CsoR Is Essential for Maintaining Copper Homeostasis in Mycobacterium tuberculosis

Abstract

Mycobacterium tuberculosis, a pathogen infecting one third of the world population, faces numerous challenges within the host, including high levels of copper. We have previously shown that M. tuberculosis CsoR is a copper inducible transcriptional regulator. Here we examined the hypothesis that csoR is necessary for maintaining copper homeostasis and surviving under various stress conditions. With an unmarked csoR knockout strain, we were able to characterize the role of csoR in M. tuberculosis as it faced copper and host stress. Growth under high levels of copper demonstrated that M. tuberculosis survives copper stress significantly better in the absence of csoR. Yet under minimal levels of copper, differential expression analysis revealed that the loss of csoR results in a cell wide hypoxia-type stress response with the induction of the DosR regulon. Despite the stress placed on M. tuberculosis by the loss of csoR, survival of the knockout strain was increased compared to wild type during the early chronic stages of mouse infection, suggesting that csoR could play an active role in modulating M. tuberculosis fitness within the host. Overall, analysis of CsoR provided an increased understanding of the M. tuberculosis copper response with implications for other intracellular pathogens harboring CsoR.

Fig 1. Construction of a nonpolar M. tuberculosis ΔcsoR strain.

(A) Diagram showing the copper-sensitive operon drawn to scale. To knock out csoR, the first gene of 4 in its operon, a 2.4kB hyg_R cassette (black bar) was inserted near the 5’ region of csoR. Grey arrowheads indicate primers flanking the genomic region amplified to be used as a probe for Southern blot. The BamHI cut site within the cso used for Southern blot is shown. (B) Southern blot targeting the csoR region after BamHI digestion of wild type (WT) or ΔcsoR::hyg_R genomic DNA (gDNA) demonstrating insertion of the hyg_R cassette. (C) The polar nature of the different ΔcsoR constructs was tested by RT-PCR of H37Rv, ΔcsoR::hyg_R, and ΔcsoR after exposure to 500μM CuCl₂, with gDNA positive and RNA negative controls from the same strains.

Fig 2. Growth kinetics of ΔcsoR under copper stress.

(A) Growth of M. tuberculosis H37Rv (circles) and ΔcsoR (squares) over the course of 15 days in Sauton’s media left untreated (filled), or (B) treated with 500μM CuCl₂ (open). (C) Growth of stationary phase M. tuberculosis H37Rv (circles) and ΔcsoR (squares) inocula over the course of 15 days in Sauton’s media left untreated (filled), or (D) treated with 500μM CuCl₂ (open). The dashed line indicates the limit of detection. Shown are one of two similar biological replicates with error bars representing standard deviation.

Fig 3. Growth of ΔcsoR in the lungs during mouse infection.

(A) Groups of BALB/c mice were infected by aerosol route with either M. tuberculosis H37Rv (filled circles) or ΔcsoR (open squares). Shown are CFU per lung for each individual mouse over the course of 25 weeks across two independent experiments. Asterisks indicate significance with the following P-values: * P < 0.015; ** P < 0.005. (B) Histopathology for mice infected with H37Rv or ΔcsoR at 4 and 8 weeks. Mouse lung sections were stained with H&E and are shown at 40× magnification (scale bar = 500μm). Insets show Ziehl-Neelson stained sections of lung tissue with pink bacilli indicated by black arrows at 1000× magnification (scale bar = 20μm).

Fig 4. Analysis of the ΔcsoR transcriptome.

(A) Counts per million (CPM) of each gene detected in our RNA-Seq study, plotted for M. tuberculosis ΔcsoR versus H37Rv wild type. Highlighted are members of the cso (yellow), members of the RicR regulon (orange), genes responsive to copper, but not part of the cso or RicR regulon (green), members of the DosR regulon (blue), and significantly differentially expressed genes as determined by an FDR ≥ 0.05 not included in the above groups (red). Two parallel, black lines demarcate the region outside of which differential expression values exceed our cutoff of a 2.0 fold difference between strains. Genes not meeting both cutoff values are shown as semi-transparent grey diamonds. Data points show the mean of two biological replicates. (B) Overlap of the induced csoR regulon (red) with 500 CuCl₂ inducible genes (green) and genes induced under the control of dosR (blue). Of the 152 genes induced in the ΔcsoR strain compared to H37Rv wild type, only 4 overlapped with the 24 genes induced in H37Rv wild type when exposed to copper stress at 500μM CuCl₂. DosR inducible genes showed substantial overlap with ΔcsoR with 44 out of 48 overlapping. The diagram is area-proportional.

Fig 5. Expression of cso genes during copper stress in the absence of csoR.

qRT-PCR was used to analyze expression levels of csoR, Rv0968, ctpV, and Rv0970. A schematic of the operon (not drawn to scale) is shown above the graph and the order of the genes in the operon corresponds to the order in which expression levels are graphed for each gene. (A) M. tuberculosis strains H37Rv (black), ΔcsoR (striped), or ΔcsoR::csoR (white) were exposed to CuCl₂ at 50 or 500μM or left untreated (0μM) for 3 hours. Values shown are the mean fold change between each gene and its untreated wild type counterpart after normalization to sigA expression levels. (B) A second comparison of the same samples was done showing the mean fold change between each gene after normalization to sigA expression levels at 50μM (striped) or 500μM (white) and its untreated counterpart (black) from the same strain. Data represent one of two similar biological replicates. Error bars represent the standard error of the mean from three technical replicates. (C) Expression levels of mmcO as determined by qRT-PCR analysis of the same samples–H37Rv (black), ΔcsoR (striped), or ΔcsoR::csoR (white)–left untreated (0μM) or stressed with 50μM or 500μM CuCl₂. Fold change is shown as expression levels of each gene relative to expression levels in untreated wild type culture after normalization to sigA expression levels. Data represent one of two similar biological replicates. Error bars represent the standard error of the mean from two technical replicates.

Fig 6. Potential consequences of ΔcsoR disrupted copper homeostasis.

Overexpression of the cso and the proteins it codes for after deletion of csoR (dark red) may result in excessive levels of copper export by CtpV (red). (A) As a result, pools of copper (orange) for metalloenzyme usage may be depleted, reducing the levels of functional proteins such as cytochrome c oxidase (yellow), as seen in the panel on the right. This could potentially slow the electron transport chain leading to the induction of the DosR (blue) regulon through the redox sensor, DosS (blue) [38]. The wild type strain is shown in the panel on the left for comparison. (B) Alternatively, the increased cytoplasmic export of copper may lead to a buildup of copper in the mycobacterial periplasmic space where MmcO (green) usually assists in copper detoxification. The increased copper stress in this region may be magnified, however, by the down regulation of mmcO in the mutant strain. Free copper can lead to the production of reactive nitrogen species, such as NO [39]. NO can trigger expression of the DosR regulon [25]. The wild type strain is shown in the panel on the left for comparison. CM; cytoplasmic membrane. MM; mycomembrane. Cu; copper. NO; nitric oxide. COX; cytochrome c oxidase. P; phosphoryl group.