Molecular and phenotypic characterization of Pseudomonas spp. isolated from milk.

Putative Pseudomonas spp. isolated predominantly from raw and processed milk were characterized by automated ribotyping and by biochemical reactions. Isolates were biochemically profiled using the Biolog system and API 20 NE and by determining the production of proteases, lipases, and lecithinases for each isolate. Isolates grouped into five coherent clusters, predominated by the species P. putida (cluster A), P. fluorescens (cluster B), P. fragi (as identified by Biolog) or P. fluorescens (as identified by API 20 NE) (cluster C), P. fragi (as identified by Biolog) or P. putida (as identified by API 20 NE) (cluster D), and P. fluorescens (cluster E). Isolates within each cluster also displayed similar enzyme activities. Isolates in clusters A, C, and D were generally negative for all three enzyme activities; isolates in cluster B were predominantly positive for all three enzyme activities; and isolates in cluster E were negative for lecithinase but predominantly positive for protease and lipase activities. Thus, only isolates from clusters B and E produced enzyme activities associated with dairy product flavor defects. Thirty-eight ribogroups were differentiated among the 70 isolates. Ribotyping was highly discriminatory for dairy Pseudomonas isolates, with a Simpson's index of discrimination of 0.955. Isolates of the same ribotype were never classified into different clusters, and ribotypes within a given cluster generally showed similar ribotype patterns; thus, specific ribotype fragments may be useful markers for tracking the sources of pseudomonads in dairy production systems. Our results suggest that ribogroups are generally homogeneous with respect to nomenspecies and biovars, confirming the identification potential of ribotyping for Pseudomonas spp.