ACA kills Mycobacterium tuberculosis and M. bovis by targeting cell wall core assembling protein CpsA Authors: Feng Huang, Mengcong Shi, Liu Chen, Xiaohui Li, Zixin Wang, Dong Jiang, Junfeng Chen, Hua Zhang, Zheng-Guo He Published: 2025 ABSTRACT Tuberculosis (TB), caused by Mycobacterium tuberculosis, still globally threatens human health. With the emergence of multidrug-resistant strains, there is an imperative need to discover new anti-TB agents and identify novel drug targets. Here, we report that the compound ACA, namely 3-azidothiophene-2-carboxylic acid, can kill M. tuberculosis by targeting the bacterial cell wall core assembling process. Whole-genome sequencing of spontaneous ACA-resistant mutants identified single-nucleotide variants in the cpsA2 ligase gene responsible for the covalent attachment of arabinogalactan and peptidoglycan. The cell wall of cpsA2 or its homolog cpsA1-deleted mutant H37Ra strains show increased permeability and drug sensitivity, which is similar to the ACA-treated mycobacteria. Both cpsA1 and cpsA2 could reverse the resistant phenotype of ACA-resistant mutants and the growth defects of the ΔcpsA1 H37Ra strain. ACA can directly bind CpsA1 and CpsA2 to inhibit significantly the pyrophosphatase activity of CpsA1 and CpsA2. Our results suggest that ACA may disturb the cell wall core assembling process and kill M. tuberculosis by targeting CpsA1 and CpsA2, which provide potential candidate drug target for controlling drug-resistant TB. Introduction Mycobacterium tuberculosis (Mtb), an ancient intracellular bacterial pathogen, still poses a major threat to global public health. Tuberculosis (TB), caused by Mtb, is the world’s second leading cause of death from a single infectious agent, following coronavirus disease (COVID-19). It accounted for nearly twice as many deaths as HIV/AIDS, totaling 1.3 million fatalities in 20221. The emergence of multidrug-resistant tuberculosis (MDR-TB) and extensively drug-resistant tuberculosis (XDR-TB) poses a major challenge to current treatment regimens and drug efficacy, underscoring the urgent need for novel anti-TB agents. In the past decade, only three drugs with innovative mechanisms of action—bedaquiline (BDQ), delamanid (DLM), and pretomanid (PMD)—have been approved for TB treatment2. Bedaquiline, a diarylquinoline, targets bacterial ATP synthase3 and is approved by the Food and Drug Administration (FDA) of the United States for the treatment of MDR-TB and XDR-TB when no other treatment options are available4. PMD is a nitroimidazole derivative similar to DLM. It is the third and most recently approved drug as a part of the BPaL regimen, which is applied to adults with XDR-TB or non-responsive or drug-intolerant MDR-TB5. The mycobacterial cell envelope, which serves as the interface with the environment, plays a critical role in bacterial growth6,7. It features a complex, multilayered assembly structure consisting of the cytoplasmic membrane, cell wall, surface lipids, a capsule, and numerous proteins, which collectively contribute to its functional properties6–9. Among these components, the cell wall is particularly notable for its unusual composition dominated by lipids and carbohydrates, which creates a permeability barrier that protects against hydrophilic drugs and is crucial for its survival and virulence10–12. The core structure of the mycobacterial cell wall comprises three key elements: a peptidoglycan polymer, a highly branched arabinogalactan, and long-chain mycolic acids13,14. As such, its biosynthesis pathway represents one of the few well-validated targets in anti-TB drug development. Isoniazid (INH) targets the synthesis of mycolic acid through inhibition of InhA15. Ethambutol (EMB) inhibits arabinosyltransferase, interfering with the synthesis of arabinogalactan and lipoarabinomannan, which affects the integrity of cell wall16. Although targeting cell wall assembly has become an attractive strategy for developing chemotherapeutics against TB10,12,17, no anti-TB compounds that directly target this process have been reported to date. The LytR-CpsA-Psr (LCP) family proteins are putative transmembrane proteins, and a typical LCP protein is composed of three distinct parts: a very short intracellular N-terminal region; a transmembrane helix; and a large extracellular portion, which contains the LCP domain18. They are involved in the cell wall core assembly that attaches the arabinogalactans (AGs) to the peptidoglycan (PG), presenting an essential antibiotic target19. Moreover, the LCP family members of the cell envelope-associated attenuators have been demonstrated to be implicated in multiple physiological processes, including β-lactam resistance, biofilm formation, various virulence factors, cell division, and cell wall maintenance20–24. The Mtb genome encodes three LCP candidate proteins: CpsA1, CpsA2, and Rv0822c. CpsA1 and CpsA2 are responsible for the covalent attachment of two major heteropolysaccharides, AG and PG, with CpsA1 playing the predominant role in this process25,26. Deletion of cpsA1 significantly inhibits the growth rate of Mtb and makes it sensitive to several antibiotics25. By contrast, the cpsA2 inactivation did not affect the bacterial growth. However, the combined inactivation of cpsA1 and cpsA2 is lethal25. CpsA2 has been shown to inhibit the host NADPH oxidase and LAP protein, antagonizing the innate immunity response27. It also enhances Mtb virulence in vivo, and this function cannot be replaced by CpsA128. Moreover, the CpsA protein, homologous to CpsA2, is indispensable for maintaining the cell wall integrity and the pathogenesis of Mycobacterium marinum18. Despite these, no known drugs are designed to target CpsA1 and CpsA2. Previous studies have demonstrated that 3-azidothiophene-2-carboxylic acid (ACA, C₅H₃N₃O₂S) effectively inhibits the growth of both M. tuberculosis H37Rv and a clinically isolated multi-drug resistant (MDR) M. tuberculosis strain29. However, the precise mechanism by which ACA exerts its antibacterial activity remains unclear. In this study, we identified ACA as a bactericidal compound against Mtb and M. bovis by targeting cell wall core assembling protein CpsA. ACA binds to CpsA1 and CpsA2, disturbing the ligation of arabinogalactan and peptidoglycan, which impairs the mycobacterial cell wall assembly and reduces bacterial growth. Interestingly, ACA enhances the bactericidal efficacy of other anti-mycobacterial drugs. Our study provides insights for developing anti-TB agents. Results ACA inhibits the growth of Mtb and M. bovis Previous screenings based on protein structure and growth inhibition in our lab found that ACA is a potential bactericidal compound against M. tuberculosis29. To further confirm this observation, we comparatively evaluated the effects of ACA on the growth of various mycobacterial and non-mycobacterial strains by determining the minimum inhibitory concentration (MIC). Exactly, ACA effectively inhibits the growth of M. bovis BCG (MIC: 17.7 μM) and M. tuberculosis H37Ra (MIC: 35.4 μM) (Table 1). By contrast, ACA exhibited a MIC of 141.9 μM against M. smegmatis MC2 155, and no antibacterial activity was observed for M. marinum and non-mycobacterial species, including Escherichia coli, Staphylococcus aureus, and Bacillus thuringiensis, under similar experiment conditions (Table 1). These results suggest that ACA possesses a specific activity for the slow-growing mycobacterial species, such as M. tuberculosis and M. bovis.Table 1The antibacterial activity of ACA against the selected microorganismsOrganism and genotypeMIC (μg/ml)MIC (μM)M. bovis BCG Pasteur 1173P2317.7M. TB H37Ra635.5M. smegmatis mc2 15524141.9ACA-resistant M. TB H37Ra (AR01)32189.2M. marinum M- ATCC #BAA-535>96>567.5Escherichia coli>96>567.5Staphylococcus aureus>96>567.5Bacillus thuringiensis>96>567.5 Table 1The antibacterial activity of ACA against the selected microorganismsOrganism and genotypeMIC (μg/ml)MIC (μM)M. bovis BCG Pasteur 1173P2317.7M. TB H37Ra635.5M. smegmatis mc2 15524141.9ACA-resistant M. TB H37Ra (AR01)32189.2M. marinum M- ATCC #BAA-535>96>567.5Escherichia coli>96>567.5Staphylococcus aureus>96>567.5Bacillus thuringiensis>96>567.5 The antibacterial activity of ACA against the selected microorganisms The kill kinetics of ACA for the BCG and H37Ra strains were further assayed to investigate the effect of this compound on mycobacterial viability. As shown in Fig. 1A, ACA exhibited a dose-dependent mycobactericidal activity, reducing the colony-forming units (CFUs) by approximately 10,000 on day 4 at 5× MIC (Fig. 1A and Supplementary Fig. 1A). And H37Ra strain exhibits a sensitivity to ACA similar to that of the BCG strain (Supplementary Fig. 1B). Subsequently, we analyzed the cytotoxicity of ACA to the bone marrow-derived macrophage cells (BMDMs) and assessed its anti-mycobacterial activity within these cells. As shown in Fig. 1B–D, ACA showed no obvious cytotoxicity (Fig. 1B) and reduced the intracellular survival of both BCG and H37Ra strains in BMDMs (Fig. 1C, D).Fig. 1ACA inhibits mycobacterial growth in vitro and in vivo.A Killing kinetics of ACA on M. bovis BCG. The chemical structure of 3-azidothiophene-2-carboxylic acid (ACA) was rendered using ChemDraw software. B LDH cytotoxicity analysis of ACA in BMDMs. BMDMs (5×104 cells per well) were seeded in a 96-well plate and incubated overnight at 37 °C with 5% CO2. The following day, different concentrations of ACA (10×, 20×, 50× MIC of M. bovis BCG) were added to the culture media and further incubated for 18 h. 14.6 μM INH (50× MIC of M. bovis BCG) was used as a control. LDH release was quantified using the Pierce LDH Cytotoxicity Assay Kit. C,D Assays for the survival of M. bovis BCG (MOI = 1) C or M. tuberculosis H37Ra (MOI = 10) D in BMDMs under different treatment conditions mentioned in (B). White bars represent ACA has no effect on intracellular survival of mycobacteria, while blue bars represent ACA has modest inhibition of mycobacterial intracellular survival in BMDMs. Data represent the mean ± SD from three independent experiments. Asterisks indicate significant differences as determined by an unpaired two-tailed Student’s t-test (*, P < 0.05; **, P < 0.01) using GraphPad Prism 5. ACA inhibits mycobacterial growth in vitro and in vivo. A Killing kinetics of ACA on M. bovis BCG. The chemical structure of 3-azidothiophene-2-carboxylic acid (ACA) was rendered using ChemDraw software. B LDH cytotoxicity analysis of ACA in BMDMs. BMDMs (5×104 cells per well) were seeded in a 96-well plate and incubated overnight at 37 °C with 5% CO2. The following day, different concentrations of ACA (10×, 20×, 50× MIC of M. bovis BCG) were added to the culture media and further incubated for 18 h. 14.6 μM INH (50× MIC of M. bovis BCG) was used as a control. LDH release was quantified using the Pierce LDH Cytotoxicity Assay Kit. C,D Assays for the survival of M. bovis BCG (MOI = 1) C or M. tuberculosis H37Ra (MOI = 10) D in BMDMs under different treatment conditions mentioned in (B). White bars represent ACA has no effect on intracellular survival of mycobacteria, while blue bars represent ACA has modest inhibition of mycobacterial intracellular survival in BMDMs. Data represent the mean ± SD from three independent experiments. Asterisks indicate significant differences as determined by an unpaired two-tailed Student’s t-test (*, P < 0.05; **, P < 0.01) using GraphPad Prism 5. Collectively, these results indicated that ACA effectively inhibits the growth of M. TB and M. bovis both in vitro and in vivo in BMDMs. cpsA2 mutation results in mycobacterial resistance to ACA Although the compound ACA was initially screened based on the structure of Rv2780, our subsequent research found that ACA does not target Rv2780 directly. To pursue the potential mycobactericidal mechanism of ACA, we next tried to screen and isolate the ACA-resistant mutant strains. M. TB H37Ra cultures (1 × 1010 CFUs) were plated on 7H10 agar plates containing 5× and 10× MIC of ACA. After several screening cycles, we finally isolated two mutant strains, AR01 and AR02, which demonstrated a more than 5-fold increase in MIC of ACA (189.2 μM) in 7H9 liquid medium (Table 1). Next, we measured the growth difference between the wild-type and mutant H37Ra strains on 7H10 medium supplemented with or without 236.5 μM ACA. As shown in the left panel of Supplementary Fig. 2A, no obvious growth difference was observed among the AR01, AR02, and wild-type strains. By contrast, both AR01 and AR02 strains grew much better than the wild-type strain under 236.5 μM ACA treatment condition (Supplementary Fig. 2A, right panel). Similarly, both mutant strains showed better growth than the wild-type strain in 7H9 medium containing 70.9 and 141.9 μM ACA (Fig. 2A).Fig. 2cpsA2 mutation is responsible for mycobacterial resistance to ACA.A Analysis of the ACA resistance of mutant strains AR01 and AR02. WT and ACA-resistant strains were cultured in 7H9 medium containing 0, 70.9, 141.9 μM ACA at 37 °C for 8 days, and then CFUs were determined. The structure of CpsA2 was predicted by AlphaFold v2 software, the red area in the middle of the right panel indicates the mutation site V221E. B Assays for the effects of CpsA2 on the ACA resistance of mutant strain AR01. WT, ACA-resistant strain AR01, and cpsA2-complemented strain were cultured in 7H9 medium supplemented with or without 70.9 μM ACA at 37 °C for 8 days. Subsequently, CFUs were quantified. C The ACA resistance of the cpsA2-deleted strain. WT, cpsA2-deleted strain and cpsA2-complemented strain were cultivated in 7H9 medium in the presence and absence of 70.9 μM at 37 °C for 9 days, followed by CFUs quantification. Data represent the mean ± SD from three independent experiments. Asterisks indicate significant differences as determined by an unpaired two-tailed Student’s t-test (*, P < 0.05; **, P < 0.01; ***, P < 0.001) using GraphPad Prism 5. cpsA2 mutation is responsible for mycobacterial resistance to ACA. A Analysis of the ACA resistance of mutant strains AR01 and AR02. WT and ACA-resistant strains were cultured in 7H9 medium containing 0, 70.9, 141.9 μM ACA at 37 °C for 8 days, and then CFUs were determined. The structure of CpsA2 was predicted by AlphaFold v2 software, the red area in the middle of the right panel indicates the mutation site V221E. B Assays for the effects of CpsA2 on the ACA resistance of mutant strain AR01. WT, ACA-resistant strain AR01, and cpsA2-complemented strain were cultured in 7H9 medium supplemented with or without 70.9 μM ACA at 37 °C for 8 days. Subsequently, CFUs were quantified. C The ACA resistance of the cpsA2-deleted strain. WT, cpsA2-deleted strain and cpsA2-complemented strain were cultivated in 7H9 medium in the presence and absence of 70.9 μM at 37 °C for 9 days, followed by CFUs quantification. Data represent the mean ± SD from three independent experiments. Asterisks indicate significant differences as determined by an unpaired two-tailed Student’s t-test (*, P < 0.05; **, P < 0.01; ***, P < 0.001) using GraphPad Prism 5. To identify the potential proteins targeted by ACA, whole-genome sequencing was performed on the wild-type and mutant strains of H37Ra using the Illumina sequencing technology. Point mutations that conferred ACA-resistance were then identified by comparing the genome sequences of the susceptible wild-type H37Ra strain with two resistant mutants (Supplementary Table 3). Interestingly, the common mutant gene within the genome of the AR01 and AR02 mutants encodes an LCP family protein CpsA2 (Rv3484) that is a cell wall ligase responsible for the covalent attachment of AG to PG; the point mutation site of CpsA2 was identified to be V221E (Table 2; Fig. 2A, right panel), which is further validated by Sanger sequencing. In addition to CpsA2, the mutant strain AR01 also contains another mutation site, UvrA-E379K (Table 2). We overexpressed uvrA in AR01 and found that UvrA could not restore the sensitivity of AR01 to ACA.Table 2Single nucleotide polymorphisms identified in the ACA-resistant H37Ra isolateSample NameSNP descriptionLocus tagGene nameSNP/Indel classCodon changeAA changeGene lengthAR01T3903739ARv3484/BCG_3548CpsA2MissensegTg/gAgV221E1.539KbAR01G1844929ARv1638/BCG_1676UvrAMissenseGag/AagE397K2.919KbAR02T3903739ARv3484/BCG_3548CpsA2MissensegTg/gAgV221E1.539Kb Table 2Single nucleotide polymorphisms identified in the ACA-resistant H37Ra isolateSample NameSNP descriptionLocus tagGene nameSNP/Indel classCodon changeAA changeGene lengthAR01T3903739ARv3484/BCG_3548CpsA2MissensegTg/gAgV221E1.539KbAR01G1844929ARv1638/BCG_1676UvrAMissenseGag/AagE397K2.919KbAR02T3903739ARv3484/BCG_3548CpsA2MissensegTg/gAgV221E1.539Kb Single nucleotide polymorphisms identified in the ACA-resistant H37Ra isolate To further explore the role of CpsA2 on the mycobacterial resistance to ACA, we transformed a pMV261 vector overexpressing the wild-type gene cpsA2 into the AR01 strain to obtain the cpsA2 complementation strain, and then determined the resistance of these recombinant strains to ACA. As shown in Fig. 2B, the cpsA2-complemented strain regained partial sensitivity to 70.9 μM of ACA. To further confirm this observation, the cpsA2-deleted H37Ra strain (H37RaΔcpsA2) (Supplementary Fig. 3A) and the cpsA2 complementation strain (ΔcpsA2/pMV361-cpsA2p-cpsA2) were constructed and subsequently cultured in the 7H9 medium with and without 70.9 μM μg/mL ACA. As shown in Fig. 2C, the cpsA2-deleted H37Ra strain grew better than the wild type H37Ra strain under 70.9 μM of ACA condition, and this phenotype was reversed when cpsA2 was complemented in the cpsA2-deleted strain (Fig. 2C, right panel); however, no obvious difference was observed in the growth of these strains in the absence of ACA (Fig. 2C, left panel), suggesting that the cpsA2 gene mutation led to the mycobacterial resistance to ACA. Taken together, these results indicated that cpsA2 mutation results in the ACA-resistance of H37Ra. CpsA1 can reverse the ACA-resistant phenotype of ΔcpsA2 mutant strain The Mtb genome contains three genes encoding LCP-like proteins: Rv3267 (CpsA1), CpsA2, and Rv0822c. Both CpsA1 and CpsA2 feature a short cytoplasmic domain, followed by a single transmembrane (TM) segment and a periplasmic catalytic domain (Fig. 3A), while Rv0822c displays a significantly longer N-terminal domain and lacks an obvious TM helix domain24. CpsA1 belongs to an AG-biosynthetic gene cluster, which was hypothesized responsible for ligating PG to AG in mycobacteria19. To determine the functional relationship of these proteins, we constructed the cpsA1 and rv0822c complementation strains in which cpsA1 and rv0822c separately were over-expressed in the AR01 mutant using the above-mentioned methods, and then assayed their resistance to ACA. As shown in Supplementary Fig. 2B and Fig. 3B, the cpsA1-complemented strain obtained a clear ACA sensitivity, similar to the phenotype of the wild-type strain. In contrast, Rv0822c hardly affected the resistant phenotype of the AR01 mutant. These results suggest that CpsA1 can reverse the defect phenotype of the AR01 mutant (Supplementary Fig. 2B and Fig. 3B).Fig. 3CpsA1 restores the ACA sensitivity of AR01 strain.A Domain organization of LCP proteins in M. tuberculosis H37Ra. Both CpsA1 and CpsA2 feature a short cytoplasmic domain, followed by a single TM segment and a periplasmic catalytic domain based on TMHMM server version 2.0 predictions. Rv0822c displays a long domain but lacks a clear TM helix using the above-mentioned prediction. B Assays for the effect of CpsA1 on the ACA sensitivity of the AR01 strain. Mycobacterial strains were cultured in 7H9 medium without or with 70.9 μM. ACA at 37 °C for 7 days. Subsequently, CFUs were determined. Data represent the mean ± SD from three independent experiments. CpsA1 restores the ACA sensitivity of AR01 strain. A Domain organization of LCP proteins in M. tuberculosis H37Ra. Both CpsA1 and CpsA2 feature a short cytoplasmic domain, followed by a single TM segment and a periplasmic catalytic domain based on TMHMM server version 2.0 predictions. Rv0822c displays a long domain but lacks a clear TM helix using the above-mentioned prediction. B Assays for the effect of CpsA1 on the ACA sensitivity of the AR01 strain. Mycobacterial strains were cultured in 7H9 medium without or with 70.9 μM. ACA at 37 °C for 7 days. Subsequently, CFUs were determined. Data represent the mean ± SD from three independent experiments. To confirm the observation, we further constructed the cpsA1-deleted strains of H37Ra (H37RaΔcpsA1) and BCG (BCGΔcpsA1) (Supplementary Fig. 3B) and their corresponding complementation strains, and then determined their differential resistance to ACA. As shown in Fig. 4A and Supplementary Fig. 2C, the H37RaΔcpsA1 strain showed an obvious growth defect compared to both the wild-type strain and the cpsA1/cpsA2 complementation strains in the absence of ACA. Interestingly, the deletion of cpsA1 did not affect the growth of the BCG strain under the similar experiment condition (Supplementary Fig. 4A). In contrast, cpsA1 deletion promoted the mycobacterial growth under 118.2 μM ACA treatment condition, and complementation of cpsA1 partially reversed this phenotype (Supplementary Fig. 4B), suggesting that CpsA1 indeed impairs the mycobacterial resistance to ACA.Fig. 4CpsA1 and CpsA2 exhibit functional similarity.A The effects of CpsA1 and CpsA2 on the growth of the cpsA1-deleted strain. WT, cpsA1-deleted and cpsA1-complemented strains were cultured in 7H9 medium at 37 °C for 7 days, and then CFUs were determined. B Analysis of the roles of CpsA1 and CpsA2 in the ACA sensitivity of the cpsA2-deleted strain. Mycobacterial strains were cultivated in 7H9 medium supplemented without or with 70.9 μM ACA at 37 °C for 7 days, followed by CFUs quantification. Data represent the mean ± SD from three independent experiments. Asterisks denote significant differences (***, P < 0.001) determined by an unpaired two-tailed Student’s t-test using GraphPad Prism 5. CpsA1 and CpsA2 exhibit functional similarity. A The effects of CpsA1 and CpsA2 on the growth of the cpsA1-deleted strain. WT, cpsA1-deleted and cpsA1-complemented strains were cultured in 7H9 medium at 37 °C for 7 days, and then CFUs were determined. B Analysis of the roles of CpsA1 and CpsA2 in the ACA sensitivity of the cpsA2-deleted strain. Mycobacterial strains were cultivated in 7H9 medium supplemented without or with 70.9 μM ACA at 37 °C for 7 days, followed by CFUs quantification. Data represent the mean ± SD from three independent experiments. Asterisks denote significant differences (***, P < 0.001) determined by an unpaired two-tailed Student’s t-test using GraphPad Prism 5. Next, we evaluated the effect of CpsA1 on the ACA resistance of the cpsA2-deleted strain. As shown in the right panels of Fig. 4B and Supplementary Fig. 2D, the cpsA1/cpsA2-complemented strains and the wild-type strain exhibited significantly poorer growth compared to the cpsA2-deleted strain in the presence of ACA; however, no obvious growth difference among these strains was observed in the absence of ACA (Fig. 4B and Supplementary Fig. 2D, left panels). Collectively, these results suggest that the CpsA1 expression can reverse the ACA-resistant phenotype of cpsA2-deleted mutant Mtb. ACA inhibits the pyrophosphatase activities of both CpsA1 and CpsA2 LCP proteins usually have pyrophosphatase activity and can specifically cleave the pyrophosphate group to release free phosphate from geranyl pyrophosphate (GPP)24. To determine whether ACA has an effect on the pyrophosphatase activity of LCP protein, we first assayed the pyrophosphatase activity of both CpsA1 and CpsA2 proteins that lacked the TM domain. As shown in Fig. 5A, the CpsA1 and CpsA2 mutants showed time-dependent and protein concentration-dependent pyrophosphatase activity. Subsequently, the effect of ACA on the pyrophosphatase activity of CpsA1 and CpsA2 was determined, respectively. As shown in Fig. 5B, 2 μg/mL (11.8 μM) ACA obviously inhibited the CpsA1 activity, and 40 μg/mL (236.5 μM) ACA showed significant inhibitory activity on CpsA2. In contrast, neither 20 nor 40 μg/mL cordycepin, used as a negative control compound, affected the activities of CpsA1 and CpsA2 under a similar condition, respectively (Fig. 5B).Fig. 5ACA binds to CpsA1 and inhibits the pyrophosphatase activity of CpsA1 and CpsA2.A Assay for the pyrophosphatase activity of CpsA1 and CpsA2. 10 μg or 20 μg of recombinant proteins (CpsA1ΔTM and CpsA2ΔTM) were added into the reaction mixtures and incubated at 37 °C for 3 h and 16 h. B Analysis of the effects of ACA on the pyrophosphatase activity of CpsA1 and CpsA2. Reaction mixtures containing 20 μg of recombinant proteins were treated with varying concentrations of ACA and incubated at 37 °C for 16 h. Cordycepin was used as a negative control. C DARTS assays for the binding ability of CpsA1 to ACA. CpsA1 was incubated with 0.5 mg/mL (2.955 mM) ACA or 0.5 mg/mL (2.447 mM) EMB (a negative control) for 30 min on ice, followed by digestion with 15 μg of pronase E for 20 min at 37 °C. Dehydroquinate synthase (DHQS) was used as a negative control protein. The reaction samples were analyzed using 12% w/v SDS-PAGE, and protein degradation was quantified relative to the undigested control using Image J. Data represent the mean ± SD from three independent experiments. Asterisks indicate significant differences (*, P < 0.05; **, P < 0.01; ***, P < 0.001) determined by an unpaired two-tailed Student’s t-test using GraphPad Prism 5. ACA binds to CpsA1 and inhibits the pyrophosphatase activity of CpsA1 and CpsA2. A Assay for the pyrophosphatase activity of CpsA1 and CpsA2. 10 μg or 20 μg of recombinant proteins (CpsA1ΔTM and CpsA2ΔTM) were added into the reaction mixtures and incubated at 37 °C for 3 h and 16 h. B Analysis of the effects of ACA on the pyrophosphatase activity of CpsA1 and CpsA2. Reaction mixtures containing 20 μg of recombinant proteins were treated with varying concentrations of ACA and incubated at 37 °C for 16 h. Cordycepin was used as a negative control. C DARTS assays for the binding ability of CpsA1 to ACA. CpsA1 was incubated with 0.5 mg/mL (2.955 mM) ACA or 0.5 mg/mL (2.447 mM) EMB (a negative control) for 30 min on ice, followed by digestion with 15 μg of pronase E for 20 min at 37 °C. Dehydroquinate synthase (DHQS) was used as a negative control protein. The reaction samples were analyzed using 12% w/v SDS-PAGE, and protein degradation was quantified relative to the undigested control using Image J. Data represent the mean ± SD from three independent experiments. Asterisks indicate significant differences (*, P < 0.05; **, P < 0.01; ***, P < 0.001) determined by an unpaired two-tailed Student’s t-test using GraphPad Prism 5. To further quantitatively evaluate the inhibitory effect of ACA on the CpsA proteins, we first determined the Michaelis constants (Km) of both CpsA1 and CpsA2. The values of Km were determined by a double reciprocal Lineweaver–Burke plot of 1/v against 1/[S], yielding the Km value for CpsA1 and CpsA2, were 115 μM and 440 μM, respectively (Supplementary Fig. 5A and Supplementary Fig. 5B). The IC50 value of ACA on pyrophosphatase activity of CpsA1 and CpsA2 was also examined. The results showed that CpsA1 and CpsA2 was inhibited by ACA in a dose-dependent manner and the IC50 was determined to be 419.3 μM and 410 μM, respectively (Supplementary Fig. 5C and Supplementary Fig. 5D). Next, we examined if there exists a direct physical interaction between ACA and CpsA1 or CpsA2. However, due to the compound’s nature itself, ACA can associate with Surface Plasmon Resonance (SPR) chips or is not compatible with the Isothermal Titration Calorimetry (ITC) system. Alternatively, we utilized the drug affinity response target stability (DARTS) assays to detect the binding ability of ACA to the CpsA1 and CpsA2 proteins. As shown in Fig. 5C, CpsA1 alone was digested by pronase E with ~30% CpsA1 remaining undigested, whereas the proteolysis of CpsA1 was decreased with remaining amount of ~60% when co-incubated with ACA, indicating that ACA obviously protected CpsA1 from degradation by pronase E. In contrast, ethambutol (EMB) did not provide similar protection to CpsA1 (Fig. 5C, lanes 1–4), and ACA exhibited no protective effect on the dehydrogenate synthase (DHQS) (Fig. 5C, lanes 5–7), implying that ACA, but not EMB, can specially bind to CpsA1. Additionally, the direct binding ability of ACA to CpsA2 was also observed under the similar experimental conditions (Supplementary Fig. 5). Taken together, these results suggest that ACA can directly bind to CpsA1 and CpsA2, thereby inhibiting their pyrophosphatase activities. ACA disturbs the mycobacterial cell wall synthesis CpsA1 and CpsA2 are implicated to play a critical role in the cell wall core assembly via attaching the AG to the PG25. To test this idea, we assayed the roles of CpsA1 and CpsA2 on the cell wall structure of H37Ra using the acid-fast staining experiment. As shown in Supplementary Fig. 6A, both ΔcpsA1 and ΔcpsA2 strains retained much less red stain than that of the wild-type and complementation strains, suggesting that CpsA1 and CpsA2 significantly affected the cell wall integrity of M. TB H37Ra. Further, we evaluated the effect of ACA on the mycobacterial cell wall using the same experiment. As shown in Supplementary Fig. 6B and C, ACA-treated strains of either H37Ra or BCG lost the acid-fast staining ability, which is consistent with the phenotype of the cpsA1/cpsA2-deleted strains. We further utilized scanning electron microscopy (SEM) to assay the effects of ACA and CpsA1/CpsA2 on the cell wall structure of M. TB H37Ra. As shown in Fig. 6A, both the ACA-treated strain and the cpsA1/cpsA2-deleted strain exhibited obvious wrinkles on the cell surface. In contrast, the surfaces of the wild-type and complementation strains maintained relatively smooth in the absence of ACA (Fig. 6A). Interestingly, the cell length of the ACA-treated strains of H37Ra and BCG was also observed to be significantly increased compared to the untreated strains (Supplementary Fig. 7).Fig. 6Both ACA and CpsA1/CpsA2 affect mycobacterial cell morphology, cell wall structure and sensitivity to antibiotics.H37Ra strains were cultured in 7H9 medium up to a mid-logarithmic growth phase in the absence and presence of 59.1 μM ACA (ACA-treated). Subsequently, the cell morphology and cell envelope of mycobacterial strains was analyzed using SEM (A) and TEM (B). Representative image in SEM was captured at 25,000× magnification with 2 μm scale bars. Red and black arrows in TEM image represent the cytoplasmic membrane (CM) and outer membrane (OM), respectively. Scale bar, 100 nm. C Analysis of the mycobactericidal efficiency of ACA, vancomycin, and their combination. Data represent the mean ± SD from three independent experiments. Both ACA and CpsA1/CpsA2 affect mycobacterial cell morphology, cell wall structure and sensitivity to antibiotics. H37Ra strains were cultured in 7H9 medium up to a mid-logarithmic growth phase in the absence and presence of 59.1 μM ACA (ACA-treated). Subsequently, the cell morphology and cell envelope of mycobacterial strains was analyzed using SEM (A) and TEM (B). Representative image in SEM was captured at 25,000× magnification with 2 μm scale bars. Red and black arrows in TEM image represent the cytoplasmic membrane (CM) and outer membrane (OM), respectively. Scale bar, 100 nm. C Analysis of the mycobactericidal efficiency of ACA, vancomycin, and their combination. Data represent the mean ± SD from three independent experiments. Transmission electron microscopy was further used to visually observe the structural change in the mycobacterial cell surface. As shown in Fig. 6B, H37Ra/ΔcpsA1 exhibited a heterogeneous population of cells, with some cells displaying a damaged outer layer and large fragments detaching from the cell surface. Similarly, the ACA-treated strain of H37Ra also showed the cell envelope defects (Fig. 6B). In contrast, both the wild type and cpsA1-complemented mycobacterial cells were homogeneous and presented a typical cell envelope with a continuous outer membrane in the absence of ACA (Fig. 6B). Given that ACA impairs the cell wall structure and increases cell permeability, we investigated whether ACA-treated mycobacteria exhibited increased sensitivity to antibiotics. To test this hypothesis, we determined the bactericidal effect of vancomycin (VAN) against H37Ra in the presence and absence of ACA. As shown in Fig. 6C, the mycobacterial growth was significantly worse under the combined treatment of ACA (29.6 μM) and VAN (0.276 μM) compared to either drug alone. These results indicated that ACA enhances the bactericidal activity of VAN in Mtb. To determine whether the combined use of ACA with VAN or other anti-TB drugs (e.g., INH) exhibits synergistic or additive bactericidal effects, we first measured the MIC values of ACA, VAN, and INH. The results demonstrated MIC90 concentrations of 35.5, 0.518, and 0.437 μM for ACA, VAN, and INH, respectively (Supplementary Fig. 8A). Subsequent combination testing of ACA with INH and ACA with VAN at varying concentrations revealed that. The ACA-VAN combination yielded a fractional inhibitory concentration index (FICI) of ≤1, indicating an additive effect between the two agents. No significant enhancement was observed for the ACA-INH combination (FICI > 1), suggesting the absence of synergistic or additive interactions (Supplementary Fig. 8B). Taken together, these results suggested that ACA may disrupt the cell wall structure and enhances the bactericidal activity of other antibiotics. Discussion With the emergence of multidrug-resistant strains, there is an imperative need to discover new anti-TB agents and identify novel drug targets. In this study, we identified a small molecule compound, ACA, with good bactericidal activity for slow-growing M. TB, and ACA targets the LCP proteins, CpsA1 and CpsA2, to inhibit their pyrophosphatase activities, which disturbs the mycobacterial cell wall assembly. These results support a model in which ACA kills M. TB through targeting CpsA proteins to impair the cell wall synthesis (Fig. 7). Our study not only identified a potential mycobactericidal compound but also elucidated its specific antibacterial mechanism.Fig. 7Schematic representation of the antimycobacterial mechanism of ACA.ACA penetrates mycobacterial cells and binds to the LCP proteins, inhibiting the ligation of AG to PG. The schematic diagram of inner membrane structure is derived from the ChemDraw library. Schematic representation of the antimycobacterial mechanism of ACA. ACA penetrates mycobacterial cells and binds to the LCP proteins, inhibiting the ligation of AG to PG. The schematic diagram of inner membrane structure is derived from the ChemDraw library. Interestingly, we found that ACA exhibits 4.0-fold lower activity against M. smegmatis compared to M. TB H37Ra, and no antibacterial activity was observed for non-mycobacterial species, such as Staphylococcus aureus, despite the presence of LCP proteins in these bacteria. The possible reasons for this phenomenon are as follows: (1) Differences in mycolic acid structure may affect the fluidity and permeability of the bilayer, which could explain the different sensitivity levels of various mycobacterial species to lipophilic inhibitors30. (2) The target LCP proteins in M. smegmatis have a different amino acid sequence, conformational structure, or expression level, leading to reduced binding affinity for ACA. Homologous sequence alignment revealed that the sequence identity between Mtb-CpsA1 and Msm-CpsA1 is 70%, and that between Mtb-CpsA2 and Msm-CpsA2 is 39%. (3) Moreover, the sequence identity between CpsA2 and the LCP family proteins of S. aureus is 15%, while that between CpsA1 and S. aureus LCP family proteins is 16%. Despite overall homology, key residues in the mycobacterial CpsA active site (V204 in CpsA1, V221 in CpsA2) are not conserved in S. aureus. In the present study, an interesting finding is that CspA1 and CspA2 show functional similarity in modulating mycobacterial sensitivity to ACA. Both CspA1 and CspA2 proteins, encoded by the Mtb genome, belong to the LCP family, which is implicated in various physiological process19–24. CspA1 is essential for mycobacterial growth and antibiotic resistance25, whereas CspA2 plays critical roles in countering the host’s innate immune response28. Notably, both CspA1 and CspA2 are involved in the core assembly of the cell wall by facilitating the attachment of arabinogalactan (AG) to peptidoglycan (PG)25,27. Consistently, in this study, we found that these proteins indeed play essential roles in maintaining cell wall integrity. The CspA2 expression can rescue the growth defects of the cpsA1-deleted strain, and CspA1 reverses the ACA-resistant phenotype of the cspA2 mutant strain. Furthermore, both CspA1 and CspA2 exhibited significant pyrophosphatase activity and were targeted by ACA. Overall, our findings together with previous studies suggest that there is a functional similarity between CpsA1 and CpsA2. Another interesting finding from the present study is that ACA kills mycobacteria by disrupting the core assembly of the cell wall. Previous studies have reported that anti-TB drugs inhibit mycobacterial growth through diverse mechanisms31. For instance, rifampin binds to DNA-dependent RNA polymerase, preventing RNA synthesis, which blocks protein synthesis and finally leads to the death of Mtb32. Pyrazinamide disrupts energy metabolism and fatty acid synthesis, consequently reducing the survival of Mtb33. Isoniazid specifically targets the fatty acid synthase II (FAS-II) system to inhibit the synthesis of mycolic acid, weakening the cell wall and finally causing mycobacterial death34. Similarly, the recently approved anti-TB drugs, PMD and DLM, inhibit mycolic acid synthesis, resulting in Mtb growth arrest2,5. Additionally, ethambutol (EMB) disrupts the AG synthesis, thus inhibiting cell wall formation and bacterial growth35. In the present study, we successfully identified an antimycobacterial compound, ACA, which targets the essential AG and PG ligation proteins, CspA1 and CspA2, inhibiting the cell wall assembly and Mtb growth. Therefore, we report a mycobactericidal compound that disrupts the AG and PG ligation in the cell wall, which represents a previously uncharacterized antimycobacterial strategy. Our experimental results demonstrate a clear additive effect between ACA and VAN. A plausible mechanistic explanation is that ACA compromises cell wall integrity, thereby enhancing VAN’s penetration to its target (D-Ala-D-Ala termini of peptidoglycan precursors), whereas ACA-INH’s indifference may reflect metabolic interference—ACA-induced cell wall stress potentially reducing INH’s KatG-dependent activation. In the past few decades of anti-TB research, the development of drugs targeting its unique cell wall metabolism process has been shown to be successful. For example, both PMD and DLM have been approved for the treatment of MDR-TB and XDR-TB owing to their bactericidal properties2. Mechanistically, PMD disrupts the cell wall by inhibiting mycolic acid and protein synthesis, causing mycobacterial cell death5. Similarly, DLM blocks the synthesis of mycolic acid, thereby compromising the structural integrity of the cell wall36. In the present study, ACA significantly impairs the cell wall structure by targeting two LCP proteins, CspA1 and CspA2, leading to the growth arrest of Mtb even at low concentrations. Moreover, our previous studies have demonstrated that the MIC of ACA against both M. tuberculosis H37Rv and multidrug-resistant M. tuberculosis (resistant to rifampicin, streptomycin, isoniazid, and ethambutol) is 0.5 μg/ml29. This result is consistent with the findings of our current study, indicating that ACA targets a mechanism distinct from those of existing first-line anti-TB drugs, thereby exhibiting potent inhibitory activity against M. tuberculosis MDR. Notably, ACA displays bactericidal specificity, preferentially inhibiting the growth of slow-growing mycobacteria such as M. tuberculosis, while showing no appreciable activity against non-mycobacterial species including E. coli, S. aureus, and B. thuringiensis. This selectivity suggests a mechanism of action potentially unique to mycobacteria. Furthermore, ACA was found to enhance the antimycobacterial efficacy of vancomycin, indicating its potential for use in combination therapies. However, the therapeutic potential of ACA has clear limitations. ACA exhibits moderate intracellular efficacy, likely due to its limited permeability across host cells. Its aqueous solubility remains a challenge, which has precluded high-resolution structural studies to unequivocally validate target engagement. Moreover, no synergistic effects were observed with front-line TB drugs such as isoniazid, underscoring the specificity of its interaction. Therefore, while ACA represents a promising chemical scaffold for anti-TB drug development, it is not yet a drug candidate itself. Future efforts should focus on addressing these limitations through systematic chemical optimization. Generating a series of structural analogs will be crucial to improving key pharmacological properties, including solubility, membrane permeability, and intracellular activity. In summary, this study characterized a mycobactericidal compound, ACA, which targets the LCP proteins (CpsA1 and CpsA2) and inhibits their phosphotransferase activity. This ultimately led to the mycobacterial death because the assembly of bacterial cell walls were disrupted and the correct physiological structure was unable to be formed. Our work provides an alternative strategy for developing novel anti-TB drugs. Materials and methods Strains, plasmids, enzymes and reagents Escherichia coli BL21 (λ DE3) and the pET28a vector (Novagen, Germany) were used for protein expression. Restriction enzymes, T4 ligase, modification enzymes, DNA polymerase, and dNTPs were sourced from TaKaRa Biotech (Japan), while PCR primers were synthesized by Invitrogen (USA). Ni-NTA agarose was purchased from Qiagen (Germany), 7H9 and 7H10 broths from Becton, Dickinson (USA), and the compound ACA (3-azidothiophene-2-carboxylic acid) was acquired from Specs (Netherlands). Cloning, expression and purification of recombinant proteins Genes in this study were amplified through polymerase chain reaction (PCR) by using their specific primer pairs (Supplementary Table 1). The PCR templates were from the genomic DNA of M. TB H37Rv. The amplified DNA fragments were cloned into the modified pET28a expression vectors to produce recombinant plasmids (Supplementary Table 2). E. coli BL21 cells were used to express the recombinant proteins. The recombinant E. coli BL21 cells were grown in 1 L Luria broth (LB) medium up to OD600 of 0.6. CpsA1ΔTM (corresponding to residues 38–498 of the full-size protein) and CpsA2ΔTM (corresponding to residues 49–512 of the full-size protein) expression was induced by the addition of 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) at 16 °C for 18 h. The harvested cells were resuspended and sonicated in ice-cold binding buffer (20 mM Tris-HCl, pH8.0, 500 mM NaCl, 5 mM imidazole). The lysate was centrifuged at 9000 g for 30 min, and the supernatant was loaded onto the affinity column (Ni-NTA agarose affinity matrix). The column-bound protein was washed with a wash buffer (20 mM Tris-HCl, pH8.0, 500 mM NaCl, 250 mM imidazole). The elution was dialyzed for 2 h and stored in buffer (20 mM Tris-HCl, pH8.0, 150 mM NaCl, 10% glycerol) at -80 °C. Protein concentration was detected using the Bradford method. MIC assays The MIC was determined using tube-broth dilution methods with several modifications37. For this assay, mycobacterial strains were cultured in 7H9 broth without Tween-80, while other bacteria were cultivated in LB media. Mid-log phase culture was diluted to 5 × 106 CFU·mL-1, and 0.1 mL of the dilution was used to inoculate 2.5 mL culture media containing various concentrations (0–567.5 μM) of ACA. Positive control (isoniazid) and growth control (no compound) were included. Tubes were incubated while shaking at 37 °C for 2 weeks for slow-growing mycobacteria and for 1–2 days for other bacteria. Bactericidal kinetic assay The bactericidal kinetics of ACA were assessed using a time-kill assay, based on the method from reference38 with the following modifications. Given that M. TB H37Ra exhibits closer genotypic and phenotypic proximity to pathogenic strains compared to M. bovis BCG, we selected H37Ra as the primary experimental model. However, observed disparities in ACA susceptibility profiles between these strains prompted us to employ BCG as a supplementary validation system. BCG was cultured in 7H9 broth (supplemented with 0.2% glycerol and 0.05% Tween-80) to mid-log phase (OD600 ≈ 1.0) and then diluted to ~1 × 106 or ~2 × 106 CFU mL-1. H37Ra was cultured in 7H9 broth (supplemented with 0.2% glycerol, 0.05% Tween-80 and 5% OADC) to mid-log phase (OD600 ≈ 1.0) and then diluted to ~2 × 106 CFU mL-1. ACA was added at 0, 1×2×, 5× or 10×MIC, and isoniazid at 10×MIC (3.64 μM). At the indicated time points, samples were taken and CFU counts were performed. Screening and characterization of ACA-resistant mutants To select for ACA-resistant mutants, log-phase M. TBs H37Ra cultures (~1 × 10¹⁰ CFUs) were plated on 7H10 agar containing 5× and 10× the MIC of ACA and incubated at 37 °C for 4 weeks. The resulting colonies were recovered and propagated in 7H9 broth supplemented with 236.5 μM ACA. Genomic DNA was then extracted using a DNeasy Blood & Tissue Kit (Qiagen, Germany). Whole-genome sequencing was performed on the Illumina platform at Shanghai Biotechnology Corporation, with libraries prepared using the Agilent SureSelectXT Target Enrichment System. The sequencing reads were aligned to the M. TB H37Ra reference genome using Burrows-Wheeler Aligner (BWA, v0.7.10) with default parameters. Finally, the Rv3484-V221E mutation identified was confirmed by Sanger sequencing. Construction of M.TB and M. bovis knock-out strains Knockouts of cpsA1 and cpsA2 gene in M. TB H37Ra and M. bovis BCG were generated using homologous recombination39,40, modified as follows. A derivative suicide plasmid, pMind, was constructed, carrying a hygromycin resistance marker and the lacZ reporter gene. The knockout primers are listed in Supplementary Table 1. Allelic exchange mutants with deletions in either the cpsA1 or cpsA2 gene were generated, and their genomic DNA was identified by PCR using primers cpsA1-ΔTM and cpsA2, along with the hygromycin gene. The primers for complement vector construction were cpsA1, p-cpsA2(cpsA2 upstream 265 bp and its open reading frame) and Rv0822c.The amplified DNA fragments were cloned into the modified overexpression vectors pMindD or integration vectors pMV361 to produce recombinant plasmids (Supplementary Table 2). ACA sensitivity assays Recombinant strains were grown for one week in Middlebrook 7H9 media, supplemented with 10% OADC, 0.05% Tween-80, 0.2% glycerol, and 30 μg/mL Kanamycin. When cultures reached an OD600 of 1.5–2.0, they were diluted in 100 mL of fresh 7H9 broth to an OD600 of ~0.15, with or without ACA. These cultures were then incubated at 37 °C with shaking at 160 rpm. At designated time points, aliquots were taken and plated on supplemented 7H10 medium for CFU enumeration⁴¹. Cytotoxicity assays Bone marrow-derived macrophages (BMDMs) were seeded in 96-well plates at 5 × 10⁴ cells per well and incubated overnight at 37 °C with 5% CO₂41. The following day, cells were treated for 18 h with ACA (0, 177.3, 354.7, or 886.7 μM) or 14.6 μM isoniazid (INH) as a control. Lactate dehydrogenase (LDH) activity in the culture supernatant was then measured using the Pierce LDH Cytotoxicity Assay Kit (Thermo Scientific, USA) according to the manufacturer’s instructions. Intracellular survival assay Primary BMDMs were generated from C57BL/6 J mice (6–8 weeks old, 16–18 g, female, purchased from Central Laboratory Animal Center of Huazhong Agricultural University). BMDMs were seeded in 24-well plates at a density of 5 × 10⁵ cells per well and subsequently infected with M. bovis BCG or M. tuberculosis H37Ra at multiplicities of infection (MOI) of 3 and 10. The infections were carried out in DMEM/F12 medium supplemented with 10% fetal bovine serum (FBS) for 4 h. Following infection, extracellular bacteria were removed by washing the cell monolayer three times with phosphate-buffered saline (PBS)42. Subsequently, fresh medium containing various concentrations of ACA (0, 177.3, 354.7, 886.7 μM ACA, 14.6 μM INH as a positive control) was added. Following an 18 h incubation, cells were lysed with 0.05% (w/v) SDS for 10 min at room temperature. Subsequently, serial dilutions of the lysates were plated on Middlebrook 7H10 agar supplemented with OADC and incubated at 37 °C. Bacterial colonies were enumerated after 14 days. Pyrophosphatase activity inhibition assays The pyrophosphatase activity of the M. tuberculosis LCP proteins (CpsA1 and CpsA2) was assayed by measuring the release of inorganic phosphate, based on the method from reference25 with the following modifications. The standard reaction mixture contained 50 mM Tris-HCl (pH8.0), 20 mM MgCl2, 270 μM GPP-Li (Sigma) as a substrate, various concentrations of ACA (0, 2, 20, 40 μg/mLACA, Cordycepin as a negative control), and 10–20 μg of recombinant LCP proteins (devoid of the TM domain) purified from E. coli. The reaction was incubated at 37 °C for 16 h. Inorganic phosphate concentrations were determined using the PiPerTM phosphate assay kit (Thermo Fisher) with absorbance measured at 560 nm in a microplate reader. For Km determination, initial reaction rates were measured by incubating purified 2 μg recombinant CpsA1 or CpsA2 with varying concentrations of GPP-Li ([0 to 400 μM]) in a reaction buffer as described above. The reaction was incubated at 37 °C for 15 min. The concentration of the reaction product, Pi, was determined using a phosphate standard curve. The reaction velocity (V) was calculated based on the Pi concentration and the reaction time. The values of Km and Vmax were derived from a Lineweaver-Burk plot, where the reciprocal of the velocity (1/V) was plotted against the reciprocal of the substrate concentration (1/[S]). For the half-maximal inhibitory concentration (IC₅₀) determination, a substrate concentration near the experimentally determined Km value was selected. The remaining enzyme activity at each ACA concentration (0–800 μM) was expressed as a percentage of the activity measured in the absence of inhibitor. The IC₅₀ was calculated by fitting the dose-response data to a variable slope using GraphPad Prism 8. The inhibitor concentration was plotted against the normalized response. Acid-fast staining assays The wild type H37Ra strain was cultured in 7H9 medium (containing 10% OADC enrichment, 0.05% Tween-80, and 0.2% glycerol) supplemented with or without 59.1 μM ACA, while the wild type BCG strain was cultivated in 7H9 medium with or without 17.7 μM ACA. After these strains were grown up to logarithmic phase, the acid-fast staining was conducted as Solarbio’s Anti-Fast Bacillus Stain Kit (Ziehl-Neelsen’s Method) instructions (Cat#G1274), and observed under a microscope for photographs. Scanning electron microscopy For SEM observation, bacterial samples were fixed with glutaraldehyde, critically point-dried, and sputter-coated with gold, following established methodologies43,44. Mycobacterial strains used for morphological observation were cultured in 7H9 media (supplemented with 10% OADC enrichment, 0.05% Tween-80, and 0.2% glycerol) until a mid-logarithmic growth phase (an OD600 of 0.8 to 1.2) was reached, and then harvested by centrifugation. Subsequent dehydration of the cells involved sequential 3 min treatments with increasing concentrations of ethanol (30, 60, and 90%). The samples were then frozen at −80 °C overnight and subjected to vacuum freeze-drying at −45 °C for 8 h. The freeze-dried bacteria were sputter-coated with gold, and observed using a scanning electron microscope and the bacterial length was measured using ImageJ software. Transmission electron microscopy For transmission electron microscopy (TEM) analysis, mycobacterial strains were cultured in 7H9 media (supplemented with 10% OADC enrichment, 0.05% Tween-80, and 0.2% glycerol) up to a mid-logarithmic growth phase (an OD600 of 0.8 to 1.2), and then harvested by centrifugation at 2000 rpm for 10 min. After discarding the supernatant, the bacteria were allowed to settle at the bottom of the tube. The deposited bacteria were then resuspended in 1 mL of a 2.5% pentanediol fixative solution and fixed at 4 °C for at least 4 h. Following fixation, the cells were centrifuged again at 2000 rpm for 10 min, and the supernatant was discarded. The prepared samples were then sent to a specialized company for further processing and were subsequently observed using transmission electron microscopy. Drug affinity-response target stability The interaction between the LCP proteins and ACA was analyzed using the drug affinity response target stability (DARTS) technique, based on the method from reference45 with the following modifications: the purified His-tag fusion proteins were co-incubated with different concentrations of ACA at 4 °C for 30 min. The incubation mixture was then treated with Proteinase E (Solarbio, CAS 9036-06-0) at 37 °C for 20 min. This reaction was terminated by the addition of SDS-PAGE loading buffer (100 mM Tris-HCl, 4% SDS, 0.2% bromophenol blue, and 20% glycerol, pH 7.0). Following termination, the samples were boiled at 100 °C for 5–10 min, and the digested protein samples were analyzed by SDS-PAGE. Ethambutol and an ACA-unrelated protein, DHQS (dehydroquinate synthase in Mycobacterium bovis BCG), were employed as controls at identical concentrations. Combination bactericidal efficacy assays H37Ra strain was grown in Middlebrook 7H9 media (supplemented with 5% OADC enrichment, 0.05% Tween-80, and 0.2% glycerol). Cells were cultured to an OD600 of approximately 0.8, and then culture was diluted in 100 mL fresh 7H9 broth to an OD600 of approximately 0.07 with 0, 35.5, 47.3 μM ACA or 0.146, 0.292, 0.437 μM INH or 0.345, 0.518 μM VAN. The cultures were cultured at 37 °C with shaking at 160 rpm. Aliquots were taken at the indicated times and plated on 7H10 medium (supplemented with 10% OADC enrichment and 0.2% glycerol) to determine CFUs46. Based on the MIC90 of individual drugs (ACA, INH, and VAN) and the fractional inhibitory concentration index (FICI) formula, we tested the following four drug combinations: ACA (7.1 μM) + INH (0.131 μM); ACA (14.2 μM) + INH (0.262 μM); ACA (7.1 μM) + VAN (0.155 μM); ACA (14.2 μM) + VAN (0.31 μM). The MIC90 determination for combination treatments followed the same protocol as single-drug testing. Statistical analysis and reproducibility Each treatment was triplicated (Data are presented as mean ± SD) or duplicated (all other experiments), and the experiments were repeated at least once to ensure reproducibility. The data were analyzed using GraphPad Prism software or Excel. Unpaired student T-tests were used to analyze the differences between treated and control groups. All statistical data are represented as mean ± SD. Statistical significance was defined as *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001. Supporting Information This article contains supporting information. A document containing Supplementary Fig. 1–11 and Supplementary Table 1, 2 and 3. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Reporting Summary Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Reporting Summary