The soxR locus of Escherichia coli K12 mediates transcriptional activation of a complex oxidative stress regulon in response to superoxide-generating (redox-cycling) agents. We have cloned the soxR locus, which is positioned near the uvrA gene at 92.2 min on the genetic map, by monitoring complementation of a delta soxR mutation. Subclones from the soxR region in the delta soxR strain simultaneously restored cellular resistance to the redox-cycling agent phenazine methosulfate and inducibility of at least two of the regulon proteins, glucose-6-phosphate dehydrogenase and endonuclease IV, by paraquat, another redox-cycling agent. DNA sequence analysis revealed the presence of two genes involved in activating the soxR regulon. These genes, named soxR and soxS, are arranged divergently with their 5' ends separated by only 85 bp. The predicted 12.9-kDa SoxS protein is related to the AraC family of one-component gene regulators, but corresponds only to the putative DNA-binding regions of these proteins. The 17.1-kDa SoxR protein bears significant homology only to the MerR family of proteins including a predicted DNA-binding helix-turn-helix and a cluster of cysteine residues positioned similarly to those that regulate the activity of MerR in response to Hg2+. This suggests that SoxR could be a metal-binding gene regulator that acts as the intracellular sensor for superoxide. SoxS is evidently the proximal activator of the regulon genes: antibiotic resistance and high-level expression of at least three of the regulon proteins was effected in vivo by the individual expression of SoxS, but not of SoxR, whether or not the cells were exposed to paraquat. These data, together with the recently reported paraquat-inducibility of the soxS gene (Wu, I., and Weiss, B. (1990) J. Bacteriol. 173, 2864-2871), indicate that SoxR and SoxS may constitute a novel type of two-component regulatory system in which the two proteins act sequentially to activate transcription of the various regulon genes in response to superoxide stress.

Escherichia coli responds to the redox stress imposed by superoxide-generating agents such as paraquat by activating the synthesis of as many as 80 polypeptides. Expression of a key group of these inducible proteins is controlled at the transcriptional level by the soxRS locus (the soxRS regulon). A two-stage control system was hypothesized for soxRS, in which an intracellular redox signal would trigger the SoxR protein as a transcriptional activator of the soxS gene and the resulting increased levels of SoxS protein would activate transcription of the various soxRS regulon genes (B. Demple and C.F. Amábile Cuevas, Cell 67:837-839, 1990). We have constructed operon fusions of the E. coli lac genes to the soxS promoter to monitor soxS transcription. Expression from the soxS promoter is strongly inducible by paraquat in a manner strictly dependent on a functional soxR gene. Several other superoxide-generating agents also trigger soxR(+)-dependent soxS expression, and the inductions by paraquat and phenazine methosulfate were dependent on the presence of oxygen. Numerous other oxidative stress agents (H2O2, gamma rays, heat shock, etc.) failed to induce soxS, while aerobic growth of superoxide dismutase-deficient bacteria triggered soxR-dependent soxS expression. These results indicate a specific redox signal for soxS induction. A direct role for SoxR protein in the activation of the soxS gene is indicated by band-shift and DNase I footprinting experiments that demonstrate specific binding of the SoxR protein in cell extracts to the soxS promoter. The mode of SoxR binding to DNA appears to be similar to that of its homolog MerR in that the SoxR footprint spans the -10 to -35 region of the soxS promoter.

No abstract

Biocides, such as herbicides, are routinely tested for toxicity but not for sublethal effects on microbes. Many biocides are known to induce an adaptive multiple-antibiotic resistance phenotype. This can be due to either an increase in the expression of efflux pumps, a reduced synthesis of outer membrane porins, or both. Exposures of Escherichia coli and Salmonella enterica serovar Typhimurium to commercial formulations of three herbicides—dicamba (Kamba), 2,4-dichlorophenoxyacetic acid (2,4-D), and glyphosate (Roundup)—were found to induce a changed response to antibiotics. Killing curves in the presence and absence of sublethal herbicide concentrations showed that the directions and the magnitudes of responses varied by herbicide, antibiotic, and species. When induced, MICs of antibiotics of five different classes changed up to 6-fold. In some cases the MIC increased, and in others it decreased. Herbicide concentrations needed to invoke the maximal response were above current food maximum residue levels but within application levels for all herbicides. Compounds that could cause induction had additive effects in combination. The role of soxS, an inducer of the AcrAB efflux pump, was tested in β-galactosidase assays with soxS-lacZ fusion strains of E. coli. Dicamba was a moderate inducer of the sox regulon. Growth assays with Phe-Arg β-naphtylamide (PAβN), an efflux pump inhibitor, confirmed a significant role of efflux in the increased tolerance of E. coli to chloramphenicol in the presence of dicamba and to kanamycin in the presence of glyphosate. Pathways of exposure with relevance to the health of humans, domestic animals, and critical insects are discussed.

No abstract