Electron microscopic observation of crystals of Escherichia coli K-12 lipopolysaccharide

Previously we showed that Salmonella minnesota Re and Ra lipopolysaccharides (LPSs) and Escherichia coli K-12 LPS formed three-dimensional crystals, either hexagonal plates or solid columns, when they were precipitated by the addition of 2 volumes of 95% ethanol containing 375 mM MgCl2 and incubated in 70% ethanol containing 250 mM MgCl2, and stored at 4 C for 10 days. Later, Escherichia coli K-12 LPS thus treated was found to form discoid crystals as well as hexagonal plate crystals and solid column crystals. Analysis by electron diffraction of the discoid crystals from the direction perpendicular to the basal plane showed that they consisted of hexagonal lattices with the a axis of 4.62 A. This result was quite the same as that of the hexagonal plate crystals. Electron micrographs of the edges of the discoid crystals revealed stacked sheets of the hexagonal plate crystals. From these results it was concluded that formation of the discoid crystals results from irregular overlapping of the hexagonal plate crystals.     

Crystallization of lipopolysaccharide from a Salmonella typhimurium semi-rough (SR) mutant.

Salmonella typhimurium SR-form lipopolysaccharide (LPS), consisting of a single repeating unit of the O-antigenic polysaccharide, linked to the R-core consisting of oligosaccharide that is, in turn, linked to lipid A, formed crystals whose shapes were hexagonal plates, discoids, and solid columns when precipitated by the addition of 2 volumes of 95% ethanol containing 375 mM MgCl2 and kept in 70% ethanol containing 250 mm MgCl2 at 4 C for 10 days. Among these crystals, the basic form is considered to be the hexagonal plates. Analyses of hexagonal plate crystals showed that they consist of hexagonal lattices with a lattice constant (a axis) of 4.62 A and longitudinal axis (c axis) of approximately 100 A. In X-ray diffraction patterns in the low-angle region, crystals of S. typhimurium SR-form LPS exhibited much less distinct reflections when compared with crystals of synthetic Escherichia coli-type lipid A. In contrast to the previous finding that S. minnesota S-form LPS possessing the O-antigenic polysaccharide does not crystallize under the same experimental conditions as used in the present study, the presence of a single repeating unit of the O-antigenic polysaccharide does not inhibit crystallization.     

Relationship between Mg2(+)-induced hexagonal assembly of R-form lipopolysaccharides and chemical structure of their R-cores

The relationship between formation of the Mg2(+)-induced hexagonal lattice structure by R-form lipopolysaccharides (LPS) and chemical structure of their R-cores was investigated using different kinds of R-form LPS from a series of mutants of Salmonella minnesota or S. typhimurium. The optimal experimental condition for formation of the hexagonal lattice structure was to suspend LPS preparations, from which cationic material was removed by electrodialysis, in 50 mM tris (hydroxymethyl) aminomethane buffer at pH 8.5 containing 10 mM MgCl2. Under this experimental condition, Rb1 LPS formed the hexagonal lattice structure with the lattice constant of 14.0 +/- 0.2 nm. Ra LPS, which possesses the full length of R-core, also formed the hexagonal lattice structure but its lattice constant was larger (18.1 +/- 0.2 nm) than that of Rb1 LPS (the lattice structure by Ra LPS was looser than that by Rb1 LPS). All the other R-form LPS preparations tested, RcP+, PcP-, Rd1P-, and Re LPS, whose R-cores are shorter than that of Rb1 LPS, did not form the hexagonal lattice structure, but formed membranous structures showing various shapes which consisted of multiple bilayer structures. Failure to form the hexagonal lattice structure was the common feature of these kinds of R-form LPS irrespective of temperature at which the LPS suspensions in 10 mM MgCl2-50 mM Tris buffer were incubated. From the results of the present study it was concluded that capability of R-form LPS to form the hexagonal lattice structure has a close correlation with the chemical structure of their R-cores.     

Polymorphism of Crystals of Salmonella minnesota Re and Ra Lipopolysaccharides

Salmonella minnesota Re and Ra lipopolysaccharides (LPSs) formed three-dimensional crystals when they were precipitated by the addition of 2 volumes of 95% ethanol containing 375 mM MgCl₂ and incubated in 70% ethanol containing 250 mM MgCl₂ at 4 C. Besides typical shapes of crystals, hexagonal plates and solid columns, which were already reported (J. Bacteriol. 172: 1516–1528 (1990)), the LPSs thus treated formed crystals possessing various shapes such as square or rectangular plate, lozenge plate, discoid, and truncated hexangular pyramid forms. Electron diffraction patterns from all these crystals except square or rectangular plate crystals obtained by electron irradiation from the direction perpendicular to the basal plane were essentially the same as those from hexagonal plate crystals, indicating that they consist of hexagonal lattices with the lattice constant of 4.62 Å. From these results as well as the results of electron microscopic observations of these crystals, it was concluded that all these crystals except square or rectangular plate crystals are composed of hexagonal plate sheets as the basic structural units. Square or rectangular crystals were assumed to correspond to the {1011} planes of solid hexagonal column crystals.     

In vitro hexagonal assembly of lipopolysaccharide of Escherichia coli K-12

We examined Escherichia coli K-12 lipopolysaccharide (LPS), which is known to be an R-form LPS, for its ability to form a hexagonal lattice structure in vitro. The LPS from E. coli K-12 strain JE1011 did not form a hexagonal lattice structure when it was precipitated by addition of two volumes of 10 mM MgCl2-ethanol, but it did form such a structure when it was electrodialyzed and then converted to the magnesium or calcium salt form. The lattice constant of the magnesium salt form was 15.2 +/- 0.3 nm and that of the calcium salt form 18.5 +/- 0.3 nm. Since prior treatment of the LPS with proteinase K in the presence of sodium dodecyl sulfate did not affect its capability of hexagonal assembly, the lattice formation by the LPS does not require the presence of proteins.     

Crystallization and Analyses of Crystals of Various Chemotypes of R-Form Lipopolysaccharides from Salmonella Spp.

Various chemotypes (Re, Rd₂, Rd₁P^–, Rd₁, RcP^–, Rc, Rb₃, Rb₂, Rb₁, and Ra) of R-form lipopolysaccharides (LPSs) of Salmonella spp. were crystallized by treatment with 70% ethanol containing 250 mM MgCl₂, and crystals of the LPSs were observed electron microscopically and analyzed by electron diffraction and synchrotron X-ray diffraction. All the LPSs tested formed three-dimensional crystals showing very similar shapes; hexagonal plate, solid column, discoid, square or rectangular plate, lozenge plate and truncated hexangular or rectangular pyramid forms. Electron diffraction patterns from the hexagonal plate crystals of all these LPSs obtained by electron irradiation from the direction perpendicular to the basal plane showed that they consist of hexagonal lattices with the lattice constant of 4.62 Å. The crystals of all the LPSs thus formed gave ring-like X-ray diffraction patterns because of their small sizes. The long-axis values were calculated from the X-ray diffraction patterns from crystals of all the LPSs in the low-angle region and they corresponded roughly to the length of the proposed primary chemical structures of the R cores of the LPSs. The volume occupied by a single molecule of all the LPSs were calculated from the molecular weights based on the proposed structures and the crystallographic data obtained by electron diffraction, X-ray diffraction, and density determination.     

Neutralization of Shwartzman-inducing activity by antibodies recognizing the Re core or lipid A structures of lipopolysaccharide from Salmonella minnesota R595 and Pseudomonas vesicularis JCM1477

Antibodies recognizing the Re core or lipid A structures of lipopolysaccharide (LPS) derived from Salmonella minnesota R595 and Pseudomonas vesicularis JCM1477 were tested for the ability to neutralize the preparatory activity of endotoxin using the local Shwartzman reaction. Shwartzman-inducing activity of R595 LPS (Re-form) was strongly suppressed when the LPS was incubated with the rabbit anti-R595 antiserum or the purified IgG antibody which recognizes core region of the LPS. The antiserum also suppressed the preparatory activity of LPS from S. typhimurium SL1102 (Re) and Escherichia coli F515 (Re), but not that of either S. typhimurium LT-2 (S) LPS or R595 lipid A. Moreover, it was found that the murine monoclonal antibody (MAb), SmRe100G (IgG2a) which recognizes the core region of R595 LPS, significantly suppressed the preparatory activity of R595 LPS. Both conventional antibodies specific to R595 lipid A, which contains a 1,4'-bisphosphorylated beta-D-glucosaminyl-alpha-D-glucosamine disaccharide structure, and JCM1477 lipid A, which contains a monophosphorylated 3-amino-D-glucosamine disaccharide structure, neutralized the preparatory activity of homologous and a closely related lipid A, but not that of LPS. In addition, it was observed that MAb Sm5G (IgG2b) specific to enterobacterial lipid A preparations (especially R595 lipid A) neutralized the preparatory activity of R595 lipid A, although the effect was somewhat weak as compared with that of rabbit antiserum. These results suggest that anti-Re LPS antibody binding to the core of Re LPS is involved in suppressing the endotoxic activity of Re LPS, and that the direct binding of anti-lipid A antibody to some specific epitopes of lipid A is important in neutralizing the endotoxic activity.     

Interaction of lipopolysaccharide with detergents and its possible role in the detergent resistance of the outer membrane of Gram-negative bacteria.

In the presence of MgCl2, amounts of detergents which disrupted phospholipid vesicles caused lipopolysaccharide I from Proteus mirabilis to aggregate and form vesicular, membrane-like structures. Vesicle formation with P. mirabilis lipopolysaccharide II containing longer O-polysaccharide chains was extremely poor. Lipopolysaccharides of Salmonella minnesota R mutants (chemotypes Ra, Rc and Re) displayed a growing tendency for vesicle formation with increasing deficiency of the R core polysaccharide. Lipopolysaccharides of chemotypes Rc and Re produced vesicles even in the absence of MgCl2 and detergent. Spherical aggregates consisting of P. mirabilis lipopolysaccharide I MgCl2 and detergent were unable to either entrap or retain [14C]-sucrose, [3H=inulin or [3H]dextran. On the other hand, S. minnesota R mutant lipopolysaccharides of chemotypes Rc and Re could entrap all three saccharides and retain them for at least short periods of time. Leakage of [3H]-inulin out of re-lipopolysaccharide vesicles was greatly retarded by addition of MgCl2 to the vesicle system. Incorporation of P. mirabilis lipopolysaccharide I or S. minnesota Rc lipopolysaccharide into phospholipid vesicles protected these model membranes from disruption by detergent. This suggested a similar protective function of lipopolysaccharide in the outer membrane of enteric bacteria against the action of surfactants occurring in their normal intestinal habitat.     

Crystallization of Synthetic Escherichia coli-Type Lipid A

Synthetic Escherichia coli-type lipid A formed hexagonal plate crystals when it was precipitated by the addition of 2 volumes of 95% ethanol containing 375 mM MgCl₂ and incubated in 70% ethanol containing 250 mM MgCl₂ at 4 C for 10 days. Analyses of crystals by electron diffraction and synchrotron X-ray diffraction showed that crystals consist of hexagonal lattices with the lattice constant (a side of the lozenge as a unit cell on the basal plane) of 4.62 Å and the longitudinal axis (perpendicular to the basal plane) of 49.3 ±1.3 Å. Results suggest that the previous finding that various kinds of R-form lipopolysaccharides crystallized but free lipid A isolated by acid hydrolysis from Re lipopolysaccharide did not crystallize under the same experimental conditions (Kato et al, J. Bacteriol., 172: 1516-1528, 1990) is due to structural changes of lipid A occurring during the procedure of isolation of free lipid A.     

大肠埃希菌脂多糖诱导的血小板降解及急性休克:O抗原在LPS诱导的免疫反应中的作用

目的探讨LPS中的0抗原部分与其它部分在血小板反应中的作用。方法给BALB／c小鼠注人大肠埃希菌野生株E．coli O8、O9、K-12（不含有O抗原）及2株重组变异的K-12株（携带编码O8、O9的O抗原rfb基因）。结果K-12的LPS引起血小板反应及急性休克能力较弱,O8及O9引起一定的反应,而这2种重组的LPS,即在K-12的LPS上带有O8或O9的O抗原．显示出极强的活性。静脉注入补体C5的阻止剂后,重组株LPS的作用消失了。而且在缺乏补体C5小鼠DBA／2中,重组的LPS能引起血小板的聚集但不能降解,也不能引起休克症状。结论诱导血小板反应及急性休克的能力依赖于LPS结构;O抗原及R核心抗原是表现活性的必要结构;LPS诱导的血小板反应及急性休克依赖补体系统。     

The effect of composition of the core region of Escherichia coli K-12 lipopolysaccharides on the surface properties of cells

The pH dependences of electrokinetic potentials (EKP) of the cells of two Escherichia coli K-12 strains (D21 and JM 103) with known lipopolysaccharide (LPS) core composition have been determined by the method of microelectrophoresis. At pH 4.6-5.2, the negative surface charge of the cells with Re core LPS was reliably higher. It was shown that the interaction of bacteria with lysozyme results in a decrease of optical density of suspensions due to higher sensitivity of the cells with complete LPS core to hypotonic shock. LPS release from bacterial cell wall depended also on bacterial LPS core composition and increased with LPS core extension. Electrokinetic measurements and the study of the interaction of cells with lysozyme suggest that higher negative surface charge of E. coli JM 103 cells (Re type LPS) is associated with higher quantity and density of LPS packing in the cell wall as compared with the cells of E. coli D21 (Ra type LPS).     

Electrophoretic and chemical characterization of lipopolysaccharides of Vibrio parahaemolyticus.

Lipopolysaccharides (LPSs) isolated from three Kanagawa-positive and three negative strains of Vibrio parahaemolyticus were characterized by using electrophoretic, immunochemical, and chemical methods. The results of this study indicated that the LPSs of all six strains of V. parahaemolyticus examined did not have an O-specific side chain. These V. parahaemolyticus LPSs appeared to have molecular weights similar to that of the rough-type (Ra) LPS of Salmonella typhimurium TV-119 and might just contain lipid A and a core region. However, the microheterogeneity of V. parahaemolyticus LPS observed was greater than that of S. typhimurium LPS. The profile of V. parahaemolyticus LPS consisted of closely spaced triplet or quadruplet bands, but that of S. typhimurium consisted of doublet bands. Slower-moving bands appeared on sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels only when large amounts of V. parahaemolyticus LPS were loaded. These bands were proven to be the aggregates of the fastest-moving low-molecular-weight bands by re-electrophoresis. The banding pattern of V. parahaemolyticus LPSs produced on nitrocellulose membranes by immunoblotting indicated that the V. parahaemolyticus LPSs did not have an O-specific side chain. The low ratio of total carbohydrate to lipid A of V. parahaemolyticus LPSs also suggested that they were like rough-type LPS. The mobility and profile of V. parahaemolyticus LPS on sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel and its chemical composition were closely related to the serotype of a specific strain but not with the Kanagawa phenomenon.     

Contributions of polysaccharide and lipid regions of lipopolysaccharide to the recognition by spike G protein of bacteriophage phi X174

A histidine-tagged G protein of bacteriophage phi X174 (HisG) bound strongly with lipopolysaccharide (LPS) of Escherichia coli C, one of a phi X174-sensitive Ra strain. The dissociation constant, Kd, was measured to be 0.16 +/- 0.04 microM by fluorometric titration. HisG showed slightly less affinity to LPSs of the insensitive Rc and Rd 2 strains having shorter R-core polysaccharide sequences than that of the sensitive Ra strains. The difference between the two types of LPS was demonstrated by CD spectra; LPSs of the sensitive strains increased the signal intensity for beta-sheet, while the insensitive strains decreased it. The chemically degraded LPS derivatives lacking a hydrophobic lipid region showed much less affinity to HisG, indicating the importance of the lipid region of LPS for strong binding with HisG. On the other hand, since the degraded derivatives increased the intensity of CD spectra, the polysaccharide region is thought to contribute to the conformation change of the protein.     

Staining of O-specific polysaccharide chains of lipopolysaccharides with ruthenium red

S-form lipopolysaccharides (LPS) from Klebsiella strain LEN-1 (O3: K1-) and from Salmonella minnesota strain 1114 were positively stained with ruthenium red, whereas R-form LPS from Klebsiella strain LEN-111 (O3-: K1-) and Ra, Rb1, RcP+, Rd1P-, and Re LPS from the respective mutant strains of S. minnesota were not or only faintly stained by such treatment. From these results it was concluded that ruthenium red stains the O-specific polysaccharide chains of LPS. The appearance of stained preparations of S-form LPS suggested that the material responsible for this positive staining corresponded to the surface projections which were seen by the negative staining technique as attached to the ribbon-like structures and spherules of the LPS.     

In vitro hexagonal assembly of R-form lipopolysaccharides: effect of pH on the Mg+2-mediated hexagonal assembly

The R-form lipopolysaccharide (LPS) from Escherichia coli K-12, from which cationic material had been removed by electrodialysis and the pH of which had fallen to 3.6, formed a rough hexagonal lattice structure with the lattice constant of about 19 nm. The rough hexagonal structure was maintained in buffers at pH 5 or lower but disintegrated into the ribbon-like structures in buffers at pH 6 or higher. However, in the presence of 10 mM Mg2+, the hexagonal lattice structure was not disintegrated even at alkaline pH levels but conversely it became more dense. At pH 8.3 to 8.9, the hexagonal lattice structure with the shortest lattice constant (15 nm) was formed. The same optimal pH levels were obtained for formation of the dense hexagonal lattice structure (lattice constant, 14 to 15 nm) by the electrodialyzed LPS from Klebsiella pneumoniae strain LEN-111 (O3-:K1-). The ability of Mg2+ to induce formation of the dense hexagonal lattice structure of the K-12 LPS depends upon the presence of buffers showing the optimal pH levels, since a very high concentration of Mg2+ such as 500 mM was required for the lattice formation in distilled water. The amount of the magnesium bound to the K-12 LPS did not significantly differ throughout the pH range of 3 to 9. Therefore, the optimal pH range is another essential factor for formation of the dense hexagonal lattice structure of the LPS in addition to binding of the magnesium to the LPS.     

Different bacterial lipopolysaccharides as toxicants and stressors in the shrimp Palaemon elegans

This study compares the in vivo haemocytic response of shrimp, Palaemon elegans (Rathke) to different types of LPS injection. In particular it investigates to what degree and speed the haemocytopenia varies between LPSs from different sources. It further compares the tolerated doses of different LPSs in these animals and finds substantial differences in the various toxicity types. The work then relates this to blood glucose levels and stress-linked variations in glycaemic status. The order of LPS decreasing toxicity determined by LD50 at 96 h was: Salmonella enteritidis, Serratia marcescens, Pseudomonas aeruginosa 10, Escherichia coli K-235 and E. coli 0111:B4. Eyestalkless animals were more sensitive to LPS. The effects of injected LPS on circulating total blood cell count (THC) was tested. The results show that LPS caused a decrease in THC 8 h after injection and then the THC returned to the initial level and this effect depended on the LPS tested. E. coli K-235 was the most effective in causing haemocytopenia followed by E. coli 0111:B4, S. enteritidis, S. marcescens, and P. aeruginosa 10. Moreover, LPS-induced increases in the blood glucose level and the time and dose related curves of response obtained depended on the type of LPS tested. E. coli K-235 LPS was again the most effective in elevating blood glucose followed by E. coli 0111:B4, S. marcescens, S. enteritidis and then P. aeruginosa 10. No significant hyperglycaemia was induced in eyestalkless animals. An inverse order relationship between toxicity (LD50) and stress responses (hyperglycaemia and THC decrease) may suggest a defensive and adaptive role of the latter in occasional septicaemia.     

Estimation of mean molecular weight of Enterobacteriaceae lipopolysaccharides using gas chromatography for determination of fatty acids

In the paper, we propose a method for estimation of the mean molecular weight of lipopolysaccharide, which is important for accuracy of endotoxin activity investigation. In our study, it was assumed that lipid A portion in Enterobacterial lipopolysaccharide is substituted by four 3-hydroxytetradecanoic acid residues. Lipopolysaccharides of S, Ra, Rc and Re chemotypes being laboratory preparations as well as purchased from Sigma were investigated. Fatty acids were determined by of gas chromatography as methyl esters according to the procedure described by Wollenweber and Rietschel. Mean molecular weight was calculated by the formula: MMW = [formula: see text]. A high agreement between the estimated and the theoretical molecular weight values was demonstrated in the case of Salmonella minnesota R595 (Re) LPS preparation. As expected, LPS heterogeneity increase together with enlargement of polysaccharide chain length which is visible in electrophoregrams also. Except for LPS mean molecular weight estimation, the method allows its detection in various preparations and samples, distinguishing of R and S LPS forms as well as the determination of mean length of O-specific chain in lipopolysaccharides which structures are known.     

Identification of lipopolysaccharide-binding proteins in 70Z/3 cells by photoaffinity cross-linking

A radioiodinated, photoactivatable derivative of Salmonella minnesota Re595 lipopolysaccharide (LPS) was used to label LPS-binding proteins in 70Z/3 cells. The labeled proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and visualized by autoradiography. 125I-Labeled-2-(p-azidosalycylamido)1,3'-dithiopropionamide S. minnesota Re595 LPS (125I-ASD-Re595) labeled a limited number of proteins. The most prominent of these had a apparent molecular mass of 18 kDa. Less prominent labeling of 25- and 28-kDa proteins was also seen. Labeling was saturated by 5 micrograms/ml 125I-ASD-Re595 and was inhibited by a 10-100-fold excess of unlabeled LPS or lipid A. Labeling was maximal within 30 min at 37 degrees C; much less labeling occurred at lower temperatures. The proteins labeled with 125I-ASD-Re595 appear to be on the surface of the cell, since they can be digested by trypsin and were found in the membrane fraction of the cell but not in the cytosol. Studies with competitive inhibitors suggested that the proteins bind to the lipid A region of the LPS molecule. Biologically inactive lipid A analogs were poor inhibitors of labeling, suggesting that the LPS-binding proteins could discriminate between active lipid A and inactive analogs. These studies suggest that the 18- and 25-kDa proteins bind specifically to the lipid A region of the LPS molecule and should be considered as candidates for a functional LPS receptor.     

Blocking of bacteriophages phi W and phi 5 with lipopolysaccharides from Escherichia coli K-12 mutants.

In the preceding paper we presented a formula for the composition of lipopolysaccharides (LPS) from Escherichia coli K-12. This formula contains four regions defined from analyses of LPS from four key strains, the parent and mutants which had lost one, two, or three regions of their carbohydrates. Support for the formula was derived from the susceptibility of the key mutants to several bacteriophages. One of these, phage phi W, was found specific for strains which had lost region 4. In this paper we described inactivation in vitro of phage phi W and its host-range mutant phi 5, using LPS devoid of regions 2 to 4. The blocking of phi W was found to require about 0.15 M concentrations of monovalent cations and to be inhibited by low concentrations of calcium and magnesium ions. One particle of phage phi W required 2 times 10-16 g of LPS devoid of region 4 for stoichiometric inactivation. Phage phi 5 was blocked by both heptose-less LPS (devoid of regions 2 to 4) and glucose-less LPS (devoid of regions 3 to 4) but was unaffected by LPS devoid of region 4. LPS from a heptose-less mutant of Salmonella minnesota showed the same inactivation ability as did LPS from heptose-less strains of E. coli K-12. Lipid A was prepared from LPS containing all four regions. Such lipid A was found to inactivate phi 5, whereas both the polysaccharide moiety as well as the intact LPS were without effect. It is suggested that lipid A is part of the receptor site for phage phi 5.     

Evaluation of the immunological activity of a vaccine from mutant Salmonella minnesota R 595 on a model of experimental Pseudomonas aeruginosa infection in mice

The comparative study of heated corpuscular vaccines prepared from S. minnesota mutant R 595 with defective lipopolysaccharide (LPS), chemotype Re, derived from S. minnesota strain SF 1111 with unchanged LPS, and from P. aeruginosa strain PA 103, was carried out. In contrast to the vaccine from S. minnesota strain SF 1111, the vaccine prepared from the mutant with chemotype Re induced the development of cell-mediated and humoral immunity to P. aeruginosa, and its immunogenicity was close to that of the vaccine from P. aeruginosa strain PA 103.