Coordinator's report: glycophorin/band 3 and associated antigens

A Balance between Transmembrane-Mediated ER/Golgi Retention and Forward Trafficking Signals in Glycophorin-Anion Exchanger-1 Interaction

Anion exchanger-1 (AE1) is the main erythroid Cl-/HCO3- transporter that supports CO2 transport. Glycophorin A (GPA), a component of the AE1 complexes, facilitates AE1 expression and anion transport, but Glycophorin B (GPB) does not. Here, we dissected the structural components of GPA/GPB involved in glycophorin-AE1 trafficking by comparing them with three GPB variants-GPBhead (lacking the transmembrane domain [TMD]), GPBtail (mainly the TMD), and GP.Mur (glycophorin B-A-B hybrid). GPB-derived GP.Mur bears an O-glycopeptide that encompasses the R18 epitope, which is present in GPA but not GPB. By flow cytometry, AE1 expression in the control erythrocytes increased with the GPA-R18 expression; GYP.Mur+/+ erythrocytes bearing both GP.Mur and GPA expressed more R18 epitopes and more AE1 proteins. In contrast, heterologously expressed GPBtail and GPB were predominantly localized in the Golgi apparatus of HEK-293 cells, whereas GBhead was diffuse throughout the cytosol, suggesting that glycophorin transmembrane encoded an ER/Golgi retention signal. AE1 coexpression could reduce the ER/Golgi retention of GPB, but not of GPBtail or GPBhead. Thus, there are forward-trafficking and transmembrane-driven ER/Golgi retention signals encoded in the glycophorin sequences. How the balance between these opposite trafficking signals could affect glycophorin sorting into AE1 complexes and influence erythroid anion transport remains to be explored.

Large red cell-derived membrane particles are major contributors to hypercoagulability in sickle cell disease

Sickle cell disease (SCD) is one of the most common inherited single gene disorders. Polymerisation of sickle hemoglobin results in erythrocytes that are inflexible and adherent, leading to coagulation, vascular and cellular activation and resultant blood vessel blockage. Previous studies have observed elevated numbers of red cell-derived particles (RCDP), also denoted extracellular vesicles, in SCD plasma. Here, imaging flow cytometry was used to quantify all RCDP in SCD plasma. A more heterogenous population of RCDP was observed than previously reported. Significantly, large right side-out red cell macrovesicles (MaV), 7 µm in diameter, were identified. Most RCDP were right side-out but a minor population of inside-out vesicles was also present. Electron micrographs confirmed the heterogenous nature of the RCDP detected. All MaV are decorated with prothrombotic phosphatidylserine (PS) and their removal from plasma lengthened clotting times by more than three-fold. Removal of all right side-out RCDP from SCD patient plasma samples resulted in a seven-fold increase in clotting time. These results indicate that MaV comprise a large area of prothrombotic membrane and are thus major contributors to hypercoagulation in SCD. Consequently, controlled removal of MaV and PS exposed RCDP from plasma could provide a novel therapy for managing this disease.

The Gerbich blood group system: old knowledge, new importance

Antigens of the Gerbich blood group system are expressed on glycophorin C (GPC) and glycophorin D (GPD), minor sialoglycoproteins of human erythrocytes. GPC and GPD help maintain erythrocyte shape of and contributes to the stability of its membrane. There are six high-prevalence Gerbich antigens: Ge2, Ge3, Ge4, GEPL (GE10), GEAT (GE11), GETI (GE12) and five low-prevalence Gerbich antigens: Wb (GE5), Ls^a (GE6), An^a (GE7), Dh^a (GE8), GEIS (GE9). Some Gerbich antigens (Ge4, Wb, Dh^a, GEAT) are expressed only on GPC, two (Ge2, An^a) are expressed only on GPD, while others (Ge3, Ls^a, GEIS, GEPL, GETI) are expressed on both GPC and GPD. Antibodies recognizing GPC/GPD may arise naturally (so-called "naturally-occurring RBC antibodies") or as the result of alloimmunization, and some of them may be clinically relevant. Gerbich antibodies usually do not cause serious hemolytic transfusion reactions (HTR); autoantibodies of anti-Ge2- or anti-Ge3 specificity can cause autoimmune hemolytic anemia (AIHA).

Autophagic vesicles on mature human reticulocytes explain phosphatidylserine-positive red cells in sickle cell disease

During maturation to an erythrocyte, a reticulocyte must eliminate any residual organelles and reduce its surface area and volume. Here we show this involves a novel process whereby large, intact, inside-out phosphatidylserine (PS)-exposed autophagic vesicles are extruded. Cell surface PS is a well-characterized apoptotic signal initiating phagocytosis. In peripheral blood from patients after splenectomy or in patients with sickle cell disease (SCD), the number of circulating red cells exposing PS on their surface is elevated. We show that in these patients PS is present on the cell surface of red cells in large (∼1.4 µm) discrete areas corresponding to autophagic vesicles. The autophagic vesicles found on reticulocytes are identical to those observed on red cells from splenectomized individuals and patients with SCD. Our data suggest the increased thrombotic risk associated with splenectomy, and patients with hemoglobinopathies is a possible consequence of increased levels of circulating mature reticulocytes expressing inside-out PS-exposed autophagic vesicles because of asplenia.

The structures of glycophorin C N-glycans, a putative component of the GPC receptor site for Plasmodium falciparum EBA-140 ligand

Glycophorins C and D are highly glycosylated integral sialoglycoproteins of human red blood cell membranes carrying the Gerbich blood group antigens. The O- and N-glycosidic chains of the major erythrocyte glycoprotein (Lisowska E. 2001, Antigenic properties of human glycophorins - an update. Adv Exp Med Biol, 491:155-169; Tomita M and Marchesi VT. 1975, Amino-acid sequence and oligosaccharide attachment sites of human erythrocyte glycophorin. Proc Natl Acad Sci USA, 72:2964-2968.) are well characterized but the structure of GPC N-glycans has remained unknown. This problem became important since it was reported that GPC N-glycans play an essential role in the interaction with Plasmodium falciparum EBA-140 merozoite ligand. The elucidation of these structures seems essential for full characterization of the GPC binding site for the EBA-140 ligand. We have employed detailed structural analysis using sequential mass spectrometry to show that many GPC N-glycans contain H2 antigen structures and several contain polylactosamine structures capped with fucose. The results obtained indicate structural heterogeneity of the GPC N-glycans and show the existence of structural elements not found in glycophorin A N-glycans. Our results also open a possibility of new interpretation of the data concerning the binding of P. falciparum EBA-140 ligand to GPC. We hypothesize that preferable terminal fucosylation of N-glycosidic chains containing repeating lactosamine units of the GPC Gerbich variant could be an explanation for why the EBA-140 ligand does not react with GPC Gerbich and an indication that the EBA-140 interaction with GPC is distinctly dependent on the GPC N-glycan structure.

Identification of binding domains on red blood cell glycophorins for Babesia divergens

BACKGROUND

Invasion of red blood cells (RBCs) is one of the critical points in the lifecycle of Babesia. The parasite does not invade other host cells. Earlier work has shown that GPA and GPB function as putative receptors during parasite invasion. The primary focus of this study was the delineation of parasite-binding domains on GPA and GPB.

STUDY DESIGN AND METHODS

The assay of choice to validate molecules that participate in invasion is an inhibition of invasion assay, in which changes in parasitemia are assessed relative to a wild-type assay (no inhibitors). Inhibition of invasion can be achieved by modification of different components of the assay or by the addition of competitors of the molecules that participate in invasion. In this study purified antibody fragments to various domains on GPA and GPB were tested for magnitude of inhibition of parasite invasion. Effects on invasion were monitored by assessment of Giemsa-stained smears every 24 hours.

RESULTS

Among 10 selected antibodies directed at various epitopes on GPA and GPB, antibodies directed against GPA(M) epitopes had the most severe effect (up to 35%) on inhibition of invasion, followed by antibodies directed against GPB(S) epitope (up to 24%).

CONCLUSION

This study confirms the role of RBC glycophorins A and B in Babesia divergens invasion and shows that the GPA(M) and GPB(S) epitopes are likely to play an important role in the entry process.

Murine monoclonal anti-s and other anti-glycophorin B antibodies resulting from immunizations with a GPB.s peptide

BACKGROUND

The blood group antigens S and s are defined by amino acids Met or Thr at position 29, respectively, on glycophorin B (GPB). Commercial anti-s reagents are expensive to produce because of the scarcity of human anti-s serum. Our aim was to develop hybridoma cell lines that secrete reagent-grade anti-s monoclonal antibodies (MoAbs) to supplement the supply of human anti-s reagents.

STUDY DESIGN AND METHODS

Mice were immunized with the GPB(s) peptide sequence TKSTISSQTNGETGQLVHRF. Hybridomas were produced by fusing mouse splenocytes with mouse myeloma cells (X63.Ag8.653). Screening for antibody production was done on microtiter plates by hemagglutination. Characterization of the MoAbs was done by hemagglutination, immunoblotting, and epitope mapping.

RESULTS

Eight immunoglobulin G MoAbs were identified. Five antibodies are specific by hemagglutination for s and two MoAbs, when diluted, are anti-S-like, but additional analyses shows a broad range of reactivity for GPB. Typing red blood cells (RBCs) for s from 35 donors was concordant with molecular analyses as were tests on RBCs with a positive direct antiglobulin test (DAT) from 15 patients. The anti-s MoAbs are most reactive with peptides containing the (31)QLVHRF(36) motif, with (29)Thr. By Pepscan analyses, the anti-S-like MoAbs reacted within the same regions as did anti-s, but independently of (29)Met. One antibody was defined serologically as anti-U; however, its epitope was identified as (21)ISSQT(25), a sequence common for both GPA and GPB.

CONCLUSION

In addition to their value as typing reagents, these MoAbs can be used to phenotype RBCs with a positive DAT without pre-test chemical modification.

Variations in Loxosceles spider venom composition and toxicity contribute to the severity of envenomation

Envenomation by Loxosceles spiders causes two main clinical manifestations: cutaneous and systemic loxoscelism. The factors contributing to the severity of loxoscelism are not fully understood. We have analysed biochemical and toxicity variations in venom of L. laeta and L. intermedia, with the aim to find a correlation with the seriousness of loxoscelism. Differences in expression of proteins, glycoproteins and sphingomyelinase activity were observed between venom from male and female spiders and between venom from the two species. These differences were reflected in the toxicity of the venoms including the capacity to induce complement-dependent haemolysis, dermonecrosis and lethality. Comparative analysis of gender and species, showed that these biological activities were more prominent in venom from female spiders, especially from L. laeta. Antiserum raised against venom from females L. laeta spiders had the highest efficacy in neutralizing venoms of males and females of both species. These results indicate that the severity of loxoscelism depends, at least partially, on the species and sex of the spider and suggest that for accidents involving L. laeta an specific serum therapy is necessary. Furthermore, it emphasizes the efficacy of the antiserum produced against L. laeta female venom in neutralizing Loxosceles venoms from different species and gender.

Analysis of the toxic potential of venom from Loxosceles adelaida, a Brazilian brown spider from karstic areas

Loxosceles adelaida spiders (Araneae, Sicariidae) are found near and inside the caves in the Parque Estadual Turistico do Alto Ribeira (PETAR), Sao Paulo, Brazil, which are visited by thousands of tourists every year. Several Loxosceles species are a public health problem in many regions of the world, by causing severe dermonecrosis and/or complement dependent haemolysis upon envenomation. The aim of this study was to characterize the biochemical and biological properties of L. adelaida venom and evaluate the toxic potential of envenomation by this non-synanthropic Loxosceles species. The biological activities of the L. adelaida venom was compared to that of Loxosceles gaucho, a synanthropic species of medical importance in Brazil. L. adelaida venom showed a similar potential to induce haemolysis, dermonecrosis and lethality as L. gaucho venom. L. adelaida crude venom was purified, yielding a 31 kDa component endowed with haemolytic and dermonecrotic activities. In conclusion, we show here that the troglophile Loxosceles species, L. adelaida, commonly found in the complex of caves from PETAR, is potentially able to cause envenomation with the same gravity of those produced by synanthropic species.

Recombinant glycophorins C and D as tools for studying Gerbich blood group antigens

BACKGROUND

The Gerbich blood group system antigens are carried on glycophorin C (GPC) and glycophorin D (GPD) and variants thereof. These glycoproteins have been expressed in a heterologous system to study the individual antigens and to determine whether Ana is antithetical to Ge2.

STUDY DESIGN AND METHODS

cDNAs encoding GPC, GPD, GPC.Yus, GPC.Ge, GPC.Lsa, and GPD.Lsa were transfected and stably expressed in a human embryonic kidney cell line (293T). Individual Gerbich antigens were analyzed with MoAbs and human polyclonal antibodies by flow cytometry and immunoblotting. Recombinant GPD and GPD.Ana were expressed transiently and analyzed for expression of Ge2 and Ana antigens.

RESULTS

All recombinant variants were detected with sialidase-resistant and -sensitive anti-Ge2, anti-Ge3, and anti-Ge4. Ge4 antigen expression was depressed in GPC.Ls(a) transfectants as well as on Ls(a+) RBCs. GPD.An(a) recombinant protein expressed Ana and Ge2 antigens.

CONCLUSION

Cell lines stably expressing glycosylated Gerbich proteins were developed in a heterologous system by transfecting individual variant forms of GPC and GPD. Unexpectedly, it was found that Ge4 antigen is reduced in both the GPC.Ls(a) recombinant and the Ls(a+) RBCs. It was also shown that Ana and Ge2 antigens were expressed on a single GPD.An(a) protein and, therefore, they cannot be antithetical.

Glycophorin C is the receptor for the Plasmodium falciparum erythrocyte binding ligand PfEBP-2 (baebl)

We report in this paper that glycophorin C (GPC) is the receptor for PfEBP-2 (baebl, EBA-140), the newly identified erythrocyte binding ligand of Plasmodium falciparum. PfEBP-2 is a member of the Duffy binding-like erythrocyte binding protein (DBL-EBP) family. Although several reports have been published characterizing PfEBP-2, the identity of its erythrocytic receptor was still unknown. Using a combination of enzymatically treated red blood cells (RBCs) and rare, variant RBCs lacking different surface proteins, we have shown that PfEBP-2 does not bind to cells lacking GPC. Additionally, we found that PfEBP-2 binds differentially to variants of GPC lacking exon 2 or exon 3, and determined that the binding domain on GPC is potentially restricted to amino acid residues 14 through 22 within exon 2. Thus PfEBP-2 is involved in a sialic acid-dependent pathway of invasion, which does not involve glycophorin A or glycophorin B and represents a novel route of entry into the RBCs.

Molecular cloning and expression of a functional dermonecrotic and haemolytic factor from Loxosceles laeta venom

The bite of spiders of the genus Loxosceles can induce a variety of biological effects, including dermonecrosis and complement-dependent haemolysis. The aim of this study was to generate recombinant proteins from the Loxosceles spider gland to facilitate structural and functional studies in the mechanisms of loxoscelism. Using "Expressed Sequencing Tag" strategy of aleatory clones from, L. laeta venom gland cDNA library we have identified clones containing inserts coding for proteins with significant similarity with previously obtained N-terminus of sphingomyelinases from Loxosceles intermedia venom [1]. Clone H17 was expressed as a fusion protein containing a 6x His-tag at its N-terminus and yielded a 33kDa protein. The recombinant protein was endowed with all biological properties ascribed to the whole L. laeta venom and sphingomyelinases from L. intermedia, including dermonecrotic and complement-dependent haemolytic activities. Antiserum raised against the recombinant protein recognised a 32-kDa protein in crude L. laeta venom and was able to block the dermonecrotic reaction caused by whole L. laeta venom. This study demonstrates conclusively that the sphingomyelinase activity in the whole venom is responsible for the major pathological effects of Loxosceles spider envenomation.

Mechanism of induction of complement susceptibility of erythrocytes by spider and bacterial sphingomyelinases

We have recently shown that the sphingomyelinase toxins P1 and P2 from the venom of the spider Loxosceles intermedia induce complement (C)-dependent lysis of autologous erythrocytes by induction of the cleavage of cell surface glycophorins through activation of an endogenous metalloproteinase facilitating the activation of the alternative pathway of C. Phospholipase D (PLD) from Corynebacterium pseudotuberculosis shows some degree of homology with the spider sphingomyelinases and can induce similar clinical symptoms to those observed after spider envenomation. The aim of this study was to investigate if the bacterial PLD-induced haemolysis of human erythrocytes was C dependent and if cleavage of glycophorins occurred. We show here that haemolysis of both PLD- and P1-treated human erythrocytes was C dependent, but while PLD-mediated haemolysis was dependent on activation of the classical pathway of C, P1 induced lysis via both the classical and alternative pathways. P1, but not PLD, induced cleavage of glycophorins and no change in expression of complement regulators was induced by either of the toxins. In both cases, annexin V binding sites were exposed, suggesting that the membrane asymmetry had been disturbed causing exposure of phosphatidylserine to the cell surface. Our results suggest that C susceptibility induced by L. intermedia and C. pseudotuberculosis PLD is a result of exposure of phosphatidylserine, and the higher potency of P1 toxin can be explained by its additional effect of cleavage of glycophorins.

Recombinant forms of Gerbich blood group antigens: expression and purification

Recombinant forms of normal glycophorin C (GPC), carrying the high frequency Gerbich blood group antigens, and its natural deletion mutants of Yus and Ge type (all combined with oligohistidyl tag) were expressed in CHO and COS 7 cells. The stable expression of all recombinant forms of GPC in CHO cells was obtained, but the level of expression was low and detectable only by flow cytometry. The high level of transient expression of GPC recombinant forms in COS 7 cells allowed their purification on Ni-NTA-agarose. The purified recombinant GPC and mutants of Yus and Ge type behaved in SDS-PAGE similarly to normal GPC forms from RBC membranes. The recombinant GPC.Yus and GPC.Ge mutants appeared as diffuse bands, suggesting the similar heterogeneity of glycosylation that was observed in natural GPC.Yus and GPC.Ge glycoproteins. The flow cytometry analysis of the transfected CHO and COS 7 cells showed that binding of anti-GPC monoclonal antibodies to GPC variants was accordant with the known fine specificity of these antibodies. The obtained recombinant forms of GPC carrying common Gerbich antigens may be useful in serology, and also as model molecules for structure-function studies.

Recombinant forms of glycophorin C as a tool for characterization of epitopes for new murine monoclonal antibodies with anti-glycophorin C specificity

Glycophorin C (GPC) and glycophorin D (GPD) are minor but important components of human RBC membranes. They carry the high-frequency antigens Ge2, Ge3 and Ge4 of the Gerbich blood group system. The epitopes for five new monoclonal antibodies (MoAbs) with anti-GPC specificity were characterized. Two antibodies (4G11 and 5B11) reacted with glycosylated N-terminal epitopes, and three reacted with internal epitopes of GPC. Pepscan analysis showed that the MoAb RB11 required for binding the EPDP sequence, occurring twice in GPC polypeptide chain. The MoAb 7F11 recognized the sequence 13PLSLEPDP20, and the MoAb RB8 did not react with synthetic peptides. Further characterization of the internal epitopes was performed in fluorescence-activated cell sorter (FACS) with the use of recombinant GPC and its variant forms transiently expressed on COS-7 cells. The results indicated that the MoAb RB11 recognized distinctly its target sequence EPDP only in a normal GPC molecule. The reactivity of the MoAb 7F11 with the PLSLEPDP sequence was confirmed and found to be enhanced by the O-glycan at the Ser15 residue. The MoAb RB8 recognized the glycopeptidic epitope in proximity to the Ser15 residue, requiring the presence of O-glycan. The combination of immunochemical techniques with the use of the recombinant forms of GPC has made it possible to define the role of sugar chains in the recognition of peptidic epitopes in glycosylated antigen and sheds new light on the Gerbich system antigens.

Antigenic Properties of Human Glycophorins - An Update

Glycophorins are complex heavily glycosylated antigens carrying peptidic and glycopeptidic epitopes. Detailed immunochemical studies showed that GPA/GPB and GPC/GPD molecules have defined sites which are particularly immunogenic. These sites include N-terminal portions of all glycophorins, internal fragments of their extracellular domains, and cytoplasmic tails. The extracellular epitopes involve directly oligosaccharide chains (e.g. blood group M- and N-related epitopes, or N-terminal epitopes of GPC) or have peptidic character, shown by the reaction of respective antibodies with synthetic peptides. Peptidic eitopes are independent of glycosylation, or are variably affected by adjacent O-glycans which may mask the epitopes or may be required for a proper exposure of an antibody binding site. Several low incidence epitopes are present on variant glycophorin molecules. Among anti-glycophorin antibodies there are the 'bispecific' ones, or antibodies recognizing an epitope formed by an interaction of two proteins (Wr(b)). Alltogether, the glycophorins serve as convenient model antigens for studying Ag-Ab interaction and a role of O-glycosylation in protein antigenic properties. Moreover, well defined specificty of monoclonal anti-glycophorin antibodies makes them more precise tools in serological investigation and identification of normal and variant antigens. Last but not least, elucidation of antigenic properties of glycophorins is important for identification and characterization of human anti-glycophorin antibodies, which in some cases create medical problems at transfusion or pregnancy.

Loxosceles intermedia spider envenomation induces activation of an endogenous metalloproteinase, resulting in cleavage of glycophorins from the erythrocyte surface and facilitating complement-mediated lysis

Loxosceles is the most venomous spider in Brazil, and envenomation causes dermonecrosis and complement (C)-dependent intravascular hemolysis. The authors studied the mechanism of induction of C-induced hemolysis. Purified Loxosceles toxins rendered human erythrocytes susceptible to lysis by human C but did not have an effect on the E-bound C-regulators DAF, CR1, or CD59. However, incubation with venom toxins caused cleavage of glycophorin from the erythrocyte (E) surface, facilitating C activation and hemolysis. The results suggest that glycophorin is an important factor in the protection of E against homologous C. Cleavage of glycophorin (GP) A, GPB, and GPC occurred at sites close to the membrane but could not be accomplished using purified GPA and purified toxins, demonstrating that cleavage was not an effect of a direct proteolytic action of theLoxosceles toxins on the glycophorins. Inhibition of the cleavage of glycophorins induced by Loxosceles venom was achieved with 1,10-phenanthroline. The authors propose that the sphingomyelinase activity of the toxins induces activation of an endogenous metalloproteinase, which then cleaves glycophorins. They observed the transfer of C-dependent hemolysis to other cells, suggesting that the Loxosceles toxins can act on multiple cells. This observation can explain the extent of hemolysis observed in patients after envenomation. Identification of the mechanism of induction of susceptibility to C-mediated lysis afterLoxosceles envenomation opens up the possibility of the development of an effective therapeutic strategy.

Anti-human red cell monoclonal antibodies produced by macaque-mouse heterohybridomas: their reactivity with human and nonhuman primate erythrocytes

Eighteen monoclonal antibodies (Mabs) against human red blood cells (RBCs) produced by macaque mouse heterobybridomas gave uniformly positive reactions with all human samples except for some with particular null phenotypes. Based on reactions with latter cells, the human antigenic targets of 11 antibodies could be identified: six were specific for glycophorin-related antigens (Wr(b), En(a), Ge4), and each of the live remaining antibodies showed one of the following specificities: CD55, CD44, CD59, Kell, and Rh proteins. Four Mabs recognized the Vc antigen of the chimpanzee V-A-B-D system. Six macaque Mabs detected polymorphisms in chimpanzee, gorilla, orangutan, and gibbon that did not correspond to any known blood group in these animals, while other Mabs gave monomorphic reactions with ape RBCs. The reagents produced by macaque hybridomas are useful tools not only for human blood grouping tests, but also for tracing the relationships among blood group antigens of man and anthropoid apes.

Fine specificities of murine anti-Mg monoclonal antibodies

The specificities of two murine anti-Mg monoclonal IgG1 antibodies, 3B10 and 2D5, were determined by pepscan analysis. The peptides which correspond to various fragments of amino-terminal portions of glycophorin A of group M (GPA-M), N (GPA-N) and Mg (GPA-Mg), and replacement analogues of some of these peptides, were synthesized on plastic pins and tested for binding of the antibodies. Both antibodies bound strongly to the N-terminal Mg octapeptide 1LSTNEVAM8, but they showed different subspecificities. The essential fragment of the epitope 2D5 are amino acid residues 2STNEV6. Replacement of any of these amino acid residues by Ala, and replacement of Glu5 residue by Gly, abolished or strongly reduced the antibody binding, but replacement of Asn4 by Thr gave only a moderate decrease of peptide activity. In contrast, the Leu1 and Asn4 residues were most essential components of the epitope 3B10, while Ser2, Thr3 and Glu5 seemed to be less important. Our present results and earlier ones on the specificity of human anti-Mg alloantibodies and monoclonal anti-M/Mg antibodies showed that antibodies reacting with Mg antigen recognize different fragments and/or different amino acid residues of the amino- terminal nonglycosylated domain of GPA-Mg. The knowledge of fine specificities of antibodies reacting with Mg antigen is interesting in view of the presence of anti-Mg alloantibodies in 1-2% of human sera.