Typing of Proteus mirabilis by bacteriocin production and sensitivity as a possible epidemiological marker.

Two bacteriocin typing methods, based on bacteriocin production and bacteriocin sensitivity, were developed to aid in the separation of strains of Proteus mirabilis. One hundred sixty-two isolates of P. mirabilis and 10 Cradock-Watson bacteriocin producers were grown inproteose peptone no. 3 broth and induced with mitomycin C under culture conditions found optimum for bacteriocin production. Crude bacteriocin lysates were spotted on 186 indicator strains, and after incubation for 18 h at 35 degrees C, positive zones of inhibition were recorded. A cluster analysis computer program was used to select 16 bacteriocin-producing and 16 indicator strains for inclusion in two bacteriocin typing sets. One hundred clinical isolates of P. mirabilis were differentiated by bacteriocin sensitivity into 41 distinct patterns, with 72% of the strains typable, whereas typing by bacteriocin production demonstrated 29 separate lysis patterns among the 80% typable strains. Combining the results of each typing method resulted in 72 individual bacteriocin production-sensitivity patterns and 91% of the isolates typed. Typing 14 epidemic strains by bacteriocin production revealed that 13 of 14 strains were identical, whereas only 2 of 14 strains were typable by bacteriocin sensitivity. Electron microscopy of partially purified bacteriocin revealed tail components of contractile bacteriophages. Standard bacteriocin lysates were destroyed by boiling for 1 h. In addition, all bacteriocin lysates tested were resistant to trypsin. The use of our bacteriocin production typing method against the presently selected 16 standard indicator strains is recommended for the investigation of any cases of suspected P. mirabilis cross-infections within hospitals.