Mycobacterium bovis frd operon phase variation hijacks succinate signaling to drive immunometabolic rewiring and pathogenicity

Abstract

Tuberculosis (TB), caused by Mycobacterium tuberculosis complex (MTBC) pathogens, remains a global health threat. While bacterial genetic adaptations during host infection are poorly understood, phase variation in genomic homopolymeric tracts (HT) may drive pathogenicity evolution. Here, we demonstrate that M. bovis exploits HT insertion mutations in the fumarate reductase-encoding frd operon to subvert host immunometabolism. In macrophages, wild-type M. bovis secretes FRD-catalyzed succinate, stabilizing hypoxia-inducible factor-1α (HIF-1α) to drive glycolytic reprogramming and IL-1β production. This activates IL-1R-dependent Th1 immunity, restraining bacterial replication. Conversely, M. bovis frd HT insertion mutants impair succinate secretion, suppressing HIF-1α/IL-1β signaling and redirecting immunity toward pathogenic Th17 responses that promote neutrophil infiltration and tissue necrosis. Mice infection models reveal that M. bovis frd mutants exhibit enhanced pathogenicity, with higher pulmonary bacterial burdens. IL-1R blockade phenocopies frd HT insertion mutation effects, exacerbating lung pathology. Crucially, conserved frd HT polymorphisms in clinical M. tb isolates suggest shared immune evasion strategies across MTBC pathogens. Our work uncovers the bacterial gene phase variation mechanism of hijacking the succinate/HIF-1α/IL-1β axis to operate host immunity, providing a framework for targeting host metabolic checkpoints in TB therapy.

In this work, authors show that Mycobacterium bovis exploits frd operon phase variation to hijack host immunity. By disrupting bacterial succinate production, identified mutations steer immunity toward harmful Th17 responses instead of protective Th1, worsening disease.

Introduction

Tuberculosis (TB) is a major public health threat caused by members of the Mycobacterium tuberculosis complex (MTBC), which exhibit significant heterogeneity in pathogenicity and transmissibility^1,2. This ancient pathogen caused 10.8 million incident human TB cases and 1.25 million deaths worldwide in 2023³. As a key member of MTBC, Mycobacterium bovis (M. bovis) infects various hosts including livestock and humans, causing substantial economic losses and posing serious zoonotic threats⁴. According to the World Health Organization (WHO), M. bovis is responsible for ~140,000 new human TB cases and 11,400 deaths annually⁵. Despite the high genomic conservation among MTBC members (nucleotide identity > 99%), they demonstrate remarkable genetic diversity during long-term host colonization and under drug selection pressure^1,6.

The adaptive evolution of MTBC in hosts involves immune evasion, metabolic adaptation, and persistent infection^1,6. For instance, M. tb isolates associated with severe TB may evade cytosolic surveillance systems through single nucleotide polymorphisms (SNPs) accumulation in ESX-1 secretion system components, reducing IL-1β production⁷. Similarly, lldD2 mutations upregulate lactate dehydrogenase activity to adapt to host-imposed lactate stress⁸. Beijing lineage of M. tb enhances virulence through ppe38 mutations that block PE_PGRS secretion⁹. Additionally, TbD1 deletion significantly improves oxidative stress and hypoxia resistance in “modern” M. tb strains, facilitating global TB dissemination¹⁰. Recent studies have identified phase variation induced by insertions/deletions (INDELs) in genomic homopolymeric tracts (HT) as a crucial adaptive mechanism in MTBC^{11, 12–13}. This slip-strand mispairing-mediated genetic “switch” not only regulates drug resistance phenotypes but may also modulate M. tb virulence^{11,14, 15–16}. However, genetic variations and adaptive evolution in clinical M. bovis isolates remain understudied.

Granulomas, characterized by caseous necrosis and lymphocyte stratification, serve as both pathological hallmarks of TB and sanctuaries for MTBC persistence^17,18. Hypoxia adaptation is crucial for mycobacterial persistence within granulomas and macrophages^17,18. MTBC ancestor acquired numerous genes through horizontal gene transfer, including those encoding transferase enzymes and hypoxia-adapted genes, such as the frd operon encoding fumarate reductase (FRD)^19,20. Our previous whole-genome sequencing (WGS) of clinical M. bovis isolates revealed prevalent insertion mutations in a HT of 7 guanines (7 G) in the frd operon²¹. Although FRD catalyzes succinate production during hypoxia, the immunopathological consequences of these phase variations remain unclear²². Notably, similar 7 G HT polymorphisms in clinical M. tb isolates have not been thoroughly investigated¹¹. As intracellular pathogens, MTBC metabolites can reshape host physiology and immunity^{23, 24–25}. We hypothesize that frd phase variations in clinical M. bovis and M. tb isolates disrupt host defense through metabolic reprogramming.

Here, we show how the frd 7 G HT insertion mutation suppresses macrophage immunometabolism and enhances pathogenicity through immune evasion. Mechanistically, wild-type (WT) M. bovis infection induces FRD-dependent succinate secretion, stabilizing hypoxia-inducible factor 1α (HIF-1α) to drive IL-1β production and subsequent IL-1 receptor (IL-1R)-mediated Th1 polarization. Conversely, M. bovis frd mutants exhibit reduced succinate secretion, skewing immunity toward pathogenic Th17 responses that promote necrotizing neutrophilic inflammation and uncontrolled bacterial replication. By revealing an immunometabolic axis linking bacterial phase variation to host immunity, our findings provide important insights into mycobacterial evasion strategies and propose alternative approaches for TB control through metabolic interference.

Results

High-frequency insertion mutations in the frd operon of clinically isolated M. bovis

Our prior study of the genetic diversity of clinical M. bovis isolates in China revealed a high prevalence of low-frequency genetic variations²¹. Notably, multiple G insertions were identified in the 7 G HT of the frd operon, which contains four genes with the 7 G HT located at the frdB terminus (Fig. 1a, b). WGS analysis of 74 M. bovis isolates showed G insertion-induced diverse mutation patterns in the 7 G HT. 56.8% of isolates displayed an 8 G HT configuration from a single G insertion, while only 6.8% retained the WT 7 G HT (Fig. 1b, c). Given the high genomic similarity between M. bovis and M. tb, we analyzed 2819 M. tb genomes in GMTV database²⁶ and detected similar G insertions in the frd operon of some isolates (Fig. 1d). As the single-base insertion causing the 8 G HT variant was predominant in M. bovis isolates, we compared the three-dimensional (3D) structures of FRD proteins encoded by WT (7 G HT) and mutant (8 G HT) operons. Structural analysis indicated that this insertion did not induce frameshift mutation but eliminated the frdB termination codon, resulting in an in-frame fusion with FrdC (Fig. 1e).

[See PDF for image]

Fig. 1

Insertion mutations in the frd HT of M. bovis clinical isolates.

a Genetic organization of the frd operon and the localization of the 7 G HT. b G-insertion patterns in the 7 G HT region of M. bovis clinical isolates. c, d Mutation frequencies in M. bovis (n  =  74) and M. tb (n  =  2819) isolates. e 3D structural models of wild-type FRD (colored subunits: FrdA in blue, FrdB in red, FrdC in green, FrdD in purple) and the 8 G HT mutant (gray). Source data are provided as a Source Data file.

To elucidate the phylogeographic distribution and evolutionary conservation of the frd 7 G HT insertion mutation in M. bovis, we analyzed genomes from Mexico, the USA, the United Kingdom, New Zealand, and Ethiopia. Notably, isolates from Mexico and the USA predominantly exhibited the 8 G HT mutation pattern observed in Chinese strains, while those from New Zealand, the United Kingdom, and Ethiopia exhibited a distinct 9 G HT mutation as the dominant genotype (Supplementary Fig. 1a–e). Structural modeling revealed that the 9 G HT insertion introduces a frameshift that fundamentally alters the FRD architecture by inducing complete translational fusion of FrdB, FrdC, and FrdD subunits into a single contiguous polypeptide chain (Supplementary Fig. 1f). This aberrant fusion results in marked structural divergence from the WT FRD, exceeding the conformational changes caused by the 8 G HT mutation. These findings suggest that high-frequency 7 G HT insertion mutations represent a conserved evolutionary strategy across geographically distinct M. bovis populations.

Functional consequences of frd HT insertion mutations in M. bovis

In Escherichia coli, the quinone-binding pocket formed by FrdB, FrdC, and FrdD accommodates menaquinone (MQ) for electron transport²⁷. Mycobacteria similarly utilize MQ as a central electron carrier, with MQ9 being the primary isoform²⁸ (Fig. 2a). To assess the functional impact of the M. bovis FrdB-FrdC in-frame fusion on FRD activity, molecular docking analyses were performed. The analyses showed that MQ9 occupies the quinone-binding pocket in WT FRD (Fig. 2b, c). Strikingly, the 8 G HT mutation caused the peptide chain to span and block the quinone-binding pocket (Fig. 2d). FRD is essential for catalyzing fumarate reduction to succinate under hypoxia in mycobacteria²² (Fig. 2e). To further evaluate frd insertion effects, we generated an M. bovis Δfrd knockout strain from M. bovis C68004 (Supplementary Fig. 2a, b). Genetic complementation with WT (7 G HT) or mutant (8 G HT) frd operons revealed that both M. bovis Δfrd and Δfrd::frd8G showed markedly reduced succinate secretion under hypoxia, but there was no significant change in growth kinetics and colony morphotypes in vitro (Fig. 2f and Supplementary Fig. 2c, d). Although classified as obligate aerobes, mycobacteria frequently encounter hypoxic niches in granulomas or macrophages^17,18. Additional experiments are therefore needed to demonstrate whether and how this variation affects mycobacterial pathogenicity.

[See PDF for image]

Fig. 2

Insertion mutations in the frd HT of M. bovis affect the 3D structure and function of FRD.

a Chemical structure of menaquinone-9 (MQ9). b Superimposed docking conformation of MQ9 (yellow) to wild-type (WT) FRD, with FrdA hidden and subunits colored: FrdB (red), FrdC (green), FrdD (purple). c Stereoscopic view showing MQ9 bound to the polar pocket formed by FrdB, FrdC, and FrdD. d The 8 G HT mutation-induced FrdB-FrdC in-frame fusion (gray) sterically obstructs the MQ9-binding pocket, with structural occlusion mediated by an orange-highlighted peptide segment. e FRD-catalyzed reduction of fumarate to succinate. f Succinate secretion by M. bovis WT, Δfrd, Δfrd::frd7G, and Δfrd::frd8G cultured in vitro under aeration or hypoxia (n  =  3 biological replicates). Data are presented as mean values ± SD. P values depicted on the graph were assessed using one-way ANOVA with Tukey’s multiple comparisons test. Source data are provided as a Source Data file.

Clinical M. bovis frd mutant exhibits enhanced pathogenicity

To evaluate the pathogenetic impact of frd operon variations, we compared clinical M. bovis isolates harboring WT (Iso 7G) or mutant (Iso 8G) frd operons in C57BL/6 mice. M. bovis Iso 8G-infected mice exhibited rapid weight loss with ~14% reduction at 21 days post-infection (dpi) (Fig. 3a). At 22 dpi, M. bovis Iso 8G-infected mice demonstrated ~2-fold higher lung-to-body weight ratios than Iso 7G-infected controls, with ~1.5-log higher pulmonary CFU levels (Fig. 3b, d and Supplementary Fig. 3a). Notably, M. bovis Iso 8G infection resulted in significantly lower spleen-to-body weight ratios than Iso 7G-infected controls, though splenic bacterial loads remained comparable (Fig. 3c, e and Supplementary Fig. 3b). This suggests that there may be a difference in the immune response induced by these two M. bovis isolates’ infection. Histopathological analysis revealed distinct immune cell infiltration and inflammatory lesions in the lung tissues of mice (Fig. 3f, g). At 12 dpi, M. bovis Iso 7G-infected lungs contained lymphocyte/macrophage-dominant inflammatory foci, whereas Iso 8G infection induced neutrophil-rich infiltrates. By 22 dpi, Iso 8G-infected lungs exhibited extensive necroinflammatory damage with pervasive neutrophil infiltration, whereas Iso 7G-infected lungs maintained localized lymphocytic infiltrates without substantial necrosis (Fig. 3g). The white pulp and red pulp of the spleen were replaced by a large number of macrophages, epithelioid cells, and lymphocytes in both groups of infected mice (Supplementary Fig. 4).

[See PDF for image]

Fig. 3

Clinical M. bovis frd mutant exhibits enhanced pathogenicity.

a–f C57BL/6 mice were infected with ~5000 CFU of two different clinical isolates intranasally for 12 and 22 days. a Body weight gain rate curves of C57BL/6 mice uninfected or infected with two different clinical isolates. b, c The ratio of lung or spleen weight to the total body weight at 12 and 22 dpi. d, e Bacterial loads (CFU) in lung or spleen. f The percentage of lung’s area occupied by inflammatory lesions. g Hematoxylin and eosin (H&E)-stained lung sections at 12 and 22 dpi. Arrows indicate lymphocytes, and arrowheads indicate neutrophils. h, i Number of cells and activated CD4⁺ T cells in draining lymph nodes of mice at 12 dpi. j, k Percentage of CD4⁺CD69⁺ T cells expressing IFN-γ or IL-17A in the draining lymph nodes of mice as in (h). Data are presented as mean values  ± SD. Sample size of n  =  5 mice was included in each group. P values depicted on the graphs were assessed using one-way ANOVA with Tukey’s multiple comparisons test (a–c) and two-tailed unpaired Student’s t-test (d–f, h–k). Scale bars, 1 mm (top) and 20 μm (bottom). Source data are provided as a Source Data file.

The host immune response during early MTBC infection determines disease outcomes. Th17 promotes neutrophil recruitment through IL-17A secretion, and excessive neutrophil infiltration during MTBC infection correlates with detrimental disease progression by establishing a permissive niche facilitating unrestricted bacterial replication^{29, 30–31}. Based on the above findings, we hypothesize that M. bovis Iso 8G infection triggers early Th17 polarization, driving neutrophil-mediated pathology and enhanced bacterial proliferation. Flow cytometry analysis of lung draining lymph nodes at 12 dpi showed comparable total cellularity and CD4⁺CD69⁺ T cell counts between groups (Fig. 3h, i). However, M. bovis Iso 8G infection exhibited elevated CD4⁺CD69⁺IL-17A⁺ T cells with concomitant reduction in CD4⁺CD69⁺IFN-γ⁺ T cells at 12 dpi, indicating M. bovis Iso 8G infection drove subsequent neutrophil infiltration and immunopathology by inducing Th1-to-Th17 immune deviation (Fig. 3j, k and Supplementary Fig. 5). Given polymorphisms in other virulence factors also present in clinical isolates, definitive validation requires isogenic M. bovis Δfrd strains complemented with 7 G HT frd or 8 G HT frd operons.

frd HT insertion mutations drive Th17 polarization and neutrophilic immunopathology in M. bovis infection

To establish a causal linkage between the frd HT insertion and enhanced pathogenicity, we infected C57BL/6 mice with isogenic M. bovis C68004 derivatives: WT, Δfrd, Δfrd::frd7G, and Δfrd::frd8G. Both M. bovis Δfrd and Δfrd::frd8G manifested rapid weight loss (~20% reduction at 21 dpi), increased lung-to-body weight ratios, and elevated pulmonary bacterial burdens compared to WT and Δfrd::frd7G controls (Fig. 4a, b, d and Supplementary Fig. 6a). Consistent with clinical M. bovis Iso 8G infections, mutant strains showed reduced splenic indices with equivalent bacterial colonization (Fig. 4c, e and Supplementary Fig. 6b). Critically, M. bovis Δfrd and Δfrd::frd8G recapitulated the immunopathological signature of clinical Iso 8G: neutrophil-dominated infiltrates at 12 dpi, as well as extensive necrotizing inflammation at 22 dpi, while WT and Δfrd::frd7G infected lungs showed a large number of lymphocyte infiltrations without significant necrosis (Fig. 4f, g). Similar to the results of clinical M. bovis isolates infection, granulomatous inflammation was found in both white pulp and red pulp of the spleen (Supplementary Fig. 7).

[See PDF for image]

Fig. 4

The frd HT insertion mutations enhance M. bovis pathogenicity.

a–f C57BL/6 mice were infected with ~5000 CFU of M. bovis intranasally for 12 and 22 days. a Body weight gain rate curves in C57BL/6 mice uninfected or infected with M. bovis. b, c The ratio of lung or spleen weight to the total body weight at 12 and 22 dpi. d, e Bacterial loads in lungs or spleen. f The percentage of lung’s area occupied by inflammatory lesions. g HE-stained lung sections at 12 and 22 dpi. Arrows indicate lymphocytes, and arrowheads indicate neutrophils. Data are presented as mean values  ± SD. Sample size of n  =  5 mice was included in each group. P values depicted on the graphs were assessed using one-way ANOVA with Tukey’s multiple comparisons test. WT, wild-type. Scale bars, 1 mm (top) and 20 μm (bottom). Source data are provided as a Source Data file.

To delineate the immunomodulatory effects of frd mutations, we assessed T cell activation and cytokine profiles in lung-draining lymph nodes of C57BL/6 mice infected with M. bovis WT, Δfrd, Δfrd::frd7G, or Δfrd::frd8G at 12 and 22 dpi. Flow cytometry revealed M. bovis Δfrd and Δfrd::frd8G infections caused reduced total cellularity compared to WT and Δfrd::frd7G controls at 12 dpi, with sustained reduction through 22 dpi (Fig. 5a, d). Intracytoplasmic staining showed that CD4⁺CD69⁺IFN-γ⁺ T cells of mice infected with M. bovis Δfrd and Δfrd::frd8G decreased significantly (Fig. 5b, e), concomitant with elevated CD4⁺CD69⁺IL-17A⁺ T cells (Fig. 5c, f). This discovery confirmed that frd insertion mutations drive Th17-skewed polarization. Serum cytokine quantification revealed elevated IL-17A and suppressed IFN-γ levels in M. bovis Δfrd- and Δfrd::frd8G-infected mice, validating the frd-dependent immune response shift from Th1 to Th17 dominance (Fig. 5g, h).

[See PDF for image]

Fig. 5

M. bovis frd HT insertion mutations reprogram lymph node T cell polarization toward Th17 responses.

a Number of cells in draining lymph nodes of mice at 12 dpi. b, c Percentage of CD4⁺CD69⁺ T cells expressing IFN-γ or IL-17A in the draining lymph nodes of mice as in (a). d Number of cells in draining lymph nodes of mice at 22 dpi. e, f Percentage of CD4⁺CD69⁺ T cells expressing IFN-γ or IL-17A in the draining lymph nodes of mice as in (d). g, h Serum IFN-γ and IL-17A in infected mice. Data are presented as mean values  ± SD. Sample size of n  =  5 mice was included in each group. P values depicted on the graphs were assessed using one-way ANOVA with Tukey’s multiple comparisons test. WT, wild-type. Source data are provided as a Source Data file.

M. bovisfrd HT insertion mutations inhibit macrophage glucose metabolism reprogramming

As primary reservoirs for mycobacterial parasitism and key effectors of anti-TB immunity, macrophages undergo metabolic reprogramming during infection¹⁸. RNA-seq profiling of mouse bone marrow-derived macrophages (BMDMs) infected with M. bovis WT or Δfrd for 12 h identified 192 upregulated and 344 downregulated differentially expressed genes in M. bovis Δfrd-infected cells (Supplementary Fig. 8a, b). Pathway analysis revealed M. bovis Δfrd infection downregulated innate immunity, response to hypoxia, and defense against Gram-positive bacteria while upregulating neutrophil chemotaxis and IL-17 signaling pathway (Supplementary Fig. 8c, d). This indicates that M. bovis Δfrd inhibits innate immunity in macrophages and promotes neutrophil recruitment, explaining early neutrophil-dominated pulmonary immunopathology (Fig. 4g).

Notably, Gapdh expression, encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a glycolytic enzyme, was significantly suppressed in M. bovis Δfrd-infected BMDMs (Supplementary Fig. 8b, e). After M. bovis infection, macrophages switch their metabolism from oxidative respiration through the TCA cycle to high-rate aerobic glycolysis (Supplementary Fig. 9a, b). Relying on glycolysis, macrophages generate sufficient ATP and biosynthetic precursors to fuel antimicrobial responses²⁴, including the production of proinflammatory cytokines like IL-1β (Supplementary Fig. 9c). We speculate that frd mutation may affect the reprogramming of glucose metabolism in M. bovis-infected macrophages.

Next, we determined the differences in central carbon metabolism in BMDMs infected with isogenic M. bovis C68004 derivatives for 12 h. The results showed that several metabolites, including primary metabolites of the TCA cycle, were significantly altered during M. bovis WT and Δfrd::frd7G infection, with marked reduction in M. bovis Δfrd and Δfrd::frd8G-infected cells (Fig. 6a, and Supplementary Data 1). Glycolytic flux, reflected by lactate accumulation and Ldha expression, was attenuated in M. bovis Δfrd-infected macrophages (Fig. 6a, b). Enrichment of immunomodulatory TCA metabolites (succinate, citrate) correlated with nitric oxide-mediated cycle disruption^24,32,33, consistent with Nos2 transcriptional induction (Fig. 6c). These findings demonstrate that frd HT insertion impairs glucose metabolism rewiring in M. bovis-infected macrophages.

[See PDF for image]

Fig. 6

M. bovis frd HT insertion mutations impair macrophage glucose metabolism reprogramming.

a–c BMDMs infected with isogenic M. bovis derivatives (multiplicity of infection (MOI)  =  10) for 12 h (n  =  3 biological replicates). a Intracellular levels of central metabolites were measured by mass spectrometry. b Relative Ldha mRNA expression. c Relative Nos2 mRNA expression. Data are presented as mean values  ± SD. P values depicted on the graphs were assessed using one-way ANOVA with Tukey’s multiple comparisons test. WT, wild-type. Source data are provided as a Source Data file.

M. bovis frd HT insertion mutations suppress macrophage immunometabolism by disrupting the succinate/HIF-1α/IL-1β signaling axis

M. bovis upregulates frd operon expression in macrophages (Supplementary Fig. 9d), promoting bacterial succinate secretion. As an inflammatory signaling metabolite, succinate stabilizes hypoxia-inducible factor-1α (HIF-1α) by inhibiting prolyl hydroxylase (PHD) activity^34,35. In BMDMs infected with WT M. bovis, we observed robust HIF-1α accumulation (Supplementary Fig. 9e). HIF-1α, functioning as a transcriptional regulator, drives glycolytic reprogramming and induces IL-1β production³⁵. Hypoxia-exposed M. bovis Δfrd and Δfrd::frd8G strains exhibited reduced succinate secretion compared to WT and Δfrd::frd7G controls, which may affect the content of HIF-1α in macrophages (Fig. 2f).

HIF-1α dynamics in infected BMDMs revealed sustained accumulation of HIF-1α protein in M. bovis WT- and Δfrd::frd7G-infected cells at 12 and 24 h post-infection, whereas M. bovis Δfrd and Δfrd::frd8G infections attenuated this response (Fig. 7a, b). Metabolite quantification revealed strain-dependent patterns: extracellular lactate (glycolysis endpoint) and IL-1β levels paralleled HIF-1α expression (Fig. 7c, d). To establish causality, we employed the HIF-1α inhibitor YC-1 [3-(5’-hydroxymethyl-2’-furyl)-1‑benzyl indazole] to eliminate infection-induced HIF-1α variability (Fig. 7e). YC-1 treatment eliminated infection-induced differences in lactate and IL-1β secretion (Fig. 7f, g), confirming frd HT insertions regulate immunometabolism through HIF-1α. We further dissected whether bacterial-derived succinate drives this phenotype. Hypoxic culture filtrates from M. bovis WT (high succinate) versus Δfrd (low succinate) were used to treat BMDMs. M. bovis WT culture filtrates induced higher HIF-1α and elevated IL-1β compared to M. bovis Δfrd (Fig. 7h), mechanistically linking bacterial succinate to host immunometabolism crosstalk.

[See PDF for image]

Fig. 7

M. bovis frd HT insertions suppress macrophage immunometabolism.

a Western blot analysis of HIF-1α protein levels in BMDMs infected with isogenic M. bovis derivatives for 0, 12, and 24 h (n  =  3 biological replicates). b BMDMs infected with FITC-labeled M. bovis (green) for 12 h to observe HIF-1α (red) expression, with nuclei (blue) stained by DAPI. Scale bar, 10 μm. c Extracellular lactate from BMDMs treated as in (a). d IL-1β ELISA in supernatants from BMDMs treated as in (a). e BMDMs were pretreated with 10 μM YC-1 for 12 h. Protein blotting analysis of HIF-1α protein levels was performed after 12 h of infection with M. bovis (n  =  3 biological replicates). f Extracellular lactate from BMDM treated as in (e). g IL-1β ELISA in supernatants from BMDM treated as in (e). h Western blot analysis of HIF-1α after incubation of BMDMs with M. bovis WT or Δfrd hypoxic culture filtrate for 12 h. Representative images are presented from three biologically independent experiments (a, b, e, h). Data are presented as mean values  ± SD. P values depicted on the graphs were assessed using one-way ANOVA with Tukey’s multiple comparisons test. WT, wild-type. Source data are provided as a Source Data file.

M. bovis frd HT insertions drive Th17-biased immunity and bacterial pathogenicity via IL-1R suppression

Mycobacteria-infected macrophages rapidly localize to the lung interstitium in an IL-1R-dependent manner, followed by rapid bacterial dissemination to draining lymph nodes and initiation of Th1 to control the bacteria^29,36. Given that frd mutations disrupt the succinate/HIF-1α/IL-1β axis, we hypothesized IL-1R signaling drives Th polarization differences between WT and mutant infections in vivo. IL-1R blockade during M. bovis infection suppressed lymph node CD4⁺CD69⁺ T cells at 12 and 22 dpi (Fig. 8a, d). Concurrently, this blockade decreased CD4⁺CD69⁺IFN-γ⁺ T cells while elevating CD4⁺CD69⁺IL-17A⁺ T cells (Fig. 8b, c, e, f). Crucially, IL-1R blockade eliminated Th1/Th17 differences between M. bovis WT and Δfrd infections (Fig. 8b–f). Mirroring M. bovis Δfrd pathology, IL-1R inhibition compromised pulmonary bacterial containment (elevated CFU and inflammatory areas) (Fig. 8g, h) and converted lymphocyte-dominated inflammation to neutrophil-driven necroinflammation in M. bovis WT-infected lungs (Fig. 8i).

[See PDF for image]

Fig. 8

M. bovis frd HT insertions drive Th17-biased immunity and bacterial pathogenicity via IL-1R suppression.

C57BL/6 mice were infected with ~5000 CFU of the indicated strain intranasally. 200 μg of anti-IL-1R1 antibody (+) or isotype control (−) was administered by i.p. on days 8, 10, and 12 dpi. a Number of CD4⁺CD69⁺ T cells in draining lymph nodes of mice at 12 dpi. b, c Number of CD4⁺CD69⁺ T cells expressing IFN-γ or IL-17A in the draining lymph nodes of mice as in (a). d Number of CD4⁺CD69⁺ T cells in draining lymph nodes of mice at 22 dpi. e, f Number of CD4⁺CD69⁺ T cells expressing IFN-γ or IL-17A in the draining lymph nodes of mice as in (d). g Bacterial loads in lungs. h The percentage of lung’s area occupied by inflammatory lesions. i HE-stained lung sections at 22 dpi. Arrows indicate lymphocytes, and arrowheads indicate neutrophils. Data are presented as mean values  ± SD. Sample size of n  =  5 mice was included in each group. P values depicted on the graphs were assessed using one-way ANOVA with Tukey’s multiple comparisons test. WT, wild-type. Scale bars, 1 mm (left) and 20 μm (right). Source data are provided as a Source Data file.

Discussion

TB remains a global health crisis, with M. bovis posing dual threats to livestock and humans through zoonotic transmission^1,2,4. A critical gap in understanding TB pathogenesis lies in how bacterial genetic adaptations, particularly phase variation, modulate host immunometabolism to drive disease progression^6,7,11. Here, we identified frd-mediated succinate secretion as a critical determinant of macrophage immunometabolic reprogramming. M. bovis frd HT insertion mutations suppress HIF-1α stabilization by reducing the secretion of succinate, thereby impairing glycolysis and IL-1β production in macrophages. This metabolic reprogramming drives a Th17-skewed immune response through the IL-1R signal, facilitating neutrophil-dominated pathology and bacterial replication (Fig. 9). While metabolic crosstalk between mycobacteria and host cells is recognized, the molecular mechanisms linking bacterial phase variation to immune evasion remain poorly defined^23,25,37. By demonstrating that frd mutations disrupt the succinate/HIF-1α/IL-1β axis, we provide a mechanistic framework for how bacterial genetic plasticity subverts host immunity.

[See PDF for image]

Fig. 9

Graphic highlight of findings.

Within infected macrophages, M. bovis FRD catalyzes succinate secretion, which stabilizes HIF-1α to reprogram host cell glucose metabolism toward glycolysis and drive IL-1β production. IL-1β subsequently activates IL-1R signaling, initiating a Th1-polarized immune response in draining lymph nodes to suppress bacterial infection. In contrast, M. bovis frd mutants disrupt FRD function, impairing the succinate/HIF-1α/IL-1β axis. This suppression of Th1 immunity promotes pathogenic Th17 polarization, resulting in uncontrolled bacterial replication and enhanced pathogenicity. Created with BioGDP.com⁷⁸.

FRD is a membrane-bound bifunctional enzyme that maintains an energized membrane in mycobacteria under anaerobic conditions²². It is hypothesized to enhance bacterial adaptation to hypoxia, potentially providing an evolutionary advantage during the transition from environmental bacteria to host-adapted pathogens⁶. Functionally, FRD catalyzes fumarate reduction to succinate, a critical reaction in anaerobic respiration across diverse organisms. However, under aerobic conditions, bacterial FRD enzymes paradoxically become major generators of reactive oxygen species, which may inflict oxidative damage on M. tb^{38, 39–40}. This dual role underscores the complexity of FRD in bacterial pathogenesis. Notably, frd operon expression may be dynamically regulated during host-pathogen interactions. Following internalization into macrophages, the M. tb H37Ra attenuated strain reveals elevated frd expression during prolonged infections compared to the virulent H37Rv strain, yet displays reduced expression in acute phases⁴¹. These temporal expression differences suggest FRD-mediated metabolic adaptations may dynamically influence bacterial virulence.

In prior studies, M. tb H37Rv frdA knockout mutant showed no significant difference in pulmonary bacterial loads compared to WT at low-dose infection (100 CFU)²², whereas our high-dose M. bovis Δfrd infection (5,000 CFU) resulted in elevated pulmonary CFU and immunopathology. This striking contrast underscores the pivotal role of infection dose in TB pathogenesis modeling. While the low-dose (100 CFU) murine model remains prevalent in TB research, emerging evidence demonstrates that higher bacterial challenges better recapitulate the immune mechanisms of active TB^{42, 43–44}. High-dose M. tb infection in C57BL/6 mice replicates the blood transcriptomic signatures observed in active TB patients—a critical feature unattainable with conventional low-dose models⁴⁴. This paradigm is further supported by non-human primate studies where aerosolized high-dose M. tb induced progressive active TB through macrophage interferon responses, mirroring the elevated IFN signatures characteristic of human active TB⁴³. Furthermore, recent work demonstrates that high-dose challenges unveil otherwise cryptic host defense mechanisms: autophagy activation in lung macrophages suppresses myeloid-derived suppressor cell accumulation, thereby potentiating T cell-mediated bacterial control⁴². Our selection of high-dose infection specifically targets to better reflect the immune response during active TB.

HIF-1α serves as a critical nexus linking macrophage metabolism to antimicrobial defense in mycobacterial infections^{45, 46–47}. Under hypoxia or inflammatory stimuli, HIF-1α stabilization reprograms macrophages toward glycolysis, enhancing ATP production and generating biosynthetic intermediates required for bactericidal functions such as reactive oxygen/nitrogen species production^24,48. Our study reveals that M. bovis frd mutants disrupt HIF-1α stabilization by reducing succinate secretion, impairing glycolytic flux and IL-1β production. This metabolic sabotage enables bacterial immune evasion. Such findings echo the previous research results, indicating that the control of mycobacteria is defective due to impaired glycolytic activity and IL-1β secretion⁴⁹.

IL-1β exhibits a dual and context-dependent role in MTBC infection. On one hand, IL-1β is critical for initiating Th1-polarized immunity and granuloma formation, as demonstrated by studies showing that IL-1R1-deficient mice exhibit defective control of M. tb infection^50,51. M. tb strains associated with severe human TB can avoid the host defense system by inducing the low expression of IL-1β⁷, a strategy mirrored in our study where M. bovis frd mutants downregulate IL-1β secretion, exacerbating pulmonary pathology through Th17-skewed inflammation. This paradoxical role extends to IL-1β overexpression, which correlates with increased lung damage in TB^52,53, highlighting the necessity for precise regulation of IL-1β levels-sufficient to sustain antimicrobial responses yet restrained to prevent immunopathology.

Notably, M. bovis frd mutants exploit this delicate balance: by disrupting the succinate/HIF-1α/IL-1β axis, they suppress Th1 immunity and promote pathogenic Th17/neutrophil cascades. This mirrors clinical observations where TB patients with elevated Th17/IL-17A signatures exhibit higher bacterial loads and cavitary lesions^54,55. The interplay between Th17 immunity and neutrophilic inflammation plays a pivotal role in TB progression. Th17 cells secrete IL-17A, which induces chemokines that recruit neutrophils to infection sites⁵⁶. While neutrophils are essential for early bacterial containment, excessive or dysregulated neutrophil infiltration exacerbates tissue damage and creates a permissive niche for mycobacterial replication^57,58. Neutrophilic extracellular traps (NETs) released by neutrophils promote M. tb replication and are strongly associated with severe lung damage^{31,59, 60–61}. The necrotic neutrophils infected by M. tb are ingested by macrophages to promote the growth of M. tb, creating a vicious cycle of inflammation and pathogen persistence⁶².

Our analysis revealed frequent insertion mutations within the frd 7 G HT of M. bovis, consistent with our previous finding of significantly higher glpK 7 C HT mutation rates in M. bovis compared to M. tb⁶³. These combined observations suggest that M. bovis may have evolved unique mechanisms to tolerate or promote replication errors. Frameshift mutations in HTs typically originate from replication-associated slipped-strand mispairing⁶⁴. In M. tb, the DNA polymerase III subunits DnaQ (3’-5’ exonuclease) and DnaE1 (replicase) physically interact with the β-clamp (DnaN), suggesting a potential coproofreading mechanism to maintain replication fidelity^65,66. Strikingly, M. bovis harbors a conserved D99G mutation in DnaN, which could promote strand slippage during replication, thereby elevating HT mutation rates, although this finding requires additional investigation.

Our study has several limitations. The C57BL/6 mouse model, while informative, may not fully recapitulate human and bovine granuloma dynamics or hypoxia gradients. The metabolomic analysis focused on central carbon pathways; broader lipidomic or fluxomic profiling could uncover additional regulatory nodes. Although molecular docking suggested that frd HT insertion mutations obstruct the quinone-binding pocket, experimental validation (e.g., cryo-electron microscopy structural analysis) is lacking. While our study establishes frd HT insertion mutations as a virulence determinant under high-dose infection, the limitations underscore the need for mechanistic depth and translational validation. Addressing these gaps through multi-omics, structural biology, and cross-species models will refine FRD’s role in mycobacterial pathogenesis and accelerate therapeutic targeting of hypoxia adaptation pathways.

In summary, we demonstrate that M. bovis frd operon HT phase variation rewires host immunometabolism, shifting from protective Th1 to pathogenic Th17 immunity. This discovery resolves a key knowledge gap by linking bacterial genetic plasticity to immune evasion, a strategy likely shared across MTBC pathogens. Screening for frd HT polymorphisms in clinical isolates may predict disease severity. Our study thus provides a roadmap for capitalizing on the pathogen’s genetic footprint to guide novel host-directed therapies. By targeting the succinate-HIF-1α-IL-1β axis and related immunometabolic checkpoints, precise recalibration of host immunity could be implemented, thereby advancing the global fight against TB.

Methods

Ethics

All experimental protocols and procedures were carried out in accordance with all relevant ethical regulations and guidelines, and were approved by the animal care and use committee (IACUC) protocols (20,110,611–01) at China Agricultural University. The animal experimental manipulations and protocols were approved by the Laboratory Animal Ethical Committee of China Agricultural University (permit number: AW02110202-2).

Sequence data collection and analysis

For this study, we downloaded raw Illumina sequence data (FastQ files) for 929 M. bovis isolates from NCBI Sequence Read Archive (SRA) (Supplementary Data 2). The data processing pipeline comprised the following steps: First, fastp (v0.20.1)⁶⁷ was employed to remove low-quality bases (Phred score < 20) and residual Illumina adapter contamination from the FastQ files. Subsequently, the filtered reads were aligned to the M. bovis AF2122/97 reference genome (NC_002945.4) using the mem algorithm of BWA v0.7.17⁶⁸. Duplicate reads were removed using Picard v2.25.5 (https://github.com/broadinstitute/picard), followed by indexing of the processed BAM files using SAMtools v1.17⁶⁹. We applied Pilon v1.24⁷⁰ with the --variant parameter to generate VCF files containing all variant calls relative to the reference genome.

Bacterial culture and infections

M. bovis C68004 was obtained from the China Institute of Veterinary Drug Control. M. bovis strains (Supplementary Table 1) were grown in Middlebrook 7H9 medium (BD, 271310, USA) supplemented with 10% oleic acid-albumin-dextrose-CAT (catalase) (OADC) (BD, 212351, USA) and 0.05% Tween-80 (Sigma-Aldrich, P1754, USA), or on Middlebrook 7H10 agar (BD, 262710, USA) supplemented with 10% OADC. The bacteria were grown for 2–3 weeks before being used for cell infection.

The M. bovis Δfrd was generated via specialized transduction following established methodology⁷¹. An allelic exchange substrate targeting the frd operon was cloned into the PacI site of phasmid phAE159 and transfected into M. smegmatis mc² 155 to produce high-titer phage particles. WT M. bovis cells were washed twice with MP buffer (50 mM Tris pH 8.0, 150 mM NaCl, 2 mM CaCl₂, 10 mM MgCl₂), then incubated with phage lysates at 37 °C for 16–18 h. Transductants were selected on Middlebrook 7H10 agar supplemented with 75 µg/mL hygromycin B at 37 °C. The deletion of frd in the M. bovis was confirmed by PCR using primers listed in Supplementary Table 2. The shuttle vector pMV361 was used to complement the strain M. bovis Δfrd with WT frd operon (M. bovis Δfrd::frd7G) or to create the strain M. bovis Δfrd::frd8G. For M. bovis Δfrd, 75 μg/ml hygromycin B was added to the culture. For M. bovis Δfrd::frd7G or M. bovis Δfrd::frd8G, 75 μg/ml hygromycin B and 50 μg/ml kanamycin were added to the culture.

For macrophage infection, BMDMs were seeded in 12-well plates at 2 × 10⁵ cells/well and allowed to adhere overnight. Cells were infected with M. bovis at MOI  =  10 for 4 h under standard culture conditions (37 °C, 5% CO₂). Extracellular bacteria were removed by three washes with pre-warmed phosphate-buffered saline (PBS; Solarbio, P1020, China). Subsequently, infected BMDMs were maintained in RPMI-1640 medium with 2% FBS for designated time intervals.

Cell culture

BMDMs were generated from femurs of 6–8-week-old C57BL/6 mice. Bone marrow cells were flushed from the femurs using RPMI-1640 medium, and bone marrow progenitors were differentiated for 6 days in RPMI-1640 medium supplemented with 20 ng/mL recombinant murine macrophage colony-stimulating factor (M-CSF; Peprotech, 315-02, USA), 10% FBS, and 1% penicillin-streptomycin. Adherent mature macrophages were harvested and seeded into experimental plates 12-16 h prior to use.

Plasmids and antibodies

Plasmids are described in Supplementary Table 1. The following antibodies were used in this study: rabbit anti-HIF-1α antibody (Bioss, bs-0737R/polyclonal, China; 1:1000 for immunoblot analysis, 1:200 for immunofluorescence), rabbit anti-IL-1β antibody (Abmart, P01583/polyclonal, China; 1:1000 for immunoblot analysis), rabbit anti-α-Tubulin antibody (Proteintech, 11224-1-AP/polyclonal, USA; 1:2000 for immunoblot analysis), donkey anti-Rabbit lgG(H + L) Secondary Antibody, YSFluor™594 (Yeasen, 34212ES60, China; 1:400 for immunofluorescence), anti-mouse IL-1R (Bio X Cell, BE0256, USA), and polyclonal Armenian hamster IgG (Bio X Cell, BE0091, USA).

Mice and infection

6–8 weeks old female Specified Pathogen Free (SPF) C57BL/6 mice were purchased from SPF Biotechnology Co., Ltd. (Beijing) and housed in ABSL-3 containment facilities under SPF conditions (12 h light/dark cycle, 25–27 °C, 50% relative humidity) with ad libitum access to food and water. Mice were intranasally challenged with 5000 CFUs of M. bovis. At 12 and 22 dpi, lungs, spleens, and draining lymph nodes were aseptically harvested. Left lung lobes and splenic halves were fixed in 10% neutral-buffered formalin and paraffin-embedded for histopathology, while contralateral tissues were homogenized in PBS using a tissue homogenizer apparatus (JXFSTPRP-24 L, Jingxin, China). Tissue lysates were serially diluted tenfold in sterile PBS and plated on Middlebrook 7H10 agar with OADC enrichment, followed by 3-4 weeks of incubation at 37 °C. Formalin-fixed tissues were sectioned (5 μm) using a Leica RM2235 microtome (Leica Biosystems, USA), stained with hematoxylin-eosin (H&E; Solarbio, G1120, China), and digitized via a TEKSQRAY SQS40P slide scanner (Shengqiang Technology, China). Standardized morphological identification of inflammatory cells was performed independently by two board-certified veterinary pathologists under double-blinded conditions^72,73. Histological images were acquired with an Olympus DP72 microscope equipped with a DS-Ri2 camera and analyzed using DPVIEW (v2.0.4) and ImageJ (v1.53t).

Single cell preparation and flow cytometric staining

Lymph nodes were passed through a 40 μm nylon membrane filter directly. After treatment with red blood cell lysis buffer (Solarbio, R1010, China), single cells were stimulated with 20 μg/mL PPD and cell stimulation cocktail (plus protein transport inhibitors, Thermo Fisher Scientific, 00-4975-93, USA) in RPMI-1640 media for 16 h in a 37 °C, 5% CO₂ humidified incubator. Stimulated cells were then washed and surface stained with anti-mouse CD4-FITC (Elabscience, E-AB-F1353C, clone RM4-5, China; 5 μL per test), and anti-mouse CD69-PE (Elabscience, E-AB-F1187D, clone H1.2F3, China; 5 μL per test). Subsequently, cells were fixed and permeabilized using intracellular fixation & permeabilization buffer set (Thermo Fisher Scientific, 88-8824-00, USA) as per the manufacturer’s instructions. Permeabilized cells were stained for intracellular cytokines using anti-mouse IFNγ-PECy7 (LIANKE, F21IFNG05, clone XMG1.2, China; 5 μL per test) and anti-mouse IL-17A-Elab Fluor^® 647 (Elabscience, E-AB-F1199M, clone TC11-18H10.1, China; 5 μL per test). All antibodies were used at a 1:20 dilution. Samples were acquired using the Fortessa X-20 and analyzed using FlowJo software (v10.9.0).

Immunoblot analysis

Cellular protein extraction was performed using RIPA lysis buffer (Beyotime, P0013B, China). Lysates were denatured in 1 × loading buffer (Solarbio, P1040, China) at 95 °C for 10 min and separated via SDS-PAGE. Separated proteins were electrotransferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, ISEQ00010, USA). Membranes were blocked with 5% bovine serum albumin in Tris-buffered saline containing 0.1% Tween-20 (TBST) for 1 h at 37 °C, then incubated with primary antibodies overnight at 4 °C. After three TBST washes, horseradish peroxidase (HRP)-conjugated secondary antibodies were applied for 1 h at 37 °C. Following three additional washes, immunoreactive bands were visualized using enhanced chemiluminescence reagent (Bio-Rad, USA).

Immunofluorescence and confocal microscopy

BMDMs were seeded on cover glasses in 24-well cell culture plates and were infected with M. bovis pre-labeled with 10 mg/mL FITC (Beyotime, ST2065, China) for 1 h. After 12 h, BMDMs were fixed, permeabilized, and blocked at room temperature. Anti-HIF-1α antibody was applied overnight at 4 °C, followed by incubation with YSFluor™594-conjugated secondary antibodies for 1 h at 37 °C. Nuclei were counterstained with DAPI (Beyotime, P0131, China). Confocal imaging was conducted on a Nikon A1HD25 system with NIS-Elements AR software.

Measurement of metabolites

The BMDMs were infected with M. bovis for 12 h and were scraped in 80% methanol. After phase separation (15,000 × g, 15 min, 4 °C), the methanol-water phase containing polar metabolites was separated and dried using a vacuum concentrator. The samples were re-dissolved in acetonitrile/water (1:1, v/v) and adequately vortexed and then centrifuged (13,500 × g, 15 min, 4 °C). The supernatant was transferred to an autosampler vial for LC-MS/MS analysis. Metabolomics data were obtained on an UHPLC (1290 Infinity LC, Agilent Technologies, USA) coupled to a QTRAP (AB Sciex 6500 + ). The Multiquant (v3.0.2) software was used to extract chromatographic peak area and retention time. Use the standard’s correct retention time to identify the metabolites. Lactate concentration was measured in supernatants using the L-Lactate Assay Kit with WST-8 (Beyotime, S0208S, China).

Enzyme-linked immunosorbent assay (ELISA)

Serums and supernatants were filtered through 0.22 µM Ultrafree centrifugal filters (Millipore, SLGP033R, USA). According to the manufacturer’s instructions, IFN-γ, IL-17A and IL-1β levels were measured by ELISA kit (NeoBioscience, EMC101g.96, EMC008.96, EMC001b.96, China).

RT-PCR analysis

Total RNA was extracted with 1 ml of Trizol reagent according to manufacturer instructions (Takara, 9108Q, Japan). The first-strand complementary DNA (cDNA) was synthesized using the HiScript III 1st Strand cDNA Synthesis Kit (Vazyme, R312-02, China) according to the manufacturer’s instructions. Lastly, qPCR analyses were performed with AceQ qPCR SYBR Green Master Mix (Vazyme, Q111-02, China) on a LightCycler 96 Instrument (Roche, Switzerland) using gene-specific primers (Supplementary Table 2).

Molecular docking analysis

The 3D structure of the FRD was predicted using AlphaFold2^74,75. The menaquinone-9 (MQ9) structure was extracted from the Mycobacterial respiratory complex I, fully-inserted quinone (PDB ID: 8E9H) using PyMOL (v2.6.2)^76,77. Molecular docking was performed with AutoDock (v4.2), where FRD served as the receptor and MQ9 as the ligand. Figures of molecular docking structures were generated using PyMOL (v2.6.2).

RNA sequencing

BMDMs were harvested 12 h post-infection with M. bovis, and RNA was harvested using Rneasy miniplus kit (Qiagen, Germany). 1.3 ug of total RNA per sample was used for library construction with SMARTER mRNA-Seg Library Prep Kit (Takara, Japan) and sequenced using the Illumina NovaSeq 6000 system, generating 150 bp paired-end reads. Raw data (raw reads) of fastq format were firstly processed through fastp (v0.20.1)⁶⁷ to obtain clean data with high quality. Clean reads were aligned to the Mus musculus reference genome (mm10) using HISAT2 (v2.0.5) with default splice-junction parameters. Gene-level quantification was performed with featureCounts (v1.5.0-p3), and differential expression analysis was conducted using DESeq2 (v1.20.0) with significance thresholds set at P value < 0.05.

Statistics and reproducibility

For pairwise comparisons, two-tailed unpaired Student’s t-tests were applied. Multi-group analyses utilized one-way ANOVA with Tukey’s multiple comparisons test. All statistical analyses were using GraphPad Prism (v9.4.1) software. In all cases, the data are presented as mean ± SD, and P value < 0.05 was statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Acknowledgements

This work was supported by the National Key Research and Development Program (project no. 2021YFD1800405 to X.M.Z.). The authors gratefully acknowledge colleagues from the State Key Laboratory of Veterinary Public Health and Safety of China Agricultural University for their valuable support.

Author contributions

Y.H.D. performed the experimental work. X.G. and Q.B.G. contributed to the acquisition and analysis of data. X.C.O., C.F.L., and W.X.F. were involved in setting up the study and in M. bovis isolate collection. Y.Z.W. and Z.Y.L. contributed to data interpretation. X.M.Z. acquired the funding for the study. Y.H.D. wrote the original draft. R.C.Y., Y.L.Z., and X.M.Z. reviewed and edited the manuscript. Y.L.Z. and X.M.Z. provided resources, insights and supervision.

Peer review

Peer review information

Nature Communications thanks Gillian Beamer, and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

The RNA-seq data generated in this study have been deposited in the Gene Expression Omnibus (GEO) database under accession number GSE291007. The metabolite profiling data generated in this study are provided in the Supplementary Data 1. Further information and requests for resources or reagents should be directed to and will be fulfilled by Yanlin Zhao ([email protected]) or Xiangmei Zhou ([email protected]).  are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-025-61824-9.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Gagneux, S. Ecology and evolution of Mycobacterium tuberculosis. Nat. Rev. Microbiol.; 2018; 16, pp. 202-213.1:CAS:528:DC%2BC1cXivVGiurY%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29456241]

2. Chiner-Oms, A et al. Genomic determinants of speciation and spread of the Mycobacterium tuberculosis complex. Sci. Adv.; 2019; 5, eaaw3307.2019SciA..5.3307C1:STN:280:DC%2BB3MrhsFaiuw%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31448322][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6691555]

3. World Health Organization. Global Tuberculosis Report 2024. Geneva, Switzerland: https://iris.who.int/bitstream/handle/10665/379339/9789240101531-eng.pdf?sequence=1(2024).

4. Zumla, A et al. Zoonotic tuberculosis-a call for an open One Health debate. Lancet Infect. Dis.; 2020; 20, pp. 642-644. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32213331]

5. World Health Organization. Global Tuberculosis Report 2020. Geneva, Switzerland: https://iris.who.int/bitstream/handle/10665/336069/9789240013131-eng.pdf?sequence=1(2020).

6. Orgeur, M et al. Evolution and emergence of Mycobacterium tuberculosis. FEMS Microbiol. Rev.; 2024; 48, fuae006.1:CAS:528:DC%2BB2cXitl2rtLvJ [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38365982][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10906988]

7. Sousa, J et al. Mycobacterium tuberculosis associated with severe tuberculosis evades cytosolic surveillance systems and modulates IL-1beta production. Nat. Commun.; 2020; 11, 2020NatCo.11.1949S1:CAS:528:DC%2BB3cXosVOqt70%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32327653][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7181847]1949.

8. Stanley, S et al. Ongoing evolution of the Mycobacterium tuberculosis lactate dehydrogenase reveals the pleiotropic effects of bacterial adaption to host pressure. PLoS Pathog.; 2024; 20, e1012050.1:CAS:528:DC%2BB2cXmtFyrs7g%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38422159][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10931510]

9. Ates, LS et al. Mutations in ppe38 block PE_PGRS secretion and increase virulence of Mycobacterium tuberculosis. Nat. Microbiol.; 2018; 3, pp. 181-188.1:CAS:528:DC%2BC1cXhtVKrtb4%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29335553]

10. Bottai, D et al. TbD1 deletion as a driver of the evolutionary success of modern epidemic Mycobacterium tuberculosis lineages. Nat. Commun.; 2020; 11, 2020NatCo.11.684B1:CAS:528:DC%2BB3cXjs1Oiur8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32019932][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7000671]684.

11. Vargas, R, Jr. et al. Phase variation as a major mechanism of adaptation in Mycobacterium tuberculosis complex. Proc. Natl. Acad. Sci. USA.; 2023; 120, e2301394120.1:CAS:528:DC%2BB3sXhsF2ktL%2FN [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37399390][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10334774]

12. Gupta, A et al. Reversible gene silencing through frameshift indels and frameshift scars provide adaptive plasticity for Mycobacterium tuberculosis. Nat. Commun.; 2021; 12, 2021NatCo.12.4702G1:CAS:528:DC%2BB3MXhslykt7fP [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34349104][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8339072]4702.

13. Worakitchanon, W et al. Comprehensive analysis of Mycobacterium tuberculosis genomes reveals genetic variations in bacterial virulence. Cell Host Microbe; 2024; 32, pp. 1972-1987.1:CAS:528:DC%2BB2cXitlCru7fO [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39471821]

14. Safi, H et al. Phase variation in Mycobacterium tuberculosis glpK produces transiently heritable drug tolerance. Proc. Natl. Acad. Sci. USA.; 2019; 116, pp. 19665-19674.2019PNAS.11619665S1:CAS:528:DC%2BC1MXhvVGnsrzP [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31488707][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6765255]

15. Vargas, R, Jr. et al. Antibiotic treatment and selection for glpK mutations in patients with active tuberculosis disease. Proc. Natl. Acad. Sci. USA.; 2020; 117, pp. 3910-3912.2020PNAS.117.3910V1:CAS:528:DC%2BB3cXjvFGksL0%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32075922][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7049102]

16. Bellerose, MM et al. Common variants in the glycerol kinase gene reduce tuberculosis drug efficacy. mBio; 2019; 10, pp. e00663-00619.1:CAS:528:DC%2BB3cXot1Kqs7k%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31363023][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6667613]

17. Scriba, TJ et al. Immunopathology in human tuberculosis. Sci. Immunol.; 2024; 9, eado5951.1:CAS:528:DC%2BB2MXhtF2ksbc%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39671470]

18. Chandra, P et al. Immune evasion and provocation by Mycobacterium tuberculosis. Nat. Rev. Microbiol.; 2022; 20, pp. 750-766.1:CAS:528:DC%2BB38XhvFGgtrrN [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35879556][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9310001]

19. Veyrier, FJ et al. The rise and fall of the Mycobacterium tuberculosis genome. Trends Microbiol.; 2011; 19, pp. 156-161.1:CAS:528:DC%2BC3MXkt1KhtL8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21277778]

20. Veyrier, F et al. Phylogenetic detection of horizontal gene transfer during the step-wise genesis of Mycobacterium tuberculosis. BMC Evol. Biol.; 2009; 9, 196. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19664275][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3087520]

21. Dong, Y et al. Genomic analysis of diversity, biogeography, and drug resistance in Mycobacterium bovis. Transbound. Emerg. Dis.; 2022; 69, pp. e2769-e2778. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35695307]

22. Watanabe, S et al. Fumarate reductase activity maintains an energized membrane in anaerobic Mycobacterium tuberculosis. PLoS Pathog.; 2011; 7, e1002287.1:CAS:528:DC%2BC3MXhtlOltLfO [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21998585][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3188519]

23. Olive, AJ et al. Metabolic crosstalk between host and pathogen: sensing, adapting and competing. Nat. Rev. Microbiol.; 2016; 14, pp. 221-234.1:CAS:528:DC%2BC28Xjs1ygt7g%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26949049]

24. Rosenberg, G et al. Immunometabolic crosstalk during bacterial infection. Nat. Microbiol.; 2022; 7, pp. 497-507.1:CAS:528:DC%2BB38XovV2qs78%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35365784]

25. Cheng, H et al. Mycobacterium tuberculosis produces D-serine under hypoxia to limit CD8(+) T cell-dependent immunity in mice. Nat. Microbiol.; 2024; 9, pp. 1856-1872.1:CAS:528:DC%2BB2cXhtFyktr7M [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38806671][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11222154]

26. Chernyaeva, EN et al. Genome-wide Mycobacterium tuberculosis variation (GMTV) database: a new tool for integrating sequence variations and epidemiology. BMC Genomics; 2014; 15, [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24767249][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4234438]308.

27. Iverson, TM et al. Structure of the Escherichia coli fumarate reductase respiratory complex. Science; 1999; 284, pp. 1961-1966.1:CAS:528:DyaK1MXktVWns7w%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10373108]

28. Dhiman, RK et al. Menaquinone synthesis is critical for maintaining mycobacterial viability during exponential growth and recovery from non-replicating persistence. Mol. Microbiol.; 2009; 72, pp. 85-97.1:CAS:528:DC%2BD1MXkslCktr0%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19220750][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4747042]

29. Lovey, A et al. Early alveolar macrophage response and IL-1R-dependent T cell priming determine transmissibility of Mycobacterium tuberculosis strains. Nat. Commun.; 2022; 13, 2022NatCo.13.884L1:CAS:528:DC%2BB38XjvV2murg%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35173157][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8850437]884.

30. Ritter, K et al. High-dose Mycobacterium tuberculosis H37Rv infection in IL-17A- and IL-17A/F-deficient mice. Cells; 2022; 11, 2875.1:CAS:528:DC%2BB38XisFemsbvK [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36139450][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9496946]

31. Chowdhury, CS et al. Type I IFN-mediated NET release promotes Mycobacterium tuberculosis replication and is associated with granuloma caseation. Cell Host Microbe; 2024; 32, pp. 2092-2111. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39637864]

32. Palmieri, EM et al. Nitric oxide orchestrates metabolic rewiring in M1 macrophages by targeting aconitase 2 and pyruvate dehydrogenase. Nat. Commun.; 2020; 11, 2020NatCo.11.698P1:CAS:528:DC%2BB3cXjs1Oiu74%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32019928][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7000728]698.

33. Lampropoulou, V et al. Itaconate links inhibition of succinate dehydrogenase with macrophage metabolic remodeling and regulation of inflammation. Cell Metab.; 2016; 24, pp. 158-166.1:CAS:528:DC%2BC28XhtFSgtLzF [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27374498][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5108454]

34. Selak, MA et al. Succinate links TCA cycle dysfunction to oncogenesis by inhibiting HIF-alpha prolyl hydroxylase. Cancer Cell; 2005; 7, pp. 77-85.1:CAS:528:DC%2BD2MXhsVait74%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15652751]

35. Tannahill, GM et al. Succinate is an inflammatory signal that induces IL-1beta through HIF-1alpha. Nature; 2013; 496, pp. 238-242.2013Natur.496.238T1:CAS:528:DC%2BC3sXkvFyjsrk%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23535595][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4031686]

36. Cohen, SB et al. Alveolar macrophages provide an early Mycobacterium tuberculosis niche and initiate dissemination. Cell Host Microbe; 2018; 24, pp. 439-446.1:CAS:528:DC%2BC1cXhsFykurnO [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30146391][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6152889]

37. Yang, H et al. Interception of host fatty acid metabolism by mycobacteria under hypoxia to suppress anti-TB immunity. Cell Discov.; 2021; 7, 90.1:CAS:528:DC%2BB3MXitFOltbrI [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34608123][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8490369]

38. Meehan, BM et al. Fumarate reductase is a major contributor to the generation of reactive oxygen species in the anaerobe Bacteroides fragilis. Microbiol. (Read.); 2012; 158, pp. 539-546.1:CAS:528:DC%2BC38XjsV2qtbg%3D

39. Imlay, JA. A metabolic enzyme that rapidly produces superoxide, fumarate reductase of Escherichia-Coli. J. Biol. Chem.; 1995; 270, pp. 19767-19777.1:CAS:528:DyaK2MXnvVCmtb8%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/7649986]

40. Adolph, C et al. Impaired succinate oxidation prevents growth and influences drug susceptibility in Mycobacterium tuberculosis. mBio; 2022; 13, e0167222. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35856639]

41. Li, AH et al. Characterization of genes differentially expressed within macrophages by virulent and attenuated Mycobacterium tuberculosis identifies candidate genes involved in intracellular growth. Microbiol. (Read.); 2008; 154, pp. 2291-2303.1:CAS:528:DC%2BD1cXhtVCrtLfN

42. Feng, S et al. Autophagy promotes efficient T cell responses to restrict high-dose Mycobacterium tuberculosis infection in mice. Nat. Microbiol.; 2024; 9, pp. 684-697.1:CAS:528:DC%2BB2cXkslahs7c%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38413834]

43. Esaulova, E et al. The immune landscape in tuberculosis reveals populations linked to disease and latency. Cell Host Microbe; 2021; 29, pp. 165-178.1:CAS:528:DC%2BB3MXkt12gsw%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33340449]

44. Moreira-Teixeira, L et al. Mouse transcriptome reveals potential signatures of protection and pathogenesis in human tuberculosis. Nat. Immunol.; 2020; 21, pp. 464-476.1:CAS:528:DC%2BB3cXltVSisbg%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32205882][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7116040]

45. Howard, NC et al. Immunometabolism during Mycobacterium tuberculosis Infection. Trends Microbiol.; 2020; 28, pp. 832-850.1:CAS:528:DC%2BB3cXovVChtL0%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32409147][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7494650]

46. Shi, L et al. Biphasic dynamics of macrophage immunometabolism during Mycobacterium tuberculosis infection. mBio; 2019; 10, pp. e02550-02518.1:CAS:528:DC%2BC1MXitlKgtL3L [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30914513][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6437057]

47. Osada-Oka, M et al. Metabolic adaptation to glycolysis is a basic defense mechanism of macrophages for Mycobacterium tuberculosis infection. Int. Immunol.; 2019; 31, pp. 781-793.1:CAS:528:DC%2BB3cXhtlertbnN [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31201418][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6839748]

48. Rodriguez-Prados, JC et al. Substrate fate in activated macrophages: a comparison between innate, classic, and alternative activation. J. Immunol.; 2010; 185, pp. 605-614.1:CAS:528:DC%2BC3cXns1Cqtro%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20498354]

49. Hackett, EE et al. Mycobacterium tuberculosis limits host glycolysis and IL-1beta by restriction of PFK-M via microRNA-21. Cell Rep.; 2020; 30, pp. 124-136 e124.1:CAS:528:DC%2BB3cXotFertLs%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31914380][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7764301]

50. Juffermans, NP et al. Interleukin-1 signaling is essential for host defense during murine pulmonary tuberculosis. J. Infect. Dis.; 2000; 182, pp. 902-908.1:CAS:528:DC%2BD3cXmvVCju7Y%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10950787]

51. Mayer-Barber, KD et al. Host-directed therapy of tuberculosis based on interleukin-1 and type I interferon crosstalk. Nature; 2014; 511, pp. 99-103.2014Natur.511..99M1:CAS:528:DC%2BC2cXhtV2qtbzO [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24990750][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4809146]

52. Zhang, G et al. Allele-specific induction of IL-1beta expression by C/EBPbeta and PU.1 contributes to increased tuberculosis susceptibility. PLoS Pathog.; 2014; 10, e1004426. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25329476][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4199770]

53. Mishra, BB et al. Nitric oxide prevents a pathogen-permissive granulocytic inflammation during tuberculosis. Nat. Microbiol.; 2017; 2, 17072.1:CAS:528:DC%2BC2sXnslOjsrs%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28504669][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5461879]

54. Shen, H et al. The crucial roles of Th17-related cytokines/signal pathways in M. tuberculosis infection. Cell Mol. Immunol.; 2018; 15, pp. 216-225.1:CAS:528:DC%2BC1cXjslOnsbs%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29176747]

55. Basile, JI et al. Outbreaks of Mycobacterium tuberculosis MDR strains induce high IL-17 T-cell response in patients with MDR tuberculosis that is closely associated with high antigen load. J. Infect. Dis.; 2011; 204, pp. 1054-1064.1:CAS:528:DC%2BC3MXhtFCjur%2FI [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21881121]

56. Mills, KHG. IL-17 and IL-17-producing cells in protection versus pathology. Nat. Rev. Immunol.; 2023; 23, pp. 38-54.1:CAS:528:DC%2BB38Xhslaitr7E [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35790881]

57. Pedrosa, J et al. Neutrophils play a protective nonphagocytic role in systemic Mycobacterium tuberculosis infection of mice. Infect. Immun.; 2000; 68, pp. 577-583.1:CAS:528:DC%2BD3cXotFOnuw%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10639420][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC97179]

58. Parker, HA et al. Antimicrobial activity of neutrophils against mycobacteria. Front. Immunol.; 2021; 12, 782495.1:CAS:528:DC%2BB38XhtlKgt7bL [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35003097][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8732375]

59. de Melo, MGM et al. Imbalance of NET and alpha-1-antitrypsin in tuberculosis patients is related with hyper inflammation and severe lung tissue damage. Front. Immunol.; 2018; 9, 3147. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30687336]

60. Muefong, CN et al. Neutrophils in tuberculosis-associated inflammation and lung pathology. Front. Immunol.; 2020; 11, 962.1:CAS:528:DC%2BB3cXitVWntr7K [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32536917][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7266980]

61. Lowe, DM et al. Neutrophils in tuberculosis: friend or foe?. Trends Immunol.; 2012; 33, pp. 14-25.1:CAS:528:DC%2BC38XlsVOhug%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22094048]

62. Dallenga, T et al. M. tuberculosis-induced necrosis of infected neutrophils promotes bacterial growth following phagocytosis by macrophages. Cell Host Microbe; 2017; 22, pp. 519-530 e513.1:CAS:528:DC%2BC2sXhs1Glsb7M [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29024644]

63. Dong, YH et al. Diversity of glpK gene and its effect on drug sensitivity in Mycobacterium bovis. Infect. Drug Resist.; 2022; 15, pp. 1467-1475.1:CAS:528:DC%2BB38Xht1Oksr7I [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35401008][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8986483]

64. Orsi, RH et al. Homopolymeric tracts represent a general regulatory mechanism in prokaryotes. BMC Genomics; 2010; 11, [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20144225][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2831843]102.

65. Gu, S et al. The beta2 clamp in the Mycobacterium tuberculosis DNA polymerase III alphabeta2epsilon replicase promotes polymerization and reduces exonuclease activity. Sci. Rep.; 2016; 6, 2016NatSR..618418G1:CAS:528:DC%2BC28XhslOmurk%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26822057][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4731781]18418.

66. Deng, MZ et al. An additional proofreader contributes to DNA replication fidelity in mycobacteria. Proc. Natl. Acad. Sci. Usa.; 2024; 121, e2322938121.1:CAS:528:DC%2BB2cXitVCmtLnJ [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39141351][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11348249]

67. Chen, S et al. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics; 2018; 34, pp. i884-i890. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30423086][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6129281]

68. Li, H et al. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics; 2009; 25, pp. 1754-1760.1:CAS:528:DC%2BD1MXot1Cjtbo%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19451168][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2705234]

69. Li, H. A statistical framework for SNP calling, mutation discovery, association mapping and population genetical parameter estimation from sequencing data. Bioinformatics; 2011; 27, pp. 2987-2993.1:CAS:528:DC%2BC3MXhtlGkur7L [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21903627][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3198575]

70. Walker, BJ et al. Pilon: an integrated tool for comprehensive microbial variant detection and genome assembly improvement. PLoS One; 2014; 9, e112963.2014PLoSO..9k2963W [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25409509][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4237348]

71. Jain, P et al. Specialized transduction designed for precise high-throughput unmarked deletions in Mycobacterium tuberculosis. mBio; 2014; 5, pp. e01245-01214. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24895308][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4049104]

72. Renne, R et al. Proliferative and nonproliferative lesions of the rat and mouse respiratory tract. Toxicol. Pathol.; 2009; 37, pp. 5S-73S.1:CAS:528:DC%2BC3cXmtFOkug%3D%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20032296]

73. Taylor, I. In A Practical Guide to the Histology of the Mouse (John Wiley & Sons, Ltd, Oxford, 2014).

74. Jumper, J et al. Highly accurate protein structure prediction with AlphaFold. Nature; 2021; 596, pp. 583-589.2021Natur.596.583J1:CAS:528:DC%2BB3MXhvVaktrrL [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34265844][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8371605]

75. Mirdita, M et al. ColabFold: making protein folding accessible to all. Nat. Methods; 2022; 19, pp. 679-682.1:CAS:528:DC%2BB38XhsVWjsLfL [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35637307][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9184281]

76. Liang, Y et al. Structure of mycobacterial respiratory complex I. Proc. Natl. Acad. Sci. USA.; 2023; 120, e2214949120.1:CAS:528:DC%2BB3sXot1Kgs7Y%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36952383][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10068793]

77. Yingke Liang, J. L. R. Mycobacterial respiratory complex I, fully inserted quinone. Protein Data Bank. https://www.rcsb.org/structure/8E9H (2022).

78. Jiang, S et al. Generic Diagramming Platform (GDP): a comprehensive database of high-quality biomedical graphics. Nucleic Acids Res.; 2025; 53, pp. D1670-D1676. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39470721]