Factor XI interacts with the leucine-rich repeats of glycoprotein Ibalpha on the activated platelet.

Factor XI (FXI) binds specifically and reversibly to high affinity sites on the surface of stimulated platelets (Kd app of approximately 10 nm; Bmax of approximately 1,500 sites/platelet) utilizing residues exposed on the Apple 3 domain in the presence of high molecular weight kininogen and Zn2+ or prothrombin and Ca2+. Because the FXI receptor in the platelet membrane is contained within the glycoprotein Ibalpha subunit of the glycoprotein 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), we utilized mocarhagin, a cobra venom metalloproteinase, to generate a fragment (His1-Glu282) of glycoprotein Ibalpha that contains the leucine-rich repeats of the NH2-terminal globular domain and excludes the macroglycopeptide portion of glycocalicin, the soluble extracytoplasmic portion of glycoprotein Ibalpha. This fragment was able to compete with FXI for binding to activated platelets (Ki of 3.125 +/- 0.25 nm) with a potency similar to that of intact glycocalicin (Ki of 3.72 +/- 0.30 nm). However, a synthetic glycoprotein Ibalpha peptide, Asp269-Asp287, containing a thrombin binding site had no effect on the binding of FXI to activated platelets. Moreover, the binding of 125I-labeled thrombin to glycocalicin was unaffected by the presence of FXI at concentrations up to 10(-5) m. The von Willebrand factor A1 domain, which binds the leucine-rich repeats, inhibited the binding of FXI to activated platelets. Thus, we examined the effect of synthetic peptides of each of the seven leucine-rich repeats on the binding of 125I-FXI to activated platelets. All leucine-rich repeat (LRR) peptides derived from glycoprotein Ibalpha were able to inhibit FXI binding to activated platelets in the following order of decreasing potency: LRR7, LRR1, LRR4, LRR5, LRR6, LRR3, and LRR2. However, the leucine-rich repeat synthetic peptides derived from glycoprotein Ibbeta and Toll protein had no effect. We conclude that FXI binds to glycoprotein Ibalpha at sites comprising the leucine-rich repeat sequences within the NH2-terminal globular domain that are separate and distinct from the thrombin-binding site.

Human factor XI (FXI), 1 a homodimeric coagulation protein, circulates in plasma as a complex with a non-enzymatic cofactor high molecular weight kininogen (HK) (1)(2)(3)(4)(5)(6)(7). FXI can bind specifically and reversibly to high affinity sites on the surface of stimulated platelets in the presence of HK (and Zn 2ϩ ions) or prothrombin (and Ca 2ϩ ions) (8,9). We have demonstrated previously that the Apple 3 domain of FXI mediates the binding of FXI to platelets (10,11), activated platelets promote optimal rates of FXI activation by thrombin in the presence of HK or prothrombin (8), and FXI binds to the platelet glycoprotein (GP) Ib-IX-V complex (12) in platelet membrane lipid rafts (13), promoting its activation by thrombin (12,13).
The platelet GPIb-IX-V complex is a large plasma membrane complex (ϳ25,000 copies/platelet) comprising four polypeptide chains, GPIb␣, GPIb␤, GPIX, and GPV, arranged in a stoichiometry of 2:2:2:1, respectively (14). This receptor is responsible for platelet adhesion to the site of injury, a function that it carries out by binding the von Willebrand factor (vWF). The GPIb-IX-V complex also binds thrombin with high affinity. Glycocalicin is the soluble form of the extracellular portion of GPIb␣, which contains the NH 2 -terminal globular domain as well as the macroglycopeptide region. GPIb␣, GPIb␤, GPIbIX, and GPIbV each contain an extracellular domain, a single transmembrane helix, and a short cytoplasmic trail. The binding sites within the complex for both vWF and thrombin reside within the first 300 amino acids of GPIb␣ (15). GPIb␣ is also a member of a family of proteins containing leucine-rich repeat (LRR) sequences including a large number of proteins that are principally involved in mediating protein-protein interactions (16). These LRR sequences are typically 22-28 amino acids long and occur in tandem repeats that are commonly flanked by disulfide loop structures (16). The interaction between vWF and GPIb␣ is mediated by the A1 domain of vWF and the NH 2 -terminal domain of the GPIb␣ chain (17)(18)(19)(20)(21). The x-ray crystal structures of the vWF A1 domain and GPIb␣ have been reported (20,21), as have those of the GPIb␣-thrombin complex (22,23). The GPIb␣ NH 2 -terminal and COOHterminal regions have been implicated in contributing to vWF binding (24,25), whereas thrombin binding is centered around an anionic region containing three tyrosine residues (Tyr 276 , Tyr 278 , and Tyr 279 ) that are post-translationally sulfated in the native receptor. This modification is essential for GPIb␣ binding to thrombin and may contribute to the interaction with vWF (26,27).
Thus, the biological relevance and structural determinants of GPIb␣ interactions with thrombin and the A1 domain of vWF are relatively well understood. In contrast, although GPIb␣ appears to comprise the platelet receptor and physiological locus for the FXI colocalization with thrombin leading to initiation of the consolidation pathway of blood coagulation (8 -13), relatively little is known about the structural determinants of FXI interactions with the GPIb-IX-V complex and its colocalization with thrombin. The present studies demonstrate that FXI and thrombin bind GPIb␣ at distinct sites, which suggest that GPIb␣ may allow the colocalization of FXI and thrombin for efficient activation of FXI on the platelet surface.  (Table I) were 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 (Foster City, CA). The potent thrombin inhibitor prolyl-phenylalanyl-arginyl-chloromethylketone (PPACK) was purchased from Calbiochem (Indianapolis, IN). Active site-inhibited thrombin was prepared by incubation of a 10-fold excess of PPACK with ␣-thrombin for 1 h at 37°C, and this mixture was dialyzed with Spectrophor tubing (3,500 M r cut off; Spectrum Medical Industries, Los Angeles, CA) overnight in phosphate-buffered saline at FIG. 1. Effect of glycocalicin and synthetic peptides on the binding of 125 I-FXI to platelets. The effects of glycocalicin fragments and synthetic peptides were examined, including glycocalicin (•), His 1 -Glu 282 (mocarhagin fragment) (‚), Thr 266 -Asp 287 (Ⅺ), and Asp 269 -Pro 280 (¹). 125 I-FXI (22 nM), gelfiltered platelets (1 ϫ 10 8 platelets/ml), ZnCl 2 (25 M), CaCl 2 (2 mM), thrombin peptide (25 M), and HK (42 nM) were incubated for 30 min at 37°C either with the designated peptide at the indicated concentration or with buffer solution. Aliquots were removed and centrifuged as described under "Experimental Procedures." Each point is an average of triplicate determinations. When 125 I-FXI was incubated with platelets at time 0 the amount of 125 I-FXI bound was Ͻ1% of the control value, and the maximum variation of counts/min bound for each observation was Ͻ2% of total counts per minute bound. One hundred percent binding of FXI represents an average of 100,250 cpm bound, whereas 0% binding of FXI represents 0% bound after subtracting 229 cpm, representing the control in which labeled FXI was incubated with platelets at time 0.

Materials-Human
FIG. 2. The binding of 125 I-FXI to glycocalicin or the mocarhagin fragment in a solid phase assay. 125 I-FXI at various concentrations was incubated in wells containing bound glycocalicin and 25 M ZnCl 2 (•) or glycocalicin only (¹) or a bound mocarhagin fragment and 25 M ZnCl 2 (⌬) or the bound mocarhagin fragment only (OE). The assay was performed and described under "Experimental Procedures." When glycocalicin or the mocarhagin fragment was not bound to the wells of microtiter plates, the amount of 125 I-FXI was Ͻ0.1% of the control value, and the maximum variation for each experimental observation was Ͻ2% of the total labeled FXI bound. Nonspecific binding, as represented by 125 I-FXI binding to wells coated with BSA (ϳ350 cpm), was subtracted from the total binding at each data point. Results shown represent the mean Ϯ S.E. of three experiments each done in duplicate.

5°C. All purified proteins appeared homogeneous after SDS-PAGE.
Radiolabeling of FXI-Purified FXI and PPACK-thrombin were radiolabeled with 125 I by a minor modification (8) of the Iodogen method to a specific activity of ϳ5 ϫ 10 6 cpm/g and ϳ1 ϫ 10 6 cpm/g, respectively. The radiolabeled FXI retained Ͼ98% of its biological activity.
Protein Analysis-Protein concentrations were determined by the Bio-Rad dye-binding assay according to the instructions provided by the manufacturer (Bio-Rad Laboratories).
Preparation of Glycocalicin and Mocarhagin Fragment His 1 -Glu 282 from Human Platelets-Glycocalicin was extracted from human platelets and purified as described previously (29). The mocarhagin fragment His 1 -Glu 282 was prepared from intact glycocalicin and purified as described previously (30).
Platelet Binding Experiments-Platelets were pre-warmed 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 recep-tor (PAR-1) activation peptide (SFLLRN-amide) as a platelet agonist (9,10,28), and HK or other proteins as designated in the legends to Figs. 1, 5, and 6. 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 (8 -10, 28). In competition binding experiments, the concentration of the competitor that displaced 50% of the bound 125 I-FXI (IC50) was determined by plotting the amount of 125 I-FXI bound to platelets versus the amount of competitor ligand added. The K i was calculated using the equation is the concentration of 125 I-FXI used in these experiments (held constant at 22 nM) and the K d was the value (ϳ10 nM) determined from direct binding experiments.
Solid Phase Binding of 125 I-FXI or 125 I-PPACK Thrombin to Glycocalicin-We utilized a modification of the method of DeCristofaro et al. (31) to examine the binding of 125 I-FXI to platelet-bound glycocalicin. Wheat germ lectin (10 g/ml) was coated on the wells of a solid phase plate (96-well polystyrene trays; Immulon high protein capacity binding) and incubated overnight at 4°C in 50-mM carbonate buffer, pH 9.50. The remaining binding capacity of the sample wells was blocked by incubation for 2 h with 1% BSA in Hepes-buffered saline (150 mM NaCl and 3.5 mM Hepes, pH 7.2). After aspiration of the BSA solution, purified glycocalicin was added to the wells at a concentration of 20 g/ml and incubated at 4°C for 1 h. After aspiration, the 125 I-FXI or 125 I-PPACK thrombin was applied to the wells and incubated for 1 h at 37°C. Each sample and blank well were washed with Hepes-buffered FIG. 3. Effect of the mocarhagin fragment and synthetic peptides on the binding of 125 I-FXI to glycocalicin in a solid phase assay. The effects of the mocarhagin fragment and synthetic peptides were examined, including the mocarhagin fragment His 1 -Glu 282 (‚), Thr 266 -Asp 287 (Ⅺ), and Asp 269 -Pro 280 (¹). 125 I-FXI (22 nM) was incubated with either the mocarhagin fragment or synthetic peptides, and the assay was performed as described under "Experimental Procedures." When glycocalicin was not bound to the wells of microtiter plates the amount of 125 I-FXI bound was Ͻ0.1% of the control value, and the maximum variation of counts per minute bound for each experimental observation was Ͻ2% of the total counts per minute bound. One hundred percent binding of 125 I-FXI represents binding in the absence of fragment or synthetic peptides (mean ϭ 29,122 cpm). Nonspecific binding, as represented by 125 I-FXI binding to wells coated with BSA (mean ϭ 492 cpm), was subtracted from the total binding at each data point. Results shown represent the mean Ϯ S.E. of three experiments, each done in duplicate.  (Table I) was accomplished by the method of Kitagawa et al. (32).
Expression of Wild-type and the Type 2B Mutant A1 Domain of vWF-The 508 -709 recombinant fragments of vWF with either the wild-type sequence or the Ile 546 -Val type 2B mutation were expressed and purified as described previously (33,34).

The Effect of Glycocalicin and Mocarhagin Fragments on the
Binding of FXI to Activated Platelets-Previous studies have determined that the FXI binding site on platelets is in the GPIb␣ subunit of the GPIb-IX-V complex for the following reasons: 1) because two GPIb␣ ligands, SZ2 (a monoclonal antibody agonist the-NH 2 terminal of GPIb␣) and bovine vWF, inhibit FXI binding to platelets; 2) because FXI was shown by surface plasmon resonance to bind specifically to glycocalicin in a Zn 2ϩ -dependent fashion; and 3) because glycocalicin could promote FXI activation by thrombin, another GPIb␣ ligand (12). It has been determined that a sulfated tyrosine/anionic sequence, Tyr 276 -Glu 282 , of GPIb␣ comprises a binding site for vWF and thrombin (21, 22, 24 -27) and that the LRRs of GPIb␣ are important for binding vWF (26,27). To determine the binding site for FXI in GPIb␣, we utilized mocarhagin, a cobra venom metalloproteinase, that cleaves GPIb␣ at Glu 282 -Asp 283 to generate a fragment, His 1 -Glu 282 , which contains the LRRs and the sulfated tyrosine/anionic sequences Tyr 276 -Glu 282 and excludes the carbohydrate-rich macroglycopeptide. We examined the effect of glycocalicin and the His 1 -Glu 282 fragment on the binding of FXI to activated platelets. Fig. 1 shows that the His 1 -Glu 282 fragment competed for binding sites on FXI-activated platelets with an IC 50 similar to that for glycocalicin (10 nM Ϯ 0.65), suggesting that FXI binds to the NH 2 -terminal portion of glycocalicin. Also, neither of the two synthetic peptides with sulfated tyrosines (Thr 266 -Asp 287 and Asp 269 -Pro 280 ) had any effect on the binding of FXI to activated platelets, suggesting that FXI does not bind to the anionic domain containing sulfated tyrosines that is utilized for binding thrombin and vWF. FXI Binding to Glycocalicin or the Mocarhagin Fragment in the Presence of ZnCl 2 -To confirm the notion that FXI is binding to the NH 2 -terminal region of GPIb␣, we performed direct binding experiments in a solid phase assay. Fig. 2 shows that FXI binds to the fragment His 1-Glu 282 with the same K d app (ϳ10 nM Ϯ 0.9) as intact glycocalicin in the presence of ZnCl 2 , whereas in the absence of ZnCl 2 no FXI-binding to either glycocalicin or the His 1 -Glu 282 fragment was observed. This confirms the conclusion that the NH 2 -terminal globular domain of GPIb mediates the binding of FXI to GPIb-IX-V␣.
The Effect of the His 1 -Glu 282 Fragment and Sulfated Peptides on the Binding of FXI to Glycocalicin-To further confirm our findings on the effect of the His 1 -Glu 282 fragment and sulfated peptides on FXI binding to activated platelets, we examined their effect on the binding of FXI to glycocalicin in a solid phase assay. Fig. 3 shows that the His 1 -Glu 282 fragment inhibited the binding of FXI to glycocalicin with an IC 50 of 20 nM Ϯ 1.8, whereas the sulfated peptides Thr 266 -Asp 287 and Asp 269 -Pro 280 had no effect.
The Effect of FXI on the Binding of Thrombin to Glycocalicin-Because both FXI and thrombin bind to the NH 2 -terminal globular domain of GPIb␣ and because FXI interacts with thrombin through anion-binding exosite I (35), we determined whether thrombin and FXI share a common binding site on GPIb␣. Fig. 4 shows that FXI at concentrations up to 4 ϫ 10 Ϫ5 M does not inhibit the binding of thrombin to glycocalicin, which suggests that FXI binds glycocalicin at sites separate and dis-tinct from those bound by thrombin.
The Effect of PPACK Thrombin on the Binding of FXI to Glycocalicin-Because FXI did not compete with thrombin for binding to glycocalicin, we examined the effect of PPACK thrombin on the binding of FXI to glycocalicin in the presence of 25 M ZnCl 2 . Our results show that PPACK thrombin at concentrations up to 3 ϫ 10 Ϫ5 M does not affect the binding of FXI to glycocalicin (data not shown), confirming the conclusion that FXI and thrombin bind to GPIb␣ at separate and distinct sites.
The Effects of the A1 Domain (Wild-type and I546V Type 2B Mutant) of vWF on the Binding of FXI to Activated Platelets-It has been determined that an interaction between the vWF-A1 domain and platelet GPIb␣ occurs in the presence of high shear stress or when vWF becomes immobilized on a surface, but not under static conditions in free solution (36). This interaction can be induced by exogenous non-physiologic modulators such as ristocetin (37,38) and by fluid shear stress (39,40), or it can be the consequence of selected mutations in the A1 domain (such as the I546V vWD type 2B mutant). Thus, the I546V A1 domain mutant binds with high affinity to GPIb␣ in the absence of modulators. To further investigate the FXI-GPIb␣ interaction, we asked whether the A1 domain of vWF affects FXI binding to activated platelets. I-FXI bound was Ͻ1% of the control value, and the maximum variation of counts per minute bound for each experimental observation was Ͻ2% of the total counts per minute bound. One hundred percent binding of FXI represents an average of 98,910 cpm bound, whereas 0% binding to FXI represents 0% bound after subtracting 150 cpm, representing the control in which 125 I-FXI was incubated with platelets at time 0. ent. Thus, as expected in the absence of ristocetin, the wildtype A1 domain did not compete with FXI for binding to activated platelets. In contrast, the vWF type 2B mutant does not require a modulator to bind GPIb␣ and was able to compete with FXI for binding to activated platelets (IC 50 , 60 nMϮ 5.8; K i , 18.8 Ϯ 2.0 nM) in the absence of ristocetin.
The Effect of LRRs on the Binding of FXI to Activated Platelets-We have determined that bovine vWF, which binds to GPIb␣ in the absence of ristocetin, inhibits the binding of FXI to activated platelets (12). Because vWF interacts with the leucine-rich repeat sequences 1, 2, 3, and 4 and may also interact with the sulfated tyrosine/anionic sequence (Tyr 276 -Glu 282 ) of GPIb␣ (24,25) and because bovine vWF and the vWF-A1 domain in the presence of ristocetin (Fig. 5) (but not the sulfated peptides, Thr 266 -Asp 287 or Asp 269 -Pro 280 ; Fig. 1) inhibit the binding of FXI to activated platelets (glycocalicin), we determined whether FXI also interacts with the LRRs of GPIb␣. Synthetic peptides corresponding to the sequence of each LRR (20 -24 amino acids) were prepared (Table I) and examined for their effects on the binding of FXI to activated platelets. Fig. 6 shows that all LRR peptides were able to inhibit FXI binding to activated platelets in the following order of decreasing potency: LRR7 (Leu 177 -Ala 200 , IC 50 of 2 ϫ 10 Ϫ8 M); LRR1 (Leu 36 -Leu 59 , IC 50 of 2.5 ϫ 10 Ϫ8 M); LRR4 (Leu 105 -Glu 128 , IC 50 of 4.5 ϫ 10 Ϫ8 M); LRR5 (Leu 129 -Lys 152 , IC 50 of 6 ϫ 10 Ϫ8 M); LRR6 (Leu 152 -Thr 176 , IC 50 of 7.5 ϫ 10 Ϫ8 M); LRR3 (Leu 82 -Val 104 , IC 50 of 9 ϫ 10 Ϫ8 M); and LRR2 (Leu 60 -Thr 81 , IC 50 of 1.25 ϫ 10 Ϫ7 M). However, a scrambled peptide composed of the same amino acids present in LRR7 did not inhibit the binding of FXI to the activated platelet (Tables I and II and Fig.  6). Finally, neither the LRR sequences from GPIb␤ (41) nor the Toll protein (Drosophila-derived protein) (42) were able to inhibit FXI binding to activated platelets ( Fig. 6 and Table I), thereby confirming the specificity of FXI interaction with the LRR of GPIb␣.

DISCUSSION
The platelet GPIb-IX-V complex is involved in several activities crucial to normal platelet function, including the initial adhesion of platelets to the exposed subendothelium, the regulation of certain cytoskeletal properties such as actin polymerization, and the response of platelets to low concentrations of thrombin (14). FXI interacts with the GPIb-IX-V complex on the platelet surface, and this interaction promotes thrombincatalyzed FXI activation (12). We have also demonstrated that FXI can interact with lipid rafts on the platelet surface and that this interaction is mediated through the GPIb-IX-V complex (13). The evidence that the FXI binding site on activated platelets consists of the GPIb␣ subunit is as follows. 1) Bernard-Soulier platelets, lacking the GPIb-IX-V complex, are deficient in FXI binding. 2) Two GPIb␣ ligands, SZ2 (a mono-clonal antibody) and bovine vWF, inhibit FXI binding to platelets; and 3) FXI interacts specifically with glycocalicin by surface plasmon resonance (12). It has been determined that the 45-kDa NH 2 -terminal domain of GPIb␣ is involved in binding thrombin primarily through a stretch of negatively charged residues, including sulfated tyrosines spanning amino acids Gly 268 -Glu 282 (15). The NH 2 -terminal portion of GPIb␣ (His 1 -Glu 282 ) also contains the LRRs, which have been implicated in regulating the binding of vWF to platelet GPIb-IX-V (16). Consistent with this conclusion, a single amino acid substitution of proline for Leu 129 within the fifth LRR is associated with the congenital bleeding disorder Bernard-Soulier syndrome in which GPIb␣ is expressed in a dysfunctional form that does not bind vWF (43). In the present study we aimed to determine the subdomains within GPIb␣ that bind FXI and regulate the activation of FXI by thrombin on the platelet surface. Aliquots were removed and centrifuged as described under "Experimental Procedures." Each point is an average of triplicate determinations. 125 I-FXI was incubated with platelets at time 0 (at the start of the incubation), the amount of 125 I-FXI bound was Ͻ1% of the control value, and the maximum variation of counts per minute bound for each observation was Ͻ2% of total counts per minute bound. One hundred percent binding of FXI represents an average of 99,565 cpm bound, whereas 0% binding of FXI represents 0% bound after subtracting 198 cpm, representing the control in which labeled FXI was incubated with platelets at time 0.   FXI binding to activated platelets, suggesting that FXI may interact with a concave surface formed by the LRR. However, the fact that LRR1 and LRR7 are the most potent inhibitors of FXI binding to platelets suggests that the major interaction of FXI with GPIb␣ involves LRR1 and LRR7. Unlike the vWF-A1 domain interaction with GPIb␣, the anionic region does not appear to be involved in FXI interaction with GPIb␣, because the sulfated peptides (Thr 266 -Asp 287 and Asp 269 -Pro 280 ) do not inhibit either FXI interaction with activated platelets or with glycocalicin ( Figs. 1 and 3). Because the vWF-A1 domain inhibits FXI interaction with activated platelets (Fig. 5), it is possible that FXI may interact with some of the same negative amino acids to which vWF binds in the LRRs of GPIb␣. We have determined the specific amino acids within the FXI-Apple 3 domain that interact with platelets and with GPIb␣ (28,44). A logical extension of the present study will be to identify the amino acid residues within the LRRs of GPIb␣ that interact with the Apple 3 domain of FXI.
In agreement with our finding that FXI interacts with the LRRs of GPIb␣ is our previous study of the interaction of FXI with activated platelets obtained from two patients with the Bernard-Soulier syndrome (12). The two subjects of these studies (T. H. and A. H.) are first cousins (both homozygotes) whose platelets have been studied by various laboratories. Platelets from both patients exhibited ϳ40% of normal vWF binding, 43% of normal FXI binding, and ϳ40% of normal GPIb-IX surface antigen (12). The mutation in the GPIb␣ gene that results in the substitution of proline for leucine at position 129 is located in the first position of the fifth LRR of the mature gene product. This mutation affects the conformation of GPIb␣ and alters FXI and vWF binding to it.
Our previous data suggest a quaternary complex model of thrombin activation of FXI on the activated platelet surface (35). Thrombin binds to the GPIb-IX-V complex via the anion binding exosite-II (ABE-II) on the posterior surface of thrombin and to FXI (A1 domain) via ABE-I on its anterior face (35). At least one thrombin-binding site of GPIb␣ has been localized to residues Asp 269 -Asp 287 (45) and is mediated by electrostatic interactions. FXI binding to GPIb␣ LRRs through its Apple 3 domain is achieved when one monomer of the homodimer is first complexed with HK (10,28). Once bound to the GPIb-IX-V complex by one of its monomers, FXI can interact with an adjacent thrombin molecule through the A1 domain of its other free monomer (12). Thus, thrombin, bound to one GPIb␣ molecule via ABE-II on its posterior surface, is properly oriented for the activation of FXI bound to a neighboring GPIb␣ molecule (via LRR) (Fig. 7). In this model, separate and distinct sites for FXI binding (through the LRRs) and thrombin binding through residues (Asp 269 -Asp 287 ) to GPIb␣ play a major role in colocalizing thrombin and FXI to form a quaternary complex on the platelet surface essential for efficient activation.