Monoclonal anti-A activity against the FORS1 (Forssman) antigen.

BACKGROUND: The FORS blood group system (originally recognized as the Apae phenotype) was discovered by sporadic activity against polyclonal anti-A reagents and activity against the lectin Helix pomatia. The extent of monoclonal anti-A reagent activity against the FORS1 antigen is serologically and immunochemically incomplete. STUDY DESIGN AND METHODS: In the absence of natural FORS1-positive red blood cells (RBCs), kodecytes were created with synthetic disaccharide and pentasaccharide Forssman function-spacer-lipid (FSL) constructs, Fsdi -kodecytes, and FORS1-kodecytes, respectively. FSL constructs were also applied to solid surfaces and used in solid-phase enzyme immunoassays. A range of characterized monoclonal anti-A and anti-B reagents were then serologically and immunochemically characterized against these Forssman antigens. Polyclonal human anti-A, anti-B, the lectin H. pomatia serologic reagents; and canine RBCs were used as serologic controls. RESULTS: None of 19 different monoclonal anti-A reagents were able to detect the pentasaccharide Forssman on FORS1-kodecytes, while three reagents were able to detect disaccharide Forssman on Fsdi -kodecytes. Most anti-A reagents were immunochemically reactive with both the di- and the pentasaccharide Forssman antigens in the solid-phase assays. Historic polyclonal human anti-A and the lectin H. pomatia reacted strongly with the FORS1-kodecytes, correlating with the discovery of the Apae phenotype and supporting the use of FORS1-kodecytes as FORS1 surrogates. CONCLUSIONS: Monoclonal anti-A reagents, despite showing reactivity against the FORS1 antigen in solid-phase assays are unlikely to cause the agglutination of FORS1 antigen-positive RBCs.

Interaction of antibody with Forssman antigen in guinea pigs. A mechanism of adaptation to antibody- and complement-mediated injury.

Forssman antigen is a glycosphingolipid with antigenic specificity determined by extra-membrane haptenic sugars similar to blood group antigens and antigens that are the main barrier to xenogeneic organ transplantation. Herein, we describe the localization of Forssman antigen in guinea pig lungs and kidneys and the consequences of its interaction with antibodies in vitro and in vivo (Forssman reaction). Exposure of cultured guinea pig aortic endothelial cells to Forssman antibodies induced rapid redistribution of antigen-antibody complexes at the cell surface, followed by shedding that occurred by blebbing of plasma membrane as vesicles or fragments, and was associated with disappearance of antigen from the cell surface (antigenic modulation). Guinea pigs surviving frequent intravenous infections of increasing amounts of antibodies, for a total of 20 to 40 lethal doses, developed a partial or complete adaptation to generalized Forssman reaction, and adaptation was associated with partial or complete modulation of Forssman antigen at the surface of the pulmonary and, in minor degree, renal endothelial and epithelial cells. These findings support the hypothesis that modulation of endothelial carbohydrate antigens contributes to adaptation of highly vascularized organs exposed to tolerable levels of allo- or xenoantibodies.

Globo-series carbohydrate antigens are expressed in different forms on human and murine teratocarcinoma-derived cells.

The glycolipids of human teratocarcinoma-derived cell line NCCIT were compared with those of 5 murine teratocarcinoma-derived cell lines. Glycolipid antigens were identified by cell surface immunofluorescence and high-performance thin-layer chromatography (HPTLC) immunostaining with a panel of monoclonal anti-carbohydrate antibodies. Human NCCIT embryonal carcinoma (EC) cells contained extended globo-series glycolipids Gb5 (galactosyl globoside) and GL7 (sialyl galactosyl globoside) recognized by antibodies to stage-specific embryonic antigens 3 and 4 (SSEA-3 and -4). SSEA-4 was not detected by immunofluorescence on the surface of any of the 5 murine teratocarcinoma-derived cell lines examined; however, SSEA-3 was detected on the surface of murine cell lines resembling primitive endoderm (JC44, NF-PE) and trophectoderm (E6496D). HPTLC analysis revealed a large amount of globoside (Gb4) in these differentiated cells, which may account for their labeling with anti-SSEA-3 antibody. Globo-series glycolipids were also detected in murine EC cells; however, differences were noted between the 2 cell lines examined. F9 cells contained primarily Gb4 and Forssman glycolipid, whereas NF-1 cells contained only minor amounts of Gb4 and lacked Forssman glycolipid entirely. Our results, coupled with the known distribution of Forssman antigen in the egg cylinder-stage mouse embryo, suggest that F9 and NF-1 murine EC cells are replicas of cells at different stages of development of the embryonic ectoderm. Glycolipids of normal mouse embryos were examined for comparison. Gb4 and Forssman glycolipid were presents in both embryonic and extra-embryonic tissues, whereas Gb5 and GL7 were restricted to visceral yolk sac and placenta. Our results demonstrate that human and murine teratocarcinoma-derived cells both synthesize extended globo-series glycolipids; however, oligosaccharide chain elongation takes different pathways in the 2 species. These differences reflect species-related and cell type-specific patterns of glycosylation.

Differences in outer membrane characteristics between gallstone-associated bacteria and normal bacterial flora.

Previous studies with scanning electron microscopy (SEM) have suggested that pigment gallstones contain bacteria. We set out to culture these bacteria and to study their membrane characteristics. We studied gallstones from 54 patients (36 men, 18 women; mean age 55.4 years) admitted consecutively to two hospitals for cholecystectomy. SEM detected bacteria in all of 14 brown pigment stones, 2 of 14 black pigment stones, and in the pigmented centres of 9 of 19 mixed cholesterol stones; no bacteria were detected in 14 pure cholesterol stones or within the cholesterol portions of mixed stones. We were able to culture bacteria from all gallstones with bacteria seen on SEM and for which sufficient material was available (n = 16). 20 bacterial species were recovered from these stones. Gallstones containing bacteria were associated with clinical sepsis and cholangitis. All bacteria obtained from gallstones agglutinated human O P1 erythrocytes, which reflects the presence of P1-specific fimbriae. 5 strains were positive for Forssman-antigen-specific fimbriae. None showed evidence of mannose-specific fimbriae. All of the organisms bound anti-Gal, a ubiquitous naturally occurring IgG specific for alpha-galactosyl residues. The presence of P1 fimbriae and alpha-galactosyl residues and the absence of mannose-specific fimbriae distinguish these organisms from gut flora. We postulate that possession of these unusual properties may enhance the ability of bacteria to colonise the biliary tree and initiate pigment gallstone formation.

Surface expression of Forssman glycosphingolipid antigen on murine bone marrow-derived macrophages is subject to both temporal and population-specific regulation and is modulated by IL-4 and IL-6.

The Forssman glycolipid antigen (Fo) has been shown to be a differentiation marker for mouse macrophages both in vivo and in vitro. In order to determine whether or not there is a relationship between stage of differentiation and Fo expression, we have analyzed the kinetics of Fo expression during the growth of cultured mouse bone marrow-derived macrophages (BMDM). BMDM were grown in serum free medium to avoid the possible influence of undefined serum factors. In this medium they could be maintained over a period of up to 20 days with cell yields comparable to those obtained with serum-supplemented media. Fo antigen was assayed with a specific antibody using both a whole cell ELISA and immunocytochemical staining of cells grown on slides. With increasing age in culture, BMDM showed a gradual quantitative increase in Fo expression and parallel increase in the Fo+ BMDM fraction from about 10% Fo+ cells on the 10th day of culture to a maximum of 50%-60% Fo+ cells between the 17th and 19th days. The temporal control over the development of the Fo+ cell fraction was intrinsic to BMDM maturation but was specific for Fo. During the same time period expression of MHC class II (Ia) remained consistently low, whereas expression of both Mac-1 (C3bR) and the macrophage-specific marker ER-BMDM-1 was always high. The interleukins IL-4 and especially IL-6 induced a premature expression of Fo at earlier stages of BMDM culture, but neither could promote further Fo expression once the intrinsically occurring maximum had been reached. No evidence in support of an autocrine regulation of Fo expression by IL-6 could be obtained, nor could a connection between cell cycle status and Fo expression be established. These data provide further evidence that Fo is a temporally regulated differentiation marker for a mouse macrophage subpopulation and for modulation of its expression by lymphokines.

Use of recombinant biotinylated aequorin in microtiter and membrane-based assays: purification of recombinant apoaequorin from Escherichia coli.

Aequorin is a calcium-dependent bioluminescent protein isolated from the hydromedusan Aequorea victoria. The gene for aequorin has been cloned and overexpressed in Escherichia coli [Prasher et al. (1985) Biochem. Biophys. Res. Commun. 126, 1259; Prasher et al. (1987) Biochemistry 26, 1326]. Higher levels of expression have recently been obtained by subcloning aequorin cDNA into the pRC23 plasmid vector such that its expression is under control of the lambda PL promoter [Cormier et al. (1989) Photochem. Photobiol. 49, 509]. Purification of recombinant apoaequorin from E. coli containing this new recombinant plasmid (pAEQ1.3) was accomplished by a two-step procedure involving gel filtration and anion-exchange chromatography on Sephadex G-100 and DEAE-Sepharose, respectively. Typically, 400-500 mg of recombinant protein was obtained from 100 L of fermentation culture. The purified recombinant apoaequorin could be converted to aequorin in high yield upon incubation with synthetic coelenterate luciferin, dissolved oxygen, and a thiol reagent with a photon yield similar to the native photoprotein. Detection of recombinant aequorin in the Dynatech ML1000 Microplate luminometer was linear between 10(-18) and 10(-12) mol, and little loss of specific activity was observed when the protein was derivatized with biotin. The biotinylated derivative was stable when frozen, lyophilized, or stored at 4 degrees C. The feasibility of using biotinylated aequorin as a nonradioactive tag was established by its application in a variety of solid-phase assay formats using the high-affinity streptavidin/biotin interaction. A microtiter-based bioluminescent immunoassay (BLIA) using biotinylated aequorin and the ML1000 luminometer was developed for the detection of subnanogram amounts of a glycosphingolipid (Forsmann antigen). In addition, nanogram to subnanogram quantities of protein antigens and DNA, immobilized on Western and Southern blots, respectively, were detected on instant and X-ray films using biotinylated aequorin.

Apical and basal Forssman antigen in MDCK II cells: a morphological and quantitative study.

We have studied the surface distribution of a glycosphingolipid (the Forssman antigen) in MDCK II and CCL39 cells. The Forssman antigen is mobile on the surface of both these cell lines. Its surface distribution is homogenous on non-polarized cells. Under conditions where MDCK II cells are well polarized, the Forssman antigen is present in equal amounts on the apical membrane and on the basal membrane and its processes. Very little Forssman antigen can be detected on the lateral membrane. The nature of the mechanism excluding the Forssman antigen from the lateral domain remains to be determined. This surface distribution is established within hours after plating and was observed with cells grown on different types of filters. The surface density of the Forssman antigen on the apical and on the basal domain has been estimated. No involvement of the basal Forssman antigen in cell attachment could be demonstrated. However, the apical Forssman antigen appears to be essential to the establishment of the cells in culture.

Subcellular localization of Forssman glycolipid in epithelial MDCK cells by immuno-electronmicroscopy after freeze-substitution.

Forssman antigen, a neutral glycosphingolipid carrying five monosaccharides, was localized in epithelial MDCK cells by the immunogold technique. Labeling with a well defined mAb and protein A-gold after freeze-substitution and low temperature embedding in Lowicryl HM20 of aldehyde-fixed and cryoprotected cells, resulted in high levels of specific labeling and excellent retention of cellular ultrastructure compared to ultra-thin cryosections. No Forssman glycolipid was lost from the cells during freeze-substitution as measured by radio-immunostaining of lipid extracts. Redistribution of the glycolipid between membranes did not occur. Forssman glycolipid, abundantly expressed on the surface of MDCK II cells, did not move to neighboring cell surfaces in cocultures with Forssman negative MDCK I cells, even though they were connected by tight junctions. The labeling density on the apical plasma membrane was 1.4-1.6 times higher than basolateral. Roughly two-thirds of the gold particles were found intracellularly. The Golgi complex was labeled for Forssman as were endosomes, identified by endocytosed albumin-gold, and lysosomes, defined by double labeling for cathepsin D. In most cases, the nuclear envelope was Forssman positive, but the labeling density was 10-fold less than on the plasma membrane. Mitochondria and peroxisomes, the latter identified by catalase, remained free of label, consistent with the notion that they do not receive transport vesicles carrying glycosphingolipids. The present method of lipid immunolabeling holds great potential for the localization of other antigenic lipids.

Regulation of Forssman antigen expression during maturation of mouse stromal macrophages in haematopoietic foci.

During a trial to develop a monoclonal antibody (mAb) specific to stromal macrophages (M phi) in haematopoietic foci, we have created a mAb, designated F10, that stains the stromal M phi more selectively than any other mAb reported. As F10 was found to react with Forssman glycosphingolipid (GSL) specifically and to give clearer immunostaining than anti-Forssman GSL IgG, we have studied Forssman antigen expression during maturation of the stromal M phi in splenic haematopoietic foci using F10 and a system of allogenic bone marrow transplantation which allows us to know the turnover of the stromal M phi in vivo. C3H/He mice (H-2k) were lethally irradiated and intravenously infused with the bone marrow cells of BALB/c nu-nu mice (H-2d). Splenic frozen-sections and cytocentrifuge preparations of splenic haematopoietic clusters from the recipient mice were stained with F10 and with mAb against major histocompatibility class I antigens. H-2d-type stromal M phi began to appear in the haematopoietic clusters at Week 5 and they gradually replaced H-2k-type stromal M phi. The percentage of Forssman+ stromal M phi gradually decreased and reached a nadir at Week 6, when most stromal M phi were already of the donor type. At Week 8, however, Forssman+ stromal M phi levels returned to normal. The delayed expression of Forssman antigen on the stromal M phi in haematopoietic foci following genotypic conversion suggests that Forssman antigen is regularly expressed on the subpopulation of stromal M phi, which mature well under specific microenvironmental factors in vivo.

Preservation and immunogold localization of lipids by freeze-substitution and low temperature embedding.

The success of post-embedding immunocytochemistry depends largely on the preparation methods. The requirements for structural preservation and immunocytochemistry are in some cases contradictory. This is especially the case in the study of lipid-rich structures and the localization of lipid components. Earlier work on freeze-substitution has shown that this method is very promising for the preservation of lipids and the immunocytochemical localization of lipids at the electron microscopical level. In this study we show that freeze-substitution in combination with low temperature embedding in Lowicryl HM20 has fulfilled this promise. Lamellar bodies in alveolar type II cells contain about 90% lipids and are very difficult to preserve in ultrathin cryosections. Lowicryl sections of freeze-substituted lung tissue shows excellent preservation of lamellar bodies in combination with immunogold localization of a hydrophobic surfactant protein. With an antibody against the Forssman glycolipid we demonstrate a highly reproducible intracellular localization of this glycolipid with high specificity and resolution. This method results in the retention of lipids and glycolipids and allows postembedding immunogold labeling.

Forssman antigen expressed on lymph node cells of MRL/MpJ-lpr/lpr mice is of a glycoprotein nature.

This paper reports the nature of abnormally expressed Forssman (F) antigen in the lymph node cells of MRL/MpJ-lpr/lpr, autoimmune mice, and also reports its autoantibody in sera. By acetylation study of the F antigen with [14C]acetic anhydride, we concluded that the F antigen was not a glycolipid but a glycoprotein. Several bands of F-active glycoproteins were identified on a nitrocellulose sheet after purification by an anti-F antibody affinity column. Hemolysis of SRBC by some sera from MRL/MpJ/lpr/lpr was inhibited by purified F glycoprotein and also by F glycolipid. The antibody in the serum, however, seemed to be more specific for F glycoproteins than F glycolipid, but the opposite was the case for rabbit anti-F glycolipid antibody. No significant difference of the SRBC hemolysis levels was observed between the sera from MRL/MpJ-lpr/lpr and its congenic MRL/MpJ-+/+ mice.

Localization of Forssman glycolipid and GM1 ganglioside intracellularly and on the surface of germ cells during fetal testicular and ovarian development of mice.

Changes in the expression pattern and intracellular localization of Forssman glycolipid (FA) and GM1 ganglioside (GM1) in fetal mouse gonads were examined during germ cell differentiation by immunofluorescence microscopy and immunoelectron microscopy. In male germ cells from the 12th to 14th day p.c., anti-FA binding was localized in granular structures aggregated on one side of the cytoplasm and/or in the plasma membrane. On day 16 p.c., some germ cells still showed patch-like positive reactions in the plasma membrane, but by day 18 p.c., positive reactions for FA had completely disappeared. The female germ cells showed granular bindings of anti-FA scattered throughout their cytoplasm during the 13th to 16th day p.c., although the positive reactions in female germ cells on day 12 p.c. tended to be found in one side of cytoplasm and/or plasma membrane similar to those in male germ cells from 12th to 14th day p.c. On day 18 p.c., positive reactions remained in the plasma membrane of some germ cells, but these positive reactions disappeared before birth. Immunoelectron microscopic observation showed that the sites of anti-FA bindings were equivalent to the "small dense bodies" (SDB) and the Golgi lamellae both in male and female germ cells. On the other hand, GM1 was not detected in male germ cells at any time during fetal testicular development, whereas an anti-GM1 reaction was detected in the plasma membrane of female germ cells from the 16th to 18th day p.c. (oocytes in the first meiotic prophase).(ABSTRACT TRUNCATED AT 250 WORDS)

Expression of Forssman glycolipid and blood group-related antigens A, Le(x), and Le(y) in human gastric cancer and in fetal tissues.

The expression of Forssman glycolipid antigen in human gastric cancers was investigated by thin-layer chromatogram immunostaining of glycolipid fractions. Incompatible blood group A antigen, Le(x) and Le(y) antigen were also studied in comparison with Forssman antigen. Forssman glycolipid as the pentaglycosylceramide was demonstrated in nine out of 12 gastric cancers, and in three out of 10 adjacent uninvolved tissues. In most cases the content of Forssman glycolipid was increased in the cancers compared with that in the uninvolved counterparts. Forssman pentaglycosylceramide was also detected in some normal gastric mucosae (two out of four), and in a fetal gastrointestinal tract tissue. In immunohistochemical examination of gastric cancer tissues, sialidase treatment revealed a positive staining with anti-Forssman antibody. Some cancer tissues from patients with blood group O were found to contain blood group A-active glycolipids, which could be distinguished from Forssman glycolipid by thin-layer chromatogram immunostaining. The incidence of incompatible A-active glycolipids was two out of 10 cancers from patients with blood group O or B. Le(x)- and Le(y)-active glycolipids were detected in most of the preparations and were not accumulated consistently in the cancers.

A blood group B-like pentaglycosylceramide is the major complex glycosphingolipid of the Madin-Darby canine kidney epithelial cell line I (MDCK I).

Two sublines of the epithelial cell line MDCK differ in glycosphingolipid composition (Hansson, G.C. et al. (1986) EMBO J. 5, 483-489). The Forssman pentaglycosylceramide was an abundant glycolipid in the MDCK II subline, but was absent in the MDCK I subline. The MDCK I line instead contained another five-sugar glycolipid in relatively large amounts. This component has now been isolated and characterized with mass spectrometry, methylation analysis, exoglycosidase digestion, and proton NMR spectroscopy. The structure was concluded to be Gal alpha 1----3Gal beta 1----4GlcNAc beta 1----3Gal beta 1----4Glc beta 1----1 Cer. This is a blood group B-like glycolipid lacking fucose, earlier found in rabbit and bovine erythrocytes.

Forssman glycosphingolipid as an immunohistochemical marker for mouse stromal macrophages in hematopoietic foci.

The immunohistochemical distribution of Forssman glycosphingolipid (GSL) in mouse hematopoietic tissue was examined by using light and electron microscopic immunoperoxidase methods with a highly specific rabbit anti-Forssman GSL antibody. Bone marrow, splenic red pulp, and thymic macrophages, which are closely associated with hematopoietic cells, were stained by the antibody, whereas hematopoietic cells, circulating cells, alveolar macrophages, Kupffer cells, peritoneal resident macrophages, and macrophages derived from granulocyte-macrophage colony-forming units cultured in the presence of L-cell-conditioned medium were not stained. In addition, thymic cortical epithelial cells, the framework of reticular cells of the cortical and medullary regions of the mesenteric lymph node and periarterial lymphoid sheath of the spleen, and some vessels of the tissues examined were also stained. After phlebotomy, the fine cytoplasmic processes of Forssman-positive splenic red pulp macrophages were distributed extensively throughout the erythroid colonies. On the other hand, after hypertransfusion, these macrophages retracted their processes, became round, and tended to aggregate. These results indicate that Forssman GSL can be used as an immunohistochemical marker for stromal macrophages in the hematopoietic foci of the bone marrow and splenic red pulp.

Isolation and characterization of the major acidic glycosphingolipids from the liver of the English sole (Parophrys vetulus). Presence of a novel ganglioside with a Forssman antigen determinant.

Acidic glycosphingolipids of the liver of English sole, Parophrys vetulus, have been isolated and characterized by 1H nuclear magnetic resonance spectroscopy, methylation analysis, and by direct probe electron-impact and fast atom bombardment mass spectrometry. In addition to the acidic glycosphingolipids with known structures (sulfatide, GM4, GM3, GM2, and GD1a), two fractions of a major monosialosylganglioside with TLC mobility slower than GM1 were isolated and characterized as having the following structure. (Formula:q see text). The structure represents a novel combination of a terminal Forssman disaccharide (GalNAc alpha 1----3GalNAc beta 1----3R) and a GM1 ganglioside core linked together. The identity of the terminal Forssman disaccharide was further established by TLC immunostaining with an anti-Forssman monoclonal antibody. This antibody showed strongly positive staining of the ganglioside only after removal of the sialic acid. Thus, the II3NeuAc residue inhibited antibody binding to the terminal disaccharide unit. Analysis of the ceramide moieties of both fractions indicated a predominance of 16:0, 22:1, 22:0, and 24:1 fatty acids in the faster migrating form and 16:0, 18:0, and 18:1 in the slower form in combination with d18:1 sphingosine.

Immunopharmacological actions of the new antiallergic drug butyl 3'-(1H-tetrazol-5-yl)oxanilate. 1st communication: effects on type I to type IV allergic reactions in animal models.

The effects of butyl 3'-(1H-tetrazol-5-yl)oxanilate (WP-833), a new antiallergic drug, on type I to type IV allergic reactions were investigated by employing various animal models. WP-833 (i.v. and p.o.) dose-dependently inhibited homologous or heterologous passive cutaneous anaphylaxis (PCA) mediated by rat or mouse immunoglobulin E (IgE) in rats. Homologous PCA caused by guinea pig IgE was also inhibited by WP-833. In addition, WP-833 had inhibitory actions upon homologous PCA induced by rat or guinea pig IgG. However, WP-833 showed no inhibition of rat skin reactions caused by histamine, serotonin and bradykinin, contrasting with the inhibition of prostaglandin E1-induced skin reaction. Furthermore, both adrenalectomy and propranolol treatment exerted no influences on the inhibition of IgE-mediated homologous PCA in rats by WP-833. In contrast to above findings demonstrating that WP-833 clearly inhibited type I allergic reaction, systemic Forssman shock in guinea pigs and reversed cutaneous anaphylaxis in rats (type II), passive Arthus reaction in rats (type III), and contact dermatitis and tuberculin reaction in mice (type IV) were unaffected by WP-833 even in higher doses than in those capable of completely inhibiting type I allergic reaction.

Polarity of the Forssman glycolipid in MDCK epithelial cells.

To determine whether epithelial plasma membrane glycolipids are polarized in a manner analogous to membrane proteins, MDCK cells grown on permeable filters were analyzed for the expression of Forssman ceramide pentasaccharide, the major neutral glycolipid in these cells. In contrast to a recent report which described exclusive apical localization of the Forssman glycolipid (Hansson, G.C., Simons, K. and Van Meer, G. (1986) EMBO J. 5, 483-489), immunofluorescence and immunoelectron microscopic staining revealed the Forssman glycolipid on both the apical and basolateral surfaces of polarized cells. Immunoblots indicated that the Forssman antigen was detectable only on glycolipids and not on proteins. Analysis of metabolically labeled glycolipids released into the apical and basal culture medium, either as shed membrane vesicles or in budding viruses, also demonstrated the presence of the Forssman glycolipid on both apical and basolateral membranes of polarized cells. Quantitation of the released glycolipid indicated that the Forssman glycolipid was concentrated in the apical membrane. These results are consistent with previous reports which described quantitative enrichment of glycolipids in the apical domain of several epithelia.

Analysis of NMR spectra of sugar chains of glycolipids by 1D homonuclear Hartmann-Hahn and NOE experiments.

We applied 1D homonuclear Hartmann-Hahn (1D-HOHAHA) and difference NOE experiments to determine the chemical structure of Forssman's antigen, a glycolipid purified from sheep red blood cells. The subspectra corresponding to the individual sugar components were extracted from overlapping proton resonances by selective excitation of the anomeric proton resonances, so that unambiguous assignments of the sugar proton resonances were accomplished. Then, difference NOE experiments were performed to determine the linkage of the sugar units. The present procedure was found to be useful for the structure determination of glycoconjugates and also reduces the amount of samples and machine time.