Adhesin–receptor interactions in Pasteurellaceae1

Authors

  • Mario Jacques,

    Corresponding author
    1. Groupe de Recherche sur les Maladies Infectieuses du Porc and Département de Pathologie et Microbiologie, Faculté de Médecine Vétérinaire, Université de Montréal, C.P. 5000, St-Hyacinthe, Québec J2S 7C6, Canada
    Search for more papers by this author
  • Sonia-Élaine Paradis

    1. Groupe de Recherche sur les Maladies Infectieuses du Porc and Département de Pathologie et Microbiologie, Faculté de Médecine Vétérinaire, Université de Montréal, C.P. 5000, St-Hyacinthe, Québec J2S 7C6, Canada
    Search for more papers by this author

  • 1

    This paper was presented at the Haemophilus, Actinobacillus, Pasteurella International Conference (HAP'96) held in Acapulco, Mexico, October 13–17, 1996.

*Corresponding author. Tel.: +1 (514) 773-8521, ext. 8348; Fax: +1 (514) 778-8108; E-mail: jacqum@ere.umontreal.ca

Abstract

The ability of bacteria to adhere to mucosal epithelium is dependent on the expression of adhesive molecules or structures, called adhesins, that allow attachment of the organisms to complementary molecules on mucosal surfaces, the receptors. Important human and animal pathogens are found among the Pasteurellaceae family which includes Haemophilus, Actinobacillus, and Pasteurella organisms. The purpose of this paper is to review the adhesin-receptor systems found in Pasteurellaceae, with an emphasis on recent developments in this specific area. Most of these organisms can employ multiple molecular mechanisms of adherence (or multiple adhesins) to initiate infection. Indeed, a wide variety of adhesins are expressed by members of the Pasteurellaceae, and different proteins (e.g. fimbriae, fibrils, outer membrane proteins) as well as polysaccharides (lipooligosaccharides, lipopolysaccharides, capsular polysaccharides) were clearly shown to play an important role in adherence. In many instances, these adhesins have proved to represent good vaccine candidates. Surprisingly, the receptors on host mucosal surfaces have yet been identified in very few cases.

1Introduction

The initial event in bacterial colonization of their host organisms is the adherence of microorganisms to the epithelial cells and/or mucus layer of the mucosal surfaces, which involves specific interactions between bacterial adhesins and host receptors [1]. Bacterial adherence endows the pathogen with the ability to withstand normal host defense cleansing mechanisms on mucosal surfaces. Adherence also confers a number of advantages on the bacterium, including enhanced toxicity to the host and increased resistance to deleterious agents. The fact that bacterial colonization and infection can be prevented by blocking adherence has stimulated research on the identification of adhesins, their molecular biology, and their binding specificities.

Members of the Pasteurellaceae are small Gram-negative rods that can colonize the mucosal surface of the respiratory and genital tracts. Important human and animal pathogens are found among this bacterial family (Table 1) [2]. The purpose of this paper is to review the adhesin-receptor systems found in Pasteurellaceae, i.e. organisms of the Haemophilus, Actinobacillus, and Pasteurella (HAP) group, with an emphasis on recent developments in this specific area. The literature review for this article ended in May 1997.

Table 1.  Members of the Pasteurellaceae family described in the present review and the diseases they cause in their respective host
OrganismHostDisease
Haemophilus  
H. ducreyihumanchancroid
H. influenzae capsular type b (Hib)humanmeningitis, septicemia, epiglottitis
H. influenzae non-typable (NTHi)humanotitis media, sinusitis, conjunctivitis, acute lower respiratory tract infection
Actinobacillus  
A. actinomycetemcomitanshumanjuvenile and adult periodontitis
A. pleuropneumoniaeswinepleuropneumonia
Pasteurella  
P. haemolyticabovine and sheeppneumonia
P. multocidaswinepneumonia, atrophic rhinitis

2Haemophilus

2.1Haemophilus ducreyi

H. ducreyi is the etiological agent of chancroid, a sexually transmitted disease that is common in developing countries and that has characteristic genital mucocutaneous ulcerative lesions (for a recent review see [3]). The disease has received renewed attention after reports that genital ulcers facilitate the transmission of the human immunodeficiency virus in endemic populations. It has been proposed [3, 4]that bacterial adherence to epithelial cells is the first step in the pathogenesis of H. ducreyi infection. Adherence is followed by the growth of bacteria on the epithelial cells and the secretion of cytotoxin which result in cell damage that may be responsible for the development of ulcers.

H. ducreyi adheres strongly to many cultured cell lines including HEp-2 (human laryngeal epidermoid carcinoma), CHO (Chinese hamster ovary), MRC-C (human embryo lung), C16 (clone of MRC-C), HeLa (human cervical carcinoma), HEC-1-B (endometrial adenocarcinoma), and A549 (human lung carcinoma), to human foreskin cells, foreskin fibroblasts, and keratinocytes [4–10]. Cell invasion has been observed with some of these cell lines [3]. In vitro, all the strains tested could survive and multiply on cell monolayers, and a small fraction of bacteria could also survive inside eukaryotic cells. However, because many cells, including epithelial cells, can phagocytose foreign particles, cell invasion may reflect the phagocytic activity of the epithelial cells rather than the invasive properties of the bacteria. H. ducreyi adheres also to extracellular matrix proteins such as fibrinogen, fibronectin, collagen, gelatin and laminin [11].

This microorganism expresses surface structures that resemble pili or fimbriae [12]. Although their role in adherence to host cells has not been demonstrated they seemed to be involved in binding to laminin [11]. These fine tangled pili are composed predominantly of a protein whose apparent molecular mass is 24 kDa [12, 13]. The gene (ftpA) coding for the major subunit has been recently cloned [13]. These pili represent a novel class of pili since the sequence lacks a cleavable signal sequence and has no homology to known pilin sequences. The major subunit FtpA does not mediate binding to laminin since an isogenic pilin mutant binds to laminin [13]. Interestingly, vaccination with a H. ducreyi pilus preparation confers protection in an animal model of chancroid, through cell-mediated immunity [14].

Lipooligosaccharides (LOS), present in the outer membrane, seem to play a role in adherence of H. ducreyi to human keratinocytes which may represent the first cells this microorganism encounters in the host [15, 16]. A Tn916 mutant of strain 3500 (mutant 1381) has LOS which lack the Gal-β-1-4-GlcNAc structure, and is deficient in adherence and invasion of human keratinocytes. The mutation is located in a rfaK-like gene (d-glycero-d-manno-heptosyltransferase). Alfa and DeGagne [17]recently showed that purified LOS was able to inhibit attachment of H. ducreyi to foreskin fibroblasts in a dose-dependent manner. In addition, proteinase K treatment of H. ducreyi significantly reduced attachment suggesting protein involvement. It appeared that H. ducreyi binds to fibronectin in the extracellular matrix of human foreskin fibroblasts since fibronectin was able to significantly reduce attachment. These authors hypothesized that the attachment of H. ducreyi involves both a protein mediator (likely pili) as well as LOS and that one or both of these bacterial surface components interacts with fibronectin to mediate attachment to human foreskin fibroblasts.

2.2Haemophilus influenzae

H. influenzae is an important human pathogen worldwide. Non-encapsulated isolates, known as non-typable H. influenzae (NTHi), are a common cause of otitis media, sinusitis, conjunctivitis and acute lower respiratory tract infections. Capsular type b isolates (Hib) cause invasive, bacteremic infections such as meningitis, septicemia, and epiglottitis, particularly in infants. H. influenzae is a typical example of microorganisms expressing multiple adhesins [18, 19].

2.2.1Capsule

Hib isolates adhere markedly less than isogenic non-encapsulated variants to cultured epithelial cells and nasal turbinates [19]suggesting that the capsule of Hib is not involved in adherence but rather masks the adhesin(s). Antibodies specific for Hib capsular polysaccharide however, were shown to block adherence of Hib to epithelial cells [20]. Swelling of the capsular material caused by these antibodies may have resulted in steric hindrance and interfered with fimbriae-mediated adherence.

2.2.2Fimbriae (pili)

Fimbriae, on the other hand, seem to play an important role and fimbriated Hib isolates adhere better to organ cultures than do non-fimbriated Hib [21–23]. Hib isolates from the nasopharynx, in contrast to blood and cerebrospinal fluid isolates, usually express fimbriae. It has been therefore postulated that expression of fimbriae is beneficial during the initial stages of colonization and infection, but disadvantageous in establishing systemic disease.

It has been shown that H. influenzae fimbriae mediate adherence to oropharyngeal epithelial cells and erythrocytes, and that the adhesive properties reside in the major fimbrial subunit [19]. The complete fimbrial gene cluster of Hib has been characterized and showed to contain five transcribed open reading frames (ORF; hifA–hifE) [24]. Using mutants that were inactivated in distinct fimbrial genes, van Ham et al. [25]have shown that the adhesive domain resides in the major subunit (HifA) and also showed that both the major and minor subunits (HifD and HifE) were required for adherence of H. influenzae to oropharyngeal epithelial cells and human erythrocytes carrying the AnWj antigen (see below). However, the role of the various fimbrial proteins in H. influenzae adhesion without interference from their function in fimbriae biogenesis was successfully established by constructing Escherichia coli recombinants expressing distinct combinations of H. influenzae fimbrial genes. A clone expressing only the major subunit HifA without coexpression of both minor subunits exhibited the specific adherence properties of H. influenzae fimbriae, implying that the minor subunits are dispensable for adherence and that the adhesive domain resides in the major subunit. The minor subunits probably play a role in adherence by raising the number of fimbriae above the minimal level required to establish adherence since insertion of a kanamycin cassette in hifD or hifE reduces the level of fimbriation. They may also play a role in adherence to other substrates via a distinct receptor-ligand interaction.

A more recent study revealed that H. influenzae fimbriae are composite structures like type 1 and P fimbriae of E. coli[26]. The structure is composed of a flexible two-stranded helical rod comprised of HifA and a short, thin, distal tip fibrillum containing HifD. In type 1 and P fimbriae, the pilus tip contains the adhesive subunit responsible for binding to the relevant host cell receptor. Results of this study raise the possibility that a minor subunit localized to the tip of H. influenzae fimbriae is the true adhesin.

The receptor for Hib fimbriae on human erythrocytes was identified as the blood group AnWj antigen [27]. Inhibition experiments with purified gangliosides revealed that GM1, GM2, GM3, and GD1a inhibited both adherence of fimbriated H. influenzae to epithelial cells and hemagglutination. The asialo derivative of GM1 (without a sialic residue) was a poor inhibitor. The glycolipid GM3 (sialyl-lactosylceramide or NeuAcβ2-3Galβ1-4Glcβ1-1-ceramide) is the minimum structure for the fimbriae-dependent binding of H. influenzae to its receptor on oropharyngeal cells and erythrocytes. The fimbriae of H. influenzae thus belong to the family of low-affinity lactosylceramide binding fimbriae [28].

The gene coding for the fimbrial subunit protein (fimbrin) of NTHi was also cloned and sequenced [29]. When the fimbrin gene was disrupted, a reduced adherence to human oropharyngeal cells was observed. Immunization with isolated fimbrial protein conferred partial protection against transbullar challenge in a chinchilla model. Their data suggest that fimbrin could be useful as a component of a vaccine to protect against otitis media. H. influenzae fimbriae are antigenically diverse however, and protection might be strain specific, although a degree of serological relatedness among fimbrin proteins was observed.

Sterk et al. [30]studied fimbriae-dependent binding to human tissue sections, and showed that the binding was to epithelial cells and other cells involved in colonization and infection by H. influenzae. Their data also suggest that a shift to the non-fimbriated form is required for bacteria in the bloodstream to escape clearance mechanisms mediated by blood cells. A study by Gilsdorf et al. [31]showed that H. influenzae isolates adhere to all of the human respiratory tract cell types they examined. Their results indicate that fimbriae of Hib as well as NTHi isolates mediate adherence to some, but not all, cells derived from human respiratory tissues, and suggest that non-pilus adhesins are also involved in adherence.

2.2.3Surface fibrils

Short, thin surface fibrils distinct from fimbriae were visualized on Hib isolates [32]. The genetic locus encoding these fibrils was characterized recently [33]and shown to be composed of one long ORF, designated hsf, which encodes a protein (Hsf) with a molecular mass of 240 kDa. The derived amino acid sequence demonstrated homology with the NTHi adhesin Hia (see below), as well as with other bacterial adherence factors including AIDA-1 (an adhesion protein expressed by some diarrheagenic E. coli strains), Tsh (a hemagglutinin produced by an avian-pathogenic E. coli strain), and SepA (a Shigella flexneri secreted protein that appears to play a role in tissue invasion). Mutagenesis of this locus resulted in the loss of expression of surface fibrils and a marked decrease in attachment to cultured human epithelial cells (Chang, HeLa and HEp-2).

In another recent study, St-Geme III and Cutter [34]evaluated the influence of fimbriae, fibrils, and capsule on adherence of Hib. They found that fimbriae and surface fibrils have distinct cellular binding specificities, suggesting that these adhesive molecules recognize different host cell receptors. Fimbriae promoted adherence to human buccal epithelial cells, but failed to recognize any of the cultured cell lines (Chang, KB, HEp-2, and HaCaT). In contrast, fibrils were associated with adherence to these cell lines, but failed to influence attachment to buccal epithelial cells. They also noted that capsular material inhibited fibril recognition of the host cell surface. A model for Hib colonization has been proposed [18, 34]. Between host, the organism must avoid desiccation in order to survive, and thus encapsulation is increased. Upon entry of the organism into a new host, the presence of a dense polysaccharide capsule necessitates that fimbriae, which extend beyond the capsule, promote the initial binding event. However, because fimbriae appear to mediate a low-affinity interaction, fibril-mediated adherence may be critical for persistent colonization. Thus, over time, encapsulation is decreased and piliation turned off in order to allow for surface fibril interaction with the appropriate host cell receptor structures. In the occasional circumstance that the organism invades the bloodstream, encapsulation again confers a survival advantage and levels are increased accordingly.

2.2.4High molecular mass adhesion proteins

Two non-fimbrial adhesins, present in NTHi, have been reported and designated HMW1 and HMW2 proteins [18, 35, 36]. These adhesins are antigenically related to the filamentous hemagglutinin of Bordetella pertussis and mediate binding to distinct cell lines. The HMW1 protein is 125 kDa in size, while the HMW2 protein has an apparent molecular mass of 120 kDa. The first 1259 base pairs of the hmw1 and hmw2 coding regions are identical; thereafter, the sequences diverge somewhat but are 80% identical overall. The derived amino acid sequences show 70% identity. The receptors for HMW proteins appear to be negatively charged glycoconjugates; sulfated glycosaminoglycans as well as a glycoprotein containing N-linked oligosaccharide chains with sialic acid in an alpha 2–3 configuration have been identified as putative receptors [37, 38].

Attachment of NTHi to the epithelial surface of the upper respiratory tract is a critical first step in colonization of the human host. In theory, interruption of this colonization process should prevent disease. Interestingly a partial protection was obtained in an experimental model of otitis media when animals were immunized with HMW1 and HMW2 adhesion proteins [39], suggesting that they might represent one component of a multicomponent NTHi vaccine. In addition, the recent identification of shared surface-exposed B-cell epitopes on HMW adhesion proteins by using monoclonal antibodies suggests the possibility of developing recombinant or synthetic peptide-based vaccines [40]. At least one surface-exposed B-cell epitope, define by monoclonal antibody AD6, was shared by most NTHi strains which express HMW1-HMW2-like proteins. This epitope mapped to the last 75 amino acids at the carboxy termini of the two proteins. However, a second family of HMW adhesion proteins expressed by NTHi HMW1/HMW2-deficient isolates was recently reported [41]. The gene, designated hia (for H. influenzae adhesin), encodes for a protein of 114 kDa. An Hia isogenic mutant was constructed and showed a reduced adherence to Chang epithelial cells. The Hsf (encoding surface fibrils) and Hia proteins were found to confer the same binding specificities (using a panel of eight cultured cell lines) [33], suggesting that hsf and hia are alleles of the same locus. Approximately 75% of NHTi isolates express HMW1/HMW2-like adhesins, and most of the remaining isolates contain an Hia homolog [41]. It might therefore be possible to develop vaccines based upon a combination of HMW1/HMW2 proteins and Hia protein which would be protective against disease caused by most or all NTHi.

2.2.5Other putative adhesins

An IgA protease-like protein (Hap) of NTHi has been shown to promote intimate interaction with human epithelial cells [42]. However, the mechanism by which Hap facilitates cellular invasion is not known [18].

H. influenzae undergoes spontaneous phase variation in colony opacity. Weiser et al. [43]have identified a locus contributing to opacity variation and containing two genes oapA and oapB. Mutagenesis of oapA resulted in loss of the ability to colonize the nasopharynx of infant rats.

Using two-dimensional thin-layer chromatography (TLC) and 3H-labeled NTHi as well as fimbriated and non-fimbriated Hib isolates, it was recently shown that H. influenzae binds to minor gangliosides of HEp-2 cells [44]. The lipid binding specificity of H. influenzae isolates was determined by TLC [45]. The 13 clinical isolates tested recognized different lipids including phosphatidylethanolamine (PE). A PE affinity matrix was used to purify an adhesin of 46 kDa from both Hib and NTHi. This adhesin was a potent inhibitor of PE binding in vitro, and polyclonal antibodies specific for this protein prevented the attachment of H. influenzae to cultured HEp-2 epithelial cells. The same group observed that after a brief heat shock treatment, NTHi strains show a long-lasting change in the binding specificity for glycolipids [46]. After exposure of the organisms to brief heat shock, Western blotting of a surface extract of H. influenzae with anti-bovine brain hsp-70 monoclonal antibody showed an increase in two protein bands at 82 and 60 kDa. In addition, this antibody was a potent inhibitor of the binding of heat-shocked H. influenzae to sulfoglycolipids. Their data indicate that cell surface hsp-70-related heat shock proteins can mediate H. influenzae attachment to sulfoglycolipids following heat shock. The authors suggest this phenomenon may be a response to fever following H. influenzae infection in humans, since a 39°C heat shock, often occurring during fever, was enough for the effect. Interestingly, heat shock proteins have recently been implicated in adherence of other microorganisms including Mycoplasma[47], Chlamydia trachomatis[48], and Helicobacter pylori[49].

The effect of mutations in genes required for LPS synthesis on the interaction of Hib with human nasal turbinate tissue maintained in an organ culture was recently examined [50]. Isogenic mutants expressing truncated LPS due to mutations in lic1lic2 or galEK genes were used. Unlike studies with H. ducreyi in which LOS are involved in adherence, their data do not support a role for LPS as an adhesin for Hib.

2.2.6Binding to mucus

Respiratory tract mucins may function as receptor molecules for NTHi and thus play an important role in colonization. Several studies have demonstrated that most NTHi isolates can bind to mucus from human respiratory tract [51–53]. However, encapsulated and some NTHi isolates failed to interact with mucins [52]. These isolates would, therefore, adhere only after disruption of the mucus layer and exposure of cellular receptors. Thus, differences in tissue toxicity and invasiveness among H. influenzae isolates may also be influenced by the mucin interactions of the isolates.

The binding of NTHi to human nasopharyngeal mucin, evaluated using an overlay assay, appeared to be mediated by outer membrane proteins (OMPs) of H. influenzae and sialic acid-containing oligosaccharides of mucin [54]. On the basis of electrophoretic mobility and Western blotting, these proteins were identified as OMPs P2 and P5. Unequivocal identification of OMPs P2 and P5 was made by employing OMPs of bacterial strains and their respective mutants lacking the respective OMP in binding assays. Miyamoto and Bakaletz [55]also observed that OMP P5 (which is in fact a fimbrin), but not OMP P2, contributes to the binding of NTHi to mucus and epithelial cells of chinchilla Eustachian tube.

2.2.7Binding to extracellular matrix components

Strains of H. influenzae adhere to the extracellular matrix (ECM) and to its isolated components, including laminin, fibronectin, and various collagens [56]. In addition, plasmin generated on H. influenzae plasminogen receptors degraded laminin and fibronectin as well as ECM from human endothelial cells. Plasmin bound on H. influenzae cells also potentiated penetration of bacteria through a basement membrane preparation reconstituted on membrane filters. Their data give evidence for a role of ECM adherence and plasminogen activation in the spread of H. influenzae through tissue barriers.

2.2.8Cell invasion

Hib as well as NTHi are able to penetrate the mucosal surface predominantly between epithelial cells (a process known as paracytosis) [57]. The bacterial cells are mainly found in clusters (or microcolonies) in crevices between the cells. The passage time of H. influenzae through cell layers of NCI-H292 lung epithelial cells (originating from a human lung mucoepidermoid carcinoma) was not influenced by the presence of capsule or fimbriae or by the ability of the bacteria to adhere to the epithelial cells [57]. However, highly adherent strains showed greater paracytosis. More recently, it has been observed that Hib occasionally exhibits highly invasive behavior with HEp-2, HeLa, and MDCK cell lines [58]. The phenomenon was not inhibited by colchicine or cytochalasin but was dependent on the presence of physiological levels of CO2. These authors proposed that the sensing by Hib of an increase of CO2 concentration may lead to phenotypic changes, such as increased invasiveness, that enhance colonization and persistence.

3Actinobacillus

3.1Actinobacillus actinomycetemcomitans

A. actinomycetemcomitans is an important periodontopathogen that has been implicated in juvenile and adult periodontitis, diseases characterized by rapid destruction of the tooth-supporting tissues. This organism possesses a large number of virulence factors which enable it to colonize the oral cavity, invade periodontal tissues, evade host defences, initiate connective tissue destruction and interfere with tissue repair (for recent reviews see [59, 60]).

Most fresh isolates of A. actinomycetemcomitans are fimbriated, and binding of this microorganism to solid surfaces (e.g. hydroxyapatite and saliva-coated hydroxyapatite) may involve fimbriae [61]. The parameters of A. actinomycetemcomitans adhesion to epithelial cells were assayed by use of different cell lines including the KB cell line, derived from a human oral epidermoid carcinoma [62–64]. Adhesion of A. actinomycetemcomitans to epithelial cells involves multiple determinants (fimbriae, OMPs, vesicles, and/or an extracellular amorphous material) and is influenced by both bacterial and host environmental conditions. Optimal adherence was observed after growth of the bacterial cells in broth under anaerobic conditions [63]. The adhesins that mediate adherence are associated with the outer membrane or are released into the surrounding medium in the form of vesicles [59, 60, 62]. As opposed to Actinobacillus pleuropneumoniae (see below), LPS does not appear to be involved in adhesion of A. actinomycetemcomitans to epithelial cells [64].

Oligopeptides were synthesized according to the amino acid sequence of the fimbrial protein of A. actinomycetemcomitans, conjugated with branched lysine polymer resin beads, and used to immunize rabbits [65]. An antiserum which reacted with a 54 kDa protein of the fimbriae from A. actinomycetemcomitans strongly inhibited the attachment of fimbriated A. actinomycetemcomitans to saliva-coated hydroxyapatite beads, buccal epithelial cells, and a fibroblast cell line, Gin-1. Such synthetic fimbrial peptide antigen might act as a vaccine for inducing an antibody response that inhibits A. actinomycetemcomitans colonization. Recently, the fimbriae associated protein (fap) gene was cloned [66]. The 228 bp ORF encoded a 7.9 kDa protein. It is not known at this moment whether the 54 kDa protein previously reported is a complex of 7.9 kDa protein or the 7.9 kDa protein is one of the components of the mature fimbriae.

Mintz and Fives-Taylor [64]have shown that, in the presence of saliva, adherence of A. actinomycetemcomitans to human oral epithelial cells is inhibited. Salivary components which interact with A. actinomycetemcomitans were investigated recently by Groenink et al. [67]. Their results indicate that IgA, the low molecular mucin MG2, parotid agglutinin, and a 300 kDa sublingual and submandibular glycoprotein, bound to the bacterial strains tested. In addition, sialic acid residues on MG2 appear to be involved in the binding.

Alugupalli and Kalfas [68]have demonstrated that lactoferrin, which is found in increased concentrations at sites of inflammation and in the gingival crevicular fluid of diseased periodontium, inhibited the adhesion of A. actinomycetemcomitans to different cell monolayers including KB cells. In A. actinomycetemcomitans, lactoferrin binds to a major 34 kDa heat-modifiable outer membrane protein [69]. This bacterium is also able to bind to the basement membrane matrix and its isolated components such as fibronectin, laminin and type IV collagen [70]. Interestingly, the above-mentioned heat-modifiable OMP has also been found to bind the basement membrane protein laminin, and lactoferrin can inhibit and displace the laminin-bacteria interaction [71]. These results indicate that lactoferrin may prevent the establishment of bacteria in periodontal tissues through adhesion counteracting mechanisms in addition to its bacteriostatic and bactericidal properties.

Once colonization has taken place, there is some evidence that A. actinomycetemcomitans can invade the tissues of the periodontium. The invasion process was studied using both KB and MDCK (Madin Darby canine kidney) epithelial cells [72]. The frequency of invasion by A. actinomycetemcomitans is comparable with the invasion of other cultured cells by known invasive microorganisms [62]. The degree of invasion by A. actinomycetemcomitans is greater in KB cells than in cells of non-oral origin or other commonly used cell lines. The work of Meyer et al. [72]revealed that invasion of epithelial cells is a multistep process which involved attachment to the epithelial cell and triggering of the movement of actin, entry into the host cell in an endosome, escape from the endosome, rapid multiplication, and intracellular and intercellular spread. Soon after entry of A. actinomycetemcomitans bacteria into epithelial cells, they may be found in protrusions which sometimes extend between neighboring epithelial cells. These protrusions are thought to mediate the cell-to-cell spread of A. actinomycetemcomitans. Meyer et al. [72]postulated that the events they described are involved in the ability of A. actinomycetemcomitans to spread to the gingival and connective tissue and cause destruction.

3.2Actinobacillus pleuropneumoniae

A. pleuropneumoniae is the causative agent of porcine pleuropneumonia, a disease found worldwide that causes tremendous economic loss to the swine industry. Several bacterial components, including hemolysins and/or cytolysins, LPS, and capsular polysaccharides, appear to contribute to the disease process (for a recent review see [73]). Fimbriae have been demonstrated on some isolates of A. pleuropneumoniae[74–76]but their role in adherence has not been established.

We have previously showed that A. pleuropneumoniae was able to agglutinate erythrocytes from various animal species [77]and adheres in vitro to porcine tracheal rings maintained in culture [78]and to porcine frozen tracheal and lung sections [79]. Dom et al. [75]evaluated the in vivo association of a serotype 2 strain with the respiratory epithelium of pigs. As soon as 30 min postinoculation, bacteria were intimately associated with the epithelium of the alveoli or the cilia of the terminal bronchioli.

Our group demonstrated that LPS were the major adhesin involved in adherence to porcine respiratory tract cells [78]and mucus [80, 81]. We observed that the degree of adherence of A. pleuropneumoniae to porcine tracheal rings was related to LPS profiles and that adherence was blocked by purified LPS [78]. Isolates with a smooth LPS profile adhered in larger numbers to tracheal rings than isolates with a semirough LPS profile. We then found, by using extracted LPS from serotypes 1 and 2, that the polysaccharidic part of this complex molecule, but not the lipidic part, was involved in binding to host cells [82]. We also showed in the latter study, using immunoelectron microscopy and flow cytometry, that LPS were well exposed at the surface of this heavily encapsulated bacterium, an essential prerequisite for a bacterial adhesin. Recent results, using mini-Tn10-generated isogenic serotype 1 LPS mutants, indicate that the core region of LPS might play a determinant role in adherence of A. pleuropneumoniae[83]. It is noteworthy that with Pseudomonas aeruginosa, LPS outer core is an important bacterial ligand involved in the binding and ingestion by airway epithelial cells expressing CFTR (cystic fibrosis transmembrane conductance regulator) [84]and by corneal cells [85]. Preliminary experiments identified two proteins of approximately 15 and 39 kDa as putative receptors for A. pleuropneumoniae LPS on porcine tracheal cells [86].

Interestingly, immunization with different adhesin (LPS)-based vaccine preparations has been shown to induce a good protection against challenge with a virulent strain of A. pleuropneumoniae in mice [87]and in pigs [88]. These preparations, however, did not induce cross-protection against other serotypes.

4Pasteurella

4.1Pasteurella haemolytica

P. haemolytica is the cause of bovine pneumonic pasteurellosis, also known as shipping fever, as well as pneumonic pasteurellosis in sheep [89, 90]. The adhesins involved in colonization of host mucosal surfaces are not known [91], although two types of fimbriae have been described [92]. It has been proposed that capsular polysaccharides (CPS), LPS, and OMPs may function as adhesins in the absence of fimbriae [91].

Interestingly, LPS are present on epithelial surfaces of experimentally infected calves [93]which might suggest a role in adherence as observed in some other Pasteurellaceae[94]. Lipopolysaccharides of P. haemolytica have also been shown to complex with lung surfactant covering the alveoli [95]. However, this mechanism may help clear certain Gram-negative bacteria from the lungs of sheep as a part of the pulmonary innate defense system. Ovine pulmonary surfactant induces killing of P. haemolytica[96].

It is hypothesized that predisposing factors, management and environmental stress factors or viral infection, alter the upper respiratory epithelium allowing P. haemolytica to colonize [90, 91]. In a recent study, Mosier et al. [97]examined the lectin histochemistry of normal and bovine herpesvirus-1 (BHV1)-infected bovine nasal mucosa. They found a greater reactivity in samples from BHV1-infected than from normal cattle for all lectins tested. The BHV1-induced alteration of nasal mucosal glycoconjugates could enhance adhesion and colonization of P. haemolytica to nasal surfaces. Indeed, using an in vitro assay, it has been observed that infection of bovine cell lines by BHV1 resulted in an increased adherence by P. haemolytica[98].

4.2Pasteurella multocida

P. multocida is an important veterinary pathogen causing diseases in a variety of domestic mammals and birds; in pigs, P. multocida is associated with atrophic rhinitis (toxigenic, capsular type D strains) and pneumonia (non-toxigenic, capsular type A strains) [99]. P. multocida is considered to be a poor colonizer and, as for P. haemolytica in bovine, a variety of factors can predispose the pig nasal cavity to colonization by toxigenic strains including treatment with a mild chemical irritant (e.g. diluted acetic acid) [100]or exposure to ammonia [101], or pre-infection with Bordetella bronchiseptica[100, 102]. B. bronchiseptica appears to facilitate upper respiratory tract colonization by P. multocida by a process which involves a low molecular mass substance, possibly the tracheal cytotoxin [103]. This toxin induces ciliostasis of the tracheal epithelium with a concomitant accumulation of mucus.

Several studies have evaluated the attachment of P. multocida to various porcine respiratory tract cells, and to porcine respiratory tract mucus (e.g. [102, 104–107]), but the adhesins have not been clearly identified. Hemagglutinins [108]and fimbriae [109]have been reported, but it was found that these two factors were independent. More recently, two types of fimbriae (curly or rigid) were observed on toxigenic strains of P. multocida[110], but piliated cells failed to attach to red blood cells and to immobilized respiratory tract mucus. It is still not known whether these structures function as adhesins in vivo. The rigid fimbriae, which are approximately 7 nm in diameter, were identified as type 4 fimbriae based on N-terminal amino acid sequence of the 18-kDa fimbrial subunit [111].

Lipopolysaccharides may be involved in adherence of P. multocida to porcine respiratory tract cells [112]. However, the partial inhibition of adherence to porcine tracheal rings maintained in culture by P. multocida LPS suggests that there may be non-LPS components that are also important in adherence. Esslinger et al. [113]found that capsular type A isolates adhered strongly to HeLa cells and to alveolar macrophages from different animal species. Interestingly, they suggested a possible role of capsular hyaluronic acid in adhesion since it was reduced by hyaluronidase or hyaluronic acid. A similar observation was made by Pruimboom et al. [114]. Their data indicate that adhesion to turkey air sac macrophages is mediated by capsular hyaluronic acid and that the receptor is a glycoprotein. The same group reported that a monocytic CD44 isoform is the receptor for capsular hyaluronic acid on cultured turkey peripheral blood monocytes [115].

A surface protein of P. multocida has also been implicated in adherence. It has been shown that the capacity of OMP preparations of P. multocida to bind to respiratory mucosal surface preparations was inhibited by antibodies against the major OMP (35 kDa, p35) of P. multocida[116]. These antibodies also cross-reacted with a 44-kDa major OMP of P. haemolytica and, according to these authors, p35 could therefore represent a potential candidate for a subunit vaccine against pneumonic pasteurellosis.

Finally, avian strains of P. multocida were shown to be able to invade porcine epithelial cells (PK15) and feline epithelial cells (CRFK) in culture, but not rabbit epithelial cells (RK13) [117]. Apparently, avian P. multocida enters polarized epithelial cells by interacting with host F-actin [118]. Invasion might be a mechanism of pathogenicity for P. multocida, contributing to colonization or virulence of avian strains.

5Conclusion

Most HAP organisms can employ multiple molecular mechanisms of adherence (or multiple adhesins) to initiate infection (Table 2). Indeed, a wide variety of adhesins are expressed by members of the Pasteurellaceae. Proteins (e.g. fimbriae, fibrils, OMPs) and/or polysaccharides (LOS, LPS, CPS) were shown to play a role in adherence of these organisms. These multiple adhesins are capable of reacting with different receptors. Furthermore, these multiple adhesins may be utilized in a stepwise fashion during colonization. It seems reasonable that evolutionary pressures have selected organisms that are capable of demonstrating more than one mechanisms of adherence [1]. The multiplicity of HAP organisms adherence interactions accounts, at least in part, for their success in colonizing the mucosal surfaces of their hosts. In many instances, adhesins of HAP organisms have been shown to represent good vaccine candidates.

Table 2.  Adhesin-receptor systems found in HAP organisms
HAP organismAdhesinReceptorReference
  1. CPS, capsular polysaccharides; ECM, extracellular matrix; HSP, heat shock protein; LOS, lipooligosaccharides; LPS, lipopolysaccharides; OM, outer membrane; OMP, outer membrane protein; PE, phosphatidylethanolamine; (?) role in adherence not yet established.

H. ducreyiLOS[15–17]
 fimbriae (?)[13]
 ECM components[11, 17]
H. influenzaefimbriae (Hib)blood group AnWj Ag, sialyl-lactosylceramide (GM1, GM2, GM3)[19, 21–25, 27, 28]
 fimbriae (NTHi)[29]
 fibrils[32, 33]
 HMW1/HMW2 adhesion proteinsglycoprotein; sulfated glycosaminoglycans[35–38]
 Hia adhesion protein[41]
 Hap (IgA protease-like protein)[42]
 OapA (opacity protein)[43]
 OMP 46 kDalipids, including PE[45]
 hsp-70-like proteinssulfoglycolipids[46]
 minor gangliosides[44]
 ECM components[56]
 mucus/mucins[51–53]
 OMP P2 and P5sialic acid-containing oligosaccharides of mucin[54]
 OMP P5mucus[55]
A. actinomycetemcomitansfimbriae[61, 65]
 protein(s) found in OM vesicles[63, 64]
 major OMP (29/34 kDa)laminin (ECM)[71]
A. pleuropneumoniaefimbriae (?)[74–76]
 LPS[78, 82]
 LPS core region[83]
 LPS15- and 39-kDa proteins of porcine tracheal cells[86]
P. haemolyticafimbriae (?)[92]
 CPS (?), LPS (?), OMPs (?)[91]
P. multocidafimbriae (?)[109–111]
 hemagglutinins[108]
 major OMP (p35)[116]
 capsular hyaluronic acid[113, 114]
 capsular hyaluronic acidCD44[115]
 LPS[112]

Many important questions remain to be answered regarding HAP adhesins, regulation of adhesins expression, and prevalence of adhesins or adhesin genes among isolates. Surprisingly, not much is known about the receptors recognized by HAP adhesins (Table 2). New information on HAP adhesins and their receptor will allow for the rational design of novel subunit or recombinant vaccines, and adhesion agonists for the control or prevention of HAP colonization and diseases.

Acknowledgements

Work in M.J. laboratory was supported by grants from the Natural Sciences and Engineering Research Council of Canada (OGP0003428). We also wish to thank the other graduate students and collaborators who have been involved in the A. pleuropneumoniae LPS project: M. Abul Milh, M. Archambault, M. Bélanger, D. Dubreuil, J. Frey, C. Galarneau, M. Gottschalk, J. Harel, M. Kobisch, J. Nicolet, and S. Rioux. We apologize to the authors whose work could not be quoted, because of space limitations.

Ancillary