Glycoconjugate Receptors for P-fimbriated Escherichia coli in the Mouse
1995; Elsevier BV; Volume: 270; Issue: 15 Linguagem: Inglês
10.1074/jbc.270.15.9017
ISSN1083-351X
AutoresBoel Lanne, Britt‐Marie Olsson, Per- Jovall, Jonas öm, Henrik Linder, Britt‐Inger Marklund, Jörgen Bergström, Karl‐Anders Karlsson,
Tópico(s)Drug Transport and Resistance Mechanisms
ResumoGlycosphingolipids were isolated from kidneys, urethers, and bladders (including urethrae) of C3H/HeN mice. Binding was studied of a clinical isolate and recombinant strains of uropathogenic P-fimbriated Escherichia coli to these glycolipids. A series of receptor-active glycolipids with Galα4Gal in common, previously shown to be recognized by these bacteria, was identified by use of specific monoclonal antibodies, fast-atom bombardment and electron-impact mass spectrometry, and proton nuclear magnetic resonance spectroscopy: galabiosylceramide (Galα4GalβCer), globotriaosylceramide (Galα4Galβ4GlcβCer), globoside (GalNAcβ3Galα4Galβ4GlcβCer), the Forssman glycolipid (GalNAcα3GalNAcβ3Galα4Galβ4GlcβCer), Galβ4GlcNAcβ6(Galβ3)GalNAcβ3Galα4Galβ4GlcβCer, and Galβ4(Fucα3)GlcNAcβ6(Galβ3)GalNAcβ3Galα4Galβ4GlcβCer.The binding pattern for mouse kidney glycolipids differed from that for kidney glycolipids of man and monkey. In particular, the dominant 8-sugar glycolipid in the mouse was not detected in the primates. A second difference was found in the binding of E. coli to kidney glycoproteins on blots after electrophoresis; the mouse showed distinct receptor-active bands while human and monkey did not. These differences may be of relevance when using the mouse as a model for clinical urinary tract infection of man. Glycosphingolipids were isolated from kidneys, urethers, and bladders (including urethrae) of C3H/HeN mice. Binding was studied of a clinical isolate and recombinant strains of uropathogenic P-fimbriated Escherichia coli to these glycolipids. A series of receptor-active glycolipids with Galα4Gal in common, previously shown to be recognized by these bacteria, was identified by use of specific monoclonal antibodies, fast-atom bombardment and electron-impact mass spectrometry, and proton nuclear magnetic resonance spectroscopy: galabiosylceramide (Galα4GalβCer), globotriaosylceramide (Galα4Galβ4GlcβCer), globoside (GalNAcβ3Galα4Galβ4GlcβCer), the Forssman glycolipid (GalNAcα3GalNAcβ3Galα4Galβ4GlcβCer), Galβ4GlcNAcβ6(Galβ3)GalNAcβ3Galα4Galβ4GlcβCer, and Galβ4(Fucα3)GlcNAcβ6(Galβ3)GalNAcβ3Galα4Galβ4GlcβCer. The binding pattern for mouse kidney glycolipids differed from that for kidney glycolipids of man and monkey. In particular, the dominant 8-sugar glycolipid in the mouse was not detected in the primates. A second difference was found in the binding of E. coli to kidney glycoproteins on blots after electrophoresis; the mouse showed distinct receptor-active bands while human and monkey did not. These differences may be of relevance when using the mouse as a model for clinical urinary tract infection of man. An important factor in the pathogenesis of bacterial infections is the ability of the bacteria to adhere to host tissues (1Beachey E.H. J. Infect. Dis. 1981; 143: 325-345Crossref PubMed Scopus (951) Google Scholar), often by means of specific binding of bacterial adhesins to host cell carbohydrates (2Mirelman D. Ofek I. Mirelman D. Microbial Lectins and Agglutinins. Wiley & Sons, New York1986: 1-19Google Scholar, 3Sharon N. Liener I.E. Sharon N. Goldstein I.J. The Lectins. Academic Press, New York1986: 493-526Crossref Google Scholar). P-fimbriated Escherichia coli are important in human urinary tract infections and are known to adhere to the epithelial cells (4Svanborg-Edén C. Hanson L.Å. Jodal U. Lindberg U. and A.S. Lancet. 1976; 2: 490-492Abstract Scopus (301) Google Scholar, 5Källenius G. Möllby R. Svenson S.B. Winberg J. Lundblad A. Svensson S. Cedergren B. FEMS Microbiol. Lett. 1980; 7: 297-302Crossref Scopus (14) Google Scholar, 6Källenius G. Möllby R. Svensson S.B. Helin I. Hultberg H. Cedergren B. Winberg J. Lancet. 1981; ii: 1369-1372Abstract Scopus (273) Google Scholar). The adhesion has been shown to depend on Galα4Gal-containing glycolipids on the host cells (5Källenius G. Möllby R. Svenson S.B. Winberg J. Lundblad A. Svensson S. Cedergren B. FEMS Microbiol. Lett. 1980; 7: 297-302Crossref Scopus (14) Google Scholar, 7Leffler H. Svanborg-Edén C. FEMS Microbiol. Lett. 1980; 8: 127-134Crossref Scopus (364) Google Scholar) and a number of glycolipid isoreceptors with Galα4Gal in terminal or internal positions have been identified (8Bock 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. J. Biol. Chem. 1985; 260: 8545-8551Abstract Full Text PDF PubMed Google Scholar). Recently, it was demonstrated for human red blood cells that Galα4Gal was absent from glycoproteins and exclusively present in the glycolipid form (9Yang Z. Bergström J. Karlsson K.-A. J. Biol. Chem. 1994; 269: 14620-14624Abstract Full Text PDF PubMed Google Scholar). The molecular genetics and biogenesis of P-fimbriae have been extensively studied, including the sequence of the three classes of adhesin, I, II, and III (10Tennent J.M. Hultgren S. Marklund B.-I. Forsman K. Göransson M. Uhlin B.E. Normark S. Iglewski B.H. Clark V.L. Molecular Basis of Bacterial Pathogenesis. Academic Press, New York1990: 79-110Google Scholar, 11Hultgren S.J. Abraham S. Caparon M. Falk St., P. Geme III, J.W. Normark S. Cell. 1993; 73: 887-901Abstract Full Text PDF PubMed Scopus (341) Google Scholar), which have slightly different affinities for various isoreceptors of glycolipids (12Strömberg N. Marklund B.-I. Lund B. Ilver D. Hamers A. Gaastra W. Karlsson K.-A. Normark S. EMBO J. 1990; 9: 2001-2010Crossref PubMed Scopus (216) Google Scholar, 13Marklund, B.-I., 1991, Structural and Functional Variation Among Gal α1-4Gal Adhesins of Uropathogenic Escherichia coli, Ph.D. thesis, University of UmeÅ, Sweden.Google Scholar, 14Marklund B.-I. Tennent J.M. Garcia E. Hamers A.B.M. Lindberg F. Gaastra W. Normark S. Mol. Microbiol. 1992; 6: 2225-2242Crossref PubMed Scopus (167) Google Scholar, 15Garcia E. Hamers A.M. Bergmans H.E.N. van der Zeijst B.A.M. Gaastra W. Curr. Microbiol. 1988; 17: 333-337Crossref Scopus (10) Google Scholar). With respect to human and dog urinary tract infection, clinical isolates of E. coli differ in adhesin class, which is related to different glycolipid patterns of the two species. In isolates of human origin, the class II adhesin predominates. This adhesin preferentially binds to globoside, 1The abbreviations and trivial names used are:globosideGb4, GalNAcβ3Galα4Galβ4GlcβCerCerceramided18:1-24:0ceramide with sphingosine and a saturated fatty acid chaingalabiosylceramideGalα4GalβCerglobotriaosylceramideGb3, Galα4Galβ4GlcβCerForssman glycolipidGalNAcα3GalNAcβ3Galα4Galβ4GlcβCerP1 antigenGalα4Galβ4GlcNAcβ3Galβ4GlcβCerGlobo-AGalNAcα3(Fucα2)Galβ3GalNAcβ3Galα4Galβ4GlcβCergangliotetraosylceramideGgO4, Galβ3GalNAcβ4Galβ4GlcβCerlactotetraosylceramideLc4, Galβ3GlcNAcβ3Galβ4GlcβCerlactoneotetraosylceramidenLc4, Galβ4GlcNAcβ3Galβ4GlcβCerLexGalβ4(Fucα3)GlcNAcβ3Galβ4GlcβCerLeyFucα2Galβ4(Fucα3)GlcNAcβ3Galβ4GlcβCerLeaGalβ3(Fucα4)GlcNAcβ3Galβ4GlcβCerLebFucα2Galβ3(Fucα4)GlcNAcβ3Galβ4GlcβCerHPLChigh-pressure liquid chromatographyNMRnuclear magnetic resonanceNOEnuclear Overhauser enhancementMSmass spectrometryFABfast-atom bombardmentEIelectron impactGCgas chromatographyPAGEpolyacrylamide gel electrophoresis. 1The abbreviations and trivial names used are:globosideGb4, GalNAcβ3Galα4Galβ4GlcβCerCerceramided18:1-24:0ceramide with sphingosine and a saturated fatty acid chaingalabiosylceramideGalα4GalβCerglobotriaosylceramideGb3, Galα4Galβ4GlcβCerForssman glycolipidGalNAcα3GalNAcβ3Galα4Galβ4GlcβCerP1 antigenGalα4Galβ4GlcNAcβ3Galβ4GlcβCerGlobo-AGalNAcα3(Fucα2)Galβ3GalNAcβ3Galα4Galβ4GlcβCergangliotetraosylceramideGgO4, Galβ3GalNAcβ4Galβ4GlcβCerlactotetraosylceramideLc4, Galβ3GlcNAcβ3Galβ4GlcβCerlactoneotetraosylceramidenLc4, Galβ4GlcNAcβ3Galβ4GlcβCerLexGalβ4(Fucα3)GlcNAcβ3Galβ4GlcβCerLeyFucα2Galβ4(Fucα3)GlcNAcβ3Galβ4GlcβCerLeaGalβ3(Fucα4)GlcNAcβ3Galβ4GlcβCerLebFucα2Galβ3(Fucα4)GlcNAcβ3Galβ4GlcβCerHPLChigh-pressure liquid chromatographyNMRnuclear magnetic resonanceNOEnuclear Overhauser enhancementMSmass spectrometryFABfast-atom bombardmentEIelectron impactGCgas chromatographyPAGEpolyacrylamide gel electrophoresis. a glycolipid that dominates in human urinary tract. E. coli isolates from dog infections, however, express mostly the class III adhesin, which binds more strongly to the Forssman glycolipid, a major glycolipid in dog kidney. Gb4, GalNAcβ3Galα4Galβ4GlcβCer ceramide ceramide with sphingosine and a saturated fatty acid chain Galα4GalβCer Gb3, Galα4Galβ4GlcβCer GalNAcα3GalNAcβ3Galα4Galβ4GlcβCer Galα4Galβ4GlcNAcβ3Galβ4GlcβCer GalNAcα3(Fucα2)Galβ3GalNAcβ3Galα4Galβ4GlcβCer GgO4, Galβ3GalNAcβ4Galβ4GlcβCer Lc4, Galβ3GlcNAcβ3Galβ4GlcβCer nLc4, Galβ4GlcNAcβ3Galβ4GlcβCer Galβ4(Fucα3)GlcNAcβ3Galβ4GlcβCer Fucα2Galβ4(Fucα3)GlcNAcβ3Galβ4GlcβCer Galβ3(Fucα4)GlcNAcβ3Galβ4GlcβCer Fucα2Galβ3(Fucα4)GlcNAcβ3Galβ4GlcβCer high-pressure liquid chromatography nuclear magnetic resonance nuclear Overhauser enhancement mass spectrometry fast-atom bombardment electron impact gas chromatography polyacrylamide gel electrophoresis. Gb4, GalNAcβ3Galα4Galβ4GlcβCer ceramide ceramide with sphingosine and a saturated fatty acid chain Galα4GalβCer Gb3, Galα4Galβ4GlcβCer GalNAcα3GalNAcβ3Galα4Galβ4GlcβCer Galα4Galβ4GlcNAcβ3Galβ4GlcβCer GalNAcα3(Fucα2)Galβ3GalNAcβ3Galα4Galβ4GlcβCer GgO4, Galβ3GalNAcβ4Galβ4GlcβCer Lc4, Galβ3GlcNAcβ3Galβ4GlcβCer nLc4, Galβ4GlcNAcβ3Galβ4GlcβCer Galβ4(Fucα3)GlcNAcβ3Galβ4GlcβCer Fucα2Galβ4(Fucα3)GlcNAcβ3Galβ4GlcβCer Galβ3(Fucα4)GlcNAcβ3Galβ4GlcβCer Fucα2Galβ3(Fucα4)GlcNAcβ3Galβ4GlcβCer high-pressure liquid chromatography nuclear magnetic resonance nuclear Overhauser enhancement mass spectrometry fast-atom bombardment electron impact gas chromatography polyacrylamide gel electrophoresis. The mouse is currently used as a model of human urinary tract infection 2H. Linder, unpublished data. 2H. Linder, unpublished data.(16Aronson M. Medalia O. Schori L. Mirelman D. Sharon N. Ofek I. J. Infect. Dis. 1979; 139: 329-332Crossref PubMed Scopus (205) Google Scholar, 17Iwahi T. Abe Y. Nakao M. Imada A. Tsuchiya K. Infect. Immun. 1983; 39: 1307-1315Crossref PubMed Google Scholar, 18Johnson J.R. Berggren T. Am. J. Med. Sci. 1984; 307: 335-339Crossref Scopus (19) Google Scholar). It is therefore of interest to investigate the carbohydrate basis for E. coli adhesion in this animal. Thus, model C3H/HeN mice were analyzed for the presence of receptor-active glycoconjugates in various parts of the urinary tract using both clinical isolates and recombinant strains of uropathogenic E. coli. It was found that the isoreceptor pattern differed significantly from that found in human and monkey urinary tract with respect to both glycolipids and glycoproteins. Female C3H/HeN mice (original breeding stock, Charles River Laboratories, Margate, Kent, UK) were kept at the animal facilities at the Department of Infection and Immunology, Astra Arcus, S-181 85 Södertälje, Sweden, where the model infections were performed.2The mice were used at 8-10 weeks of age. The animals were anesthetized with diethyl ether and tissues were removed using sterile instruments. Kidneys from 60 animals, and urethers, bladders, and urethrae (in one piece) from 10 animals were obtained. The tissues were immediately frozen using cryostat spray and stored at −70°C until analysis was performed. Before lipid extraction the urethers were separated from bladders plus urethrae. Lipids were extracted with a Soxhlet apparatus using chloroform/methanol (2:1 for 24 h, and 1:9 for 24 h). The combined extracts were subjected to mild-alkaline hydrolysis (0.2 M KOH in methanol, 3 h), dialyzed against water, and purified on a silica straight-phase column. No further purification of lipids from urethers and bladders plus urethrae, 2.1 and 2.7 mg, respectively, was carried out, while the purification of the mouse kidney preparation continued as described by Karlsson (19Karlsson K.-A. Methods Enzymol. 1987; 138: 212-220Crossref PubMed Scopus (183) Google Scholar). The kidney glycosphingolipids were separated into two main fractions, a neutral one containing uncharged lipids, 8.5 mg, and an acid fraction containing gangliosides and other charged lipids. Fractionation of neutral glycolipids was achieved by HPLC chromatography on a Spherisorb S10W silica column, 250 × 8-mm inner diameter, particle size 10 μm (Phase Separation Ltd., Queensferry, UK). The glycolipid mixture, 6 mg, was applied and eluted with a gradient starting with chloroform/methanol/water, 60:35:8 (by volume unless otherwise stated), and ending with 10:10:3, 2 ml/min. Mild hydrolysis of acid glycosphingolipids to remove sialic acid was achieved in 1 ml of acetic acid/water, 1:100 (300 μg of lipid), 100°C, 1 h. Reference glycosphingolipids were obtained as follows: globotriaosylceramide, globoside, and P1 glycolipid from human erythrocytes (20Stults C.L.M. Sweeley C.C. Macher B.A. Methods Enzymol. 1989; 179: 167-214Crossref PubMed Scopus (233) Google Scholar), Leaand Lebfrom human small intestine (21Björk S. Breimer M.E. Hansson G.C. Karlsson K.-A. Leffler H. J. Biol. Chem. 1987; 262: 6758-6765Abstract Full Text PDF PubMed Google Scholar), the Forssman glycolipid (22Vance W.R. Shook III, C.P. McKibbin J.M. Biochemistry. 1966; 5: 435-445Crossref PubMed Scopus (32) Google Scholar) and Ley(23McKibbin J.M. Spencer W.A. Smith E.L. MJ -E. Karlsson K.-A. Samuelsson B.E. Li Y.-T. Li S.-C. J. Biol. Chem. 1982; 257: 755-760Abstract Full Text PDF PubMed Google Scholar) from dog intestine, and gangliotetraosylceramide from feces of axenic mouse (24Hansson G.C. Karlsson K.-A. Leffler H. Strömberg N. FEBS Lett. 1982; 139: 291-294Crossref PubMed Scopus (28) Google Scholar). Reference mixtures of glycolipids were obtained from dog intestine, calf brain, human kidneys, human erythrocytes (blood group A), and human primary liver cancer, which contains Lexstructures (20Stults C.L.M. Sweeley C.C. Macher B.A. Methods Enzymol. 1989; 179: 167-214Crossref PubMed Scopus (233) Google Scholar). For analysis of glycolipid mixtures using bacteria and antibodies, the glycolipids were separated on TLC with chloroform/methanol/water, 60:35:8, on aluminum backed silica nano plates of 0.2-mm phase thickness (Merck, Germany). Chemical detection was done with anisaldehyde (25Waldi D. Stahl E. Dunnschichts-Chromatographie. Springer Verlag, Berlin1962: 496-515Google Scholar) or resorcinol staining (26Svennerholm L. J. Neurochem. 1963; 10: 613-623Crossref PubMed Scopus (1297) Google Scholar). The procedure used for incubation of TLC plates with labeled biological reagents has been described (27Karlsson K.-A. Strömberg N. Methods Enzymol. 1987; 138: 220-231Crossref PubMed Scopus (120) Google Scholar). Radioactivity was detected by autoradiography (Kodak XAR-5 film, Eastman Kodak). Dried residues (20 mg) remaining after the lipid extraction of C3H/HeN mice and cynomolgus monkey (Macaca fascicularis) kidneys were mixed with 1 ml of 50 m M Tris-HCl buffer (pH 8.0) containing 2.5% SDS. The mixture was incubated at room temperature overnight (gentle shaking), heated to 95°C for 10 min, and then centrifuged at 10,000 × g for 10 min. To extract residual lipids from the SDS-solubilized material, this was treated with 1-butanol/water, 9:1, three times. Prior to SDS-PAGE of the supernatant, the samples were diluted to 2-4 mg of protein/ml (determined by BCA Protein Assay, Pierce), and 2-mercaptoethanol (5%) was added. SDS-PAGE (gradient gel, 8-25%) and Coomassie R-350 (PhastGel™Blue R, Pharmacia) staining were carried out with a Pharmacia Phast System™(Pharmacia, Sweden) according to the protocols of the manufacturer. Briefly, samples were heated to 95°C for 5 min and centrifuged at 10,000 × g for 2 min before electrophoresis to remove unsolubilized material. After electrophoresis the gel was either stained for protein or sugar (Glycan detection kit, Boehringer, Germany), or electroblotted onto a nitrocellulose membrane (0.45 μm) in 20% methanol containing 192 m M glycine and 25 m M Tris at pH 8.3. For incubation with bacteria, the nitrocellulose membrane with electroblotted proteins was preincubated in blocking solution (3% bovine serum albumin, 50 m M Tris-HCl, 200 m M NaCl, 0.1% NaN3, pH 8.0) for 1.5 h. The membrane was then incubated with 35S-labeled E. coli in phosphate-buffered saline (0.14 M phosphate buffer, pH 7.2, 0.14 M NaCl, 5 m M KCl). After 1.5-2 h the membrane was washed in 50 m M Tris-HCl, 0.2 M NaCl, 0.05% Tween 20 (pH 8.0), dried at room temperature, and exposed to x-ray film overnight. Mass spectra were obtained with a ZAB-2F/HF (VG Analytical, Manchester, UK) and a Jeol SX 102A (Jeol, Tokyo, Japan), both sector instruments, either in the positive-EI or negative-FAB mode (Xe atom bombardment, 8 kV). Triethanolamine was used as matrix. Methylation was performed according to Ref. 28Larson G. Karlsson H. Hansson G.C. Pimlott W. Carbohydr. Res. 1987; 161: 281-290Crossref PubMed Scopus (96) Google Scholar (and references therein) and gas chromatography-mass spectroscopic analysis of partially methylated alditol acetates was done according to Refs. 29Stellner K. Saito H. Hakomori S.-i. Arch. Biochem. Biophys. 1973; 155: 464-472Crossref PubMed Scopus (672) Google Scholar and 30Yang H. Hakomori S. J. Biol. Chem. 1971; 246: 1192-1200Abstract Full Text PDF PubMed Google Scholar. For analysis of partially methylated alditol acetates, a quadropole MS (Trio-II, VG Masslab, Altricham, UK) was used. The capillary column was a DB-1, 15 m × 0.25 mm, inner diameter, 0.2-μm film thickness (J& Scientific). FAB-MS analysis of glycolipids was performed directly from the TLC plate as described (31Karlsson K.-A. Lanne B. Pimlott W. Teneberg S. Carbohydr. Res. 1991; 221: 49-61Crossref PubMed Scopus (28) Google Scholar) with the first MS of a Jeol HX/HX110A instrument. The conditions for the aquisition of EI spectra were: 70 eV electron ionization potential, 10 kV accelerating voltage, 300 μA trap current, ion source temperature 375°C, and scan time 26 s. 1H NMR spectra on deuterium-exchanged glycolipid fractions were acquired at 11.74 telsa on a JEOL ALPHA-500 (Jeol, Tokyo, Japan). Samples were dissolved in 0.5 ml of Me2SO/D2O, 98:2, and spectra recorded at 30°C with a digital resolution of 0.4 Hz. Chemical shifts are given relative to tetramethylsilane using the internal solvent peak. Nuclear Overhauser enhancements were measured using the standard software of the instrument. The recycle time was at least five times the longitudinal relaxation time (T1). Four different P-fimbriated E. coli strains were used, HB101/pPIL291-15 (a gift from Dr. I. van Die, Vrije Universiteit, Amsterdam, The Netherlands, and Dr. B. Westerlund, University of Helsinki, Finland (32de Ree J.M. Schwillens P. van den Bosch J.F. FEMS Microbiol. 1985; 29: 91-97Google Scholar)), HB101/pPAP5 (33Lindberg F.P. Lund B. Normark S. EMBO J. 1984; 3: 1167-1173Crossref PubMed Scopus (110) Google Scholar), HB101/pDC1 (34Clegg S. Infect. Immun. 1982; 38: 739-744Crossref PubMed Google Scholar), HB101/pPAP601 (35Lund B. Lindberg F. Marklund B.-I. Normark S. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 5898-5902Crossref PubMed Scopus (187) Google Scholar), and DS-17 (a gift from Dr. K. Tullus, S:t Görans barnsjukhus, Stockholm, Sweden (36Tullus K. Hörlin K. Svenson S.B. Källenius G. J. Infect. Dis. 1984; 150: 728-736Crossref PubMed Scopus (89) Google Scholar)). The bacteria were cultivated on colonization factor agar plates supplemented with [35S]methionine (400 μCi/10 ml, Amersham International, UK) at 37°C overnight. They were collected by centrifugation, washed twice with phosphate-buffered saline, and resuspended in phosphate-buffered saline to approximately 1 × 109colony-forming units/ml. The bacteria were diluted to an activity of approximately 1 × 106cpm/ml. The following mouse monoclonal antibodies were used: anti-Lex(BL-G15, Monosan, Bio-Zak, Järfälla, Sweden), three antibodies binding to terminal Galα4Gal sequences, pk002, P001, and MC2102 = 87:5 (all obtained from Accurate Chemical & Science Corp., New York); and anti-Forssman (MAS033b, Seralab, Göteborgs Termometerfabrik, Sweden). The secondary antibodies used (rabbit anti-mouse immunoglobulins, Z109, DAKO A/S, Denmark) were labeled with 125I (37Aggarwal B.B. Eessalu T.E. Haas P.E. Nature. 1985; 318: 665-667Crossref PubMed Scopus (737) Google Scholar). Minimum energy conformers of the GL-III glycolipid, identified below as Galβ4(Fucα3)GlcNAcβ6(Galβ3)GalNAcβ3Galα4Galβ4GlcβCer, were calculated within the Biograf molecular modeling program (Molecular Simulations Inc.) using the Dreiding-II force field (38Mayo S.L. Olafson B.O. Goddard III, W.A. J. Phys. Chem. 1990; 94: 8897-8909Crossref Scopus (5061) Google Scholar) on a Silicon Graphics 4D/35TG workstation. Partial atomic charges were generated using the charge equilibration method (39Pappé A.K. Goddard III, W.A. J. Phys. Chem. 1991; 95: 3358-3363Crossref Scopus (2658) Google Scholar), 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 Dhb was set to 4 kcal/mol. The acid and neutral glycosphingolipids of C3H/HeN mouse kidneys were isolated and purified separately. TLC separations of the neutral glycosphingolipids visualized chemically with anisaldehyde are shown in Fig. 1 A. For comparison, glycolipids from human kidneys, as well as purified globoside, were included in the chromatogram. The neutral glycolipids from mouse kidneys are dominated by a very slow-moving band with more than 6 sugars (GL-III, Fig. 1) although several more weakly-staining bands, particularly in the 3-4-sugar region, are also seen. The mixture of acid glycolipids from mouse kidneys and the crude preparations of lipids from urethers and bladders plus urethrae were used for bacterial binding studies. The latter preparations were not purified sufficiently for chemical staining (the bands seen in Fig. 1 of urethers and bladders plus urethrae do not have the green color characteristic of glycoconjugates stained with anisaldehyde). Initially P-fimbriated E. coli, HB101/pPIL291-15, which binds very strongly to Galα4Gal-containing glycolipids, was used to screen the glycolipid preparations as shown in Fig. 1 B. Globotriaosylceramide, globoside, and the Forssman glycolipid were applied in the first three lanes as references. The bacteria bound strongly to neutral glycolipids from mouse kidneys in the 4-5-sugar region. In addition, four slower moving bands were strongly bound by the bacteria (GL-I, GL-II, GL-III, and GL-IV). The urethers and bladders plus urethrae both contained binding glycolipids in the 4- and 5-sugar regions, while in the preparation of bladders plus urethrae weak binding is also seen to a band in the 3-sugar region and also to a band comigrating with GL-II. For the urether a weak binding band was obtained which had the same mobility as GL-III. The acid glycolipids obtained from the mouse kidneys showed two distinct binding bands (Fig. 1 B) which had the same mobility as GL-II and GL-III. No chemical identification was made of these glycolipids. Antibodies were used to reveal the presence of terminally placed Galα4Gal structures in glycolipids from mouse kidneys. In the 2-sugar region, the binding of the antibody MC2001 showed that Galα4GalβCer was present in mouse kidneys (Fig. 2 A). The three bands probably differed in ceramide composition. Dog intestine was included in the analysis as a positive reference. 3K.-A. Karlsson, unpublished data. In the 3-sugar region, the antibody pk002 bound weakly, Fig. 2 B, but strongly to the reference globotriaosylceramide. The lack of a clear binding by monoclonal antibody P001 indicated that the P1 antigen is absent from mouse kidneys (Fig. 2 C). Further investigation with an antibody specific for the Forssman glycolipid (Fig. 2 D), showed that this glycolipid was present in the mouse but not in human kidneys. A weak band was also detected in the preparation of mouse bladders plus urethrae but not in the urethers. To elucidate the structure of GL-II, GL-III, and GL-IV in the mouse kidneys, the neutral glycolipid mixture was subjected to HPLC fractionation on a silica column. The separation conditions were chosen to optimize separation of slow-moving compounds. Analysis by TLC of the fractions obtained is shown in Fig. 3. Fractions 4 and 5 were analyzed by FAB-MS in two ways after separation by TLC. First, the 3- and 4-sugar regions were scraped off the plates separately, extracted in chloroform/methanol, 2:1, and their mass spectra were collected. Second, they were analyzed by direct TLC FAB-MS. The 2-sugar region gave spectra consistent with (Hex)2Cer, the 3-sugar region with (Hex)3Cer, and the 4-sugar region gave spectra showing that the glycolipid was HexNAcHexHexHexCer. All three compounds had similar ratios of hydroxylated to unhydroxylated ceramide (1:1). HPLC fractions 8 (GL-II), 11 (GL-III), and 15 (GL-IV) were subjected to negative-ion FAB-MS analysis (Fig. 4, A-C, respectively). The glycolipid in fraction 8 was composed of five hexoses (of which at least one was terminal) and two internal N-acetylhexosamines. One of the N-acetylhexosamines was located next to a terminal hexose. The FAB-MS obtained from fraction 11 (Fig. 4 B) showed the presence of five hexoses, two N-acetylhexosamines, and one Fuc. The Fuc and at least one hexose were terminally placed. This glycolipid had two dominating ceramide species which differed by two carbon atoms. The mass spectrum of fraction 15 closely resembled that of fraction 11. However, fraction 15 contained one dominating ceramide species with six carbons less than the light ceramide species of fraction 11. This conclusion was also consistent with the difference in mobility on the TLC plate. Because of the high intensity of m/z 1736.8 and 1387.8 relative to m/z 1590.8 and 1225.7, it is probable that the molecule contained two branching points. This irregular intensity of ions is also seen for fraction 11 (Fig. 4 B). Fractions 11 and 12 were pooled and analyzed by EI-MS as the methylated derivative (Fig. 4 D). Molecular ions of m/z 2347.1 and 2375.2 were obtained which correspond to ceramides with sphingosine and 22- and 24-carbon fatty acids, respectively. These ceramides are also seen at m/z 632.6 and 660.6. Large peaks appear at 638.3, 1087.5, and 1291.5, which correspond to the following oxonium ions: HexHexNAcFuc, (Hex)2(HexNAc)2Fuc, and (Hex)3(HexNAc)2Fuc, respectively. These three oxonium ions, [Ox]+, lose either Fuc [Ox-Fuc+H]+or Hex [Ox-Hex+H]+giving rise to the following sets of ions: m/z 450.2 and 420.2 from 638.3, m/z 899.4 and 869.4 from 1087.5, and m/z 1073.5 and 1103.5 from 1291.5. Ions from terminal Hex (219.1 and 219.1-32 = 187.1) and Fuc (189.1 and 189.1-32 = 157.1) were also obtained. The presence of a set of m/z at 692.6 (Cer), 2173.4 (M-Hex), and 2203.1 (M-Fuc), indicates that minor amounts of the glycolipid might carry phytosphingosine. Gas chromatography-mass spectroscopic analysis of partially methylated alditol acetates was performed on the pooled fractions 13-15 in parallel to analysis of the reference compounds gangliotetraosylceramide and Ley. The following monosaccharides were identified; terminal Gal and Fuc, 4-substituted Gal and Glc, 3-substituted Gal, and 3,4-substituted GlcNAc. A second di-substituted HexNAc, with a longer retention time, but similar MS, was obtained. The anomeric region of the 1H NMR spectrum (Fig. 5 C, Table I) of GL-III (HPLC fraction 10) shows two α-signals (J ≤ 4 Hz) at 4.82 and 4.80 ppm (a and b in Fig. 5 C) corresponding to Fucα3 and Galα4, respectively. These signals can be conclusively assigned from the nuclear Overhauser enhancement (NOE, magnetic dipole coupling through space) experiments. The NOE between the 4.80 ppm signal and a typical Gal H4 signal at 3.82 ppm (40Scarsdale J.N. Yu R.K. Prestegard J.H. J. Am. Chem. Soc. 1986; 108: 6778-6784Crossref Scopus (49) Google Scholar, 41Dabrowski J. Hanfland P. Egge H. Biochemistry. 1980; 19: 5652-5658Crossref PubMed Scopus (129) Google Scholar), as well as a quartet at 3.76 ppm (H2 of Galα), are shown in Fig. 5 B. The Fuc H5 gave rise to a quartet of 4.67 ppm (c in Fig. 5 C). Of the two HexNAcβ signals at 4.61 and 4.49 ppm (d and e in Fig. 5C) the former showed NOEs (Fig. 5 A) to Galα H5 (4.19 ppm) and Galα H4 (3.98 ppm) and also a large NOE to a signal at 3.59 ppm. This corresponds well with published values for the Galα H3 (40Scarsdale J.N. Yu R.K. Prestegard J.H. J. Am. Chem. Soc. 1986; 108: 6778-6784Crossref Scopus (49) Google Scholar, 41Dabrowski J. Hanfland P. Egge H. Biochemistry. 1980; 19: 5652-5658Crossref PubMed Scopus (129) Google Scholar), and thus confirms the 4.61 ppm signal as arising from the GalNAcβ3 of a globo-core structure. The signal at 4.49 ppm is within the region for GlcNAc shifts, see for example, Ref. 42Hanfland P. Egge H. Dabrowski U. Kuhn S. Roelcke D. Dabrowski J. Biochemistry. 1981; 20: 5310-5319Crossref PubMed Scopus (86) Google Scholar. Two well separated Galβ signals were seen at 4.29 and 4.26 ppm (f and g in Fig. 5C), respectively, the latter arises from the internal Galβ4 linked to Glc, which in turn was seen at 4.19 ppm, overlapping two other signals. The former Gal is consistent with a terminal Galβ4 of a Lexterminal (43Levery S.B. Nudelman E.D. Andersen N.H. Hakomori S.-i. Carbohydr. Res. 1986; 151: 311-328Crossref PubMed Scopus (52) Google Scholar). The other signals at 4.19 ppm (overlapping the Glc anomer) were triplets from the H5 of Galα and a terminal Galβ3 (41Dabrowski J. Hanfland P. Egge H. Biochemistry. 1980; 19: 5652-5658Crossref PubMed Scopus (129) Google Scholar). NOEs were measured from all anomeric protons (not shown), and they confirmed the saccharide sequence as determined above. Since most non-anomeric signals are unassigned, the NOE analysis was based on the multiplicity of signals and their chemical shifts compared to model structure
Referência(s)