MOLECULAR ANALYSIS OF THE VIRULENCE DETERMINANTS OF CLOSTRIDIUM-PERFRINGENS ASSOCIATED WITH FOAL DIARRHEA

During an epidemiological study of foal diarrhoea, over half of the ca ses yielded Clostridium perfringens which was significantly associated with disease (Netherwood et al., 1996b). However, the association cou ld not be accounted for by enterotoxigenic isolates which had a low pr evalence (Netherwood et al, 1997). Nonetheless, we have hypothesized t hat the association may be caused by a pathogenic sub-population which would be significantly more common amongst C. perfringens-positive ca ses compared with C. perfringens-positive healthy controls if it acted as a pathogen when present. Conversely, if foal diarrhoea caused by C . perfringens was dependent on a predisposing factor, then such an ass ociation might not be evident. As a first step to determine if a molec ular marker was more frequently to be found in C. perfringens-positive cases than controls, we have genotyped the study isolates (up to five per foal) by polymerase chain reaction (PCR) based on the published g ene sequences for the major lethal toxins alpha, beta, epsilon and iot a as well as for theta toxin, large and small sialidases, hyaluronidas e and virulence regulation. Isolates of major toxin types B, C, D and E, or isolates which were untypeable, were isolated from less than 15% of C. perfringens-positive foals and these were not associated with d iarrhoea nor were they more commonly found in C. perfringens-positive cases. Isolates of type A were found in more than 90% of all C. perfri ngens-positive foals. A number of different genotypes were identified by their different patterns of gene possession but types without any o f the genes for theta toxin, large and small sialidases, hyaluronidase and virulence regulation were found in only 10% of positive foals. On ly type A isolates with all of these genes were associated with diarrh oea overall but they were not more commonly isolated from C. perfringe ns-positive cases than controls. In conclusion, genotyping by the sequ enced virulence genes did not identify a marker for a sub-population o f C. perfringens which may be acting more frequently as a pathogen whe n present.