RNase HI Depletion Strongly Potentiates Cell Killing by Rifampicin in Mycobacteria

Abstract

Multidrug-resistant (MDR) tuberculosis (TB) is defined by the resistance of Mycobacterium tuberculosis, the causative organism, to the first-line antibiotics rifampicin and isoniazid. Mitigating or reversing resistance to these drugs offers a means of preserving and extending their use in TB treatment. R-loops are RNA/DNA hybrids that are formed in the genome during transcription, and they can be lethal to the cell if not resolved. RNase HI is an enzyme that removes R-loops, and this activity is essential in M. tuberculosis: knockouts of rnhC, the gene encoding RNase HI, are nonviable. This essentiality makes it a candidate target for the development of new antibiotics. In the model organism Mycolicibacterium smegmatis, RNase HI activity is provided by two enzymes, RnhA and RnhC. We show that the partial depletion of RNase HI activity in M. smegmatis, by knocking out either of the genes encoding RnhA or RnhC, led to the accumulation of R-loops. The sensitivity of the knockout strains to the antibiotics moxifloxacin, streptomycin, and rifampicin was increased, the latter by a striking near 100-fold. We also show that R-loop accumulation accompanies partial transcriptional inhibition, suggesting a mechanistic basis for the synergy between RNase HI depletion and rifampicin. A model of how transcriptional inhibition can potentiate R-loop accumulation is presented. Finally, we identified four small molecules that inhibit recombinant RnhC activity and that also potentiated rifampicin activity in whole-cell assays against M. tuberculosis, supporting an on-target mode of action and providing the first step in developing a new class of antimycobacterial drug.

Keywords: R-loop; RNase HI; antibiotic development; antibiotic resistance; antibiotic synergy; rifampicin.

FIG 1

Twin domain model of transcription and location of enzymes (and their inhibitors) involved in topological modification or R-loop metabolism. RNAP, RNA polymerase I.

FIG 2

R-loop accumulation in wild-type, ΔrnhA, and ΔrnhC strains of M. smegmatis mc²155. (A) Dot-blot analysis of R-loop accumulation. Total nucleic acid from each strain was spotted onto the membrane and detected using the S9.6 RNA/DNA hybrid-specific monoclonal antibody. Control nucleic acid extracts were treated with E. coli RNase HI before spotting. (B) R-loops in the wild-type, ΔrnhA, and ΔrnhC strains, and their respective complemented strains were quantitated using Image Lab (Bio-Rad). Relative amounts are shown normalized to the wild type. (C) Uncut genomic DNA from the wild-type (WT), ΔrnhC (C−), or ΔrnhC-complemented strain (C+) was separated on a 1% agarose gel in duplicate. One set was stained with ethidium bromide as a record, and the other was transferred to a nylon membrane. (D) Genomic DNA from panel C (bottom panel) or applied as a dot blot (top panel) was probed with S9.6 monoclonal antibody to detect RNA/DNA hybrids. The bottom image was flipped horizontally to conform to the lane positions in panel C. (E) Extended exposure of an experimental repeat, showing smear of fragmented RNA:DNA hybrids migrating more quickly than bulk chromosomal DNA. The data shown are representative of the average of three independent experiments, with standard deviations indicated by error bars. The statistical significance of the differences was assessed using Student’s unpaired t test calculated using GraphPad Prism 7. *, P < 0.05.

FIG 3

Expression of gyrase (gyrB) and topoisomerase I (topA) promoters fused to lacZ in wild-type, ΔrnhC, and complemented strains of M. smegmatis mc²155. The relative activities of the β-galactosidase reporters are shown, normalized to the wild type. The data shown are representative of the average of three independent experiments, with standard deviations indicated by error bars. The statistical significance of the differences was assessed using Student’s unpaired t test calculated using GraphPad Prism 7. *, P = 0.02; **, P = 0.002.

FIG 4

R-loop quantitation in wild-type and ΔrnhC strains of M. smegmatis mc²155 after exposure to various concentrations of rifampicin. (A) Dot blot analysis of R-loop formation. (B) Quantification of R-loops observed in the dot blot, normalized to the wild-type strain in the absence of rifampicin. The data shown are representative of the average of three independent experiments, with standard deviations indicated by error bars. The statistical significance of the differences was assessed using Student’s unpaired t test calculated using GraphPad Prism 7. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 5

(A and B) Dose-response curves of rifampicin killing in (A) wild-type and ΔrnhC and (B) wild-type and ΔrnhA strains of M. smegmatis mc²155 and the respective complemented strains. The data shown are representative of the average of three independent experiments, with standard deviations indicated by error bars. When included, vitamin B₁₂ was added to the growth medium at 10 μg mL⁻¹.

FIG 6

(A to D) Dose-response curves of antibiotic killing for (A) moxifloxacin, (B) streptomycin, (C) amsacrine, and (D) isoniazid in wild-type and ΔrnhC strains of M. smegmatis mc²155. The data shown are representative of the average of three independent experiments, with standard deviations indicated by error bars.

FIG 7

(A and B) Chemical structures (A) and dose-response curves (B) for NSC353720 (i), NSC600285 (ii), NSC18806 (iii), and NSC99726 (iv) against recombinant M. tuberculosis RNase HI. The data shown are the average of two independent experiments, with standard deviations indicated by error bars.

FIG 8

Domino effect model of R-loop formation under rifampicin stress. (A) Negatively supercoiled DNA (- - -) opens between a stationary RNAP (such as one in the promoter region) and a mobile RNAP (motion indicated by “((”). This danger zone for R-loop formation is policed by topoisomerase I (blue; shown only once for clarity), which competes with mRNA (red line; only shown twice for clarity) for underwound DNA and by RNase HI (green). Zones between mobile RNAP are more likely to be neutrally wound (+ -). (B) Under rifampicin stress, the promoter-bound RNAP cannot start elongating, the danger zone becomes critically underwound, and an R-loop forms and stalls its cognate RNAP. A new danger zone opens up between this stalled RNAP and its adjacent mobile RNAP. A new R-loop forms, stalling its RNAP, and the cycle continues until all mobile RNAP molecules are stalled by R-loops. Ribosomes are not shown.