Members of the AraC protein family regulate a broad range of activities important in the pathogenicity of E. coli: bacterial stress response, invasion of intestinal mucosa and toxin release. The proteins MarA and SoxS (approximately 45% identical to MarA), members of an AraC sub- family, are transcriptional activators that mediate resistance to superoxides, organic solvents and a broad range of antibiotics in a number of gram-negative bacteria. The binding site or marbox to which MarA and SoxS bind is located in the region immediately upstream of the binding signals (-10 and -35 sites) for RNA polymerase (RNP). The marbox is a highly degenerate sequence and although we had previously solved the 3-dimensional structure of MarA bound to an oligonucleotide corresponding in sequence to the marbox of the mar promoter (S. Rhee, D. Davies and J. L. Rosner), the precise nature of the consensus had not been defined. This, in part because the X-ray crystallographic analysis demonstrated that the principle interactions between protein and DNA are van der Waals interactions, often to the phosphate backbone, rather than hydrogen bonds to specific bases. Detailed analysis of a number of promoters of the regulon have demonstrated: i) that the minimal binding sequence is 20 bp in length; ii) that when the binding site (which is asymmetric) overlaps the -35 signal for RNP it orients the MarA in one direction (so that the amino terminal portion of MarA faces downstream) and that the separation between the -10 signal and the binding sequence is 18 or 19 bp; iii) that with one exception, when the binding site lies upstream of the -35 signal, it orients MarA in the opposite direction (so that the amino terminal portion of MarA faces upstream) and that the separation between the ?35 signal and the binding site is 15 or 16 bp, or one additional turn of the double helix, i.e. 26 or 27 bp; and iv) that in the exceptional case, the binding site lies upstream of the -35 signal by 7 bp and orients MarA in the same orientation as those binding sites that overlap the -35 signal. In addition, we have shown that in all cases, the binding sites are located such that MarA is bound to the same surface of the DNA, consistent with the hypothesis that interaction between MarA and RNP is necessary for activation. In collaboration with J. L. Rosner and W. Gillette we have demonstrated that while MarA and SoxS bind to and activate the same family of genes (the mar/sox regulon) induction by MarA and SoxS are not equivalent: although maximal induction by MarA and SoxS result in similar levels of antibiotic resistance, SoxS is far more effective than MarA in inducing resistance to superoxides. We have demonstrated that this differential activation of resistances reflects a differential induction of members of the regulon. The ability of the promoters to distinguish between MarA and SoxS we have termed discrimination and have shown that the crucial element in discrimination is the sequence of the marbox. Furthermore, we have shown that discrimination is the result of the differential binding of MarA and SoxS to the different members of the marbox family. Preliminary results suggest: i) that for promoters in which the marbox overlaps the -35 signal, MarA interacts with the sigma 70 subunit of RNP; and ii) that the activation of a promoter that is absolutely dependent upon sigma 38 for activity behaves in a manner similar to sigma 70-dependent promoters. In collaboration with B. Dangi, J. L. Louis and A. M. Gronenborn, we are studying the interaction of MarA with a number of different marbox sequences. Preliminary results suggest that hydrogen bonds are not formed between protein and DNA in solution. - activation, repression, antibiotic resistance, control, E. coli, regulation, stress response
Showing the most recent 10 out of 14 publications
