Nonheritable Antibiotic Resistance

Diabetes, Digestive, Kidney Diseases

Investigators

Linked publications, trials & patents

Abstract

The MarA, SoxS and Rob transcriptional activators of Escherichia coli regulate the expression of 17 promoters known as the marA/soxS/rob regulon. Overexpression of the activators engenders multiple antibiotic resistance, superoxide resistance and organic solvent tolerance. The activators bind to a 20 bp asymmetric sequence that is highly degenerate and which is functional only when in certain configurations relative to the RNA polymerase binding site. Because of this and because we cannot readily identify the RNAP binding sites, it is difficult to identify additional members of the regulon. Recent microarray studies from two laboratories identified 153 genes/promoters as part of the regulon, including 12 promoters previously known to be directly activated. Surprisingly, only 27 of the153 were implicated by both studies and eight of these were previously known regulon promoters. We analyzed the regulation of the remaining 19 genes in this group using both conventional genetic and biochemical approaches and an algorithm which we devised for detecting activator binding sites. We found that only seven of the 19 putative regulon members were in fact regulated by MarA, SoxS or Rob. We estimate that there are only 30-40 directly activated promoters in the regulon. Thus, caution is required in evaluating the results of microarray studies. We previously demonstrated that MarR represses the transcription of the marRAB operon by binding as a dimer to inverted repeats at each of two palindromic DNA sites. The operon is derepressed by salicylate which binds MarR and decreases its affinity for the DNA. We are now examining the translation of MarR which does not appear to have a good Ribosome Binding Site (RBS) or the usual initiator codon AUG. We measured transcription and translation of marR with appropriate lacZ fusions. When inserted just after a GTG codon (corresponding to nucleotides (nt) 28-30 of the mRNA), the translational fusion had 1,000-fold less activity than the transcriptional fusion suggesting that the GTG codon might not initiate translation. Additional in-frame fusions following different codons downstream of the GTG, showed that translation increased dramatically after a GAT codon (nt 43-45) of the mRNA. When the GTG was replaced by a non-initiator codon (GGG) in this nt 45 fusion, activity was severely reduced; when replaced by the preferred ATG initiator codon, activity was appreciably increased. This is good evidence that the GTG codon is the initiator codon for MarR. RNA modeling indicates that when the mar mRNA is 30-42 nt long, it forms stem-loop structures which might hinder the attachment of ribosomes to the GUG codon. When the mRNA is 45 or more nt, a new structure may form which exposes the GUG codon to the ribosomes. When the putative stem-loop structure was disrupted by a 5 bp transversion mutation and lacZ was inserted just after the GTG codon, translation was increased. Thus it appears that MarR translation is limited by the lack of a good RBS, the use of the weaker GTG initiator codon and the folding of the mRNA. Perhaps this maintains MarR at a minimal level in the cell without interfering with the transcription and translation of MarA that occurs upon derepression.