Selection and fitness in bacteriocin-producing bacteria

Abstract

Bacteriocins are proteinaceous anticompetitor molecules produced by bacteria against closely related species. A number of theoretical models have been used to explain experimental data that indicate high polymorphisms among bacteriocins and a frequency-dependent nature of selection for bacteriocin-producing strains. The majority of these experimental data were, however, obtained from investigations into the colicin group of bacteriocins produced by Gram-negative bacteria. The conclusions drawn from these models have been extrapolated to other bacteriocins and allelopathic compounds in general. Examination of more recent experimental investigations into the bacteriocins of Gram-positive bacteria indicate a lower degree of polymorphism and a less frequency-dependent mode of selection among these strains than among the colicin-producing strains. Here we examine these contradictions in the light of the assumptions and conclusions of the theoretical models and reported data. We show that fitness costs (as indicated by decreased relative maximum growth rate) associated with bacteriocin production may be much lower in many cases than is assumed in the present models. A lower fitness cost associated with bacteriocin production adequately explains the newer data from Gram-positive bacteria cited here, and indicates that extrapolation of existing models to all bacteriocins and other allelopathic compounds is not appropriate.