Synthetic Salmonella-type lipid A antigen with high serological specificity

Structural and biological characterization of highly purified hepta-acyl lipid A present in the lipopolysaccharide of the Salmonella enterica sv. Minnesota Re deep rough mutant strain R595

One major component of the Salmonella enterica sv. Minnesota Re deep rough mutant (strain R595) lipopolysaccharide is hepta-acyl lipid A (LAhepta). In a recent publication [Tanamoto K-I, Azumi S. Salmonella-type heptaacylated lipid A is inactive and acts as an antagonist of lipopolysaccharide action on human line cells. J Immunol 2000; 164: 3149—3156] the corresponding synthetic hepta-acyl lipid A (compound 516) was reported to be agonistically inactive but to rather suppress pro-inflammatory activation by the endotoxichexa-acyl lipid A (LAhexa, compound 506) and S-form LPS from Escherichia coli in the human macrophage-like cell lines THP-1 and U937. These results, however, were in contrast to previous findings with human mononuclear cells (hMNC) isolated from peripheral blood, in which compound 516 was found to be an agonist, expressing low, but significant,cytokine-inducing activity as compared to LAhexa. We have investigated the structure of natural LA hepta from the S. enterica sv. Minnesota Re deep rough mutant strain (R595) by TLC immunoblot, MALDI-TOF mass spectrometry and NMR spectroscopy. Using these techniques, the structural identity between LAhepta and the synthetic compound 516 was confirmed. In corroboration of previous findings with studies employing compound 516, purified LA hepta was found to induce the production of TNF- , IL-1 and IL-6 in hMNC, thus displaying moderate agonistic activity. Furthermore, we showed that LAhepta agonistically activated nuclear translocation of NF- B in THP-1 cells, thus clearly ruling out the possibility that LAhepta is an antagonist and that its biological activity is influenced by the type of human myeloid cells used for testing endotoxicity(hMNC versus THP-1 cells).

Cellular recognition of trimyristoylated peptide or enterobacterial lipopolysaccharide via both TLR2 and TLR4

Evidence for specific and direct bacterial product recognition through toll-like receptors (TLRs) has been emphasized recently. We analyzed lipopeptide analogues and enterobacterial lipopolysaccharide (eLPS) for their potential to activate cells through TLR2 and TLR4. Whereas bacterial protein palmitoylated at its N-terminal cysteine and N-terminal peptides derived thereof are known to induce TLR2-mediated cell activation, a synthetic acylhexapeptide mimicking a bacterial lipoprotein subpopulation for which N-terminal trimyristoylation is characteristic (Myr(3)CSK(4)) activated cells not only through TLR2 but also through TLR4. Conversely, highly purified eLPS triggered cell activation through overexpressed TLR2 in the absence of TLR4 expression if CD14 was coexpressed. Accordingly, TLR2(-/-) macrophages prepared upon gene targeting responded to Myr(3)CSK(4) challenge, whereas TLR2(-/-)/TLR4(d/d) cells were unresponsive. Through interferon-gamma (IFNgamma) priming, macrophages lacking expression of functional TLR4 and/or MD-2 acquired sensitivity to eLPS, whereas TLR2/TLR4 double deficient cells did not. Not only TLR2(-/-) mice but also TLR4(-/-) mice were resistant to Myr(3)CSK(4) challenge-induced fatal shock. d-Galactosamine-sensitized mice expressing defective TLR4 or lacking TLR4 expression acquired susceptibility to eLPS-driven toxemia upon IFNgamma priming, whereas double deficient mice did not. Immunization toward ovalbumin using Myr(3)CSK(4) as adjuvant was ineffective in TLR2(-/-)/TLR4(-/-) mice yet effective in wild-type, TLR2(-/-), or TLR4(-/-) mice as shown by analysis of ovalbumin-specific serum Ig concentration. A compound such as Myr(3)CSK(4) whose stimulatory activity is mediated by both TLR2 and TLR4 might constitute a preferable adjuvant. On the other hand, simultaneous blockage of both of the two TLRs might effectively inhibit infection-induced pathology.

Conformational studies of synthetic lipid A analogues and partial structures by infrared spectroscopy

Synthetic lipid A analogues and partial structures were analyzed and compared with natural hexaacyl lipid A from E. coli applying Fourier transform infrared spectroscopy. The investigations comprised (i) the measurement of the beta <=> alpha phase transition of the acyl chains via monitoring of the symmetric stretching vibration of the methylene groups, (ii) an estimation of the supramolecular aggregate structures evaluating vibrations from the interface like ester carbonyl and applying theoretical calculations (iii) a determination of the inter- and intramolecular conformations monitoring functional groups from the interface and the diglucosamine backbone (ester carbonyl, phosphate). The phase transition temperature Tc was found to be nearly a linear function of the number of acyl chains for most bisphosphoryl compounds indicating comparable packing density, whereas the deviating behaviour of some samples indicated a higher packing density. From the determination of the supramolecular aggregate structures (cubic, HII) of natural hexaacyl lipid A by X-ray small-angle diffraction, the existence of the same aggregate structures also for the synthetic hexaacyl lipid A was deduced from the nearly identical thermotropism of the ester carbonyl band. From this, a good approximation of the supramolecular structures of all synthetic samples was possible on the basis of the theory of Israelachvili. The analysis of the main phosphate band, together with that of the Tc data and former colorimetric results, allowed the establishment of a model of the intermolecular conformations of neighbouring lipid A/LPS molecules. The biological relevance of the findings is discussed in terms of the strongly varying biological activity (between high and no activity) of the samples.

Legionella pneumophila growth restriction and cytokine production by murine macrophages activated by a novel Pseudomonas lipid A

Peritoneal exudate macrophages from A/J mice activated by purified lipid A preparations from Pseudomonas vesicularis, which contain 2,3-diamino-2,3-dideoxy-D-glucose disaccharide phosphomonoester as the lipid A backbone, restricted the growth of Legionella pneumophila, an intracellular opportunistic bacteria which readily grows in otherwise permissive macrophages from susceptible A/J mice and induced production of the proinflammatory cytokines interleukin 1 and tumor necrosis factor alpha. Activation of the macrophages was similar to that which occurred after stimulation with more conventional lipid A from other bacteria such as salmonellae. A purified fraction A3 preparation from the Pseudomonas lipid A, which lacked only 1 mol of amide-linked fatty acid, in comparison with another fraction (A2), which contained the fatty acid, also markedly activated the usually permissive macrophages from susceptible A/J mice to resist growth of the legionellae. The fraction A3 also induced both interleukin and tumor necrosis factor alpha. These results show that this novel lipid A from P. vesicularis can activate macrophages to resist infection with an opportunistic bacterium in a manner similar to that induced by conventional enterobacterial lipid A and that the hydrophobic portion of this Pseudomonas molecule may have an important role in activation of macrophages.

Legionella pneumophila growth restriction in permissive macrophages cocultured with nonpermissive lipopolysaccharide-activated macrophages

Macrophages can be activated by lipopolysaccharides (LPS) from gram-negative bacteria to evince a number of biological activities, including increased resistance to intracellular infection by opportunistic bacteria. In the present study, intraperitoneal injection of LPS into A/J mice activated peritoneal macrophages so that they resisted subsequent in vitro infection with Legionella pneumophila. Coculture of these macrophages with those from nontreated A/J mice converted the entire population of cells from permissive to nonpermissive. This effect did not appear to be mediated by soluble factors released from the LPS-treated macrophages, since the levels of interleukins-1 and -6 and tumor necrosis factor alpha produced by the macrophages were not found to be markedly elevated at the time when the macrophages from the LPS-treated mice were most effective in converting normal macrophages to nonpermissiveness. Furthermore, macrophages from mice injected intraperitoneally with either interferon or tumor necrosis factor alpha did not evince nonpermissiveness and also did not have the ability to convert normal spleen cells to nonpermissiveness. Polymyxin B, a known inactivator of LPS activity, did not inhibit the macrophages from the LPS-treated mice from inducing this resistance. It seemed unlikely that free LPS released from the macrophages mediated this effect. The results of this study thus showed that macrophages activated by LPS in vivo can evince nonpermissiveness for Legionella growth in vitro and also can induce macrophages from normal, permissive mice to become nonpermissive for Legionella growth in vitro.

An artificial glycoconjugate containing the bisphosphorylated glucosamine disaccharide backbone of lipid A binds lipid A monoclonal antibodies

Monoclonal antibodies (MAbs) against lipid A, the endotoxic component of lipopolysaccharide (LPS) of gram-negative bacteria, are presently discussed as therapeutic agents against lethal gram-negative infections; however, their binding specificities are controversial. We have isolated from the LPS of Escherichia coli J-5 the 1,4'-bisphosphorylated beta 1-->6-linked glucosamine disaccharide backbone of its lipid A moiety, which was covalently linked to bovine serum albumin. It was shown by solid-phase enzyme immunoassay that one antibody (MAb A6) bound equally well to the glycoconjugate and synthetic E. coli-type lipid A over a broad range of antigen concentrations whereas two other MAbs (IC3 and S1-15) bound better to the conjugate at low antigen concentrations and better to the lipid A when high concentrations of antigen were used. This proves in a direct way that there exist lipid A MAbs with different specificities which bind to epitopes in the hydrophilic backbone of lipid A and which do not require the presence of fatty acids.

Characterization of the epitope specificity of murine monoclonal antibodies directed against lipid A

A series of monoclonal antibodies directed against lipid A was characterized by using synthetic lipid A analogs and partial structures. These compounds vary in phosphate substitution, acylation pattern (type, number, and distribution of fatty acids), and, in the case of monosaccharides, in their backbone glycosyl residue. The monoclonal antibodies tested could be subdivided into five groups according to their reactivity patterns. One group reacted exclusively with 1,4'-bisphosphoryl lipid A, and a second also reacted with 4'-monophosphoryl lipid A. Two further groups recognized either 4-phosphoryl or 1-phosphoryl monosaccharide partial structures of lipid A. The fifth group reacted with 4-phosphoryl monosaccharide structures and with phosphate-free compounds. Antibodies reactive with monosaccharide structures also recognized their epitopes in corresponding phosphorylated disaccharide compounds. Both groups of monosaccharide and monophosphoryl lipid A-recognizing antibodies have access to their epitopes in bisphosphoryl compounds as well. Because of this unidirectional reactivity with more complex structures, the various specificities cannot be distinguished by using bisphosphoryl lipid A (e.g., Escherichia coli lipid A) as a test antigen. The epitopes recognized by the various monoclonal antibodies all reside in the hydrophilic backbone of lipid A, and there was no indication that fatty acids were part of the epitopes recognized. Nevertheless, the reactivities of compounds in the different test systems are strongly influenced by their acylation patterns; i.e., acyl groups may modulate the exposure of lipid A epitopes.

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.

Structural requirements of lipid A for endotoxicity and other biological activities

For the past ten years, several groups were engaged in synthetic studies of lipid A, namely the lipid portion of bacterial lipopolysaccharides (LPS) that has been assumed to be the bioactive center of LPS, but has not been unanimously approved. Among them, Shiba, Kusumoto, and colleagues, Osaka, Japan have synthesized most energetically and successfully a variety of counterparts of lipid As, biosynthetic lipid A precursors, and their analogs. The endotoxic and related bioactivities of these synthetic compounds were studied by Japanese and German groups, including ours. In 1985, one of the compounds, having an acylation and phosphorylation pattern in beta(1-6)-D-glucosamine disaccharide which was proposed for Escherichia coli F515 lipid A was found to be exhibit full endotoxic and related bioactivities identical to those of the bacterial product. The study was extended by synthesis and examination of bioactivities of variously acylated D-glucosamine di- and monosaccharide phosphates, which correspond to structural components of lipid As, and their analogs or derivatives. Thus, structural requirements have been fairly well elucidated. In this article, first we will review the progress of synthetic and biological studies, with particular emphasis on chemical structure--bioactivities relationships of lipid As, and then we will discuss possible usefulness of some less or nontoxic lipid A-related synthetic compounds in clinical and preventive medicine.

Production of antibody to lipopolysaccharide (LPS) after immunization with a LPS-polymyxin B-agarose immunogen

A method was devised to produce antibodies to lipopolysaccharide (LPS) in guinea-pigs following a single immunization. The antigen was prepared by mixing polymyxin B-agarose with LPS from Escherichia coli O55:B5. Use of the agarose support allowed purification of the complex by simple washing procedures. Twenty-nine days after a single injection of the immunogen mixed with Freund complete adjuvant all animals demonstrated antibody to the LPS portion of the complex. No antibodies were detected to the polymyxin B component. Typical titres of LPS as measured by ELISA were 2(11). After, a booster immunization, titres of LPS antibody were further increased and a greater avidity was noted. In contrast to other methods which have been employed for production of antibody to LPS, use of the polymyxin B-agarose complex has the following advantages: ease of antigen preparation, ready purification of the complex, potent immunostimulation, and under the conditions employed here, LPS-specific antibody production, without accompanying antibody to polymyxin B.

Tumor necrosis factor-inducing activities of lipid A preparations from Pseudomonas diminuta and Pseudomonas vesicularis

Tumor necrosis factor (TNF)-inducing activities of lipid A preparations from P. diminuta and P. vesicularis, which contain mainly 2 mol of 2,3-diamino-2,3-dideoxy-D-glucose and 1 mol of nonglycosidic phosphate as the backbone component and have partly different fatty acid compositions, were examined. TNF was induced by injecting various lipid A fractions into mice that had previously been sensitized with Mycobacterium bovis BCG vaccine. A major component of lipid A of both strains, referred to as A3 fraction, exhibited stronger TNF-inducing activity than A2 fraction having incomplete acyl residues. The removal of ester-linked fatty acyl groups by mild hydrazinolysis of the P. diminuta lipid A results in a marked decrease of the activity. These results suggest that the structure of the hydrophobic part, including the amide-linked acyloxyacyl group(s), of the lipid A molecule play an important role in inducing TNF in the sera of mice.

Identification of Re lipopolysaccharide-binding protein on murine erythrocyte membrane

Our recent studies have suggested that bacterial lipopolysaccharide (LPS) attaches to Pronase-sensitive proteins on the murine erythrocyte membrane. In the present study, in order to identify the LPS-binding protein on the murine erythrocyte membrane, a unique method to detect LPS-binding protein on a nitrocellulose membrane was developed. Murine erythrocyte membrane proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, then transferred electrophoretically onto a nitrocellulose membrane. The membrane was incubated with LPS of Salmonella minnesota R595 (Re LPS) in phosphate-buffered saline (PBS), after the remaining sites were blocked with gelatin in PBS. We were able to obtain a non-background stain by adding the nonionic detergent octylglucoside at the low concentration of 0.1% to the Re LPS solution. The Re LPS bound to the protein on the nitrocellulose membrane was exposed to affinity purified anti-Re LPS antibodies (IgG) and then to alkaline phosphatase-conjugated anti-IgG. The alkaline phosphatase was detected on the membrane by an enzymatic reaction. This method demonstrated that Re LPS was bound to an erythrocyte protein of 96 kDa. Treatment of erythrocytes with Pronase led to disappearance of the Re LPS-binding protein on the erythrocyte membrane. There was no difference between LPS-responder and LPS-nonresponder murine erythrocyte membranes in amount and molecular weight of the Re LPS-binding protein.

The immunogenicity and antigenicity of lipid A are influenced by its physicochemical state and environment

We investigated the immunogenicity and antigenicity of synthetic lipid A and partial structures thereof. Included in the study were compounds which varied in the position of phosphate (1-mono-, 4'-mono-, and 1,4'-bisphosphates) and in the acylation (type, number, and distribution of fatty acids) and, in the case of monosaccharide compounds, the nature of the backbone sugar (D-glucosamine, D-glucose, 3-amino-3-deoxy-D-glucose, and 2,3-diamino-2,3-dideoxy-D-glucose). With the aid of the passive-hemolysis and passive-hemolysis-inhibition assays and by absorption experiments, five distinct antibody specificities were detected in polyclonal rabbit antisera raised against sheep erythrocyte-coated lipid A and lipid A incorporated into the membrane of liposomes (liposome-incorporated immunogens). Three antibody specificities reacted with disaccharide antigens specific for a 1-mono-, 4'-mono-, and 1,4'-bisphosphorylated beta-1,6-linked D-glucosamine disaccharide. Two antibodies reacted with either 1- or 4-phosphates of acylated D-gluco-configured monosaccharides and exhibited no cross-reaction with each other. However, they cross-reacted with disaccharide antigens with phosphate groups in the appropriate positions. We found that the physicochemical state and the environment of lipid A modulated its immunoreactivity. The immunogenicity was best expressed by erythrocyte-coated and liposome-incorporated immunogens. The antigenicity of lipid A was also greatly influenced by its physical surroundings. The reaction pattern of the above antibodies was highly specific in the hemolysis assay and in absorption experiments (the antibody reacted with antigen embedded in a cell membrane), whereas some cross-reactivities were observed in inhibition studies (the antibody reacts with antigen in aqueous solution). By using liposome-incorporated antigens as inhibitors, nonspecific reactions were avoided and specific ones were enhanced. Thus the antibodies described above against lipid A recognize epitopes in the hydrophilic backbone, the exposure of which depends on the intrinsic physicochemical properties of lipid A on the one hand and the physical environment on the other.

Pseudomonas diminuta LPS with a new endotoxic lipid A structure

Lipid A that contains mainly 2,3-diamino-2,3-dideoxy-D-glucose, phosphate and fatty acids in the molar ratio 2:1:5-6 was found in Pseudomonas diminuta lipopolysaccharide. The lipid A was considered to have a diamino-sugar disaccharide structure that carries a nonglycosidic phosphomonoester group and amide-bound acyloxyacyl and 3-hydroxy fatty acyl groups. The lipopolysaccharide exhibited endotoxic activities including lethal toxicity, pyrogenicity, local Shwartzman activity, body weight-decreasing toxicity and Limulus activity. The free lipid A was also endotoxic.

Low endotoxic activities of synthetic Salmonella-type lipid A with an additional acyloxyacyl group on the 2-amino group of beta (1-6) glucosamine disaccharide 1,4'-bisphosphate

A synthetic lipid A (Salmonella type, compound 516), beta (1-6)-linked D-glucosamine disaccharide 1,4'-bisphosphate, with three acyloxyacyl groups and one hydroxyacyl group, i.e., (R)-3-hexadecanoyloxytetradecanoyl, (R)-3-hydroxytetradecanoyl, (R)-3-dodecanoyloxytetradecanoyl, and (R)-3-tetradecanoyloxytetradecanoyl groups at the 2-amino, 3-hydroxyl, 2'-amino, and 3'-hydroxyl groups, respectively, was less biologically active than the synthetic Escherichia coli-type lipid A (compound 506), which has only two acyloxyacyl groups at the 2' and 3' positions and is substituted with a (R)-3-hydroxytetradecanoyl group at the 2-amino group. Compound 516 exhibited considerably weaker pyrogenic and leukopenic activity than compound 506, and it scarcely prepared rabbit skin for the Shwartzman reaction and lacked lethal toxicity on chicken embryos, although its lethal toxicity in galactosamine-loaded mice was as strong as that of compound 506. Compound 516 was also less active than compound 506 or natural E. coli lipid A (from Restrain F515) in other biological test systems, such as the Limulus test, stimulation of macrophages and lymphocytes, and interferon-inducing activity but not for interleukin-1 induction or complement activation. This observation suggests that there is an optimal number of acyloxyacyl groups on the glucosamine backbone for producing the biological activities of lipid A, especially the endotoxic activities. The 4'-monophosphate analog (compound 514) of compound 516 in general had significantly weaker activity than compound 516 in the above assays, most probably because of its greater hydrophobicity and consequently lower solubility in assay systems. Bacterial R595 lipid A derived from S. minnesota Re-mutant, which is a mixture of compounds 516 and 506, their 4'-monophosphate analogs and other compounds, exerted intermediate degrees of activity between compounds 506 and 516 in the various test systems employed.