Pulsotype Diversity of Clostridium botulinum Strains Containing Serotypes A and/or B Genes

Clostridium botulinum strains are prevalent in the environment and produce a potent neurotoxin that causes botulism, a rare but serious paralytic disease. In 2010, a national PulseNet database was established to curate C. botulinum pulsotypes and facilitate epidemiological investigations, particularly for serotypes A and B strains frequently associated with botulism cases in the United States. Between 2010 and 2014 we performed pulsed-field gel electrophoresis (PFGE) using a PulseNet protocol, uploaded the resulting PFGE patterns into a national database, and analyzed data according to PulseNet criteria (UPGMA clustering, Dice coefficient, 1.5% position tolerance, and 1.5% optimization). A retrospective data analysis was undertaken on 349 entries comprised of type A and B strains isolated from foodborne and infant cases to determine epidemiological relevance, resolution of the method, and the diversity of the database. Most studies to date on the pulsotype diversity of C. botulinum have encompassed very small sets of isolates; this study, with over 300 isolates, is more comprehensive than any published to date. Epidemiologically linked isolates had indistinguishable patterns, except in four instances and there were no obvious geographic trends noted. Simpson's Index of Diversity (D) has historically been used to demonstrate species diversity and abundance within a group, and is considered a standard descriptor for PFGE databases. Simpson's Index was calculated for each restriction endonuclease (SmaI, XhoI), the pattern combination SmaI-XhoI, as well as for each toxin serotype. The D values indicate that both enzymes provided better resolution for serotype B isolates than serotype A. XhoI as the secondary enzyme provided little additional discrimination for C. botulinum. SmaI patterns can be used to exclude unrelated isolates during a foodborne outbreak, but pulsotypes should always be considered concurrently with available epidemiological data.