Comparison of Repetitive Sequence-based PCR Characterization and PFGE for the Molecular Typing of Salmonella in Food Matrices [POSTER]

Salmonella causes estimated 1.3 million cases of food-borne infections annually in the U.S. with high morbidity and great economic consequences. Currently, Pulsed-Field Gel Electrophoresis (PFGE) is the standard tool for Salmonella subtyping in food-borne illness outbreaks. However, PFGE is labor intensive, requires highly skilled personnel, and takes 3 to 4 days to complete. In contrast, the Biomerieux DiversiLab (DL), a commercially available, semi-automated, repetitive sequence based DNA fingerprinting system (rep-PCR), has been shown to be considerably less labor intensive than PFGE and provides rapid strain typing results in one work-day.

To assess the ease of use and accuracy of DiversiLab System, its results were compared to PFGE cluster analysis generated by the BioNumerics software for samples of known serotypes. A total of 23 well-characterized Salmonella isolates from various food and environmental matrices were tested. The isolates included 4 different serotypes and grouped by their PFGE patterns relatedness. The DiversiLab Salmonella kit was used for rep-PCR amplification of intergenic repetitive elements in the bacterial genome. The amplicons were separated using the Agilent Bioanalyzer that incorporated a DNA microfluidics chip (LabChip) for fragment separation and detection.

Data analysis was performed using the DL web-based software, which included Kullback-Leibler method for similarity calculation and UPGMA for creating the dendrogram. Finally, DiversiLab data were compared to PFGE cluster analysis generated by the BioNumerics software per CDC PulseNet protocol of XbaI restriction enzyme digest of Salmonella. Comparison results demonstrated rep-PCR and PFGE interpretations were concordant in 23/23 (100%) isolates tested.

Furthermore, rep-PCR dendrograms showed > 98% similarity for isolates having indistinguishable PFGE patterns. However, rep-PCR results also showed > 98% similarity for isolates with 1 or 2 band difference(s) in their PFGE clusters. Although the correlation between rep-PCR and PFGE was highly concordant, data also suggested PFGE was more discriminatory in differentiating highly related strains.