Separation of the M and N blood-group antigens of the human erythrocyte

Glycobiology of the cell surface: Its debt to cell electrophoresis 1940-65

This Review describes how in the period 1940-1959 cell electrophoresis (in the earlier literature often referred to as 'microelectrophoresis') was used to explore the surface chemistry of cells. Using the erythrocyte as a suitable model for the study of biological membranes, the early investigators were agreed on the presence of negatively charged groups at the surface of this cell. The contemporary dogma was that these were phosphate groups associated with phospholipids. Work in the 1960s, particularly on changes in the electrokinetic properties of erythrocytes following treatment with proteolytic enzymes, lead to the realization that the negatively charged groups at the red cell surface are predominantly due to sialic acids carried on glycoproteins. It quickly became apparent from cell electrophoresis that sialic acids have a ubiquitous presence on the surface of animal cells. This finding required that any complete model of the plasma membrane must include glycosylated molecules at the cell periphery, thus laying the foundations for the field termed 'Glycobiology of the Cell Surface'.

Identification and function of transmembrane glycoproteins--the red cell model

Transmembrane glycoproteins in the red cell membrane traverse the plasma membrane, have their carbohydrate moieties on the extracellular surface, are sialyated (except for band 3) and are tethered to the membrane cytoskeleton proteins on the cytoplasmic surface. This linkage between the transmembrane proteins and the skeletal membrane proteins provides a two-way communication between the extracellular surface and the interior of the red cell; i.e., a transmembrane effect can be initiated from either side. These interactions are discussed in this review, including the example of sickle cell anemia in which the membrane bound hemoglobin may exert a transmembrane effect to change the conformation or distribution of transmembrane glycoproteins and and hence the extracellular surface receptors. This, in turn, may explain why sickle cells adhere to endothelium in vitro. Although the RBC transmembrane sialoglycoproteins may function in communication, regulation of cell shape, and adhesion, uncertainties exist regarding many of their functions. To study these sialoglycoproteins, we have developed a double staining technique (Dzandu et al., 1984) that differentially stains human RBC membrane sialoglycoproteins and asialoproteins in SDS-polyacrylamide gels. This should aid in elucidating the conformational structure and function of transmembrane glycoproteins.

Biosynthesis of human blood group T-, N- and M-specific immunodeterminants on human erythrocyte antigens

We synthesized on Tn erythrocytes with human sera, UDP-Gal, and activators T-specific haptenic structures in satisfactory yield. The specificity of this biosynthesis was ascertained by agglutination with human and animal anti-T, by specific absorption of human anti-T as well as by agglutination inhibition assays. With isolated human erythrocyte T antigen as substrate we synthesized N- and M-specific structures with sera from individual human donors in presence of CMP-sialic acid by incubation for 24 hr at 37 degrees C. Serology on the recovered product was carried out with nineteen monospecific human and animal sera under strictly standardized and controlled conditions with the mandatory tube assay. All M- as well as N-derived T antigens tested acquired N specificity with all transferase sera of all MN types. In contrast, M-activation of M- and N-drived T antigens tested acquired N specificity with all transferase sera of all MN types. In contrast, M-activation of M- and N-derived T antigens occurred only if the transferase donor had the M gene. The nine M transferase sera used all gave M-activation of MM- and NN-derived T antigens. None of twelve transferase sera from NN donors M-activated any T antigen. NN antigen was transformed to a M-specific one by all transferase sera from MM donors but by none from NN donors. We have not yet established the biochemical-genetic relation of M to N; N may be the immediate precursor of M or M may originate directly from T. The sialyltransferase responsible for M activation may be a N transferase 'modified' by the M gene product or an entirely different sialyltransferase.

Effect of periodate oxidation on specific activities and carbohydrate components of human blood group N- and M-specific glycoproteins and glycopeptides

Mild as well as strong periodate oxidation of isolated erythrocyte N and M glycoproteins and glycopeptides gave extensive to complete destruction of N-specificities as measured with Vicia graminea extracts and of N- as well as M- activities determined with all but 1 of 8 animal anti-N and 13 anti-M sera. Results with human antisera differed somewhat, while the specificity of mildly oxidized N-glycoprotein was completely destroyed as determined with all 8 human anti-N sera used, that of strongly oxidized N-active substance was completely inactivated towards one of the human antisera; the remainder showed 63--94% destruction, and 2 sera indicated no effect of oxidation. Similarly, while 10 of 14 human anti-M indicated complete inactivation of M-specific glycoproteins and glycopeptides after mild or strong oxidation, 2 showed partial inactivation and 2 human anti-M sera showed no inactivating effect of oxidation. The most relevant findings of quantitative carbohydrate analysis of periodate oxidized N- and M-specific substance were extensive transformation of N-acetylneuraminic acid (NAN) to its C8 and C1 analogues on mild oxidation and pronounced destruction of NAN and its analogues on strong oxidation; however, some intact NAN always remained. In all instances galactose (Gal) was destroyed to a much larger extent in N-derived than in M-derived glycoproteins and glycopeptides.

Mechanisms of intercellular adhesion

Investigations into cellular adhesion, both of a biochemical and biophysical nature, have not yet produced an established theory or widely accepted hypothesis to explain the mechanics of this fundamental biological process although much information concerning the structure and function of the mammalian cell surface has been gained. At the present time there is increasing evidence to suggest that cellular adhesion is mediated by specific cell surface macromolecules which are capable of forming protein-carbohydrate complexes possibly resembling those found between plant lectins and their carbohydrate substrates.

Asymmetric binding of cytochrome b5 to the membrane of human erythrocyte ghosts

The intact, amphipatic form of cytochrome b5 could bind to unsealed ghosts, but not to resealed ghosts, suggesting that the cytochrome could bind only to the inner (cytoplasmic) surface of the ghost membrane. This was further confirmed by the finding that the cytochrome could bind to closed, inside-out vesicles prepared from the ghosts. This asymmetric binding was not due to the exclusive localization of sialic acid and sugar chains on the outer surface of the ghosts membrane, because the cytochrome could not bind to ghosts even after enzymatic removal of these components. Although liposomes consisting of phosphatidylcholine or both phosphatidylcholine and sphingomyelin could effectively bind the cytochrome, this binding capacity was progressively decreased as increasing amount of cholesterol was included in the composition of phosphatidylcholine liposomes. Removal of cholesterol from resealed ghosts by incubation with egg phosphatidylcholine liposomes resulted in the binding of cytochrome b5 to the outer surface of the treated ghosts. The possibility is discussed that the asymmetric binding is due to preferential localization of cholesterol in the outer leaflet of the lipid bilayer that constitutes the ghost membrane.

Perturbational effects of inorganic cations on human erythrocyte membranes

The perturbational effects of monovalent and divalent cations on human erythrocyte membranes were analyzed by examining their influence on kinetic and structural characteristics of trinitrobenzenesulfonic acid (TNBS) incorporation into the amino groups of protein and phospholipid structural components. The stimulatory effects of monovalent cations on TNBS incorporation, which were size-independent and attributed to nonspecific membrane alterations resulting from ionic strength factors, contrasted with the more pronounced stimulatory properties of divalent cations which were markedly size-dependent. These stimulatory effects of cations on TNBS incorporation were associated with alterations not only in rate but also in activation energy in incorporation. Changes in activation energy produced by divalent cations paralleled their ability to perturb membrane protein components and probably reflected changes in probe permeation. The rate of TNBS incorporation exhibited a dependence on divalent cation ionic radius which paralleled ion-induced perturbations in the labelling of the membrane amino phospholipid phosphatidylethanolamine. Divalent cations differed both in the relative extent and in the characteristics of protein and phospholipid perturbation. Alkaline earth cations behaved as a rather homogeneous group while Ni++, Co++ and Mn++ constituted a second heterogeneous group. The influence of monovalent and divalent cations on the hemolytic behavior of intact erythrocytes paralleled their effects on TNBS incorporation into isolated membranes rather closely. It is suggested that TNBS incorporation may provide a valuable means of analyzing functionally relevant cation-induced alterations in biological membranes in general.

Identification of a sialoglycopeptide released by self-digestion from human erythrocyte membranes

Membranes from human O Rhesus-positive erythrocyte 'ghosts' were tested in vitro for their ability to digest their own glycoproteins. 'Ghost' membranes incubated in Tris/HCl buffer, pH 7.4, release a sialoglycopeptide, which contains glucosamine, galactosamine, galactose and mainly polar amino acids. Chemical composition, molecular size and aggregation properties suggest that this glycopeptide may be a fragment of glycophorin.

Sialoglycopeptides from bovine milk fat globule membrane

Milk fat globule membrane was shown to contain sialic acid, all of which could be released without disruption of the fat globule. Sialoglycopeptides were cleaved from the surface of intact fat globules by Pronase and fractionated on Sephadex G-50. Further fractionation of the major sialoglycopeptide peak on DEAE-Sephadex gave two groups of sialoglycopeptides eluted with 0.1 M NaCl (Group A) and 0.5 M NaCl (Group B), respectively. Refractionation gave a major sialoglycopeptide from each of the two groups together with a total of three minor sialoglycopeptides. All five sialoglycopeptides eluted as single peaks using shallow salt gradients on DEAE-Sephadex and contained a hydrophilic peptide chain together with galactose, mannose, N-acetylgalactosamine, N-acetylglucosamine, and sialic acid. Glycopeptides of Group A but not Group B contained fucose. The major sialoglycopeptide of Group B released 35% of its hexose and hexosamine on treatment with alkaline borohydride leaving a sialoglycopeptide which had reduced serine and threonine and elevated alanine levels and in addition contained 2-aminobutyric acid. An oligosaccharide fraction containing N-acetylgalactosaminitol, galactose and sialic acid in a molar ratio of 1:1:2 was partially characterised from the clevage mixture. The major sialoglycopeptide of Group A had a more complex carbohydrate structure and showed no released carbohydrate on treatment with alkaline borohydride. The sialoglycopeptides of milk fat globule membrane show many similarities with those of erythrocyte membrane and have a potential use in comparative and structural studies.

Dissociation of MNSs antigens using sodium deoxycholate fraction of erythrocyte membranes

A dissociation of the MNSs blood group system has been demonstrated by sodium deoxycholate fractionation of erythrocyte membrane preparations. The MN antigens were localized in a fraction that sedimented at 30,000 g while the Ss antigens remained soluble after 100,000 g centrifugation for donors who were heterozygous at the MN locus. In contrast, fractions derived from erythrocytes of 'homozygous' M or N donors showed M or N activity in both the 30,000 g pellet and 100,000 g supernatant and Ss remained in the 100,000 g supernatant.

Chemical characterization and pronase susceptibility of the Na:K pump-associated phosphoprotein of human red blood cells

The phosphoproteins formed by incubation of red cell ghosts with [gamma-(32)P]ATP in the presence of Mg and Na + Mg have been characterized by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis. The (32)P-labeled phosphoprotein was seen as a single peak confined to the region of the diffuse 90,000 dalton polypeptide band; labeling with Na + Mg considerably increased the quantity of (32)P-phosphoprotein contained in this band relative to labeling with Mg alone. Treatment of intact cells with Pronase known to partially hydrolyze the glycoproteins and the 90,000 daltons polypeptide did not change either the amount or the position of the (32)P-phosphoprotein present in the gels. The molecular weight of the (32)P-phosphoprotein is estimated to be 103,000. Pronase treatment of intact cells also did not significantly alter any of the transport parameters of the membrane such as the K pump flux, ouabain binding, or Na,K-ATPase. In contrast, treatment of ghosts with Pronase not only resulted in drastic alteration of the transport parameters but also inhibited the formation of the phosphoprotein under all conditions. Thus, while the Na:K pump is not intrinsically resistant to Pronase, those elements of the pump which are susceptible are not accessible from the outside of the cell. Further, SDS-polyacrylamide gel electrophoresis after Pronase treatment of intact cells results in a substantial increase in the purification of the phosphoprotein relative to that which was previously possible in ghosts.

Chemical characterization and surface orientation of the major glycoprotein of the human erythrocyte membrane

The major glycoprotein of the human erythrocyte membrane has been isolated by treatment with lithium di-iodosalicylate and found to be a single polypeptide chain with a molecular weight of about 50,000. This molecule, which is 60% carbohydrate and 40% protein, carries multiple blood-group antigens, the receptors for influenza viruses, and various plant agglutinins. Four unique carbohydrate-containing peptides (alpha-1, alpha-2, alpha-3, and beta) are produced by tryptic digestion of the isolated glycoprotein; their order in the molecule has been determined by sequential tryptic digestion of intact erythrocyte membranes and partially digested glycoprotein fragments. Cleavage of the native protein with cyanogen bromide produces five fragments; two of these (C-5 and C-1) contain most of the carbohydrate in the molecule and are derived from the N-terminal half of the polypeptide chain. The nonpolar amino acids of this glycoprotein are located predominantly in the C-terminal fragment (C-2). Phytohemagglutinin conjugated to ferritin has been used to map the distribution of glycoprotein receptors over the surfaces of intact erythrocytes by freeze-etching and electron microscopy. This label localizes to sites on the membrane that overlie the intramembranous particles. These findings suggest that the glycoprotein is oriented at the cell surface with its oligosaccharide-rich N-terminal end exposed to the exterior, while its C-terminal segment interacts with other components in the interior of the membrane to form intramembranous particles.

The structure of erythrocyte membranes studied by freeze-etching. II. Localization of receptors for phytohemagglutinin and influenza virus to the intramembranous particles

The distribution of specific glycoprotein receptors on the external surfaces of red cells was mapped, by the freeze-etching technique, to determine if the receptors coincided with the underlying 75-A intramembranous particles. Phytohemagglutinin, ferritin-conjugated phytohemagglutinin, and influenza virus were used as labeling agents since they can be seen by freeze-etching techniques and each reacts with a different site on the same glycoprotein molecule. The distribution of these labels was studied on intact human red cells, isolated ghost membranes, and trypsin-treated ghost membranes. The results show that the receptors for these labels are distributed uniformly over the surfaces of normal red cell membranes in the same apparent distribution as that of the 75-A particles within the membrane. The association between the external receptors and the underlying particles is especially evident when trypsin-treated ghost membranes are labeled: the labeled receptor sites and the intramembranous particles both form sharply defined, reticulated networks, which overlap. These results provide further support for the idea that membrane-bound glycoproteins are oriented so that their carbohydrate-rich segments, which bear the antigenic sites and receptors, are exposed to the external medium, while hydrophobic segments of the same molecules interact with lipids, and possibly other proteins, to form the intramembranous particles.

Chemical modification of membranes. I. Effects of sulfhydryl and amino reactive reagents on anion and cation permeability of the human red blood cell

Four different amino-reactive reagents, 4-acetamido-4'-isothiocyano-stilbene-2,2'-disulfonic acid (SITS),(1) 1-fluoro-2,4-dinitrobenzene (FDNB), 2,4,6-trinitrobenzene sulfonic acid (TNBS), and 2-methoxy-5-nitrotropone (MNT) decrease the anion permeability of the human red blood cell, as measured by sulfate fluxes, whereas the sulfhydryl agent, parachloromercuriphenyl sulfonic acid (PCMBS), does not. In contrast, PCMBS increases the cation permeability as measured by K(+) leakage, whereas SITS does not. Of the other agents, FDNB increases the cation permeability to the same extent as PCMBS but MNT and TNBS produce smaller increases. PCMBS does not protect against FDNB as it does against other sulfhydryl agents (X-irradiation) and the FDNB effect on cations is attributed to amino groups. Studies of the binding of SITS indicate that it does not penetrate into the membrane and its failure to influence cation permeability is attributed to its inability to reach an internal population of amino groups. It is concluded that two ion permeability barriers, both involving proteins, are present in the red blood cell. The more superficial barrier contains amino groups and controls anion flow; the more internal barrier contains sulfhydryl and amino groups and controls cation flow. The amino groups contribute to the control of permeability by virtue of their positive charges, but the role of sulfhydryl groups is not clear. Only a small fraction of the membrane protein amino and sulfhydryl is involved in the barriers.

Inside-out red cell membrane vesicles: preparation and purification

Plasma membranes purified from human red cells were Converted into small vesicleS by disruption in alkaline buffer of low ionic strength. Most of these vesicles were inside-out. The presence of divalent cations prevented this inversion. The inside-out vesicles were separatcd from right-side-out vesicles by centrifugration to equilibrium in dextran density gradients.

The electrophoretic mobility of normal and leukaemic cells of mice

1. The pH-mobility relationships for saline-washed cells from a mouse strain of acute lymphoblastic leukaemia were examined before and after treatment with lower aldehydes, diazomethane and neuraminidase (EC 3.2.1.18). 2. The content of sialic acid released into the supernatant fluid of neuraminidase-treated cells was measured. 3. The stability of the charge-determining structures to temporary changes in environment (pH and ionic strength) was established. 4. Similar measurements were made on lymph-node cells obtained from non-leukaemic mice (a resistant and a leukaemia-susceptible strain were examined). 5. It is deduced that both the malignant and the non-malignant cell possess two dissociable acid functions at the cell surface, a carboxyl group of sialic acid and another acidic group(s), probably carboxyl, of pK 3.0-4.5. The malignant cells, however, have a basic dissociable function not present in the non-malignant types. 6. Suggestions are made as to how the difference in surface chemistry may be related to the problem of malignancy.