The Molecular Basis Of Antibiotic Resistance

Diabetes, Digestive, Kidney Diseases

Investigators

Linked publications & trials

Abstract

Derepression of the mar operon in E. coli leads to low level resistance to a broad spectrum of antibiotics, superoxide generating compounds and organic solvents. This phenotype is mediated through the synthesis of a transcriptional activator, MarA. MarA functions to activate a number of genes in E. coli, the mar regulon. Most transcriptional activators in prokaryotic organisms are believed to function by site-specifically binding to DNA and 'recruiting' RNA polymerase (RNAP) to the promoter. We have shown that MarA does not seem to follow this rule. Rather, MarA appears to form complexes with polymerase before interacting with DNA. In direct support of this we demonstrated that a complex between the two is formed in solution in the absence of DNA. Other experimnents involving bacterial physiology in vivo also support the hypothesis. Specifically, the over production of a mutant MarA that binds DNA but not RNAP does not inhibit the stimulation of transcription by low levels of the wild-type MarA. We have now identified the amino acids on the surface of MarA involved in this interaction and are examining which of the amino acids of the alpha subunit of RNAP they interact with. In a separate study we have been examining potential promoter sites in E. coli. It is well established that transcription begins at specific sites upstream of ORFs and that the signals for RNAP recognition are highly degenerate. So degenerate, that they are virtually impossible to identify unless they are immediately upstream of ORFs. For this reason, it has only recently been possible to identify small RNA molecules synthesized from inter-ORF regions not by developing new algorithms for RNAP binding, but rather by recognizing that these inter-ORF small RNAs are highly preserved among the prokaryotes. Based on the degeneracy of the known RNAP binding signals we predicted that there should be significant matches to the -35 and -10 sequences (appropriately separated by 16 to 18 nucleotides) at random around the genome spaced approximately once every 150 bp. We therefore digested E. coli DNA with two 4 bp restriction enzymes into fragments of varying sizes and assayed the fractions for their ability to be retarded on gel electrophoresis by RNAP. As predicted it appeared that approximately 2/3rds of the DNA was retarded at a fragment size of ~150 bp (expected for 1 binding site by the Poisson distribution). We next cloned 105 of these fragments in a plasmid with a promotorless lacZ reporter gene, sequenced each, and assayed the fusions for beta-galactosidase activity. About 1/3rd of the sequences had no more activity than the control plasmid, but the remainder had minimal to significant activity. The sequences were randomly distributed around the chromosome mostly in, but occasionally between, ORFs and oriented in both directions with respect to the ORF. In collaboration with Dr. Gisella Storz and Mitsuoki Kawano we have been analyzing these sequences for chromosomal as well as plasmid expression in vivo. To date none of the sequences appear to function from the chromosome (although they do from the plasmid) for reasons that remain to be elucidated.