Studies on the membrane glycoprotein defect of En(a-) erythrocytes. I. Biochemical aspects

Complete Deficiency of Glycophorin A in Red Blood Cells From Mice With Targeted Inactivation of the Band 3 (AE1) Gene

Glycophorin A is the major transmembrane sialoglycoprotein of red blood cells. It has been shown to contribute to the expression of the MN and Wright blood group antigens, to act as a receptor for the malaria parasite Plasmodium falciparum and Sendai virus, and along with the anion transporter, band 3, may contribute to the mechanical properties of the red blood cell membrane. Several lines of evidence suggest a close interaction between glycophorin A and band 3 during their biosynthesis. Recently, we have generated mice where the band 3 expression was completely eliminated by selective inactivation of the AE1 anion exchanger gene, thus allowing us to study the effect of band 3 on the expression of red blood cell membrane proteins. In this report, we show that the band 3 −/− red blood cells contain protein 4.1, adducin, dematin, p55, and glycophorin C. In contrast, the band 3 −/− red blood cells are completely devoid of glycophorin A (GPA), as assessed by Western blot and immunocytochemistry techniques, whereas the polymerase chain reaction (PCR) confirmed the presence of GPA mRNA. Pulse-label and pulse-chase experiments show that GPA is not incorporated in the membrane and is rapidly degraded in the cytoplasm. Based on these findings and other published evidence, we propose that band 3 plays a chaperone-like role, which is necessary for the recruitment of GPA to the red blood cell plasma membrane.

Complete Deficiency of Glycophorin A in Red Blood Cells From Mice With Targeted Inactivation of the Band 3 (AE1) Gene

Glycophorin A is the major transmembrane sialoglycoprotein of red blood cells. It has been shown to contribute to the expression of the MN and Wright blood group antigens, to act as a receptor for the malaria parasite Plasmodium falciparum and Sendai virus, and along with the anion transporter, band 3, may contribute to the mechanical properties of the red blood cell membrane. Several lines of evidence suggest a close interaction between glycophorin A and band 3 during their biosynthesis. Recently, we have generated mice where the band 3 expression was completely eliminated by selective inactivation of the AE1 anion exchanger gene, thus allowing us to study the effect of band 3 on the expression of red blood cell membrane proteins. In this report, we show that the band 3 −/− red blood cells contain protein 4.1, adducin, dematin, p55, and glycophorin C. In contrast, the band 3 −/− red blood cells are completely devoid of glycophorin A (GPA), as assessed by Western blot and immunocytochemistry techniques, whereas the polymerase chain reaction (PCR) confirmed the presence of GPA mRNA. Pulse-label and pulse-chase experiments show that GPA is not incorporated in the membrane and is rapidly degraded in the cytoplasm. Based on these findings and other published evidence, we propose that band 3 plays a chaperone-like role, which is necessary for the recruitment of GPA to the red blood cell plasma membrane.

Role of N-glycosylation in the expression of human band 3-mediated anion transport

The human erythrocyte anion transporter (band 3; AE1) has a single N-linked glycosylation site at amino residue Asn-642. To investigate the functional role of the N-glycan in band 3 (b3) we have constructed mutant b3 cDNAs in which this residue has been replaced by Gly, Ser or Thr, and the expression of these mutants was examined in Xenopus oocytes. Chymotrypsin treatment of intact oocytes was used to assess surface b3. Similar amounts of cleavage were observed with both glycosylated and unglycosylated b3. Greater cleavage of b3 was obtained when human red cell glycophorin A (GPA) was co-expressed with either glycosylated or unglycosylated b3. The co-expression of GPA with either glycosylated or unglycosylated b3 increased the stilbene disulphonate-sensitive chloride transport into oocytes at low cRNA concentrations. In both the presence or absence of GPA, a higher b3-mediated chloride influx into oocytes was observed on expression of glycosylated b3 cRNA compared with similar amounts of unglycosylated b3 cRNA. We suggest that glycosylation is not essential for the expression of functional b3 in oocytes, but may play a role in enabling the protein to acquire its correct folding with the highest anion transport activity.

A mouse monoclonal antibody against glycophorin A produced by in vitro stimulation with human red cell membranes

Human erythrocyte membranes were used as antigen for production of mouse monoclonal antibodies against blood group related structures by in vitro immunization. Culture medium supernatant of PHA and PMA stimulated mouse thymus cells was used as source of cytokines. The selected antibody designated 124,D-7 (isotype IgM) was found to directly agglutinate all human red cells, except the rare erythrocytes En(a-) which lack glycophorin A. Immunoblotting showed faint bands in the positions of glycophorin A, whereas no binding occurred to glycophorin B. Inhibition of agglutination with purified glycophorin A and peptides suggests that the epitope is located within the amino acid residues 35-40. Rat and chicken erythrocytes also reacted with the antibody, whereas mouse erythrocytes were only agglutinated at very low dilutions of ascitic fluid.

SAT, a 'new' low frequency blood group antigen, which may be associated with two different MNS variants

A new private blood group antigen, SAT, was identified in an NFLD-Japanese woman as a result of testing 10,480 blood donors with a serum containing anti-NFLD and anti-SAT. Three other sera were subsequently also shown to contain anti-SAT. The donor's family showed that SAT is inherited as a dominant character and may be associated with a weak M antigen. Serological and immunochemical analysis revealed no other aberrations in the MNS system. Study of a second SAT+Japanese blood donor and his family suggested that SAT is associated with an unusual MNS variant resulting from a hybrid glycophorin comprising the N-terminus of glycophorin A and the C-terminus of glycophorin B. The propositus appears to be homozygous for the gene that produces the putative hybrid, which differs from previously described glycophorin (A-B) hybrids by expressing no S, s or U antigen. SAT antigen, therefore, may be associated with two different MNS variants in the only two families in which it has been identified.

Blood group-active surface molecules of the human red blood cell

The surface of the human red blood cell is dominated by a small number of abundant blood group active proteins. The major proteins are the anion transport protein (band 3) which has AB(H) activity, and Glycophorin A which has MN activity. Band 3 and Glycophorin A are of equal abundance in the normal red cell membrane (approximately 10(6) copies of each) and the two proteins may associate together as a complex. The glucose transporter (band 4.5) had AB(H) activity and there are about 5 x 10(5) copies/red cell. Several polypeptides associate together to form the Rh complex. The major components of this complex (abundance 1-2 x 10(5) copies/red cell) are polypeptides of Mr 30,000, polypeptides of Mr 45,000-100,000 and Glycophorin B. The antigens of the Rh blood group system appear to be associated with the polypeptides of Mr 30,000 and those of Mr 45,000-100,000 (the latter also express AB(H) activity). Glycophorin B expresses the blood group 'N' antigen and the Ss antigens. Glycophorins C and D carry the Gerbich antigens and, together, these polypeptides comprise approximately 10(5) copies/red cell. The complete protein sequence of all the above-mentioned proteins is known, except for the Mr 30,000 and Mr 45,000-100,000 polypeptides of the Rh complex for which only partial sequences are available, and Glycophorin D, the sequence of which can be inferred from that of Glycophorin C. Several of the minor blood group active proteins at the red cell surface (abundance less than 1.2 x 10(4)/red cell) have been the subject of recent studies. The polypeptide expressing Cromer-related blood group antigens has been identified as decay-accelerating factor and that carrying the Ina/Inb antigens as CD44. The protein sequence of both of these proteins has been deduced form nucleotide sequencing. The polypeptides expressing Kell antigens, Lutheran antigens, Fy antigens, and LW antigens have also been identified and partially characterised.

Alteration of the genes for glycophorin A and B in glycophorin-A-deficient individuals

Glycophorins A and B are homologous glycoproteins of the red cell membrane which carry the blood-group MN and Ss antigens, respectively, and are encoded by two distinct genes closely linked on chromosome 4, which are probably derived from each other by duplication during evolution. The lack of glycophorin A is associated with the rare phenotype En(a-), indicating individuals who are defective for MN antigens, as well as for the Ena antigens, also located on this glycoprotein. The En(a-) condition is heterogenous and includes two categories of variants exemplified by the Finnish and the English types referred to as En(Fin) and En(UK), respectively. By Southern blot and preliminary genomic clone analyzes we have compared the status of the genes for glycophorins A and B, as well as that of the gene encoding glycophorin C, another unrelated red cell membrane glycoprotein, in the En(a-) variants and in the En(a+) control donors. Our data indicate that the En(Fin) variant is homozygous for a complete deletion of the glycophorin A gene without any detectable abnormality of the genes encoding glycophorins B or C. In the genome of the En(UK) variant, with the presumed genotype Mk/En(UK), and where the Mk condition abolishes the expression of MN and Ss antigens, we have identified several abnormalities of the glycophorin A and B genes, but the glycophorin C gene was unaffected. Our results strongly support the view that in Mk chromosome the glycophorin A and B genes are largely deleted, whereas the En(UK) chromosome probably contains a gene fusion product encoding a hybrid glycoprotein AM-B, composed of the N-terminal portion of a blood group M-type glycophorin A and of the C-terminal portion of glycophorin B. The determination of the 5' and 3' limits of the hybrid gene and elucidation of the mechanism involved will require sequencing of the rearranged DNA of the variant and a full knowledge of the organization of the glycophorin A and B genes.

Major O-glycosylated sialoglycoproteins of human hematopoietic cells: differentiation antigens with poorly understood functions

All human hematopoietic cells seem to contain a major, heavily O-glycosylated sialoglycoprotein. Glycophorin A is specific for the erythroid lineage of cells, and leukocytes have a major sialoglycoprotein, also called leukosialin or sialophorin. Cell differentiation results in patterns of O-glycosylation in these proteins, which reflect the stage of differentiation within a cell lineage as well as lineage specificity. The altered carbohydrate compositions may influence the interactions of the cells with external ligands. Healthy individuals lacking glycophorin A in their red cells are known, whereas a deficiency of the leukocyte sialoglycoprotein may result in immunological disease. Although little is known about the physiological functions of these proteins, they form interesting models for studies on regulation of glycosylation, biosynthesis of O-glycosylated glycoproteins, and function of cell surface receptors.

Characterization of cDNA clones for human glycophorin A. Use for gene localization and for analysis of normal of glycophorin-A-deficient (Finnish type) genomic DNA

Glycophorin A is the major membrane sialoglycoprotein of human erythrocytes and represents a typical example of a transmembrane glycoprotein. The functional role of this cell-surface component is not known but it represents a receptor for viruses, bacteria and parasites like Plasmodium falciparum. 1. Two cDNA clones encoding glycophorin A have been characterized from human fetal cDNA libraries. The longer cDNA extended from the coding region of glycophorin A (residues 4-131) to the 3' untranslated region which included two polyadenylation signals and a poly(A) tail. 2. The structural gene for glycophorin A is located on chromosome 4, q28-q31 as shown by in situ hybridization, thus confirming the previous localization by genetic linkage analysis. 3. Three distinct mRNA species (1.0 kb, 1.7 kb and 2.2 kb) have been identified in erythroid spleen. Northern blot analyses with a probe directed against the 3' untranslated region of the mRNAs indicated that all these species share a homologous 3' non-coding region and that the first polyadenylation signal downstream the stop codon is not used. 4. Preliminary studies by Southern blot analysis of the genomic DNA from normal En(a+) and rare En(a-) donors suggest that the glycophorin A gene has a complex organization and is largely deleted in donors of the En(a-) phenotype (Finnish type) who lack glycophorin A on their red cells.

The Dantu erythrocyte phenotype of the NE variety. I. Dodecylsulfate polyacrylamide gel electrophoretic studies

Red cell membranes from patient NE, Mr. Dantu and 16 additional Black individuals, positive for the low-frequency MNSs-system antigen Dantu, were studied by dodecylsulfate polyacrylamide gel electrophoretic techniques. The content of the major, blood group M- or N-active sialoglycoprotein (glycophorin A, GP A) was found to be decreased by about 57%. The blood group S- or s-active sialoglycoprotein (GP B) was decreased by about 51% in membranes from proven Dantu/U heterozygotes and not detectable in those from patient NE and other Dantu+U- individuals. Donor NE was shown to exhibit the genotype Dantu/u. Dantu-positive cells exhibit a proteinase-resistant GP B-GP A hybrid with an apparent molecular mass of 29 KDa whose intramembraneous and cytoplasmic domains were shown to be similar to those of GP A. The molar hybrid: GP A ratio in all cells was found to be about 2.4: 1, indicating that the NE variety of the Dantu phenotype is much more frequent than the Ph or MD types. The significance of an additional minor 'new' component (molecular mass 21 KDa) in Dantu+ membranes and the minor component J (molecular mass 22 KDa) occurring in normal and Dantu+U+ GP preparations, but not in those from Dantu+U- cells, has not been resolved. The apparent molecular mass of the anion channel protein (band 3) in all cells of the NE variety was shown to be decreased by about 3 KDa, due to a shortening of carbohydrate chains. This suggests that the hybrid, just like GP A, might form a complex with band 3.

The Dantu erythrocyte phenotype of the NE variety. II. Serology, immunochemistry, genetics, and frequency

Red cells (RBC) possessing the low-frequency MNSs antigen Dantu from 36 Black individuals (21 propositi) were found to exhibit the NE variety of this phenotype, as judged from the electrophoretic glycophorin (GP) pattern, described in an accompanying article, and/or from the polybrene test which detects the decreased NeuAc level of these RBC. All known DantuNE RBC (53) exhibit the phenotype M+N+. This finding as well as family studies and immunochemical investigations demonstrate that the DantuNE allele encodes a blood group M-specific GP A. Thus, the strongly decreased GP A level of RBC from DantuNE heterozygotes represents the product of the Dantu allele and its normal counterpart. It is suggested that the formation of a complex with the anion channel protein (band 3) represents the prerequisite for optimum incorporation of GP A into normal RBC membranes. The hybrid GP in DantuNE RBC, produced in large quantity, might suppress the incorporation of GP A in a cis and trans manner via the formation of a complex with band 3. The hybrid GP in DantuNE RBC lacks U activity, but expresses N activity and a qualitatively altered s antigen, thus proving its GP B-GP A hybrid nature in conjunction with data described in the accompanying article. Screening of ficin-treated RBC with Vicia lectin revealed that the Dantu phenotype exhibits a frequency of about 0.005 in American Blacks and less than 0.001 in Germans.

Finnish En(a-) propositus with anti-EnaFS and anti-EnaFR: in vitro and in vivo characteristics

A third Finnish En(a-) individual (ERP) with alloanti-Ena has been identified. Subsequent family studies revealed that ERP is distantly related to the 2 previous Finnish En(a-) propositi. Serologically, ERP's erythrocytes were found to be M-N-'N'+S-s+U+En(a-)Wr(a-b-) and were sialic acid deficient. SDS-PAGE studies confirmed that the red cell membranes lacked the MN-sialoglycoprotein, glycophorin-A. ERP's serum contained an IgG antibody which demonstrated two separable specificities, a ficin-sensitive specificity (anti-EnFS) and a ficin-resistant specificity (anti-EnFR), which reacted by the indirect antiglobulin technique. The antibodies were probably pregnancy induced, as ERP had never been transfused. A 51Cr-labelled red cell survival study showed that the antibodies were capable of causing significant destruction of incompatible red cells.

Two examples of low ionic strength-dependent autoagglutinins with anti-Pr1 specificity

Two low ionic strength-dependent autoagglutinins were studied and found to have anti-Pr1 specificity. This specificity was determined by studies with enzyme-treated and neuraminidase-treated human red blood cells (RBCs), animal RBCs and chemically-modified sialoglycoproteins, all suspended in a low ionic strength solution (LISS). Both IgM complement-binding cold agglutinins had a wide thermal range and caused in vitro hemolysis of some LISS-suspended RBCs at 37 degrees C. Compatible blood was found for these patients by using techniques that did not employ LISS.

Erythrocyte membrane rigidity induced by glycophorin A-ligand interaction. Evidence for a ligand-induced association between glycophorin A and skeletal proteins

Erythrocyte skeletal proteins are known to play an important role in determining membrane deformability. In order to see whether transmembrane proteins also influence deformability and, if so, whether this influence is mediated by an interaction with the membrane skeleton, we examined the effect on deformability of ligands specific for transmembrane proteins. We found membrane deformability markedly reduced in erythrocytes that were pretreated with glycophorin A-specific ligands. In contrast, ligands specific for band 3 and A and B blood group antigens had no effect. The increase in membrane rigidity appeared to depend upon a transmembrane event and not upon a rigidity-inducing lattice on the outside surface of the cell in that a monovalent Fab of antiglycophorin IgG caused decreased deformability. We therefore looked for a ligand-induced association of glycophorin and the skeletal proteins and found, in Triton X-100-insoluble residues, a partitioning of glycophorin with the skeletal proteins only after preincubation with a ligand specific for glycophorin. We then studied cells and resealed membranes with skeletal protein abnormalities. In spectrin-deficient and protein 4.1-deficient erythrocytes and in 2,3-diphosphoglycerate-treated resealed membranes, the antiglycophorin IgG was only one-third as effective in decreasing deformability as it was in normal cells. Thus, normal skeletal proteins appear to be essential for liganded glycophorin to affect membrane deformability maximally. Taken together, these observations indicate that there is a ligand-induced interaction between glycophorin A and skeletal proteins and that this interaction can directly influence membrane deformability.

Glycophorin A on normal and leukemia cells detected by monoclonal antibodies, including a new monoclonal antibody reactive with glycophorins A and B

A new hemagglutinating monoclonal antibody, MoAb31, detected glycophorins A and B in Western blots. Results with enzyme-modified erythrocytes indicated the MoAb31 determinants were sialic acid dependent, and resided on glycophorin A on the trypsin-resistant, ficin-sensitive segment, and on glycophorin B on the ficin-sensitive segment. Another new monoclonal antibody, MoAb36, detected the Wrb antigen, located on the non-glycosylated segment of glycophorin A near its insertion into the lipid bilayer. Immunofluorescent staining of normal hematopoietic and leukemia cells with these and other monoclonal antibodies to glycophorin A demonstrated glycophorin A on erythroid cells only. Cytofluorograph analysis showed the majority of cells of the erythroleukemia cell lines K562 and HEL expressed glycophorin A, as indicated by reactivity with the monoclonal glycophorin A antibodies R10, R18, 6A7 and 10F7. However, reactivity with monoclonal antibodies to glycosylated determinants (MoAb31 and R1.3) and to the non-glycosylated segment near the membrane insertion (MoAb36, and R7.1) was reduced or absent. Expression of "missing" glycophorin A antigens on K562 and HEL could not be induced using a variety of chemical and biologically active modifiers. We conclude that glycophorin A of erythroleukemia cell lines K562 and HEL differs from glycophorin A at the surface of normal, mature erythrocytes with respect to reactivity with monoclonal glycophorin A antibodies.

Altered membrane sialoglycoproteins in human erythrocytes lacking the Gerbich blood group antigens

The sialoglycoproteins (glycophorins) in human red cell membranes of rare individuals lacking totally (Ge-1,-2,-3 phenotype) or partially (Ge-1,-2,3 phenotype) the Gerbich (Ge) blood group antigens and two Ge-1,-2,-3 heterozygotes were studied by dodecylsulfate polyacrylamide gel electrophoretic techniques. Two sialoglycoproteins (components D and E) were not detectable in the membranes from the homozygotes and found to be decreased by about 50% in those from the heterozygotes. Ge--1,-2,-3 and Ge-1,-2,3 cells were found to contain a 'new' component (mol. masses about 29 and 30 kDa, respectively) possibly representing a D/E hybrid molecule. This sialoglycoprotein was not detectable in membranes from the Ge-1,-2,-3 heterozygotes, suggesting that the Ge-1,-2,-3 phenotype may be caused by at least two different alleles at the Ge blood group antigen locus. Hemagglutination or hemagglutination inhibition tests involving anti-Ge 1,2,3 and -Ge 1,2 as well as native and enzyme-treated normal red cells (phenotype Ge 1,2,3) or membrane and sialoglycoprotein fractions from normal erythrocytes indicate that the receptors of these sera are located within the glycosylated domain(s) of the D and/or E sialoglycoprotein(s). Our data suggest that the Ge locus encodes the polypeptide sequences of the D and E sialoglycoproteins.

High frequency antigens of human erythrocyte membrane sialoglycoproteins. I. Ena receptors in the glycosylated domain of the MN sialoglycoprotein

The specificity of various allo- and autoantibodies, which agglutinate normal erythrocytes, but do not react with En(a-) red cells and normal erythrocytes, treated with trypsin (anti-EnaTS) or ficin (anti-EnaFS), was investigated. Various fragments and modification products of the major (MN) red cell membranes sialoglycoprotein were used in hemagglutination inhibition assays. Six anti-EnaFS sera were found to be directed against the residues approx. 46-56 of the molecule. Five of these require the carbohydrate unit, attached to Thr50, for binding. One anti-EnaTS serum was found to be directed against the residues approx. 36-42. Another antibody with anti-EnaTS specificity was shown to react with the residues 31-39 in some of the MN sialoglycoprotein molecules, namely those not glycosylated at a certain position (probably Thr33). A third anti-EnaTS serum, directed against the sequence domain around Lys30, was also found to react only with a fraction of the molecules, apparently due to the variable attachment of oligosaccharides in that region. The heterogeneity of glycosylation, detected by these two sera, appears to account for the partial tryptic and chymotryptic cleavage in this domain of the MN sialoglycoprotein, which has been described previously. Heterogeneity of the glycosylation at various positions of the molecule could be established by the isolation and analysis of peptides.

Differentiation of human erythroid cells is associated with increased O-glycosylation of the major sialoglycoprotein, glycophorin A

Glycophorin A, the major human erythrocyte sialoglycoprotein, is found exclusively on cells of the erythroid lineage. The amino acid sequence is known, and glycophorin A isolated from mature erythrocytes contains a single N-glycosidic and 15 O-glycosidic oligosaccharides. Monoclonal antibodies against erythrocyte glycophorin A reacted weakly with erythroid precursors while a monospecific rabbit antiserum reacted strongly with immature and mature red cells. Glycophorin A was isolated from cells representing various stages of erythropoiesis in normal bone marrow, from blood cells of neonates with erythroblastosis fetalis, and from the erythroleukemic cell lines K562 and HEL before and after induced differentiation. Analysis of the oligosaccharides showed less O-glycosylation of glycophorin A in erythroid precursors. The degree of glycosylation increased concomitantly with differentiation.

Mouse monoclonal anti-N. I. Production and serological characterization

Two examples of mouse monoclonal anti-N are described. The antibodies were derived from mice immunized with sialoglycoprotein extracts of group O MM ss erythrocyte membranes and the probable stimulus for immunization was glycophorin B associated N-antigen. Both antibodies reacted as direct agglutinins but appeared to recognize different epitopes with one having a greater dependence on sialic acid. The antibodies could prove to be valuable alternatives to those reagents used currently for N-blood grouping.

Immunofluorescent detection of erythrocyte sialoglycoprotein antigens on murine erythroid cells

A sialoglycoprotein fraction isolated from murine (DBA/2) erythrocytic ghosts (see companion article, Sarris and Palade, 1982, J. Cell. Biol. 93:583-590) was used to raise antibodies in rabbits. By immune-IgG (serum)-[125I] protein A overlays, the antibodies were found to react positively with the four sialoglycoprotein monomers (gp-2.1, gp-2.2, gp-3.1, and gp-3.2) of the original fraction, with the sialoglycoproteins detected in erythrocytic ghosts (gp-2.1 and gp-3.1), with a diffuse component (probably a macroglycolipid) trailing around gp-3.1 in SDS polyacrylamide gel electrophoretograms of solubilized ghosts, and with a minor sialoglycoprotein hidden under this trail. IgG's isolated from immune and nonimmune rabbit sera were conjugated to tetramethylrhodamine isothiocyanate and used to survey, by fluorescence microscopy, the distribution of the cognate antigens on the three different erythroid lines known to succeed each other during the life span of the mouse. In the peripheral blood of the adult, the antibodies recognized only mature erythrocytes; they did not crossreact with either platelets, monocytes, or different types of granulocytes. In the spleen of adult anemic mice, the antibodies reacted weakly with proerythroblasts and strongly with all types of erythroblasts. In enucleating erythroblasts, antigens were preferentially segregated on the cell membrane of the nascent reticulocyte. In the 10-day-old embryo, antigens were already present on the primitive nucleated erythrocytes (produced by the blood islets of the yolk sack), and in the 14-d fetus they were found on all hepatic erythroblasts and derived non-nucleated erythrocytes. A positive immunoreaction was also obtained on Friend erythroleukemic cells, before or after induction by dimethyl sulfoxide. Nonimmune serum, or nonimmune IgGs gave negative reactions in all cases. The antibodies were species-specific: they did not crossreact with either human or rat erythrocytes.

Monoclonal antibodies to human erythrocytes

Eight monoclonal antibodies from mouse hybridomas raised to normal human erythrocytes were tested with a panel of null-type erythrocytes, enzyme-treated normal cells, and by inhibition with human erythrocyte sialoglycoproteins. Two antibodies reacted poorly or not at all with RhNULL cells. These antibodies are of considerable interest since it may be possible to use them to elucidate the chemical nature of the antigens of the Rhesus blood group system. Four other antibodies were inhibited by sialoglycoprotein preparations. The antigens recognized were, respectively, two different determinants on the major sialoglycoprotein alpha (glycophorin A) and one determinant which is probably common to sialoglycoproteins alpha and delta (glycophorins A and B). Another antibody had anti-Wrb specificity. One of these antibodies is of considerable potential value for the further characterization of erythrocyte sialoglycoproteins.

Role of sialic acid in the mobility of membrane proteins containing O-linked oligosaccharides on polyacrylamide gel electrophoresis in sodium dodecyl sulfate

The major sialoglycoprotein of the human red-cell membrane, glycophorin A, contains 15 O-glycosidically linked oligosaccharides and one N-glycosidic oligosaccharide. The protein shows a decreased mobility on polyacrylamide gel electrophoresis in sodium dodecyl sulfate after neuraminidase treatment of the non-denatured protein. The molecular mechanism behind this phenomenon has been elucidated. Neuraminidase treatment of glycophorin A in intact cells or after solubilization in buffers containing Triton X-100 resulted in conversion of the predominant tetrasaccharide N-acetylneuraminosyl alpha 2-3galactosyl beta 1-3(N-acetylneuraminosyl alpha 2-6)-N-acetylgalactosamine to the trisaccharide galactosyl beta 1-3(N-acetylneuraminosyl alpha 2-6)-N-acetylgalactosamine and the disaccharide galactosyl beta 1-3-N-acetylgalactosamine. After denaturation with sodium dodecyl sulfate, Vibrio cholerae neuraminidase also liberated the N-acetylgalactosamine-bound sialic acids. Such treatment resulted in increased electrophoretic mobility. The results show that distal sialic acids linked to galactose are readily available to neuraminidase, and that their negative charge gives an increased electrophoretic mobility in polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. In contrast, most of the N-acetylgalactosamine-linked sialic acids of glycophorin A are not liberated by neuraminidase without denaturation of the substrate. Like sialic acids of complex-type oligosaccharides the decreased electrophoretic mobility caused by them is exclusively due to their mass while no significant contribution by the charge was seen.

Organization of membrane lipids and proteins in human En(a-) erythrocytes that lack the major sialoglycoprotein, glycophorin A. A spin-label study

Membrane fluidity was studied by electron-spin-resonance techniques in human En(a-) erythrocytes that lack the major membrane sialoglycoprotein, glycophorin A. By using stearic acid spin labels with a doxyl group in the C-12 or C-15 positions, we demonstrated that the hydrophobic core in these cells was more fluid than in normal cells. Surface-located regions in isolated En(a-) membranes, when probed with stearic acid labelled in the C-5 position, appeared more stable than in normal membranes. In isolated En(a-) membranes, protein motion was decreased when probed with a nitroxide derivative of maleimide. After incubation with anti-(glycophorin A) antibodies protein motion and membrane fluidity were increased in normal membranes. This effect was observed also after spectrin depletion, which by itself increased protein motion but decreased membrane fluidity in the hydrophobic core of the membrane. The results show that membrane proteins influence the fluidity of membrane lipids.

Immunochemical properties of Mg erythrocytes

The major erythrocyte membrane (MN) sialoglycoprotein in Mg red cells was found to exhibit a slightly decreased sodium-dodecyl-sulphate polyacrylamide gel electrophoretic molecular weight and periodic and/Schiff staining intensity. Mg antigen activity was shown to be associated with this molecule. As judged from chemical modification experiments, no carbohydrate but the glycoprotein's N-terminal amino acid is involved in the Mg receptor site. The endgroup of the glycoprotein was found to leucine and studies involving Staphylococcus aureus V8 protease suggest that a glutamic acid is located at the fifth position of its peptide chain. This indicates that the Mgs gene complex evolved from a mutation of an Ns allele. An amino acido substitution or deletion at the second, third and/or fourth position(s), preventing the glycosylation of all or some of these amino acids, provides an explanation for the properties of Mg erythrocytes.

An auto-anti-Ena, inhibitable by MN sialoglycoprotein

An auto-anti-Ena, which reacts with trypsin, but not with ficin-treated red blood cells, and which can be totally inhibited with sialoglycoprotein (SGP) isolates from red blood cells, is described. From comparative studies on this antibody and on the four known examples of allo-anti-Ena, it is clear that the term "anti-Ena" describes a heterogeneous group of related but not identical specificities. The specificities contained within the auto-anti-Ena described are different from those within any of the sera containing allo-anti-Ena. Several of the specificities that have been included under the blanket term, anti-Ena, complex with the MN SGP of normal red blood cells, but recognize different portions of that polypeptide.

Rotational diffusion of band 3 proteins in membranes from En(a-) and neuraminidase-treated normal human erythrocytes

Recent experiments have demonstrated an association between band 3 and glycophorin A in the human eythrocyte membrane (Nigg, E.A., Bron, C., Girardet, M. and Cherry, R.J. (1980) Biochemistry 19, 1887-1893). Here, the influence of sialoglycoproteins on the rotational diffusion of band 3 in the human erythrocyte membrane was investigated by studying membranes from En(a-) and neuraminidase-treated erythrocytres. Rotational diffusion was measured by observing flash-induced transient dichroism of eosin-labeled band 3. Although erythrocytes of the rare phenotype En(a-) lack the major sialoglycoprotein, glycophorin A, no significant difference in band 3 rotation at pH 7.4 and 37 degrees C could be detected between En(a-) and normal erythrocyte membranes. Band 3 rotation at pH 7.4 was also insensitive to the enzymatic removal of sialic acid from the surface of normal erythrocytes. Moreover, the existence of an essentially similar temperature-dependent equilibrium between band 3 proteins with different mobilities was observed in normal, En(a-) and neuraminidase-treated erythrocytes. From these results it is concluded that glycophorin A contributes less than 15% to the cross-sectional diameter of the band 3-glycophorin A complex in the plane of the normal membrane. The rotation of the complex at pH 7.4 is not significantly influenced by either steric or electrostatic interactions involving the oligosaccharide moiety of glycophorin A.

Band 3-glycophorin A association in erythrocyte membrane demonstrated by combining protein diffusion measurements with antibody-induced cross-linking

A new approach to the study of molecular protein interactions in biological membranes is presented. The technique is based on measuring the rotation of a membrane protein in the presence and absence of specific antibodies directed toward a purported complex partner. As a first illustration of the method, the putative association of band 3 with glycophorin A in the human erythrocyte membrane was investigated. The rotational diffusion of band 3 was strongly reduced following cross-linking of glycophorin A with divalent antibodies. However, little or no effect on band 3 rotation was produced by monovalent antiglycophorin A Fab fragments, antispectrinor nonspecific antibodies, ruling out major effects on band 3 mobility due to steric hindrance, unspecific antibody adsorption, or transmembrane interactions involving spectrin. It is concluded that immobilization of band 3 by antiglycophorin A antibodies is directly caused by cross-linking of a preexisting band 3-glycophorin A complex in the human erythrocyte membrane.

Identification of blood group A-active glycoproteins in the human erythrocyte membrane

Normal human erythrocytes of blood groups A1, A2, B and O, and En (a-) erythrocytes lacking glycophorin A, but with A1B-activity, were surface-labeled with tritiated sodium borohydride after oxidation of terminal galactosyl and N-acetylgalactosaminyl residues with galactose oxidase. A1 cells were also labeled by lactoperoxidase catalyzed iodination. After solubilization in Triton X-100, the blood group A-active glycoconjugates were isolated using the A-specific lectin from Vicia cracca coupled to Sepharose. No radioactivity was bound from erythrocytes of B and O blood groups. The glycoconjugates from A cell membranes which bound to the lectin and were eluted with 0.01 M N-acetyl-D-galactosamine were analyzed using cylindrical or slab gel electrophoresis in the presence of sodium dodecyl sulfate. The A-active glycoproteins included the major integral glycoprotein, band 3, and many minor, previously poorly defined components. Glycophorins A and B did not contain A-activity.

Structure of the Ss blood group antigens. I. Isolation of Ss-active glycopeptides and differentiation of the antigens by modification of methionine

The Ss blood group antigen determinants were found to be associated with the N-terminal tryptic and chymotryptic glycopeptides (residues 1--35 or 1--32) of the Ss sialoglycoprotein from human erythrocyte membranes. The N-terminal portion (residues 1--26) of these peptides is largely identical with that of the MN sialoglycoprotein. Therefore, and since the Ss activity of tryptic glycopeptides was higher than that of chymotryptic fragments, it is concluded that the structural difference between the S and s antigens is located on the C-terminal part (residues 27--32) of these peptides. Chemical modification of sialoglycoproteins by various methods suggests that Glu residue(s) (positions 29 or 28, 31) and possibly alpha-GalNAc-Thr (residue 25) are recognized by anti-S and -s. Carboxymethylation, performic acid, hydrogen peroxide and cyanogen bromide treatment destroy the S antigen, but have no effect on the s receptor. This suggests that the S antigen is determined by a methionyl-residue.

Glycophorin A as a cell surface marker of early erythroid differentiation in acute leukemia

We show here that the leukemic blast cells from three patients of a total of 15 subsequently diagnosed as having acute leukemia express on their surface the major red cell sialoglycoprotein, glycophorin A (GP-A). This was demonstrated (1) by indirect immunofluorescence using rabbit anti GP-A anti-serum and (2) by immune precipitation of GP-A from surface radiolabelled leukemic cells. Since GP-A is exclusively present on erythroid cells and their precursors, these findings indicate that a higher proportion of the blast leukemias than previously recognized show features of early erythroid differentiation.

Biosynthesis of the major human red cell sialoglycoprotein, glycophorin A, in a continuous cell line

During biosynthesis of glycophorin A in K562 cells a precursor is rapidly transferred through the endoplasmic reticulum membrane with the COOH-terminal remaining in the cytoplasm. This is glycosylated within the cell and appears at the cell surface after about 30 min. The biosynthetic pathway resembles that described for viral membrane glycoproteins.

Phospholipid composition and external labeling of aminophospholipids of human En(a--) erythrocyte membranes which lack the major sialoglycorprotein (glycophorin A)

Erythrocytes of the rare human blood group En(a--) lack the major sialoglycoprotein, glycophorin A, and the cell population heterozygous for the En(a) antigen contain half the normal amount of glycophorin A. With such cells we have studied whether glycophorin A influences the phospholipid composition and the availability of aminophospholipids to external labeling reagents. We here demonstrate that the amounts of all phospholipids are closely similar in normal and variant membranes. However, using the amino-reactive reagent trinitrobenzenesulfonate, we show that phosphatidylethanolamine is more easily labeled in intact En(a--) cells as compared to normal cells, whereas phosphatidylethanolamine shows an intermediate labeling in En(a) heterozygous cells.

K562--a human erythroleukemic cell line

We have studied the surface membrane properties of the human leukemic cell line K562 which previously has been reported to represent an early stage of granulocyte maturation. The surface glycoprotein pattern of the K562 cells obtained after galactose oxidase-NaB[3H]4 labelling and slab gel electrophoresis shows striking similarities with that of normal erythrocytes but is completely different from the patterns of normal and malignant cells of various stages of the myeloblast to granulocyte differentiation. Moreover, the K562 cell expressed the major red cell sialoglycoprotein, glycophorin, on its surface as shown by immunofluorescence and by immunoprecipitation from labelled membrane preparations. As glycophorin is exclusively found on erythroid cells in human bone marrow we conclude that the K562 is a human erythroleukemic line.

Genetic variants involving the major membrane sialoglycoprotein of human erythrocytes. Studies on erythrocytes of type Mk, Miltenberger class V and Mg

1. Membranes from erythrocytes heterozygous for the Mk and Miltenberger Class V (Mi.V) condition and membranes from erythrocytes homozygous for the Mg condition were studied by polyacrylamide-gel electrophoresis by using the periodate/Schiff stain binding of radioiodinated lectins and labelling with lactoperoxidase. 2. Both the Mk and Mi.V conditions are associated with a decreased content of the major blood-group-MN-active sialoglycoprotein. 3. An unusual blood-group-M-active membrane component was found in Mi.V cells of appropriate genotype. No comparably component was found in Mk erythrocytes. 4. The Mg antigen appears to result from a modification of the MN-active sialoglycoprotein found in normal cells. Our results suggest that the Mg sialoglycoprotein contains fewer sialotetrasaccharides than does the normal sialglycoprotein. This may result from changes in the amino acid sequence of the protein. 5. The results are discussed in relation to differences in the antigenic properties of Mk, Mi.V and Mg cells and their possible influence on the structure of the surface of each of these cells.

Further studies on the membrane glycoprotein defects of S--s--and En(a--)-erythrocytes

Sodium dodecylsulfate polyacrylamide gel electrophoretic methods were applied to S--s--and En(a--) red cells as well as to erythrocytes from individuals being heterozygous for these defects. The results demonstrate more conclusively than previous data, that the glycosylated part of the MN glycoprotein is lakcing in En(a--) red cell membranes. S--s--U--erythrocytes either lack the Ss glycoprotein completely or contain a defective molecule which is devoid of the glycosylated part. Conversely, S--s--U+ cells exhibit small amounts of Ss glycoprotein which could only be detected when large amounts of extracted glycoproteins were separated. It is shown that this molecule possesses the 'N' antigenic determinant.

Structural properties of the human MN blood group antigen receptor sites

It is shown that the MN blood group antigen determinant of the major human erythrocyte membrane (MN) sialoglycoprotein is located on its N-terminal octaglycopeptide. The only analytically detectable difference between peptides from MM and NN cells are Ser/Leu and Gly/Glu polymorphisms at the first and fifth positions, respectively. Destruction of the antigens by removal of the N-terminal residues suggests that these amino acids represent a part of the receptor areas for various anti-M or -N reagents. Evidence is presented that the N-terminal structure of the Ss glycoprotein is identical with that of MN glycoprotein from NN red cells up to the fifth residue. This provides an explanation for the 'N' antigen on this molecule and direct support for the earlier proposal that the MNSs locus is represented by homologous genes.

Studies on the membrane glycoprotein defect of En(a-) erythrocytes. III. N-terminal amino acids of sialoglycoproteins from normal and En(a-) red cells

Sodium dodecylsulphate polyacrylamide gel electrophoretic methods and quantitative analyses of the N-terminal amino acids were applied to the sialoglycoprotein mixture and glycoprotein fractions from normal erythrocyte membranes, as well as preparations from red cells of individuals belonging to the English and Finnish En(a-) families. The data confirm the observation by alternative methods that SS cells exhibit a higher Ss glycoprotein content than ss erythrocytes. The results of end-group analyses suggest that the N-terminal amino acids serine and leucine represent the structures differentiating the MN and the 'M' and 'N' antigens on the MN and Ss glycoproteins respectively. Data from peptide sequence analyses confirm that the glycine/glutamic acid polymorphism at the fifth position of the MN glycoprotein's peptide chain is closely or absolutely linked with the serine/leucine polymorphism at its N-terminal position. As normal (EnaEna) red cells exhibiting 'M' antigenic properties have not been detected, the hypothesis is proposed that the Ss glycoprotein of English En(a-) erythrocytes possesses an MN-Ss hybrid polypeptide chain analogous to those of the delta-beta Lepore haemoglobins.