An Investigation of the Distinguishing Features of C. Jejuni Strains that have Host/Disease Associations

Final report: The overall aim of this project was to investigate the relationship between C. jejuni strain, host and virulence. As very little is known about the virulence of C. jejuni a multi-pronged, fundamental approach was adopted. Multi-locus sequence typing was applied to veterinary and human isolates to investigate host-specific groupings within the population of C. jejuni. In addition three potential virulence mechanisms were investigated at the molecular level to try to understand variations in virulence that have been observed between strains and to identify markers of pathogenicity.
<P>
Campylobacter jejuni and C. coli are the most common causes of acute bacterial enteritis in Great Britain, with over 54000 reported cases in England and Wales, in 1999. The consumption of contaminated food; particularly poultry meat is believed to be the main cause of human campylobacteriosis. However, recent evidence from both molecular typing schemes and population genetic studies, has suggested that poultry may not be the sole source of human infection and other potential sources include domestic and companion animals, water and milk. These data also suggest that sub-populations of C. jejuni exist that appear to be associated with a particular host or disease and that not all campylobacters are pathogenic for humans. The presence of non-pathogenic strains, particularly in foods, could confound or bias epidemiological studies, which primarily use data from total bacterial counts. Thus indicators of the relationships between strains, pathogenicity and host are urgently needed. A two-pronged approach was adopted in order to investigate these relationships:
<P>1: a study of the population structure of C. jejuni, using MLST, to determine whether the populations of human and veterinary isolates overlap; and
<P>2: to investigate putative pathogenic mechanisms of C. jejuni at the molecular level with the aim of identifying markers of pathogenicity.
<P>
1. MLST approach. Data from the MLST study of 266 isolates of veterinary and human origin indicated that the populations of veterinary and human isolates do overlap and so campylobacters from all sources should be considered as potentially pathogenic to man. However, there was some evidence for the association of certain ST complexes with a particular source, but there were no complexes that contained just isolates from a single source. A number of porcine isolates did appear to cluster together and may represent a pig-adapted group of strains. In addition another group of highly-related strains from various sources and geographical locations was identified which may well represent another stable strain of C. jejuni. This group of strains is currently being investigated further in a Seedcorn project.
<P>
2. Molecular pathogenesis approach. The mechanism by which C. jejuni causes disease in humans is still not fully understood. Invasion of the host epithelium, with prior, transient adhesion, and cytolethal distending toxin (CDT) are the main virulence mechanisms suggested to date. Variation in invasion and CDT activity is known to exist between strains, the molecular basis for which was investigated in this project.
<P>
Invasion: A number of hyper-invasive strains were available from a previous project. Transposon mutagenesis was applied to one such strain and a screen of the initial batch of 96 mutants revealed 27 mutants with reduced invasion compared to the wild-type. The location of the transposon has been determined in 24 of these mutants and all but one are located in genes already present in NCTC11168, most of which are involved with metabolism and survival in vivo. The exception, is an insertion into the dtpT gene, identified originally in strain 81116 following subtractive hybridisation. Two of the genes were selected for further characterisation. In NCTC11168 Cj0490 encodes the C-terminus of the aldehyde dehydrogenase gene, whereas Cj0489 encodes the N-terminus. In 01/51 and two of the other hyper-invasive strains that possess this gene Cj0489 and Cj0490 are part of the same gene. A single nucleotide deletion from NCTC11168, generating a stop codon, appears to have resulted in the two open reading frames. The relevance of this polymorphism to invasion will be investigated further with the use of a probe based approach. The other gene, Cj0497, encodes a putative lipoprotein which when disrupted results in a reduction in adhesion to and invasion of INT407 cells, but does not affect the motility. Cj0497 may encode a putative adhesin.
A study of genome complement, as determined by PCR, in strains with known invasion potential did not reveal any association between the presence of known virulence-related genes, or any of the genes identified within this project, and the invasion potential. Neither was there any correlation with MLST type and invasion potential. Evidence was generated to suggest that the mechanism of uptake used by the hyperinvasive strains was different from that utilised by strains with lower invasiveness, although further work is needed to confirm this.
<P>
Toxin: The molecular basis for the lack of CDT activity in certain strains was determined to be due either to a deletion between cdtA and cdtB or to two point mutations in cdtB. Neutralisation assays using pooled human and chicken antisera suggested that CDT is expressed and is antigenic during human infection. However, CDT is not recognised by chicken sera, despite apparent expression in the chicken as determined by RT-PCR on caecal contents. Recent evidence indicated that the observed variation in CDT activity between strains may be due to polymorphisms in the promoter region and this is currently being investigated further. In addition to the CDT a cytolethal rounding toxin activity has been observed in a number of isolates of C. jejuni.
<P>
Surface structures: As well as the two widely-accepted virulence mechanisms described above the appearance of fimbriae-like structures on the surface of C. jejuni was also investigated. These structures were observed in the absence of bile salts, and mutagenesis of the pspA gene, previously described as having a role in fimbriae expression in C. jejuni, failed to have an effect on the expression of the structures observed in our study. Attempts at purification of this material resulted in the identification of a hypothetical protein that appears to be essential to the biology of C. jejuni, although its role, if any, in the expression of fimbriae is unclear.
<P>
Finally, transcript analysis was carried out on two variants of the same strain, both with different phenotypic properties, including chick colonisation, as well as motility, morphology and invasiveness. The expression of a large number of genes was found to differ between the two variants, including a large number of genes involved with the adaptation to microaerobic growth.
<P>
In summary, the relationship between strains, pathogenicity and host appears to be complex. As C. jejuni is highly diverse both at the genetic level and phenotypically it is not unsurprising that no clear correlation has been identified. Despite this, a number of novel, putative virulence genes have been identified, some of which are highly conserved amongst C. jejuni. It would be interesting to investigate polymorphisms in such genes to determine their use as additional markers to supplement current sequence based typing schemes.

<P>