Glycoprotein Ib -Mediated Platelet Adhesion and Aggregation to Immobilized Thrombin Under Conditions of Flow

【摘要】  Objectives- Thrombin interacts with platelets via the protease-activated receptors (PARs) 1 and 4, and via glycoprotein Ib (GPIb ). Recently, it was shown that platelets are able to adhere to immobilized thrombin under static conditions via GPIb.

Methods and Results- Here, we show that platelets are also able to adhere to and form stable aggregates on immobilized thrombin under conditions of flow. Adhesion and aggregation to thrombin was dependent on the interaction with GPIb, as addition of glycocalicin or an antibody blocking the interaction between thrombin and GPIb inhibited platelet adhesion. Additionally, platelet adhesion to recombinant thrombin mutants, which are unable to bind GPIb, was severely suppressed. Furthermore, platelet adhesion to thrombin was dependent on activation of PARs, and partly on granule secretion and thromboxane-A2 synthesis. Immobilization of thrombin on a fibrin network resulted in substantially increased adhesion compared with fibrin alone. The adhesion to fibrin alone was completely abolished by addition of dRGDW, whereas fibrin-bound thrombin still showed substantial platelet adhesion in the presence of dRGDW, indicating that fibrin-bound thrombin is able to directly capture platelets under flow.

Conclusion- These results indicate that platelets are able to adhere to thrombin under flow conditions, which is dependent on the interaction with GPIb.

Recently, it was shown that platelets are able to adhere to immobilized thrombin under static conditions via GPIb. Flow studies reveal that platelets are also able to adhere to thrombin immobilized on fibrin or directly on a glass coverslip, which is dependent on the interaction with GPIb.

【关键词】  thrombin platelet fibrin GPIb flow conditions

Introduction

Thrombin has a central role in hemostasis. It activates platelets, cleaves fibrinogen into fibrin, and activates factor XIII. Furthermore, thrombin enhances coagulation by activating factors V, VIII, and XI, but it also inhibits coagulation by activating protein C, and attenuates fibrinolysis by activating thrombin activatable fibrinolysis inhibitor (TAFI). On clot formation, thrombin is immobilized to the fibrin clot, 1 and this binding to fibrin could be important in localizing thrombin to the site of vascular injury. Fibrin-bound thrombin is protected against inactivation by the heparin-antithrombin complex, 2,3 but the active site still remains accessible, as fibrin-bound thrombin is still capable of cleaving fibrinogen and activating factor XI. 4

Thrombin can activate platelets via the protease-activated receptors (PARs) PAR1 and PAR4, which are generally assumed to account for the moderate- and low-affinity binding sites for thrombin, respectively. 5 GPIb is described to be the high-affinity receptor for thrombin. 6,7 GPIb consists of 2 subunits, GPIb and GPIbß, and is expressed in platelets as a complex with GPIX and GPV in a 2:2:2:1 stoichiometry. However, there are &25 000 copies of GPIb on the platelet surface, but only a small number (&100 to 1000) appear to be involved in the high-affinity binding of thrombin. 8 The localization of the GPIb-IX-V complex in rafts has proved to be important in platelet activation by von Willebrand Factor (vWF), 9 and it has been postulated that raft association may also account for the difference in high-affinity binding sites for thrombin and GPIb copies on the platelet. 10

Thrombin contains 2 anion binding sites or exosites referred to as exosite I and exosite II, a catalytic pocket and a Na + binding site. Exosite I is important in the binding of multiple substrates, including fibrin and fibrinogen 11 and PAR1, 12 whereas exosite II is referred to as the heparin binding site. 13 The catalytic pocket is responsible for the actual cleavage of the substrates, and the amount of Na + bound to the Na + binding site regulates the affinity of thrombin for its substrates (reviewed by Di Cera 14 ). Recently, site-directed mutagenesis has indicated the involvement of many basic exosite II residues in GPIb binding. 15,16 In addition, the crystal structures of thrombin bound to GPIb reported by Celikel et al 17 and Dumas et al 18 revealed the importance of both exosites of GPIb in the binding of thrombin. Although there were many discrepancies between the 2 structures, which resulted in fundamentally different functional interpretations, both structures showed that 2 thrombin molecules can interact with a single GPIb molecule, one via exosite I and the other via exosite II. It was proposed that the first thrombin molecule binds via its exosite II to GPIb, which is followed by a conformational change in GPIb, after which a second thrombin molecule can bind via its exosite I to a different location on GPIb. The interpretation of Celikel involved dimerization of 2 GPIb molecules on the same platelet via thrombin, whereas Dumas described the possibility of aggregation of platelets via 2 GPIb molecules bridged by thrombin (reviewed by Sadler 19 ).

It was previously shown that platelets are able to adhere to immobilized thrombin under static conditions. 20 However, it is unclear whether this interaction is sufficiently strong to resist shear forces. In this study, we investigated whether platelets could interact with immobilized thrombin under flow conditions. Furthermore, the role of fibrin-bound thrombin is not yet completely understood, and in this study we investigated whether fibrin-bound thrombin contributes to platelet adhesion. We show that thrombin immobilized either directly on a glass coverslip or on fibrin induces platelet adhesion and aggregate formation under flow conditions, which is dependent on its interaction with GPIb.

Methods

For Methods, please see the data supplement, available online at http://atvb.ahajournals.org.

Results

Platelet Adhesion and Aggregate Formation to Immobilized Thrombin Under Conditions of Flow

To investigate whether thrombin immobilized on a surface is able to interact with platelets under conditions of flow, reconstituted blood was perfused over immobilized thrombin for 5 minutes at a shear rate of 300 s -1. Real time perfusion experiments showed rapid adhesion of single platelets to the surface, followed by the formation of large aggregates at sites of primary platelet adhesion (for movie capture, please see http://atvb.ahajournals.org). The aggregates were stable, and embolization only occurred sporadically. Figure 1 A shows a microscopic picture of platelet adhesion and aggregate formation to immobilized thrombin after 5 minutes of perfusion at a shear rate of 300 s -1. Scanning Electron Microscope analysis supports our observations that large stable aggregates were formed on top of initially spread platelets ( Figure 1 B). Platelet adhesion to coverslips that were coated only with blocking buffer (BSA) was virtually absent ( Figure 2 ). In contrast to platelet adhesion to vWF, 21 platelets did not roll on immobilized thrombin before firm adhesion, but rather attached instantly. Perfusion experiments using a range of shear rates (100 s -1 to 4000 s -1 ) showed platelet adhesion at all shear rates tested, with optimal adhesion at a shear rate of 300 s -1 (data not shown). For further perfusion experiments, a shear rate of 300 s -1 was used, which is comparable to venous shear rates.

Figure 1. Platelet adhesion to immobilized thrombin. Reconstituted blood was perfused over a coverslip coated with 25 µg/mL thrombin at a shear rate of 300 s -1 for 5 minutes at 37°C using a single-pass perfusion chamber. After perfusion, coverslips were fixed and stained with May-Grünwald/Giemsa and examined by light microscopy (original magnification 400 x; A), or coverslips were fixed and examined by Scanning Electron Microscopy (original magnification, 2500 x; B). Representative images of at least 6 independent experiments are shown.

Figure 2. Determinants of platelet adhesion to immobilized thrombin. A, Reconstituted blood was perfused over immobilized thrombin (IIa) in the presence of agents interfering with thrombin binding to GPIb (left section), platelet activation via PARs (mid section), or other platelet receptors (right section). After perfusion, coverslips were stained with May-Grünwald/Giemsa and examined by light microscopy. Graph shows mean surface coverage of at least 3 independent experiments performed in triplicate. ** P <0.01. Error bars indicate standard deviation. B, Reconstituted blood was perfused over immobilized thrombin in the absence (top) or presence (bottom) of dRGDW (200 µmol/L). After perfusion, coverslips were fixed and stained with May-Grünwald/Giemsa and examined by light microscopy (original magnification 400 x ). Representative images of at least 3 independent experiments are shown.

Platelet Adhesion to Immobilized Thrombin Is Mediated by GPIb

To investigate whether GPIb is involved in platelet adhesion to immobilized thrombin, we perfused reconstituted blood over immobilized thrombin in the presence of an antibody directed against the thrombin-binding site of GPIb (LJIb-10, 100 µg/mL). As shown in Figure 2 A, surface coverage is substantially and significantly reduced on addition of LJIb-10. Also, addition of glycocalicin (GC, 50 µg/mL), a proteolytic fragment of the extracellular domain of GPIb, inhibited platelet adhesion to immobilized thrombin ( Figure 2 A). Platelet adhesion to immobilized thrombin was not dependent on the GPIb-vWF interaction, as an antibody against vWF (RAG-35, used in a dilution of 1:250), interfering with the interaction of vWF with GPIb, did not show a reduction in platelet adhesion ( Figure 2 A).

Platelet Adhesion to Immobilized Thrombin Requires Activation of PARs, Secretion of ADP and Thromboxane A2 Synthesis

Next, we investigated the role of PAR1 and PAR4 in platelet adhesion to immobilized thrombin. Platelet adhesion to immobilized thrombin could be blocked by addition of an inhibitory antibody against PAR1 ( Figure 2 A). Platelets desensitized for either PAR1 or PAR4 with the PAR1 activating peptide SFLLRN (15 µmol/L, 30 minutes, 37°C) or the PAR4 activating peptide GYPGQV (1 mmol/L, 30 minutes, 37°C) had a reduced capacity to adhere to immobilized thrombin ( Figure 2 A). Furthermore, platelet adhesion to thrombin was almost completely abolished when thrombin was preincubated for 30 minutes at 37°C with 50 µmol/L PPACK, which blocks the active site of thrombin.

Inhibitors of the ADP-receptors P2Y1 (A3P5P, 300 µmol/L) and P2Y12 (AR-C69931MX, 1 mmol/L) partly inhibited platelet adhesion and aggregate formation to immobilized thrombin ( Figure 2 A). Also, the thromboxane-receptor analog SQ30741 (10 µmol/L) partly inhibited platelet aggregation to immobilized thrombin. Platelet adhesion to immobilized thrombin did not involve the integrin IIb ß 3, as platelets pretreated with dRGDW (200 µmol/L), a peptide which blocks ligand binding to IIb ß 3, readily adhered to thrombin. Although in the presence of dRGDW aggregate formation was absent, spreading features such as pseudopod formation and filopodia extension were present ( Figure 2 B).

Platelet Adhesion to Immobilized Thrombin Is Dependent on the Interaction Between Thrombin Exosite II and GPIb

To investigate the interaction between thrombin exosites and the role of GPIb in the adhesion of platelets to immobilized thrombin, we performed perfusion experiments using the recombinant thrombin (exosite II) mutants R98A and R89A/R93A/E94A, which were previously shown to lack the ability to bind GPIb. 22 When reconstituted blood was perfused over these recombinant thrombin mutants, platelet adhesion was almost completely absent, whereas platelet adhesion and aggregation to recombinant wild-type thrombin was similar to that observed using plasma-derived thrombin ( Figure 3 ). Also, platelet adhesion to a recombinant thrombin with a mutation in the Na + binding site (E229A) was abolished. Y71A, which has a reduced binding capacity for exosite I ligands such as fibrinogen, 11 supported platelet adhesion and aggregation comparable to wild-type thrombin ( Figure 3 ).

Figure 3. Platelet adhesion to immobilized thrombin is dependent on exosite II. Reconstituted blood was perfused for 5 minutes at a shear rate of 300 s -1 over a coverslip coated with 25 µg/mL wild-type (WT) thrombin or 25 µg/mL thrombin mutant. After perfusion, coverslips were stained with May-Grünwald/Giemsa and examined by light microscopy. Graph shows mean surface coverage of at least 3 independent experiments performed in triplicate. ** P <0.01. Error bars indicate standard deviation.

To investigate whether the decrease in platelet adhesion is caused by a reduced potential to activate platelets, platelets were aggregated in suspension with the recombinant thrombin mutants. Details of these experiments will be published elsewhere (Myles et al, unpublished data, 2005). The platelet aggregatory potential of Y71A (EC50=4.3±0.1 nmol/L), R98A (EC50=9.7±0.3 nmol/L), and R89A/R93A/E94A (EC50=3.0±0.4 nmol/L) was moderately reduced compared with wild-type (EC50=1.2±0.2 nmol/L). The observation that Y71A and R89A/R93A/E94A have a similar EC50 for aggregation in suspension, but that only R89A/R93A/E94A has a substantially decreased response to immobilized thrombin under flow, indicates that flow-mediated adhesion requires exosite II but not exosite I. However, E229A showed an EC50-value of 39.9±0.1 nmol/L, again indicating that the so-called slow form of thrombin has a severely impaired capacity to activate PAR1.

Fibrin-Bound Thrombin Contributes to Platelet Adhesion and Aggregate Formation

Subsequently, we investigated platelet adhesion to immobilized fibrin and fibrin-bound thrombin. Reconstituted blood was perfused for 5 minutes at a shear rate of 300 s -1 over fibrin-coated coverslips, which were incubated with thrombin (25 µg/mL) or vehicle. Platelets readily adhered to fibrin as shown in Figure 4 A. Platelet adhesion and aggregate formation substantially increased on fibrin with bound thrombin compared with fibrin alone as shown in Figure 4 B and increased with increasing thrombin concentrations with half-maximum effect obtained at 15 µg/mL thrombin and maximum effect reached at 50 µg/mL thrombin (data not shown). Platelet adhesion to fibrin is fully dependent on IIb ß 3, and therefore after addition of dRGDW (200 µmol/L) adhesion was abolished ( Figure 4 C). However, Figure 4 D shows that in the presence of dRGDW platelets did adhere to fibrin-bound thrombin, indicating that fibrin-bound thrombin is able to directly bind platelets. Figure 4 E shows the surface coverage results of Figure 4A through 4 D.

Figure 4. Fibrin-bound thrombin contributes to platelet adhesion and aggregation. Reconstituted blood was perfused for 5 minutes at shear rate of 300 s -1 over a fibrin-coated coverslip in the absence (A) or presence (C) of dRGDW (200 µmol/L), or over a fibrin-coated coverslip, which was incubated with 25 µg/mL thrombin in the absence (B) or presence (D) of dRGDW (200 µmol/L). After perfusion, coverslips were stained with May- Grünwald/Giemsa and examined by light microscopy (original magnification 400 x ). E, Surface coverage results of A through D. Graph shows mean surface coverage of at least 3 independent experiments performed in triplicate. * P <0.05; ** P <0.01. Error bars indicate standard deviation.

To study whether platelet adhesion to fibrin with thrombin is also dependent on GPIb, platelets were pretreated with the snake venom Nk (5 µg/mL), which sheds GPIb from the platelet surface. Platelet adhesion to fibrin alone is partially inhibited by Nk, indicating the involvement of GPIb in adhesion to fibrin ( Figure 5 ), which is in agreement with experiments performed by Hantgan et al, who showed that platelet adhesion to fibrin is in part dependent on GPIb. 23 The increase in adhesion of fibrin-bound thrombin was inhibited completely by Nk, as seen in Figure 5. Furthermore, PPACK and an inhibitory antibody against PAR1 inhibited the increase in surface coverage obtained by exposing the fibrin to thrombin ( Figure 5 ).

Figure 5. Platelet adhesion to fibrin-bound thrombin is dependent on GPIb and PAR1. Reconstituted blood was perfused over fibrin-bound thrombin (Fb+IIa) or fibrin alone (Fb) after pretreatment with Nk (left section), PPACK (mid section), or an inhibitory antibody against PAR1 (right section). After perfusion, coverslips were stained with May-Grünwald/Giemsa and examined by light microscopy. Graph shows mean surface coverage of at least 3 independent experiments performed in triplicate. Adhesion to fibrin alone is indicated as 100% relative coverage. ** P <0.01; *** P <0.001; ns indicates not significant. Error bars indicate standard deviation.

Recombinant thrombin mutants could also increase adhesion to fibrin as compared with wild-type thrombin (please see Figure I, available online at http://atvb.ahajournals.org). However, only wild-type thrombin was capable of inducing large aggregates, whereas immobilization of thrombin mutants only increased surface coverage. The thrombin mutant E229A could not contribute to platelet adhesion to fibrin but also did not respond in aggregation experiments. In the presence of dRGDW, platelet adhesion to thrombin mutants R98A, R89A/R93A/E94A, and E229A was significantly reduced compared with wild-type, indicating that platelet adhesion to fibrin-bound thrombin also requires binding of GPIb and activation of PARs (please see Figure I, available online at http://atvb.ahajournals.org). Also, Y71A failed to induce platelet adhesion on fibrin in the presence of dRGDW. In the presence of dRGDW, platelet adhesion to the recombinant thrombin mutants was not significantly increased compared with fibrin alone.

Discussion

This study shows that thrombin immobilized on a coverslip or on fibrin is able to capture platelets under conditions of flow. The capacity of thrombin to function as a platelet adhesive protein has not been recognized previously. Perfusion of reconstituted blood over immobilized thrombin resulted in rapid platelet adhesion and the formation of large stable aggregates. Platelet adhesion was shown to be dependent on GPIb and the proteolytic activity of thrombin. We propose the following sequence of events leading to the formation of a stable aggregate when thrombin is immobilized on a surface. Immobilized thrombin is able to capture platelets from flowing blood via GPIb. Subsequently, intracellular signaling occurs in response to thrombin binding to GPIb and activation of PAR1 and PAR4, resulting in the formation of thromboxane A2 and secretion of ADP and the activation of IIb ß 3. These processes are responsible for the stable adhesion to thrombin and the formation of aggregates.

Activation of IIb ß 3 is not mandatory for primary platelet adhesion to thrombin, as platelets readily adhere in the presence of dRGDW, whereas signal transduction via PARs, thromboxane A2, and ADP is required for primary adhesion. These processes result in inside-out signaling to GPIb 24 or relocation of GPIb into lipid rafts, 9 which we hypothesize to be required for a firm GPIb -thrombin interaction. However, we cannot exclude that other (unknown) receptors contribute to stable platelet adhesion to thrombin.

Platelet adhesion to immobilized thrombin is dependent on the interaction between thrombin exosite II and GPIb on the platelet surface. This is demonstrated by the fact that antibodies against the thrombin binding site on GPIb inhibit platelet adhesion to immobilized thrombin. Furthermore, recombinant thrombins with mutations in exosite II, which virtually abolish the interaction with GPIb, did not induce platelet adhesion and aggregate formation when immobilized directly on a coverslip, whereas the exosite I mutant does support adhesion. Taken together, these results provide strong evidence that exosite II is essential for platelet adhesion to immobilized thrombin mediated by GPIb, and that the interaction of GPIb with exosite I apparently is not required or capable of inducing platelet adhesion under flow conditions. This is in correspondence with the observations of Celikel et al, 17 who reported that thrombin first binds to GPIb via exosite II, after which a second molecule can bind via exosite I.

Although we show that thrombin can act as a platelet adhesive protein, thrombin is usually not present as a surface-bound protein but functions in hemostasis primarily as a soluble protein. Nevertheless, on clot formation thrombin is immobilized to the fibrin clot 1 and this binding to fibrin may be important in localizing thrombin to the site of vascular injury. Our results show that when thrombin is bound to fibrin, platelet adhesion and aggregate formation are substantially enhanced. Although relatively high concentrations of thrombin are required for this process, it would make sense to believe that local thrombin concentrations bound to fibrin could rise to high levels and thereby contribute in the post-recruitment of platelets to the fibrin-clot. Also, in the presence of dRGDW, which completely blocks adhesion to fibrin, platelets readily adhere to fibrin-bound thrombin. This suggests that fibrin-bound thrombin not only increases platelet adhesion and aggregation by enhancement of platelet activation mediated by PARs, but also is able to directly capture platelets via GPIb. This is further demonstrated by the experiments shown in Figure 5, which show that the increase in platelet adhesion to fibrin with bound thrombin is abolished when platelets are depleted from GPIb after pretreatment with Nk.

When bound to fibrin, exosite II mutants contributed to platelet adhesion in the absence of dRGDW, which most likely reflects enhancement of platelet activation via PARs. As these exosite mutants could not initiate platelet adhesion in the presence of dRGDW, we conclude also that fibrin-bound thrombin is able to capture platelets via GPIb. Although Y71A has a reduced binding capacity for fibrin, it can still contribute to platelet adhesion to fibrin. However, it is unable to induce platelet adhesion in the presence of dRGDW. It appears that in our experimental setup the amount of Y71A, which has bound to fibrin, still has the potential to activate PARs and contribute to platelet adhesion to fibrin but is present in insufficient amounts to directly capture platelets via GPIb.

It is important to note that when thrombin is immobilized on a surface, probably thrombin exosites are not both available for ligand binding. Whether the fibrin-bound thrombin is bound to fibrin via exosite I or exosite II is still a matter of debate. Extensive reviews have been dealing with this controversy (Huntington, 25 Mosesson, 26 and Lane 27 ). Although from the results in the present study we cannot confirm the exact mechanism of how thrombin is bound to fibrin, the fact that adhesion of GPIb -depleted platelets to fibrin-bound thrombin is strongly diminished compared with control platelets, combined with the observation that platelet adhesion to thrombin itself is mediated by thrombin exosite II, suggests that thrombin is bound to fibrin via exosite I and contributes to platelet adhesion via a GPIb -dependent interaction with exosite II. Although exosite I interacts with fibrin, and thrombin also interacts via exosite I with PAR1, PAR1 can still be hydrolyzed by fibrin-bound thrombin. This seems contradictory, but Myles et al already described that the ability of thrombin mutants to activate PAR1 or clot fibrinogen differ profoundly, indicating the involvement of different exosite I residues in PAR1 activation and fibrin(ogen) binding. 12

In conclusion, these experiments show that immobilized thrombin can act as an adhesive surface and is able to directly capture and activate platelets under flow conditions. This platelet adhesion is dependent on the interaction of thrombin with GPIb, the activation of PARs, and the secretion of ADP and thromboxane A2. Platelet adhesion to fibrin-bound thrombin could be a novel target for new antithrombotic drugs, which could now more specifically interfere with the action of thrombin on platelets on the actual site of thrombosis.

Acknowledgments

This research was supported in part by grants from the Netherlands Organization for Scientific Research (NWO; VENI 916.56.076) and the Netherlands Thrombosis Foundation (No. 2003-3) to T.L.

【参考文献】
  Nilsen DW, Brosstad F, Kierulf P, Godal HC. Binding of various thrombin fractions to fibrin and the influence of AT-III on their adsorption. Thromb Haemost. 1986; 55: 352-356.

Weitz JI, Hudoba M, Massel D, Maraganore J, Hirsh J. Clot-bound thrombin is protected from inhibition by heparin-antithrombin III but is susceptible to inactivation by antithrombin III-independent inhibitors. J Clin Invest. 1990; 86: 385-391.

Becker DL, Fredenburgh JC, Stafford AR, Weitz JI. Exosites 1 and 2 are essential for protection of fibrin-bound thrombin from heparin-catalyzed inhibition by antithrombin and heparin cofactor II. J Biol Chem. 1999; 274: 6226-6233.

dem Borne PA, Meijers JC, Bouma BN. Effect of heparin on the activation of factor XI by fibrin-bound thrombin. Thromb Haemost. 1996; 76: 347-353.

Kahn ML, Nakanishi-Matsui M, Shapiro MJ, Ishihara H, Coughlin SR. Protease-activated receptors 1 and 4 mediate activation of human platelets by thrombin. J Clin Invest. 1999; 103: 879-887.

Okumura T, Hasitz M, Jamieson GA. Platelet glycocalicin. Interaction with thrombin and role as thrombin receptor of the platelet surface. J Biol Chem. 1978; 253: 3435-3443.

Harmon JT, Jamieson GA. The glycocalicin portion of platelet glycoprotein Ib expresses both high and moderate affinity receptor sites for thrombin. A soluble radioreceptor assay for the interaction of thrombin with platelets. J Biol Chem. 1986; 261: 13224-13229.

De Marco L, Mazzucato M, Masotti A, Fenton JW, Ruggeri ZM. Function of glycoprotein Ib alpha in platelet activation induced by alpha-thrombin. J Biol Chem. 1991; 266: 23776-23783.

Shrimpton CN, Borthakur G, Larrucea S, Cruz MA, Dong JF, Lopez JA. Localization of the adhesion receptor glycoprotein Ib-IX-V complex to lipid rafts is required for platelet adhesion and activation. J Exp Med. 2002; 196: 1057-1066.

Vanhoorelbeke K, Ulrichts H, Romijn RA, Huizinga EG, Deckmyn H. The GPIbalpha-thrombin interaction: far from crystal clear. Trends Mol Med. 2004; 10: 33-39.

Hall SW, Gibbs CS, Leung LL. Identification of critical residues on thrombin mediating its interaction with fibrin. Thromb Haemost. 2001; 86: 1466-1474.

Myles T, Le Bonniec BF, Stone SR. The dual role of thrombin?s anion-binding exosite-I in the recognition and cleavage of the protease-activated receptor 1. Eur J Biochem. 2001; 268: 70-77.

Sheehan JP, Sadler JE. Molecular mapping of the heparin-binding exosite of thrombin. Proc Natl Acad Sci U S A. 1994; 91: 5518-5522.

Di Cera E. Thrombin interactions. Chest. 2003; 124: 11S-17S.

Li CQ, Vindigni A, Sadler JE, Wardell MR. Platelet glycoprotein Ib alpha binds to thrombin anion-binding exosite II inducing allosteric changes in the activity of thrombin. J Biol Chem. 2001; 276: 6161-6168.

De Cristofaro R, De Candia E, Landolfi R, Rutella S, Hall SW. Structural and functional mapping of the thrombin domain involved in the binding to the platelet glycoprotein Ib. Biochemistry. 2001; 40: 13268-13273.

Celikel R, McClintock RA, Roberts JR, Mendolicchio GL, Ware J, Varughese KI, Ruggeri ZM. Modulation of alpha-thrombin function by distinct interactions with platelet glycoprotein Ibalpha. Science. 2003; 301: 218-221.

Dumas JJ, Kumar R, Seehra J, Somers WS, Mosyak L. Crystal structure of the GpIbalpha-thrombin complex essential for platelet aggregation. Science. 2003; 301: 222-226.

Sadler JE. Structural biology. A menage a trois in two configurations. Science. 2003; 301: 177-179.

Adam F, Guillin MC, Jandrot-Perrus M. Glycoprotein Ib-mediated platelet activation. A signalling pathway triggered by thrombin. Eur J Biochem. 2003; 270: 2959-2970.

Savage B, Saldivar E, Ruggeri ZM. Initiation of platelet adhesion by arrest onto fibrinogen or translocation on von Willebrand factor. Cell. 1996; 84: 289-297.

Yun TH, Baglia FA, Myles T, Navaneetham D, Lopez JA, Walsh PN, Leung LL. Thrombin activation of factor XI on activated platelets requires the interaction of factor XI and platelet glycoprotein Ib alpha with thrombin anion-binding exosites I and II, respectively. J Biol Chem. 2003; 278: 48112-48119.

Hantgan RR, Hindriks G, Taylor RG, Sixma JJ, de Groot PG. Glycoprotein Ib, von Willebrand factor, and glycoprotein IIb:IIIa are all involved in platelet adhesion to fibrin in flowing whole blood. Blood. 1990; 76: 345-353.

Englund GD, Bodnar RJ, Li Z, Ruggeri ZM, Du X. Regulation of von Willebrand factor binding to the platelet glycoprotein Ib-IX by a membrane skeleton-dependent inside-out signal. J Biol Chem. 2001; 276: 16952-16959.

Huntington JA. Molecular recognition mechanisms of thrombin. J Thromb Haemost. 2005; 3: 1861-1872.

Mosesson MW. Fibrinogen and fibrin structure and functions. J Thromb Haemost. 2005; 3: 1894-1904.

Lane DA, Philippou H, Huntington JA. Directing thrombin. Blood. 2005; 106: 2605-2612.

作者单位：Thrombosis and Haemostasis Laboratory (C.W., J.A., P.G.d.G., T.L.), Department of Haematology, University Medical Centre Utrecht, the Netherlands; the Institute of Biomembranes (C.W., P.G.d.G., T.L.), Utrecht University, Utrecht, the Netherlands; and the Division of Hematology (T.M.), Department of