Adhesion

Adhesion involves surface interactions between specific receptors on the mammalian cell membrane (carbohydrates, proteins or glycolipids) and ligands (usually proteins or glycoproteins) on the bacterial surface. The presence or absence of specific receptors on mammalian cells contributes significantly to tissue specificity of infection. Non-specific surface properties of the bacterium, including surface charge and hydrophobicity, also contribute to the initial stages of the adhesion process. Several different mechanisms of bacterial adherence have evolved, all utilizing specialized cell surface organelles or macromolecules that help to overcome the natural forces of repulsion that exist between the pathogen and its target cell.

Fimbrial adhesins

Electron microscopy of the surface of many Gram-negative and some Gram-positive bacteria reveals the presence of numerous thin, rigid, rod-like structures called fimbriae, or pili, that are easily distinguishable from the much thicker bacterial flagella (see p. 18). Fimbriae are involved in mediating attachment of some bacteria to mammalian cell surfaces. Different strains or species of bacteria may produce different types of fimbriae, which can be identified on the basis of antigenic composition, morphology and receptor specificity (Table 13.1). A broad division can be made between fimbriae, in which adherence in vitro is inhibited by D-mannose (mannose-sensitive fimbriae) and those unaffected by this treatment (mannose-resistant fimbriae).

Table 13.1 Examples of fimbriae produced by Gram-negative pathogens

The antigenic composition of fimbriae can be complex. For instance, two fimbrial antigens called colonization factor antigen (CFA) I and II have been detected in enteropathogenic E. coli strains. CFA-II consists of three distinct fimbrial antigens designated as coli surface (CS) antigens 1, 2 and 3. Another E. coli strain, E8775, has been found to produce three other CS antigens, CS4, CS5 and CS6. Pyelonephritogenic E. coli isolates produce a group of adhesins called X-adhesins; two fimbrial types designated S and M on the basis of receptor specificity have been identified in this group. The genes encoding a second type of fimbrae, the Chaperone Usher (CU) fimbriae, also found in E. coli, are found as multiple homologous, and often cryptic, variants, potentially with specificities for different carbohydrate receptors.

The evolutionary significance of such heterogeneity may be that the ability of an individual bacterium to express several different types of fimbriae allows different target receptors to be used at different anatomical sites of the infected host. In vitro, production of fimbriae is influenced by cultural conditions such as incubation temperature and medium composition, which may switch off the production of fimbriae or induce a phase change from one fimbrial type to another.

For some fimbriae the association with infection is clear. Thus the K88 fimbrial antigen is clearly associated with the ability of E. coli K88 to cause diarrhoea in pigs; pigs lacking the appropriate intestinal receptors are spared the enterotoxigenic effects of E. coli strains of this type. In many other instances the association between production of fimbriae and infection remains putative at present. Production of fimbriae is controlled by either chromosomal or plasmid genes.

The structure of one of these fimbrial types – type 1 or common fimbriae – has been studied in detail. These consist of aggregates of a structural protein subunit called fimbrillin (or pilin) arranged in a regular helical array to produce a rigid rod-like structure 7 nm in diameter, with a central hole running along its length.

A highly conserved minor protein, located at both the tip and at intervals along the length of the fimbriae, mediates specific adhesion. Type 1 fimbriae bind specifically to D-mannose residues. Their role in vivo remains controversial; however, they may be involved in the pathogenesis of urinary tract infections.

Other Gram-negative bacteria, including those of the genera Pseudomonas, Neisseria, Bacteroides and Vibrio, produce fimbriae that share some homology, especially in the amino-terminal region of the fimbrillin subunits (the so-called N-methylphenylalanine fimbriae). These fimbriae have been shown to act as virulence determinants for Ps. aeruginosa and N. gonorrhoeae.

Non-fimbrial adhesins

Non-fimbrial adhesins include protein or polysaccharide structures that are surface exposed on the bacterial cell and/or secreted. Protein-based adhesins include the filamentous haemagglutinin of B. pertussis, a mannose-resistant haemagglutinin from Salmonella enterica serotype Typhimurium and a fibrillar haemagglutinin from Helicobacter pylori. Outer membrane proteins are involved in the adherence of many, if not most, pathogenic bacteria.

Exopolysaccharides present on the surface of some Gram-positive bacteria are also involved in adhesion. For example, Streptococcus mutans, which is involved in the pathogenesis of dental caries, synthesizes a homopolymer of glucose that anchors the bacterium to the tooth surface and contributes to the matrix of dental plaque. Actinomyces may adhere to other oral bacteria – a process called co-aggregation. Teichoic acid and surface proteins of coagulase-negative staphylococci mediate adherence of the bacterium to prosthetic devices and catheters, contributing to increasing numbers of hospital-acquired infections. Continued growth following attachment to these biomaterials may result in formation of a biofilm (see Ch. 4), which may hinder successful antibiotic treatment by restricting access of drugs to the bacterium.

Flagella act as adhesins in Vibrio cholerae and Campylobacter jejuni. Bacterial motility is also thought to be important in chemotaxis of these organisms and H. pylori towards intestinal cells, and in penetration of these bacteria through the mucous layer during colonization.

Binding to connective tissue proteins

Binding of pathogenic bacteria to a number of connective tissue proteins, including laminin, vitronectin and collagen, have been described. The best-studied of these host proteins is fibronectin, a complex multifunctional glycoprotein found in plasma and associated with mucosal cell surfaces, where it promotes numerous adhesion functions. Many pathogenic bacteria bind fibronectin at the bacterial surface, and for some organisms fibronectin has been shown to act as the cell surface receptor for bacterial adhesion. In Streptococcus pyogenes, lipoteichoic acid mediates attachment of the bacterium to the amino terminus of the fibronectin molecule. Attachment of S. aureus to cell surfaces also involves the amino terminus of fibronectin, but the bacterial ligand appears to be protein in this instance. T. pallidum also binds fibronectin. The significance of the interaction with fibronectin in the pathogenesis of syphilis and many other bacterial diseases needs further clarification.

Consequences of adhesion

In addition to preventing loss of the pathogen from the host, adhesion induces structural and functional changes in mucosal cells, and these may contribute to disease. For example, adherence of enteropathogenic E. coli (EPEC) to epithelial cells induces rearrangements of the cell cytoskeleton causing loss of microvilli and localized accumulation of actin, without subsequent invasion of the host cell. In contrast, adherence of H. pylori to gastric epithelial cells causes enhanced production of the pro-inflammatory chemokine interleukin (IL)-8, which contributes to gastric pathology. In both cases, these changes involve the induction of intracellular signalling pathways triggered after binding of the bacteria to specific receptors on the epithelial cell surface. Adhesion of bacteria to mammalian cells may also induce changes in bacterial gene expression.

Uptake into host cells

The initial phase of cellular invasion involves penetration of the mammalian cell membrane; rather than eliciting mechanisms for host cell penetration themselves, many intracellular pathogens induce phagocytic processes in the host cell to gain access.

Shigellae invade colonic mucosal cells but rarely penetrate deeper into the host tissues. Inside the cell, they are surrounded by a membrane-bound vesicle derived from the host cell. Soon after entry, this vesicle is lysed by the action of a plasmid-encoded haemolysin (haemolysis is just one way of detecting a membrane-damaging toxin), and the bacteria are released into the cell cytoplasm. Listeria monocytogenes produces a heat shock protein with a similar function, termed listeriolysin. Once free in the cell cytoplasm, shigellae multiply rapidly with subsequent inhibition of host cell protein synthesis. Several hours later the host cell dies, and bacteria spread to adjacent cells where the process of invasion is repeated.

Other bacteria remain within membrane-bound vesicles but modify these cellular compartments, preventing them from maturing and fusing with lysosomes. The pathogenic Neisseria, for example, elaborate a secreted toxin, immunoglobulin A protease, which in addition to cleaving immunoglobulin A during colonization of the mucosa (see below), is also capable of cleaving a key endocytic vesicle protein, LAMP1, which is responsible for acidification of the endosome. In this way the intracellular pathogen modifies the environment within the vesicle to favour its own survival. Most salmonellae proceed through the superficial layers of the gut and invade deeper tissues, in particular cells of the reticulo-endothelial system such as macrophages. Salmonellae also occupy a host-derived vesicle that does not lyse. In this case several vesicles coalesce to form large intracellular vacuoles. These vacuoles traverse the cytoplasm to reach the opposite side of the cell and initiate spread to adjacent cells and deeper tissues. Although invasion of host cells is essentially a bacterium-directed but a host-mediated process, the active participation of the bacterium and novel bacterial protein synthesis is needed.

Role of cell receptors

The availability of specific receptors defines the types of host cell that are involved. As a result, some pathogens can invade a wide range of cell types, whereas others have a more restricted invasive potential. Specific host receptors for some of the invasive pathogens have been identified. For example, Legionella pneumophila and Mycobacterium tuberculosis adhere to complement receptors on the surface of phagocytic cells. A specific receptor for Yersinia pseudotuberculosis belongs to a family of proteins termed integrins that form a network on the surface of host cells to which host proteins such as fibronectin can bind. Mimicry of the amino acid sequence (Arg-Gly-Asp) of fibronectin that mediates attachment to the integrins may represent a common mechanism of effecting intracellular entry. The ability to utilize integrins may not be restricted to intracellular bacteria. The filamentous haemagglutinin of B. pertussis may use the fibronectin integrin to mediate attachment in the respiratory tract.

A number of bacterial pathogens capable of crossing the blood–brain barrier, including N. meningitidis, Haemophilus influenza and Streptococcus pneumonia, as well as a number of neurotropic viruses and prions, have been shown to bind to the laminin receptor protein, suggesting that binding to this host receptor protein confers a neurotropism on these pathogens.

Capsules

Many bacterial pathogens need to avoid phagocytosis; production of an extracellular capsule is the most common mechanism by which this is achieved. Virtually all the pathogens associated with meningitis and pneumonia, including H. influenzae, N. meningitidis, E. coli and S. pneumoniae, have capsules, and non-capsulate variants usually exhibit much reduced virulence. Most capsules are polysaccharides composed of sugar monomers that vary among different bacteria. Polysaccharide capsules reduce the efficiency of host defences in a number of ways:

Streptococcal M protein

The M protein present on the surface of S. pyogenes is not a capsule but functions in a similar manner to prevent complement deposition at the bacterial surface. The M protein binds both fibrinogen and fibrin, and deposition of this material on the streptococcal surface hinders the access of complement activated by the alternative pathway.

Meningococcal factor H-binding protein

The complement system is a powerful weapon against bacterial pathogens, but can potentially damage host cells as well. The host must, therefore, protect itself against the damaging activities of complement. The serum glycoprotein factor H is a negative regulator of the complement system that protects host cells by binding to glycosamino glycans present on the surface of host but not to bacterial cells. It can downregulate the activity of complement and thus protect the host cell. N. meningitidis expresses a factor H-binding protein on its surface, which recruits this complement regulator, thus affording similar protection to the bacterial pathogen.

Resistance to killing by phagocytic cells

Some pathogens not only survive within macrophages and other phagocytes, but may actually multiply intracellularly. The normal sequence of events following phagocytosis involves fusion of the phagosome in which the bacterium is contained with lysosomal granules present in the cell cytoplasm. These granules contain enzymes and cationic peptides involved in oxygen-dependent and oxygen-independent bacterial killing mechanisms (see p. 117).

Different organisms use different strategies for survival (Table 13.2). M. tuberculosis is thought to resist intracellular killing by preventing phagosome–lysosome fusion; other bacteria are able to resist the action of such lysosomal components after fusion. Some organisms stimulate a normal respiratory burst but are intrinsically resistant to the effects of the potentially toxic oxygen radicals produced. Production of catalase by S. aureus and N. gonorrhoeae is thought to protect these organisms from such toxic products. The smooth lipopolysaccharide of many bacterial pathogens is also thought to contribute to their resistance to the effects of bactericidal cationic peptides present in the phagolysosome.

Table 13.2 Some strategies adopted by bacteria to avoid intracellular killing

Antigenic variation

Variation in surface antigen composition during the course of infection provides a mechanism of avoidance of specific immune responses directed at those antigens. This strategy is most highly developed in blood-borne parasitic protozoa, such as trypanosomes, but is also exhibited by bacteria. Pathogenic Neisseria, for example, are capable of changing surface antigens using three highly efficient mechanisms. These are mutation of individual amino acids, phase variation (switching genes on and off) and horizontal exchange of DNA material. N. meningitidis can avoid the killing effect of antibodies against its major porin (PorA) by mutating amino acids and/or acquiring parts of or all of its porA gene from another meningococcal strain. The organism can switch off the expression of its capsule or its immunogenic proteins by shifts in the nucleotide sequence encoding them. The latter varies as a result of recombination or mutation during DNA replication.

Another mechanism of antigen variation in Neisseria is the genetic rearrangements demonstrated in the fimbriae. Usually only one complete fimbrillin gene is expressed, although there may be several incomplete ‘silent’ gene sequences present on the chromosome. Movement of parts of the incomplete gene sequences, either from within the genome of the expressing strain or from DNA released by a co-infecting strain to an expression locus, results in synthesis of a protein that may differ antigenically from the original.

The borreliae that cause relapsing fever use a similar strategy to generate antigenic variation in their outer surface proteins. Other bacteria show strain-specific antigenic variability, for example group A streptococci produce up to 75 antigenically distinct serotypes of M protein.

The capacity for variation in surface antigens allows for longer survival of an individual organism in a host and means that antibody produced in response to infection by one strain of a pathogen may not protect against subsequent challenge with a different strain of that bacterium. This makes the variable antigens elusive targets for protective antibodies, and the development of vaccines based on inhibition of attachment or generation of opsonic or bactericidal antibodies is particularly difficult for these organisms.

Immunoglobulin A proteases

Several species of pathogenic bacteria that cause disease on mucosal surfaces produce a protease that specifically cleaves immunoglobulin A (IgA), the principal antibody type produced at these sites. These proteases are specific for human IgA isotype I. Nearly all of the pathogens causing meningitis possess an IgA protease and a polysaccharide capsule enabling them to persist on the mucosal surface and resist phagocytosis during the invasive phase of the disease.

Serum resistance

To survive in the bloodstream, bacteria must be able to resist lysis as a result of deposition of complement on the bacterial surface. In the Enterobacteriaceae, resistance is primarily due to the composition of the lipopolysaccharide present in the bacterial outer membrane. Smooth colonial variants that possess polysaccharide ‘O’ side chains are more resistant than rough colonial variants that lack such side chains (see below). The side chains sterically hinder deposition of complement components on the bacterial surface. Conversely, however, some O-chain polysaccharides activate complement by an alternative pathway leading to lysis of the bacterial cell. In N. meningitidis group B and E. coli K1, sialic acid capsules prevent efficient complement activation and, in N. gonorrhoeae, complement binds but forms an aberrant configuration in the bacterial outer membrane so that it is unable to effect lysis.

Iron acquisition

The concentration of free iron in body fluids and secretions is below that required for bacterial growth because it is bound (chelated) by high-affinity mammalian iron-binding proteins such as transferrin and lactoferrin. To multiply in body fluids or on mucous membranes bacterial pathogens have evolved efficient mechanisms for scavenging iron from mammalian iron-binding proteins. Bacteria such as E. coli, Klebsiella pneumoniae and some staphylococci produce extracellular iron chelators called siderophores for this purpose. Others, including Campylobacter jejuni and N. meningitides, do not produce siderophores themselves but are able to obtain iron from siderophores produced by other species or use host molecules such as noradrenalin, which has siderophore-like activity, as a source of iron. In an alternative strategy pathogens including N. meningitidis, Haemophilus parainfluenzae, H. influenzae type b, Staphylococcus epidermidis and S. aureus have specific receptors for transferrin and/or lactoferrin, on their surfaces, and are able to bind these proteins and remove the bound iron directly from these host proteins. Production of siderophores, their cell surface receptors, and receptors for transferrin, lactoferrin and other mammalian iron-binding proteins is iron-regulated and occurs mainly under conditions of iron restriction.

Two other mechanisms of iron acquisition from mammalian iron chelators have been described. Some Bacteroides species remove iron by proteolytic cleavage of the chelator. In L. monocytogenes, reduction of the Fe³⁺ ion to Fe²⁺ reduces the affinity for the chelator sufficiently for it to be removed by the bacterium.

Many bacteria express receptors for binding and/or internalizing other mammalian iron-containing molecules, such as haem, haemoglobin and haemoglobin–haemopexin complexes. These mammalian molecules are located intracellularly, where they may be available to intracellular pathogens; they may also be released by bacterial haemolysins that lyse the red cells.

Endotoxin

Endotoxin, also called lipopolysaccharide or lipo-oligosaccharide, is a component of the outer membrane of Gram-negative bacteria and is released from the bacterial surface via outer membrane vesicles (blebs), which may be released from the bacterial cell surface, or by lysis and disintegration of the organism. Lipopolysaccharide is anchored into the bacterial outer membrane through a unique molecule termed lipid A (Fig. 13.1). Covalently linked to lipid A is an eight-carbon sugar, ketodeoxyoctonate, in turn linked to the chain of sugar molecules (saccharides) that form the highly variable O antigen structures of Gram-negative bacteria. On solid bacteriological media, bacteria carrying lipopolysaccharide with O antigen form smooth colonies with hydrophilic surfaces; in contrast those carrying lipopolysaccharide without the O antigen form rough colonies with hydrophobic surfaces.

Fig. 13.1 Diagrammatic representation of the structure of bacterial lipopolysaccharide.

After Rietschel ET, Galanos C, Lüderlitz O 1975 Structure, endotoxicity and immunogenicity of the lipid A component of bacterial lipopolysaccharide. In: Schlessinger D (ed.) Microbiology, pp. 307–314. American Society for Microbiology, Washington, DC.

The term endotoxin was originally introduced to describe the component of Gram-negative bacteria responsible for the pathophysiology of endotoxic shock, a syndrome with a high mortality rate, particularly in immunocompromised or otherwise debilitated individuals. Endotoxin activates complement via the alternative pathway, but most of the biological activity of the molecule is attributable to lipid A. Both endotoxin and lipid A are potent activators of macrophages, resulting in the induction of a range of cytokines involved in the regulation of immune and inflammatory responses (see Ch. 9). Although endotoxin from Gram-negative organisms remains central to our understanding of septic shock, other structural and secreted components of bacteria that interact with pattern recognition receptors (see Ch. 9) can contribute to the pathogenesis of this clinical syndrome. Thus the multi-system effects, often involving complement, blood clotting factor and kinin activation together with extensive cytokine release, should not be seen as exclusive to Gram-negative infection.

Exotoxins

Exotoxins, in contrast to endotoxin, are diffusible proteins secreted into the external medium by the pathogen. Most pathogens secrete various protein molecules that facilitate adhesion to, or invasion of, the host. Many others cause damage to host cells. The damage may be physiological, for example cholera toxin promotes electrolyte (and fluid) excretion from enterocytes without killing the cells, or pathological, where the toxin (e.g. diphtheria toxin) inhibits protein synthesis and induces cell death. Exotoxins vary in their molecular structure, biological function, mechanism of secretion and immunological properties. The list of bacterial exotoxins is vast and increasing; however, they are often classified by their mode of action on animal cells:

Examples of exotoxins and their effects on target cells are shown in Table 13.3. Bacteria secrete toxins and other proteins using a number of distinct mechanisms that are understood to varying extents. Some of these systems are listed in Gram-negative bacteria as types I–VI, although this number does not reflect the total number of pathways that are known to mediate secretion of proteins across the bacterial envelope. Figure 13.2 shows the basic components of the types I and V pathways which are relatively well characterized. In type I, at least three proteins get together to form a channel through which large molecules (such as haemolysin of E. coli) are exported. In type V, however, a single precursor protein that consists of three domains is secreted sequentially across the inner and outer membranes. Although these latter proteins are called autotransporters, it is now known that both secretion steps require the activity of a number of additional cellular proteins. A typical example of an autotransporter is the IgA1 protease of Neisseria spp.

Table 13.3 Some effects of bacterial exotoxins

^a Separation of epidermis from dermis.

Fig. 13.2 Diagrammatic representation of type I and type V (autotransporter) secretion of exotoxins across the bacterial cell membrane. ADP, adenosine diphosphate; ATP, adenosine triphosphate; Hly, haemolysin; Tol, special receptor.

Enterotoxins cause symptoms of gastrointestinal disease, including diarrhoea, dysentery and vomiting. In some cases the disease is caused by ingestion of preformed toxin in food, but in most cases colonization of the intestine is required before toxin is made.

Cholera toxin and heat-labile toxins of enterotoxigenic E. coli (ETEC) do not induce inflammatory changes in the intestinal mucosa, but perturb the processes that regulate ion and water exchange across the intestinal epithelium (see Chs 26 and 30). In contrast, the enterotoxins of Clostridium difficile, C. perfringens type A and Bacillus cereus cause structural damage to epithelial cells, resulting in inflammation. Another gastrointestinal pathogen, enteropathogenic E. coli (EPEC), mediates damage by a process that involves secretion of a protein, Tir, directly across both bacterial membranes and the host cell membrane in a single step. Tir is subsequently inserted into the host cell membrane where it acts as a receptor for a bacterial adhesin, intimin. Binding of intimin to Tir results in phosphorylation of the latter protein; this in turn leads to a signalling cascade and cytoskeletal rearrangements within the host cell.

B. pertussis, the causative agent of whooping cough, produces various extracellular products, including a tracheal cytotoxin that inhibits the beating of cilia on tracheal epithelial cells, pertussis toxin, which exhibits several systemic effects, and an adenylate cyclase that interferes with phagocyte function.

Another group of toxins causes damage to subepithelial tissues following penetration and multiplication of the pathogen at the site of infection. Many of these toxins also inhibit or interfere with components of the host immune system. Membrane-damaging toxins such as staphylococcal α and β toxins, streptolysin O and streptolysin S, and C. perfringens α and θ toxins inhibit leucocyte chemotaxis at subcytolytic concentrations, but cause necrosis and tissue damage at higher concentrations.

Systemic effects of toxins

Some toxins cause damage to internal organs following absorption from the focus of infection. Included in this category are the toxins causing diphtheria, tetanus and botulism, and those associated with streptococcal scarlet fever, staphylococcal toxic shock syndrome and haemolytic uremic syndrome associated with Shiga toxin-producing enteropathogens. The diphtheria toxin, the gene for which is bacteriophage-encoded, inhibits protein synthesis in mammalian cells. Tetanus toxin, in contrast, exerts its effect by preventing the release of inhibitory neurotransmitters whose function is to prevent overstimulation of motor neurones in the central nervous system, resulting in the convulsive muscle spasm characteristic of tetanus. Diphtheria and tetanus toxins represent the sole determinant of disease and are neutralized by specific antitoxin antibody. As a result, vaccination with toxoids (formalin-inactivated toxins) derived from these toxins is highly effective (see Ch. 70).

Botulism results from the ingestion of preformed toxin produced by Clostridium botulinum in food contaminated with this bacterium, and is not a true infectious disease. The toxic activity is due to a family of serologically distinct polypeptide neurotoxins that prevent release of acetylcholine at neuromuscular junctions, resulting in the symptoms of flaccid paralysis. These toxins have been used clinically in treating squints and muscle spasm.

Other toxins cause disseminated multi-system organ damage. Such pathology is seen in staphylococcal toxic shock syndrome caused by certain strains of S. aureus that produce a toxin designated toxic shock syndrome toxin 1 (TSST-1). This toxin belongs to a group of functionally related proteins collectively referred to as superantigens, which include the staphylococcal enterotoxins, staphylococcal exfoliative toxin and streptococcal pyrogenic exotoxin A. These molecules are potent T cell mitogens whose reactivity with lymphocytes induces cytokine release; they may initiate tissue damage by mechanisms similar to those postulated to account for Gram-negative endotoxic shock (see p. 165).