VapC36-mediated intrinsic stress adaption is required for copper tolerance and pathogenesis of Mycobacterium tuberculosis

Abstract

Here, we have used a comprehensive, system-wide approach to delineate the function of the virulence-associated proteins B and C 36 (VapBC36) toxin-antitoxin (TA) system in Mycobacterium tuberculosis pathophysiology. We show that VapC36 interacts with cognate and noncognate antitoxins. We demonstrate that serX is degraded in vitro by VapC36. We show that VapC36 binds copper, which enhances its ribonuclease activity. Furthermore, we report that deletion of vapBC36 or vapC36 enhances the susceptibility of M. tuberculosis upon exposure to copper in vitro. Compared to the wild-type strain, ΔBC36 was attenuated for growth in guinea pigs. However, deletion of either vapC36 or vapBC36 did not affect M. tuberculosis growth in mice. Compared to the wild-type strain, the in vivo susceptibility of ΔC36 and ΔBC36 strains was increased in copper-fed mice. Together, these findings have enhanced our understanding of the regulation and contribution of TA systems in mycobacterial pathophysiology.

VapC36 interacts with noncognate VapB antitoxins, degrades serX tRNA, and contributes to copper tolerance in M. tuberculosis.

INTRODUCTION

Mycobacterium tuberculosis, the causative agent of tuberculosis (TB) is responsible for approximately 1.3 million annual deaths (1). M. tuberculosis is a highly successful intracellular pathogen that has a remarkable ability to adapt to different stress conditions and survive in host tissues (2). Studies have shown that M. tuberculosis requires several metabolic pathways, such as stringent response, two-component system, sigma factors, and toxin-antitoxin (TA) systems to establish disease in mice and guinea pigs (3–10). TA systems are widely distributed in the genomes of bacteria and archaea and comprise a stable toxin and a labile antitoxin (11). TA systems have been categorized into eight types based on the nature of antitoxins and their mechanism to inhibit the function of toxins (12, 13). Among these, type II is the most prevalent and extensively studied subfamily of TA systems (12, 13). Most toxins belonging to the type II TA systems cleave either mRNA, tRNA, or ribosomal RNA (rRNA) and inhibit bacterial growth in a bacteriostatic manner (11, 12, 14). In addition to transcription, toxins belonging to type II TA systems also interfere with other essential cellular processes such as translation, replication, and cell wall biosynthesis (11, 12). The acquisition and maintenance of a huge repertoire (~90) of TA systems in M. tuberculosis indicates their contribution to M. tuberculosis pathophysiology (15, 16). The high prevalence of TA systems in M. tuberculosis raises the possibility that cross-talk might exist among noncognate TA systems (16–18). Intriguingly, few studies have reported that cross-talk exists between noncognate TA systems in different bacteria, including M. tuberculosis (17, 19–21). It has also been reported that a subset of TA systems is differentially expressed upon exposure of M. tuberculosis to disease-relevant stress conditions and in macrophages (4, 15, 22).

Among these M. tuberculosis TA systems, most belong to the virulence-associated proteins B and C (VapBC) subfamily (15, 16, 18). VapC toxins harbor a PIN domain and require Mg²⁺ or Mn²⁺ for their ribonuclease activity (14, 23–26). Studies have shown that VapC-mediated tRNA cleavage in M. tuberculosis results in ribosomal stalling, translation inhibition, and proteome or transcriptional reprogramming (3, 5, 20, 25, 27, 28). We have earlier reported that the transcriptional profiles obtained upon the ectopic expression of either VapC11, VapC13, VapC21, VapC22, or VapC26 toxins were similar to the profiles obtained from M. tuberculosis exposed to various stress conditions, such as low oxygen and nutritional stress (3, 5, 20, 28). In another study, it was demonstrated that the ectopic expression of VapC4 toxin results in increased levels of transcripts encoding for proteins implicated in the M. tuberculosis copper-resistance pathway (25). It has been suggested that this transcriptional reprogramming enables M. tuberculosis to adapt to stress conditions it might encounter in the host environment. It has also been reported that M. tuberculosis strains harboring deletions in a subset of toxins or TA systems are more susceptible to killing upon exposure to oxidative stress (3, 5, 28–30). Another study showed that the simultaneous deletion of eight TA systems reduced mycobacteriophage resistance and Mycobacterium smegmatis survival in nutrient-limiting growth conditions (31). We have also previously reported that simultaneous deletion of either mazF3, mazF6, and mazF9, or menT3 and menT4, or individual toxins (higB1, menT2, and vapC22) or TA systems (vapBC3, vapBC4, and vapBC11) impaired M. tuberculosis growth in vivo (3–5, 30, 32, 33). However, the molecular mechanisms by which TA systems contribute to M. tuberculosis pathophysiology are not well understood.

Here, we have performed a comprehensive study to ascertain the contribution of the VapBC36 TA system in the pathophysiology of M. tuberculosis. We show that VapC36 binds to cognate (VapB36) and noncognate (VapB28) antitoxins with comparable affinity. RNA sequencing (RNA-seq) studies revealed that overexpression of VapC36 in M. tuberculosis results in transcriptional reprogramming and increased expression of transcripts encoding for proteins associated with ribosomes or implicated in the copper-resistance pathway. Furthermore, we report that VapC36 binds to copper and preferentially degrades serX in vitro. The deletion of vapBC36 enhances M. tuberculosis susceptibility to copper. We further demonstrate that deletion of vapBC36 attenuated M. tuberculosis growth in guinea pigs, and this defect associated with ΔBC36 in vivo is a copper-sensitive effect. We also report that, in addition to VapBC36, individual deletion of either vapC22 or mazF3, mazF6, and mazF9 also enhanced M. tuberculosis susceptibility upon exposure to copper in vitro. Cumulatively, we conclude that VapBC36 forms a complex network of TA interactions and contributes to copper tolerance and pathogenesis of M. tuberculosis.

RESULTS

VapC36 binds to VapB36 and VapB28 with comparable affinity

The extraordinary expansion of TA systems in the M. tuberculosis genome suggests that they might interact and form a complex regulatory network. In the present study, we performed experiments to determine cross-interactions between VapC36 and noncognate antitoxins. For cross-interaction studies, vapB, relE, mazE, and parD antitoxins were cloned into pLam12, an acetamide-inducible vector, and electroporated in M. smegmatis that harbored an integrative copy of anhydrotetracycline (Atc)–inducible VapC36 (Fig. 1A). The expression of VapC36 and antitoxins in recombinant M. smegmatis strains was induced by the addition of Atc and acetamide, respectively, and the growth patterns of recombinant strains were monitored by measuring optical density at 600 nm (OD_600nm) or bacterial numbers at regular intervals (Fig. 1A). We observed that overexpression of VapC36 toxin decreased the bacterial counts compared to the vector control at 18 and 36 hours postinduction (Fig. 1B). As expected, the growth defect associated with VapC36 overexpression in M. smegmatis was reversed by coexpression of cognate antitoxin VapB36 (Fig. 1B). As shown in Fig. 1C and fig. S1A, we observed that the growth defect associated with VapC36 overexpression in M. smegmatis was also restored by coexpression of several noncognate antitoxins belonging to the VapBC subfamily. However, in our repeat experiments, growth restoration was consistently observed in M. smegmatis strains coexpressing VapC36 along with VapB36, VapB28, VapB30, VapB34, or VapB42 (Fig. 1D). These results imply that, in addition to VapB36, VapC36 interacts with noncognate antitoxins belonging to the VapBC subfamily of TA systems.

Fig. 1. VapC36 cross-interacts with the noncognate antitoxins.

(A) The expression of VapC36 and antitoxins in early log phase cultures of M. smegmatis was induced by adding Atc and acetamide, respectively. The growth of various recombinant strains was determined by measuring OD_600nm at different time points. Created in BioRender. Varadarajan, R. (2025) https://BioRender.com/mkjsxkc. (B) The growth kinetics of M. smegmatis strains harboring pTetR-int or pTetR-int-vapC36 in the absence or presence of VapB36 were determined by colony-forming unit (CFU) enumeration. hrs., hours. (C) The heatmap depicts the growth patterns of M. smegmatis strains coexpressing VapC36 in the absence or presence of VapB antitoxins. The scale bar denotes the mean OD_600nm values obtained from at least two independent experiments. (D) Bacterial enumeration of M. smegmatis harboring pTetR-int or pTetR-int-vapC36 in the absence or presence of noncognate antitoxins. The data shown in (B) and (D) are the means ± SD of log₁₀ CFU obtained from two independent experiments performed in duplicates. (E) The binding strength of the interaction between VapC36 and interacting antitoxins was determined using MST. (F) The data shown are the means ± SD of apparent dissociation constants for the binding of VapC36 with various antitoxins and are obtained from two independent experiments, each performed in duplicates. (G) The binding of the vapBC36 promoter-operator DNA fragment with in vitro–reconstituted TA complexes at a stoichiometry of antitoxin: toxin (2:1) was determined using MST. (H) The K_d, as a function of binding of the vapBC36 promoter operator region with various ratios of in vitro–constituted TA complexes, was measured using MST. The reported means ± SD of apparent dissociation constants are derived from two independent experiments, each performed in duplicates. The data and the fit shown in (E) and (G) are obtained from two independent experiments performed in duplicates. The data obtained were statistically analyzed using one-way analysis of variance (ANOVA). ****P < 0.0001.

We next performed thermal stability experiments to confirm interactions between VapC36 with cognate and noncognate VapB antitoxins. For in vitro assays, purified VapC36 toxin, either alone or in complex with various antitoxins at a stoichiometric ratio of 1:1, was subjected to thermal denaturation, and the apparent temperature at which 50% of are protein are unfolded (T_m) was determined. We noticed that the in vitro–assembled VapBC36 TA complex (T_m ~76.2°C) had higher thermal stability compared to VapC36 (T_m ~47°C) (fig. S1B). Notably, we observed that relative to purified VapC36, the in vitro–constituted noncognate (VapB28:VapC36, VapB30:VapC36, VapB34:VapC36, and VapB42:VapC36) TA complexes exhibited greater thermal stability, with T_m values in the range of 54° to 66°C (fig. S1B). In agreement with our coexpression studies, we observed that the thermal stability profiles of VapC36 and the in vitro–constituted noncognate TA complexes (VapB3:VapC36 and VapB4:VapC36) were similar (fig. S1B). These findings are in agreement with earlier studies showing that TA complexes are thermally more stable compared to free toxin proteins (34, 35). Intriguingly, as previously reported, toxins and antitoxins belonging to VapBC36, VapBC28, VapBC30, VapBC34, and VapBC42 are phylogenetically closer and grouped in subcluster III (fig. S2, A and B) (16). As shown in fig. S2 (C and D), we observed that the predicted structures for VapC36, VapC28, VapC30, VapC34, and VapC42 are highly conserved with a backbone RMSD (root mean square deviation) in the range of 0.64 to 3.46 Å. In comparison, the backbone RMSD between predicted structures of VapB36, VapB28, VapB30, VapB34, and VapB42 ranged between 1.39 and 4.03 Å (fig. S2, E and F). Next, we performed microscale thermophoresis (MST) experiments to compare the strength of interactions between the VapC36 toxin with VapB36 and interacting noncognate VapB antitoxins. VapC36 exhibited comparable affinities for binding with cognate antitoxin VapB36 and noncognate antitoxin VapB28 (Fig. 1, E and F). Among the noncognate antitoxins, VapB30 displayed the weakest binding with VapC36 (Fig. 1, E and F). The binding constants for the interaction of VapC36 with VapB36, VapB28, VapB30, VapB34, and VapB42 were ~1.85, 1.35, 435, 3.5, and 6.0 nM, respectively (Fig. 1F). Together, these findings reveal a diverse range of interactions among noncognate TA systems from M. tuberculosis.

In vitro–reconstituted noncognate TA complexes bind to the promoter region of the VapBC36 TA system

The antitoxins belonging to type II TA systems have a DNA binding motif at the N terminus and bind to their native promoter with a modest affinity (36–40). The C terminus domain of antitoxin binds to its cognate toxin, and the TA complex binds to the native promoter with a higher affinity (36, 37). This phenomenon, where the stoichiometric ratios of toxins and antitoxins regulate the transcriptional response, is called conditional cooperativity (41). This regulatory mechanism maintains the ratios of toxins and antitoxins and prevents the random activation of toxins. As shown in fig. S3A, we observed that VapB36 antitoxin had a binding affinity of 124 μM for the vapBC36 promoter operator region. Further, we determined the extent of binding between the vapBC36 promoter operator and in vitro–reconstituted cognate VapBC36 TA complex at VapB to VapC ratios of 1:2, 1:1, and 2:1 (Fig. 1, G and H). We found that the VapBC36 complex had the lowest binding affinity toward promoter-operator at a 1:2 stoichiometric ratio of antitoxin-to-toxin (Fig. 1H). As expected, the binding affinity of the VapBC36 complex with the vapBC36 promoter operator increased with an increase in the antitoxin:toxin ratios (Fig. 1H). At a stoichiometric ratio of 2:1, the VapBC36 TA complex’s dissociation constant (K_d) value for binding to the vapBC36 promoter operator region was ~1.7 μΜ. The apparent binding affinity of the VapB36:VapC36 complex (2:1) with the vapBC36 promoter operator region was ~5.0-fold higher in comparison to the VapBC36 TA complex at a stoichiometric ratio of 1:2 (Fig. 1H). Given that VapC36 interacts with noncognate antitoxins VapB28, VapB34, and VapB42 that share structural and sequence similarities with VapB36, we speculated that these noncognate TA complexes might also be able to bind to the vapBC36 promoter operator region. As shown in Fig. 1 (G and H), noncognate TA complexes were able to bind to the vapBC36 promoter-operator region. Among these, the VapB28:VapC36 showed the highest affinity for the vapBC36 promoter-operator region (Fig. 1H). The binding affinities of VapB28:VapC36, VapB30:VapC36, VapB34:VapC36, and VapB42:VapC36 complexes with the vapBC36 promoter operator region at a stoichiometric ratio of 2:1 were ~1.7, 8.0, 2.5, and 11.0 μM, respectively (Fig. 1H). The obtained data suggest that the TA complexes formed between VapC36 toxin and noncognate antitoxins also bind to the vapBC36 promoter operator region. Further, to validate the specificity of our promoter binding assays, we also determined the binding of the VapBC34 TA complex with the vapBC34 and vapBC36 promoter operator region at different stoichiometric ratios. We observed that the VapBC34 TA complex specifically binds to the vapBC34 promoter operator region, while no binding was observed with the vapBC36 promoter-operator region (fig. S3, B and C). We also observed that the VapB36:VapC34 complex at a stoichiometric ratio of 1:1 was unable to bind to either the vapBC34 promoter operator or the vapBC36 promoter operator region (fig. S3, B and C). In agreement with our promoter binding results, we also observed that VapB36 and VapC34 are unable to interact with each other (fig. S3D). On the basis of these observations, we hypothesize that interactions between VapC36 and noncognate antitoxins might result in reduced cognate VapBC28, VapBC30, VapBC34, and VapBC42 TA complex levels, followed by derepression from their respective promoter operator region. Together, we speculate that interactions between toxins and antitoxins belonging to noncognate TA systems regulate M. tuberculosis growth during unfavorable growth conditions.

Overexpression of VapC36 alters the transcriptional profiles of M. tuberculosis

To better investigate the effects of VapC36 overexpression on the M. tuberculosis transcriptome, we compared the transcriptional profiles of M. tuberculosis strains harboring either pTetR or pTetR-vapC36 at 24 hours postinduction. Using a cutoff log₂ (fold change) of ≥1.0 or ≤ −1.0 and a P value of ≤ 0.05, we identified 791 genes that were differentially expressed (DEGs) in the VapC36 overexpression strain compared to the vector control strain (Fig. 2A). Among these DEGs, the transcript levels encoding for 438 and 353 proteins were increased and decreased, respectively, in the VapC36 overexpression strain compared to the vector control strain (table S1 and Fig. 2A). DAVID (Database for Annotation, Visualization and Integrated Discovery) analysis of the top 200 differentially expressed transcripts with increased expression revealed that these transcripts encode for proteins associated with biological processes such as translation, transcription regulation, and response to metal ions (Fig. 2B) (42). In comparison to the vector control, we observed that the VapC36 overexpression strain exhibited considerably increased levels of transcripts encoding for noncognate toxins and antitoxins, including vapB28 and vapC28 (Fig. 2D and table S1). These observations support our hypothesis that interactions between noncognate antitoxins and toxins might result in transcriptional cross-activation of noncognate TA systems. We also observed that the levels of transcripts encoding for either the 30S subunit or 50S subunit ribosomal proteins were higher in the VapC36 overexpression strain compared to the vector control strain (Fig. 2D and table S1). These observations are consistent with previous studies that reported increased expression of ribosomal proteins upon overexpression of VapC toxins in M. tuberculosis or Mycobacterium abscessus (20, 28, 43). Similar to the VapC4 overexpression strain, we also observed relatively higher levels of transcripts encoding for proteins belonging to the copper-responsive RicR (regulated in copper response) regulon (cysK2, mymT, Rv2963, lpqS, and Rv0846c) and CsoR (copper-sensitive operon repressor) regulon (csoR, ctpV, and Rv0968) in the VapC36 overexpression strain compared to the vector control strain (Fig. 2D and table S1) (25, 44–46). The differential expression of a subset of noncognate toxins, antitoxins, and copper tolerance genes in the VapC36 overexpression strain relative to the vector control strain was verified by quantitative polymerase chain reaction (qPCR) (fig. S4, A and B). Further, we also noticed that the top 200 transcripts with reduced levels in the VapC36 overexpression strain belonged to biological processes that are involved in M. tuberculosis’s energy metabolism such as adenosine 5′-triphosphate (ATP) synthesis, phosphate ion transport, aerobic respiration, and fatty acid biosynthesis (Fig. 2C). These included transcripts encoding for subunits of ATP synthase and NADH dehydrogenase (Fig. 2D and table S1). We also observed that in comparison to vector control, overexpression of VapC36 resulted in reduced levels of transcripts encoding for proteins involved in lipid biosynthesis and transport (Fig. 2D and table S1). Together, these findings imply that ectopic expression of VapC36 leads to increased and decreased relative levels of transcripts encoding for proteins involved in the copper resistance pathway and energy metabolism, respectively. On the basis of these observations, we hypothesized that VapC36 might contribute to copper tolerance in M. tuberculosis.

Fig. 2. VapC36 overexpression results in transcriptional reprogramming in M. tuberculosis.

(A) The volcano plot represents the transcripts with increased (red dots), decreased (blue dots), or unaltered (black dots) levels in the VapC36 overexpression strain relative to the vector control strain. The x and y axes depict log₂ fold change and −log₁₀ P value, respectively. (B and C) DAVID functional analysis for the transcripts with increased (B) or decreased (C) relative levels in the VapC36 overexpression strain relative to the control strain. The color and area of each circle denote the −log₁₀ P value and the number of transcripts observed in each category, respectively. (D) The heatmap depicts the relative expression of transcripts that encode for toxins and antitoxins belonging to TA systems, 50S and 30S ribosomal proteins, proteins involved in copper tolerance, ATP synthesis and NADH dehydrogenases, and lipid biosynthesis and transport in the VapC36 overexpression strain relative to the control strain of M. tuberculosis. The scale bar in these panels represents the log₂ of the observed raw counts. (E to I) VapC36 cleaves the anticodon region of serX in vitro. (E and H) The schematic in these panels represents the cloverleaf model of either serX wild type (E) or serX mutated (serX_M, H). Created in BioRender. Varadarajan, R. (2025) https://BioRender.com/puviu1f. (F) This panel shows the log₂ (protection ratio) for serX in the VapC36 overexpression strain relative to the vector control strain. The x axis and y axes in this panel represent the nucleotide position and log₂ protection ratios of tRNA in the VapC36 overexpression strain relative to the vector control strain. The data shown in these panels are obtained from three biological replicates. (G and I) In vitro tRNA degradation assay with serX wild type (G) and mutated serX tRNA (I) was performed using purified (His)₆-VapC36, resolved on 15% UREA–polyacrylamide gel electrophoresis (PAGE) and visualized by EtBr staining. The data shown in these panels are representative of two independent experiments.

VapC36 toxin cleaves serX in M. tuberculosis

Numerous studies have shown that VapC toxins are site-specific endoribonucleases that mostly cleave the anticodon loop of tRNA (14, 24–26). To ascertain the cellular target of VapC36 in its native host, we analyzed RNA-seq data and calculated the log₂ protection ratio for 44 tRNAs in the VapC36 overexpression strain relative to the vector control strain as previously reported (28, 35). We observed that the VapC36 overexpression strain had the lowest log₂ protection ratio of −4.23 and −3.05 for tRNA^ArgTCT (argW) and tRNA^IleGAT (ileT), respectively, compared to the vector control strain (fig. S5, A, B, D, and E, and table S2). However, neither argW nor ileT tRNA could be cleaved by purified (His)₆-VapC36 in vitro (fig. S5, C and F). Previously, Winther et al. (24) reported that phylogenetically related toxins cleave the same cellular target. VapC28 and VapC30 have been demonstrated to cleave the serine encoding tRNA^SerCGA (serX) and tRNA^SerTGA (serU) of M. tuberculosis, respectively (24). Given that VapC36, VapC30, and VapC28 are phylogenetically closer, we hypothesized that the cellular target of VapC36 might be either serX or serU. In comparison to the vector control strain, the log₂ protection ratios of serX and serU were −1.55 and −2.33, respectively, in the VapC36 overexpression strain (Fig. 2, E and F; fig. S5, G and H; and table S2). The cleavage profile of serX demonstrated a log₂ protection ratio of less than 1.0 throughout the sequence (Fig. 2F). As shown in fig. S5H, we noticed a reduction in the log₂ protection ratio of serU only toward the end of the nucleotide sequence. In agreement with our hypothesis, we observed that purified (His)₆-VapC36 degraded serX in vitro (Fig. 2G). However, we did not notice any changes in the log₂ protection ratio or VapC36-mediated degradation for other serine-encoding tRNAs such as serT, or serV (fig. S5, I to O). We hypothesized that argW, ileT, serU, and other tRNA molecules with reduced log₂ protection ratios might be the cellular targets for noncognate toxins whose expression was increased in the VapC36 overexpression strain relative to the vector control strain (tables S1 and S2). We also observed that VapC36 is unable to degrade serX tRNA harboring mutations (CGA to AAA) in the anticodon loop (Fig. 2, H and I). Together, these observations suggest that VapC36 degrades at the anticodon loop of serine encoding tRNA, serX, in M. tuberculosis.

VapBC36 TA system contributes to copper tolerance in M. tuberculosis

To gain a better understanding of the contribution of the VapBC36 TA system in the pathophysiology of M. tuberculosis, we constructed the M. tuberculosis strain harboring a deletion in the vapBC36 locus (Fig. 3A) (47). The replacement of vapBC36 with the hygromycin resistance gene in the mutant strain was verified by Southern blot. The locus-specific DIG (digoxigenin)-labeled probe hybridized to 4.4- and 5.6-kb regions of Msc I digested genomic DNA isolated from wild-type and ΔBC36 mutant strains, respectively (Fig. 3B). We observed that the growth patterns in 7H9 medium, colony morphology, and ability to form biofilms of M. tuberculosis remained unchanged upon deletion of vapBC36 (fig. S6, A and B). Furthermore, the survival of both wild-type and ΔBC36 mutant strains was comparable following exposure to stress conditions such as oxidative, nitrosative, and nutritional stress (fig. S6, C to E). Since the transcript levels encoding for proteins involved in copper tolerance were increased in the VapC36 overexpression strain compared to the vector control strain, we hypothesized that VapBC36 might contribute to copper tolerance in M. tuberculosis. We first quantified the transcript levels of vapB36 and vapC36 in untreated and copper-treated cultures of M. tuberculosis. As shown in Fig. 3C, in comparison to untreated cultures, we observed that the transcript levels of vapB36 and vapC36 were increased by ~8.0-fold and 5.0-fold, respectively, in copper-treated M. tuberculosis cultures. It has also been reported that M. tuberculosis–infected macrophages accumulate copper inside phagosomes as an innate strategy to eliminate intracellular bacteria (48, 49). We previously reported that vapC36 transcript levels were increased in M. tuberculosis–infected macrophages (4). Next, we compared the survival of wild-type and ΔBC36 strains harboring an empty vector after exposure to copper. We observed enhanced susceptibility of ΔBC36 strain upon exposure to copper compared to the wild-type strain (Fig. 3D). The susceptible phenotype observed in the mutant strain was partially restored in the complemented strain (Fig. 3D). These data unequivocally demonstrate that the VapBC36 TA system plays a pivotal role in the copper tolerance pathway of M. tuberculosis.

Fig. 3. Deletion of vapBC36 or vapC36 increases the susceptibility of M. tuberculosis upon exposure to copper.

(A and G) Schematic representation of vapBC36 (A) and vapC36 (G) locus in the wild-type and respective mutant strains of M. tuberculosis. (B and H) The genomic DNA isolated from various strains was digested with Msc I (B) and Sma I (H), respectively. The replacement of the vapBC36 (B) and vapC36 (H) with the hygromycin resistance cassette in the M. tuberculosis genome was confirmed by Southern blot. (C) The relative levels of vapB36 and vapC36 transcripts were measured by qPCR in M. tuberculosis upon exposure to copper sulfate (100 μM) for 6 hours. The data obtained were normalized to the relative levels of the housekeeping gene, sigA. The data in this panel are the means ± SD of fold change from three independent experiments performed in duplicates. (D and I) The growth of wild-type strain, ΔBC36 (D) and ΔC36 (I) strains, each harboring empty vector, pJEB402, designated as V and their respective complemented strains ΔBC36-CT (D) and ΔC36-CT (I) was determined by spotting assays on 7H11 medium in the absence or presence of copper. The images shown in (D) and (I) are representative of two independent experiments. (E) In vitro binding and dissociation constants between VapC36 and different concentrations of various salts were determined using MST. The data and the fit shown in this panel are the means ± SD of normalized fluorescence obtained from two independent experiments, each performed in duplicates. (F) In vitro serX degradation assays were performed using purified (His)₆-VapC36 in the absence or presence of magnesium and copper metal ions. The in vitro reactions were resolved on 15% UREA-PAGE and visualized by EtBr staining. The data shown are representative of two independent experiments.

VapC toxins are Mg²⁺-dependent ribonucleases, and the divalent ion binds to the acidic pocket of the PIN domain (50, 51). In agreement, we observed that VapC36 binds to Mg²⁺ with the highest affinity and exhibits a K_d value of ~70 μM (Fig. 3E). Several studies have shown that M. tuberculosis proteins implicated in copper tolerance bind directly to copper ions (45, 46, 52, 53). We also observed that VapC36 binds to Cu²⁺ (CuSO₄ and CuCl₂) and has a K_d value of ~735 and 450 μM, respectively (Fig. 3E). Furthermore, in comparison to Mg²⁺ and Cu²⁺, VapC36 exhibited a marginal affinity for Zn²⁺ with a K_d value of ~2.6 mM (Fig. 3E). We did not observe any binding of VapC36 with either Ca²⁺ or Fe²⁺ ions. Next, we determined the binding of Cu²⁺ with other VapC toxins that belong to subcluster III (16). In our binding assays, VapC28 was unable to bind to Cu²⁺ (fig. S7A). We observed that VapC30, VapC34, and VapC42 can also bind to CuSO₄ and CuCl₂ with K_d values between 700 to 800 μM and 900 to 1100 μM, respectively (fig. S7, B to D). The K_d values for the binding of VapC28, VapC30, VapC34, and VapC42 with Mg²⁺ ranged from 75 to 100 μM (fig. S7, A to D). Previous studies have shown that Cu²⁺ regulates protein activity by displacing metal ions from their active site (54, 55). We further performed experiments to determine whether copper binding regulates the ribonuclease activity of VapC36. We observed VapC36-mediated serX degradation in the presence of Cu²⁺ ions (Fig. 3F). The degradation of serX by VapC36 was further increased upon addition of Mg²⁺ and Cu²⁺ ions to the assay buffer (Fig. 3F). On the basis of these observations, we conclude that copper binding enhances the in vitro ribonuclease activity of VapC36.

We next performed experiments to investigate whether the deletion of vapC36 enhances the susceptibility of M. tuberculosis upon exposure to copper. We generated ΔC36 mutant strain of M. tuberculosis using temperature-sensitive mycobacteriophages (Fig. 3G). The replacement of vapC36 with the hygromycin resistance cassette in the mutant strain was verified by Southern blot (Fig. 3H). As observed in the case of ΔBC36, the growth pattern, colony morphology, and biofilm formation of the wild-type and ΔC36 strains were comparable (fig. S6, F and G). As shown in fig. S6 (H to J), both wild-type and ΔC36 strains showed similar survival upon exposure to oxidative, nitrosative, and nutrient-limiting growth conditions. Consistent with the data obtained for the ΔBC36 strain, we observed that ΔC36 also exhibited decreased growth in copper-supplemented 7H11 medium in comparison to the wild-type strain (Fig. 3I). This susceptible phenotype associated with ΔC36 deletion strain was partially complemented in the vapC36 complemented strain (Fig. 3I). In addition, we noticed that the growth of wild type, ΔC36, and ΔBC36 was similar on 7H11 medium supplemented with either zinc or iron (fig. S8). These findings imply that the growth defect observed upon deletion of either vapC36 or vapBC36 is a copper-specific effect. To the best of our knowledge, this is the first study to demonstrate that a toxin belonging to the TA system binds to copper and contributes to copper resistance in M. tuberculosis.

Proteome reprogramming in the wild-type, ΔBC36, and ΔC36 strains upon copper treatment

To further delineate the contribution of the VapBC36 TA system in the M. tuberculosis copper tolerance pathway, we performed label-free quantitative mass spectrometry (MS) using data-dependent acquisition to compare the proteome of mid-log phase cultures from untreated and copper-treated wild-type, ΔBC36, and ΔC36 strains (Fig. 4A). Using a cutoff value of ± 2.0 fold change and a P value of < 0.05, we observed that the relative expression of 54 and 8 proteins was increased and decreased, respectively, in the mid-log phase cultures of ΔBC36 mutant strain compared to the wild-type strain (Fig. 4B and table S3). Among these, in comparison to the wild-type strain, mid-log phase cultures of ΔBC36 showed lower relative levels of proteins involved in cell wall–related processes, including EccB5, EccC5, ESAT-6, EccD5, and EsxB (table S3). It has been demonstrated that EccA5-EccD5 are members of the ESX-5 type VII secretion system and contribute to PPE protein secretion, cell wall integrity and virulence of M. tuberculosis (56, 57). We also observed that the relative levels of various conserved hypothetical proteins, proteins involved in intermediary metabolism and respiration, information pathways, cell wall processes, and lipid metabolism were higher in the mid-log phase of ΔBC36 compared to the wild-type strain (table S3) (58, 59). Previous studies have reported that M. tuberculosis alters its transcriptional profiles upon exposure to copper ions (60, 61). Since deletion of vapBC36 or vapC36 enhanced the susceptibility of M. tuberculosis to copper-mediated killing, we further compared the proteome profiles of untreated and copper-treated mid-log phase cultures of wild-type, ΔBC36, and ΔC36 strains. As previously reported, the relative levels of copper-responsive genes, including MymT, CtpV, CysK2, Rv0968, and Rv0849, were qualitatively higher in copper-treated wild-type strain in comparison to the untreated cultures (table S4) (45, 46, 52, 53). The differentially expressed proteins were considered as qualitatively altered, where the intensity value for peptides belonging to a particular protein was detected either in untreated or copper-treated cultures. In addition, we observed that copper-treated cultures of wild-type strain also had qualitatively higher levels of Hsp (ribosomal binding protein) and Rv3054c (heat shock–inducible protein) compared to untreated cultures (table S4) (62). It has previously been reported that Hsp is probably a molecular chaperone and is involved in initiating translation at high temperatures (62). It has also been reported that metallochaperones are important for intracellular metal ion homeostasis (63). This qualitative increase seen in copper-treated cultures suggests that Hsp might function as a copper-responsive molecular chaperone in M. tuberculosis. We also noticed that the relative levels of copper-responsive proteins such as CysK2, Rv0968, Rv0849, and CtpV were qualitatively increased in samples prepared from copper-treated ΔBC36 and ΔC36 strains, respectively, compared to untreated cultures (table S4).

Fig. 4. Proteome analysis of untreated and copper-treated wild-type, ΔC36, and ΔBC36 cultures from M. tuberculosis.

(A) For proteomics experiments, mid-log phase cultures (OD_600nm ~0.8 to 1.0) of various strains were either untreated or exposed to 100 μM copper sulfate for 24 hours. At designated time points, cell-free protein lysates were prepared, acetone precipitated, and subjected to liquid chromatography–mass spectrometry as described in the Materials and Methods. hr., hours. (B to E) The volcano plots represent the comparison of the proteome profiles obtained from mid-log phase cultures of untreated cultures of wild-type strain and ΔBC36 (B), untreated and copper-treated wild-type strain (C), untreated and copper-treated ΔBC36 strain (D), or untreated and copper-treated ΔC36 strain (E). The x and y axes in these panels depict log₂ fold change and −log₁₀ P value, respectively. The red, blue, and black colored dots represent the proteins whose levels were either increased, decreased, or remained unchanged compared to the control group. (F to H) DAVID functional analysis and enriched biological pathways for proteins whose relative levels are reduced in samples prepared from copper-treated cultures of wild-type (F), ΔBC36 (G), and ΔC36 (H) M. tuberculosis strains, relative to their respective untreated cultures. The color and area of each circle represent the P value and the number of proteins observed in each category, respectively. (I and J) Heatmaps represent the levels of various 50S and 30S ribosomal proteins and CRP obtained from untreated and copper-treated cultures of ΔBC36 (I) or ΔC36 (J) strains. The scale bar in (I) and (J) represents the log₂ of the average raw values observed in samples prepared from untreated or copper-treated cultures. The data shown in (B) to (J) are obtained from three or four biological replicates.

The detailed quantitative analysis of proteomics data revealed that relative levels of 95 proteins were reduced in copper-treated cultures relative to untreated cultures of the wild-type strain (Fig. 4C and table S5). The proteins with lower relative levels were associated with energy metabolism–related biological processes, including glycolytic processes, tricarboxylic acid cycle, iron response, and fatty acid β-oxidation (Fig. 4F). In comparison, the quantitative analysis of the proteomics data revealed that the relative levels of 184 and 105 proteins were decreased in from copper-treated ΔBC36 and ΔC36 cultures, respectively, in comparison to their respective untreated cultures (Fig. 4, D and E, and tables S6 and S7). In contrast to the wild-type strain, translation was the most significantly enriched biological process for proteins with reduced levels in copper-treated ΔBC36 and ΔC36 cultures relative to untreated cultures (Fig. 4, G and H). In addition to the biological process of translation, we observed that in comparison to untreated cultures, the copper-treated ΔBC36 and ΔC36 cultures had lower relative levels of proteins associated with fatty acid biosynthesis and tricarboxylic acid cycle (Fig. 4, G and H). As shown in Fig. 4 (I and J), compared to untreated cultures, the relative levels of several 50S and 30S ribosomal proteins were reduced in copper-treated cultures of ΔBC36 and ΔC36 strains (tables S6 and S7). We also observed that copper-treated ΔBC36 and ΔC36 strains had reduced relative levels of adenosine 3′,5′-monophosphate–responsive protein (CRP, Rv3676) in comparison to untreated cultures (Fig. 4, I and J, and tables S6 and S7). We observed that untreated and copper-treated cultures of wild-type strain had comparable relative levels of ribosomal proteins and CRP. Previously, it has been shown that deletion of the crp gene increases the susceptibility of Mycobacterium bovis BCG (Bacillus Calmette-Guérin) upon exposure to copper (64). It has also been shown that metal ions stabilizes the structure of ribosomes, and ribosomal proteins might contribute to metal ion homeostasis (65, 66). Overall, these findings suggest that (i) exposure of wild type, ΔBC36, and ΔC36 to copper resulted in increased expression of proteins implicated in copper tolerance and (ii) VapBC36 TA system might regulate the relative levels of CRP and ribosomal proteins during copper exposure.

VapBC36 TA system contributes to M. tuberculosis virulence

Previously, it has been suggested that copper-mediated killing is more effective in low-oxygen growth conditions (49, 67). Since infection of guinea pigs with M. tuberculosis results in the formation of hypoxic granulomas and closely resembles the disease reported in humans, we next investigated the growth patterns of wild-type, ΔBC36, and ΔBC36-CT strains in aerosol-infected guinea pigs (Fig. 5A) (67, 68). We observed that at both 4 and 8 weeks postinfection, guinea pigs infected with ΔBC36 strain had ~22.0- to 25.0-fold lower lung bacillary loads in comparison to wild-type strain–infected animals (Fig. 5, B and C). We also observed that the splenic bacillary burdens were reduced by ~8.0- and ~18.0-fold, respectively, in ΔBC36 strain–infected guinea pigs at 4 and 8 weeks postinfection compared to wild-type strain–infected guinea pigs (Fig. 5, B and C). This defect associated with the in vivo growth of the ΔBC36 strain was partially restored in the ΔBC36 complemented strain–infected guinea pigs (Fig. 5, B and C). We observed that the lungs of guinea pigs infected with the ΔBC36 strain had fewer tubercles in comparison to the lung tissues of guinea pigs infected with the wild-type and ΔBC36-CT strains at 8 weeks postinfection (Fig. 5F). Next, we also compared the growth of wild-type, ΔBC36, and ΔBC36-CT strains in aerosol-infected mice (Fig. 5A). We observed that the lung bacillary loads in BALB/c mice infected with either wild-type, ΔBC36, and ΔBC36-CT strains were similar at 4 and 8 weeks postinfection (Fig. 5, D and E). At 8 weeks postinfection, the splenic bacillary burdens in mice infected with the ΔBC36 mutant strain were lower by 6.0-fold relative to mice infected with the wild-type strain (Fig. 5E). The splenic bacterial loads were comparable in wild-type, ΔBC36, and ΔBC36-CT strains–infected mice at 4 weeks postinfection (Fig. 5D). In agreement with the bacterial burdens, gross pathology seen in lungs from mice infected with various strains was comparable at 8 weeks postinfection (Fig. 5G). These observations imply that the VapBC36 TA system is essential for M. tuberculosis to establish disease in guinea pigs but not in mice.

Fig. 5. VapBC36 TA system is required for the pathogenesis of M. tuberculosis in guinea pigs.

(A) Female guinea pigs (Dunkin Hartley, 6 to 8 weeks) or mice (BALB/c) were infected with different M. tuberculosis strains via the aerosol route and disease progression was determined at 4 and 8 weeks postinfection. (B to G) The bacterial burdens were determined in the lungs and spleens of guinea pigs (B and C) and mice (D and E) infected with either wild-type, ΔBC36, or ΔBC36-CT strain at either 4 (B and D) or 8 weeks (C and E) postinfection. The data shown in (B) to (E) are the means ± SD of log₁₀ CFU obtained from five or six animals per strain per time point. (F and G) The representative micrograph of gross pathology of lungs from guinea pigs (F) and mice (G) infected with various strains at 8 weeks postinfection. The images shown in these panels are representative of five or six animals per group. The data obtained were statistically analyzed using one-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

In agreement with our data, it has been previously reported that compared to the wild-type strain, the copper-sensitive ΔmctB mutant strain exhibited a more prominent phenotype in guinea pigs compared to mice (67). This lack of an attenuated phenotype in mice can be attributed to the fact that the granulomas formed in BALB/c mice are not hypoxic, and copper has been suggested to be more bactericidal in low-oxygen growth conditions, where it remains in the toxic reduced (Cu⁺) state (49, 67). To correlate the in vitro phenotype of ΔC36 and ΔBC36 strains with in vivo copper susceptibility, we further compared the growth of wild type, ΔvapBC36, and ΔvapBC36-CT in aerosol-infected mice BALB/c mice that were fed with copper sulfate (200 mg/liter) in drinking water (Fig. 6A). We noticed that the lung bacillary loads in ΔBC36-infected animals were reduced by ~3.0-fold compared to wild-type strain–infected mice at both 4 and 8 weeks postinfection (Fig. 6, B and C). In addition, we observed that compared to wild-type strain–infected mice, the splenic bacillary loads were ~38.0- and 35.0-fold lower in ΔBC36-infected mice at 4 and 8 weeks postinfection, respectively (Fig. 6, B and C). Most ΔBC36 strain–infected mice had splenic bacillary loads below the detection limit at 4 and 8 weeks postinfection (Fig. 6, B and C). This growth defect associated with ΔBC36 was partially restored in mice infected with ΔBC36 complemented strain (Fig. 6, B and C). Together, these data suggest that guinea pigs are a better model to investigate the contribution of copper-sensitive M. tuberculosis mutant strains in disease pathogenesis.

Fig. 6. The attenuated phenotype of ΔBC36 is a copper-specific effect.

(A) Female BALB/c mice were infected with various strains via the aerosol route. The infected mice were fed with copper sulfate (200 mg/liter) in their drinking water, and the bacterial burdens were determined in lungs and spleens at 4 and 8 weeks postinfection. (B to E) The data shown in these panels are the means ± SD of log₁₀ CFU obtained from lungs and spleens of mice infected with various strains at 4 weeks (B and D) and 8 weeks (C and E) postinfection. The data shown are obtained from five animals per strain per time point. The CFU readings shown for the wild-type strain in (B) and (D) are derived from the same experiment. The CFU readings shown for the wild-type strain in (C) and (E) are derived from the same experiment. The dotted line in (B) to (E) represents the limit of detection of CFU enumeration. The data obtained were statistically analyzed using one-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Since we observed that the deletion of vapC36 toxin also rendered M. tuberculosis more susceptible toward copper stress, we further compared the growth of wild-type, ΔC36, and ΔC36-CT strains in normal and copper-fed mice (fig. S9A and Fig. 6A). We noticed that tissue damage and bacterial burdens in the lungs and spleens of wild-type, ΔC36, and ΔC36-CT aerosol–infected mice were comparable at 4 and 8 weeks postinfection (fig. S9, B to D). However, compared to wild-type strain–infected copper-fed mice, the lung bacillary loads in ΔC36 strain–infected copper-fed mice were reduced by ~2.5-fold at 8 weeks postinfection (Fig. 6E). In comparison to wild-type strain–infected copper-fed mice, the splenic bacillary burdens in ΔC36-infected copper-fed mice were reduced by ~14.0- and 8.0-fold, respectively, at 4 and 8 weeks postinfection (Fig. 6, D and E). Together, these findings imply that the VapBC36 TA system is required for copper tolerance and pathogenesis of M. tuberculosis. These findings suggest that the in vivo–attenuated phenotype of M. tuberculosis harboring deletions in either vapBC36 or vapC36 is a copper-sensitive effect.

VapC22 and MazF3, MazF6, and MazF9 toxins also contribute to copper tolerance in M. tuberculosis

Given that deletion of either vapBC3, vapBC4, vapBC11, or vapC22, as well as mazF3, mazF6 and mazF9, menT2, menT3 and menT4 resulted in attenuation of M. tuberculosis growth in vivo, we hypothesized that these TA systems or toxins might also be associated with copper tolerance of M. tuberculosis (3–5, 29, 30, 32). To evaluate our hypothesis, we compared the growth of wild-type, ΔBC3, ΔBC4, ΔBC11, ΔC22, ΔmazF3Δ9Δ6, ΔmenT3, ΔmenT4, and ΔmenT4ΔT3 M. tuberculosis strains in spotting assays in the absence or presence of copper. As shown in Fig. 7A, ΔmazF3Δ9Δ6 and ΔC22 strains also exhibited reduced growth in copper-supplemented plates compared to the wild-type strain. We also observed that the growth of wild-type and ΔC22 complemented strains was similar in the copper-supplemented plates (Fig. 7A). However, the growth of wild-type, ΔBC3, ΔBC4, ΔBC11, ΔmenT3, ΔmenT4, and ΔmenT4ΔT3 strains was comparable in growth media in the absence or presence of copper (fig. S10, A and B). Previous studies have reported that the copper-sensitive M. tuberculosis mutant strains, such as ΔmctB and ΔmymT were specifically sensitive to the Cu(I) ions (52, 67). To further investigate whether the ΔBC36, ΔC36, ΔmazF3Δ9Δ6, and ΔC22 strains are specifically susceptible to Cu(I) ions, we compared the growth of these strains upon exposure to copper in the presence of bathocuproinedisulfonic acid (BCS), a Cu(I)-specific chelator (67). We observed that the growth of wild-type strain, ΔBC36, ΔC36, ΔmazF3Δ9Δ6, and ΔC22 was comparable on 7H11 plates supplemented with 100 μM copper and 100 μM BCS (Fig. 7B). These findings imply that, in addition to VapBC36, VapC22 and MazF3, MazF6, and MazF9 toxins also contribute to the copper tolerance of M. tuberculosis. To the best of our knowledge, this is the first study demonstrating that TA systems might play a pivotal role in the copper tolerance mechanism of M. tuberculosis.

Fig. 7. VapC36, VapBC36, VapC22, MazF3, MazF6, and Maz9 are specifically required for Cu(I) tolerance of M. tuberculosis.

(A) The growth of wild-type, ΔC22, ΔC22-CT, and ΔmazF3Δ9Δ6 was determined by spotting assays on MB7H11 plates in the absence or presence of copper. (B) The growth of wild type, ΔC36, ΔBC36, ΔC22, and ΔmazF3Δ9Δ6 was determined by performing spotting assays on MB7H11 plates in the absence or presence of either copper or Cu(I)-specific chelator BCS. The images shown in (A) and (B) are representative of two independent experiments. (C) Proposed model for the contribution of the VapBC36 TA system in copper tolerance and pathophysiology of M. tuberculosis. In this present study, we speculate that exposure to copper results in the degradation of VapB36 by cellular proteases. The degradation of VapB36 results in derepression and increased expression from the promoter of the vapBC36 TA complex. Further, binding of VapC36 with copper results in copper quenching and enhanced degradation of serX tRNA. We propose that increased degradation of serX by VapC36 results in transcriptional reprogramming. This transcriptional reprogramming results in (i) increased levels of transcripts encoding for proteins involved in copper tolerance, toxins and antitoxins belonging to noncognate TA systems and ribosome-associated proteins, and (ii) decreased levels of transcripts encoding for proteins involved in lipid biosynthesis and transport. We propose that VapC36-mediated transcriptional reprogramming provides an intrinsic mechanism that contributes to copper tolerance and M. tuberculosis pathogenesis.

DISCUSSION

In the present study, we have functionally characterized the VapBC36 TA system from M. tuberculosis. Using coexpression and biophysical experiments, we demonstrate that VapC36 interacts with VapB28, VapB30, VapB34, and VapB42 antitoxins in addition to VapB36. We show that VapC36 had a similar binding affinity for cognate and noncognate antitoxins belonging to subcluster III, particularly VapB28. On the basis of these findings, we hypothesize that the interaction between noncognate antitoxins and VapC36 might result in the formation of the higher-order oligomeric complex. The autoregulation of the TA promoter operator region by cognate TA complexes emphasizes the necessity of maintaining a precise equilibrium between toxin and antitoxin levels to regulate bacterial growth and responses to stress conditions (39, 69, 70). We demonstrate that cognate VapBC36 and noncognate VapB28C36 complexes bind to the vapBC36 promoter operator region with comparable affinity. Also, in concordance with previous findings, an increased antitoxin-to-toxin ratio within these complexes enhanced promoter binding, indicating conditional cooperativity and a fine-tuned regulatory mechanism (35, 71). To the best of our understanding, this is the first study to demonstrate that the binding affinities of cognate and noncognate TA complexes to the promoter-operator region are comparable. These findings imply that complex regulatory networks exist between TA systems in M. tuberculosis.

Furthermore, we demonstrate that expression of VapC36 using Atc-based vector results in transcriptional reprogramming as previously reported upon overexpression of other VapC and MazF toxins in M. tuberculosis (3, 5, 20, 25, 27, 28, 35). We observed that overexpression of VapC36 resulted in transcriptional cross-activation of noncognate TA systems, including VapBC28. This might be attributed to the interactions between noncognate VapC36 and VapB28 and subsequent derepression from the vapBC28 promoter-operator region. The observed large-scale transcriptional changes in the VapC36 overexpression strain might be attributed to increased expression of regulatory proteins and noncognate TA proteins. We show that, similar to VapC28 and VapC30, serX is the cellular tRNA target for VapC36 (24). As reported in other toxin overexpression strains (VapC5, VapC22, and VapC21), we also observed that the relative levels of transcripts encoding for ribosomal proteins were higher in the VapC36 overexpression strain in comparison to the vector control strain (3, 20, 43). We hypothesize that degradation of serX tRNA in M. tuberculosis is likely to be associated with widespread ribosomal stalling at serine UCG codons and subsequent increased expression of ribosomal proteins, as reported earlier (43). Future experiments will include performing 3′ end RNA-seq of the total RNA isolated from the VapC36 overexpression strain to determine its precise cleavage site in its native host (25, 26).

As reported earlier for the VapC4 overexpression strain, we also observed that compared to vector control, overexpression of VapC36 resulted in increased relative levels of transcripts encoding for proteins involved in the copper resistance pathways of M. tuberculosis (25, 44, 46, 72–74). Copper is essential at physiological concentrations, but at higher concentrations, it inhibits microbial growth by irreversibly damaging essential macromolecules such as nucleic acids, proteins, and lipids (73, 75, 76). Previous studies have reported accumulation of copper in M. tuberculosis–infected phagosomes (48). We observed that the transcript levels of vapB36 and vapC36 were increased in M. tuberculosis upon exposure to copper. On the basis of these observations, we hypothesize that higher intracellular copper levels might be responsible for the increased transcript levels of vapC36 in M. tuberculosis–infected macrophages as previously described (4). In concordance with the increased relative levels of transcripts encoding for proteins involved in copper tolerance in the VapC36 overexpression strain, we observed that both ΔC36 and ΔBC36 strains exhibited enhanced susceptibility to copper exposure when compared to the wild type and their respective complemented strains. We also provide compelling evidence that compared to the wild-type strain, strains harboring deletions in either vapC22 or mazF3, mazF6, and mazF9 were more susceptible to copper exposure. Furthermore, the restoration of the growth defect associated with ΔC36, ΔBC36, ΔC22, and ΔmazF3Δ9Δ6 in the presence of Cu(I)–chelating BCS suggests that Cu(I) is the copper species responsible for the observed growth defect in vitro. It has also been reported that the MqsR toxin belonging to the MqsRA TA system from Xylella fastidiosa contributes to persister formation in the presence of copper (77). Several studies have shown that most proteins involved in the copper tolerance pathway bind copper (45, 46, 52, 53). We also report that VapC36, VapC30, VapC34, and VapC42 bind Cu²⁺; however, their binding constants for Cu²⁺ were ~6.0 to 14.0-fold lower compared to Mg²⁺. VapC36-mediated serX cleavage was increased in the presence of Cu²⁺ ions. On the basis of these findings, we speculate that in the presence of high intracellular copper concentration, free VapC36 binds with copper. Subsequently, enhanced serX cleavage by copper-bound VapC36 results in transcriptional reprogramming that enables the bacteria to adapt to high intracellular copper concentrations (Fig. 7C).

The detailed proteomic analysis revealed that copper-treated cultures of wild-type, ΔBC36, and ΔC36 M. tuberculosis strains showed qualitatively higher relative levels of proteins belonging to RicR and CsoR regulons in comparison to their respective untreated cultures (44–46, 60). In addition to these proteins, the relative levels of a molecular chaperone, Hsp, were also increased in copper-treated wild type cultures compared to untreated cultures. Since chaperone proteins are important for metal homeostasis, we speculate that Hsp might function as a copper-specific metallochaperone in M. tuberculosis (63). Compared to the untreated cultures, the relative levels of various ribosomal proteins and CRP were significantly reduced in copper-treated ΔBC36 and ΔC36 cultures. It has been proposed that CRP might regulate the bacterial response to copper by binding to the promoter region of a putative copper exporter, CtpB (64). In another study, it has been reported that metal ions such as zinc and magnesium stabilize the structure of ribosomal proteins, and these proteins also contribute to metal ion homeostasis (65, 66). We hypothesize that the enhanced susceptibility of ΔBC36 and ΔC36 upon exposure to copper might be associated with reduced expression of ribosomal proteins and CRP. Studies have shown that compared to the wild-type strain, copper-sensitive M. tuberculosis strains such as ΔctpV, ΔmctB, and a strain with an inactive ricR regulon exhibited attenuated growth in vivo (44, 67, 78). Further, it has been reported that the copper-dependent antimicrobial activity is enhanced in hypoxic growth conditions and is associated with increased levels of copper transporter CTR1 and its subsequent delivery to phagosomes (79). Given that infection of guinea pigs with M. tuberculosis results in the formation of hypoxic granulomas and increased levels of copper in lung lesions, it is considered an appropriate animal model to assess the in vivo phenotype of copper-sensitive mutants (68, 79). In agreement with the reported phenotype for ΔmctB, we observed that in comparison to the wild-type strain, ΔBC36 strain also displayed attenuated growth in aerosol-infected guinea pigs compared to mice (67). Furthermore, in comparison to the wild-type strain, ΔBC36 and ΔC36 strains were significantly attenuated for growth in the spleens of copper-fed mice. Together, these findings imply that the attenuated growth of ΔBC36 in vivo is a copper-sensitive effect. Overall, the present study expands our knowledge of the (i) regulatory networks that exist among TA systems from M. tuberculosis and (ii) the molecular mechanisms by which TA systems might contribute to the pathogenesis of M. tuberculosis.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions

The detailed list of bacterial strains and plasmids used in the present study is shown in table S8. In the present study, all experiments were performed in either Escherichia coli XL1 blue, E. coli HB101, E. coli BL-21 (λDE3, pLysE), M. smegmatis mc²155, M. tuberculosis H₃₇Rv, or M. tuberculosis Erdman strains. The culturing of E. coli and mycobacterial strains was performed as reported earlier (20, 33). When necessary, hygromycin (150 μg/ml for E. coli and 50 μg/ml for mycobacteria), kanamycin (25 μg/ml for both E. coli and mycobacteria), tetracycline (10 μg/ml for E. coli), ampicillin (50 μg/ml for E. coli), and chloramphenicol (34 μg/ml for E. coli) were used.

Construction of overexpression, knockout, and complemented strains

For overexpression studies, vapC36 was PCR amplified using M. tuberculosis genomic DNA and cloned in either Atc-inducible integrative (pTetR-Int) or episomal (pTetR) vectors (4, 80). M. tuberculosis strains harboring deletions in either vapC36 or vapBC36 were generated using temperature-sensitive mycobacteriophages as reported earlier (47). Briefly, 800 base pairs (bp) of upstream and downstream regions of either vapC36 or vapBC36 were PCR amplified and cloned in pYUB854 flanking a hygromycin resistance cassette. The resulting cosmids, pYUB854-ΔvapC36 and pYUB854-ΔvapBC36 were Pac I digested and packaged into pYUB159 using MaxPlax lambda packaging extract. The recombinant phagemids were electroporated into M. smegmatis to generate high-titer temperature-sensitive mycobacteriophages. To generate mutant strains, mid-log phase cultures of M. tuberculosis H₃₇Rv (OD_600nm ~0.8 to 1.0) were transduced with high-titer temperature-sensitive mycobacteriophages as reported earlier (47). The generation of mutant strains was verified by Southern blot using locus-specific probes. For the generation of complemented strains, vapBC36 was PCR amplified, cloned into pJEB402, and subsequently electroporated into mutant strains. The empty vector, pJEB402, was also electroporated in the wild-type, ΔC36, and ΔBC36 strains for the in vitro copper spotting assays. The construction of complemented strains was validated by qPCR using gene-specific primers. For qPCR studies, total RNA was isolated from mid-log phase cultures of various strains, and cDNA was prepared as previously described (7). The synthesized cDNA was subjected to qPCR using gene-specific primers and SYBR green mix as per the manufacturer’s recommendations. The data obtained were normalized to the transcript levels of the housekeeping gene sigA.

Bioinformatic analysis of the predicted VapB antitoxins and VapC toxins structures

The phylogenetic analysis of VapB antitoxins and VapC toxins was performed using the molecular evolutionary genetics analysis platform (81). The structures of VapC toxins and VapB antitoxins were predicted using SWISS-MODEL and AlphaFold 2.0 (82–84). Briefly, evolutionary-related structures matching the target were identified from the SWISS-MODEL template library using BLAST and HHBlits (85, 86). On the basis of the target-template alignment, various protein models were generated using ProMod3 and their quality (both global and per residue) was assessed using GMQE (Global Model Quality Estimate) and QMEAN scoring function (82). The model with the highest GMQE score was finally selected. The predicted toxin and antitoxin structures were also obtained from the AlphaFold Protein Structure Database using MMseq2 and Foldseek (87). To determine the structural conservation among different predicted VapB antitoxins and VapC toxins structures, the backbone RMSD was calculated using a pairwise structure alignment tool integrated with the Research Collaboratory for Structural Bioinformatics Protein Data Bank (88).

Coexpression and growth inhibition studies in M. smegmatis

For coexpression experiments, antitoxins belonging to either VapB, MazE, Rel, and ParD subfamilies were PCR amplified and cloned into acetamide inducible vector, pLam12, as reported earlier (20). The recombinant plasmids were individually electroporated in M. smegmatis harboring an integrative copy of Atc-inducible VapC36. Subsequently, the expression of VapC36 and antitoxin in early-log phase culture of M. smegmatis (OD_600nm ~ 0.025) was induced by adding Atc (50 ng/ml) and 0.2% acetamide, respectively. The growth patterns of recombinant strains were determined by measuring either OD_600nm or bacterial counts at regular intervals. For colony-forming unit (CFU) enumeration, 10.0-fold serial dilutions were prepared and plated on MB7H11 plates at 37°C for 2 to 3 days.

RNA-seq experiments

For RNA-seq experiments, early-log phase cultures of M. tuberculosis (OD_600nm ~ 0.2 to 0.3) harboring either pTetR or pTetR-vapC36 were induced for 24 hours by adding Atc. At designated time points, induced cultures were harvested, washed with 1× PBS (phosphate-buffered saline) and resuspended in TRIzol. The cells were lysed by bead beating, and total RNA was extracted from the clarified lysates as previously described (7). The purified RNA was treated with deoxyribonuclease (DNase) I and shipped to Medgenome for RNA-seq. Further, rRNA was removed, and cDNA libraries were prepared for sequencing. The prepared cDNA libraries were sequenced on the Illumina HiSeqX platform to generate 30 million, 2 × 150 bp reads per sample. The quality of the raw data obtained was assessed using FastQC v0.11.4 as reported earlier (20). The raw data were further analyzed using the reference M. tuberculosis H37Rv genome that was retrieved from the ENSEMBL database, indexed using Bowtie2, and aligned to the reference genome with TopHat v2.1.1 (89, 90). To determine the cleavage profiles for various tRNAs, SAMtools depth was used to calculate the coverage at each base for each replicate in reads obtained from the vector control and VapC36 overexpression strain (91). The coverage counts were normalized, and protection ratios at each nucleotide position in the overexpression strain relative to the vector control were computed as reported earlier (28, 35).

Cloning, expression, and purification of proteins

To purify individual toxins and antitoxins, genes were PCR amplified and cloned into either pET28b or pET15b vector. The recombinant plasmids were transformed into E. coli BL-21 (λDE3, pLysE). The recombinant strains were cultured in TB medium till OD_600nm ~0.8 and expression of recombinant proteins was induced by the addition of 1.0 mM isopropyl-β-d-thiogalactopyranoside overnight at 20°C. The induced cultures were harvested by centrifugation, bacterial pellets were resuspended and sonicated in lysis buffer [10 mM Hepes (pH 8.0), 100 mM NaCl, 100 mM l-arginine, and 10% glycerol containing 1× protease inhibitor cocktail]. The clarified lysates were prepared and incubated with 2 ml of Ni–nitrilotriacetic acid Sepharose 6 Fast Flow for 4 hours at 4°C. The unbound protein fractions were removed by washing the resin with 25 volumes of the wash buffer [10 mM Hepes (pH 8.0), 100 mM NaCl, 100 mM l-arginine, 50 mM imidazole, and 10% glycerol]. Last, the recombinant protein was eluted with elution buffer [10 mM Hepes (pH 8.0), 100 mM NaCl, 100 mM l-arginine, 10% glycerol, and a gradient of imidazole (100 to 900 mM)]. The purity of the eluted fractions was assessed by 15% Tricine SDS–polyacrylamide gel electrophoresis (PAGE), and the protein concentration was determined by absorbance measurements. The purified fractions were dialyzed, concentrated, and stored in storage buffer [10 mM Hepes (pH 8.0), 100 mM NaCl, 100 mM l-arginine, 500 mM imidazole, 10% glycerol, and 1× protease inhibitor cocktail) at −80°C till further use.

Thermal stability analysis by nanoDSF studies

Thermal unfolding of purified VapC36 and in vitro–constituted TA complexes was performed using 10 μM protein concentration by nano differential scanning fluorimetry (nanoDSF) (Prometheus NT.48) as previously described (92, 93). Briefly, the samples were loaded into the capillaries of the apparatus and heated at a ramp rate of 1°C/min from 20° to 90°C. Further, the samples were excited at 280 nm, and emission was detected at 330 and 350 nm to determine F₃₅₀/F₃₃₀ ratio. The first derivative of F₃₅₀/F₃₃₀ was plotted as a function of temperature, to determine the apparent T_m for different protein complexes as previously described (92, 93).

MST experiments

For MST experiments, VapC36 protein was buffer exchanged in 10 mM Hepes (pH 8.4) and labeled using the Monolith Protein labeling kit RED-NHS (NanoTemper Technologies) according to the manufacturer’s instructions. To perform protein-protein binding studies, a dimeric concentration of 40 nM VapC36 was labeled and titrated with different concentrations of unlabeled antitoxins. For metal ion binding studies, 200 nM of labeled VapC36 toxin was titrated with varying concentrations of CaCl₂, CuSO₄, CuCl₂, MgSO₄, FeSO₄, and ZnSO₄. For promoter binding studies, 40 nM of labeled VapC36 was incubated with various concentrations of cognate and noncognate antitoxins to form in vitro–constituted TA complexes of different stoichiometry. These in vitro–constituted complexes were incubated with various concentrations of the vapBC36 promoter-operator DNA fragment. The binding studies were performed at 25°C in 10 mM Hepes (pH 8.4). The measurements were performed using a light-emitting diode/excitation power setting of 40 to 60%, and two MST power settings, medium and high. The data were analyzed using MO Affinity Analysis software (version 2.2.5, NanoTemper Technologies) at different standard MST-off times, as reported earlier (92, 94, 95).

In vitro transcription and tRNA cleavage experiments

For in vitro transcription, tRNA-encoding wild-type or mutant oligonucleotide sequences were synthesized with a T7 overhang in the forward primer. These oligos were annealed, and PCR amplification was performed using the T7 promoter and a second primer corresponding to the 3′ end of the tRNA gene. In vitro transcription for tRNA molecules was performed using T7 RNA polymerase and the PCR product as a template according to the manufacturer’s instructions. The tRNA preparation was DNase I treated, precipitated with ammonium acetate, and finally resuspended in nuclease-free water. For in vitro tRNA cleavage assays, 2 μM of purified VapC36 was incubated with purified tRNA in 1× cleavage buffer [20 mM tris (pH 7.4), 60 mM NaCl, 1 mM dithiothreitol (DTT), and 10 mM MgCl₂] at 37°C for 40 mins. The enzymatic reactions were inactivated by adding 1× RNA formamide dye and resolved on 15% UREA-PAGE as per standard protocols (5).

In vitro growth, stress, and copper susceptibility experiments

The growth patterns of wild-type and mutant strains were compared by measuring OD_600nm at regular intervals till the stationary phase was attained. For biofilm experiments, mid-log phase cultures (OD_600nm ~ 0.8 to 1.0) of various strains were harvested, washed, and diluted in Sauton’s medium [aspartate (0.5 g/liter), KH₂PO₄ (1.0 g/liter), Na₂HPO₄ (2.5 g/liter), ferric ammonium citrate (50 mg/liter), MgSO₄.7H2O (0.05 g/liter), CaCl₂ (0.05 g/liter), and ZnSO₄ (0.01 g/liter)]. The diluted cultures were added to six-well plates, sealed with parafilm, and kept at 37°C without shaking in a CO₂ incubator for 3 to 4 weeks. The susceptibility of various strains to different stress conditions, such as oxidative, nitrosative, and nutritional stress, was performed using protocols as reported earlier (7). Briefly, early-log phase cultures of various strains (OD_600nm ~ 0.2 to 0.3) were exposed to either oxidative stress (5 mM H₂O₂ for 1 day), nitrosative (5 mM NaNO₂ for 1 day), or nutritional stress [1× tris-buffered saline (pH 7.4) containing 0.05% tween 80 for 7 days]. For bacterial enumeration, at designated time points, 10.0-fold serial dilutions were prepared and plated on MB7H11 agar at 37°C for 3 to 4 weeks. For copper susceptibility assays, mid-log phase cultures of various strains (OD_600nm ~ 0.8 to 1.0) were harvested and washed twice with 1× PBS. Subsequently, 10.0-fold serial dilutions were prepared and spotted on MB7H11 agar in the absence or presence of copper sulfate (in the absence or presence of 100 μM BCS), zinc sulfate, and iron chloride at 100 μM concentration. The spotting plates were incubated at 37°C for 3 to 4 weeks.

In vivo animal studies

The animal experiments were approved by the Institutional Animal Ethics Committee of Translational Health Science and Technology Institute (THSTI), Faridabad, India (IAEC/THSTI/106 and IAEC/THSTI/214). The experiments were performed according to the guidelines proposed by the committee for the purpose of control and supervision of experiments on animals. The animals were housed in individually ventilated cages in the biosafety level 3 facility as previously described (30). For aerosol infection, single cell suspension from mid-log phase cultures of various strains (OD_600nm ~ 0.8 to 1.0) was prepared, and female guinea pigs (Dunkin Hartley strain, 6 to 8 weeks old) or mice (BALB/c, 6 to 8 weeks old) were infected via aerosol route using an inhalation exposure system (Glascol). The aerosol infection resulted in the implantation of ~50 to 100 bacilli per lung at day 1 postinfection. In the copper treatment experiment, aerosol-infected mice were fed copper sulfate (200 mg/liter) in drinking water from the first day postinfection. For bacterial enumeration, lungs and spleens were homogenized in 2 ml of normal saline, and 10.0-fold serial dilutions were prepared and plated on MB7H11 agar at 37°C for 3 to 4 weeks.

LC-MS/MS, data curation, and analysis

For proteomics experiments, mid-log phase cultures (OD_600nm ~ 0.8 to 1.0) of various strains were either untreated or exposed to 100 μM copper sulfate for 24 hours. At designated time points, cultures were harvested, washed twice with 1× PBS, and resuspended in lysis buffer (50 mM triethyl bicarbonate buffer, 2% SDS, and 1× protease inhibitor cocktail). Subsequently, the clarified lysates were prepared using bead-beating, as reported earlier (3). For liquid chromatography–tandem MS (LC-MS/MS) analysis, 500 μg of clarified protein lysates were acetone precipitated, washed with ice-cold acetone and subjected to overnight trypsin digestion at 37°C following reduction and alkylation using DTT and iodoacetamide, respectively. The resulting peptide mixture was acidified with formic acid, desalted, and vacuum concentrated. LC-MS/MS was performed using an Eksigent nano-LC 415 connected to a Sciex TripleTOF 5600 mass spectrometer (96). The trypsin-digested samples were loaded onto a trap column (C18-CL, 300 μm by 10 mm, 5 μm, 120 Å) and separated on a reverse-phase analytical column (C18-CL, 300 μm by 150 mm, 3 μm, 120 Å). The mobile phase for high-performance LC consisted of the following components: water, acetonitrile, and formic acid (solvent A, 98, 2, and 0.1% and solvent B, 2, 98, and 0.1%). For LC-MS/MS analysis, 10 μg of processed samples were injected onto the column with a flow rate of 5 μl/min for a 90 min gradient run (5% solvent B for 5 mins, 25% solvent B for 67 mins, 35% solvent B for 77 mins) followed by washing and equilibration with solvent A for the remaining minutes. The sample data were acquired using an electrospray ionization ion source with optimized settings for voltage, curtain gas, nebulizer gas, heater gas, and the source temperature. The MS was operated with a resolution of at least 30,000 full width at half maximum for time-of-flight MS scans with a mass range of mass/charge ratio (m/z) 350 to 1250. The survey scans with a cycle time of 2.3 s were acquired for 30 product ions covering a mass range of m/z 100 to 1500. Using the in-built Andromeda algorithm in MaxQuant software, the protein identification in the samples was performed through database searching against the UniProt M. tuberculosis (strain ATCC 25618/ H37Rv, July 2024) (97). The following search parameters were used for data analysis: (i) variable modifications, oxidation (M), and acetyl (protein N-term); (ii) fixed modifications, carbamidomethyl (C); (iii) label-free quantification; (iv) digestion, trypsin/P; and (v) max. missed cleavages 2. To determine the false discovery rate (FDR) for peptide identification, a reverse database search approach was used and the proteins were identified with an FDR of less than 1%. Heatmaps were constructed using Graph Pad Prism 9.5.1 by calculating the log₂ fold of the average raw values of proteins quantified in the present study. NCBI-based tool DAVID was used to obtain significantly enriched biological processes in the present study (42).

Statistical analysis

Graph Pad Prism 9.5.1 software (GraphPad Software Inc., CA, USA) was used for the preparation of graphs and statistical analysis. The P of < 0.05 was considered to be statistically significant. The number of replicates and statistical tests used are mentioned in the respective figure legends.

Supplementary Materials

The PDF file includes:

Legends for tables S1 to S3, and S5 to S7

Tables S4 and S8

Figs. S1 to S10

Other Supplementary Material for this manuscript includes the following:

Tables S1 to S3, and S5 to S7