The role of SOS boxes in enteric bacteriocin regulation

Abstract

Bacteriocins are a large and functionally diverse family of toxins found in all major lineages of Bacteria. Colicins, those bacteriocins produced by Escherichia coli, serve as a model system for investigations of bacteriocin structure-function relationships, genetic organization, and their ecological role and evolutionary history. Colicin expression is often dependent on host regulatory pathways (such as the SOS system), is usually confined to times of stress, and results in death of the producing cells. This study investigates the role of the SOS system in mediating this unique form of toxin expression. A comparison of all the sequenced enteric bacteriocin promoters reveals that over 75 % are regulated by dual, overlapping SOS boxes, which serve to bind two LexA repressor proteins. Furthermore, a highly conserved poly-A motif is present in both of the binding sites examined, indicating enhanced affinity of the LexA protein for the binding site. The use of gene expression analysis and deletion mutations further demonstrates that these unique LexA cooperative binding regions result in a fine tuning of bacteriocin production, limiting it to times of stress. These results suggest that the evolution of dual SOS boxes elegantly accomplishes the task of increasing the amount of toxin produced by a cell while decreasing the rate of uninduced production, effectively reducing the cost of colicin production. This hypothesis may explain why such a promoter motif is present at such high frequencies in natural populations of bacteriocin-producing enteric bacteria.

Fig. 1

Bayesian tree indicating the structural relationship among the 300 bp promoters located just upstream of the colicin genes. Indicated are the NJ/MP/PP values obtained using the different tree-building algorithms (NJ, 10 000 bootstrap; MP, 1000 bootstrap). One major clade from pColN to pCol10 is recognized and separated from all other promoter structures. Nodes indicated with a clover-leaf symbol form one node using the NJ and MP algorithms. Similarly, nodes indicated with ‡ are one node using the NJ and MP algorithms. See Table 3.

Fig. 2

Inter-bacteriocin consensus for 22 bacteriocin LexA binding sites. Sequences were produced using the WebLogo service at http://weblogo.berkeley.edu. Overlapping LexA binding sites are shown by the solid and dotted lines below the figure.

Fig. 3

Light emission of ce1a (strain pDEW-E1/7) and ciba (strain pDEW-Ib/18) promoter fusions in response to (a) various concentrations of MitC measured at highest response ratio 2.5 h following induction, (b) no induction, and (c) 2 mg MitC l⁻¹ over 5.5 h. Each point is the mean of at least three replicates.

Fig. 4

Light emission of ce1a (strain pDEW-E1/7) and recA (strain pDEW238) wild-type and mutants (strains pDEW-E1/3738185, pDEW-E1/3940222 and pDEW-E1/3740345) constructs in response to (a) 0.5 mg MitC l⁻¹ over 2.5 h, (b) no induction, and (c) various concentrations of MitC measured at the highest response ratio 2.5 h following induction. Each point is the mean of at least three replicates.