Polymorphism of human platelet membrane glycoprotein IIb associated with the Baka/Bakb alloantigen system

The human Baka/Bakb alloantigen system has been implicated in the pathogenesis of post-transfusion purpura and neonatal alloimmune thrombocytopenic purpura. Human alloantisera specific for either the Baka or Bakb allele have been shown to react exclusively with the heavy chain of membrane glycoprotein (GP) IIb. To investigate the structure of the Bak epitopes, we used the polymerase chain reaction (PCR) to amplify GPIIb cDNA synthesized from platelet RNA samples prepared from individuals of known serologic phenotype. Subsequent DNA sequence analysis of amplified GPIIb cDNAs derived from one Baka homozygous individual and one Bakb homozygous individual revealed a single nucleotide base difference near the 3' end of the mRNA encoding the GPIIb heavy chain. Short 13 base allele-specific oligonucleotides (ASO) containing the putative phenotype-specific base in the middle were then synthesized, end-labeled with digoxigenin-11-dUTP using terminal transferase, and used as probes in subsequent dot-blot hybridization experiments. Platelet RNA was prepared from a panel made up of four Baka/a, three Bakb/b, and two Baka/b individuals, and the mRNA encoding GPIIb was amplified using PCR and spotted onto nylon membranes. ASO hybridization showed that the nucleotide base difference identified above segregated with Bak phenotype in all nine individuals examined (P = .002). The base pair substitution results in an amino acid polymorphism at residue 843 of the mature heavy chain. The Baka form of GPIIb encodes an isoleucine at this position, whereas the Bakb allele contains a serine. Identification of the polymorphism associated with this clinically important alloantigen system should permit new therapeutic and diagnostic approaches for treating and managing patients with alloimmune thrombocytopenic disorders.     

Posttransfusion Purpura Associated with an Anti-Bakb

We describe a 46-year-old white woman with typical clinical features of posttransfusion purpura (PTP) whose serum held a platelet-specific alloantibody reactive with an antigen antithetical to Baka, i.e. anti-Bakb. The specificity of the antibody was confirmed by family analysis, a population study (expected versus observed gene frequency: 0.3651 versus 0.3984; n = 105) and localization of the antigen on glycoprotein IIb in radioimmunoprecipitation. Typing of family members and blood donors for platelet antigens disclosed that the patient had been preimmunized by two blood transfusions in 1981, while fetomaternal incompatibility for Bakb was ruled out (her three children and their father were Bakb negative). Treatment of PTP with corticosteroids and platelet transfusions was ineffective, but infusion of high-dose intravenous IgG resulted in a rapid increase in the platelet count.     

Posttransfusion purpura associated with an autoantibody directed against a previously undefined platelet antigen

Although alloantibody against the PLA1 platelet antigen is usually found in patients with posttransfusion purpura (PTP), the mechanism of destruction of the patient's own PLA1-negative platelets is unexplained. We used a sensitive immunoblot technique to detect antiplatelet antibodies in a patient with classic PTP. The patient's acute-phase serum contained antibodies against three proteins present in control (PLA1-positive) platelets: an antibody that bound to a previously unrecognized platelet protein of mol wt 120,000 [glycoprotein (GP) 120], antibodies that bound to PLA1 (mol wt 90,000), and an epitope of GP IIb (mol wt 140,000). The antibodies against PLA1 and GP IIb did not react with the patient's own PLA1-negative platelets, control PLA1-negative platelets, or thrombasthenic platelets. In contrast, the antibody against GP 120 recognized this protein in all three platelet preparations, but not in Bernard-Soulier or Leka (Baka)-negative platelets. Antibody against GP 120 was not detected in the patient's recovery serum, although the antibodies against PLA1 and GP IIb persisted. F(ab)2 prepared from the patient's acute-phase serum also bound to GP 120. These results suggest that in PTP, transient autoantibody production may be responsible for autologous (PLA1-negative) platelet destruction. In addition, alloantibodies against more than one platelet alloantigen may be found in this disease. The nature of the GP 120 autoantigen and the GP IIb-related alloantigen defined by our patient's serum remains to be determined.     

Comparative studies on an autoantigen in ITP and the Baka alloantigen on platelet membrane glycoprotein IIb

Platelet membrane glycoprotein (GP) IIb carries not only the Baka alloantigen but also an autoantigen in some patients with chronic idiopathic thrombocytopenic purpura (ITP). We previously reported one ITP patient (RY) with anti-GPIIb autoantibody in her plasma. In this report, we investigated whether the epitope of the autoantigen on GPIIb (case RY) is identical to that of the Baka alloantigen or not. The anti-Baka antibody used in this study was obtained from a mother with a child suffering from neonatal alloimmune thrombocytopenic purpura as reported by Okada et al. Immunoblotting showed that the anti-Baka antibody bound to GPIIb from Baka-positive platelets only. In contrast, the anti-GPIIb autoantibody in RY plasma bound to GPIIb, irrespective of Baka phenotype. In addition, after neuraminidase treatment of GPIIb on nitrocellulose paper, the anti-Baka antibody failed to bind to Baka-positive GPIIb, while the autoantibody still bound to GPIIb. From these data, we conclude that the epitope of the autoantigen in RY patient is different from that of the Baka alloantigen.     

Immunochemical characterization of an autoantigen on platelet glycoprotein IIb in chronic ITP: comparison with the Baka alloantigen

Employing an immunoblotting procedure, we have identified and characterized an autoantigen carried on glycoprotein (GP) IIb in a patient with chronic idiopathic thrombocytopenic purpura (ITP), and have compared the location of the autoantigen with that of the platelet-specific alloantigen Baka. Immunoblots, using the partially purified GP IIb/IIIa complex as the target antigen, indicated that GP IIb alpha carried both the ITP autoantigen and the Baka alloantigen. The ITP plasma contained another antibody against a 100 kD protein (P100), a trace contaminant in the GP IIb/IIIa sample, which is probably a proteolytic fragment of an internal 124 kD protein. After chymotrypsin treatment, the auto- and alloantigen were found to be located on 65 kD fragments detectable under reducing conditions. In addition, immunoblots made after two-dimensional nonreduced-reduced SDS-polyacrylamide gel electrophoresis (SDS-PAGE) directly demonstrated that both 65 kD fragments had a molecular weight of 80 kD under nonreducing conditions; this provides evidence that these fragments were one and the same, and were derived from GP IIb alpha. Immunoblots of platelets digested in situ with chymotrypsin indicated that the 65 kD fragment of GP IIb alpha was retained by the platelet membrane. We conclude, therefore, that a 65 kD fragment, which represents the membrane side of the chymotrypsin cleavage site on GP IIb alpha, carries a clinically important determinant(s) recognized not only by the anti-Baka alloantibody, but also by the ITP autoantibody.     

Effect of single amino acid substitutions on the formation of the PlA and Bak alloantigenic epitopes.

The subunits that comprise the platelet-specific integrin alpha IIb beta 3 are polymorphic in nature, with several allelic forms present in the human gene pool. Minor changes in the secondary and tertiary structures of platelet membrane glycoproteins (GP) IIb and IIIa encoded by these alleles can result in an alloimmune reaction after transfusion or during pregnancy. To better understand the molecular structure of the PlA alloantigen system, located on GPIIIa, and the Bak alloantigen on GPIIb, we used a heterologous mammalian expression system to express these integrin subunits in their known polymorphic forms. An expression vector containing the PlA1 form of a GPIIIa cDNA, which encodes a leucine at amino acid 33 (Leu33), was modified to express the PlA2-associated form encoding a proline at amino acid 33 (Pro33). Similarly, a Baka GPIIb cDNA expressing an isoleucine at amino acid 843 (IIe843) was modified to express the Bakb form containing a serine at the same position (Ser843). Transfection of these vectors into COS cells resulted in the synthesis of GPIIb and GPIIIa molecules that were identical in size to those present in platelet lysates. Immunoprecipitation of the GPIIIa-transfected COS lysates with PlA)-specific alloantisera indicated that the Leu33 form was recognized only by anti-PIA1 sera while the Pro33 form was bound only by anti-PlA2 sera, showing that single amino acid polymorphisms are necessary and sufficient to direct the formation of the PlA1 and PlA2 alloepitopes. Similar experiments with Bak allele-specific expression vectors indicated that while the amino acid polymorphism (IIe843 in equilibrium Ser843) was necessary, posttranslational processing of pro-IIb was required for efficient exposure of both the Baka and Bakb alloepitopes.     

Probable genetic linkage between genes coding for platelet-specific antigens of the PlA and Bak systems

Studies of the inheritance of the platelet-specific alloantigens PlA1/PlA2 and Baka/Bakb in informative families provided evidence that the genes determining PlA and Bak antigens are in close genetic linkage (P = .004). In 154 normal, unrelated Caucasian subjects it was found that the antigen PlA1 is associated with Baka and the antigen PlA2 with Bakb more often than would be expected by chance (P less than .05). Bak and PlA antigens are carried on platelet membrane glycoproteins IIb and IIIa, respectively, which are decreased or absent in the inherited platelet disorder Glanzmann's thrombasthenia. Thus, the findings suggest that the defect in this disease involves two closely linked genes.     

Immunochemical characterization of the platelet-specific alloantigen Leka: a comparative study with the PlA1 alloantigen

We report the immunochemical characterization of a new platelet- specific alloantigen detected using an IgG antibody isolated from the serum of a patient with posttransfusion purpura (PTP). In indirect immunoprecipitation experiments, the antibody, termed anti-Leka, predominantly precipitated glycoprotein (GP) IIb from Triton X-100 lysates of normal human platelets. In an immunoblot procedure, which involved the transfer of platelet polypeptides separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis to nitrocellulose membrane, anti-Leka bound exclusively to GP IIb. Under identical conditions, four anti-PlA1 antibodies each reacted with GP IIIa. No binding of anti-Leka IgG occurred to Leka (-) platelets or to their separated polypeptides although GP IIb was normally detected by Coomassie blue staining. After electrophoresis of reduced platelet proteins, the Leka determinant was localized to the IIb alpha chain. Thus, unlike the PlA1 antigen, the Leka determinant was not destroyed by disulfide reduction. Analysis of platelets from a patient with Glanzmann's thrombasthenia revealed little or no binding in the GP IIb position. Anti-Leka permitted the identification of 76,000 and 60,000 dalton fragments of GP IIb retained by the platelet following chymotrypsin treatment. Our results further highlight the immunogenicity of the GP IIb-IIIa complex. They also suggest that antibodies against GP IIb can cause the thrombocytopenia observed in PTP and that anti-PlA1 antibodies do not account exclusively for the pathophysiology of this immune disorder.     

Platelet Anti-BAKb Antibody Associated with Neonatal Alloimmune Thrombocytopenia

A platelet alloantibody of IgG class but unknown specificity was detected in the serum from the mother of an infant with neonatal alloimmune thrombocytopenia. The antibody reacted by indirect immunofluorescence with 33/65 random donors, 10/11 Baka-negative donors and 23/54 Baka-positive donors. The mother's platelet phenotype was PLA1, Koa, Kob, Baka and the father's was PLA1, Koa, Kob, Baka. Immunoblotting and immunoprecipitation of maternal serum with paternal platelets produced a band at molecular weight 140 kilodaltons identical to the band obtained with Baka antiserum. Family studies confirmed the allelic distribution of Baka and the unknown antigen. The platelet-specific antibody in this patient meets the criteria for an antibody to the new platelet antigen, Bakb.     

Human Monoclonal Fab Fragments Recovered from a Combinatorial Library Bind Specifically to the Platelet HPA-1a Alloantigen on Glycoprotein IIb–IIIa

Background and objectives: Certain clinical conditions are related to the presence of platelet-specific alloantibodies in the patient's serum. We studied the molecular diversity of HPA-1a antibodies to analyze some peculiarities of this antibody response. Materials and methods: Human antibody Fab fragments that bind to the platelet alloantigen HPA-1a on glycoprotein IIb–IIIa (GPIIbIIIa) were generated by using a recombinant phage display system. We established an immunoglobulin G1, kappa combinatorial library from the peripheral blood lymphocytes of a person undergoing a severe posttransfusion purpura. Results: Characterization of Fab clones selected from the fifth round of antigen-specific panning of this library demonstrates a highly specific reactivity to the HPA-1a alloantigen. The nucleotide sequence analysis of representative HPA-1a-specific clones reveals at least 3 distinct V_L and 3V_H gene segments that present an extensive degree of mutation as demonstrated by comparison of gene usage and homologies to the nearest germline genes. Conclusions: These human HPA-1a-specific Fab reagents should allow us to better understand the molecular mechanism involved in HPA-1a alloimmunization.     

Post-transfusion purpura associated with alloimmunization against the platelet-specific antigen, Baka

Post-transfusion purpura (PTP) with severe thrombocytopenia occurred eight days after transfusion in a 28-year-old woman and responded to treatment with prednisone and plasma exchange. In contrast to nearly all previously studied cases of PTP, the patient's platelets were PlA1-positive and anti-PlA1 antibody could not be detected in serum obtained during the thrombocytopenic episode. Her serum was found to contain an antibody specific for a recently described platelet-specific alloantigen, Baka, in addition to multiple HLA-specific antibodies. The patient's platelets, typed following recovery, were Baka-negative. These findings indicate that post-transfusion purpura can occur in association with alloimmunization to platelet-specific antigens other than PlA1. In performing the serologic studies, a close relationship and possible identity between Baka and another recently reported platelet antigen, Leka, was observed. A method for analyzing mixtures of cytotoxic platelet-reactive antibodies without separating the individual antibodies is described.     

Alloantigenic composition of the endothelial vitronectin receptor

Endothelial cells synthesize a heterodimeric adhesion molecule, the vitronectin receptor (VnR), which is similar to the platelet glycoprotein (GP)IIb/IIIa complex. The subunits of the endothelial VnR (VnR alpha and GPIIIa) have been studied for their ability to express alloantigens associated with platelet GPIIb and IIIa. We previously showed that endothelial GPIIIa can express the platelet alloantigen Zwa or PIA1, which is associated with GPIIIa. We studied the relationship between the expression of Zwa on platelets and endothelial cells in neonates (n = 13). Using immunoprecipitation and immunofluorescence techniques, we showed that the Zwa antigen is either expressed or absent from both platelets and endothelial cells of the same individual. This finding indicates that in both cell types the same gene is expressed. We also showed that Zwa-negative endothelial cells express Zwb (PIA2), in analogy to Zwa-negative platelets. Moreover, our results strongly suggest expression on endothelial cells of Yukb, a recently described platelet alloantigen, also located on GPIIIa. However, we could not demonstrate expression on the endothelial VnR alpha subunit of Baka, an alloantigen located on platelet GPIIb. These findings are in agreement with the concept that the endothelial GPIIIa subunit is more closely related to its platelet counterpart than to the endothelial VnR alpha subunit.     

Uncoupling in the expression of platelet GP IIb/IIIa in human endothelial cells and K562 cells: absence of immunologic crossreactivity between platelet GP IIb and the vitronectin receptor alpha chain [see comments]

Platelet glycoproteins (GP) IIb and IIIa exist as noncovalently associated Ca++-dependent heterodimer complexes within the platelet membrane and express the major platelet alloantigens Leka (Baka) and PIA1 (Zwa), which are genetic markers of GP IIb and GP IIIa, respectively. Since heterodimers immunologically related to platelet GP IIb/IIIa have been identified in a number of nucleated cell types, we tested anti-Leka and anti-PIA1 antiserum, polyclonal anti-platelet GP IIb/IIIa IgG, as well as a panel of 28 monoclonal anti-GP IIb, GP IIIa, or complex dependent anti-GP IIb/IIIa antibodies on endothelial cells, peripheral blood mononuclear cells, and the erythroleukemic cells HEL and K562 in order to determine whether nucleated cell GP IIb/IIIa related proteins and platelet GP IIb/IIIa are immunologically related. Using immunofluorescence, immunoblotting, and immunoprecipitation experiments, evidence is presented that (1) the alloantigen Leka is not expressed in endothelial cells of an individual whose platelets are of the Leka/PIA1 phenotype, whereas the PIA1 alloantigen is readily detectable in these cells, (2) that in contrast to HEL cells, which express platelet GP IIb/IIIa and are of the Leka/PIA1 phenotype, platelet GP IIb is immunologically undetectable in 12-O-tetradecanoyl- phorbol-13-acetate (TPA)-treated K562 cells despite the presence of platelet GP IIIa, and (3) that peripheral blood mononuclear cells do not express platelet GP IIb or GP IIIa on their cell surface.     

Platelet specific alloantigens on the platelet glycoprotein Ia/IIa complex

The majority of platelet alloantigens are located on platelet glycoproteins IIb/IIIa. This report describes a codominant allelic system carried on the glycoprotein Ia/IIa complex, which we originally designated as Zava/Zavb but which is identical to the Bra/Brb system. Furthermore Zava was found to be identical to Hca. The alloantigens could not be detected using a direct binding enzyme immunoassay (EIA) with intact platelets, but were readily detected using a glycoprotein capture EIA and by radioimmunoprecipitation techniques. The two index cases (designated as homozygous Zava and Zavb) had alloantibodies against the corresponding antigen and did not react with their own platelets. Using these alloantibodies and a monoclonal antibody that reacts with the platelet glycoprotein Ia/IIa complex (12F1), we demonstrated that all Ia/IIa molecules carry either Zava or Zavb and we found that Zava and Zavb are on discrete populations of Ia/IIa. Following immunodepletion using either anti-Zava or anti-Zavb, all detectable Ia/IIa complexes from the respective homozygous platelets were removed. Immunodepletion of heterozygous Zava/Zavb with either anti-Zava or anti-Zavb did not reduce the amount of Ia/IIa complexes precipitable using the alternate alloantiserum. Population studies (n = 50) indicated the phenotypic frequency of Zava/Zava is less than 1%; Zava/Zavb is 18% and Zavb/Zavb is 82%. Four different alloantisera that had either anti-Zava or anti-Zavb reactivity also carried reactivity against the Baka or Bakb antigens which may suggest an association in the immune response to these alleles.     

Brb, a Platelet Alloantigen Involved in Neonatal Alloimmune Thrombocytopenia

Serum from a pregnant woman with the May-Hegglin anomaly contained a platelet-specific antibody. The serum reacted in the platelet indirect immunofluorescence test (PIIFT) with 97.6% of random donor platelets and those of the father but not with the mother's own platelets. This antibody induced a moderate thrombocytopenia in the infant that responded to infusion of intravenous immunoglobulin concentrates. The platelet phenotypes were PLA1+, Baka+, Bra+/Brb- for the mother, PLA1+, Baka+, Bra-/Brb+ for the father, and PLA1+, Bra+/Brb+ for the neonate. Analysis of the maternal serum with an immunoassay based on monoclonal antibody immobilization of platelet antigens (MAIPA) and immunoprecipitation techniques demonstrated the absence of antibodies directed against HLA class I antigens and that the antigen recognized was located on the platelet-GpIa/IIa complex. This antigen was present on 113/115 random donor platelets, in 7 of the 7 unrelated May-Hegglin platelets, and only absent in 3/24 Bra+ individuals, including the mother. No platelet-specific antibody was present in the serum of the 7 unrelated May-Hegglin subjects. The antigen recognized by this platelet-specific antibody thus meets the criteria defining the antithetic allele of Bra, i.e. the Brb alloantigen.     

A dinucleotide deletion in exon 4 of the PlA2 allelic form of glycoprotein IIIa: implications for the correlation of serologic versus genotypic analysis of human platelet alloantigens

Platelets from a patient with a suspected case of posttransfusion purpura were subjected to alloantigen phenotyping and found to express the PlA1, but not the PlA2, allelic form of human platelet membrane glycoprotein (GP) IIIa on the platelet surface. However, genotyping showed unambiguously that the patient carried the genes for both of these GPIIa alleles. Based on these results, we postulated that the PlA2 allele was silent, ie, that this patient was a carrier for Glanzmann thrombasthenia (GT). Quantitative analysis of GPIIb-IIIa surface expression showed only 20,000 GPIIb-IIIa receptors/platelet, approximately half of the value obtained with control platelets. Southern blot analysis showed no large deletions or insertions within the GPIIIa gene, and amplification of all 14 exons encoding GPIIIa resulted in the production of normal sized polymerase chain reaction (PCR) products in all cases. DNA-sequence analysis showed an AG dinucleotide deletion affecting codons 210 and 211 within exon 4 of the GPIIIa gene, leading to a change in reading frame and the creation of a stop codon 38 nucleotides down-stream. The predicted truncated protein consists of only the first 223 of the normal 762 amino acids, thus accounting for the failure to express the PlA2 allele on the platelet surface. While encountered only rarely, carriers of either GT or Bernard Soulier syndrome that are at the same time heterozygous for human platelet alloantigenic epitopes found on GPIb, GPIIb, or GPIIIa have the possibility to give discrepant results when comparing genotypic versus phenotypic analysis. In such situations, the combination of serologic and DNA-based evaluation contributes complementary and beneficial diagnostic information than either one alone are able to provide.     

Early Immunization against Platelet Glycoprotein Ilia in a Newborn Glanzmann Type I Patient

Alloimmunization against platelet glycoprotein IIb and/or IIIa is a complication rarely observed during the evolution of type I Glanzmann's thrombasthenic patients. The occurrence of such alloantibodies is usually due to repeated blood transfusion and greatly complicates the treatment of these patients since they prevent effective platelet transfusion and might, theoretically, cause posttransfusion purpura. We describe the case of a newborn thrombasthenic patient who developed an IgG platelet allo-antibody 1 month after birth. The diagnosis of Glanzmann's thrombasthenia was complicated by the rare platelet phenotype (PLA1-negative PLA2-positive) of the healthy mother, which was probably heterozygous for the abnormal thrombasthenic gene. Immunofluorescence and immunoblotting techniques demonstrated that the patient antibody was principally directed against the platelet glycoprotein IIIa. Surprisingly, this patient had only received four blood transfusions (fresh frozen plasma on days 1 and 2, and standard red blood cell concentrates on days 5 and 6) before the discovery of the antibody, suggesting prior in utero sensitization. This study emphasizes the need for early diagnosis of the disease. Thrombasthenic patients should be transfused with deleukocyted platelet-free blood products.     

Demonstration of the heterogeneity of epitopes of the platelet-specific alloantigen, Baka

It is well known that the platelet-specific alloantigen, Baka is carried on glycoprotein (GP) IIb, but little is known about the biochemical characteristics of its epitopes. To clarify the characteristics of the epitopes, we examined the interaction of four anti-Baka sera (Yam, Lin, Kl and MO) with their epitopes, either with or without modifications by sodium dodecyl sulphate (SDS) and/or neuraminidase. By immunoprecipitation, all four antisera bound to the intact GP IIb/IIIa complex from a Baka-positive subject. In contrast, immunoblotting demonstrated that Yam, Lin and Kl bound to SDS-denatured GP IIb, while MO did not. When blotted GP IIb was treated with neuraminidase, Yam and Lin did not bind to desialylated GP IIb, while Kl still did. When the purified GP IIb/IIIa complex or washed platelets were treated first with neuraminidase followed by immunoblotting, the molecular weight of GP IIb decreased from 145 kD to 138 kD; Yam did not bind to desialylated GP IIb, but Kl did. Furthermore, to eliminate the effect of SDS, we examined the interaction of Yam and Lin with neuraminidase-treated platelets using flow cytometry. The results were the same as those obtained using immunoblotting. Our results thus demonstrate that the expression of the Baka epitopes is not uniform and that sialic acid contributes to the expression of some actual allogenic epitopes.     

Leka, a New Platelet Antigen Absent in Glanzmann''s Thrombasthenia

The serum of a patient who developed a posttransfusion purpura contained antibodies directed against a previously undescribed platelet antigen Lek a. The antiplatelet activity was present in the IgG fraction and was detected by immunofluorescence, 51Cr lysis and 14C-serotonin release. The frequency of the Lek a phenotype in the French population is 98.18%. Lek a does not appear to be sex-linked and seems to be closely related to the Bak a antigen. The Lek a antigen is not expressed on thrombasthenic platelets but is found on platelets from patients with the Bernard-Soulier syndrome which suggests that this antigen is carried by platelet glycoproteins IIb and/or IIIa.     

Characterization of platelet-specific alloantigens by immimoblotting: localization of Zw and Bak antigens

The glycoprotein localization of the platelet-specific antigens Zwa, Zwb and Baka and their presence on tryptic fragments of glycoproteins was studied by immunoblotting. Human platelets were solubilized and pre-cleared from platelet-associated IgG. The glycoproteins were separated on SDS polyacrylamide gels, transferred to nitrocellulose and incubated with platelet antibodies, followed by 125I-radiolabelled anti-human Ig antibodies. Glycoprotein IIb/IIIa were isolated from platelet lysates by immuno-affinity chromatography. These proteins were subjected to trypsin digestion, and then used for the immunoblot procedure with platelet antibodies. A glycoprotein specifically reacting with either anti-Zwa or anti-Zwb was found, with an apparent molecular weight of 88 kDa. This protein co-migrated, and was probably identical with, glycoprotein IIIa. After trypsin digestion the smallest fragment, reactive with IgG anti-Zwa or IgM anti-Zwb, had a molecular weight of approximately 23 kDa. IgG anti-Baka and anti-Leka antibodies reacted with a protein of 130 kDa from platelets of Bak(a+) donors. This protein was identified as glycoprotein IIb.