Analysis of pulsed field gel electrophoresis profiles using multiple enzymes for predicting potential source reservoirs for strains of Salmonella Enteritidis and Salmonella Typhimurium isolated from humans

Abstract We reported previously on a highly discriminatory pulsed field gel electrophoresis-based (PFGE) subtyping scheme for Salmonella enterica serovar Enteritidis (SE) and Salmonella Typhimurium (ST) that relies on combined cluster analysis of up to six restriction enzymes. This approach allowed for the high-resolution separation of numerous poultry-derived SE and ST isolates into several distinct clusters that sorted along several geographical and host-linked boundaries. In this study, 101 SE and 151 ST strains isolated from poultry, swine, beef, mouse, and produce origins were combined with 62 human SE and ST isolates of unknown sources. PFGE profiles were generated across six restriction enzymes ( Xba I, Bln I, Spe I, Sfi I, Pac I, and Not I) for human SE and ST isolates. The combined six-enzyme UPGMA trees of SE and ST revealed six separate origins of North American human SE isolates including one association with a “cosmopolitan” cluster of SEs from poultry originating in Scotland, Mexico, and China. In the case of ST, human isolates assorted readily along host lines rather than geographical partitions with the majority of human STs clustering in a larger group of STs of potential porcine origin. Such observations may underscore the ecological importance of poultry and pork reservoirs for SE and ST transmission to humans, respectively. In an examination of the relationship between enzyme diversity and congruence among enzymes, pairwise genetic diversity ranged from 6.5% to 9.7% for SE isolates and, more widely, from 17.5% to 27.4% for ST isolates. Phylogenetic congruence measures singled out Xba I, Bln I, and Sfi I as most concordant for SE while Xba I and Sfi I were most concordant among ST strains. Thus, these data provide the first proof of principal for concatenated PFGE, when coupled with sufficient enzyme numbers and combinations, as one effective means for predicting geographical and food source reservoirs for human isolates of these two highly prevalent Salmonella serovars.