Mycobacterium tuberculosis response to cholesterol is integrated with environmental pH and potassium levels via a lipid metabolism regulator

Abstract

Successful colonization of the host requires Mycobacterium tuberculosis (Mtb) to sense and respond coordinately to disparate environmental cues during infection and adapt its physiology. However, how Mtb response to environmental cues and the availability of key carbon sources may be integrated is poorly understood. Here, by exploiting a reporter-based genetic screen, we have unexpectedly found that overexpression of transcription factors involved in Mtb lipid metabolism altered the dampening effect of low environmental potassium concentrations ([K+]) on the pH response of Mtb. Cholesterol is a major carbon source for Mtb during infection, and transcriptional analyses revealed that Mtb response to acidic pH was augmented in the presence of cholesterol and vice versa. Strikingly, deletion of the putative lipid regulator mce3R had little effect on Mtb transcriptional response to acidic pH or cholesterol individually, but resulted specifically in loss of cholesterol response augmentation in the simultaneous presence of acidic pH. Similarly, while mce3R deletion had little effect on Mtb response to low environmental [K+] alone, augmentation of the low [K+] response by the simultaneous presence of cholesterol was lost in the mutant. Finally, a mce3R deletion mutant was attenuated for growth in foamy macrophages and for colonization in a murine infection model that recapitulates caseous necrotic lesions and the presence of foamy macrophages. These findings reveal the critical coordination between Mtb response to environmental cues and cholesterol, a vital carbon source, and establishes Mce3R as a transcription factor that crucially serves to integrate these signals.

Fig 1. Environmental K⁺ levels modulate Mtb transcriptional response to acidic pH.

(A and B) rv2390c’::GFP reporter response to acidic pH is dampened in the presence of low [K⁺]. Mtb(rv2390c’::GFP) was grown in 7H9, pH 7.0 or pH 5.85 media ([K⁺] = 7.35 mM), or in K⁺-free 7H9, pH 5.85 medium, supplemented with 0.1 mM K⁺. Samples were taken at indicated time points and either fixed for analysis of GFP induction by flow cytometry (A) or OD₆₀₀ measured (B). Data are shown as means ± SEM from 3 experiments. p-values were obtained with an unpaired t-test with Welch’s correction and Holm-Sidak multiple comparisons, and compare the 7H9, pH 5.85 to the 0.1 mM K⁺, pH 5.85 condition. N.S. no significant, ** p<0.01. (C) Dampening of Mtb response to acidic pH by low K⁺ is concentration dependent. Mtb(rv2390c’::GFP) was grown in 7H9, pH 5.85 medium (“control”), or in K⁺-free 7H9, pH 5.85 medium, supplemented with indicated amounts of K⁺. Samples were taken 9 days post-assay start, fixed, and GFP induction analyzed by flow cytometry. Data are shown as means ± SEM from 3 experiments. (D) Mtb response to low [K⁺] is not affected by environmental pH. Mtb(kdpF’::GFP) was grown in 7H9, pH 7.0 or pH 5.7, or in K⁺-free 7H9, pH 7 or pH 5.7 media. Samples were taken at indicated time points, fixed, and GFP induction analyzed by flow cytometry. Data are shown as means ± SEM from 3–4 experiments. For all flow cytometry assays, mean GFP fluorescence/Mtb was obtained from measurement of 10,000 Mtb cells in each experimental run. The numerical data underlying the graphs shown in this figure are provided in S1 Data.

Fig 2. A rv2390c’::luciferase reporter transcription factor overexpression screen identifies lipid metabolism regulators as modulators of the interplay between environmental [K⁺] and Mtb pH response.

(A and B) Lipid metabolism regulator hits from reporter-based, inducible transcription factor (TF) overexpression screen. A library of inducible TF overexpression plasmids (P₁’::TF-FLAG-tetON) in the background of a Mtb(rv2390c’::luciferase) strain was screened for their response to acidic pH in the presence of low [K⁺]. TF overexpression was induced by adding 200 ng/ml of anhydrotetracycline (ATC) 1 day before Mtb was exposed to K⁺-free 7H9, pH 7 medium supplemented with 0.05 mM K⁺, or K⁺-free 7H9, pH 5.85 media supplemented with 1.6 mM, 0.1 mM, or 0.05 mM K⁺. 9 days post-exposure (continuous ATC presence), light output (relative light units, RLU) and OD₆₀₀ were measured. Fold induction compares RLU/OD₆₀₀ in each condition to RLU/OD₆₀₀ in the control 0.05 mM K⁺ 7H9, pH 7 condition. (A) shows results of the lipid metabolism regulator hits, together with an empty vector plasmid control. (B) shows validation of the screen hits in (A), with each hit TF compared to its uninduced control (ethanol, “EtOH”, as a carrier control). In (B), data are shown as means ± SEM from three experiments, and p-values were obtained with an unpaired t-test with Welch’s correction. N.S. not significant, * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001. (C) Mce3R modulates the interplay between environmental [K⁺] and Mtb pH response. Log-phase WT, Δmce3R and mce3R* (complemented mutant) were exposed for 4 hours to K⁺-free 7H9, pH 7 medium supplemented with 0.05 mM K⁺ or K⁺-free 7H9, pH 5.85 media supplemented with 1.6 mM or 0.05 mM K⁺, before RNA was extracted for qRT-PCR analysis. Fold change is as compared to the 0.05 mM K⁺ 7H9, pH 7 condition. sigA was used as the control gene, and data are shown as means ± SEM from 3 experiments. p-values were obtained with an unpaired t-test with Welch’s correction and Holm-Sidak multiple comparisons. N.S. not significant, * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001. The numerical data underlying the graphs shown in this figure are provided in S1 Data.

Fig 3. Mtb transcriptional response to low [K⁺] and cholesterol are linked.

(A) Low [K⁺] dampens Mtb response to cholesterol. Log-phase Mtb was exposed for 4 hours to (i) 7H9, pH 7, (ii) cholesterol, pH 7, or (iii) K⁺-free cholesterol, pH 7 media, before RNA was extracted for qRT-PCR analysis. [K⁺] = 7.35 mM in both conditions (i) and (ii). (B) Cholesterol augments Mtb response to low [K⁺]. Log-phase Mtb was exposed for 4 hours to (i) 7H9, pH 7, (ii) K⁺-free 7H9, pH 7, or (iii) K⁺-free cholesterol, pH 7 media, before RNA was extracted for qRT-PCR analysis. Fold change is as compared to the 7H9, pH 7 condition in all cases. sigA was used as the control gene, and data are shown as means ± SEM from 3 experiments. p-values were obtained with an unpaired t-test with Welch’s correction. *** p<0.001, **** p<0.0001. The numerical data underlying the graphs shown in this figure are provided in S1 Data.

Fig 4. Mtb transcriptional response to acidic pH and cholesterol are linked globally.

(A and D) Global changes in Mtb response to cholesterol and acidic pH in the simultaneous presence of both signals. Log-phase Mtb was exposed for 4 hours to (i) 7H9, pH 7, (ii) 7H9, pH 5.7, (iii) cholesterol, pH 7, or (iv) cholesterol, pH 5.7 media, before RNA was extracted for RNAseq analysis. Log₂-fold change compares gene expression in each indicated condition to the 7H9, pH 7 condition. Genes marked in red had a log₂-fold change ≥ 0.6 between the two conditions compared (p<0.05, FDR<0.01 in both sets, with log₂-fold change ≥1 in the single condition set). (B and C) Effect of acidic pH on Mtb response to cholesterol is concentration dependent. Log-phase Mtb was exposed to the indicated conditions, along with the control 7H9, pH 7 condition, for 4 hours, before RNA was extracted for qRT-PCR analysis. (E and F) Effect of cholesterol on Mtb response to acidic pH is concentration dependent. For all qRT-PCR data, fold change is as compared to the 7H9, pH 7 condition, and data are shown as means ± SEM from 3 experiments. The numerical data underlying the graphs shown in this figure are provided in S1 Data.

Fig 5. Mce3R regulates Mtb response to cholesterol only in the context of acidic pH.

Log-phase WT, Δmce3R and mce3R* (complemented strain) Mtb were exposed for 4 hours to (i) 7H9, pH 7, (ii) 7H9, pH 5.7, (iii) cholesterol, pH 7, or (iv) cholesterol, pH 5.7 media, before RNA was extracted for RNAseq (A-C) or qRT-PCR (D-F). For RNAseq data, log₂-fold change compares gene expression in each indicated condition to the 7H9, pH 7 condition. Genes marked in red had a log₂-fold change ≥ 0.6 between Δmce3R and WT strains in the indicated condition (p<0.05, FDR<0.01 in both sets, with log₂-fold change ≥1 in WT). For qRT-PCR data, fold change is as compared to the 7H9, pH 7 condition. Data are shown as means ± SEM from 3 experiments. p-values were obtained with an unpaired t-test with Welch’s correction and Holm-Sidak multiple comparisons. N.S. not significant, ** p<0.01, *** p<0.001, **** p<0.0001. The numerical data underlying the graphs shown in this figure are provided in S1 Data.

Fig 6. Mce3R deletion dampens Mtb response to K⁺ in the presence of cholesterol.

Log-phase WT, Δmce3R, and mce3R* (complemented strain) Mtb were exposed for 4 hours to (A) K⁺-free 7H9 or cholesterol media at pH 7, or (B) cholesterol or K⁺-free cholesterol media at pH 7, along with 7H9, pH 7 as the control condition. ([K⁺] = 7.35 mM in cholesterol, pH 7 medium.) Fold change is as compared to the 7H9, pH 7 condition in all cases. sigA was used as the control gene, and data are shown as means ± SEM from 3 experiments. For clarity of comparison, the cholesterol medium, pH 7 data in Fig 5D are shown in Fig 6B again. p-values were obtained with an unpaired t-test with Welch’s correction and Holm-Sidak multiple comparisons. N.S. not significant, * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001. The numerical data underlying the graphs shown in this figure are provided in S1 Data.

Fig 7. A Δmce3R Mtb mutant is attenuated for host colonization in the presence of lipids.

(A) Δmce3R Mtb is attenuated for macrophage colonization. Murine bone marrow-derived macrophages untreated or pre-treated with oleate for 24 hours to induce foamy macrophages were infected with WT, Δmce3R, or mce3R* (complemented mutant) Mtb, and colony forming units (CFUs) tracked over time. Data are shown as means ± SD from 3 wells, representative of 3 independent experiments. p-values were obtained with an unpaired t-test with Welch’s correction, comparing Δmce3R to WT Mtb in untreated macrophages (*) or foamy macrophages (#). */# p<0.05, ** p<0.01. (B) Mtb is present in foamy macrophages 6 weeks post-infection in C3HeB/FeJ mice. C3HeB/FeJ mice were infected with Mtb constitutively expressing mCherry, and animals sacrificed at 2 or 6 weeks post-infection. Lungs were fixed and processed for confocal microscopy imaging. Images shown are z-slices from reconstructed 3D images. Nuclei are shown in grayscale (DAPI), all bacteria are marked in red (mCherry), f-actin is shown in blue (phalloidin), and lipid droplets are shown in green (Bodipy 493/503). Scale bar 10 μm. (C) Δmce3R Mtb is attenuated for colonization in a murine infection model. C3HeB/FeJ mice were infected with WT, Δmce3R, or mce3R* Mtb, and lung homogenates plated for CFUs 2 or 6 weeks post-infection. p-values were obtained with a Mann-Whitney statistical test. N.S. not significant, **p<0.01. The numerical data underlying the graphs shown in this figure are provided in S1 Data.

Fig 8. Summary of the relationships between Mtb response to acidic pH, cholesterol, and low [K⁺], and the role of Mce3R.

A summary model and table of the interplay between acidic pH, cholesterol, and low [K⁺] is shown in (A), with the effect of mce3R deletion on the various relationships shown in (B). Shaded boxes in (A) are the single conditions. As there was no effect on the response of Mtb to low [K⁺] in the presence of acidic pH, the effect of mce3R deletion on this relationship was not tested here (shaded box in B).