Chromosomal multidrug resistance (MDR) in bacteria is a serious clinical problem. Our studies have shown that Escherichia coli becomes resistant to a variety of antibiotics, organic solvents and superoxides when the activities of any of three paralogous, but differently regulated, transcriptional activators, MarA, SoxS and Rob, are increased. These activators bind a sequence called the marbox which lies upstream of the promoters of a set of about 40 chromosomal genes called the marA/soxS/rob regulon. tolC and acrAB are regulon genes that have critical functions in multiple antibiotic resistance since their protein products constitute the most important MDR pump in E. coli. The major goals of this project are to understand the regulation of these activators, the mechanisms whereby they activate the regulon promoters, and the mechanisms whereby the MDR is generated. Upregulation of the transcriptional activators, MarA, SoxS and Rob can be effected by treating the cells with certain chemicals. Phenolic compounds derepress the marRAB operon;superoxides activate SoxR which in turn activates SoxS;and bile salts and other compounds activate the Rob protein directly. Thus, the upregulation of these activators can indicate the presence of such substances in the environment or in the cell. We have known for some time that MarA and SoxS are similarly potent in activating antibiotic-resistance (by activating the expression of the micF, acrAB and tolC genes which control influx and efflux of antibiotics and other xenobiotics). However, SoxS is 2- to 23-fold more potent than MarA as an activator of superoxide-resistance genes such as zwf and fpr. A major increase in our understanding of this was obtained by comparing MarA, MarA mutants, and SoxS for their abilities to activate a variety of regulon promoters and to bind their cognate marbox binding sites (ref 2). Replacement of the MarA glutamic acid (E) residue 89 with alanine (A) greatly increased the marbox binding and activation of many class I promoters (where the marbox is upstream of the promoter). Like cells constitutive for SoxS, cells expressing MarA with the E89A mutation were more resistant to superoxides than those harboring WT MarA. The activities of several other E89 substitutions ranked as follows: E89A >E89G >E89V >WT >E89D. Increased binding and activation occurred only at class I promoters when the 12th base of the promoter's marbox (a position at which there is no known interaction between the marbox and MarA) was not a T residue. Furthermore, WT MarA binding to a synthetic marbox in vitro was enhanced when the phosphate group between positions 12 and 13 was eliminated on one strand. The results demonstrate that relatively minor changes in a single amino acid side chain (e.g., alanine to valine or glutamic acid to aspartic acid) can strongly influence activity despite any evidence that the side chain is involved in positive interactions with either DNA or RNA polymerase. We have presented a model which attributes the differences in binding and activation to the steric interference between the β- and γ-carbons of the amino acid at position 89 and the phosphate group between positions 12 and 13. This differential activation of genes by MarA and SoxS makes sense in terms of the agents which induce their expression. The presence of superoxides which induces SoxS expression results in efficient activation of superoxide defense genes. In contrast, defense against phenolics does not appear to require the superoxide defense genes so their expression is limited when MarA is induced by phenolics. Previously, we found that E. coli tolC mutants, which do not have the TolC outer membrane channel, have elevated levels of transcription of marRAB and soxS and have elevated activity of Rob protein. Since TolC is a vital component of eight known efflux pumps in E. coli and plays important roles in ridding bacteria of multiple antibiotics, bile salts, organic solvents and other xenobiotics, we concluded (1) that in the absence of TolC, intracellular metabolic waste products accumulate and trigger the upregulation of the activators and (2) that TolC is normally involved in the efflux of cellular metabolites and not merely of xenobiotics. Recently, we have subjected tolC mutants and mutants defective in all the nine known TolC-dependent efflux pumps to microarray assays. Our analysis showed increased expression of the spy gene in both cases. This gene has recently been shown by others to encode a novel ATP-independent periplasmic chaperone. Using spy::lacZ transcriptional fusions, we have found 10-fold increases in transcription of spy in tolC mutants or in mutants defective for only 4 of theTolC-dependent pumps. Why defective efflux increases Spy chaperone expression remains to be determined. One possibility is that in the absence of efflux certain proteins are not folded correctly unless Spy is present in increased amounts. We have also shown that TolC plays a role in extreme acid survival (ref 1). Enterobacteria such as E. coli have mechanisms that enable them to survive passage through the highly acidic stomach. One major mechanism is expression of glutamate decarboxylase encoded in the gadA and gadB genes. tolC mutants survive pH 2 very poorly compared to wild-type (<1E-5). The TolC-dependent mdtB emrB genes are largely responsible for the poor survival. In the tolC mutant, the transcription of gadA, and the expression of GadA and GadB proteins were severely decreased. Plasmid complementation studies show that overexpression of GadA or GadB or of GadE (an activator of gadA and gadB in addition to other genes) rescues the tolC mutant from extreme acid. This suggests that the absence of TolC (and possibly the TolC-dependent MdtB and EmrB pumps) decreases expression of GadE. Again, this suggests that the antibiotic-resistance pumps serve other, more general, interests of the cell.
