The DosR regulon of M. tuberculosis and antibacterial tolerance

Abstract

Adaptation of Mycobacterium tuberculosis to an anaerobic dormant state that is tolerant to several antibacterials is mediated largely by a set of highly expressed genes controlled by DosR. A DosR mutant was constructed to investigate whether the DosR regulon is involved in antibacterial tolerance. We demonstrate that induction of the regulon is not required for drug tolerance either in vivo during a mouse infection or in vitro during anaerobic dormancy. Thus, drug tolerance observed in these models is due to other mechanisms such as the bacilli simply being in a non-replicating or low metabolic state. Our data also demonstrate that the DosR regulon is not essential for virulence during chronic murine infection. However, decreased lung pathology was observed in the DosR mutant. We also show that the DosR regulon genes are more highly conserved in environmental mycobacteria, than in pathogenic mycobacteria lacking a latent phase or environmental reservoir. It is possible that the DosR regulon could contribute to drug tolerance in human infections; however, it is not the only mechanism and not the primary mechanism for tolerance during a mouse infection. These data suggest that the regulon evolved not for pathogenesis or drug tolerance but for adaptation to anaerobic conditions in the environment and has been adapted by M. tuberculosis for survival during latent infection.

Figure 1

The DorKO mutant (KO) failed to induce two genes of the DosR regulon in response to nitric oxide as measured by qrt-PCR compared to H37Rv (Rv) and the DorCO (CO) strains. White bars indicate induction of Rv2626c and black bars induction of Rv1738. sigA transcripts were used to normalize RNA levels between samples.

Figure 2

Survival of the DorKO mutant during murine infection. C57BL/6 mice were inoculated with a low-dose (50–100 bacilli) aerosol of M. tuberculosis strains H37Rv (circles), DorKO (squares), or DorCO (triangles). The chronic (dashed lines) and Cornell model (solid lines) groups ran parallel for the initial 4 weeks of infection, at which point mice of the Cornell model groups were treated with the antibacterials INH and RIF (administered in drinking water). After 12 weeks, antimicrobial therapy was discontinued and four weeks after cessation of treatment, immune suppression was initiated by administration of dexamethasone for four weeks. CFU counts were determined for (A) the lung or (B) the spleen. For the earlier time points every group consisted of five mice, while for later time points groups consisted of 8–10 mice.

Figure 3

Survival of H37Rv, the DorKO, or the DorCO during the Wayne model of dormancy after antibacterial challenge. H37Rv (white bars), DorKO (horizontal stripes), or DorCO (vertical stripes) were challenged with H₂O, Isoniazid (INH), Rifampin (RIF), or Metronidazole (MET) on (a) day four, (b) day 12, or (c) day 20 into the model. CFU were determined by plating onto DTA plates five days post addition of drug.

Figure 4

A graphical representation of DosR regulon conservation in Mycobacteriaceae and beyond. Rows represent M. tuberculosis ORFs and columns represent sample taxa, organized by phylogenetic relationship, where known[51]. Represented taxa included M. avium (A), M. paratuberculosis (P), M. leprae (L), M. bovis (B), M. tuberculosis (T), M. ulcerans (U), M. smegmatis (S), M. gilvum (G), M. vanbaalenii (V), Mycobacterium sp. KMS (K), M. marinum (M), Mycobacterium sp. JLS (J), Escherichia coli (Ec) and Bacillus subtilis (Bs). Homologs of M. tuberculosis genes were detected using tblastn to search each strain’s genome. Genomes containing a top hit greater than e⁻⁵ are shaded grey, and greater than e⁻¹⁰ are striped to indicate the presence of one or more homologous sequences. The ORF identified by the top hit in each species was then used to blastp the M. tuberculosis proteome. If it identified the initial M. tuberculosis gene, this “reciprocal blast pair” was classified as a probable ortholog pair and the corresponding cell colored black.