The aim of the study is to examine different thresholds of similarity coefficients for reliable clustering of clinical and dairy enterococci originating from various geographical regions in Bulgaria, based on randomly amplified polymorphic DNA polymerase chain reaction (RAPD-PCR) unweighted pair group method with arithmetic mean (UPGMA) analysis. For creating subsets of different strains on a relatively independent basis, was created a selection based on the presence of some bacteriocin genes. A collection of 110 Enterococcus faecalis and 51 Enterococcus faecium nosocomial strains from in-patients in five hospitals, covering the major cities of Bulgaria, and additional 11 Enterococcus faecalis and 15 Enterococcus faecium strains isolated from home-made and artisanal cheeses in different mountain regions of the country was screened for the presence of five commonly found bacteriocin genetic determinants – enterolysin A (EntlA) and bacteriocin AS-48 for E. faecalis and enterocin A (EntA), enterocin B (EntB) and enterocin P (EntP) for E. faecium. Strains carrying the same bacteriocin gene were subjected separately to RAPD-PCR, followed by UPGMA analysis. The obtained results showed that clustering should be performed with similarity coefficients above 95% in order to separate the starins based on cheeses’ or clinical origin and according to their geographical origins. Our findings indicate that the largely accepted similarity coefficient of 70% for clonal relatedness should be reconsidered and the threshold raised up.
Dairy and clinical Enterococcus spp., RAPD-PCR, UPGMA, clustering thresholds