Structure of bacterial lipopolysaccharides

Abstract

Bacterial lipopolysaccharides are the major components of the outer surface of Gram-negative bacteria They are often of interest in medicine for their immunomodulatory properties. In small amounts they can be beneficial, but in larger amounts they may cause endotoxic shock. Although they share a common architecture, their structural details exert a strong influence on their activity. These molecules comprise: a lipid moiety, called lipid A, which is considered to be the endotoxic component, a glycosidic part consisting of a core of approximately 10 monosaccharides and, in “smooth-type” lipopolysaccharides, a third region, named O-chain, consisting of repetitive subunits of one to eight monosaccharides responsible for much of the immunospecificity of the bacterial cell.

Introduction

The history of lipopolysaccharides, recently documented by Rietschel and Westphal,¹ began in the eighteenth century with the search for the fever- and disease-producing substance that was associated with unhygienic conditions. This was variously referred to as the pyrogenic material, putrid poison, or toxin. By 1872, the growing awareness of a possible role for living organisms in human diseases allowed Klebs, a German bacteriologist, to attribute a majority of military war deaths to a pyrogenic substance from microorganisms which he called “Microsporon septicum”. Two years later, a Danish pathologist named Panum reported a non-volatile, heat-resistant, water-soluble, pyrogenic toxin obtained from putrid matter. Later, with the pure-culture techniques developed by Koch,² it was possible to show that different diseases were caused by specific bacteria. From the same laboratory in 1892, Pfeiffer reported that the agent of cholera, Vibrio cholerae, produced a pyrogenic, non-secreted toxin that was heat-stabile, in addition to a secreted, heat-labile toxin. He called it endotoxin, a term still used today for the lipopolysaccharides that were later found to constitute them. Nevertheless, we know at present that endotoxins are not all toxic, just as bacteria are not all pathogenic. The endotoxins were soon shown to characterize the major group of Gram-negative bacteria, i.e., those having a second, outer membrane.

The development of techniques for extracting and preparing endotoxins sufficiently pure for structural studies was slow. In the 1930s and 1940s, fairly pure preparations were reported to consist of a polysaccharide, a lipid part, and a small amount of protein, and received the name lipopolysaccharide (LPS).¹ The 1950s and 1960s saw the introduction of the extraction methods most commonly used today.3., 4.

The enterobacterial LPSs (especially those of Escherichia coli and Salmonella enterica serovar typhimurium) were most thoroughly studied. Strains of these species gave colonies with either a rough or a smooth appearance. By 1964, the former were found to produce LPS containing glucosamine, glucose, galactose, the recently characterized l-glycero-d-manno-heptose and 3-deoxy-oct-2-ulosonic-acid (Kdo), as well as phosphate and the C₁₂ and C₁₄ lauric, myristic, and hydroxymyristic (C₁₄OH) acids. Strains giving rise to colonies with a smooth aspect also produced LPSs with these components, but with many other sugars as well.

The LPSs resisted structural characterization by their amphipathic nature. Their strong tendency to form aggregates made it difficult to determine their molecular weight. In addition to hydrophobic bonds, there were problems of intermolecular crosslinking of acid groups (phosphates, pyrophosphates, and acid sugars) via divalent cations. Chelation and electrodialysis removed some of the latter as well as monovalent (Na+, K+) ions.5., 6. Nevertheless, such techniques, with or without the replacement of divalent cations by the more soluble ammonium or triethylammonium salts, were useful in lowering the size of aggregates.⁷ Molecular-weight estimation by SDS-polyacrylamide gel electrophoresis, which gave good resolution, suffered from the lack of appropriate molecular-weight standards. In any case, most LPS preparations were (and still are) heterogeneous. Precise molecular masses of native LPS molecules were obtained only after progress was made in the preparation of LPS for mass spectrometry. Nevertheless, the natural heterogeneity of most LPS preparations still complicates the interpretation of their spectra.

The finding that the lipid region could be separated after weak acid hydrolysis helped establish the general architecture comprising two or three regions:⁸ a lipid region, a core, and the third region in bacteria yielding smooth colonies. The use of rough mutants having cores of different lengths facilitated the determination of the Salmonella core structure.9., 10. In the early 1980s, the structures of both the lipid A and the core were established. The regions going from the bacterial membrane toward the outside are, in order: a lipid (called lipid A) linked to Kdo of the core oligosaccharide, itself linked to the second glycoside if any, consisting of a sequence of repetitive subunits and called O-chain or O-specific antigen (Fig. 1). The endotoxins were later recognized as the matrix of the external leaflet of the bacterial outer membrane. It forms the major component (45%) occupying 75% of the surface of the bacterium. E. coli is estimated to have 10⁶ molecules per cell.¹¹ The polysaccharide moiety is directed outwards from the bacterium surface, extending up to 10 nm from the surface (Fig. 1).

The biosyntheses of the lipid A and core are finely orchestrated and the combination put into place followed by the O-chain. The O-chains extending outward from the bacterial cell surface are exposed to the environment and the defence system of a potential host.

It is not surprising that most LPSs selected for analysis have been those of medical or veterinary interest. In many cases, bacteria that were not considered to be human pathogens were found in infected, immuno-compromised patients. In other cases, bacteria that were pathogenic for other mammals, became pathogens for humans, immuno-compromised or not, and vice versa. At present, considerable attention is being given to the LPSs of Gram-negative, nitrogen-fixing bacteria. These are not pathogens of course, even though one refers to the interaction between plant and bacterium that results in their symbiosis as an infection.

Although not secreted by the cells, small amounts of the LPS are liberated into the medium under some circumstances such as cell division. Larger amounts are released by bacteria killed by antibiotics, phagocytosis, the complement complex, or treatment with divalent cation chelators. In an infected host, small amounts of LPS can be protective by stimulating the immune system, they have e.g., been used to shrink tumours. Large amounts, however, induce high fever, increase heart rate, and lead to septic shock and death by lung and kidney failure, intravascular coagulation, and systemic inflammatory response.

Most of the biological activities have been associated with the lipid moiety of the molecule. However, the role of the polysaccharide moiety is not negligible. This has been illustrated by the stronger biological activities induced by Re-type LPS consisting of lipid A carrying only two Kdo residues compared to those of isolated or synthetic lipid A.12., 13. The conformation of at least part of the lipid A is modified by these sugars and their charges.¹⁴ A highly purified lipid A was unable to induce secretion of interleukin-1 by human monocytes,¹⁵ and a charged, attached Kdo unit was shown to be necessary for this activity.16., 17. The importance of the polysaccharide moiety also lies in its antigenic properties, its effect on the solubility of the LPS molecule, and its charge. Other activities which require an oligosaccharide include mitogenicity (the stimulation of B cell division) and the activation of human macrophage cell lines.17., 18.