Mycobacterium tuberculosis senses host-derived carbon monoxide during macrophage infection

Abstract

Mycobacterium tuberculosis (MTB) expresses a set of genes known as the dormancy regulon in vivo. These genes are expressed in vitro in response to nitric oxide (NO) or hypoxia, conditions used to model MTB persistence in latent infection. Although NO, a macrophage product that inhibits respiration, and hypoxia are likely triggers in vivo, additional cues could activate the dormancy regulon during infection. Here, we show that MTB infection stimulates expression of heme oxygenase (HO-1) by macrophages and that the gaseous product of this enzyme, carbon monoxide (CO), activates expression of the dormancy regulon. Deletion of macrophage HO-1 reduced expression of the dormancy regulon. Furthermore, we show that the MTB DosS/DosT/DosR two-component sensory relay system is required for the response to CO. Together, these findings demonstrate that MTB senses CO during macrophage infection. CO may represent a general cue used by pathogens to sense and adapt to the host environment.

Fig. 1. Heme oxygenase is induced in macrophages and mice by MTB infection

A. Naïve bone marrow derived macrophages were infected with MTB, and global macrophage gene expression determined by microarray. Each time point (Cy5) was compared to a pooled reference (Cy3) and normalized to time zero. Shown is the average response of mouse HO-1 from four independent experiments. Scale bar indicates the mean of the log ratio (635/532). One-way ANOVA resulted in a P value of <0.0001, and comparison of time 0 with 4, 8, 16 and 24 hours yielded P values <0.05 for all comparisons. B. Naïve macrophages were either mock treated (-) or infected with MTB (+) and HO-1 accumulation determined by western blotting with polyclonal anti-HO-1 antibodies and anti-GAPDH antibody as loading control. C. BALB/c mice were infected with MTB by IV injection, and organs harvested after 10 days. Paraffin sections were incubated with rabbit anti-HO-1 antibodies followed by biotinylated goat anti-rabbit antibody and DAB staining, which stains positive cells brown. Cells were counterstained with hematoxylin. Arrows indicate Kupffer cells (3) and alveolar macrophages (4). Scale bar = 200 μm (1, 2) or 50 μm (3, 4).

Fig. 2. Carbon monoxide induces the MTB dormancy regulon in vitro

A. MTB growing in vitro was exposed to 20 (16nM), 200 (160 nM), and 2000 (1.6 μm) ppm of CO gas in the headspace for 48 hours, RNA harvested and gene expression determined by microarray analysis. CO-treated MTB samples were labeled with Cy5 (red) and from untreated MTB samples with Cy3 (green). Green and red spots indicate a gene whose transcription was repressed or induced at least 2-fold relative to untreated samples in at least 2 arrays. Scale bar indicates the mean of the log ratio (635/532). Two representative experiments of more than 4 similar experiments. B. Quantitative PCR (qPCR) analysis of two MTB transcripts, fdxA (blue) and hspX (red) from MTB grown in the presence of CO. Each sample was normalized to 16S RNA as an internal control. Fold induction (±SD) was calculated by dividing all samples by the untreated sample (CO 0 ppm = 1 Fold). Representative of three similar experiments. One-way ANOVA resulted in a P value of <0.0001. * p <0.05, ◊ p <0.01, and † p <0.001 compared to time zero.

Fig. 3. The MTB DosS/T/R signal transduction system transmits the CO signal

A. MTB wild type, ΔdosR and ΔdosR compdosR were treated with DETA-NO (100 μM) or CO (20,000 ppm, 16 μM) for 60 minutes. Shown are the genes of the dormancy regulon only. Scale bar indicates the mean of the log ratio (635/532). Representative of two similar experiments. B. MTB wild type, ΔdosS, ΔdosS comp dosS, ΔdosT, ΔdosT comp dosT, ΔdosS/ΔdosT and ΔdosS/ΔdosT comp dosT were treated with DETA-NO (100 μM) or CO (20,000 ppm, 16 μM) for 60 minutes. Shown are the genes of the dormancy regulon only. Scale bar indicates the mean of the log ratio (635/532). Representative of two similar experiments. C, D. Response of wild type, ΔdosS, and ΔdosT mutants to either DETA-NO (C) or CO (D) for 3 hours at the indicated doses as determined by qPCR for fdxA. Each sample was normalized to 16S RNA as an internal control. Fold induction (±SD) was calculated by dividing each experimental sample by the untreated sample for its respective genotype. One-way ANOVA resulted in a P value of <0.0001 for both C and D. * p <0.001 versus MTB wild type at each CO concentration. Representative of two similar experiments.

Fig. 4. Absence of macrophage heme oxygenase reduces MTB dormancy gene induction

A. Bone marrow derived macrophages from wild type, HO1^-/-, or NOS2^-/- mice were infected with MTB. Twenty-four hours after infection, macrophages were lysed, bacterial RNA isolated and amplified antisense RNA generated. Samples were labeled with Cy5 and a pooled reference of all samples was labeled with Cy3 and hybridized against MTB microarrays. All samples were normalized by the MTB arrays from WT macrophages. Scale bar indicates the mean of the log ratio (635/532). B. aRNA prepared as described above was subjected to qPCR for three MTB genes of the dormancy regulon: Rv2630, fdxA, and hspX. Genes were normalized to 16S RNA as an internal control. Included for comparison is the gene expression of MTB growing in vitro. One of two similar experiments. One way ANOVA gave P values of <0.0001 for all three genes. The asterisk (*) indicates p<0.001 compared to WT macrophages. C. Bone marrow macrophages were infected as described above, and HO-1 detected by western blotting after 24h of infection.