Candidate transmission survival genome of Mycobacterium tuberculosis

Abstract

Mycobacterium tuberculosis (Mtb), a leading cause of death from infection, completes its life cycle entirely in humans except for transmission through the air. To begin to understand how Mtb survives aerosolization, we mimicked liquid and atmospheric conditions experienced by Mtb before and after exhalation using a model aerosol fluid (MAF) based on the water-soluble, lipidic, and cellular constituents of necrotic tuberculosis lesions. MAF induced drug tolerance in Mtb, remodeled its transcriptome, and protected Mtb from dying in microdroplets desiccating in air. Yet survival was not passive: Mtb appeared to rely on hundreds of genes to survive conditions associated with transmission. Essential genes subserving proteostasis offered most protection. A large number of conventionally nonessential genes appeared to contribute as well, including genes encoding proteins that resemble antidesiccants. The candidate transmission survival genome of Mtb may offer opportunities to reduce transmission of tuberculosis.

Keywords: Mycobacterium; aerosol; biophysics; transmission; tuberculosis.

Fig. 1.

Fluids and atmospheres modeling three stages of transmission. (A) Schematic. Part of the cartoon is adapted from Manna and Bourouiba in ref. . (B) Components of MAF. Organic metabolites present both in caseum and in Minimum Essential Medium-α (MEM- α) are in red. Those found in caseum but lacking in MEM-α are in green. Inorganic salts present in MEM-α are highlighted in purple, along with the organic compound bicarbonate, given its substantial contribution to the osmolality of MEM-α. Lipids are colored by class. Concentrations of each component are given in SI Appendix, Table S2 A–F. (C) Osmolality of MAF, 7H9 supplemented with glycerol and the latter further supplemented with tyloxapol, oleic acid, albumin, dextrose, and catalase at the standard concentrations listed in SI Appendix, Table S2 A–F. Means ± SEM for three measurements. The dotted line indicates the osmolality of MEM-α, which mimics human extracellular fluid. The solid line indicates values reported for mouse spleen and shaded area highlights values reported for BCG-infected mouse tissue (see text). (D) Dynamic viscosity $\eta$ and surface tension ${\gamma }_{\mathit{GL}}$ of MAF, MEM-α and its six major salts, and contact angle ${\theta }_0$ of those fluids on a tissue culture-treated polystyrene surface were measured and compared to MilliQ deionized water. Density is reported in SI Appendix, Fig. S1D. Means ± SD for at least three measurements; individual values not shown where n > 10. (E) Cough cloud generated by the exhalation mimic system ExhaleSimulator (32) with a trajectory here shown to span 3 m from the source. (F) Total droplet counts within size range 0.5 to 20 µm, measured at 1 m from the source in (E). (G) Cumulative droplet size distribution of MAF particles counted in (F). Dashed lines indicate that ~94% were ≤3 µm in diameter.

Fig. 2.

Effects of MAF on Mtb’s survival, drug tolerance, and transcriptome in bulk culture. (A) Survival of Mtb preincubated in MAF under 0.2% O₂, 5% CO₂ or 21% O₂, 5% CO₂ for 2 or 8 wk. (B–F) Survival of Mtb from log phase replicating cultures in 7H9 under 21% O₂, 5% CO₂ or following preincubation in MAF under the indicated levels of O₂ and CO₂ for 2 or 8 wk and treated with the indicated antibiotics. Kan, kanamycin. INH, isoniazid. Moxi, moxifloxacin. RIF, rifampicin. DMSO, dimethylsulfoxide vehicle (1%). (A–F) Data are means ± SEM of two independent experiments. P values were determined by the unpaired t test. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001. (G) Principal component analysis of gene expression profiles of Mtb from log phase replicating culture in 7H9 under or 21% O₂, 5% CO₂ or after incubation for 2 wk in 7H9 or MAF under the indicated O₂ and CO₂ concentrations. (H) Volcano plot of differentially expressed genes from Mtb incubated under O.2% O₂, 5% CO₂ in 7H9 or MAF for 2 wk vs. a log phase replicating culture in 7H9 under 21% O₂, 5% CO₂. (I) Comparison of upregulated genes in Mtb incubated under O.2% O₂, 5% CO₂ in 7H9 vs. MAF for 2 wk. Selected top genes upregulated in MAF and belonging to different functional categories are listed. (J) Multi-dimensional scaling (MDS) of the Mtb H37Rv transcriptome in log phase replicating culture in 7H9 under 21% O₂, 5% CO₂ or after incubation in 7H9 or MAF under the indicated O₂ and CO₂ tensions for 2 or 6 wk, compared with the following: 34 samples of Mtb-containing caseum from 26 pulmonary lesions in five rabbits infected with Mtb HN878 (duplicate samples were analyzed from eight of the lesions) having CFU burdens ranging from 7.45 × 10(4) to 1.10 × 10⁸ per gram; Mtb incubated in a caseum mimic prepared from a lysate of human macrophage-like THP1 cells made foamy by incubation with stearic acid; and Mtb-infected THP1 cells. The MDS was kindly generated in the Sherman lab (University of Washington) using the data for Mtb H37Rv in MAF and comparing it to their own results for Mtb HN878 in caseum, caseum mimic, and THP1 cells. The data were batch corrected using ComBat-seq (48) prior to dimensionality reduction.

Fig. 3.

Properties of desiccating droplets. (A) Survival of Mtb and BCG taken from log phase culture in 7H9 and transferred to the indicated fluids for 1 d in bulk culture before desiccation for 1 d in 2 μL droplets. Means ± SEM of two independent experiments, each in triplicate. *P ≤ 0.05, ****P ≤ 0.0001 by the unpaired t test. (B) Normalized volume, V/V₀, as it changes from V₀ = 2 μL in sessile droplets of MAF, MEM-α, six salts, deionized water, and MAF, with and without BCG, over normalized time, t/t_CCRL, compared to theoretical predictions taking the droplet temperature, T_d, to be either ambient temperature, T_a, (dashed black line) or measured droplet surface temperature, T_m (solid black line); see SI Appendix, Fig. S3 C and D. Phase I lasted about 15 min at relative humidity ≤5% and ambient temperature T_a = 22 °C. Inset shows void fraction, $V_{\mathit{void}}/V_r=(V_r-V_s)/V_r$, where $V_r$ is the terminal residue volume and $V_s$ is the compact volume of dry solutes. Shown for 10 measurements per condition (mean ± SE). (C) Time-lapse photographs capturing side views of evaporating droplets in Phase I and the transition to Phase II, where the residues of the evaporated droplets formed. The edge of the droplets remained pinned except for MEM-α, which receded to the partially pinned edge toward one side. (Scale bar, 500 μm.) (D) Bottom views of the residue of droplets in Phase II. (Scale bar, 500 μm.) (E) Height profile of droplets in Phase II, also quantified in SI Appendix, Fig. S3E, serves as basis, with SI Appendix, Fig. S3F, to compute the terminal residue volume, $V_r$, used in Phase II Inset in (B) for each fluid. (Scale bar, 500 μm.) (F) Brightfield and fluorescence photomicrographs of 2-µL desiccated droplets of the indicated fluids containing Mtb-Mrx1roGFP2. (Scale bars, 500 μm.) (G) Scanning electron micrographs of the surfaces of desiccated droplets containing BCG. Crumpled BCG cells (false-colored red) were found on the surface of desiccated droplets of six salts and MEM-α. Some nondeformed BCG (false-colored blue) were also found on the surface of desiccated droplets of MEM-α. No BCG were found on the surface of desiccated droplets of MAF. (Scale bar, 5 μm.)

Fig. 4.

Genome-wide CRISPRi screen of Mtb during three modeled stages of transmission. (A) Schematic of the screen across transmission stages (created with BioRender.com) and summary of comparisons. Each stage allows for multiple possible comparisons where significant differences may be detected. Stage 1: 12 possible comparisons against the input library across two screens, each testing three predepletion times (5, 10, 20 generations) with and without outgrowth in 7H9 in 21% O₂, 5% CO₂ after exposure to the conditions of the stage. Stage 2: 16 possible comparisons, combining comparisons to the input library and stage-to-stage comparisons. Stage 3: 20 possible comparisons. “Matched” comparisons maintain the same predepletion time and outgrowth condition between stages (e.g., comparing Stage 2 vs. Stage 1 at 5 generations, no outgrowth). (B) Variation in frequency of conventionally essential genes in MAF as a function of the NOSDs for each gene across all stages of the CRISPRi screen. Genes with the highest NOSDs are listed at the Left and those with fewest NOSDs or none to the Right. Within each set of 100 consecutive genes, the proportion of genes is indicated that were previously categorized as essential in standard culture conditions in 7H9 or 7H10 media. (C–E) Volcano plots for 1st, 2nd, and 3rd stages, in which the median log₂ FC for all significant differences in expression of a given gene at any stage compared to the starting libraries or the earlier stages is plotted against the median false discovery rate for those differences. Symbols for each gene are sized in proportion to the NOSDs (“Counts”). (F) Frequency distribution of the most negative log₂ FC for each gene at any stage according to whether the gene has been characterized as essential (red bars) or nonessential (blue bars) under conventional conditions. For ease of visualization, the height of each histogram is normalized to equalize the areas under the curve for genes of the two classes. (G) Schematic of CFU survival assay described in Fig. 4 H–J (created with BioRender.com). (H–J) Survival assay for CRISPRi knockdown strains of (H) devR, (I) otsA, and (J) otsB2. Data are means ± SEM of two independent experiments. P values were determined by the unpaired t test. *P ≤ 0.05; **P ≤ 0.01; ****P ≤ 0.0001; ns = nonsignificant.

Fig. 5.

Prominence of proteostasis genes in Mtb’s survival during modeled transmission. (A) Comparison of NOSDs for all stages vs. 3rd stage. Each dot corresponds to a gene with NOSD ≥ 1 in the 3rd stage. (B–E) Survival assay for Mtb knockdown or knockout strains of (B) clpP1P2, (C) prcBA, (D) dnaJ2, and (E) groEL2. Data are means ± SEM of two independent experiments. P values were determined by the unpaired t test. *P ≤ 0.05; ****P ≤ 0.0001; ns = nonsignificant.

Fig. 6.

Identification of hydrophilin-like proteins as desiccation stage candidates in the transmission survival genome of Mtb. (A) Analysis of the full Mtb proteome in three descriptor dimensions: glycine content, hydrophilicity (averaged across all residues in each protein), and ratio of residues predicted to be in intrinsically disordered regions. Highlighted cluster contains 50 proteins. (B) Proteins in the highlighted cluster whose genes had NOSD ≥ 1 in the desiccation stage.