Ca2+-dependent Binding of a Synthetic Arg-Gly-Asp (RGD) Peptide to a Single Site on the Purified Platelet Glycoprotein IIb-IIIa Complex*

The platelet glycoprotein IIb-IIIa complex (GP IIb- IIIa) is a member of the integrin receptor family that recognizes adhesive proteins containing the Arg-Gly- Asp (RGD) sequence. In the present study the binding characteristics of the synthetic hexapeptide Tyr-Asn- Arg-Gly-Asp-Ser (YNRGDS, a sequence present in the fibrinogen a-chain at position 570-575) to purified GP IIb-IIIa were determined by equilibrium dialysis. The binding of "'1-YNRGDS to GP IIb-IIIa was specific, saturable, and reversible. The apparent dissociation constant was 1.0 f 0.2 p ~ , and the maximal binding capacity was 0.92 & 0.02 mol of '261-YNRGDS/mol of GP IIb-IIIa, indicating that GP IIb-IIIa contains a single binding site for RGD peptides. The binding of "'1- YNRGDS to purified GP IIb-IIIa showed many of the characteristics of fibrinogen binding to activated platelets: the binding was inhibited by fibrinogen, by the monoclonal antibody AzAe, and by the dodecapep- tide from the C terminus of the fibrinogen y-chain. In addition, the binding of "'I-YNRGDS to GP IIb-IIIa was divalent cation-dependent. Platelet Aggregation-Platelets were isolated from fresh human plasma by gel filtration on Sepharose CL-2B (Pharmacia, Uppsala) in Tyrode's buffer containing 0.2% bovine serum albumin, 1 mM CaC12, and 1 mM MgC12 (27). The gel-filtered platelets (1.5 X 10a/ml) were incubated in the presence of 0.6 p~ fibrinogen and various amounts of RGDS or YNRGDS. Aggregation was initiated by addition of 10 p~ ADP and measured in a dual-channel aggregometer (ELVI, Milan, Italy) at 37 "C. The concentration of peptide required for 50% inhibition of aggregation was designated as ICm.

RGD-recognizing integrins interact only with a single RGDcontaining ligand. The conformation of the RGD domain of each ligand appears to be important for binding selectivity (6).
The platelet GP IIb-IIIa complex demonstrates less ligand selectivity because it binds the RGD-containing proteins fibrinogen, fibronectin, von Willebrand factor, and vitronectin (7). Synthetic RGD-containing peptides inhibit the binding of all four ligands (8,9), supporting the hypothesis that the four adhesive proteins interact with GP IIb-IIIa via their RGD domains. If so, the question arises whether the four ligands bind to a common RGD-recognition site on GP IIb-IIIa. The results of two independent investigations, measuring binding of lZ5I-labeled RGD-containing peptides to intact platelets, are inconsistent since in one study 6 X lo' binding sites per platelet for RGD peptides have been reported (lo), whereas in the other a 6-fold greater number of binding sites was obtained (11). The synthetic dodecapeptide, HHLGGAKQAGDV, corresponding to the C terminus of the fibrinogen y-chain, also inhibits the binding of fibrinogen, fibronectin, and von Willebrand factor (12,13), although the latter two ligands contain no sequences homologous to the dodecapeptide of fibrinogen. In addition to their inhibitory activity, RGDS and the dodecapeptide are capable of inducing a conformational change of GP IIb-IIIa in detergent solution (14) as well as in intact platelets resulting in the expression of a neoantigenic site (15). These results taken together with the observation that GP IIb-IIIa, when bound to an RGDS affinity matrix, can be specifically eluted with the dodecapeptide (11), led to the assumption that RGD peptides and the dodecapeptide may bind to the very same site(s) on GP IIb-IIIa. Studies using RGD-and dodecapeptide-containing crosslinkers, however, demonstrated that the binding sites for the two inhibitory peptides may not be identical (16). On the surface of stimulated platelets the RGD-containing crosslinker specifically labeled both subunits of GP IIb-IIIa to a similar extent, whereas the photoaffinity derivative of the dodecapeptide predominantly labeled GP IIb (16). In another study, using a slightly different protocol for chemical crosslinking, it was found that an RGD peptide was coupling mainly to the GP IIIa subunit and only poorly to GP IIb (17).
Subsequently, the region of GP IIIa where the RGD peptide became cross-linked was identified (18).
The purpose of the present study was to compare the binding properties of GP IIb-IIIa for a synthetic RGD-containing peptide with those described for fibrinogen and to determine the number of binding sites of purified GP IIb-IIIa for RGD peptides. Our results indicate that the binding of a lZ5I-labeled RGD-containing peptide to GP IIb-IIIa has many features in common with fibrinogen binding and that platelet GP IIb-IIIa contains a single binding site for RGD peptides.

EXPERIMENTAL PROCEDURES
Peptides and Peptide Affinity Matrix-Peptides were synthesized by the classical technique (19) using various coupling procedures and a combination of acid-labile protecting groups. The peptides were purified by preparative high performance liquid chromatography (HPLC) using a LiChrosorb RP-18 column (Merck, Darmstadt, FRG). The purity of each peptide exceeded 95% as assessed by thin layer chromatography, analytical HPLC, mass spectroscopy, and amino acid analyses using a Liquimat 111 analyzer (Labotron, Munich, FRG).
The peptide Tyr-Asn-Arg-Gly-Asp-Ser (YNRGDS) was labeled with ' "I (2 mCi) by the chloramine-T method (20). The labeled material was injected into a C18 HPLC column (Vydac) pre-equilibrated in 0.1% aqueous trifluoroacetic acid and acetonitrile (95:5, v/ v). The hexapeptide, the mono-and the di-iodinated hexapeptide were separated by applying a gradient of 5 to 30% acetonitrile. The monoiodinated form of YNRGDS was finally lyophilized from a solution containing 2% D-lactose, 0.05% bovine serum albumin, and 3 mM butanesulfonic acid in 100 mM sodium phosphate buffer. The '251-labeled YNRGDS was stored at 4 "C and used within 2 weeks. The specific activity was 2200 Ci/mmol peptide (81 TBqjmmol). The affinity matrix was prepared by incubating 120 mg of aminoethylglycine-Arg-Gly-Asp-Ser (Aeg-RGDS) with 10 ml of CNBractivated Sepharose (Pharmacia, Uppsala, Sweden) according to the method described by Pytela et al. (9). The spacer aminoethylglycine (Aeg) was introduced to facilitate coupling. G P IIb-IIIa Purification-Outdated, washed human platelets were lysed with 1% Triton X-100, 0.15 M NaCl, 20 mM Tris-HC1, 1 mM CaC12, 1 mM MgC12, 0.02% NaN3, 10 p M leupeptin, and 0.5 mM phenylmethylsulfonyl fluoride, pH 7.3, at 4 "C for 15 h. The glycoproteins were isolated at room temperature by concanavalin A affinity chromatography according to Fitzgerald et al. (21) using 0.1% Triton X-100,0.15 M NaCl, 20 mM Tris-HC1,l mM CaC12, 1 mM MgC12, and 0.05% NaN3, pH 7.0, as column buffer (buffer A). The concanavalin A-retained glycoproteins were eluted with buffer A containing 100 mM a-methyl-D-mannose and applied onto an Aeg-RGDS-Sepharose column. The bound GP IIb-IIIa complexes were eluted by including RGDS (3 mM) in buffer A. To remove the tetrapeptide the purified GP IIb-IIIa was extensively dialyzed at 4 "C versus buffer A. Removal of RGDS was monitored by including a trace amount of Iz5I-labeled YNRGDS in the dialysis bag.
For electrophoresis, aliquots of the platelet lysate, of different column fractions, and of the purified GP IIb-IIIa were treated with sodium dodecyl sulfate (SDS), reduced with j3-mercaptoethanol, and examined on 7.5% SDS-polyacrylamide gels (22). The gels were stained with Coomassie Brilliant Blue R-250, and the relative amounts of GP IIb and GP IIIa were determined by densitometry using a TLC Scanner I1 (Camag, Switzerland) connected to a Sanyo personal computer.
Peptide Binding Assay-Binding of '=I-labeled YNRGDS to purified GP IIb-IIIa was determined by equilibrium dialysis a t pH 7.4 using a 20-cell dialyzer (Dianorm, FRG). Dialysis membranes (Spectrapor) with a molecular weight cut-off of 12,000-14,000 presoaked in buffer A were used to separate each Teflon cell into two 200-pl compartments. The GP IIb-IIIa complex was introduced in one cell half, whereas '%I-labeled YNRGDS (1-5 X lo5 cpm) was added to both compartments in equal concentrations. In some experiments, various concentrations of monoclonal antibodies or other proteins such as fibrinogen (IMCO, Stockholm, Sweden), fibronectin (Sigma), or bovine serum albumin (Sigma) were added to the same compartment as the GP IIb-IIIa complex. Peptides, however, were routinely included in both compartments. The cells were slowly rotated for 4 h in a water bath at 25 "C. Following dialysis the two compartments of each cell were emptied and the '"1 radioactivity was determined in a y-counter (Kontron, Switzerland). Before and after dialysis the protein concentrations were determined by the method of Bradford (23). In addition, the total amino acid content of each GP IIb-IIIa stock solution was determined.
For the binding experiments in presence of various Ca2+ concentrations, the GP IIb-IIIa complexes were dialyzed into buffer A without M$+. The amounts of EGTA and Ca2+ required to achieve the desired free Ca2+ concentrations at pH 7.4 and 25 "C, in the presence or absence of 1 mM M e , were calculated with an iterative computer program as described previously (24). The extent of dissociation of the GP IIb-IIIa complex into subunits was analyzed immediately following dialysis by sedimentation through linear 5 to 25% sucrose gradients as described by Fitzgerald and Phillips (25).
Monoclonal Antibodies-The monoclonal antibody A2Ag was kindly provided by Drs. J. Bennett and S. Shattil, University of Pennsylvania School of Medicine, Philadelphia. The monoclonal antibody pl-21 was produced in our laboratory by standard techniques (26). Both monoclonal antibodies are directed against the GP IIb-IIIa complex and belong to the immunoglobulin G class.
Platelet Aggregation-Platelets were isolated from fresh human plasma by gel filtration on Sepharose CL-2B (Pharmacia, Uppsala) in Tyrode's buffer containing 0.2% bovine serum albumin, 1 mM CaC12, and 1 mM MgC12 (27). The gel-filtered platelets (1.5 X 10a/ml) were incubated in the presence of 0.6 p~ fibrinogen and various amounts of RGDS or YNRGDS. Aggregation was initiated by addition of 10 p~ ADP and measured in a dual-channel aggregometer (ELVI, Milan, Italy) at 37 "C. The concentration of peptide required for 50% inhibition of aggregation was designated as ICm.

RESULTS
Purification of GP IIb-IIIa-Concanavalin A and Aeg-RGDS affinity chromatography were used to purify GP IIb-IIIa from outdated human platelets. Fig. 1 shows the SDSpolyacrylamide gel separations of various fractions. Densitometric scans of the material eluted from the Aeg-RGDS affinity column showed that GP IIb and GP IIIa represented >98% of the Coomassie-stained protein ( reapplied onto another Aeg-RGDS column of the same size, 20 times less GP IIb-IIIa were bound by this second column than by the first column. This indicates that 90-95% of the G P IIb-IIIa eluted from the concanavalin A affinity column was incapable of binding to the Aeg-RGDS column. For the binding studies, purified G P IIb-IIIa was dialyzed to remove the tetrapeptide RGDS used for elution from the Aeg-RGDS column. When the dialyzed protein was reapplied onto an Aeg-RGDS column more than 95% of G P IIb-IIIa rebound. Analysis of the dialyzed protein by sedimentation through linear 5-25% sucrose gradients showed that more than 95% of GP IIb and GP IIIa sedimented as the heterodimer complex (data not shown).

Inhibition of Platelet Aggregation by RGDS and YNRGDS-
The peptide YNRGDS, present in the fibrinogen a-chain a t position 570-575 (28), was used as a prototype of the RGDcontaining peptides. To compare the inhibitory activity of YNRGDS with that of RGDS, ADP-induced aggregation of gel-filtered platelets was performed. In presence of 0.6 p~ fibrinogen, 32 ~L M RGDS and 39 ~L M YNRGDS were required for 50% inhibition (IC50), indicating that platelets had similar affinities for both RGD-containing peptides.
lZ5Z-YNRGDS Binding Assay-The '251-labeled YNRGDS was routinely added to both compartments of the dialysis cells, and dialysis was performed for 4 h at 25 "C as described under "Experimental Procedures." Control experiments showed that equilibrium was reached within 3 h, when the labeled peptide was introduced in only one cell half.
Nonspecific binding of lZ5I-YNRGDS was measured in the presence of an excess of unlabeled peptide. In the presence of 0.82 p~ lZ5I-YNRGDS a 50-fold excess of unlabeled peptide inhibited total binding by 90%, whereas in the presence of 0.12 p~ lZ5I-labeled peptide a 200-fold excess was required to approximate the level of nonspecific binding (Fig. 2). Therefore, according to the '251-YNRGDS concentration used, a 50-200-fold excess of unlabeled peptide was used to determine nonspecific binding. Specific '251-YNRGDS binding to purified GP IIb-IIIa was calculated by subtracting nonspecific binding from total binding.
To determine whether iodination altered the binding characteristics of unlabeled YNRGDS, GP IIb-IIIa was incubated with various ratios of '251-YNRGDS to unlabeled YNRGDS, while maintaining the total peptide concentration constant (Fig. 3). The control value (100% bound) was obtained with  and the unlabeled peptide bound to G P IIb-IIIa with similar association constants (29). The reversibility of '251-YNRGDS binding to G P IIb-IIIa was determined by adding excess unlabeled YNRGDS to preformed protein-peptide complexes. Purified G P IIb-IIIa (40 pg/ml) was first incubated with 0.82 p~ '251-YNRGDS for 4 h. Unlabeled YNRGDS or buffer A was then added to both compartments, and dialysis was continued for another 16 h.
Analysis of the data showed that 98 +. 3% of the specific lZ5I-YNRGDS binding was reversed by 50 p~ unlabeled YNRGDS. Thus, the binding of lZ5I-YNRGDS to purified G P IIb-IIIa was fully reversible.

Specificity of the Interaction between lZ5Z-YNRGDS and GP
ZZb-ZZZa-The specificity of lZ5I-YNRGDS binding to purified GP IIb-IIIa was examined in three ways. First, we investigated the effect of peptides, structurally related to YNRGDS, on the specific binding of lZ5I-labeled YNRGDS. As shown in Table I, the tetrapeptide RGDS had a similar inhibitory activity as the hexapeptide YNRGDS, ICs0 = 3.8 and 4.2 phi, respectively. However, substitutions of one amino acid of the tetrapeptide, histidine for arginine or alanine for glycine, decreased the inhibitory activity by a factor of 8 for HGDS and more than 10 for RADS. The relative inhibitory potency of these synthetic peptides was similar in the platelet aggregation assay (Table I). Second, the inhibitory activities of two RGD-containing ligands of G P IIb-IIIa, fibrinogen and fibronectin, were compared (Fig. 4). Fibrinogen, at a concentration of 1 p~, inhibited 72% of specific '251-YNRGDS binding, whereas 1 pM fibronectin only inhibited 21% of this binding. Bovine serum albumin a t a concentration of 1 pM was without any effect (data not shown). Third, we investigated the effect  Fig. 5 shows that the antibody A2A9 inhibited the specific binding of lZ5I-YNRGDS t o purified G P IIb-IIIa in a dose-dependent manner; 161 nM AzAg inhibited 88% of this binding. These data reinforce the possibility that the binding sites for fibrinogen and for the synthetic fibrinogen Aa-chain sequence YNRGDS are identical or a t least interdependent. The antibody pl-21 inhibited platelet aggregation with a similar ICs0 as A2A9 (data not shown). However, this antibody had no effect on lZ5I-YNRGDS binding, indicating that pl-21 does not recognize the same epitope on G P IIb-IIIa as A2Ag, although they both inhibit platelet aggregation.
Equilibrium Binding Constants-The equilibrium binding constants for specific lZ5I-YNRGDS binding to purified G P IIb-IIIa were determined by incubating G P IIb-IIIa with various concentrations of lZ5I-YNRGDS. The dissociation constant (&)   Effect of the Dodecapeptide from the Carboxyl Terminus of the Fibrinogen y-Chain-It has recently been shown that the dodecapeptide from the C terminus of the fibrinogen y-chain (HHLGGAKQAGDV) can elute solubilized G P IIb-IIIa that is bound to an RGDS affinity column (11). We therefore determined whether increasing concentrations of the dodecapeptide have an effect on the binding of lZ5I-YNRGDS to purified G P IIb-IIIa. At a concentration of 30 PM the dodecapeptide inhibited the specific binding by 70% (Fig. 7). The concentration required for 50% inhibition of specific lZ5I-YNRGDS binding (ICso) was 14.2 WM for the dodecapeptide as compared with 4.0 PM for YNRGDS.
Effect of Ca2+ and Mg2+ on '251-YNRGDS Binding to G P IIb-IIIa-Fibrinogen binding to activated platelets requires the presence of approximately 1 mM of either Ca2+ or Mg2+ ions (35, 36). To determine whether the binding of lZ5I-YNRGDS to purified G P IIb-IIIa is also divalent cationdependent, we studied the effect of various free Ca2+ concentrations in the presence or absence of 1 mM M$+ on specific lZ5I-YNRGDS binding (Fig. 8). In the presence of 1 mM M$+, lZ5I-YNRGDS binding at 10 and 1 $M free Ca2+ was slightly higher than the control value and decreased to 37% at 0. through linear 5-25% sucrose gradients to determine the percentage of GP IIb-IIIa still present as heterodimeric complexes. In all the samples incubated at Ca'+ concentrations of 10 p~ or higher, at least 90% of GP IIb and GP IIIa sedimented as the heterodimer complex, whereas significant dissociation of the GP IIb-IIIa complex was observed in the samples incubated at 0.1 PM Ca'+. The complete inhibition of lZ5I-YNRGDS binding at 10 p~ Ca'+ in the absence of Mg'+ was therefore not caused by dissociation of the GP IIb-IIIa complex. The reduced binding of "5-YNRGDS at 0.1 p~ Ca2' and 1 mM M$+ (37% of control), however, might be explained by dissociation because under these conditions only 40% of GP IIb and GP IIIa remained complexed. These results suggest that for binding of lZ5I-YNRGDS to purified GP IIb-IIIa two classes of divalent cation binding sites must be occupied; one class is specific for Ca2+ and is saturated at 1 p~ Ca2+ and the other is less specific and reaches saturation at 1 mM of either Ca2+ or M$+.

DISCUSSION
The glycoprotein IIb-IIIa complex is receptor for the adhesive proteins fibrinogen, fibronectin, von Willebrand factor, and vitronectin on the surface of stimulated platelets. Peptides containing the RGD sequence inhibit the binding of all four macromolecules to GP IIb-IIIa. Because each ligand contains at least one RGD sequence, the possibility exists that the four ligands might interact with the receptor via their RGD-containing domains. In the present study, the hexapeptide YNRGDS has been used as prototype for an RGDcontaining peptide. Preliminary studies indicated that whole platelets contain many low affinity binding sites for YNRGDS, so that an exact analysis of the binding data was not possible.' We therefore determined the binding properties of purified GP IIb-IIIa for YNRGDS by the method of equilibrium dialysis. Our results demonstrated that the binding of the RGD-containing peptide exhibits many of the properties characteristic for fibrinogen binding to stimulated platelets and that the GP IIb-IIIa complex contains a single binding site for RGD peptides.
The specificity of lZ5I-YNRGDS binding to purified GP IIb-IIIa was demonstrated by three different approaches. First, specific lZ5I-YNRGDS binding was inhibited by peptides containing the RGD-sequence such as YNRGDS and RGDS but not by related peptides such as RADS. Second, fibrinogen or fibronectin, two RGD-containing adhesive proteins, competed with YNRGDS for binding to GP IIb-IIIa, whereas bovine serum albumin or mouse immunoglobulins were without any effect. Third, the monoclonal antibody AzA9 directed against GP IIb-IIIa (30) inhibited the specific binding of lZ5I-YNRGDS to purified GP IIb-IIIa, while pl-21, another monoclonal antibody directed against GP IIb-IIIa, did not inhibit this binding.
The specific binding of lZ5I-YNRGDS to GP IIb-IIIa was saturable and fully reversible. The equilibrium binding data were therefore analyzed by the method of Scatchard (31). An apparent dissociation constant ( K d ) of 1.0 f 0.2 pM was obtained. A similar K d (0.38 p~) has been reported for the binding of an RGD-containing peptide, 13 amino acids in length, to 59,990 sites on thrombin-stimulated platelets (10). In another study, however, an RGD-containing peptide (14 amino acids) bound to about 350,000 sites with a K d of 119 PM (11).
Many of the properties of YNRGDS binding to purified GP IIb-IIIa were comparable with those of fibrinogen binding to stimulated platelets.
(i) The number of binding sites on GP IIb-IIIa for fibrinogen and the hexapeptide YNRGDS appears to be identical. This is apparent from the Scatchard analysis of the binding data which indicated that the maximum binding capacity was 0.92 mol of "51-YNRGDS/mol of GP IIb-IIIa. When the purified and dialyzed GP IIb-IIIa was reapplied onto the Aeg-RGDS-Sepharose column, more than 90% of GP IIb-IIIa rebound to this affinity matrix. This indicates that most of the GP IIb-IIIa used for the lZ5I-YNRGDS binding assays was capable of binding at least one RGD-containing peptide. Thus, it can be concluded that the GP IIb-IIIa complex contains a single binding site for RGD peptides. The binding of fibrinogen to GP IIb-IIIa also shows a 1:l stoichiometry as calculated from the number of fibrinogen-binding sites (37) and of GP IIb-IIIa complexes on the surface of stimulated platelets (38).
(ii) Binding of the hexapeptide YNRGDS to GP IIb-IIIa requires the presence of millimolar concentrations of either Ca2+ or M$+, which is consistent with the divalent cation requirement for fibrinogen binding to stimulated platelets Therefore, it appears that the low affinity binding site(s) for these cations are probably located on GP 1%-IIIa and not on fibrinogen. The two cations either induce the binding sites for YNRGDS and fibrinogen on GP IIb-IIIa or they are directly involved in bridging GP IIb-IIIa to the two ligands.
In addition to these low affinity binding sites we found a class of higher affinity on GP IIb-IIIa that is specific for Ca2+. Incubation of purified GP IIb-IIIa in the presence of 1 mM Mg2+ and varying concentrations of free Ca2+ indicated that maximal binding of YNRGDS required 1 pM Ca2+, since at 0.1 p~ Ca2+ YNRGDS binding decreased to 37%. Concomitantly with the loss of binding a significant extent of dissociation of the heterodimeric complex was observed. It has been reported that in the presence of Mg2+ maximal fibrinogen binding to activated platelets (41) and isolated platelet membranes (39) also occurs only at Ca2+ concentrations 2 1 pM. However, in intact platelets the dissociation of the GP IIb-IIIa complex is not the cause of the observed loss of fibrinogen binding at low Ca2+ concentrations, because the binding experiments were performed at 22 "C (41), a temperature where Ca2+ concentrations in the nanomolar range do not induce dissociation of the GP IIb-IIIa complex (25,42,43). Although the binding of YNRGDS to GP IIb-IIIa could be regulated by these Ca2+ ions identically to the binding of fibrinogen, we cannot exclude the possibility that the dissociation of the GP IIb-IIIa complex was causing the decrease in the YNRGDS binding.
Brass and Shattil (44, 45) have reported that GP IIb-IIIa contains two high affinity ( K d = 9 nM) and approximately six low affinity ( K d = 400 nM) binding sites for Caz+. The present study suggests that GP IIb-IIIa might also contain at least one very low affinity binding site (Kd -200 p~) for either Ca2+ or M P . (iii) The binding sites on GP IIb-IIIa for fibrinogen and for the hexapeptide YNRGDS appear to overlap. This conclusion is supported by the finding that the monoclonal antibody A2A9, which has previously been shown to inhibit fibrinogen binding to activated platelets (30), also inhibited YNRGDS binding to purified GP IIb-IIIa. Recently it has been reported that the synthetic tetrapeptide RGDS does not inhibit AzAg binding to stimulated platelets (46). The reason for this apparent difference is presently not known, but the size of the RGD-containing peptides used could be critical for its inhibitory effect.
(iv) The binding of lZ5I-YNRGDS to GP IIb-IIIa was inhibited by the dodecapeptide HHLGGAKQAGDV as is fibrinogen binding to intact platelets (12,13) and to purified GP IIb-IIIa incorporated into vesicles (40). There is convincing evidence that RGD-containing peptides and the dodecapeptide corresponding to the C terminus of the fibrinogen y-chain are binding to GP IIb-IIIa in a mutually exclusive manner (11, 16, 17, 46). Our data confirm this finding. Two explanations for the mutually exclusive binding of the two peptides have been discussed in the literature (16,46); either the binding of one peptide induces a conformational change in GP IIb-IIIa so that the other peptide cannot bind anymore or both peptides bind to the same or to an overlapping site on GP IIb-IIIa. Further studies are required to clearly differentiate between these two possibilities.
In summary, we have demonstrated that the hexapeptide YNRGDS, corresponding t o a sequence present in the fibrinogen a-chain at position 570-575, binds to purified GP IIb-IIIa with properties very comparable with fibrinogen binding to activated platelets. Our data strongly suggest that the RGD-containing domains within the adhesive proteins are indeed responsible for the binding to GP IIb-IIIa. This con-clusion is supported by the recently reported results that antibodies specifically recognizing the RGD-containing domains within von Willebrand factor (47) and fibronectin (48) inhibit the binding of these adhesive proteins to activated platelets.