Numerical analysis of DNA microarray data of Campylobacter jejuni strains correlated with survival, cytolethal distending toxin and haemolysin analyses.

Molecular epidemiological studies of the enteric pathogen Campylobacter jejuni have suggested that not all animal isolates are equally pathogenic to humans. We examined the use of numerical analysis of whole-genomotype data as a potential tool for evaluating C. jejuni virulence potential. Whole-genome microarray analysis was used to determine the gene-level complementarity of 12 Danish strains to the pathogenic, genome-sequenced strain NCTC 11168. Cytolethal distending toxin (CDT) and haemolysin activities, and survival characteristics under aerobic conditions at room temperature were also determined. Among the strains examined, 439 genes were polymorphic. Numerical analysis of these data by use of the squared Euclidean distance coefficient and Ward's clustering method clearly delineated strains into two clusters. CDT and haemolysin activities of cluster 1 strains were not statistically significantly different from cluster 2 strains. However, viability during aerobic incubation of cluster 1 strains was statistically significantly lower than corresponding estimates of cluster 2 strains. The number of missing or highly divergent genes in cluster 1 strains with respect to NCTC 11168 was also statistically significantly greater compared with those of cluster 2 strains. Sixty-seven genes present in NCTC 11168 were characteristically missing or divergent among cluster 1 strains. Of these, 53 genes were localised within 11 major gene clusters, of which eight were associated with surface structures and included flagellar, lipo-oligosaccharide, and membrane transport proteins. Our data indicate a correlation between C. jejuni genomic content, particularly in surface-coding regions, and its capacity for environmental survival, and may help explain why certain serotypes are more commonly reported in human disease.