳ2,000 copies per platelet) is partitioned into lipid rafts in resting platelets, and 26% (ϳ6,500 copies per platelet) is partitioned into lipid rafts in von Willebrand factor-activated platelets (14). In contrast, of the 1,500 -2,000 FXI molecules bound per platelet, ϳ8% (ϳ120 -160 molecules per platelet) is partitioned into rafts of TRAP-activated platelets. This suggests that the GPIb-IX-V complex is not saturated with FXI, which may occupy only ϳ10% of the available raft-associated GPIb-IX-V complexes. Another ligand known to bind to and signal through the GPIb-IX-V complex is thrombin (32). The present studies strongly suggest the possibility that the FXI-thrombin-GPIb-IX-V complex is partitioned within membrane lipid microdomains for efficient thrombin-catalyzed FXI activation. Consistent with this possibility is our observation that disruption of lipid rafts by cholesterol depletion with M␤CD does not prevent platelet activation leading to FXI binding to platelets but does prevent FXI localization within lipid rafts and FXI activation by thrombin.

VOLUME 277 (2002) PAGES 1662-1668
Factor XI binding to the platelet glycoprotein Ib-IX-V complex promotes factor XI activation by thrombin.  Table 1 have been retracted by the authors for the following reasons.
All of the authors with the exception of F. A. Baglia retract the specific data listed above because recent experiments conducted by Dipali Sinha, Sergei Shikov, Wenman Wu, and Syed Ahmad in the laboratory of Peter N. Walsh failed to confirm the conclusion that activated platelets promote the activation of factor XI by thrombin. All of the other results reported in the paper are valid. A detailed explanation of the chronology of events leading to this retraction and the retraction of a paper from Biochemistry (Baglia, F. A., and Walsh, P.N. (1998) Prothrombin is a cofactor for the binding of factor XI to the platelet surface and for platelet-mediated factor XI activation by thrombin. Biochemistry 37, 2271-2281) has been published in the journal Biochemistry (manuscript bi-2007-01501k, accepted July 27, 2007). We apologize to the readers, reviewers, and editors of the Journal of Biological Chemistry for publishing these erroneous data.

VOLUME 275 (2000) PAGES 20514 -20519
Thrombin-mediated feedback activation of factor XI on the activated platelet surface is preferred over contact activation by factor XIIa or factor Xia.

PAGES 20514 -20519:
The authors are retracting the entire article for the following reasons. All of the authors with the exception of F. A. Baglia retract the paper listed above because recent experiments conducted by Dipali Sinha, Sergei Shikov, Wenman Wu, and Syed Ahmad in the laboratory of Peter N. Walsh failed to confirm the conclusion that activated platelets promote the activation of factor XI by thrombin. A detailed explanation of the chronology of events leading to this retraction and to the retraction of a paper from Biochemistry (Baglia, F. A., and Walsh, P. N. (1998) Prothrombin is a cofactor for the binding of factor XI to the platelet surface and for platelet-mediated factor XI activation by thrombin. Biochemistry 37, 2271-2281) has been published in the journal Biochemistry (manuscript bi-2007-01501k, accepted July 27, 2007). We apologize to the readers, reviewers, and editors of the Journal of Biological Chemistry for publishing these erroneous data.

VOLUME 278 (2003) PAGES 21744 -21750
The glycoprotein Ib-IX-V complex mediates localization of factor XI to lipid rafts on the platelet membrane. All of the authors with the exception of F. A. Baglia retract the specific data listed above because recent experiments conducted by Dipali Sinha, Sergei Shikov, Wenman Wu, and Syed Ahmad in the laboratory of Peter N. Walsh failed to confirm the conclusion that activated platelets promote the activation of factor XI by thrombin. All other results reported in this paper are valid. A detailed explanation of the chronology of events leading to the retraction and the retraction of a paper from Biochemistry (Baglia, F. A., and Walsh, P. N. (1998) Prothrombin is a cofactor for the binding of factor XI to the platelet surface and for platelet-mediated factor XI activation by thrombin. Biochemistry 37, 2271-2281) has been published in the journal Biochemistry (manuscript bi-2007-01501k, accepted July 27, 2007). We apologize to the readers, reviewers, and editors of the Journal of Biological Chemistry for publishing these erroneous data.

VOLUME 278 (2003) PAGES 48112-48119
Thrombin activation of factor XI on activated platelets requires the interaction of factor XI and platelet glycoprotein Ib␣ with thrombin anion-binding exosites I and II, respectively.  Table 1 has been retracted by the authors for the following reasons. All of the authors with the exception of F. A. Baglia retract the specific data listed above because recent experiments conducted by Dipali Sinha, Sergei Shikov, Wenman Wu, and Syed Ahmad in the laboratory of Peter N. Walsh failrd to confirm the conclusion that activated platelets promote the activation of factor XI by thrombin. All of the other results reported in this paper are valid. A detailed explanation of the chronology of events leading to this retraction and the retraction of a paper from Biochemistry (Baglia, F. A., and Walsh, P. N. (1998) Prothrombin is a cofactor for the binding of factor XI to the platelet surface and for platelet-mediated factor XI activation by thrombin. Biochemistry 37, 2271-2281) has been published in the journal Biochemistry (manuscript bi-2007-01501k, accepted July 27, 2007). We apologize to the readers, reviewers, and editors of the Journal of Biological Chemistry for publishing these erroneous data.