Rifamycin inhibition of WT and Rif-resistant Mycobacterium tuberculosis and Escherichia coli RNA polymerases in vitro.

Mycobacterium tuberculosis (MTB) infects over 9 million people globally and claims approximately 2 million lives annually. Rifampin (Rif) is one of the first-line anti-tuberculosis drugs that inhibits transcription by binding to the β subunit (encoded by the rpoB gene) of the prokaryotic RNA polymerase (RNAP). A highly conserved 81 base pair core region among the β subunit of prokaryotes harbors most of the point mutations leading to rifamycin-resistant (RifR) mutations, where the majority of the clinically relevant MTB RifR mutations result from amino acid substitutions of one of the following three amino acids: βAsp435, βHis445, and βSer450 (MTB numbering). In this study, to determine the direct effect of rifamycins on the MTB RNAP, co-overexpression vectors were constructed to co-express the core subunits of wild-type and RifR mutants of MTB RNAP. The three aforementioned amino acids were each mutated to the most prevalent substitution found in the MTB clinical isolates (Asp435Val, His445Tyr, Ser450Leu) in the rpoB gene via site-directed mutagenesis. After purification via two-step column chromatography, the in vitro activity of the wild-type and RifR mutant MTB RNAPs was assessed via rolling circle transcription assay. The apparent IC(50) values for three key rifamycins (rifampin (Rif), rifabutin (Rbn), and rifaximin (Rfx)) were determined and these results indicate that the mutant RNAPs demonstrate approximately 10(3)-fold or greater loss of affinities for rifamycins relative to wild-type MTB RNAP. Along with the MTB RNAPs, rifamycin inhibition of the Escherichia coli RNAP counterparts was also assessed. Previously, it has been reported that Gram-positive bacteria (particularly mycobacteria) are more sensitive to rifamycins than Gram-negative bacteria. Under our experimental conditions, the rifamycin IC(50)s for wild-type and RifR mutants of MTB and E. coli RNAPs (wild-type and corresponding mutants) were very similar; therefore, the difference in sensitivity toward rifamycins does not reside in the RNAP itself. The correlation between the sensitivity of rifamycins and permeability into cells was evaluated using the wild-type E. coli strains (TG2 and DH5α) and a mutant E. coli strain with efflux pump defects (EC2880, tolC(-)/imp(-)). The MICs were drastically lower in the EC2880 strain, consistent with previous reports that the differential sensitivity of MTB and E. coli to rifamycins is not related to the RNAP, but rather has to do with efflux pumps in E. coli. Future work will focus on the elucidation of the molecular interaction of these MTB RifR mutants with rifamycins to provide insight to the design of novel rifamycins.