Agonist-specific Structural Rearrangements of Integrin αIIbβ3
2002; Elsevier BV; Volume: 277; Issue: 42 Linguagem: Inglês
10.1074/jbc.m205886200
ISSN1083-351X
AutoresMarı́a J. Calzada, María Victoria Álvarez, José González-Rodrı́guez,
Tópico(s)Biochemical and Structural Characterization
ResumoConcrete structural features of integrin αIIbβ3 on the surface of platelets (at rest and after activation) have been obtained from epitope maps based on cross-competition among monoclonal antibodies directed against the αIIb subunit calf-2 domain and the β3 subunit βA domain of αIIbβ3. At rest, the observed intersubunit interface is formed by the sequence stretches β3-(150–216), αIIb light chain-(1–92), and αIIb heavy chain-(826–856); and the αIIb interchain interface is formed by the two latter sequence stretches, disulfide-bonded between αIIb heavy chain Cys826 and αIIb light chain Cys9. These structural features agree with those observed in the αIIbβ3 rudimentary connectivity map in solution and with the αvβ3 V-shaped crystal structure (Xiong, J.-P., Zhang, R., Dunker, R., Scott, D. L., Joachimiak, A., Goodman, S. L., and Arnaout, M. A. (2001) Science 294, 339–345), but they disagree with the domain disposition suggested by the actual ultrastructural model. The epitope maps in platelets activated by ADP, thrombin receptor activation peptide, and arachidonic acid differ not only from those in platelets at rest, but also among themselves. The structural rearrangements observed confirm the presence in activated platelets of the crystallographically observed knee and argue against the switchblade mechanism proposed for activation (Beglova, N., Blacklow, S. C., Takagi, J., and Springer, T. A. (2002) Nat. Struct. Biol. 9, 282–287), demonstrate the existence of αIIbβ3 agonist-specific activation states, explain the specificity for ligand binding and functional inhibition for some agonists, and predict the existence of agonist-specific final effectors and receptor activation mechanisms. The distinct non-reciprocal competition patterns observed at rest and after activation support the agonist-specific activation states and the existence of intrasubunit and intersubunit allosteric effects, previously proposed as the mechanism for αIIbβ3 transmembrane activation. Concrete structural features of integrin αIIbβ3 on the surface of platelets (at rest and after activation) have been obtained from epitope maps based on cross-competition among monoclonal antibodies directed against the αIIb subunit calf-2 domain and the β3 subunit βA domain of αIIbβ3. At rest, the observed intersubunit interface is formed by the sequence stretches β3-(150–216), αIIb light chain-(1–92), and αIIb heavy chain-(826–856); and the αIIb interchain interface is formed by the two latter sequence stretches, disulfide-bonded between αIIb heavy chain Cys826 and αIIb light chain Cys9. These structural features agree with those observed in the αIIbβ3 rudimentary connectivity map in solution and with the αvβ3 V-shaped crystal structure (Xiong, J.-P., Zhang, R., Dunker, R., Scott, D. L., Joachimiak, A., Goodman, S. L., and Arnaout, M. A. (2001) Science 294, 339–345), but they disagree with the domain disposition suggested by the actual ultrastructural model. The epitope maps in platelets activated by ADP, thrombin receptor activation peptide, and arachidonic acid differ not only from those in platelets at rest, but also among themselves. The structural rearrangements observed confirm the presence in activated platelets of the crystallographically observed knee and argue against the switchblade mechanism proposed for activation (Beglova, N., Blacklow, S. C., Takagi, J., and Springer, T. A. (2002) Nat. Struct. Biol. 9, 282–287), demonstrate the existence of αIIbβ3 agonist-specific activation states, explain the specificity for ligand binding and functional inhibition for some agonists, and predict the existence of agonist-specific final effectors and receptor activation mechanisms. The distinct non-reciprocal competition patterns observed at rest and after activation support the agonist-specific activation states and the existence of intrasubunit and intersubunit allosteric effects, previously proposed as the mechanism for αIIbβ3 transmembrane activation. monoclonal antibodies thrombin receptor activation peptide arachidonic acid fluorescein isothiocyanate heavy chain light chain Integrin αIIbβ3, a 230-kDa Ca2+-dependent heterodimeric protein peculiar to megakaryocytes and blood platelets, serves as the receptor for fibrinogen and other adhesive proteins upon platelet stimulation; also known as glycoprotein IIb/IIIa, it plays a crucial role in platelet adhesion to and spreading on the subendothelium, platelet aggregation, and clot retraction (1Plow E.F. Ginsberg M.H. Prog. Hemostasis Thromb. 1989; 9: 117-156PubMed Google Scholar, 2Kieffer N. Phillips D.R. Ann. Rev. Cell Biol. 1990; 6: 329-358Crossref PubMed Scopus (261) Google Scholar, 3Humphries M.J. Biochem. Soc. Trans. 2000; 28: 311-339Crossref PubMed Google Scholar). The genetic, biochemical, immunochemical, molecular dynamic, functional, pathological, and pharmacological characterization of αIIbβ3 is more advanced than similar knowledge about any other member of the integrin family. To acquire the receptor capacity (viz. recognition and binding of adhesive proteins), αIIbβ3requires some induction mechanism to take place, which is physiologically linked to platelet activation by a variety of agonists and whose final step consists most probably of a change in the quaternary structure of the integrin (4Shattil S.J. Kashiwagi H. Pampori N. Blood. 1998; 91: 2645-2657Crossref PubMed Google Scholar). At present, the best molecular picture of αIIbβ3 in solution is that derived from protein chemical analysis (5Calvete J.J. Henschen A. González-Rodrı́guez J. Biochem. J. 1989; 264: 561-568Crossref Scopus (53) Google Scholar, 6Calvete J.J. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 274: 63-71Crossref PubMed Scopus (158) Google Scholar, 7Calvete J.J. Arias J. Alvarez M.V. López M.M. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 273: 457-463Crossref Scopus (49) Google Scholar, 8Calvete J.J. Arias J. Alvarez M.V. Lopez M.M. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 273: 767-775Crossref PubMed Scopus (34) Google Scholar, 9Calvete J.J. Mann K. Alvarez M.V. López M.M. González- Rodrı́guez J. Biochem. J. 1992; 282: 523-532Crossref PubMed Scopus (46) Google Scholar, 10Calvete J.J. Schäfer W. Mann K. Henschen A. González-Rodrı́guez J. Eur. J. Biochem. 1992; 206: 759-765Crossref PubMed Scopus (38) Google Scholar); x-ray crystallography of the extracellular domain of αvβ3 (11Xiong J.-P. Zhang R. Dunker R. Scott D.L. Joachimiak A. Goodman S.L. Arnaout M.A. Science. 2001; 294: 339-345Crossref PubMed Scopus (1100) Google Scholar) and the NMR structure of the epidermal growth factor-3 domain of β2 (12Beglova N. Blacklow S.C. Takagi J. Springer T.A. Nat. Struct. Biol. 2002; 9: 282-287Crossref PubMed Scopus (259) Google Scholar); mapping of antibody epitopes (13Bachelot C. Rendu F. Gulino D. Semin. Thromb. Hemostasis. 1995; 21: 23-36Crossref PubMed Scopus (21) Google Scholar), ligand-mimetic cross-linking sites (10Calvete J.J. Schäfer W. Mann K. Henschen A. González-Rodrı́guez J. Eur. J. Biochem. 1992; 206: 759-765Crossref PubMed Scopus (38) Google Scholar, 14D'Souza S.E. Ginsberg M.H. Burke T.A. Plow E.F. J. Biol. Chem. 1990; 265: 3440-3446Abstract Full Text PDF PubMed Google Scholar), and natural and site-directed mutations (15O'Tool T.E. Madelman D. Forsyth J. Shattil S.J. Plow E.F. Ginsberg M.H. Science. 1991; 254: 845-847Crossref PubMed Scopus (272) Google Scholar, 16Huges P.E. Diaz-González F. Leong L., Wu, C. McDonnald J.A. Shattil S.J. Ginsberg M.H. J. Biol. Chem. 1996; 271: 6571-6574Abstract Full Text Full Text PDF PubMed Scopus (505) Google Scholar); electron microscopy (17Carrel N.A. Fitzgerald L.A. Steiner B. Erickson H.P. Phillips D.R. J. Biol. Chem. 1985; 260: 1743-1749PubMed Google Scholar); and molecular size and shape (18Rivas G.A. Aznarez J.A. Usobiaga P. Saiz J.L. González-Rodrı́guez J. Eur. Biophys. J. 1991; 20: 335-345Crossref Scopus (11) Google Scholar). Integrins in solution appear in electron micrographs as a globular head with two legs, with the total length from tip to tip including the globular head being ∼40 nm. The legs of αIIbβ3 are flexible and emerge from the same side of the head, with their tips either separated but never reaching a diametrical disposition (the head-free tail shapes), touching (the empty oval shapes) or overlapping (the bilobular shapes) each other, or fully folded on themselves and on the head (the filled globular shapes) (18Rivas G.A. Aznarez J.A. Usobiaga P. Saiz J.L. González-Rodrı́guez J. Eur. Biophys. J. 1991; 20: 335-345Crossref Scopus (11) Google Scholar). This high degree of flexibility of the tails of the heterodimer, seen in electron micrographs and deduced from the crystallographic structure, accounts for the short rotational correlation times measured in Triton X-100 solutions by time-resolved fluorescence anisotropy. 1M. P. Lillo, G. A. Rivas, A. U. Acuña, and J. González-Rodrı́guez, unpublished data. 1M. P. Lillo, G. A. Rivas, A. U. Acuña, and J. González-Rodrı́guez, unpublished data.Given its high segmental mobility, αIIbβ3should adopt a large variety of shapes also in the membrane, where its lateral and rotational mobility has been measured (19González-Rodrı́guez J. Acuña A.U. Alvarez M.V. Jovin T.M. Biochemistry. 1994; 33: 266-274Crossref PubMed Scopus (27) Google Scholar, 20Schootemeijer A. van Willigen G. van der Vuurst H. Tertoolen L.G.J. de Laat S.W. Akkerman J.W.N. Thromb. Haemostasis. 1997; 77: 143-149Crossref PubMed Scopus (11) Google Scholar). The integrin model that locates the subunit N-terminal domains at the globular head and the transmembrane and cytoplasmic domains at the tail ends was based on a combination of the morphological results and secondary structure prediction (21Nermut M.V. Green N.M. Eason P. Yamada S.S. Yamada K.M. EMBO J. 1988; 7: 4093-4099Crossref PubMed Scopus (193) Google Scholar) and confirmed by the crystallographic structure. Electron micrographs of fibrinogen-αIIbβ3 complexes suggest that the fibrinogen-binding site(s) is located at the integrin globular head (22Weisel J.W. Nagaswami C. Vilaire G. Bennett J.S. J. Biol. Chem. 1992; 267: 16637-16643Abstract Full Text PDF PubMed Google Scholar), and the cysteine-rich and proteinase-resistant core of β3 (β3-(423–622)) was tentatively localized in one of the heterodimer tails (23Wippler J. Kouns W.C. Schlaeger E.-J. Kuhn H. Hadvary P. Steiner B. J. Biol. Chem. 1994; 269: 8754-8761Abstract Full Text PDF PubMed Google Scholar). Finally, the sequences of the epitopes for monoclonal antibodies B1B5 (located at the C-terminal end of the αIIb heavy chain) and LIBS2 (located within β3-(602–690)) were also assigned to the ends of one of the tails (22Weisel J.W. Nagaswami C. Vilaire G. Bennett J.S. J. Biol. Chem. 1992; 267: 16637-16643Abstract Full Text PDF PubMed Google Scholar, 24Du X., Gu, M. Weisel J.W. Nagaswami C. Bennett J.S. Bowditch R. Ginsberg M.H. J. Biol. Chem. 1993; 268: 23087-23092Abstract Full Text PDF PubMed Google Scholar). The latter sequence was thought to be joined by the Cys406–Cys655 disulfide bond to the mid-region of β3 (not confirmed in the crystal structure), which was also joined to the N-terminal end by the Cys5–Cys435 disulfide bond (6Calvete J.J. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 274: 63-71Crossref PubMed Scopus (158) Google Scholar). Recently, the crystal structure of αvβ3 confirmed the metal ion-dependent adhesion site proposed earlier in the N-terminal region of β3 (MIDAS domain) (25Tozer E.C. Liddington R.C. Sutcliffe M.J. Smeeton A.H. Loftus J.C. J. Biol. Chem. 1996; 271: 21978-21984Abstract Full Text Full Text PDF PubMed Scopus (161) Google Scholar) and the β propeller with seven blades predicted in the N-terminal domain of the α subunits (26Springer T.A Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 65-72Crossref PubMed Scopus (386) Google Scholar). However, discrepancies appeared in the disulfide arrangement of β3 proposed by protein chemical methods, x-ray diffraction, and NMR. Based on several pieces of evidence, a two-state allosteric model of integrin activation had been proposed (27Loftus J.C. Liddington R.C. J. Clin. Invest. 1997; 99: 2302-2306Crossref PubMed Google Scholar). The final effector(s) for the proposed allosteric transition has not been identified (4Shattil S.J. Kashiwagi H. Pampori N. Blood. 1998; 91: 2645-2657Crossref PubMed Google Scholar), and information on the structure and disposition of the integrin in the membrane, both at rest and after activation, is lacking. It has been known for some time that ADP- and thrombin-activated platelets (1Plow E.F. Ginsberg M.H. Prog. Hemostasis Thromb. 1989; 9: 117-156PubMed Google Scholar), as well as platelets activated with some monoclonal antibodies (28Frelinger III, A.L. Du X.-P. Plow E.F. Ginsberg M.H. J. Biol. Chem. 1991; 266: 17106-17111Abstract Full Text PDF PubMed Google Scholar), differ in their specificity for the binding of adhesive proteins and monoclonal antibodies. Additionally, antibodies AP3 and Tab completely abolish ADP- induced platelet aggregation and secretion without impairing fibrinogen binding; however, thrombin-activated platelets neither aggregate nor bind fibrinogen in the presence of both antibodies (29Newman P.J. McEver R.P. Doer M.P. Kunicki T.J. Blood. 1987; 69: 668-676Crossref PubMed Google Scholar). Similarly, some anti-αIIb and anti-β3 antibodies show agonist-dependent immunochemical inhibition of platelet aggregation in vitro (30de Castellarnau C. Cullaré C. Alvarez M.V. Muñiz-Diez E. Calzada M.J. González-Rodrı́guez J. Platelets. 1997; 8: 243-253Crossref PubMed Scopus (5) Google Scholar). Together, these observations suggest the existence of agonist-specific activated states of αIIbβ3 or different agonist-dependent membrane vicinities of the activated receptor. The existence of activated states differing in ligand affinity has been shown before for other integrins (31Masumoto A. Hemler M.E. J. Biol. Chem. 1993; 268: 228-234Abstract Full Text PDF PubMed Google Scholar, 32Ortlepp S. Stephens P.E. Hogg N. Figdor C.G. Bobinson M.K. Eur. J. Immunol. 1995; 25: 637-643Crossref PubMed Scopus (58) Google Scholar, 33Luque A. Gomez M. Puzon W. Takada Y. Sánchez-Madrid F. Cabañas C. J. Biol. Chem. 1996; 271: 11067-11075Abstract Full Text Full Text PDF PubMed Scopus (267) Google Scholar, 34Garcı́a A.J. Takagi J. Boettiger D. J. Biol. Chem. 1998; 273: 34710-34715Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar, 35Chen L.L. Whitty A. Lobb R.R. Adams S.P. Pepinsky R.B. J. Biol. Chem. 1999; 274: 13167-13175Abstract Full Text Full Text PDF PubMed Scopus (70) Google Scholar), although we still lack information on the structural basis of these different activated states or receptor vicinities. On the other hand, “activated” forms of αIIbβ3 have been generated in cells, in whole platelets, and in solution by site-directed mutagenesis (36O'Tool T.E. Loftus J.C., Du, X. Glass A.A. Ruiggeri Z.M. Shattil S.J. Plow E.F. Ginsberg M.H. Cell Reg. 1991; 1: 883-893Crossref Scopus (224) Google Scholar), interaction with RGD peptides (24Du X., Gu, M. Weisel J.W. Nagaswami C. Bennett J.S. Bowditch R. Ginsberg M.H. J. Biol. Chem. 1993; 268: 23087-23092Abstract Full Text PDF PubMed Google Scholar) and monoclonal antibodies (28Frelinger III, A.L. Du X.-P. Plow E.F. Ginsberg M.H. J. Biol. Chem. 1991; 266: 17106-17111Abstract Full Text PDF PubMed Google Scholar), chemical reduction of disulfides (37Yan B. Smith J.W. Biochemistry. 2001; 40: 8861-8897Crossref PubMed Scopus (114) Google Scholar), etc.; however, we lack information on the structural differences between the resting and activated forms. Agonist-specific activation states of αIIbβ3 in the cell can be demonstrated by agonist-specific antibody recognition patterns. In this work, we have used a collection of monoclonal antibodies (mAbs)2 directed against the calf-2 domain of αIIb and the βA domain of β3, whose epitopes have been located in their primary structures by protein chemical and immunochemical analyses (Refs. 7Calvete J.J. Arias J. Alvarez M.V. López M.M. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 273: 457-463Crossref Scopus (49) Google Scholar, 8Calvete J.J. Arias J. Alvarez M.V. Lopez M.M. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 273: 767-775Crossref PubMed Scopus (34) Google Scholar, and 38Melero J.A. González-Rodrı́guez J. Eur. J. Biochem. 1984; 141: 421-427Crossref PubMed Scopus (20) Google Scholar and this work). Competition among these antibodies for binding to whole platelets at rest and to ADP-, thrombin receptor activation peptide (TRAP)-, and arachidonic acid (AA)-activated platelets was quantified by bound-phase separation assays. The structural features deduced from the antibody competition patterns in resting platelets differ from those deduced from the patterns in ADP-, TRAP-, and AA-activated platelets, and the latter differ among themselves. We have compared these two-dimensional low resolution topographical maps and discussed them in relation to the structural and functional information currently available for integrin αIIbβ3, and we have concluded that the maps are consistent with the existence of conformationally different agonist-dependent activation states of the integrin and with the bent structure both at rest and after activation and therefore inconsistent with the switchblade mechanism proposed for integrin activation (12Beglova N. Blacklow S.C. Takagi J. Springer T.A. Nat. Struct. Biol. 2002; 9: 282-287Crossref PubMed Scopus (259) Google Scholar). Blood samples were collected from healthy volunteers, and 9 volumes of blood were anticoagulated with 1 volume of 0.129 m trisodium citrate. Platelet-rich plasma was prepared by sedimentation of red and white blood cells at 150 × g for 10 min. Washed platelets were prepared by gel filtration on a Sepharose 2B column equilibrated with Tyrode's buffer, 0.2% bovine serum albumin, and 5 mm glucose without calcium. Integrin αIIbβ3 was prepared from the membrane fraction of outdated human platelets as described (39Rivas G.A. Calvete J.J. González-Rodrı́guez J. Protein Expression Purif. 1991; 2: 248-255Crossref PubMed Scopus (12) Google Scholar). The purified protein, free from salts and detergent, was freeze-dried and stored in liquid nitrogen. The αIIb and β3 subunits were prepared from pure αIIbβ3 as described previously (39Rivas G.A. Calvete J.J. González-Rodrı́guez J. Protein Expression Purif. 1991; 2: 248-255Crossref PubMed Scopus (12) Google Scholar). The murine mAbs specific for β3 (P6, P23-7, P37, P40, P95-1, P95-2, and P97) and αIIb (Mβ1, Mβ2, Mβ3, Mβ4, Mβ5, M3, M6, M9, M10, and Mα2), all of the IgG class, were prepared according to immunization and fusion protocols and screening assays described previously (38Melero J.A. González-Rodrı́guez J. Eur. J. Biochem. 1984; 141: 421-427Crossref PubMed Scopus (20) Google Scholar). The antibodies were purified from ascites fluid after 50% saturated (NH4)2SO4 precipitation. The 50% saturated (NH4)2SO4 precipitate was subjected to affinity chromatography on protein A-Sepharose (Amersham Pharmacia Biotech) according to the manufacturer's instructions. The purification process was monitored by SDS-PAGE and enzyme immunoassay. Purified antibodies were labeled with fluorescein isothiocyanate (FITC) at pH 8.0–9.0 and filtered through a Sephadex G-50 column (1 × 40 cm) equilibrated with phosphate-buffered saline to separate the free dye. After clearing by centrifugation, the protein and bound dye in the FITC-antibody conjugates were measured spectrophotometrically at 280 and 495 nm, respectively, with the dye/protein molar ratio being between 2 and 5. The labeled antibodies were monitored by enzyme immunoassay and SDS-PAGE, and their affinity for αIIbβ3 in solution was measured by enzyme immunoassay (40Friguet B. Chaffote A.F. Djavadi-Ohaniance L. Goldberg M.E. J. Immunol. Methods. 1985; 11: 305-319Crossref Scopus (1125) Google Scholar). Concentrated solutions of antibodies were stored in liquid nitrogen. Proteins (41Markwell M.A.K. Haas S.M. Bieber L.L. Tolbert N.E. Anal. Biochem. 1978; 87: 206-210Crossref PubMed Scopus (5297) Google Scholar) and free sulfhydryl groups and disulfide bonds (42Thannhauser T.W. Konishi Y. Scheraga H.A. Methods Enzymol. 1987; 143: 115-119Crossref PubMed Scopus (195) Google Scholar) were determined as described previously. SDS-PAGE was carried out according to Laemmli (43Laemmli U.K. Nature. 1970; 277: 680-685Crossref Scopus (206271) Google Scholar). For immunoblotting, gel electrophoresis bands were transferred to nitrocellulose or Immobilon by a standard procedure (44Towbin H. Staehelin T. Gordon J. Proc. Natl. Acad. Sci. U. S. A. 1979; 76: 4350-4354Crossref PubMed Scopus (44724) Google Scholar). The goat anti-mouse IgG secondary antibody was affinity-purified and peroxidase-conjugated. The early 23- and 80-kDa tryptic fragments obtained by digestion of the β3subunit were prepared and characterized as described previously (7Calvete J.J. Arias J. Alvarez M.V. López M.M. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 273: 457-463Crossref Scopus (49) Google Scholar). The selective chemical reduction of the interchain disulfide bond of αIIb and the isolation and analysis of the two chains were performed following earlier work (8Calvete J.J. Arias J. Alvarez M.V. Lopez M.M. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 273: 767-775Crossref PubMed Scopus (34) Google Scholar). The affinity of the antibodies for αIIbβ3 in whole platelets and the number of binding sites per platelet were measured as follows. Increasing concentrations of the FITC-antibody conjugates were mixed with 5–10 μl of platelet-rich plasma (1.5–3 × 106 platelets) and brought to 50 μl with Tyrode's buffer, 0.1% bovine serum albumin, and 1 mmCaCl2 (equivalent to 5–10 nmαIIbβ3). After a 1-h incubation at 20 °C, the mixtures were loaded on top of 20% sucrose in 400-μl microcentrifuge tubes and centrifuged for 10 min at 5000 rpm. After, the tip of the tubes was cut and isolated using a razor blade, the supernatant was aspirated off the upper part, which was removed after thoroughly being washed and dried. The pellet was solubilized in 300 μl of 0.1% SDS, and FITC-bound mAbs were measured in an SLM-AMINCO 8000 spectrofluorometer (excitation at 492 nm and emission at 530 nm). In experiments with activated platelets, 100 μmADP, 20 μm TRAP, or 1 mm AA (final concentrations) was added with brief and gentle mixing (to avoid platelet aggregation) before antibody addition, and the samples were left resting for 1 h at room temperature. In the latter experiments, visual controls under the microscope were routinely done to check that the platelets had aggregated, but only after thorough mixing. Binding isotherms were constructed and fitted to a single binding site model to obtain the K d and the number of antibody-binding sites per platelet using Origin Version 5 (Microcal). FITC-labeled monoclonal antibody and competitive unlabeled antibody were added to 5–10 μl of human platelet-rich plasma (adjusted to 3 × 105 platelets/μl) to reach final concentrations just below the K d (dissociation constant) for the labeled antibody and 5–10-fold greater than theK d for the unlabeled antibody. The mixture was brought to a final incubation volume of 50 μl with Tyrode's buffer, 0.1% bovine serum albumin, and 1 mm CaCl2. Competition among antibodies for binding to human platelets was assayed by spectrofluorometric quantification of the platelet-bound FITC-labeled monoclonal antibodies after separation of the free and bound FITC-labeled antibodies by sedimentation as described above. Every antibody in the set was labeled and subjected to binding competition with the rest of the unlabeled antibodies. The results of the competition experiments in percent inhibition of binding of labeled antibodies by unlabeled antibodies were introduced in double-entry square tables: one entry for labeled (analysis) antibodies and the other for unlabeled (competition) antibodies. We considered paired antibodies to have competed sterically with each other when the percent inhibition for the pair in both entries was >75%. From the binary square matrix of data, which indicate the ability of paired antibodies to bind simultaneously or not to the antigen, the x-y coordinates for each antibody epitope were adjusted according to the experimental data to construct the two-dimensional map of the relative distribution of the antibody epitopes on the receptor surface. For this, we considered an average epitope area of 7 nm2 and assigned the 0,0 coordinates to the epitope of the antibody that competed reciprocally with the highest number of antibodies in the set. The representation was done with Grapher Version 4 (Golden Software), with the position of non-overlapping epitopes being totally arbitrary. The extent and rate of platelet aggregation in platelet-rich plasma, prepared as indicated above, were measured at 37 °C under constant stirring in a spectrophotometer. Inhibition by anti-αIIb and anti-β3 monoclonal antibodies of platelet aggregation induced by ADP, TRAP, and AA was expressed as a percentage of the decrease in the extent of aggregation with respect to the extent of aggregation of sample controls without antibody. Tryptic digestion of β3 in solution allows the selective and sequential cleavage of the Arg–X and Lys–X peptide bonds of β3(Arg216–Asp217, Arg261–Leu262, and Lys302–Asn303), which are more accessible in solution. After selective reduction and alkylation of the β3 Cys5–Cys435 disulfide bond in the digestion mixture, the following fragments were sequentially obtained: 80 kDa (β3-(217–762)), 23 kDa (β3-(1–216)), 70 kDa (β3-(262–762)), 17 kDa (β3-(1–150)), and 52 kDa (β3-(303–762)), as shown previously (7Calvete J.J. Arias J. Alvarez M.V. López M.M. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 273: 457-463Crossref Scopus (49) Google Scholar). Using enzyme immunoassay and immunoelectroblotting, it was found that the epitope for P6 is resistant to degradation, and it was located in the 80-, 70-, and 50-kDa peptides. Given that it is exposed in whole platelets, its final location was limited to β3-(303–692). The epitopes for P23-7 and P37 were found in the 23-kDa N-terminal fragment (β3-(1–216)), as described previously (7Calvete J.J. Arias J. Alvarez M.V. López M.M. Henschen A. González-Rodrı́guez J. Biochem. J. 1991; 273: 457-463Crossref Scopus (49) Google Scholar), where the epitopes for P40, P95-1, P95-2, and P97 were now also located. Of all these epitopes, that for P23-7 was the only one remaining in the 17-kDa fragment (β3-(1–150)) (Fig.1 A). Given that these epitopes are resistant to reduction and alkylation and given that the 23- and 17-kDa fragments are not internally cleaved, as demonstrated by SDS-PAGE after full reduction and alkylation, the epitopes for P37, P40, P95-1, P95-2, and P97 were now located in β3-(150–216). Immunoelectroblotting and enzyme immunoassay detected the location of the M9 epitope in the heavy chain of αIIb, whereas the epitopes for Mβ2, Mβ3, Mβ4, and Mβ5 were located in the light chain of αIIb (Fig. 1 B). Given that the latter antibodies bind to whole platelets, their epitopes were located in the N terminus and extracellular domain of the light chain (αIIbL2-(1–92)). On the other hand, the epitope for M10 was destroyed by selective reduction of the interchain disulfide bond αIIbH Cys826–αIIbL2Cys9 of the αIIb subunit, but not by denaturation; therefore, we located this epitope in the interchain region formed by the C terminus of αIIbH and the N terminus of αIIbL. The K d values for the binding of the anti-αIIb and anti-β3 FITC-mAb complexes in intact platelets at rest and after ADP, TRAP, and AA activation together with the number of mAb-binding sites per platelet and the epitope localization for each antibody are given in TableI. Examples of isotherms for binding for some anti-αIIb and anti-β3 antibodies to resting and activated platelets are given in Fig.2. For a given mAb, theK d is generally higher for the isolated subunits (data not shown) and lower for the heterodimer in whole platelets, as expected, given that these antibodies were raised by animal immunization with isolated subunits or subunit fragments. TheK d values in whole platelets ranged from 5 to 700 nm. An unexpected result is the wide range found for the number of mAb-binding sites per platelet for the different antibodies, ranging from 40 to 80 × 103. The total number of molecules of αIIb and β3/platelet determined in platelet homogenates by enzyme immunoassay was 98,382 ± 11,818 and 113,612 ± 11,800, respectively.Table IEpitope location of the anti-β3 and anti-αIIbmonoclonal antibodies (obtained by protein chemical and inmunochemical analysis), dissociation constants, and the number of antibody molecules bound per platelet (obtained from the isotherms of antibody binding to whole platelets at rest and after activation by ADP, TRAP, and AA)mAbEpitope locationKdMolecules× 10−3/plateletnRestADPTRAPAARestADPTRAPAAnmP6β3(302–692)aThis work.,bRefs. 7 and 8.45 ± 545 ± 446 ± 544 ± 360 ± 245 ± 248 ± 242 ± 25P23–7cAfter platelet incubation at 37 °C in 1 mmEDTA for 1 h.β3(1–150)aThis work.1843 ± 51480 ± 114P37β3(150–216)aThis work.9 ± 243 ± 110 ± 315 ± 445 ± 250 ± 251 ± 339 ± 48P40cAfter platelet incubation at 37 °C in 1 mmEDTA for 1 h.β3(150–216)aThis work.1237 ± 19083 ± 65P95–1β3(150–216)aThis work.19 ± 310 ± 29 ± 19 ± 250 ± 343 ± 245 ± 134 ± 46P95–2β3(150–216)aThis work.6 ± 17 ± 17 ± 120 ± 148 ± 333 ± 134 ± 139 ± 36P97β3(150–216)aThis work.9 ± 320 ± 311 ± 29 ± 255 ± 439 ± 335 ± 135 ± 36Mβ1αIIbL2-(4–24)bRefs. 7 and 8.381 ± 64332 ± 95625 ± 260105 ± 1759 ± 345 ± 453 ± 835 ± 15Mβ2αIIbL2-(1–92)aThis work.443 ± 44359 ± 100325 ± 85231 ± 8546 ± 239 ± 444 ± 335 ± 147Mβ3αIIbL2-(1–92)aThis work.680 ± 157722 ± 30365 ± 542 ± 44Mβ4αIIbL2-(1–92)aThis work.89 ±
Referência(s)