MarA, SoxS and Rob are three homologous but distinct proteins that can transcriptionally activate the 25 or so genes of the mar/sox/rob regulon of Escherichia coli. Such activation results in multiple antibiotic resistance, superoxide resistance and organic solvent tolerance. We have previously described details of the mechanism whereby salicylate derepresses the marRAB operon, resulting in high levels of MarA. We have also characterized the sites to which MarA binds and have presented a high resolution structure of MarA bound to its cognate site. Others have described the regulation of SoxS. A. Here we report that Rob activity in vivo can be increased 10-fold by treating cells with either 2,2'- or 4,4'-dipyridyl (DIP). Normally, there are about 10,000 molecules of Rob per cell but their activity is low. The action of DIP is primarily post-translational as shown with constructs that are missing the normal rob gene's transcriptional and translational signals. Furthermore, the C-terminal domain of Rob is essential for this effect. In the absence of the C- terminal domain, Rob has the normal low level of activity seen with wild-type Rob but this does not increase upon treatment with DIP. NMR studies in vitro show that 4,4'-DIP interacts with, and perturbs the structure of, Rob but does not do so with Rob lacking the C-terminal domain. Since DIP is not a normal constituent of E. coli or its environment, we assume that DIP mimics a natural substance which is yet to be identified. The determination of the crystal structure of a Rob-DIP-DNA complex should help us understand how DIP activates Rob. B. The mechanism of transcriptional activation in bacteria is commonly believed to occur by the activator binding first to the promoter DNA and then to RNA polymerase (RNAP) thereby recruiting RNAP to the DNA. We now have evidence for an alternate mechanism: activator binding first to RNAP and then this complex scanning the DNA for promoters with an appropriately oriented and spaced adjacent activator binding site. We have isolated MarA-RNAP and SoxS-RNAP complexes in vitro in the absence of DNA. Rob complexes with RNAP are less stable, but Rob-DNA complexes are usually more stable, than the corresponding MarA-DNA or SoxS-DNA complexes. This suggests answers for two puzzling questions about MarA/SoxS/Rob activation. (1) How do the small number of activator molecules (estimated at 350 for SoxS and 2,300 for MarA after induction) find the 25 correct DNA binding sites out of the thousands that are not functional? We have found by computer analysis of the E. coli genome and by gel shifts of genomic DNA with activators that there are about 9,000 of the 20 bp sequences of the highly degenerate activator binding sites per chromosome but very few of these are correctly oriented and spaced relative to promoters. We propose that by simultaneously scanning the DNA with both RNAP and activator reading heads, the complex can recognize and bind to correctly configured sites. (2) How do MarA and SoxS compete with Rob for the functional binding sites? We find that MarA and SoxS activation of the regulon is similar whether 10,000 molecules of Rob are present or not. We suggest that the tighter binding of MarA and SoxS to RNAP, and of these complexes to DNA, is responsible for their ability to bind the functional promoters even in the presence of greater numbers of Rob. C. MarA recognizes a very degenerate, asymmetrical DNA consensus sequence. What is responsible for this flexibility? The crystal structure of MarA has shown two helix-turn-helix motifs per monomer that bind the major groove of the target DNA. We have now used various NMR techniques to probe the MarA structure when bound to 4 different promoter DNAs. This has revealed that MarA exists in at least 2 forms; that it does not make certain H-bonds with DNA that were inferred from the X-ray structure; and that a portion of the N-terminus, not visible in the X-ray structure, may help bind the target DNA. Thus, MarA is in a highly dynamic state when interacting with different DNA targets allowing for small but significant side chain or backbone rearrangements to bind the different DNAs.
Showing the most recent 10 out of 19 publications
