Binding of thrombin to the G-protein-linked receptor, and not to glycoprotein Ib, precedes thrombin-mediated platelet activation.

The roles of the G-protein-linked thrombin receptor and platelet glycoprotein Ib (GPIb) as α-thrombin-binding sites on platelets remain controversial. α-Thrombin has been proposed to bind to both GPIb and the hirudin-like domain of the G-protein-linked receptor (from which it cleaves the NH2-terminal extracellular domain to release a 41-mer peptide (TR-(1-41), where TR is α-thrombin receptor)) to initiate platelet activation. Using affinity-purified rabbit anti-human TR-(1-41) IgG and immunoblotting, we demonstrated TR-(1-41) release from platelets suspended in Tyrode's buffer containing 2 mM CaCl2 and incubated with ≥0.5 nM α-thrombin for 10-60 s at 37°C. As quantified by enzyme-linked immunosorbent assay, 0.32-0.59 nM TR-(1-41) was released from washed platelets (5 × 1011 platelets/liter) after their incubation with 10 nM α-thrombin for 10 s. Parallel binding of α-thrombin to and activation of the platelets were confirmed by flow cytometry. A monoclonal antibody against the hirudin-like domain of the G-protein-linked receptor abrogated α-thrombin binding to platelets, cleavage of TR-(1-41), and platelet activation by ≤1.0 nM (but not 10 nM) α-thrombin. Proteolysis of platelet GPIb with Serratia marcescens protease or O-sialoglycoprotein endopeptidase had no effect on α-thrombin binding to platelets or their subsequent activation. In contrast, chymotrypsin, which cleaves both GPIb and the G-protein-linked receptor, abrogated α-thrombin binding to platelets, TR-(1-41) release, and platelet activation. Furthermore, monoclonal antibodies directed against the reported α-thrombin-binding site on GPIb inhibited neither α-thrombin binding to nor activation of the platelets. Thus, α-thrombin binds to and cleaves the G-protein-linked receptor when it activates platelets, and GPIb does not appear to serve as an important binding site when α-thrombin activates platelets.

Another ␣-thrombin receptor on platelets, a member of the superfamily of G-protein-linked receptors and also found on endothelial cells, smooth muscle cells, and fibroblasts, has been cloned (9, 17, 20 -22). ␣-Thrombin binds to and cleaves this receptor at Arg-41/Ser-42, releasing a 41-mer activation peptide (called TR-  in this study, where TR is ␣-thrombin receptor) and exposing a new NH 2 -terminal domain, which then binds to an undefined part of the same receptor to activate the platelets (9,20,21). Whether interactions of ␣-thrombin with this G-protein-linked thrombin receptor, GPIb, or both are required for platelet activation by ␣-thrombin remains an unresolved question. Some investigators consider the G-proteinlinked ␣-thrombin receptor to be the moderate-affinity binding site since there are 1700 copies of this ␣-thrombin receptor/ platelet (23,24), and Bernard-Soulier platelets have normal numbers of this receptor (24). However, monoclonal antibodies that bind to the hirudin-like domain of the G-protein-linked thrombin receptor abrogate the responses of platelets to Յ1.0 nM ␣-thrombin (25)(26)(27)(28). This level of ␣-thrombin would be expected to bind preferentially to its high-affinity binding sites on platelets. It is possible that GPIb, by initiating ␣-thrombin binding to platelets, could localize ␣-thrombin to sites on platelets where the cleavage of the G-protein-linked ␣-thrombin receptor would be facilitated to cause platelet activation (19,23).
This study examined whether the cleavage of the G-proteinlinked thrombin receptor necessarily occurs when platelets are activated with 0.5, 1.0, and 10 nM ␣-thrombin. Affinity-purified polyclonal antibodies against the 41-mer activation peptide (TR- ) released from the G-protein-linked thrombin receptor by ␣-thrombin were used to detect cleavage of this receptor and release of the 41-mer activation peptide from platelets incubated with ␣-thrombin. Binding of ␣-thrombin to and activation of the same platelets were assessed by flow cytometry (28). Whether ␣-thrombin binding to GPIb is a prerequisite for platelet activation by ␣-thrombin was also explored by cleaving GPIb from platelets with three proteases known to cleave this platelet glycoprotein (13)(14)(15)(16) and by using a panel of monoclonal anti-GPIb antibodies previously reported to inhibit ␣-thrombin binding to platelets (11,17,18).
Antibodies-The following monoclonal antibodies were gifts: TM60 from Dr. Naomasa Yamamoto (27). Polyclonal antibodies against the 41-amino acid peptide released from platelets after ␣-thrombin cleaves the G-protein-linked thrombin receptor (and subsequently called TR-  in this study) were raised by immunizing rabbits and chickens with 20 g of the synthetic peptide corresponding to the first 41 amino acid residues of the G-protein-linked thrombin receptor at biweekly intervals. The IgG fraction was isolated from the rabbit antisera by chromatography on a protein G-Sepharose 4B column, and specific rabbit anti-human TR-(1-41) IgG was isolated by immunoaffinity chromatography of the IgG on a TR-(1-41)-Sepharose 4B column. IgG isolated from egg yolk (30) was also subjected to affinity chromatography on a TR-(1-41)-Sepharose 4B column to isolate specific chicken anti-human TR-(1-41) IgG. The other antibodies used in the flow cytometric studies were phycoerythrinconjugated monoclonal anti-GMP-140 IgG (Becton Dickinson Advanced Cellular Biology, San Jose, CA) and polyclonal rabbit anti-human ␣-thrombin IgG isolated from rabbit anti-␣-thrombin serum and biotinylated as described previously (28). Fluorescein isothiocyanate-labeled monoclonal anti-native GPIIb-IIIa and anti-activated GPIIb-IIIa antibodies were also obtained from Becton Dickinson Advanced Cellular Biology.
Preparation of Platelets and Platelet-rich and Platelet-poor Plasmas-Venous blood of healthy volunteers who had not taken any medication for 7 days was collected into 38 g/liter sodium citrate (9 parts blood to 1 part sodium citrate). Platelet-poor plasmas were isolated by centrifugation at 1500 ϫ g for 20 min at 4°C. Pooled normal plasma was obtained by adding equal volumes of platelet-poor plasmas from at least 20 normal healthy subjects and was stored at Ϫ50°C. Washed platelets were prepared using the procedures of Mustard et al. (32). Briefly, blood from healthy volunteers not on any medication was collected into ACD anticoagulant solution containing 5 mM citric acid, 85 mM trisodium citrate, and 111 mM glucose at a ratio of 6 volumes of blood and 1 volume of ACD. Platelet-rich plasma was isolated by centrifuging blood at 190 ϫ g for 15 min at ϳ23°C, followed by a second centrifugation at 2500 ϫ g for 15 min at ϳ23°C. After discarding the plasma, the platelet pellet was washed twice in a modified Tyrode's buffer containing 3.5 g/liter bovine serum albumin, 5 mM HEPES, 2 mM CaCl 2 , 1 mM MgCl 2 , 1 g/liter glucose, and apyrase adjusted to pH 7.35 at 37°C. The washed platelets were resuspended in the modified Tyrode's buffer (1 ϫ 10 12 platelets/liter) as the stock platelet suspension for use in all experiments requiring platelets resuspended in buffer or plasma. Platelet-rich plasmas were made by resuspending the washed platelets in pooled normal plasma to a final platelet count of 2 ϫ 10 11 /liter. To delay any fibrin formed from polymerizing, 5 nM GPRP (final concentration) was added to the platelet-rich plasmas (28). All experiments were performed at 37°C.
Treatment of Platelets with Chymotrypsin, S. marcescens Protease, or O-Sialoglycoprotein Endopeptidase-Washed platelets resuspended in the modified Tyrode's buffer (2 ϫ 10 11 platelets/liter) were incubated at 37°C with 50 nM chymotrypsin for 30 min, with 25 mg/liter S. marc-escens protease for 30 min, or with 10 mg/liter O-sialoglycoprotein endopeptidase for 60 min. Aliquots of the platelet suspensions were fixed in 1% paraformaldehyde for 10 min at ϳ23°C and subsequently analyzed by flow cytometry to estimate markers of platelet activation and alterations in platelet glycoproteins. Other periodic aliquots were also taken into the modified Tyrode's buffer containing 1 M hirudin and the supernatants to estimate the cleavage of TR-(1-41) from platelets by dot blotting as detailed below.
Platelet Activation by ␣-Thrombin-Washed control platelets in the modified Tyrode's buffer or platelets preincubated with a protease (2 ϫ 10 11 platelets/liter) were incubated with 0, 0.5, 1, or 10 nM ␣-thrombin at 37°C for up to 30 min. In some experiments, the washed platelets were resuspended in the modified Tyrode's buffer without 2 mM CaCl 2 and incubated with ␣-thrombin for up to 30 min at 37°C. Periodic aliquots were fixed in 10 g/liter paraformaldehyde for flow cytometric analysis or were added to the modified Tyrode's buffer containing 1 M hirudin to inactivate the added ␣-thrombin, followed by detection of TR-(1-41) release from the platelets as detailed below. In other experiments, the following monoclonal anti-platelet glycoproteins (at the final concentrations shown in parentheses) were added to control washed platelets resuspended in the modified Tyrode's buffer for 10 min at 37°C prior to the addition of ␣-thrombin and flow cytometric analysis: TM60 (100 mg/liter), LJ-IB10 (120 mg/liter), 6D1 (100 mg/liter), and ATAP-138 (150 mg/liter). These experiments determined how each monoclonal antibody influenced the binding of ␣-thrombin to platelets and the subsequent responsiveness of the platelets to the bound ␣-thrombin.
Immunoblotting Analysis-Washed platelets preincubated with a protease were centrifuged at 15,500 ϫ g for 1 min at 37°C to determine the cleavage and release of TR-(1-41) from the extracellular domain of the G-protein-linked thrombin receptor into the supernatants. The supernatants were recovered and subjected to dot blotting by loading 10 l of each supernatant onto strips of nitrocellulose membrane, which were then air-dried at room temperature and incubated in 10 g/liter gelatin dissolved in a buffer containing 20 mM Tris, 500 mM NaCl, 0.2 g/liter trisodium azide, and 0.5 g/liter Tween 20, pH 7.4 (TBS-Tween) overnight. After washing twice with TBS-Tween, the membranes were incubated with 2 mg/liter biotinylated rabbit anti-human TR-(1-41) IgG (in TBS-Tween containing 1 g/liter gelatin) for 2 h. After washing the membrane four times with the above buffer, blots containing TR-(1-41) were identified using alkaline phosphatase-conjugated streptavidin, followed by color development with 5-bromo-4-chloro-3-indolyl phosphate, p-toluidine, and nitro blue tetrazolium.
Quantification of TR-  Release by ␣-Thrombin and Chymotrypsin-The release of TR-(1-41) resulting from the incubation of platelets resuspended in Tyrode's buffer with 1.0, 10.0, and 50 nM ␣-thrombin or with 10 and 50 nM chymotrypsin was quantified by an ELISA for TR- . In this ELISA, 200 l of affinity-purified chicken antihuman TR-(1-41) dissolved in 0.1 M NaHCO 3 , pH 9.6 to a concentration of 10 g of IgG/liter was added to each well of microtiter platelets and incubated at 4°C for 16 h. The free IgG was removed by suction, and free sites on the wells of the microtiter plates were blocked with 1 g/liter fatty acid-free bovine serum albumin in a buffer containing 0.01 M Tris-HCl, 0.15 M NaCl, and 0.5 g/liter Tween 20, pH 8.0 (TBS-T). After four washes with a buffer containing 0.01 M Na 2 HPO 4 , 0.145 M NaCl, and 0.5 g/liter Tween 20, pH 7.4 (PBS-T), a standard curve for estimating the concentration of TR-(1-41) was constructed as follows. 100 l of increasing concentrations of TR-(1-41) (20 pM to 5 nM) in TBS-T containing 10 g/liter bovine serum albumin were added to microtiter wells and incubated at 37°C for 60 min. Each well then received four washes with PBS-T, followed by the addition of 100 l of biotinylated chicken anti-human TR-(1-41) (100 g/liter) for a 60-min incubation at 37°C. After four washes with PBS-T, 100 l of alkaline phosphate-conjugated streptavidin diluted 1:10,000 were then added, and the plates were incubated at 37°C for 1 h. After another four washes with PBS-T, 100 l of 1 g/liter p-nitrophenyl phosphate were added for a 40-min incubation at 37°C, and the color yield at 405 nM was quantified.
To quantify the release of TR-(1-41) from platelets, washed platelets (5 ϫ 10 11 /liter) were resuspended in the modified Tyrode's buffer supplemented with bovine serum albumin to a final concentration of 10 g/liter. ␣-Thrombin (1.0 or 10.0 nM) or chymotrypsin (10 or 50 nM) was then added to the platelet suspensions. Periodic aliquots were removed and added to 0.02 volume of 50 M D-Phe-Pro-ArgCH 2 Cl (for platelets incubated with ␣-thrombin) or 1.5 mM phenylmethylsulfonyl fluoride (for platelets incubated with chymotrypsin). The concentrations of TR-(1-41) released from the platelets were estimated by ELISA immediately after the platelets had been centrifuged at 10,000 ϫ g for 10 min at 22°C.
Flow Cytometric Analysis of Platelets-The procedures described previously were used to estimate ␣-thrombin binding to platelets (28). Briefly, platelets fixed in 10 g/liter paraformaldehyde for 10 min at 22°C were centrifuged at 1175 ϫ g for 15 min and resuspended in 154 mM NaCl. The fixed platelets were next incubated with biotinylated rabbit anti-human thrombin IgG (at a concentration of 1 g/liter) for 30 min at 22°C and then washed twice with 154 mM NaCl containing 1 g/liter bovine serum albumin. After resuspension in FACSFlow fluid (Becton Dickinson, Mississauga, Canada), the platelets were incubated with phycoerythrin-conjugated Z-avidin for 30 min, washed twice with 154 mM NaCl containing 1 g/liter bovine serum albumin, and finally resuspended in FACSFlow fluid. The percentage of 10,000 platelets that had bound ␣-thrombin and the associated fluorescence intensity were determined using a FACScan argon ion flow cytometer operating at 488 nm and at 15-milliwatt power using LysisII software. The instrument was set up to measure the size (forward scatter), granularity (side scatter), and platelet fluorescence. All parameters were collected using a 4-decade logarithmic amplification. The data are reported as thrombin fluorescence intensity on the platelets (mean channel fluorescence in arbitrary units).
Similar procedures were used to quantify the expression of GMP-140 (P-selectin), CD63, and the resting and activated conformers of GPIIb-IIIa on platelets using the appropriate monoclonal antibodies, except that the data are reported as the percentage of platelets expressing the marker under study. The panel of monoclonal anti-GPIb antibodies (TM60, LJ-IB10, and 6D1) was also used to detect GPIb on control washed platelets and washed platelets preincubated with chymotrypsin, S. marcescens protease, or O-sialoglycoprotein endopeptidase.

Cleavage of the G-protein-linked Thrombin Receptor by
␣-Thrombin-To determine whether activation of platelets by ␣-thrombin necessarily coincided with cleavage of the G-protein-linked thrombin receptor, washed platelets resuspended in the modified Tyrode's buffer were incubated with up to 10 nM ␣-thrombin for up to 1 min. Both the binding of ␣-thrombin to the platelets and activation of the same platelets were estimated. Platelet activation was estimated by quantifying GMP-140 ( Fig. 1), CD63, and the activated conformer of GPIIb-IIIa expression on platelets by flow cytometry. As shown in Fig. 1, dose-dependent binding of ␣-thrombin to the platelets and expression of GMP-140 on the activated platelets were observed, beginning 10 s after Ն0.5 nM ␣-thrombin addition. The fluorescence intensity of GMP-140 associated with each concentration of added ␣-thrombin remained unchanged during the next 50 s of incubation. However, both ␣-thrombin binding to platelets and expression of GMP-140 thereon had decreased by ϳ20% when the incubation of platelets with ␣-thrombin was increased to 30 min (Table I).
We also explored the response of platelets to a second addition of ␣-thrombin. In these experiments, washed platelets were incubated with 0.5 or 1.0 nM ␣-thrombin for 60 s, followed by the addition of 10 nM ␣-thrombin for a 10-s incubation. The percentage of platelets expressing P-selectin was used to estimate the responses of the platelets to the first and subsequent ␣-thrombin additions. The addition of 0.5 and 1.0 nM ␣-thrombin to these platelets for 60 s resulted in 31 and 76% of the platelets, respectively, expressing P-selectin. The subsequent addition of 10 nM ␣-thrombin to these platelets resulted in Ͼ95% of the platelets expressing P-selectin. Thus, platelets unactivated following the addition of suboptimal concentrations of ␣-thrombin respond to a second addition of ␣-thrombin.
Similar dose-dependent expression of CD63 and the activated conformer of GPIIb-IIIa on the platelets was also observed after ␣-thrombin addition (data not shown). Using affinity-purified rabbit anti-human TR-(1-41) IgG and dot blotting, release of TR-(1-41) from platelets incubated with the three concentrations of ␣-thrombin was observed as TR-  was detected in the supernatants of platelets incubated with Ն0.5 nM ␣-thrombin for 10 s (Fig. 2). The staining intensity of TR-(1-41) seen at 10 s remained unchanged for the next 50 s (data not shown).
The concentrations of TR-(1-41) released from washed platelets incubated with 1.0 or 10 nM ␣-thrombin or with 10 or 50 nM chymotrypsin were quantified by ELISA, and the results are summarized in Table II. This ELISA could quantify Ն20 pM synthetic TR-(1-41) with an intra-assay variability of Ϯ7% Effects of ATAP-138 on Thrombin-mediated Platelet Activation-ATAP-138, a monoclonal antibody against the hirudinlike domain of the G-protein-linked thrombin receptor (27), abrogates thrombin-mediated activation of platelets suspended in buffers or plasma by preventing the binding of ␣-thrombin to platelets (27,28). We explored whether abrogation of platelet activation by ATAP-138 was associated with the inhibition of the release of TR-(1-41) from the platelets by ␣-thrombin. ATAP-138 at 150 mg/liter abrogated the binding of both 0.5 and 1.0 nM ␣-thrombin to and the associated activation of washed platelets (Fig. 3), TR-(1-41) release from the platelets (Fig. 2), and the expression of CD63 and the activated GPIIb-IIIa conformer on the platelets (data not shown). However, 10 nM ␣-thrombin normally bound to and activated washed platelets preincubated with 150 mg/liter ATAP-138 (Fig. 3), and both events in this case were associated with cleavage of this thrombin receptor and release of TR-(1-41) into the supernatant (Fig. 2). The concentration of TR-(1-41) released by 10 nM ␣-thrombin from platelets preincubated with ATAP-138 was similar to that released by 1 nM ␣-thrombin from control platelets. ATAP-138 abrogated the binding of 1 or 10 nM ␣-thrombin to and the subsequent activation of platelets resuspended in pooled normal plasmas (Fig. 4).
Effects of Platelet GPIb Cleavage on ␣-Thrombin-mediated Platelet Activation-How the enzymatic degradation of GPIb influences ␣-thrombin binding to and activation of platelets was next determined. O-Sialoglycoprotein endopeptidase and S. marcescens protease specifically cleave GPIb, while chymotrypsin probably cleaves other platelet glycoproteins (13)(14)(15)(16). None of these three proteases directly activated the platelets or altered the initial fluorescence of the resting and activated GPIIb-IIIa conformers on platelets (data not shown). However, each protease completely removed GPIb from the platelets or markedly altered the tertiary structure of GPIb since none of the monoclonal anti-GPIb antibodies (TM60, LJ-IB10, or 6D1) bound to platelets preincubated with any of these three proteases (data not shown).
In spite of this observation, 0.5, 1, or 10 nM ␣-thrombin normally bound to and activated platelets preincubated with S. marcescens protease and O-sialoglycoprotein endopeptidase (Fig. 1). Additionally, neither protease inhibited the expression of CD63 or the activated GPIIb-IIIa conformer on platelets following 0.5, 1.0, or 10 nM ␣-thrombin addition (data not shown). In contrast to platelets incubated with these two proteases, ␣-thrombin neither bound to nor activated platelets preincubated with chymotrypsin (Fig. 1). Immunoblotting confirmed the release of fragment(s) of the G-protein-linked thrombin receptor that reacted with anti-human TR-(1-41) IgG from platelets preincubated with chymotrypsin for 30 s (Fig. 2). Cleavage of this receptor by chymotrypsin abrogated the activation of the platelets by ␣-thrombin.
Effects of Monoclonal Anti-GPIb Antibodies on ␣-Thrombin Binding to Platelets-In additional experiments exploring the role of GPIb in ␣-thrombin-mediated platelet activation, the effects of the three monoclonal anti-GPIb antibodies on TABLE I Effects of TM60, LJ-IB10, and ATAP-138 on ␣-thrombin binding to washed platelets and their subsequent activation The monoclonal antibody or saline was added to washed platelets (5 ϫ 10 11 /liter) resuspended in Tyrode's buffer, with (first four columns) or without (last four columns) CaC 2 , for incubation at 37°C. ␣-Thrombin (0.5 nM) was then added to each platelet suspension. Aliquots were withdrawn at 10 s or 30 min into 1% paraformaldehyde, fixed for 10 min, and subjected to flow cytometry to estimate the percentage of platelets that had bound ␣-thrombin and the percentage of platelets that had become activated (by expressing GMP-140 or P-selectin). The percentage of control platelets not exposed to ␣-thrombin but expressing thrombin at 10 s and 30 min was 5.1%; the percentage of control platelets expressing GMP-140 at 10 s and 30 min were 5.0 and 4.9%, respectively. Each result was obtained after the platelet suspensions from three to four experiments had been pooled prior to centrifugation and flow cytometry.  ␣-thrombin binding to and activation of platelets were also determined. TM60 and LJ-IB10 are monoclonal antibodies against the high-affinity ␣-thrombin-binding domain on GPIb (7,18), while 6D1 is directed against the von Willebrand factorbinding domain of GPIb (11,12). Thus, unlike TM60 and LJ-IB10, 6D1 was not expected to inhibit the interactions of ␣-thrombin with platelets. The binding of each monoclonal antibody to the platelets was verified by the positive and maximal staining of the platelets with either fluorescein isothiocyanate-or phycoerythrin-conjugated goat anti-mouse antibodies (data not shown). In spite of the above observation, none of the three anti-GPIb antibodies inhibited ␣-thrombin binding to or the subsequent activation of washed platelets (Fig. 3) or washed platelets resuspended in pooled normal plasma (Fig. 4).
The binding of ␣-thrombin to and the subsequent activation of washed platelets resuspended in Ca 2ϩ -free Tyrode's buffer were also investigated using 0.5 nM ␣-thrombin. This concentration of ␣-thrombin was chosen to ensure that only the highaffinity binding sites for ␣-thrombin on platelets would be occupied by the enzyme. As reported previously (28), ␣-thrombin bound to ϳ20% fewer platelets in the absence than in the presence of Ca 2ϩ (Table I). In the absence of Ca 2ϩ , LJ-IB10 significantly inhibited ␣-thrombin binding to platelets and their activation 10 s and 30 min after 0.5 nM ␣-thrombin had been added to the washed platelets. However, TM60 did not inhibit ␣-thrombin binding to platelets or their activation as effectively as LJ-IB10. Thus, Ca 2ϩ enhances the binding of ␣-thrombin to platelets and in a manner that decreases any requirement for GPIb for directing the initial binding of ␣-thrombin to platelets and their subsequent activation.

DISCUSSION
Platelets have ϳ25,000 copies of GPIb, the platelet glycoprotein proposed to provide ϳ50 high-affinity binding sites for ␣-thrombin (K d ϳ 1 nM) since platelets of Bernard-Soulier patients (and thus congenitally deficient in GPIb) aggregate slowly, but demonstrate normal dense body release in response to subnanomolar ␣-thrombin. Additionally, cleavage of GPIb or occupancy of GPIb by some monoclonal anti-GPIb antibodies inhibits platelet aggregation and release by Յ1.0 nM ␣-thrombin, but not by 10 nM ␣-thrombin (1, 5, 7, 8, 11, 12-16). A G-protein-linked thrombin receptor on platelets to which ␣-thrombin binds (probably via the hirudin-like domain of this receptor) and cleaves off the first 41 amino acid residues (called TR-  in this study) has been described (9, 10, 19 -21). There are ϳ1700 copies of this receptor/platelet (27), and some investigators have assigned the moderate-affinity ␣-thrombinbinding sites (K d ϳ 10 nM) on platelets to this receptor (11,23,24). Antibodies against the hirudin-like domain of this G-protein-linked receptor inhibit the responsiveness of platelets to ␣-thrombin (27,28,33). Thus, the primary site on platelets to which ␣-thrombin binds to initiate platelet activation remains unclear.
In this study, cleavage of platelet TR-(1-41) by ␣-thrombin was directly monitored, as were ␣-thrombin binding to platelets and the subsequent activation of the same platelets. No attempt was made in this study to quantify the number of ␣-thrombin molecules/platelet or the concentrations of markers of platelet activation that became expressed on activated platelets. Rather, the percentages of platelets that rapidly bound ␣-thrombin and subsequently expressed surface P-selection, CD63, and the activated conformer of GPIIb-IIIa for each concentration of the enzyme were quantified. We have presented data demonstrating the parallel binding of ␣-thrombin to platelets, cleavage and release of TR-(1-41) from the platelets, and activation of the same platelets with each concentration of ␣-thrombin. There was a similar (ϳ1:1) relationship between the binding of ␣-thrombin to platelets and the expression of each of the three markers of platelet activation within 60 s of ␣-thrombin addition. This study also confirmed the observation by Norton et al. (33) that ␣-thrombin releases TR-(1-41) from platelets. It is unclear why 1.0 nM ␣-thrombin did not release TR-(1-41) as effectively as 10 nM ␣-thrombin when both concentrations of the enzyme activated Ն75% of the washed platelets ( Figs. 1 and 3). We eliminated the possibility that this ␣-thrombin receptor became inaccessible to ␣-thrombin following the exposure of platelets to suboptimal concentrations of ␣-thrombin. Specifically, we demonstrated that platelets preincubated with 0.5 or 1 nM ␣-thrombin responded appropriately to a subsequent addition of ␣-thrombin. Thus, the fraction of the thrombin receptor not previously occupied by suboptimal concentrations of ␣-thrombin remained accessible to added ␣-thrombin. Since ϳ2.0 nM TR-(1-41) could be theoretically released from platelets (27), the fact that 10 nM ␣-thrombin fully activated the platelets but released only Յ0.6 nM TR-  suggests that complete cleavage of the receptor is not required for maximum platelet activation. Nonetheless, partial cleavage of this ␣-thrombin receptor is required to initiate platelet activation since abrogation of thrombin-mediated cleavage of this receptor by ATAP-138 also abrogated platelet activation.
A likely reason for the failure of ␣-thrombin to quantitatively cleave all available TR-(1-41) from platelets may reside in the ability of ␣-thrombin to induce endocytosis of this receptor, as demonstrated for two megaloblastic cell lines, namely human erythroleukemia cells and Children's Hospital Research Foundation cell line 288 (34 -36). This failure of up to 10 nM ␣-thrombin to fully cleave the G-protein-linked thrombin receptor and to release TR-(1-41) from platelets parallels the effects ␣-thrombin has on fibrinogen and fibrin has on the enzymatic activity of ␣-thrombin. Similar to the release of TR-(1-41), ␣-thrombin cleaves fibrinogen in a dose-dependent manner, with fibrinopeptide A release proceeding to the maximum extent achievable with each ␣-thrombin concentration within 60 s (37). ␣-Thrombin binding to fibrin also clearly impairs the ability of this enzyme to release fibrinopeptide A from fibrinogen (37). Binding of ␣-thrombin to the cleaved receptor (which   from platelets incubated with ␣-thrombin and chymotrypsin Washed platelets (5 ϫ 10 11 /liter) isolated from the blood of three healthy volunteers (A, B, and C) were resuspended in the modified Tyrode's buffer containing 2 mM CaC 2 and 10 mg/ml bovine serum albumin. Following the addition of ␣-thrombin or chymotryspin, aliquots were withdrawn into 0.02 volume of 50 M D-Phe-Pro-ArgCH 2 C (␣-thrombin) or 50 mM phenylmethylsulfonyl fluoride (chymotrypsin). Following centrifugation at 10,000 ϫ g for 10 min, the concentration of TR-(1-41) released into the supernatant was quantified by ELISA. The concentration of TR-(1-41) measured in each control platelet supernatant was 0.02 nM. Each incubation of platelets with an enzyme was conducted in triplicate, and the platelet supernatants were pooled prior to quantifying the release of TR-(1-41) by ELISA. then becomes phosphorylated (17,38)) may similarly impair the ability of the bound enzyme to cleave nearby receptors. Continued tight binding of ␣-thrombin to this site may be important, and one study has reported that continued occupancy of the G-protein-linked receptor by ␣-thrombin is required to propagate tyrosine phosphorylation. Specifically, Lau et al. (38) have reported that addition of hirudin to platelets preincubated with ␣-thrombin for 60 s does not deaggregate the platelets, but inhibits specific tyrosine phosphorylation and simultaneously accelerates specific tyrosine dephosphorylation. Occupancy of this receptor by ␣-thrombin at the hirudinlike domain of the receptor is clearly crucial for platelet activation since ATAP-138 abrogates the binding of 0.5 or 1 nM ␣-thrombin to platelets, release of TR-(1-41) from the platelets, and activation of the platelets. As previously reported by Brass et al. (27), we found that 10 nM ␣-thrombin binds to and activates washed platelets in the presence of a saturating concentration ATAP-138. The high-affinity binding sites for ␣-thrombin on GPIb are reportedly located within the M r 45,000 NH 2 -terminal domain of GPIb␣ (3, 5, 7, 18, 39 -42), and removal of GPIb from platelets by chymotrypsin, S. marcescens protease, or elastase yields platelets with a lower sensitivity to Յ1.0 nM ␣-thrombin (13)(14)(15)(16). This study has demonstrated that platelets with this putative high-affinity ␣-thrombin-binding domain on GPIb removed (by protease digestion) bound normally to ␣-thrombin.
In further experiments, two monoclonal antibodies against this putative high-affinity ␣-thrombin-binding domain on GPIb (TM60 and LJ-1B10) that inhibit the responses of platelets to Յ1 nM ␣-thrombin (7, 18, 39 -42) were used in another attempt to prevent ␣-thrombin binding to platelets via GPIb. In the presence of 2 mM CaCl 2 , ␣-thrombin bound normally to and activated platelets that had been preincubated with either monoclonal anti-GPIb antibody.
Therefore, we conclude that GPIb does not normally participate in the initial interactions of ␣-thrombin with platelets and that cleavage(s) by chymotrypsin additional to GPIb abrogate the responsiveness of platelets to ␣-thrombin. Chymotrypsin cleaves the G-protein thrombin-linked receptor at a point distal to Arg-41/Ser-42 (43,44). This cleavage may explain why only Ͻ30 pM TR-(1-41) was detected by the ELISA for TR-(1-41). Using a chimeric fusion protein consisting of glutathione S-transferase and residues 25-97 corresponding to the NH 2terminal extracellular domain of the G-protein-linked thrombin as the substrate, Bouton et al. (44) reported that the glycocalicin portion of GPIb did not alter the kinetics describing the cleavage of this fusion protein by ␣-thrombin, whereas fibrinogen fragment E, thrombomodulin, and hirudin fragment 54 -65 did. These results suggest minimal rapid binding interactions between ␣-thrombin and the extracellular domain of GPIb when the enzyme normally cleaves the G-protein-linked thrombin receptor.
There are three reasons why we could not ascribe a critical role to GPIb for mediating ␣-thrombin binding to platelets in the time required for ␣-thrombin to optimally activate platelets. (i) We used 10 -60-s incubations to demonstrate optimal binding of ␣-thrombin to platelets, compared with Ն30-min incubations used in some of the previous studies (7,18,24). The incubation times of 10 and 60 s were chosen as activation of platelets by ␣-thrombin proceeds to the maximum extent Tyrode's buffer were preincubated at 37°C for 10 min with TM60 (100 g/ml), LJ-IB10 (120 g/ml), 6D1 (100 g/ml), or ATAP-138 (150 g/ml). Following 0.5, 1.0, or 10 nM ␣-thrombin addition to the washed platelets for 10 s at 37°C, aliquots of the platelet suspensions were fixed in 1% paraformaldehyde for 10 min, and ␣-thrombin binding to the platelets and their activation were estimated by flow cytometry. Results are the means Ϯ S.D. (Յ5%) of four or more determinations. 3-6% of the platelets not incubated with ␣-thrombin were positive for both ␣-thrombin and GMP-140. achievable with each concentration of thrombin in Յ60 s (28). This choice was also justified by the demonstration of decreased binding of ␣-thrombin to platelets after the enzyme was incubated with platelets for 30 min (compared with 10 s), as shown in Table I. (ii) The platelets used in this study were fixed with 10 g/liter paraformaldehyde after their incubation with ␣-thrombin to immobilize the enzyme on platelets. Fixation of the platelets also inactivated ␣-thrombin and halted further platelet reactions resulting from ␣-thrombin binding to the platelets. Fixation does not alter the binding of ␣-thrombin to platelets (40). (iii) We also estimated ␣-thrombin binding to platelets resuspended in CaCl 2 -containing media, while the previous studies were without addition of this salt. CaCl 2 enhances the binding of ␣-thrombin to platelets and stabilizes the expression of P-selectin on the activated platelets (28), as was confirmed in this study. Additionally, two monoclonal anti-GPIb antibodies (LJ-IB10 and TM60) inhibited ␣-thrombin binding to washed platelets and their activation, but only in the absence of added CaCl 2 ( Table I). Inhibition of ␣-thrombin binding to platelets by these two monoclonal anti-GPIb antibodies (in the absence of Ca 2ϩ ) has been reported by many other investigators (11, 12, 39 -42).
Previous reports have hypothesized that GPIb and the Gprotein-linked thrombin receptor form a functional complex on platelets. Specifically, interactions of ␣-thrombin with GPIb localize ␣-thrombin to sites that facilitate cleavage of nearby G-protein-linked thrombin receptors during the activation process (18,23). Our results do not support significant interactions between ␣-thrombin and GPIb to effect ␣-thrombin binding to platelets, in the presence of Ca 2ϩ , to initiate platelet activation. The G-protein-linked thrombin receptor appears to be the primary site to which ␣-thrombin binds to initiate platelet activation. Our observations, however, do not exclude GPIb modulating additional signaling events, including changes in extracellular Ca 2ϩ and aggregation resulting from ␣-thrombin binding to the platelets (12,40,41).