TY - JOUR
AU - Karlsson,, Karl-Anders
AB - Abstract The possible role of glycosphingolipids as adhesion receptors for the human gastric pathogen Helicobacter pylori was examined by use of radiolabeled bacteria, or protein extracts from the bacterial cell surface, in the thin-layer chromatogram binding assay. Of several binding specificities found, the binding to lactosylceramide is described in detail here, the others being reported elsewhere. By autoradiography a preferential binding to lactosylceramide having sphingosine/phytosphingosine and 2-D hydroxy fatty acids was detected, whereas lactosylceramide having sphingosine and nonhydroxy fatty acids was consistently nonbinding. A selective binding of H.pylori to lactosylceramide with phytosphingosine and 2-D hydroxy fatty acid was obtained when the different lactosylceramide species were incorporated into liposomes, but only in the presence of cholesterol, suggesting that this selectivity may be present also in vivo. Importantly, lactosylceramide with sphingosine and hydroxy fatty acids does not bind in this assay. Furthermore, a lactosylceramide-based binding pattern obtained for different trisaccharide glycosphingolipids is consistent with the assumption that this selectivity is due to binding of a conformation of lactosylceramide in which the oxygen of the 2-D fatty acid hydroxyl group forms a hydrogen bond with the Glc hydroxy methyl group, yielding an epitope presentation different from other possible conformers. An alternative conformation that may come into consideration corresponds to the crystal structure found for cerebroside, in which the fatty acid hydroxyl group is free to interact directly with the adhesin. By isolating glycosphingolipids from epithelial cells of human stomach from seven individuals, a binding of H.pylori to the diglycosylceramide region of the non-acid fraction could be demonstrated in one of these cases. Mass spectrometry showed that the binding-active sample contained diglycosylceramides with phytosphingosine and 2-D hydroxy fatty acids with 16–24 carbon atoms in agreement with the results related above. bacterial adhesion, ceramide preference, Helicobacter pylori, lactosylceramide, liposomes Introduction Recognition of cell surface carbohydrates is a common motif for microorganisms, and microbial toxins, and plays an essential role for colonization and initiation of infection (Beachey, 1981). Carbohydrate receptors have thus been identified for a large number of bacteria (Beachey, 1981; Mirelman and Ofek, 1986; Karlsson, 1989). Among these, numerous observations of bacteria, both commensals and pathogens, which on thin-layer chromatograms appear to bind in a lactosylceramide-related manner, i.e., to lactosylceramide with a ceramide containing sphingosine/phytosphingosine and 2-D hydroxy fatty acids, isoglobotriaosylceramide, gangliotriaosylceramide, and gangliotetraosylceramide, have been reported (Karlsson, 1989). This binding specificity has also been characterized in detail for some species (Strömberg et al., 1988; Strömberg and Karlsson, 1990a,b). The mechanisms utilized by the newly discovered human gastric pathogen H.pylori for adherence and colonization of the epithelial cells of the human stomach have been the subject of many studies in recent years, and several putative receptors have been reported, such as, for example, sulfatide and the ganglioside GM3 (Saitoh et al., 1991), phosphatidylethanolamine (Lingwood et al., 1992), and the Leb antigen (Borén et al., 1993). Thus, the binding specificities identified to date for H.pylori are diverse and epitope carriers may be of both glycoprotein and glycolipid origin, which most likely is a reflection of the varying conditions this bacterium goes through during different stages of colonization and development of infection (Doig et al., 1992). In a series of studies of carbohydrate receptors for H.pylori, the binding of radiolabeled bacteria to mixtures of glycosphingolipids separated on thin-layer plates was examined. Thereby it was found that this microorganism also recognizes lactosylceramide, with a preference for the species with sphingosine/phytosphingosine and 2-D hydroxy fatty acids. In addition, several other distinct binding specificities were detected, as will be reported separately. By use of a binding assay for glycosphingolipids incorporated into liposomes, it was shown that H.pylori specifically recognizes the lactosylceramide species having phytosphingosine and 2-D hydroxy fatty acids. Molecular modeling indicates that this selectivity may result from recognition of a conformer in which the Glc hydroxy methyl group forms a hydrogen bond with the 2-D hydroxyl group of the fatty acid or, alternatively, the conformer analogous to the cerebroside crystal conformation. Furthermore, binding of H.pylori to the diglycosylceramide region of one non-acid glycosphingolipid fraction isolated from the epithelial cells of human stomach was obtained, and the presence of diglycosylceramides with phytosphingosine and 2-D hydroxy fatty acids in this fraction was demonstrated by mass spectrometry. Results Chromatogram binding assay By incubating 35S-labeled H.pylori, or 125I-labeled protein extracts from the bacterial cell surface, on thin-layer chromatograms with chromatographed mixtures of naturally occurring glycosphingolipids, followed by visualization of the bound bacteria or proteins by autoradiography, a classical binding pattern for lactosylceramide-binding bacteria (Strömberg et al., 1988; Karlsson, 1989; Strömberg and Karlsson, 1990a,b) was detected. In addition, several other distinct binding specificities were found, which will be reported separately. Fig. 1. Open in new tabDownload slide Binding of 35S-labeled Helicobacter pylori to glycosphingolipids separated on thin-layer plates. (A) Glycosphingolipids detected with anisaldehyde. (B) Glycosphingolipids detected by autoradiography after binding of radiolabeled H.pylori strain 1139. The assay was done as described in the Materials and methods section. The lanes contained the following glycosphingolipids: lane 1, non-acid glycosphingolipids from human blood group A erythrocytes, 40 µg; lane 2, non-acid glycosphingolipids from dog small intestine, 40 µg; lane 3, non-acid glycosphingolipids from guinea pig erythrocytes, 40 µg; lane 4, non-acid glycosphingolipids from mouse feces, 40 µg; lane 5, non-acid glycosphingolipids from human meconium, 40 µg; lane 6, non-acid glycosphingolipids from human stomach of a blood group A(Rh+)p individual, 40 µg; lane 7, non-acid glycosphingolipids from human stomach of a blood group A individual, 40 µg; lane 8, non-acid glycosphingolipids from human stomach of a blood group O individual, 40 µg; lane 9, non-acid glycosphingolipids from mucosal cells of human stomach of a blood group O(Rh-) individual, 40 µg; lane 10, non-acid glycosphingolipids from nonmucosal residue of human stomach of a blood group O(Rh-) individual, 40 µg. The solvent system was chloroform/methanol/water (60:35:8, by volume). Autoradiography was for 12 h. Fig. 1. Open in new tabDownload slide Binding of 35S-labeled Helicobacter pylori to glycosphingolipids separated on thin-layer plates. (A) Glycosphingolipids detected with anisaldehyde. (B) Glycosphingolipids detected by autoradiography after binding of radiolabeled H.pylori strain 1139. The assay was done as described in the Materials and methods section. The lanes contained the following glycosphingolipids: lane 1, non-acid glycosphingolipids from human blood group A erythrocytes, 40 µg; lane 2, non-acid glycosphingolipids from dog small intestine, 40 µg; lane 3, non-acid glycosphingolipids from guinea pig erythrocytes, 40 µg; lane 4, non-acid glycosphingolipids from mouse feces, 40 µg; lane 5, non-acid glycosphingolipids from human meconium, 40 µg; lane 6, non-acid glycosphingolipids from human stomach of a blood group A(Rh+)p individual, 40 µg; lane 7, non-acid glycosphingolipids from human stomach of a blood group A individual, 40 µg; lane 8, non-acid glycosphingolipids from human stomach of a blood group O individual, 40 µg; lane 9, non-acid glycosphingolipids from mucosal cells of human stomach of a blood group O(Rh-) individual, 40 µg; lane 10, non-acid glycosphingolipids from nonmucosal residue of human stomach of a blood group O(Rh-) individual, 40 µg. The solvent system was chloroform/methanol/water (60:35:8, by volume). Autoradiography was for 12 h. The lactosylceramide binding specificity was revealed by a selective binding to the di- and triglycosylceramide region of the non-acid fraction from dog small intestine (Figure 1B, lane 2), the triglycosylceramide region of the non-acid fraction from guinea pig erythrocytes (Figure 1B, lane 3), and to a slow-migrating compound of the non-acid fraction from mouse feces (Figure 1B, lane 4). The binding to these compounds was not influenced by the culture conditions, since binding was obtained both after growth in liquid media and on solid phase. The same binding pattern was obtained when 125I-labeled crude protein extracts from the bacterial cell surface were used (not shown). By binding of radiolabeled bacteria to pure and structurally defined glycosphingolipids on thin-layer plates (Figures 2, 3; summarized in Table I) the indicated lactosylceramide binding property of H.pylori was confirmed. Thus, the bacteria bound to lactosylceramide with sphingosine/phytosphingosine and 2D-hydroxy fatty acids (nos. 5–7 of Table I, lanes 2–5 of Figure 2, and lane 1 of Figure 3), but did not bind to lactosylceramide having sphingosine and nonhydroxy fatty acids (no. 4, lane 1 of Figure 2, and lane 2 of Figure 3). Substitution of the terminal galactose of lactosylceramide by an α-galactose in 3-position (isoglobotriaosylceramide; nos. 11 and 12 of Table I, and lanes 6 and 7 of Figure 2) was tolerated with retained binding activity. For isoglobotriaosylceramide a preferential binding to the molecular species with phytosphingosine and hydroxy fatty acid with 16 carbon atoms (no. 12 of Table I, lane 7 of Figure 6) was obtained. A β-N-acetylgalactosamine in 4-position of the terminal galactose of lactosylceramide (gangliotriaosylceramide; no. 14 of Table I and lane 3 of Figure 3) was also tolerated. However, some other substitutions, such as, for example, an α-galactose in 4-position of the galactose (globotriaosylceramide; nos. 9 and 10 of Table I, and lanes 8 and 9 of Figure 2), or an α-linked sialic acid in 3-position (GM3 ganglioside; nos. 16 and 17), abolished the binding. The gangliotriaosylceramide structure could be further elongated by a β-galactose in 3-position of the N-acetylgalactosamine (gangliotetraosylceramide; nos. 20 and 21 of Table I and lane 5 of Figure 3), and also by Fuca2Galβ3 (fucosylgangliotetraosylceramide; no. 25), without loss of binding activity. Fig. 2. Open in new tabDownload slide Binding of 35S-labeled Helicobacter pylori to different molecular species of lactosylceramide, isoglobotriaosylceramide and globotriaosylceramide on thin-layer chromatograms. (A) Glycosphingolipids detected with anisaldehyde reagent. (B) Glycosphingolipids detected by autoradiography after binding of 35S-labeled H.pylori strain 032. The assay was performed as described in Materials and methods. The lanes contained the following glycosphingolipids: lane 1, lactosylceramide from human granulocytes with d18:1–16:0 and 24:1, 2 µg; lane 2, lactosylceramide from human meconium with d18:1-h22:0, 2 µg; lane 3, lactosylceramide from dog intestine with t18:0-h24:0 and d18:1-h16:0, 2 µg; lane 4, lactosylceramide from dog intestine with d18:1-h16:0, 2 µg; lane 5, lactosylceramide from dog intestine with t18:0-h16:0, 2 µg; lane 6, isoglobotriaosylceramide from dog intestine with d18:1-h16:0 and t18:0-h24:0, 2 µg; lane 7, isoglobotriaosylceramide from dog intestine with t18:0-h16:0, 2 µg; lane 8, globotriaosylceramide from human erythrocytes with d18:1–16:0, 2 µg; lane 9, globotriaosylceramide from rabbit intestine with t18:0-h24:0, 2 µg. The solvent system was chloroform/methanol/water (65:25:4, by volume). Autoradiography was for 12 h. Fig. 2. Open in new tabDownload slide Binding of 35S-labeled Helicobacter pylori to different molecular species of lactosylceramide, isoglobotriaosylceramide and globotriaosylceramide on thin-layer chromatograms. (A) Glycosphingolipids detected with anisaldehyde reagent. (B) Glycosphingolipids detected by autoradiography after binding of 35S-labeled H.pylori strain 032. The assay was performed as described in Materials and methods. The lanes contained the following glycosphingolipids: lane 1, lactosylceramide from human granulocytes with d18:1–16:0 and 24:1, 2 µg; lane 2, lactosylceramide from human meconium with d18:1-h22:0, 2 µg; lane 3, lactosylceramide from dog intestine with t18:0-h24:0 and d18:1-h16:0, 2 µg; lane 4, lactosylceramide from dog intestine with d18:1-h16:0, 2 µg; lane 5, lactosylceramide from dog intestine with t18:0-h16:0, 2 µg; lane 6, isoglobotriaosylceramide from dog intestine with d18:1-h16:0 and t18:0-h24:0, 2 µg; lane 7, isoglobotriaosylceramide from dog intestine with t18:0-h16:0, 2 µg; lane 8, globotriaosylceramide from human erythrocytes with d18:1–16:0, 2 µg; lane 9, globotriaosylceramide from rabbit intestine with t18:0-h24:0, 2 µg. The solvent system was chloroform/methanol/water (65:25:4, by volume). Autoradiography was for 12 h. However, binding of H.pylori to the de-N-acylated gangliotriaosyl-and gangliotetraosylceramide structures did not occur (nos.15 and 22 of Table I, and Figure 3, lanes 4 and 6), indicating thatthe acetamido moiety of these compounds may be essential forbinding (see Discussion). Apart from lactosylceramide, no dependence of a certain ceramide composition for binding to occur was found for the compounds where species with both nonhydroxy and hydroxy ceramides were available for testing. Binding to gangliotetraosylceramide was obtained both for the species with sphingosine, and the 20 carbon analog of sphingosine, and nonhydroxy 18:0 fatty acid (no. 20 of Table I), and the species with phytosphingosine and hydroxy 16:0 and 24:0 fatty acids (no. 21 of Table I). On the other hand, no binding of the bacteria to globotriaosylceramide with phytosphingosine and hydroxy 24:0 fatty acid (no. 10 of Table I, and lane 9 of Figure 2), or NeuGc-GM3 ganglioside with sphingosine and hydroxy 16:0 and 24:0 fatty acids (no. 17 of Table I) was observed. Open in new tabDownload slide Open in new tabDownload slide Open in new tabDownload slide Open in new tabDownload slide Liposome assay In order to investigate whether the selective binding of lactosylceramide also persists when inserted into a membrane bilayer, the different forms of this glycosphingolipid, i.e., with sphingosine/ phytosphingosine and fatty acids with or without a 2D hydroxyl group, were incorporated into liposomes to which binding of four different strains of H.pylori was tested. Figure 4 shows typical results of aggregation between the bacterial strain NCTC 11638 and liposomes having lactosylceramide with sphingosine and nonhydroxy fatty acids (A) or liposomes having lactosylceramide with phytosphingosine/sphingosine and 2D-hydroxy fatty acids (B), and Table II summarizes the results for the different strains of H.pylori that were used. It should be noted that bacteria grown on solid media consistently self-aggregated (both in the absence and presence of liposomes devoid of lactosylceramide) in contrast to bacteria cultured in liquid media where no self-aggregation was found except occasionally for strain 032. The latter conditions were therefore used throughout for this assay From the findings given in Table II it is quite clear that a selectivity in favor of ceramide species having sphingosine/phytosphingosine and 2D-hydroxy fatty acids results when lactosylceramide is incorporated into a bilayer, suggesting that this is the case also in vivo. These results were extended by binding of H.pylori to purified lactosylceramide species from dog small intestine and infant rabbit small intestine incorporated into liposomes, showing that lactosylceramide with phytosphingosine and 2D-hydroxy fatty acids is the structure preferentially recognized by the bacteria (Table II). Furthermore, by varying the relative amounts of phosphatidylcholine and cholesterol, while keeping the total lipid content constant, it was found that the latter membrane component is vital in order to obtain maximal binding of H.pylori. Thus, in the absence of cholesterol only a weak binding, corresponding to one + on the scale used here, was found for binding of strain 002 to lactosylceramide from dog small intestine containing a mixture of ceramides species having sphingosine/phytosphingosine and hydroxy fatty acids. The binding then increased continuously up to a cholesterol content of ∼40% (mol) where it leveled off at +++(+). Binding of H.pylori to glycosphingolipids from human stomach epithelium At the initial stage only glycosphingolipids isolated from the whole human gastric wall were available. When these fractions were examined for H.pylori binding activity using the chromatogram binding assay, a binding of the bacteria to the diglycosylceramide region was obtained (Figure 1, lanes 6–8). However, since these binding-active diglycosylceramides might be derived from the nonepithelial part of the human stomach, acid and non-acid glycosphingolipids were subsequently isolated from epithelial scrapings from the fundus region of human stomach (obtained fresh at surgery) from seven individuals. In two cases glycosphingolipids were also isolated from the nonepithelial residues. The glycosphingolipid fractions obtained were tested for H.pylori binding activity by the chromatogram binding assay. No binding to glycosphingolipids from the nonepithelial residues (Figure 1, lane 10) or to the acid glycosphingolipids from the epithelial cells (not shown) was observed. In one of the seven non-acid glycosphingolipid fractions from the epithelial cells a distinct binding in the diglycosylceramide region was detected, as shown in Figure 1, lane 9. This fraction also contained a binding-active glycosphingolipid which migrated in the hepta-glycosylceramide region. Fig. 3. Open in new tabDownload slide Binding of Helicobacter pylori to pure glycosphingolipids on thin-layer plates. (A) Chemical detection with anisaldehyde. (B) Autoradiogram obtained by binding of 35S-labeled H.pylori strain 032. The assay was done as described in Materials and methods. Autoradiography was for 12 h. The lanes were: lane 1, Galβ4Glcβ1Cer from dog small intestine, 4 µg; lane 2, Galβ4Glcβ1Cer from human erythrocytes, 4 µg; lane 3, GalNAcβ4Galβ4Glcβ1Cer from guinea pig erythrocytes, 4 µg; lane 4, GalNH2β4Galβ4Glcβ1Cer, ∼4 µg; lane 5, Galβ3GalNAcβ4Galβ4Glcβ1Cer from mouse feces, 4 µg; lane 6, Galβ3GalNH2β4Galβ4Glcβ1Cer, ∼4 µg; lane 7, GalNAcβ3Galα4Galβ4Glcβ1Cer from human erythrocytes, 4 µg. Fig. 3. Open in new tabDownload slide Binding of Helicobacter pylori to pure glycosphingolipids on thin-layer plates. (A) Chemical detection with anisaldehyde. (B) Autoradiogram obtained by binding of 35S-labeled H.pylori strain 032. The assay was done as described in Materials and methods. Autoradiography was for 12 h. The lanes were: lane 1, Galβ4Glcβ1Cer from dog small intestine, 4 µg; lane 2, Galβ4Glcβ1Cer from human erythrocytes, 4 µg; lane 3, GalNAcβ4Galβ4Glcβ1Cer from guinea pig erythrocytes, 4 µg; lane 4, GalNH2β4Galβ4Glcβ1Cer, ∼4 µg; lane 5, Galβ3GalNAcβ4Galβ4Glcβ1Cer from mouse feces, 4 µg; lane 6, Galβ3GalNH2β4Galβ4Glcβ1Cer, ∼4 µg; lane 7, GalNAcβ3Galα4Galβ4Glcβ1Cer from human erythrocytes, 4 µg. Fig. 4. Open in new tabDownload slide Open in new tabDownload slide Photographs of typical aggregations of Helicobacter pylori and lactosylceramide-containing liposomes as seen by phase-contrast microscopy. Lactosylceramide from (A) human erythrocytes having sphingosine and nonhydroxy fatty acids and (B) dog small intestine having phytosphingosine/sphingosine and 2-D hydroxy fatty acids were incorporated into liposomes and allowed to aggregate with H.pylori strain NCTC 11638 as outlined in Materials and methods. Fig. 4. Open in new tabDownload slide Open in new tabDownload slide Photographs of typical aggregations of Helicobacter pylori and lactosylceramide-containing liposomes as seen by phase-contrast microscopy. Lactosylceramide from (A) human erythrocytes having sphingosine and nonhydroxy fatty acids and (B) dog small intestine having phytosphingosine/sphingosine and 2-D hydroxy fatty acids were incorporated into liposomes and allowed to aggregate with H.pylori strain NCTC 11638 as outlined in Materials and methods. EI mass spectrometry Mass spectra of permethylated and permethylated/reduced derivatives of the non-acid glycosphingolipid mixture from the epithelial cells of human stomach, with H.pylori binding activity in the diglycosylceramide region, are reproduced in Figure 5. Simplified interpretation formulae, representing the species with phytosphingosine and hydroxy 24:0 fatty acid, are shown above the spectra. In both spectra ions deriving from two different glycosphingolipids can be discerned. A terminal Hex is indicated by the ions at m/z 219 and 187 (219 minus 32), seen in both spectra, while the ion at m/z 423 in Figure 5B indicates a terminal Hex-Hex sequence. The ion at m/z 292 (291+1) is a rearrangement ion, containing the carbohydrate part and part of the fatty acid, stemming from monohexosylceramide, and the corresponding ion of dihexosylceramide can be seen at m/z 496 (495+1) in Figure 5A. Fig. 5. Open in new tabDownload slide Open in new tabDownload slide (A) EI mass spectrum of permethylated non-acid glycosphingolipids isolated from mucosal scrapings of the fundus region of human stomach. Above the spectrum a formula for interpretation, representing the one- and two-sugar species with phytosphingosine and 2-D hydroxy 24:0 fatty acid, is shown. The analytical conditions were: sample amount, 10 µg; electron energy, 70 eV; trap current, 200 µA and acceleration voltage, 8 kV. Starting at 200°C, the temperature was increased by 6°C/min. The spectrum was recorded at 350°C. (B) EI mass spectrum of permethylated and LiAlH4-reduced non-acid glycosphingolipids isolated from mucosal scrapings of the fundus region of human stomach. Above the spectrum a formula for interpretation, representing the one- and two-sugar species with phytosphingosine and 2-D hydroxy 24:0 fatty acid, is shown. The analytical conditions were: sample amount, 10 µg; electron energy, 70 eV; trap current, 500 µA and acceleration voltage, 8 kV. Starting at 240°C, the temperature was increased by 6°C/min. The spectrum was recorded at 280°C. Fig. 5. Open in new tabDownload slide Open in new tabDownload slide (A) EI mass spectrum of permethylated non-acid glycosphingolipids isolated from mucosal scrapings of the fundus region of human stomach. Above the spectrum a formula for interpretation, representing the one- and two-sugar species with phytosphingosine and 2-D hydroxy 24:0 fatty acid, is shown. The analytical conditions were: sample amount, 10 µg; electron energy, 70 eV; trap current, 200 µA and acceleration voltage, 8 kV. Starting at 200°C, the temperature was increased by 6°C/min. The spectrum was recorded at 350°C. (B) EI mass spectrum of permethylated and LiAlH4-reduced non-acid glycosphingolipids isolated from mucosal scrapings of the fundus region of human stomach. Above the spectrum a formula for interpretation, representing the one- and two-sugar species with phytosphingosine and 2-D hydroxy 24:0 fatty acid, is shown. The analytical conditions were: sample amount, 10 µg; electron energy, 70 eV; trap current, 500 µA and acceleration voltage, 8 kV. Starting at 240°C, the temperature was increased by 6°C/min. The spectrum was recorded at 280°C. Two series of immonium ions are present in the spectrum of the permethylated and reduced derivative (Figure 5B). These ions, formed by loss of part of the long-chain base, give information about the carbohydrate sequence and fatty acid composition. The series at m/z 546–658 originates from a monohexosyl in combination with h16:0 to h24:0 fatty acids, while the ions at m/z 805, 833, and 863 indicate a dihexosyl combined with 22:0, 24:0, and h24:0 fatty acids, respectively. The immonium ions at m/z 844 and 876 in the spectrum of the permethylated sample (Figure 5A) indicate a Hex-Hex saccharide together with 24:1 and h24:0 fatty acids, while the ions at m/z 892 (1133 minus 241) and 920 (1161 minus 241) show the presence of t18:0 long-chain base combined with h22:0 and h24:0 fatty acids. Molecular ions of monohexosylceramide with d18:1-h24:0, t18:0-h24:1, and t18:0-h24:0 ceramides are seen at m/z 911, 941 and 943 in Figure 5B. The ions at m/z 986, 1048, 1096, 1128, and 1158 in the spectrum of the permethylated derivative (Figure 5A) are molecular ions of dihexosylceramide with d18:1–16:0, t18:0-h16:0, d18:1–24:1, d18:1-h24:0, and t18:0-h24:0, respectively, while molecular ions of dihexosylceramide with d18:1–16:0, t18:0-h16:0, t18:0-h18:0, and t18:0-h24:1 fatty acids are found at m/z 973, 1035, 1063, and 1145 in the spectrum of the reduced derivative (Figure 5B). Additional information about the ceramide composition is given by the series of ions at m/z 548, 580, 610, 692, 720, and 750 (Figure 5A), demonstrating the presence of d18:1–16:0, t18:0–16:0, t18:0-h16:0, t18:0–24:0, t18:0-h24:1, and t18:0–26:0 ceramides. In mass spectra of the permethylated derivative recorded at higher temperatures (not shown) carbohydrate sequence ions denoting a Hex-HexN-Hex-Hex sequence are found at m/z 219 and 187 (219 minus 32), 464, 668, and 872. A Hex-(Fuc-)HexN-Hex-Hex structure is indicated by ions at m/z 189 and 157 (189 minus 32), 219 and 187 (219 minus 32), 638, 842, and 1046. From the relative intensities of the ceramide ions, immonium ions and molecular ions, in the spectra of both derivatives, it was concluded that the major ceramides of both compounds were d18:1–24:0, d18:1-h24:0, t18:0–24:0, t18:0-h24:0, and t18:0-h24:1. Fig. 6. Open in new tabDownload slide Molecular models of lactosylceramide having phytosphingosine and 2-D hydroxy fatty acid. The minimum energy conformers shown are: lactosylceramide according to (A) conformer no. 4 (extended conformation), (B) conformer no. 5 ("cerebroside" conformation), and (C) conformer no. 7 (Glc 6-OH - fatty acid 2-O′ hydrogen bond) as detailed in Table III. In (B) the lactose moiety is pointing away from the viewer. Fig. 6. Open in new tabDownload slide Molecular models of lactosylceramide having phytosphingosine and 2-D hydroxy fatty acid. The minimum energy conformers shown are: lactosylceramide according to (A) conformer no. 4 (extended conformation), (B) conformer no. 5 ("cerebroside" conformation), and (C) conformer no. 7 (Glc 6-OH - fatty acid 2-O′ hydrogen bond) as detailed in Table III. In (B) the lactose moiety is pointing away from the viewer. Thus, by mass spectrometry it was demonstrated that the glycosphingolipid fraction with H.pylori binding activity from human gastric epithelial cells contained dihexosylceramides with both sphingosine and phytosphingosine, and both hydroxy and non-hydroxy fatty acids. In addition, monohexosylceramide and glycosphingolipids with Hex-HexN-Hex-Hex and Hex-(Fuc-) HexN-Hex-Hex sequences were found, while the identity of the H.pylori-binding compound with mobility in the heptaglycosylceramide region could not be established. Conformational aspects The results from the liposome binding assay suggest that the selective H.pylori binding to one of the three forms of lactosylceramide investigated stems from the presence of the fatty acid 2-D hydroxyl group and the 4-D hydroxyl group of phytosphingosine. They could also be indicative of a requirement for hydrogen bond interactions between the fatty acid 2-D hydroxyl group and the Glc 2-OH or 6-OH for a correct binding epitope presentation (see Figure 6) and/or a direct involvement of the fatty acid 2-D hydroxyl group in the binding process. It is at present not known which parts of the lactosyl moiety that constitute the epitope, but the requirement of the terminal Gal at the disaccharide level is clearly indicated by the nonbinding of both Glcβ1Cer and Galβ1Cer (nos. 1 and 2 of Table I). Parts of the internal Glc are probably also involved in the epitope since this residue is quite exposed in the conformation most likely responsible for a correct presentation of the epitope as discussed below. A previous molecular modeling study of Glcβ1Cer revealed that nine different low energy conformers are obtainable by varying the dihedral angles of the Glcβ1Cer linkage (see Table III) and that the presence of a 2-D hydroxyl group on the fatty acid has very little influence on the population distribution among these conformers (Nyholm and Pascher, 1993). However, these calculations were performed in vacuo and need not reflect the situation in vivo, particularly when considering that the presence of cholesterol in the liposome assay was necessary for binding of H.pylori to occur. Furthermore, it is not uncommon that a conformer representing a minority species may be responsible for binding (Imberty et al., 1993). It could, moreover, be argued that all conformers having a lactose–ceramide angle below a certain value, when the glycosphingolipid is membrane-bound, would be disallowed due to steric interference between the lactose part and the membrane surface (Nyholm and Pascher, 1993). Thus, on this ground conformers 6, 8, and 9, where the lactosyl moiety is at an angle of 90° or more relative to the membrane normal, would most likely have to be excluded as candidates responsible for the selective H.pylori binding, even in the presence of cholesterol. Open in new tabDownload slide Open in new tabDownload slide Aid in determining which of the remaining conformers that is responsible for the observed preferential binding of H.pylori to lactosylceramide having phytosphingosine and hydroxy fatty acids may be obtained from the binding or nonbinding of the trisaccharide glycosphingolipids given in Table I. Figure 7 shows isoglobotriaosylceramide (top row), globotriaosylceramide (middle row), and lactotriaosylceramide (bottom row) with Glcβ1Cer conformations according to conformer nos. 1, 2, 5, and 7 in Table III. Comparison of the three glycosphingolipids in these four conformations suggests that conformer no. 7, and possibly also conformer no. 5, displays a conformation compatible with both access to the lactose epitope and the actual binding results. Thus, in the isoglobotriaosylceramide structure having Glcβ1Cer conformations 1 or 2 the terminal Galα3 residue most likely blocks the binding epitope for an adhesin approaching from above, contrary to what would be expected from binding data, whereas in conformer no. 7, and to a lesser degree also conformer no. 5, the epitope is accessible (Figure 7A–D). In globotriaosylceramide the terminal Gala4 residue does not block access to the binding epitope in the no. 1 and no. 2 conformations, but does so in the no. 5 and no. 7 conformations in accordance with binding data (Figure 7E–H). For lactotriaosylceramide binding of H.pylori appears to be blocked by the terminal GlcNAcβ3 in either of the conformations shown in accordance with binding data (Figure 7I–L). Similar comparisons for the remaining Glcβ1Cer conformations given in Table III (nos. 3 and 4) show that neither of these are compatible with both access to the binding epitope and the binding results. H.pylori does bind to gangliotriaosylceramide, but since this binding was concluded to be dependent on the presence of an intact acetamido moiety of the terminal GalNAcβ4 residue (see nos. 14 and 15 in Table I), the question of a lactosylceramide-based binding is irrelevant. Discussion The ability to interact with lactosylceramide, and related structures, has been described for a large number of bacteria of various origins (Karlsson, 1989). This list of lactosylceramide-binding bacteria may be further extended by the results of the present study, where a lactosylceramide-related binding of the human gastric pathogen H.pylori was demonstrated by binding of 35S-labeled bacteria to lactosylceramide and isoglobotriaosylceramide on thin-layer plates. Binding of gangliotriaosylceramide, gangliotetraosylceramide and fucosyl-gangliotetraosylceramide, on the other hand, was by de-N-acylation of the former two structures shown to be dependent on the presence of the acetamido moiety and thus should not be included within the lactosylceramide binding specificity. Furthermore, the dominating ceramide species of gangliotriaosylceramide contain nonhydroxy fatty acids, arguing that a lactosylceramide-based recognition definitely can be ruled out in this case. This is further supported by the gangliotetraosylceramide binding, which is independent of the degree of hydroxylation. The appearance of a positive charge for the de-N-acylated compounds may augment the nonbinding characteristics of these two structures, but loss of interactions with the acetamido moiety are likely to be mainly responsible for the observed results as judged by the effects of de-N-acylation on globoside binding to Verotoxin VT2e (Nyholm et al., 1996). Of the set of glycosphingolipids earlier included in the lactosylceramide-related binding pattern (Karlsson, 1989), lactosylceramide itself is the only relevant receptor candidate for human infections since isoglobotriaosylceramide, a major non-acid glycosphingolipid of dog small intestine (Hansson et al., 1983), has not been identified in human tissues. The biological significance of the binding of H.pylori to lactosylceramide is suggested by binding of these bacteria to the diglycosylceramide region of the non-acid glycosphingolipid fraction from the epithelial cells of human stomach, found in one of seven samples examined. By mass spectrometry it was demonstrated that this fraction contained diglycosylceramides with sphingosine and phytosphingosine and 2-D hydroxy fatty acids with 16–24 carbon atoms. Although no H.pylori binding activity was detected in the non-acid glycosphingolipid fractions isolated from the stomach epithelium of the other six individuals, EI mass spectrometry after permethylation of two of these fractions demonstrated the presence of diglycosylceramides with phytosphingosine and 2-D hydroxy fatty acids as minor species, while the main diglycosylceramides had sphingosine and nonhydroxy 16:0 and 24:0 fatty acids (not shown). Thus, in these cases the relative amounts of the relevant lactosylceramide were in all likelihood too low to permit binding of H.pylori in the chromatogram binding assay. Fig. 7. Open in new tabDownload slide Molecular models of isoglobotriaosylceramide, globotriaosylceramide and lactotriaosylceramide having different Glcβ1Cer conformations. The top row shows minimum energy conformers of isoglobotriaosylceramide (A–D) having the same Glcβ1Cer conformations as conformers 1, 2, 5, and 7 in Table III while the middle and bottom rows show corresponding conformers for globotriaosylceramide (E–H) and lactotriaosylceramide (I–L), respectively. The Galα3Gal and Galα4Gal glycosidic dihedral angles (Φ/Ψ.) in (A–D) and (E–H) were found to be -47/-60 and -20/48, respectively. The GlcNAcβ3Gal glycosidic dihedral angles in (I–L) are 63/55, corresponding to the global minimum energy conformer in which the carbonyl oxygen of the GlcNAcβ3 acetamido group forms a hydrogen bond with the Gal 4-OH and where the ring oxygen of GlcNAcβ3 forms a hydrogen bond with the Gal 2-OH. The energetically closest conformer is found 0.2 kcal/mol above the global minimum (Φ/Ψ = 32/-60) with hydrogen bonds between the Gal 2-OH and 4-OH to the carbonyl and ring oxygens of GlcNAcβ3, respectively. Adoption of this or other higher energy conformations does not alter the conclusions made in the Results section. Fig. 7. Open in new tabDownload slide Molecular models of isoglobotriaosylceramide, globotriaosylceramide and lactotriaosylceramide having different Glcβ1Cer conformations. The top row shows minimum energy conformers of isoglobotriaosylceramide (A–D) having the same Glcβ1Cer conformations as conformers 1, 2, 5, and 7 in Table III while the middle and bottom rows show corresponding conformers for globotriaosylceramide (E–H) and lactotriaosylceramide (I–L), respectively. The Galα3Gal and Galα4Gal glycosidic dihedral angles (Φ/Ψ.) in (A–D) and (E–H) were found to be -47/-60 and -20/48, respectively. The GlcNAcβ3Gal glycosidic dihedral angles in (I–L) are 63/55, corresponding to the global minimum energy conformer in which the carbonyl oxygen of the GlcNAcβ3 acetamido group forms a hydrogen bond with the Gal 4-OH and where the ring oxygen of GlcNAcβ3 forms a hydrogen bond with the Gal 2-OH. The energetically closest conformer is found 0.2 kcal/mol above the global minimum (Φ/Ψ = 32/-60) with hydrogen bonds between the Gal 2-OH and 4-OH to the carbonyl and ring oxygens of GlcNAcβ3, respectively. Adoption of this or other higher energy conformations does not alter the conclusions made in the Results section. The lactosylceramide-binding bacteria were the first case where the requirement of certain ceramide species for binding to a glycosphingolipid to occur was described previously (Strömberg et al., 1988; Strömberg and Karlsson, 1990a,b). However, recently a similar preferential binding to glycosphingolipids with certain ceramide species has been described for other microbial systems, i.e., the binding of Escherichia coli K99 to NeuGc-GM3 (Yuyama et al., 1993) and verotoxins to globotriaosylceramide (Kiarash et al., 1994). Antibody binding to tumor-associated murine gangliotriaosylceramide represents another example where a ceramide dependence linked to crypticity has been suggested (Kannagi et al., 1983). Although the lactosylceramide-binding property is often found when binding of different bacteria to glycosphingolipids is examined (Karlsson, 1989), there are exceptions. No binding to lactosylceramide has, for example, been observed for E.coli with K99(Teneberg et al., 1994) or P-fimbriae (Bock et al., 1985). Among the very complex factors governing the availability of glycosphingolipid binding epitopes are lipid composition, cholesterol content, and presence of further natural membrane constituents such as proteins and other glycolipids which determine such phenomena as lateral movement, patching and crypticity (Boggs, 1987; Maggio, 1994). Of these, cholesterol has in several systems been found to have a dramatic effect on the glycosphingolipid epitope accessibility (Brûlet and McConnell, 1977; Balakrishnan et al., 1982; Utsumi et al., 1984), in line with our observation that H.pylori only discriminates between sphingosine- and phytosphingosine-containing lactosylceramides in the presence of this substance in liposomes. However, this discrimination is not found when bacteria are allowed to bind to lactosylceramide on thin-layer chromatogram plates, probably due to less conformational restrictions around the Glcβ1Cer linkage. The molecular interactions responsible for the effects of cholesterol are at present not well understood (Boggs, 1987). However, a 2H NMR study of the effect of fatty acid hydroxylation of sphingosine-containing galactosylceramide, incorporated into liposomes lacking cholesterol, revealed insignificant differences in average motional behavior of the sugar residue when comparing hydroxy and nonhydroxy forms, thus indicating similar conformational preferences (Singh et al., 1992). Similar results were later found in the presence of cholesterol (Morrow et al., 1995). Furthermore, the average motional behavior of the oligosaccharide part of larger glycosphingolipids, such as globoside, gangliotetraosylceramide, and GM1, has also been found to be perturbed only to a minor degree by the presence of cholesterol (Singh et al., 1995), suggesting that this may be the case also for lactosylceramide. Removal of the 4,5 double bond in sphingosine-containing lactosylceramide, on the other hand, has been found to lead to a reduced mobility at the C-2 level of the nonhydroxy fatty acid (Fenske et al., 1991). This indicates that an orientational change involving the fatty acid 2-D hydroxyl group could be present also in lactosylceramide containing phytosphingosine which possibly could favor hydrogen bond interactions with the Glc 6-OH. The fact that a rather high percentage of cholesterol (∼40%) is required for maximal binding of H.pylori to lactosylceramide having phytosphingosine and hydroxy fatty acids, coupled to the low amount of incorporated glycosphingolipid (1%), could indicate that glycosphingolipid interactions with the phosphatidylcholine lipid heads are diminished to such a degree that the binding epitope is allowed to be presented correctly. A plausible role for the 4-D hydroxyl group of phytosphingosine would then be to enhance the sterol–ceramide interactions. It is not likely, however, that the 4-D hydroxyl group interacts directly with the adhesin since binding was obtained in the thin-layer chromatogram assay for lactosylceramide with sphingosine and C-16 hydroxy fatty acid (no. 5 of Table I). As regards the 2-D hydroxyl group of the fatty acid, it cannot be excluded that it may, apart from providing hydrogen bond interactions with the Glc 6-OH in the conformation suggested to be responsible for binding (Figure 6C), interact directly with the adhesin although being partly shielded by the glucose residue. In the alternative conformation (Figure 6B) the 2-D hydroxyl group is unshielded and would in this case have to interact directly with the adhesin in order to explain the crucial role of this moiety. The restricted tissue and cell tropism of H.pylori has been well described previously (Bode et al., 1988; Falk et al., 1993). In contrast, lactosylceramide with phytosphingosine and 2-D hydroxy fatty acids is found in the gastrointestinal tract of several species, e.g., the epithelial cells of dog (Hansson et al., 1983) and rat (Breimer et al., 1982) small intestine. Although absent in the epithelial cells of human small intestine (Björk et al., 1987), lactosylceramide with phytosphingosine and 2-D hydroxy fatty acids is found in small amounts in the epithelial cells of human colon (Holgersson et al., 1991). This broad distribution makes lactosylceramide with the relevant ceramide composition an unlikely determinant of H.pylori tissue specificity. Furthermore, the lactose sequence has so far only been identified in glycosphingolipids. Target tissue specificity is thus most likely achieved by binding of the microorganisms to other specific receptors on glycoproteins and glycosphingolipids. Once tissue-specific binding has been established, however, a second-step binding to lactosylceramide may occur, conferring a more membrane-close attachment and leading to increased local concentrations of bacterial toxins and other injurious factors at the cell membrane surface (Beachey, 1981). Materials and methods Bacterial strains, culture conditions, and labeling During this study five H.pylori strains, 1139, 17874, and 17875 (obtained from Culture Collection, University of Göteborg (CCUG), Sweden; strains 17874 and 17875 are identical with strains 11637 and 11638, respectively, from the National Collection of Type Cultures (NCTC), London) and the clinical isolates 002 (isolated from a gastritis case) and 032 (isolated from a duodenal ulcer case), were used in parallel. The strains were stored at -80°C in tryptic soy broth containing 15% glycerol (by volume). The bacteria were initially cultured on GAB-CAMP agar (Soltesz et al., 1988) in a humid (98%) microaerophilic atmosphere (O2: 5–7%, CO2: 8–10%, N2: 83–87%, and H2: less than 0.5%) at 37°C for 48–72 h. For radiolabeling, colonies were inoculated on GAB-CAMP agar and 50 µCi 35S-methionine (Amersham, U.K.) diluted in 0.5 ml phosphate-buffered saline (PBS), pH 7.3, was sprinkled over the plates. After incubation for 12–24 h at 37°C under microaerophilic conditions, the cells were scraped off, washed three times with PBS, and resuspended to 1 × 108 CFU/ml in PBS. Alternatively, colonies were inoculated (1 × 105 CFU/ml) in Ham's F12 medium (Gibco BRL, U.K.), supplemented with 10% heat-inactivated fetal calf serum (Sera-Lab) and, for labeling, with 50 µCi 35S-methionine per 10 ml medium, and incubated with shaking under microaerophilic conditions for 24 h. Bacterial cells were harvested by centrifugation and purity of the cultures as well as a low content of coccoidal forms was ensured by phase-contrast microscopy. After two washes with PBS, the cells were resuspended to 1 × 108 CFU/ml in PBS. Both labeling procedures resulted in suspensions with specific activities of approximately 1 c.p.m. per 100 H.pylori organisms. For use in the liposome binding assay unlabeled bacteria were resuspended to 1 × 109 CFU/ml in PBS. Preparation of crude bacterial cell surface protein extracts Before the extraction procedure, H.pylori strains NCTC 11637, NCTC 11638, and 1139 were cultured on horse blood agar for 2–3 days under microaerophilic conditions, harvested, and washed once with PBS. The bacterial cells were then incubated with 1 M LiCl in PBS at 45°C for 2 h (Lelwala-Guruge et al., 1992), followed by centrifugation. The supernatants obtained were dialyzed overnight against PBS. The protein concentrations of the extracts were 0.3–1.5 mg/ml, as determined by using an acidic solution of Coomassie brilliant blue G-250 dye reagent (Bio-Rad, Richmond, U.K.). From each extract aliquots of approximately 0.1 mg protein were taken out and labeled with 125I by the Iodogen method (Aggarwal et al., 1985) to a specific activity of 2–5 × 103 c.p.m./µg. Glycosphingolipid preparations Acid and non-acid glycosphingolipid fractions were isolated by the previously reported method (Karlsson, 1987). The pure non-acid glycosphingolipids used in the binding studies were in general obtained by acetylation of the non-acid fractions (Handa, 1963), followed by repeated separations on Iatrobeads (6RS-8060, Iatron Laboratories, Tokyo, Japan) columns. The acid glycosphingolipids were isolated by DEAE-Sepharose column chromatography, followed by separations on Iatrobeads columns. The isolated glycosphingolipids were characterized by mass spectrometry (Samuelsson et al., 1990), 1H NMR spectroscopy (Falk et al., 1979a–c; Koerner et al., 1983), and degradation studies (Yang and Hakomori, 1971; Stellner et al., 1973). For isolation of individual ceramide species of lactosylceramide, 25 mg of total non-acid glycosphingolipids from the small intestine of a 10-day-old rabbit was acetylated, and separated on a 25 g Iatrobeads column eluted with a linear gradient of methanol in chloroform (0–2%, by volume). Each 2 ml fraction was analyzed by thin-layer chromatography (see below), and pooling of the fractions containing glycosphingolipids with mobility in the diglycosylceramide region yielded 9.3 mg. This material was further separated on a 10 g Iatrobeads column eluted with a linear gradient of methanol in chloroform (0–1.5%, by volume). Pooling of diglycosylceramide-containing fraction was followed by deacetylation, and the 5.8 mg thus obtained was separated on a 10 g Iatrobeads column eluted with a linear gradient of chloroform/methanol/water (85:15:1 to 60:30:5, by volume). Thus, 0.6 mg of lactosylceramide with sphingosine and nonhydroxy 16:0–24:0 fatty acids, and 0.5 mg of lactosylceramide with phytosphingosine and hydroxy 16:0–24:0 fatty acids (according to EI mass spectrometry of permethylated derivatives) were obtained. Lactosylceramide with sphingosine and hydroxy 16:0 fatty acid was isolated from 200 mg of total non-acid glycosphingolipids from epithelial scrapings of dog small intestine, which were separated with a Beckman HPLC system (model Prep-350 LC System, Beckman Instruments Inc., San Ramon, CA) with a 5.0×50 cm silica column (YMC Micron 60A Silica, 25/40 µm particles, Skandinaviska Genetec, Kungsbacka, Sweden). The column was equilibrated in chloroform/methanol/water (80:20:1, by volume; Solvent A), and eluted (100 ml/min) with linear gradients of chloroform/methanol/water (40:40:12, by volume; Solvent B) in Solvent A. The percentage of Solvent B in Solvent A was raised from 0% to 25% during 5 min, from 25% to 45% during 23 min, and from 45% to 100% during 22 min. Each 250 ml fraction was analyzed by thin-layer chromatography, using anisaldehyde staining for detection of carbohydrates (see below). The diglycosylceramides eluted in tubes 7–10, and tube 10 contained mainly diglycosylceramides with phytosphingosine and hydroxy 16:0 to 24:0 fatty acids, and sphingosine and hydroxy 16:0 fatty acids. This fraction (6.4 mg) was further separated by HPLC (model Beckman System Gold) on a 1.0×25 cm silica column (Kromasil 5, 5 µm particles, Skandinaviska Genetec) eluted (2 ml/min) with chloroform/methanol/water (80:20:0.5, by volume) over 180 min. Aliquots from each 2 ml fraction were analyzed by thin-layer chromatography and anisaldehyde staining, and the diglycosylceramide-containing fractions were analyzed by negative ion FAB mass spectrometry. Thus, tube 12 contained diglycosylceramide with both phytosphingosine and hydroxy 24:0 fatty acid, and sphingosine and hydroxy 16:0 fatty acid, while pure diglycosylceramide with sphingosine and hydroxy 16:0 fatty acid eluted in tube 13, and diglycosylceramide with mainly phytospingosine and hydroxy 16:0 fatty acid eluted in tube 14. Thin-layer chromatography Thin-layer chromatography was performed on glass- or aluminum-backed silica gel 60 HPTLC plates (Merck, Darmstadt, Germany), using chloroform/methanol/water (60:35:8, by volume) or chloroform/methanol/water (65:25:4, by volume) as solvent systems. Chemical detection was accomplished by the anisaldehyde reagent (Waldi, 1962). De-N-acylation The GalNAcβ4 residue in gangliotriaosylceramide and gangliotetraosylceramide was selectively de-N-acylated by treatment with anhydrous hydrazine (Keith et al., 1978). The pertinent glycosphingolipid (200 µg) was dissolved in 300 µl of freshly distilled anhydrous hydrazine (Pierce) and sonicated for 30 s, whereafter the reaction was allowed to proceed for 72 h at 76°C. The hydrazine was subsequently removed using N2(g) at 40°C, followed by two cycles of redissolution in toluene and evaporation prior to redissolution of the residue in approximately 100 µl of methanol with the aid of sonication. After addition of water (10 ml) the solution was passed through a 500 mg C18 Extract-Clean column (Alltech Assoc., Inc.) prewashed in chloroform and chloroform/methanol. The column was then washed with 5 ml of water followed by elution with methanol and chloroform/methanol. The reaction products in the gangliotriaosylceramide case were verified by negative ion FAB mass spectrometry and 1H NMR and for gangliotetraosylceramide using the former technique only (not shown). The ceramide fatty acyl chain is hydrolyzed more slowly than the acetamido group under the conditions described above probably due to the fact that this part of the molecule is less exposed to the polar reagent. However, deacylation of the acetamido group is also dependent upon accessibility since the reaction went to completion for gangliotriaosylceramide but only 30% of gangliotetraosylceramide was converted to the corresponding amine. The product mixture of the latter glycosphingolipid was therefore chromatographed on a HPTLC plate and the amine band was scraped off and extracted from the silica gel with chloroform/methanol/water (40:40:12, by volume). Chromatogram binding assay The chromatogram binding assay was performed as described previously (Hansson et al., 1985). Glycosphingolipid mixtures (20–40 µg/lane) or pure compounds (0.1–4 µg/lane) were chromatographed on aluminum-backed silica gel plates. The chromatograms were treated with 0.3–0.5% polyisobutylmethacrylate in diethylether/n-hexane (1:3, by volume) for 1 min, dried, and subsequently soaked in PBS containing 2% bovine serum albumin and 0.1% NaN3 (Solution A) for 2 h. A suspension of radiolabeled bacteria (diluted in PBS to 1 × 108 CFU/ml and 1–5 × 106 c.p.m./ml) was thereafter sprinkled over the chromatograms and incubated for 2 h, followed by rinsings with PBS. After drying, the chromatograms were exposed to XAR-5 x-ray films (Eastman Kodak Co., Rochester, NY) for 12–72 h. For the binding assays using bacterial surface proteins, blocking was done with Sol. A supplemented with 0.1% Tween 20 (w/v) for 2 h. The chromatograms were then covered with 125I-labeled protein extracts diluted in PBS containing 0.1% Tween 20 (∼2 × 106 c.p.m./ml) for 2 h, washed, and autoradiographed. Liposome assay Large unilamellar liposomes were prepared according to the "reverse phase evaporation. method (Szoka and Papahadjopoulos, 1978) as follows (Magnusson et al., 1982): ∼20 µmol lipid (total content) dissolved in chloroform was dried by N2(g) and subsequently dissolved in 15 ml of chloroform/diethylether (1:1, by volume). To this solution 4 ml of PBS was added and emulsified by sonication for 3 min at room temperature in a sonicator bath. The organic phase was then evaporated off, also at room temperature, in a rotary evaporator in order for the liposomes to be formed. The liposomes had the following standard composition: glycosphingolipid/egg phosphatidylglycerol/cholesterol/egg phosphatidylcholine 0.2:0.5:10:10 (mol/ mol). The three latter compounds were all obtained from Sigma. Aggregation of bacteria with liposomes was achieved in micro-titer wells by adding 100 µl of the bacterial suspension (1 × 109 CFU/ml) to 50 µl of the liposome suspension containing 1% (mol) glycosphingolipid. Samples were allowed to stand at room temperature for 1 h whereafter the extent of aggregation was evaluated by eye and photographically using a light microscope. Experiments were also performed using liposomes prepared with varying relative amounts of cholesterol and phosphatidylcholine while keeping the total amount lipid constant. Mass spectrometry Before mass spectrometry the non-acid glycosphingolipid fraction from mucosal cells of human stomach was permethylated (Larson et al., 1987), and part of the permethylated sample was reduced using LiAlH4 in diethyl ether (Karlsson, 1974). Mass spectra were obtained with a ZAB 2F/HF mass spectrometer (VG Analytical, Manchester, U.K.) using the in beam technique (Breimer et al., 1980). The analytical conditions are given in the figure captions. Negative ion FAB mass spectra were acquired on a JEOL SX 102A (JEOL, Tokyo, Japan) using Xe atom bombardment at 8 kV and triethanolamine as matrix. Molecular modeling Minimum energy conformations of the glycosphingolipids used in this study were calculated within the Biograf molecular modeling program (Molecular Simulations Inc.) using the Dreiding-II force field (Mayo et al., 1990) on a Silicon Graphics Inc. 4D/35TG workstation. Partial atomic charges were generated using the charge equilibration method (Rappé and Goddard, 1991), and a distance dependent dielectric constant ε = 3.5r was used for the Coulomb interactions. In addition a special hydrogen bonding term was used in which the maximal interaction (Dhb) was set to -4 kcal/mol (Mayo et al., 1990). The conformation of the lactose moiety was not varied, since according to NMR the 3-OH of Glc predominantly forms a hydrogen bond with 5-O of Gal (Poppe et al., 1990), and the Galβ4Glc glycosidic dihedral angles were set to Φ/Ψ = 41/-12. The Φ/Ψ/θ dihedral angles of the Glcβ1Cer linkage of the three different lactosylceramide species investigated were adopted according to Nyholm and Pascher (1993) yielding nine different low energy conformers for each species as shown in Table III for lactosylceramide having phytosphingosine and hydroxy fatty acid. The glycosidic dihedral angles of the terminal sugar of the four different trisaccharide glycosphingolipids investigated here were found to be close to those found by Bock et al. (1985) and Imberty et al. (1991). Acknowledgments This work was supported by grants from the Swedish Medical Research Council (Nos. 3967, 10435 and 16x-04723 (T.W.)) and from Symbicom Ltd. Abbreviations Abbreviations The glycosphingolipid nomenclature follow the recommendations by the IUPAC-IUB Commission on Biochemical Nomenclature (CBN for Lipids:Eur. J. Biochem. (1977) 79, 11–21; J. Biol. Chem. (1982) 257, 3347–3351; and J. Biol. Chem. (1987) 262, 13–18). It is assumed that Gal, Glc, GlcNAc, GalNAc, and NeuAc are of the d-configuration, Fuc of the l-configuration, and that all sugars are present in the pyranose form. The abbreviations used are: CFU colony forming units Hex hexose HexN N-acetyl-hexosamine EI electron ionization FAB fast atom bombardment In the shorthand nomenclature for fatty acids and bases, the number before the colon refers to the carbon chain length and the number after the colon gives the total number of double bonds in the molecule. Fatty acids with a 2-D hydroxy group are denoted by the prefix h before the abbreviation, e.g., h16:0. For long chain bases, d denotes dihydroxy and t trihydroxy. Thus, d18:1 designates sphingosine (1,3-dihydroxy-2-aminooctadecene) and t18:0 phytosphingosine (1,3,4-trihydroxy-2-aminooctadecene). References Aggarwal B.B. , Eessalu T.E. , Hass P.E. . Characterization of receptors for human tumour necrosis factor and their regulation by γ-interferon , Nature , 1985 , vol. 318 (pg. 665 - 667 ) Google Scholar Crossref Search ADS PubMed WorldCat Balakrishnan K. , Mehdi S.Q. , McConnell H.M. . Availability of dinitrophenylated lipid haptens for specific antibody binding depends on the physical properties of host bilayer membranes , J. Biol. Chem. , 1982 , vol. 257 (pg. 6434 - 6439 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Beachey E.H. . Bacterial adherence: adhesin-receptor interactions mediating the attachment of bacteria to mucosal surfaces , J. Infect. Dis. , 1981 , vol. 143 (pg. 325 - 345 ) Google Scholar Crossref Search ADS PubMed WorldCat Björk S. , Breimer M.E. , Hansson G.C. , Karlsson K.-A. , Leffler H. . Structures of blood group glycosphingolipids of human small intestine. A relation of the expression of fucolipids of epithelial cells and the ABO, Le and Se phenotype of the donor , J. Biol. Chem. , 1987 , vol. 262 (pg. 6758 - 6765 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Bock K. , Breimer M.E. , Brignole A. , Hansson G.C. , Karlsson K.-A. , Larson G. , Leffler H. , Samuelsson B.E. , Strömberg N. , Svanborg-Edén C. , Thurin J. . Specificity of binding of a strain of uropathogenic Escherichia coli to Galα1→4Gal-containing glycosphingolipids , J. Biol. Chem. , 1985 , vol. 260 (pg. 8545 - 8551 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Bode G. , Malfertheiner P. , Ditschuneit H. . Pathogenic implications of ultrastructural findings in Campylobacter pylori related gastroduodenal disease , Scand. J. Gastroenterol. , 1988 , vol. 23 142 (pg. 25 - 39 ) Google Scholar Crossref Search ADS WorldCat Boggs J.M. . Lipid intermolecular hydrogen bonding: influence on structural organization and membrane function , Biochim. Biophys. Acta , 1987 , vol. 906 (pg. 353 - 404 ) Google Scholar Crossref Search ADS PubMed WorldCat Borén T. , Falk P. , Roth K.A. , Larson G. , Normark S. . Attachment of Helicobacter pylori to human gastric epithelium mediated by blood group antigens , Science , 1993 , vol. 262 (pg. 1892 - 1895 ) Google Scholar Crossref Search ADS PubMed WorldCat Breimer M.E. , Hansson G.C. , Karlsson K.-A. , Larson G. , Leffler H. , Pascher I. , Pimlott W. , Samuelsson B.E. . Quayle A. . Fingerprinting of lipid-linked oligosaccharides by mass spectrometry , Advances in Mass Spectrometry , 1980 , vol. 8 London Heyden & Son (pg. 1097 - 1108 ) Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Breimer M.E. , Hansson G.C. , Karlsson K.-A. , Leffler H. . Glycosphingolipids of rat tissues. Different composition of epithelial and non-epithelial cells of small intestine , J. Biol. Chem. , 1982 , vol. 257 (pg. 557 - 568 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Breimer M.E. , Hansson G.C. , Karlsson K.-A. , Leffler H. , Pimlott W. , Samuelsson B.E. . Selected ion monitoring of glycosphingolipid mixtures. Identification of several blood group type glycolipids in the small intestine of an individual rabbit , Biomed. Mass Spectrometry , 1979 , vol. 6 (pg. 231 - 241 ) Google Scholar Crossref Search ADS WorldCat Brûlet P. , McConnell H.M. . Structural and dynamical aspects of membrane immunochemistry using model membranes , Biochemistry , 1977 , vol. 16 (pg. 1209 - 1217 ) Google Scholar Crossref Search ADS PubMed WorldCat Doig P. , Austin J.W. , Kostrzynska M. , Trust T.J. . Production of a conserved adhesin by the human gastroduodenal pathogen Helicobacter pylori , J. Bacteriol. , 1992 , vol. 174 (pg. 2539 - 2547 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Falk K.-E. , Karlsson K.-A. , Samuelsson B. E. . Proton nuclear magnetic resonance analysis of anomeric structure of glycosphingolipids. The globo series (one to five sugars) , Arch. Biochem. Biophys. , 1979 , vol. 192 (pg. 164 - 176 ) Google Scholar Crossref Search ADS PubMed WorldCat Falk K.-E. , Karlsson K.-A. , Samuelsson B.E. . Proton nuclear magnetic resonance analysis of anomeric structure of glycosphingolipids. Blood group ABH-active substances , Arch. Biochem. Biophys. , 1979 , vol. 192 (pg. 177 - 190 ) Google Scholar Crossref Search ADS PubMed WorldCat Falk K.-E. , Karlsson K.-A. , Samuelsson B.E. . Proton nuclear magnetic resonance analysis of anomeric structure of glycosphingolipids. Lewis-active and Lewis-like substances , Arch. Biochem. Biophys. , 1979 , vol. 192 (pg. 191 - 202 ) Google Scholar Crossref Search ADS PubMed WorldCat Falk P. , Roth K.A. , Borén T. , Westblom T.U. , Gordon J.I. , Normark S. . An in vitro adherence assay reveals that Helicobacter pylori exhibits cell lineage-specific tropism in the human gastric epithelium , Proc. Natl. Acad. Sci. USA , 1993 , vol. 90 (pg. 2035 - 2039 ) Google Scholar Crossref Search ADS WorldCat Fenske D.B. , Hamilton K. , Jarrell H.C. , Florio E. , Barber K.R. , Grant C.W.M. . Glycosphingolipids: 2H NMR study of the influence of carbohydrate headgroup structure on ceramide acyl chain behavior in glycolipid-phospholipid bilayers , Biochemistry , 1991 , vol. 30 (pg. 4503 - 4509 ) Google Scholar Crossref Search ADS PubMed WorldCat Handa S. . Blood group active glycolipid from human erythrocytes , Jpn. J. Exp. Med. , 1963 , vol. 33 (pg. 347 - 360 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Hansson G.C. , Karlsson K.-A. , Leffler H. , Strömberg N. . Gangliotetraosylceramide is a major glycolipid of epithelial cells of mouse small intestine , FEBS Lett. , 1982 , vol. 139 (pg. 291 - 294 ) Google Scholar Crossref Search ADS PubMed WorldCat Hansson G.C. , Karlsson K.-A. , Larson G. , McKibbin J.M. , Strömberg N. , Thurin J. . Isoglobotriaosylceramide and the Forssman glycolipid of dog small intestine occupy separate tissue compartments and differ in ceramide structure , Biochim. Biophys. Acta , 1983 , vol. 750 (pg. 214 - 216 ) Google Scholar Crossref Search ADS PubMed WorldCat Hansson G.C. , Karlsson K.-A. , Larson G. , Strömberg N. , Thurin J. . Carbohydrate-specific adhesion of bacteria to thin-layer chromatograms: A rationalized approach to the study of host cell glycolipid receptors , Anal. Biochem. , 1985 , vol. 146 (pg. 158 - 163 ) Google Scholar Crossref Search ADS PubMed WorldCat Holgersson J. , Jovall P.-Å. , Breimer M.E. . Glycosphingolipids of human large intestine: detailed structural characterization with special reference to blood group compounds and bacterial receptor structures , J. Biochem. , 1991 , vol. 110 (pg. 120 - 131 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Imberty A. , Delage M.-M. , Bourne Y. , Cambillau C. , Pérez S. . Data bank of three-dimensional structures of disaccharides. II. N-acetyllactosaminic N-glycans. Comparison with the crystal structure of a biantennary octasaccharide , Glycoconj. J. , 1991 , vol. 8 (pg. 456 - 483 ) Google Scholar Crossref Search ADS PubMed WorldCat Imberty A. , Bourne Y. , Cambillau C. , Rougé P. , Pérez S. . Oligosaccharide conformation in protein carbohydrate complexes , Adv. Biophys. Chem. , 1993 , vol. 3 (pg. 71 - 118 ) OpenURL Placeholder Text WorldCat Kannagi R. , Stroup R. , Cochran N.A. , Urdal D. L. , Young W.W. Jr. , Hakomori S.-i. . Factors affecting expression of glycolipid tumor antigens: influence of ceramide composition and coexisting glycolipid on the antigenicity of gangliotriaosylceramide in murine lymphoma cells , Cancer Res. , 1983 , vol. 43 (pg. 4997 - 5005 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Karlsson K.-A. . Carbohydrate composition and sequence analysis of a derivative of brain disialo ganglioside by mass spectrometry, with molecular weight at m/e 2245. Potential use in the specific microanalysis of cell surface components , Biochemistry , 1974 , vol. 13 (pg. 3643 - 3647 ) Google Scholar Crossref Search ADS PubMed WorldCat Karlsson K.-A. . Preparation of total non-acid glycolipids for overlay analysis of receptors for bacteria and viruses and for other studies , Methods Enzymol. , 1987 , vol. 138 (pg. 212 - 220 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Karlsson K.-A. . Animal glycosphingolipids as membrane attachment sites for bacteria , Annu. Rev. Biochem. , 1989 , vol. 58 (pg. 309 - 350 ) Google Scholar Crossref Search ADS PubMed WorldCat Keith D.D. , Tortora J.A. , Yang R. . Synthesis of l-2-amino-4-methoxy-trans-but-3-enoic acid , J. Org. Chem. , 1978 , vol. 43 (pg. 3711 - 3713 ) Google Scholar Crossref Search ADS WorldCat Kiarash A. , Boyd B. , Lingwood C.A. . Glycosphingolipid receptor function is modified by fatty acid content. Verotoxin 1 and Verotoxin 2c preferentially recognize different globotriaosyl ceramide fatty acid homologues , J. Biol. Chem. , 1994 , vol. 269 (pg. 11138 - 11146 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Koerner T.A.W. Jr. , Prestegard J.H. , Demou P.C. , Yu R.K. . High-resolution proton NMR studies of gangliosides. 1. Use of two-dimensional spin-echo J-correlated spectroscopy for determination of residue composition and anomeric configurations , Biochemistry , 1983 , vol. 22 (pg. 2676 - 2687 ) Google Scholar Crossref Search ADS PubMed WorldCat Larson G. , Karlsson H. , Hansson G.C. , Pimlott W. . Application of a simple methylation procedure for the analysis of glycosphingolipids , Carbohydr. Res. , 1987 , vol. 161 (pg. 281 - 290 ) Google Scholar Crossref Search ADS PubMed WorldCat Ledeen R. , Salzman K. . Structure of the Tay-Sachs' ganglioside , Biochemistry , 1965 , vol. 4 (pg. 2225 - 2233 ) Google Scholar Crossref Search ADS WorldCat Ledeen R.W. , Yu R.K. . Methods for isolation and analysis of gangliosides , Res. Methods Neurochem. , 1978 , vol. 4 (pg. 371 - 410 ) OpenURL Placeholder Text WorldCat Leffler H. , Hansson G.C. , Strömberg N. . A novel sulfoglycosphingolipid of mouse small intestine, IV3-sulfogangliotetraosylceramide, demonstrated by negative ion fast atom bombardment mass spectroscopy , J. Biol. Chem. , 1986 , vol. 261 (pg. 1440 - 1444 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Lelwala-Guruge J. , Nilsson I. , Ljungh Å. , Wadström T. . Cell surface proteins of Helicobacter pylori as antigens in ELISA and a comparison with three commercial ELISA , Scand. J. Infect. Dis. , 1992 , vol. 24 (pg. 457 - 465 ) Google Scholar Crossref Search ADS PubMed WorldCat Lingwood C.A. , Huesca M. , Kuksis A. . The glycerolipid receptor for Helicobacter pylori (and Exoenzyme-S) is phosphatidylethanolamine , Infect. Immun. , 1992 , vol. 60 (pg. 2470 - 2474 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Macher B.A. , Clock J.C. . Isolation and chemical characterization of neutral glycosphingolipids of human neutrophils , J. Biol. Chem. , 1980 , vol. 255 (pg. 2092 - 2096 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Maggio B. . The surface behavior of glycosphingolipids in biomembranes: a new frontier of molecular ecology , Prog. Biophys. Mol. Biol. , 1994 , vol. 62 (pg. 55 - 117 ) Google Scholar Crossref Search ADS PubMed WorldCat Magnusson K.-E. , öhman L. , Stendahl O. , Kihlström E. . Aggregation of lipid vesicles (liposomes). A versatile method to study sugar exposure on biological membranes and sugar affinity of bacteria , Acta Chem. Scand. , 1982 , vol. B36 (pg. 337 - 340 ) Google Scholar Crossref Search ADS WorldCat Mayo S.L. , Olafson B.D. , Goddard W.A. III . Dreiding: a generic force field for molecular simulations , J. Chem. Phys. , 1990 , vol. 94 (pg. 8897 - 8909 ) Google Scholar Crossref Search ADS WorldCat Mirelman D. , Ofek I. . Mirelman D. . Introduction to microbial lectins and agglutinins , Microbial Lectins and Agglutinins. , 1986 New York John Wiley & Sons (pg. 1 - 19 ) Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Morrow M.R. , Singh D.M. , Grant C.W.M. . Glycosphingolipid head-group orientation in fluid phospholipid/cholesterol membranes: similarity for a range of glycolipid fatty acids , Biophys. J. , 1995 , vol. 69 (pg. 955 - 964 ) Google Scholar Crossref Search ADS PubMed WorldCat Nyholm P.-G. , Pascher I. . Orientation of the saccharide chains of glycolipids at the membrane surface. Conformational analysis of the glucose-ceramide and the glucose-glyceride linkages using molecular mechanics (MM3) , Biochemistry , 1993 , vol. 32 (pg. 1225 - 1234 ) Google Scholar Crossref Search ADS PubMed WorldCat Nyholm P.-G. , Magnusson G. , Zheng Z. , Norel R. , Binnington-Boyd B. , Lingwood C.A. . Two distinct binding sites for globotriaosyl ceramide on verotoxins: identification by molecular modelling and confirmation using deoxy analogues and a new glycolipid receptor for all verotoxins , Chem. Biol. , 1996 , vol. 3 (pg. 263 - 275 ) Google Scholar Crossref Search ADS PubMed WorldCat Poppe L. , Dabrowski J. , von der Lieth C.-W. , Koike K. , Ogawa T. . Three-dimensional structure of the oligosaccharide terminus of globotriaosylceramide and iso globotriaosylceramide in solution. A rotating frame NOE study using hydroxyl groups as long-range sensors in conformational analysis by 1H-NMR spectroscopy , Eur. J. Biochem. , 1990 , vol. 189 (pg. 313 - 325 ) Google Scholar Crossref Search ADS PubMed WorldCat Rappé A.K. , Goddard W.A. III . Charge equilibration for molecular dynamics simulations , J. Chem. Phys. , 1991 , vol. 95 (pg. 3358 - 3363 ) Google Scholar Crossref Search ADS WorldCat Saitoh T. , Natomi H. , Zhao W. , Okuzumi K. , Sugano K. , Iwamori M. , Nagai Y. . Identification of glycolipid receptors for Helicobacter pylori by TLC-immunostaining , FEBS Lett. , 1991 , vol. 282 (pg. 385 - 387 ) Google Scholar Crossref Search ADS PubMed WorldCat Samuelsson B.E. , Pimlott W. , Karlsson K.-A. . Mass spectrometry of mixtures of intact glycosphingolipids , Methods Enzymol. , 1990 , vol. 193 (pg. 623 - 646 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Seyama Y. , Yamakawa T. . Chemical structure of glycolipid of guinea pig red blood cell membrane , J. Biochem. (Tokyo) , 1974 , vol. 75 (pg. 837 - 842 ) OpenURL Placeholder Text WorldCat Singh D. , Jarrell H.C. , Barber K.R. , Grant C.W.M. . Glycosphingolipids: 2H NMR study of the influence of ceramide fatty acid characteristics on the carbohydrate headgroup in phospholipid bilayers , Biochemistry , 1992 , vol. 31 (pg. 2662 - 2669 ) Google Scholar Crossref Search ADS PubMed WorldCat Singh D.M. , Shan X. , Davis J.H. , Jones D.H. , Grant C.W.M. . Oligosaccharide behavior of complex natural glycosphingolipids in multicomponent model membranes , Biochemistry , 1995 , vol. 34 (pg. 451 - 463 ) Google Scholar Crossref Search ADS PubMed WorldCat Soltesz V. , Schalen C. , MÅrdh P.A. . Kaijser B. , Falsen E. . New selective medium for Campylobacter pylori , Proceedings of the Fourth International Workshop on Campylobacter Infections. , 1988 Kungälv, Sweden, Goterna (pg. 433 - 436 ) Stellner K. , Saito H. , Hakomori S.-i. . Determination of aminosugar linkages in glycolipids by methylation. Amino sugar linkages of ceramide pentasaccharides of rabbit erythrocytes and of Forssman antigen , Arch. Biochem. Biophys. , 1973 , vol. 155 (pg. 464 - 472 ) Google Scholar Crossref Search ADS PubMed WorldCat Strömberg N. , Deal C. , Nyberg G. , Normark S. , So M. , Karlsson K.-A. . Identification of carbohydrate structures which are possible receptors for Neisseria gonorrhoeae , Proc. Natl. Acad. Sci. USA , 1988 , vol. 85 (pg. 4902 - 4906 ) Google Scholar Crossref Search ADS WorldCat Strömberg N. , Karlsson K.-A. . Characterization of the binding of Propionibacterium granulosum to glycosphingolipids adsorbed on surfaces. An apparent recognition of lactose with a dependence on the ceramide structure , J. Biol. Chem. , 1990 , vol. 265 (pg. 11244 - 11250 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Strömberg N. , Karlsson K.-A. . Characterization of the binding of Actinomyces naeslundii (ATCC 12104) and Actinomyces viscosus (ATCC 19246) to glycosphingolipids using a solid-phase overlay approach , J. Biol. Chem. , 1990 , vol. 265 (pg. 11251 - 11258 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Svennerholm L. . The chemical structure of normal brain and Tay-Sachs gangliosides , Biochem. Biophys. Res. Commun. , 1962 , vol. 9 (pg. 436 - 441 ) Google Scholar Crossref Search ADS PubMed WorldCat Svennerholm L. . Chromatographic separation of human brain gangliosides , J. Neurochem. , 1963 , vol. 10 (pg. 613 - 623 ) Google Scholar Crossref Search ADS PubMed WorldCat Szoka F. Jr. , Phapadjopouos D. . Procedure for preparation of liposomes with large internal aqueous space and high capture by reverse-phase evaporation , Proc. Natl. Acad. Sci. USA , 1978 , vol. 75 (pg. 4194 - 4198 ) Google Scholar Crossref Search ADS WorldCat Teneberg S. , Willemsen P. , de Graaf F.K. , Karlsson K.-A. . Receptor-active glycolipids of epithelial cells of small intestine of young and adult pigs in relation to susceptibility to infection with Escherichia coli K99 , FEBS Lett. , 1990 , vol. 263 (pg. 10 - 14 ) Google Scholar Crossref Search ADS PubMed WorldCat Teneberg S. , Willemsen P.T.J. , de Graaf F.K. , Stenhagen G. , Pimlott W. , Jovall P.-Å. , Ångström J. , Karlsson K.-A. . Characterization of gangliosides of epithelial cells of calf small intestine, with special reference to receptor-active sequences for enteropathogenic Escherichia coli K99 , J. Biochem. (Tokyo) , 1994 , vol. 116 (pg. 560 - 574 ) OpenURL Placeholder Text WorldCat Thurin J. , Brodin T. , Bechtel B. , Jovall P.-Å. , Karlsson H. , Strömberg N. , Teneberg S. , Sjögren H.-O. , Karlsson K.-A. . Novel isoglobo-neolactoseries hybrid glycolipid detected by a monoclonal antibody is a rat tumor-associated antigen , Biochim. Biophys. Acta , 1989 , vol. 1002 (pg. 267 - 272 ) Google Scholar Crossref Search ADS PubMed WorldCat Utsumi H. , Suzuki T. , Inoue K. , Nojima S. . Haptenic activity of galactosyl ceramide and its topographical distribution on liposomal membranes. Effects of temperature and phospholipid composition , J. Biochem. (Tokyo) , 1984 , vol. 96 (pg. 97 - 105 ) OpenURL Placeholder Text WorldCat Waldi D. . Stahl E. . Sprühreagentien für die Dünnschicht-chromatographie , Dünnschicht-Chromatographie. , 1962 Berlin Springer-Verlag (pg. 496 - 515 ) Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Yamakawa T. , Irie R. , Iwanaga M. . The chemistry of lipid of posthemolytic residue or stroma of erythrocytes. 9. Silicic acid chromatography of mammalian stroma glycolipids , J. Biol. Chem. , 1960 , vol. 48 (pg. 490 - 507 ) OpenURL Placeholder Text WorldCat Yamakawa T. , Kiso N. , Hansa S. , Makita A. , Yokoyama S. . On the structure of brain sulfuric ester and ceramide dihexoside of erythrocytes , J. Biol. Chem. , 1962 , vol. 52 (pg. 226 - 227 ) OpenURL Placeholder Text WorldCat Yamakawa T. . Glycolipids of mammalian red blood cells , Colloq. Ges. Physiol. Chem. , 1966 , vol. 16 (pg. 87 - 111 ) OpenURL Placeholder Text WorldCat Yang H. , Hakomori S.-i. . A sphingolipid having a novel ceramide and lacto-N-fucopentose-III , J. Biol. Chem. , 1971 , vol. 246 (pg. 1192 - 1200 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Yuyama Y. , Yoshimatsu K. , Ono E. , Saito M. , Naiki M. . Postnatal change of pig intestinal ganglioside bound by Escherichia coli with K99 fimbriae , J. Biochem. , 1993 , vol. 113 (pg. 488 - 492 ) Google Scholar PubMed OpenURL Placeholder Text WorldCat Author notes 5 Present address: Astra Hässle AB, Kärragatan 5, S-431 83 Mölndal, Sweden © 1998 Oxford University Press
TI - The lactosylceramide binding specificity of Helicobacter pylori
JO - Glycobiology
DO - 10.1093/glycob/8.4.297
DA - 1998-04-01
UR - https://www.deepdyve.com/lp/oxford-university-press/the-lactosylceramide-binding-specificity-of-helicobacter-pylori-5frvWxJs96
SP - 297
EP - 309
VL - 8
IS - 4
DP - DeepDyve
ER -