Competition between H4PteGlu and H2PtePAS Confers para-Aminosalicylic Acid Resistance in Mycobacterium tuberculosis

Tuberculosis remains a serious challenge to human health worldwide. para-Aminosalicylic acid (PAS) is an important anti-tuberculosis drug, which requires sequential activation by Mycobacterium tuberculosis (M. tuberculosis) dihydropteroate synthase and dihydrofolate synthase (DHFS, FolC). Previous studies showed that loss of function mutations of a thymidylate synthase coding gene thyA caused PAS resistance in M. tuberculosis, but the mechanism is unclear. Here we showed that deleting thyA in M. tuberculosis resulted in increased content of tetrahydrofolate (H4PteGlu) in bacterial cells as they rely on the other thymidylate synthase ThyX to synthesize thymidylate, which produces H4PteGlu during the process. Subsequently, data of in vitro enzymatic activity experiments showed that H4PteGlu hinders PAS activation by competing with hydroxy dihydropteroate (H2PtePAS) for FolC catalysis. Meanwhile, over-expressing folC in ΔthyA strain and a PAS resistant clinical isolate with known thyA mutation partially restored PAS sensitivity, which relieved the competition between H4PteGlu and H2PtePAS. Thus, loss of function mutations in thyA led to increased H4PteGlu content in bacterial cells, which competed with H2PtePAS for catalysis by FolC and hence hindered the activation of PAS, leading to decreased production of hydroxyl dihydrofolate (H2PtePAS-Glu) and finally caused PAS resistance. On the other hand, functional deficiency of thyA in M. tuberculosis pushes the bacterium switch to an unidentified dihydrofolate reductase for H4PteGlu biosynthesis, which might also contribute to the PAS resistance phenotype. Our study revealed how thyA mutations confer PAS resistance in M. tuberculosis and provided new insights into studies on the folate metabolism of the bacterium.


Introduction
Tuberculosis (TB), caused by M. tuberculosis, is an ancient infectious disease.Recent data released by the World Health Organization show that around 10 million people fell in with the disease every year worldwide [1].The increasing spread of drug-resistant M. tuberculosis makes TB treatment more difficult, and drug resistance has become one of the major challenges.The best way to solve the above problem is to introduce new anti-TB drugs.However, no new first-line drug has been introduced in clinical TB treatment for more than 50 years, since rifampicin [2].Therefore, rational use of existing anti-tuberculosis drugs is necessary.In addition, researchers also have made efforts in using phages as an individual or supplementary therapy to treat M. tuberculosis infections [3].
Folate is an essential nutrient for all sorts of life.Bacteria need to synthesize folate de novo, but mammals are unable to synthesize it, which makes the bacterial de novo folate biosynthesis pathway an ideal target for developing new antibacterial drugs [4].As is well known, dihydropteroate (H 2 Pte) is synthesized by dihydropteroate synthetase (DHPS, FolP) using para-aminobenzoic acid (pABA) and 7,8-dihydropterin pyrophosphate (H 2 PtePP) as substrates, which is further converted into dihydrofolate (H 2 PteGlu) by FolC (Figure 1) [5].Dihydrofolate reductase (DHFR, DfrA or RibD) and thymidylate synthase (ThyA or ThyX) maintain the interconversion and balance between H 2 PteGlu, H 4 PteGlu and 5, 10-methylenetetrahydrofolate (5, 10-m-H 4 PteGlu) (Figure 1).PAS was first used as a first-line anti-TB drug in 1946 [6], and is presently still used for treating multiple drug-resistant TB [7].The mechanism of action of PAS had been gradually discovered over 70 years of clinical utilization.As a structural analogue of pABA, PAS is firstly catalyzed by the FolP1 of M. tuberculosis to form H 2 PtePAS, an analogue of H 2 Pte.Subsequently, H 2 PtePAS was further catalyzed by the FolC, yielding H 2 PtePAS-Glu [5] (Figure 1).Ultimately, H 2 PtePAS-Glu inhibited the activity of M. tuberculosis DfrA (Figure 1), resulting in bacterial growth inhibition and cell death [8].
Antibiotics 2023, 12, x FOR PEER REVIEW 2 of 13 tuberculosis drugs is necessary.In addition, researchers also have made efforts in using phages as an individual or supplementary therapy to treat M. tuberculosis infections [3].
Folate is an essential nutrient for all sorts of life.Bacteria need to synthesize folate de novo, but mammals are unable to synthesize it, which makes the bacterial de novo folate biosynthesis pathway an ideal target for developing new antibacterial drugs [4].As is well known, dihydropteroate (H2Pte) is synthesized by dihydropteroate synthetase (DHPS, FolP) using para-aminobenzoic acid (pABA) and 7,8-dihydropterin pyrophosphate (H2PtePP) as substrates, which is further converted into dihydrofolate (H2PteGlu) by FolC (Figure 1) [5].Dihydrofolate reductase (DHFR, DfrA or RibD) and thymidylate synthase (ThyA or ThyX) maintain the interconversion and balance between H2PteGlu, H4PteGlu and 5, 10-methylenetetrahydrofolate (5, 10-m-H4PteGlu) (Figure 1).PAS was first used as a first-line anti-TB drug in 1946 [6], and is presently still used for treating multiple drugresistant TB [7].The mechanism of action of PAS had been gradually discovered over 70 years of clinical utilization.As a structural analogue of pABA, PAS is firstly catalyzed by the FolP1 of M. tuberculosis to form H2PtePAS, an analogue of H2Pte.Subsequently, H2PtePAS was further catalyzed by the FolC, yielding H2PtePAS-Glu [5] (Figure 1).Ultimately, H2PtePAS-Glu inhibited the activity of M. tuberculosis DfrA (Figure 1), resulting in bacterial growth inhibition and cell death [8].Although the mechanism of PAS action has been elucidated, its mechanisms of resistance still await investigation.Until the present, confirmed molecular markers associated with PAS resistance in M. tuberculosis clinical isolates included mutations of folC [9][10][11], thyA [9,[11][12][13], and ribD [8,9,11].Among them, folC or thyA gene mutations were the main reasons for PAS resistance, accounting for two-thirds of the PAS resistant clinical isolates [9,11,14].Molecular mechanisms of PAS resistance caused by folC and ribD mutations have been elucidated [8,10].Our previous research showed that H2Pte binding pocket variants of FolC failed to activate H2PtePAS to H2PtePAS-Glu, hindering the activation of PAS and hence conferring resistance to PAS [10].On the other hand, ribD could serve as an alternative for DHFR, as mutations in the promoter region of the gene could cause over-expression of ribD, and thus lead to PAS resistance [8].However, the molecular Although the mechanism of PAS action has been elucidated, its mechanisms of resistance still await investigation.Until the present, confirmed molecular markers associated with PAS resistance in M. tuberculosis clinical isolates included mutations of folC [9][10][11], thyA [9,[11][12][13], and ribD [8,9,11].Among them, folC or thyA gene mutations were the main reasons for PAS resistance, accounting for two-thirds of the PAS resistant clinical isolates [9,11,14].Molecular mechanisms of PAS resistance caused by folC and ribD mutations have been elucidated [8,10].Our previous research showed that H 2 Pte binding pocket variants of FolC failed to activate H 2 PtePAS to H 2 PtePAS-Glu, hindering the activation of PAS and hence conferring resistance to PAS [10].On the other hand, ribD could serve as an alternative for DHFR, as mutations in the promoter region of the gene could cause overexpression of ribD, and thus lead to PAS resistance [8].However, the molecular mechanism of PAS resistance caused by thyA mutations still remains unclear, though the association between thyA mutations and PAS resistance has been established for nearly two decades [13].
According to the data of epidemiological analysis, thyA mutations were identified in about 1/3 of the PAS resistant M. tuberculosis clinical isolates [9,11,12].Thus, unravelling the mechanism of PAS resistance caused by thyA mutations will broaden our understanding of folate metabolism in M. tuberculosis and be useful for guiding the clinical administration of PAS.To elucidate how thyA mutations caused PAS resistance in M. tuberculosis, the thyA gene was deleted in H37Ra using the phage-mediated allelic exchange method, and a clinical PAS resistant isolate F461 harboring the thyA R235P mutation was selected [14].Subsequently, the effect of thyA deletion on bacterial H 4 PteGlu content was determined by UPLC-MS/MS.Then, the competition for catalysis of FolC between H 4 PteGlu and H 2 PtePAS was analyzed by in vitro enzymatic activity assays.Meanwhile, folC was overexpressed in the thyA deletion mutant and the selected PAS resistant clinical isolate, PAS susceptibilities of these two strains were tested.The level of FolC in ThyA deficiency strain was explored by RNA-seq and Western blot assays.The results are presented herein.

thyA Deletion Leads to High Level PAS Resistance in M. tuberculosis
Considering the genetic complexity of clinical isolates, and also high similarity of mechanisms of PAS action and resistance between H37Ra and H37Rv [10], we constructed the thyA deletion strain in H37Ra to elucidate the molecular mechanism of how thyA mutations lead to PAS resistance in M. tuberculosis.H37Ra ∆thyA showed a significant growth defect (Figure S1), which is consistent with the observation in H37Rv ∆thyA [15].Subsequently, the susceptibility to PAS was determined.The results showed that thyA deletion led to a hundreds of times increase in minimum inhibitory concentration (MIC) of PAS to M. tuberculosis (Table 1), which is consistent with clinical data [13].After that, recombinant plasmids carrying thyA or thyX genes from M. tuberculosis H37Ra were used to transform H37Ra and H37Ra ∆thyA, respectively.Plasmid-borne expression of thyA restored PAS sensitivity of the thyA deletion strain, but that of thyX could not (Table 1).We noticed that over-expression of thyA and thyX both caused an eight times increase in PAS MIC (Table 1).

folC Over-Expression Partially Restores PAS Sensitivity in thyA Functional Deficient Strains
Previous researches have confirmed that blocking the incorporation of PAS into folate synthesis pathway leads to high level resistance to PAS in M. tuberculosis [8,10].To assess whether the high-level resistance to PAS of the thyA deletion strain was related to the efficiency of PAS incorporation, core genes folP1, folC, and dfrA of the folate biosynthesis pathway were over-expressed in H37Ra and H37Ra ∆thyA.The results showed that plasmid-borne expression of folP1 and folC in H37Ra led to increased sensitivity to PAS, as demonstrated by the reduced MICs (four times for folP1 over-expression and two times for folC over-expression) (Table 2).As the target for bio-activated PAS, dfrA over-expression increased the PAS MIC by thousands of times (Table 2).Over-expression of folP1 in H37Ra ∆thyA also led to a four-times decrease in PAS MIC, which was consistent with that in H37Ra (Table 2).However, over-expressing folC in H37Ra ∆thyA led to a 16-times decrease in PAS MIC, and over-expressing dfrA in H37Ra ∆thyA did not change the PAS MIC (Table 2).To further prove that over-expressing folC could reverse the high-level PAS resistance phenotype in thyA functional deficient strains, folC was over-expressed in the PAS resistant clinical isolate harboring the thyA R235P mutation.As shown in Table 2, folC over-expression also led to a 10-times decrease in PAS MIC in the clinical isolate.

The Expression Level of folC Gene and FolC Protein Remain Unchanged in H37Ra ∆thyA
To further explore the role of folC in PAS resistance caused by ThyA functional deficiency, we detected the expression level of folC in wild-type and thyA deletion strain.Western blot assay was performed to compare the expression level of FolC between wildtype and thyA deletion strain, and the results showed that the FolC expression level was not significantly changed in the thyA deletion strain (Figure 2A,B).Meanwhile, RNA-seq data also showed that the expression level of folC was not significantly changed in the thyA deletion strain (Figure 2C).
with that in H37Ra (Table 2).However, over-expressing folC in H37Ra ΔthyA led to a 16times decrease in PAS MIC, and over-expressing dfrA in H37Ra ΔthyA did not change the PAS MIC (Table 2).To further prove that over-expressing folC could reverse the high-level PAS resistance phenotype in thyA functional deficient strains, folC was over-expressed in the PAS resistant clinical isolate harboring the thyA R235P mutation.As shown in Table 2, folC over-expression also led to a 10-times decrease in PAS MIC in the clinical isolate.

The Expression Level of folC Gene and FolC Protein Remain Unchanged in H37Ra ΔthyA
To further explore the role of folC in PAS resistance caused by ThyA functional deficiency, we detected the expression level of folC in wild-type and thyA deletion strain.Western blot assay was performed to compare the expression level of FolC between wildtype and thyA deletion strain, and the results showed that the FolC expression level was not significantly changed in the thyA deletion strain (Figure 2A,B).Meanwhile, RNA-seq data also showed that the expression level of folC was not significantly changed in the thyA deletion strain (Figure 2C).

thyA Deletion Leads to Increased H 4 PteGlu Content in Bacterial Cells
There are two types of thymidylate synthase, ThyA and ThyX, in M. tuberculosis [15], and the thymidylate synthase function is mainly performed by ThyA.ThyA uses 5, 10m-H 4 PteGlu as methyl donor to generate H 2 PteGlu and maintain the balance of folate metabolism (Figure 1) [15,16], and ThyX uses 5, 10-m-H 4 PteGlu as methyl donor to generate H 4 PteGlu (Figure 1) [16,17].After the loss of ThyA function, the bacterium relies on ThyX for synthesizing thymidylate [15].Thus, we speculated that the H 4 PteGlu content would increase in ThyA deficient strains.As expected, we observed an obvious increase in H 4 PteGlu content in the thyA deletion strain compared to the wild-type strain (Figure 3).
ments were repeated at least three times, and were performed three biological replicates each time.(B) Relative quantitative of FolC product by Western blot assay.ns, no significance.(C) Comparison of the transcriptional level of the gene folC during the exponential phase in H37Ra (WT) and H37Ra ΔthyA (thyA − ) by RNA-seq.ns, no significance.

thyA Deletion Leads to Increased H4PteGlu Content in Bacterial Cells
There are two types of thymidylate synthase, ThyA and ThyX, in M. tuberculosis [15], and the thymidylate synthase function is mainly performed by ThyA.ThyA uses 5, 10-m-H4PteGlu as methyl donor to generate H2PteGlu and maintain the balance of folate metabolism (Figure 1) [15,16], and ThyX uses 5, 10-m-H4PteGlu as methyl donor to generate H4PteGlu (Figure 1) [16,17].After the loss of ThyA function, the bacterium relies on ThyX for synthesizing thymidylate [15].Thus, we speculated that the H4PteGlu content would increase in ThyA deficient strains.As expected, we observed an obvious increase in H4PteGlu content in the thyA deletion strain compared to the wild-type strain (Figure 3).

Comparison of Catalytic Efficiency of FolC on H2Pte, H4PteGlu and H2PtePAS
FolC was demonstrated to be a bifunctional enzyme in Escherichia coli (E.coli) which not only converted H2Pte into H2PteGlu, but also added glutamic acid tail to H4PteGlu [18,19].Therefore, we speculated that H4PteGlu would also compete with H2PtePAS for catalysis activity of FolC in M. tuberculosis, thus hindering the activation process of PAS.To test this speculation, catalytic efficiency of FolC on H2Pte, H4PteGlu, and H2PtePAS was compared.The results showed that, under the same reaction conditions, FolC could convert about 85% H2Pte and 50% H4PteGlu, but only about 12% H2PtePAS (Figure 4).

Comparison of Catalytic Efficiency of FolC on H 2 Pte, H 4 PteGlu and H 2 PtePAS
FolC was demonstrated to be a bifunctional enzyme in Escherichia coli (E.coli) which not only converted H 2 Pte into H 2 PteGlu, but also added glutamic acid tail to H 4 PteGlu [18,19].Therefore, we speculated that H 4 PteGlu would also compete with H 2 PtePAS for catalysis activity of FolC in M. tuberculosis, thus hindering the activation process of PAS.To test this speculation, catalytic efficiency of FolC on H 2 Pte, H 4 PteGlu, and H 2 PtePAS was compared.The results showed that, under the same reaction conditions, FolC could convert about 85% H 2 Pte and 50% H 4 PteGlu, but only about 12% H 2 PtePAS (Figure 4).

H4PteGlu Hinders the Activation of PAS by FolC
To further demonstrate whether H4PteGlu could hinder the activation of PAS by FolC, H2PtePAS was synthesized by purified recombinant FolP1 using H2PtePP and PAS as substrates [10].H2PtePAS was analyzed by UPLC-MS/MS (Figure 5A, Supplementary

H 4 PteGlu Hinders the Activation of PAS by FolC
To further demonstrate whether H 4 PteGlu could hinder the activation of PAS by FolC, H 2 PtePAS was synthesized by purified recombinant FolP1 using H 2 PtePP and PAS as substrates [10].H 2 PtePAS was analyzed by UPLC-MS/MS (Figure 5A, Supplementary Table S1).FolC catalytic activity was analyzed using H 2 PtePAS instead of H 2 Pte as a substrate.Consistent with previous reports [5,10], FolC could catalyze the ligation of L-glutamic acid to H 2 PtePAS generating H 2 PtePAS-Glu, which was confirmed by HPLC-MS/MS (Figure 5B, Supplementary Table S1).We then sought to understand the effect of H 4 PteGlu on H 2 PtePAS activation by FolC, and different concentrations (10 µM and 50 µM) of H 4 PteGlu were added into the FolC reaction mixture using H 2 PtePAS as substrate.As shown in Figure 5C, when H 4 PteGlu was added into the reaction mixture, the catalytic efficiency of FolC for H 2 PtePAS decreased remarkably.

H4PteGlu Hinders the Activation of PAS by FolC
To further demonstrate whether H4PteGlu could hinder the activation of PAS by FolC, H2PtePAS was synthesized by purified recombinant FolP1 using H2PtePP and PAS as substrates [10].H2PtePAS was analyzed by UPLC-MS/MS (Figure 5A, Supplementary Table S1).FolC catalytic activity was analyzed using H2PtePAS instead of H2Pte as a substrate.Consistent with previous reports [5,10], FolC could catalyze the ligation of L-glutamic acid to H2PtePAS generating H2PtePAS-Glu, which was confirmed by HPLC-MS/MS (Figure 5B, Supplementary Table S1).We then sought to understand the effect of H4PteGlu on H2PtePAS activation by FolC, and different concentrations (10 µM and 50 µM) of H4PteGlu were added into the FolC reaction mixture using H2PtePAS as substrate.As shown in Figure 5C, when H4PteGlu was added into the reaction mixture, the catalytic efficiency of FolC for H2PtePAS decreased remarkably.

Discussion
Folates, especially derivatives of H 4 PteGlu, are one carbon carriers required by the biosynthesis of purines, thymidylate, methionine, serine, and glycine, thus making them essential for all sorts of lives [20,21].Bacteria must synthesize these essential cofactors de novo, while mammal can intake them from their diet [4].This difference makes the bacterial de novo folate biosynthesis pathway an ideal target for developing new antibacterial drugs [4].Although thousands of folates antagonists have been designed for folate biosynthesis pathway heretofore, PAS is the only one used for TB treatment with a unique mode of action only observed in M. tuberculosis complex.Thus, better understanding the mechanisms of PAS resistance in M. tuberculosis will benefit the development of new antifolates against this bacterium.
As the first molecular marker for PAS resistance in M. tuberculosis clinical isolates, thyA gene mutations have been identified for nearly two decades [13], but the molecular mechanism of how these mutations lead to PAS resistance remains unknown.Ten years later, when probing the molecular mechanism of PAS resistance caused by folC mutation [10], we noticed that though FolC could also catalyze the conversion of H 2 PtePAS to H 2 PtePAS-Glu, but the catalytic efficiency was much lower than that of the natural substrate H 2 Pte, implying that the bio-activation process of PAS might be vulnerable to interference of natural metabolite of folate biosynthesis.Indeed, exogenous H 2 Pte made M. tuberculosis more resistant to PAS [10].Previous studies have showed that FolC could not only convert H 2 Pte into H 2 PteGlu, but also add glutamic acid tail to H 4 PteGlu in E. coli [18,19].In this study, we found that M. tuberculosis FolC is also bifunctional.In addition, its catalytic efficiency for H 2 PtePAS is remarkably lower than that for H 4 PteGlu (Figure 4), implying intracellular H 4 PteGlu may interfere the activation of PAS by FolC.As expected, the in vitro biochemical experiments showed that H 4 PteGlu hinders the conversion of H 2 PtePAS to H 2 PtePAS-Glu in a concentration-dependent manner (Figure 5C).Since M. tuberculosis is not able to intake exogenous H 4 PteGlu, it is not possible to test the effect of exogenous H 4 PteGlu on PAS susceptibility.Alternatively, we compared the H 4 PteGlu content between H37Ra and the thyA deletion mutant, and found that the H 4 PteGlu content in the thyA deletion mutant was significantly higher than that of the wild-type strain (Figure 3).This is not surprising since the bacterium has to solely rely on ThyX to synthesize thymidylate in the absence of ThyA, and utilization of the former yields H 4 PteGlu.Since the expression level of FolC remained unchanged in the thyA deletion mutant, increased H 4 PteGlu content could hinder the conversion of H 2 PtePAS since they compete for the same protein.Correspondingly, this competition could be mitigated by over-expression of the target protein FolC.As expected, over-expression of folC could reverse the PAS resistance phenotype caused by thyA deletion or clinical thyA R235P mutation (Table 2).We noticed that the PAS resistance phenotype caused by thyA deletion or mutation could only be partially restored by folC over-expression, suggesting the existence of other mechanisms for PAS resistance caused by functional deficiency of ThyA.
When assessing whether the resistance to PAS of the thyA deletion mutant was related to the efficiency of PAS activation, we over-expressed folP1, folC, and dfrA in H37Ra and H37Ra ∆thyA.To our surprise, over-expression of dfrA in the thyA deletion mutant did not affect the susceptibility to PAS (Table 2), suggesting either the lack of DfrA protein or loss of function of DfrA in the thyA deletion mutant.Previous works also showed that thyA and dfrA double deletion mutants had been identified in M. tuberculosis clinical isolates from different countries [11,22].Thus, in the absence of thyA, M. tuberculosis discards the commonly used DHFR Rv2763c (DfrA), and switches to another alternative to synthesize H 4 PteGlu.Although RibD was shown to be an alternative DHFR in M. tuberculosis, previous research revealed that RibD could only replace DfrA when it was highly over-expressed in a multi-copy plasmid [8], suggesting that the dihydrofolate reductase activity of RibD is quite low, which was confirmed by subsequent biochemical analysis [23].Zheng et al. found that mutations in the promoter region of ribD could cause over-expression of ribD [8].To determine whether RibD is the alternative DHFR in the absence of ThyA in M. tuberculosis, we further analyzed genome sequences of isolates with frameshift or deletion mutations in thyA or dfrA genes from previous studies and NCBI database.The results showed that there was no mutation in either the promoter region (300 bp upstream start codon) or the coding sequence (CDS) of the ribD gene in ThyA or DfrA deficient clinical isolates (Supplementary Table S2).Moreover, our RNA-seq data also showed that the expression level of ribD remained unchanged in the thyA deletion mutant (Figure S2).Therefore, RibD is not the alternative DHFR in the absence of ThyA.What the alternative DHFR is in the absence of ThyA requires further investigation.It will be important to test if the alternative DHFR would be more resistant to the inhibition of H 2 PtePAS-Glu, since over-expressing folC could only partially restore PAS sensitivity to the thyA deletion mutant.
Previous studies already showed that the C −16 T mutation in the upstream regulatory region of thyX could lead to increased expression of thyX and PAS resistance in M. tuberculosis [24,25].Thus, it is not surprising to see that over-expressing thyX led to PAS resistance in H37Ra.The fact that over-expressing thyX in the thyA deletion mutant did not affect PAS susceptibility of the latter indicated that over-expressing thyX and deleting thyA in Mtb might share the same mechanism of PAS resistance.In addition, over-expressing thyA in H37Ra also led to low level PAS resistance.Considering the role of ThyA in folate salvage, we speculated that the intracellular H 2 PteGlu content might be increased when over-expressing thyA; this would in turn reduce the demand for dihydrofolate biosynthesis through FolC.Previous studies showed that FolC was critical for the bio-activation of PAS, and decreased FolC enzymatic activity caused PAS resistance [8,10].
In conclusion, our results showed that functional deficiency of ThyA led to increased H 4 PteGlu content of the bacterial cells, which competed with H 2 PtePAS for FolC catalysis, thus hindered the activation of PAS and conferred PAS resistance in M. tuberculosis.Meanwhile, our study also suggested that M. tuberculosis could switch from Rv2763c to a yet unknown alternative DHFR in the absence of thyA, and further investigation is required to identify the protein and elucidate its role on PAS resistance caused by ThyA functional deficiency.Our study broadens the understanding of folate metabolism in M. tuberculosis and might be useful for guiding the clinical administration of PAS.
A modified strategy for phage-mediated allelic exchange [26] was used to construct M. tuberculosis H37Ra ∆thyA mutant.Briefly, the native copy of thyA was deleted by specialized transduction using phAE159 containing a hygromycin resistance cassette.All primers used are listed in Supplementary Table S3.

Purification of Recombinant FolP1 and FolC
FolP1 and FolC proteins were purified as previously reported [10].Briefly, folP1 and folC were amplified from M. tuberculosis H37Ra genomic DNA using specific primers (Supplementary Table S3) and separately cloned into pET28a to yield pET28a::folP1 to introduce an N-terminal hexa-histidine tag and into pMAL-c2X to yield pMAL-c2X::folC to introduce an N-terminal maltose-binding protein (MBP) tag linked with a factor Xa cleavage site.The sequence-confirmed recombinant plasmids were transformed into E. coli BL21 (DE3).The cells were grown at 37 • C in LB broth containing 150 µg mL −1 ampicillin or 100 µg mL −1 kanamycin to an OD 600 of ~0.6.Isopropyl-β-D-thiogalactopyranoside (IPTG, Acmec, China) was added to 0.25 mM, then the cells were incubated further at 16 • C for 20 h.The bacterial cells were harvested by centrifugation, disrupted by sonication, and clarified by centrifugation.
Recombinant FolC proteins were first purified over an amylose resin column (New England BioLabs).The FolC protein obtained from the first purification contains MBP tag.To remove the MBP tag, the purified samples were incubated with factor Xa at 4 • C overnight in reaction buffer (20 mM HEPES (pH 8.0), 100 mM NaCl, 2 mM CaCl 2 , and 10% glycerol).Then, the cleavage mixtures were dialyzed against 50 mM phosphate buffer (pH 8.0).The samples were loaded on a HiTrap DEAE FF column (GE Healthcare), and a step gradient from 50 mM to 1 M NaCl in phosphate buffer was applied to elute FolC.The fractions were then analyzed by SDS-PAGE.Recombinant FolC was eluted with 300 mM NaCl.

Western Blot Assay
The H37Ra and H37Ra ∆thyA strains were cultured at 37 • C in 10 mL of 7H9 medium and harvested at logarithmic phase by centrifugation.For Western blot analysis, bacterial cells were resuspended in phosphate buffer saline (PBS, pH 7.0), then lysed using zirconium beads.Protein samples acquired from the supernatant after centrifugation.The protein concentration of the supernatant was determined using the NanoDrop2000 (Thermo, Waltham, MA, USA).Then, the protein samples were separated by SDS-PAGE and immediately transferred to a polyvinylidene difluoride membrane (Merck Millipore, Darmstadt, Germany) by a Bio-Rad SD device (Bio-Rad Laboratories, Hercules, CA, USA) at 15 V for 30 min.Finally, the proteins were probed with rabbit FolC polyclonal antibody (ABclonal biotechnology, Wuhan, China, Cat.No. WG-00133D).

RNA-Seq Analysis
Mycobacterial strains were grown in 7H9 to mid logarithmic phase and were collected by centrifugation.Total RNA was extracted using RNeasy mini kit (Qiagen, Hilden, Germany).Library constructions were prepared using TruSeq Stranded Total RNA Sample Preparation kit (Illumina, San Diego, CA, USA), and RNA sequencing was conducted on Illumina NovaSeq6000 at Beijing Novogene Corporation.The insert size conformation of purified libraries was validated by an Agilent 2100 bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).Bowtie2 was used to map the cleaned reads to the M. tuberculosis H37Ra genome acquired from the National Center for Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/nuccore/CP000611.1)(Accessed on 25 May 2023).Then, HTSeqv0.6.1 was run with a reference annotation to generate fragments per kilobase of exon model per million mapped reads values for estimation of fold changes.Three biological (v/v) formic acid).The bacterial biomasses of the individual samples were determined by colony counting method.All data obtained by metabolomics were averaged from the independent sextuplicates.p-values (p) were calculated using t-tests.The graphs for the determination of H 4 PteGlu in vivo were prepared using GraphPad Prism.4.10.Comparative Analysis of Variants in M. tuberculosis Genomes M. tuberculosis clinical isolates with complete or partial deletion of thyA or dfrA were extensively collected from previous studies [11,22,27] and the NCBI database (https://www.ncbi.nlm.nih.gov/genome/browse#!/prokaryotes/mycobacterium%20tuberculosis) (Accessed on 7 December 2022).A total of 31 M. tuberculosis genomes from clinical isolates were obtained, and the mutations in the promoter region (300 bp upstream start codon) or the CDS of ribD were analyzed in these isolates (Supplemental Table S2).All of the raw reads were available.The acquired reads were subjected to quality assessment using FastQC v.0.11.9.Subsequently, low-quality sequences were removed and trimmed using fastp.Reads shorter than 50 bp were discarded, the last 10 bp were trimmed, and bases with an average quality below 25 were removed using a sliding window of 20 bp.Finally, variant calling against the M. tuberculosis H37Rv (NC_000962.3)genome was performed using the Snippy pipeline.

Statistical Analysis
GraphPad Prism 8.0.1 was used to analyze all experimental data, adopting the twotailed unpaired t-test method.Mean ± standard deviation (SD) was adopted to express the experimental data.

Figure 2 .
Figure 2. The expression of folC remains unchanged in ThyA functional deficient strain.(A) Comparison of the expressional level of FolC during the exponential phase in H37Ra (WT) and H37Ra ∆thyA (thyA − ) by Western blot assay.Upper part: Total protein was normalized to 25 µg of each strain, then electrophoresed by SDS-PAGE and stained by Coomassie brilliant blue.Lower part: Western blot analysis of total protein immunoblotted with rabbit FolC polyclonal antibody.Experiments were repeated at least three times, and were performed three biological replicates each time.(B) Relative quantitative of FolC product by Western blot assay.ns, no significance.(C) Comparison of the transcriptional level of the gene folC during the exponential phase in H37Ra (WT) and H37Ra ∆thyA (thyA − ) by RNA-seq.ns, no significance.

Table 2 .
Over-expression of folC gene reverses the PAS resistance phenotype.

Table 2 .
Over-expression of folC gene reverses the PAS resistance phenotype.