AC2P20 selectively kills M. tuberculosis at acidic pH by depleting free thiols

Mycobacterium tuberculosis (Mtb) senses and adapts to host immune cues as part of its pathogenesis. One environmental cue sensed by Mtb is the acidic pH of its host niche in the macrophage phagosome. Disrupting the ability of Mtb to sense and adapt to acidic pH has the potential to reduce survival of Mtb in macrophages. Previously, a high throughput screen of a ∼220,000 compound small molecule library was conducted to discover chemical probes that inhibit Mtb growth at acidic pH. The screen discovered chemical probes that kill Mtb at pH 5.7 but are inactive at pH 7.0. In this study, AC2P20 was prioritized for continued study to test the hypothesis that it was targeting Mtb pathways associated with pH-driven adaptation. RNAseq transcriptional profiling studies showed AC2P20 modulates expression of genes associated with redox homeostasis. Gene enrichment analysis revealed that the AC2P20 transcriptional profile had significant overlap with a previously characterized pH-selective inhibitor, AC2P36. Like AC2P36, we show that AC2P20 kills Mtb by selectively depleting free thiols at acidic pH. Mass spectrometry studies show the formation of a disulfide bond between AC2P20 and reduced glutathione, supporting a mechanism where AC2P20 is able to deplete intracellular thiols and dysregulate redox homeostasis. The observation of two independent molecules targeting free thiols to kill Mtb at acidic pH further supports that Mtb has restricted redox homeostasis and sensitivity to thiol-oxidative stress at acidic pH.


Introduction 44
Mtb pathogenesis is driven by its ability to exploit and adapt to the intracellular host 45 environment. During pathogenesis, Mtb encounters a variety of stressors including nitrosative, 46 oxidative, acidic pH, and hypoxic stress [1]. In response to these stresses, Mtb alters its physiology 47 in order to survive the hostile macrophage environment and modulate expression of virulence 48 genes critical for its pathogenicity. Acidic pH is an initial environmental cue that Mtb senses upon 49 infection of the host macrophage [2,3]. For survival within the resting macrophage, Mtb inhibits 50 fusion of the phagosome and lysosome and resides in a mildly acidic phagosome (pH 6.4) [4]. 51 Activation of the macrophage leads to phagosome acidification and Mtb resists this acid stress, 52 maintaining a relatively neutral cytoplasmic pH, even at pH <5.0 [5][6][7][8]. In addition to expressing 53 mechanisms to survive acid stress, Mtb also exhibits pH-and-carbon source dependent growth 54 adaptations. Mtb will completely arrest its growth in minimal media buffered to pH 5.7 with 55 glycerol as the sole carbon source [9]. During this growth arrest, Mtb exhibits carbon specificity, 56 and will only arrest growth on glycolytic carbon sources (i.e. glucose and glycerol) [9]. However, 57 when given specific carbon sources (i.e. phosphoenolpyruvate, pyruvate, acetate, oxaloacetate, and 58 cholesterol), Mtb resuscitates its growth at pH 5.7 in minimal media, and thus, exhibits direct 59 metabolic remodeling during pH stress [9]. Collectively, these studies show that in response to 60 acidic pH, Mtb has multiple mechanisms in place whereby it alters its physiology for survival and 61 virulence. 62 When Mtb is cultured at acidic pH or in macrophages, the bacterium has an imbalanced 63 redox state with a more reduced cytoplasm [9,10], a phenomenon referred to as reductive stress 64 [3,11]. It is hypothesized that acidic pH may cause redox imbalances due to adaptations of the 65 electron transport chain that promote oxidative phosphorylation while maintaining cytoplasmic In this study, we report on a new chemical probe isolated from a prior screen, AC2P20 (N-113 1,3-benzothiazol-2-yl-2-[(4,6-dioxo-5-phenyl-1,4,5,6-tetrahydropyrimidin-2-yl)thio]acetamide) 114 ( Figure 1A), that selectively kills Mtb at acidic pH. AC2P20 was identified as a phoPR-115 independent, pH-selective inhibitor of Mtb growth. Through transcriptional profiling we observed 116 that genes modulated by AC2P20 treatment significantly overlap with genes modulated by 117 AC2P36 treatment. Although both compounds are structurally distinct, like AC2P36, AC2P20 also 118 exhibits killing of Mtb at pH 5.7, disrupts thiol homeostasis by depleting intracellular free thiol 119 pools, and increases reactive oxygen ROS production. Thus, AC2P20 is a second structurally 120 unique pH-selective chemical probe that exhibits thiol-depletion as a mechanism-of-action for 121 killing at acidic pH. This finding further reinforces the vulnerability of Mtb to perturbations of 122 Mtb CDC1551 and Mtb Erdman strains were grown to an OD600 of 0.6-1.0, spun down, and 139 resuspended in 7H9 media buffered to pH 5.7. Mtb cells were plated at 10 9 cells/mL on 7H10 agar 140 media buffered to pH 5.7 and supplemented with 10 µM, 20µM or 40 µM AC2P20. Plates were 141 incubated at 37°C for over 12 weeks without any significant isolated colonies appearing. This 142 experiment was performed three times with similar results. 143 144

Transcriptional profiling and data analysis 145
Mtb cultures were grown at 37°C and 5% CO2 in standing T-25 culture flasks to an OD600 of 0.5 146 in 8 mL of 7H9 buffered media. Treatment conditions examined include i) 20µM AC2P20 at pH 147 5.7 and ii) an equivalent volume of DMSO at pH 5.7 as the baseline control. Each culture was 148 incubated for 24 hours and treatment conditions were conducted in two biological replicates. 149 Following incubation, total bacterial RNA was extracted as previously described [2,9] and 150 sequencing data was analyzed using SPARTA (ver. 1.0) [30]. Genes identified filtered based on 151 log2CPM < 5 and log2FC < 1. A Chi-Square analysis with Yates Correction was conducted to test 152 the statistical relationship between gene overlap with the AC2P36 transcriptional profile as 153 described by Coulson et al. [29]. The RNAseq data has been deposited at the GEO database 154 (Accession # GSE151884). 155

Half-maximal effective concentration (EC50) determination and spectrum of activity in other 157
mycobacteria 158 Mtb cultures were incubated in buffered 7H9 media (pH 5.7 or pH 7.0) at a starting OD600 of 0.2, 159 with 200 uL aliquoted into 96-well microtiter assay plates (CoStar #3603). Cultures were treated 160 with a 2.5-fold dose-response of AC2P20 (80 µM-0.13 µM) and incubated standing for 6 days at 161 37 °C and 5% CO2, with bacterial growth assessed by optical density (OD600 Mtb was initially cultured in 7H9 media (pH 5.7 or 7.0) at a starting OD600 of 0.2 in 96-well assay 172 plates. Cultures were treated with a 2.5-fold dose-response of AC2P20 (80 µM-0.33 µM). An 173 equivalent volume of DMSO was included as a control. Each treatment condition was conducted 174 in triplicate and incubated for 7 days. Following incubation, treated wells were serially diluted in 175 1X Phosphate-Buffered Saline (PBS) and plated for colony forming units (CFUs) on 7H10 agar 176 plates supplemented with 10% OADC and glycerol. Bactericidal activity was determined by 177 comparing CFUs from the initial inoculum to CFUs following treatment. 178

Cytoplasmic pH-homeostasis 180
Mtb washed with PBS (pH 7.0) was labelled with Cell Tracker 5'-chloromethylfuoroscein 181 diacetate (CMFDA) and analyzed using methods previously described [31]. Mtb treated with 182 AC2P20 in PBS (pH 5.7) was assayed for cytoplasmic pH changes over the course of 24-hours. 183 Excitation ratio results were converted to pH via a standard curve generated using Nigericin- chemical library of >220,000 small molecules was previously screened, with compound hits being 232 defined as those that inhibited reporter fluorescence or Mtb growth. These compounds were further 233 classified as TCS target inhibitors or growth inhibitors. The screens only differed in the reporter 234 strain used and the pH of the medium, which was neutral or acidic in the DosRST and PhoPR 235 inhibitor screens, respectively. Comparing growth inhibiting hits from these two screens identified 236 a subset of compounds that selectively inhibited Mtb growth at acidic pH independent of PhoPR 237 signaling. These compounds were classified as pH-selective growth inhibitors if they exhibited 238 >50% growth inhibition at acidic pH and < 10% inhibition at neutral pH. AC2P20 (N-1,3-239 benzothiazol-2-yl-2-[(4,6-dioxo-5-phenyl-1,4,5,6-tetrahydropyrimidin-2-yl)thio]acetamide) 240 ( Figure 1A) exhibited >5-fold selectivity at acidic pH and was characterized as one of these pH-241 selective inhibitors of Mtb growth. The pH-dependent activity of AC2P20 was confirmed by 242 determining its half-maximal effective concentration (EC50). Mtb treated with an 8-point dose-243 response of AC2P20 for six days at pH 5.7 results in dose-dependent growth inhibition with an 244 EC50 of 4.3 μM, however, has a >10-fold higher EC50 of ~60 μM at pH 7.0 ( Figure 1B). AC2P20 245 also exhibits mycobacterial selectivity for Mtb compared to M. smegmatis, which has an EC50 > 246 80 μM at acidic pH and does not exhibit growth inhibitory activity at neutral pH ( Figure S1A). 247 Time-dependent and concentration-dependent killing assays were conducted to define whether 248 AC2P20 is bactericidal or bacteriostatic. Mtb treated with 20 μM AC2P20 exhibits pH-selective 249 inhibition of Mtb growth in acidic conditions and results in approximately 2-log fold reduction in 250 CFUs over 5 days ( Figure 1C). In contrast, DMSO controls and AC2P20 treatment in neutral 251 conditions have no impact on growth. The concentration-dependent killing assay shows that 252 AC2P20 is bactericidal at ~32 μM and bacteriostatic at 12 μM ( Figure 1D). Cytoplasmic pH was 253 measured to determine whether AC2P20 functions as an ionophore. Treatment with AC2P20 does 254 not modulate the cytoplasmic pH of Mtb compared to the nigericin positive control ( Figure S1B). 255 Together, these data show that AC2P20 activity is pH-dependent, bactericidal, and does not alter 256 Mtb cytoplasmic pH homeostasis. 257 258 AC2P20 induces a thiol oxidative stress response similar to AC2P36. To isolate resistant mutants 259 and thereby find potential targets for AC2P20, 10 9 Mtb cells were plated on 7H10 agar media 260 buffered to pH 5.7 containing 10 µM, 20 µM or 40 µM AC2P20. Despite several weeks of 261 incubation each time at 37°C, no spontaneous mutants were isolated from multiple rounds of 262 screening for resistant mutants to AC2P20. Following our resistance screening attempts, 263 transcriptional profiling was conducted to define Mtb physiologies targeted following AC2P20 264 treatment. Mtb CDC1551 cultures were prepared in rich media (pH 5.7) and treated with 20µM 265 AC2P20 or DMSO control for 24 hours. Mtb treated with AC2P20 caused induction of 156 genes 266 (>2-fold, q < 0.05) and repression of 81 genes (>2-fold, q < 0.05) (Figure 2A, Table S1). Using 267 MycoBrowser [35] to classify gene function, we found that the functional pathway most induced 268 by AC2P20 (excluding conserved hypotheticals) was intermediary metabolism and respiration 269 ( Figure 2B, Table S1A, Table S1B). Differentially induced genes included genes involved in sulfur metabolism (cysT, sirA, mec), transcriptional regulation of the stress response (sigH, sigB, rshA), 271 and redox homeostasis (katG, trxB1, trxC) ( Figure 2C, Table S1B). Notably, differentially 272 regulated genes from AC2P20 treated cells overlapped with differential gene expression profiles 273 previously characterized for the pH-selective Mtb growth inhibitor, AC2P36 [29]. Gene 274 enrichment analysis showed a statistically significant overlap between groups AC2P20 and 275 category most induced (excluding conserved hypotheticals) was intermediary metabolism and 300 respiration, the same as AC2P20. However, major differences were noted between categories of 301 both induced gene sets for AC2P20 and AC2P36. For example, induction of lipid metabolism 302 genes comprised roughly 3.33% of the total genes induced by AC2P36 compared to 12.82% for 303 AC2P20 ( Figure 2B, Figure S2A). Noticeably, AC2P20 appeared to upregulate several mycolic 304 acid biosynthesis pathway and operon genes (fas, acpM, kasA, accD6) ( Figure S2B). In contrast, 305 these genes were repressed following AC2P36 treatment. Other lipid metabolism genes not 306 observed in AC2P20 transcriptional data, but actively repressed by AC2P36 include scoA/B, 307 accD1, Rv3087, and fadE35 [29]. Additionally, transcriptional profiling showed that methylcitrate 308 synthase and methylcitrate dehydratase genes (prpC and prpD, respectively) were oppositely 309 modulated in both regulons; AC2P20 repressed prpC/D expression while their expression was 310 induced by AC2P36 ( Figure S2B). Other functional categories that saw large quantitative changes 311 between both transcriptional profiles include cell wall and cell wall processes and virulence, 312 detoxification and adaptation. Fewer cell wall and cell wall processes genes were induced by 313 AC2P36 compared to AC2P20, while the number of virulence, detoxification and adaptation 314 functional genes were increased following AC2P36 treatment ( Figure S2A). The transcriptional 315 differences observed between both regulons demonstrates that despite the shared similarities in 316 regulation of thiol-redox homeostasis and regulatory genes, distinct differences exist between how 317 pathways are modulated following AC2P20 and AC2P36, with lipid metabolism being most 318 notable. 319 320 AC2P20 forms an adduct with the low molecular weight thiol, GSH. Although AC2P36 and 321 AC2P20 have distinct structures, both compounds contain a similar thiol-containing pyrimidine 322 group. In AC2P36, it is thought that the methylsulfone moiety acts as an electron-withdrawing 323 group which allows a thiolate anion to undergo nucleophilic attack on the C-2 carbon of the 324 pyrimidine ring in order to release methanesulfinic acid or methanesulfinate ( Figure S1C) [29]. 325 This interaction is thought to result in the formation of a sulfide bond and depletion of available 326 free thiols. Indeed, heteroaromatic sulfones have been recently described as tunable agents for 327 cysteine-reactive profiling [45,46]. Based on these observations with AC2P36, and the noted 328 similarity with the thiol-containing pyrimidine group, we hypothesized that AC2P20 may have a 329 similar mechanism of action and undergo covalent modification of free thiols. To test this 330 hypothesis, 80µM AC2P20 was incubated with 100µM reduced glutathione (GSH) for one hour 331 in basic, neutral, and acidic conditions and analyzed via mass spectrometry. Incubation of AC2P20 332 with GSH resulted in the formation of an adduct at pH 5.7 with an exact molecular weight of ~529 333 Da ( Figure 3A, Figure 3C, Figure S5). There is also adduct formation in neutral and basic 334 conditions ( Figure S3A, Figure S3B) although with lower peak intensity. AC2P20 incubated with 335 DMSO does not appear to fragment in the absence of GSH in any of these conditions. ( Figure 3B, 336 Figure S3C, Figure S3D). In the positive ESI mode ( Figure 3C), a neutral fragment of 129 Da is 337 lost from the adduct with a peak seen at ~401 Da, consistent with a loss of the glutamate fragment 338 from GSH [47]. Fragmentation of AC2P20 is also observed when incubated with GSH at pH 5.7, with peaks at ~222 Da, ~206 Da, ~194 Da, and ~178 Da aligning with possible fragments of the 340 pyrimidine group of AC2P20 ( Figure S5). The peak observed at ~391 Da is a mass spectrometry 341 plasticizer and common contaminant that can be used for mass calibration [48]. We also looked at 342 N-acetylcysteine (NAC), a derivative of GSH, and its ability to form an adduct with AC2P20. A 343 peak was observed at ~384 Da, aligning with the formation of an AC2P20-NAC adduct (Figures 344 S4A, Figure S5). Interestingly, higher peak intensities of these adducts were observed at neutral 345 and basic conditions ( Figure S4B, Figure S4C). This is possibly due to NAC having a pKa ~9.5, 346 and therefore favoring the adduct reaction with AC2P20 under these conditions. Together, these 347 findings support that AC2P20 reacts with low molecular weight thiols and thiol groups. 348 Additionally, we looked at whether AC2P20 still form an adduct with GSH in the presence of the 349 oxidant, H2O2. It was thought that H2O2 may cause the formation of intermediate sulfenic acid and 350 oxidize GSH, resulting in the formation of glutathione (GSSG) [49]. After incubating AC2P20 351 with both GSH and H2O2, we still observed disulfide bond formation between AC2P20 and GSH, 352 indicating that GSSG is probably not being produced ( Figure S4D). These results suggest that 353 AC2P20 is capable of forming a disulfide bond with low molecular weight thiols. 354

AC2P20 depletes free thiols and causes an accumulation in ROS in Mtb at acidic pH. Given that 356
an adduct is able to form between AC2P20 and GSH, we sought to test the ability of AC2P20 to 357 deplete free thiols in Mtb. For this assay, Mtb was treated with AC2P20 (2 μM and 20 μM) in both 358 acidic and neutral conditions for 24 hours. Auranofin (20 μM) was used as a positive control 359 because it inhibits Mtb's thioredoxin reductase, TrxB2, thereby disrupting thiol-and redox-360 homeostasis [23]. AC2P36 (20 μM) was also included in the assay to compare thiol depleting 361 activities of both compounds. Following AC2P20 treatment, a statistically significant reduction in free thiol concentrations was observed intracellularly in Mtb at pH 5.7 where free thiols are 363 reduced by ~2.8-fold to ~133nM compared to the DMSO vehicle control at ~380 nM ( Figure 4A). 364 As expected, we also see free thiol depletion in Mtb following treatment with both positive 365 controls, supporting the observation seen with AC2P20. In contrast to Auranofin, AC2P20 366 treatment at neutral pH does not exhibit any statistically significant reduction in free thiols, 367 supporting the pH-selective activity of this compound. Interestingly, AC2P36 does exhibit some 368 activity in neutral conditions. This is possibly due to AC2P36 still exhibiting some growth 369 inhibitory activity at neutral pH at ~30 μM, whereas AC2P20 requires much higher concentrations 370 (~60 μM) to see a similar inhibitory effect. 371 Depletion of total free thiols will result in disrupted redox homeostasis and therefore may 372 result in enhanced ROS accumulation. To test this hypothesis, we conducted an assay measuring 373 intracellular ROS production in Mtb. Mtb was incubated with 2 μM and 20 μM AC2P20 for 24 374 hours, treated with CellROX fluorescent dye for 1 hour, and then assayed for relative fluorescence 375 and optical density. AC2P36 (2 μM and 20 μM) was included as the positive control, because it 376 has previously been shown to accumulate intracellular ROS following treatment. At acidic pH, 20 377 μM AC2P20 exhibits ~3-fold increase in intracellular ROS production compared to DMSO (Figure  378 4B). AC2P20 (20 μM) also increases ROS accumulation ~3-fold greater in acidic conditions 379 compared to neutral pH, where there is little ROS accumulation compared to DMSO. AC2P36 (20 380 μM) also increases ROS production ~2-fold at pH 5.7, which is consistent with previous 381 observations. These data support a mechanism whereby enhanced ROS accumulation can be 382 driven by pH stress and is further exacerbated by AC2P20 treatment. 383 384 Discussion 386 Based on the chemical structure of AC2P20 and the adduct it forms with GSH at pH 5.7, 387 we propose a reaction model where the benzothioazole-mercaptoacetamide group covalently 388 modifies free thiols, forming stable adducts. Shown here is a potential mechanism for the 389 generation of adducts observed by mass spectrometry (Figure 3A, Figure 3C). Disulfide bond 390 formation between GSH (307.32 Da) and the free benzothioazole-mercaptoacetamide group 391 (223.29 Da) results in a molecule mass of 529 Da, which can be observed in both positive and 392 negative ESI modes ( Figure 5A, Figure S5). Loss of the neutral glutamate fragment from the 393 AC2P20-GSH adduct results in a peak at 401 Da (ESI+). We suspect AC2P20 may be undergoing 394 hydrolysis, however, we do not observe the phenyl-dioxopyrimidine fragment (204 Da). We do 395 observe a fragmented phenyl-dioxopyrimidine group at 178 Da which may be due to the sample's 396 molecules breaking into charged fragments during mass spectrometry. The absence of a 204 Da 397 fragment may also suggest that adduct formation could be occurring via a different chemical 398 process. However, the observation of an adduct supports that the formation of disulfide bonds 399 between AC2P20 and other thiol-containing molecules could be occurring in Mtb ( Figure 5B). 400 Although, both AC2P20 and AC2P36 function by depleting free thiols, the two scaffolds 401 are distinctly different and engage glutathione (GSH) in different ways. AC2P36 is itself an 402 electrophile, by virtue of the reactive methanesulfonyl moiety on the pyrimidine. GSH can add 403 directly to AC2P36 on the pyrimidine, followed by elimination of the excellent leaving group 404 methanesulfinic acid [29]. On the other hand, AC2P20 is not itself reactive to GSH in an analogous 405 fashion, as evidenced by a lack of MS ion for a direct adduct of GSH to the pyrimidine dione 406 moiety. Instead, AC2P20 has to get hydrolyzed to the free thiol, after which it forms a disulfide with GSH. Therefore, AC2P20 and AC2P26 have different chemical mechanisms of action, and 408 the GSH adducts are chemically distinct (e.g. disulfide vs thiopyrimidine). imbalance. Thus, targeting redox-homeostasis represents an important new approach to treating 418 TB. Like AC2P36, we have discovered a second, albeit novel, pH-selective compound (AC2P20) 419 that directly targets free thiols to perturb redox homeostasis. Both AC2P36 and AC2P20 deplete 420 free thiol pools and increase intracellular ROS as part of their killing mechanisms. Interestingly, 421 AC2P20 depletes less free thiols than AC2P36, but has a greater increase in intracellular ROS. 422 This suggests that although both appear to target Mtb free thiols, there are differences in their 423 mechanisms. One hypothesis is that release of the phenyl-dioxopyrimidine group could also be 424 targeting a secondary unknown Mtb physiology, possibly explaining the higher ROS increase that 425 is observed compared to AC2P36 ( Figure 4B). Both compounds also form adducts with the low 426 molecular weight thiol, GSH; however, there are major chemical scaffold differences. AC2P36 427 captures thiols with the release of methylsulfinate while AC2P20 is cleaved to generate 428 benzothioazole-mercaptoacetamide, which then goes on to form disulfide bonds. Although 429 AC2P20 and AC2P36 compounds are structurally unique and have distinct mechanisms-of-action, they do exhibit similar physiological effects on Mtb, supporting the conclusion that thiol redox 431 homeostasis is specifically vulnerable to inhibition at acidic pH. 432 Several studies in Mtb show a link between low pH-and oxidative stress responses 433 [7,9,29,50,52]. At acidic pH in vitro, Mtb exhibits a more reduced cytoplasm and a shift from 434 glycolysis to fatty acid synthesis [9]. This metabolic remodeling is thought to occur in order to 435 generate more oxidized cofactors to mitigate reductive stress. However, a more reduced cytoplasm 436 in Mtb may also play a role in protecting Mtb against oxidative stress. A recent study comparing 437 the RNAseq profiles of reduced MSH redox potential (EMSH-reduced), intraphagosmal Mtb, and 438 pH stress supports this claim and shows that EMSH-reduced transcriptome has significant overlap 439 with the pH-regulon[50]. When we compare the EMSH-reduced, intraphagosomal Mtb, and pH 440 stress regulons with AC2P20 and AC2P36 transcriptional profiles, we again see overlap in redox 441 sensitive genes (i.e. katG, trxB2, and whiB3) which are important for protection against oxidative 442 stress. 443 While both AC2P20 and AC2P36 share these similar gene induction characteristics, there 444 are differences in specific thiol-related genes. For example, methionine synthesis (i.e. metK, metA, 445 metC) appears modulated by AC2P36 treatment, but induction of these genes is absent in AC2P20 446 transcriptional data. Likewise, AC2P20 strongly induces sulfate reduction via APS (cysH, nirA), 447 however, these genes are not modulated by AC2P36. These differences may reflect differences in 448 how these compounds sequester free thiols and which free thiols in particular are being modified. 449 While mycothiol is the most abundant free thiol in Mtb (present in millimolar amounts) [53], it is 450 plausible AC2P20 targets other low molecular weight thiols such as ergothioneine (ERG) [32] 451 or gamma-glutamylcysteine (GGC) [54]. Our mass spectrometry data also supports AC2P20 may 452 be generally targeting free thiols, forming adducts with both GSH and NAC, which would indicate that 1) AC2P20 can target a thiol group in general, and 2) it can directly target a cysteine derivative. A) The chemical structure of AC2P20 ((N-1,3-benzothiazol-2-yl-2-[(4,6-dioxo-5-phenyl-1,4,5,6-509 tetrahydropyrimidin-2-yl)thio]acetamide) 510 B) Mtb growth is inhibited in a dose-dependent manner when treated with AC2P20 at pH 5.7 and 511 exhibits an EC50 of 4.3 μM following six days of treatment. Treatment with AC2P20 at pH 7.0 512

Author Contributions
Mtb requires concentrations >60 μM to see growth inhibitory effects. 513 C) Mtb treated with 20 μM of AC2P20 and grown in buffered 7H9 media (pH 5.7) for 5 days 514 shows time-dependent killing as indicated by ~100-fold reduction in viability compared to the 515 DMSO control. Time-dependent killing is not observed in neutral conditions. 516 D) Mtb was treated with a dose-response of AC2P20 at pH 5.7 for 7 days, then assessed for 517 dose-dependent killing by plating for colony-forming units (CFUs). The dotted line indicates the 518 CFUs plated on Day 0. 519 520 Figure 2. AC2P20 treatment promotes a thiol-oxidative and redox stress response. 521 A) Mtb differential gene expression data after being treated for 24 hours with 20 µM AC2P20 at 522 pH 5.7. Genes indicated include those involved in sulfur metabolism, transcriptional regulation, 523 and redox homeostasis. Statistically significant genes (q < 0.05) are highlighted in red. 524 B) A pie chart depicting the functional classification breakdown of significantly induced genes 525 (>2-fold, q < 0.05) following the analysis of AC2P20-treated Mtb RNA-seq profile. 526 C) Heatmap comparing 16 upregulated genes (between AC2P20 and AC2P36 at pH 5.7 that are 527 involved in sulfur metabolism, transcriptional regulation, and redox homeostasis . Genes were 528 annotated with the H37Rv genome. 529 between AC2P20-treated and AC2P36-treated Mtb 29 . 531 532 Figure 3. AC2P20 forms adducts with free thiols at acidic pH. 533 A) AC2P20 was incubated in Tris-HCl buffer, pH 5.7 with reduced glutathione (GSH) for one 534 hour. An AC2P20-GSH adduct (~528 Da) was confirmed via mass spectrometry. Samples were 535 run in duplicate and observed in negative ESI mode. 536 B) In the absence of GSH, AC2P20 incubated with DMSO does not fragment at pH 5.7. Only the 537 parent molecule is observable at a molecular weight of ~409 Da. Samples were run in duplicate 538 and observed in negative ESI mode. 539 C) AC2P20-GSH adduct formation at pH 5.7 (~530 Da) was also observed in positive ESI mode, 540 as well as adduct loss of the glutamate fragment (~401 Da) and subsequent fragmentation of the 541 AC2P20 molecule and its pyrimidine fragments. Samples were run in duplicate. 542 543 Figure 4. AC2P20 depletes free thiols and induces intracellular ROS accumulation. 544 A) Treatment of Mtb with AC2P20 leads to a pH-dependent decrease in free thiols. Free thiol 545 depletion is observed at pH 5.7 with AC2P20 treatment. AC2P36 is a pH-dependent chemical 546 probe known to deplete free thiol pools and serves as a positive control. Statistical significance 547 was calculated using a two-way ANOVA (*p<0.05). 548 B) ROS accumulate under AC2P20 treatment at acidic conditions. Mtb treatment with 549 AC2P20 leads to a pH-dependent increase in intracellular reactive oxygen species (ROS). ROS 550 was detected using a final concentration of 5 µm fluorescent dye, CellROX Green, and normalized to an OD595. DMSO was used as a control. Statistical significance was calculated using a one-way 552 ANOVA (*p<0.05) 553 554 Figure 5. Proposed mechanism for AC2P20 adduct formation. 555 A) Proposed reaction mechanism for the formation of a disulfide bond between AC2P20 and GSH 556 at pH 5.7. 557 B) Proposed stable covalent bond formation between AC2P20 and free thiols in Mtb during redox 558 cycling. 559 560 Supplemental Figures: 561 S1A. Dose-response curve for AC2P20 inhibition of M. smegmatis GFP fluorescence. 562 S1B. AC2P20 does not modulate Mtb cytoplasmic pH at pH 5.7. DMSO and Nigericin served as 563 negative and positive controls, respectively. 564 S1C. The chemical structure of AC2P36 (5-chloro-N-(3-chloro-4-methoxyphenyl)-2-565 methylsulfonylpyrimidine-4-carboxamide) [29]. 566 S2A. A pie chart depicting the functional classification breakdown of significantly induced genes 567 (>2-fold, q < 0.05) following the analysis of AC2P36-treated Mtb RNA-seq profile. 568 S2B. Heatmap comparing the contrast between 8 differentially-regulated genes (between AC2P20 569 and AC2P36 at pH 5.7) that are involved in lipid metabolism and central metabolism. Genes were 570 annotated with the H37Rv genome. 571 S3A. Mass spectrometry data showing adduct formation between AC2P20 and GSH at pH 7.0. 572 Spectra were analyzed in negative ESI mode.

S3B.
Mass spectrometry data showing adduct formation between AC2P20 and GSH at pH 8.5. 574 Spectra were analyzed in negative ESI mode. 575 S3C. AC2P20 incubated with DMSO does not fragment in the absence of GSH at pH 7.0. Spectra 576 were analyzed in negative ESI mode. 577 S3D. AC2P20 incubated with DMSO does not fragment in the absence of GSH at pH 8.5. Spectra 578 were analyzed in negative ESI mode. 579 S4A. Mass spectrometry data showing adduct formation between AC2P20 and N-acetylcysteine 580 at pH 5.7. Spectra were analyzed in negative ESI mode. 581 S4B. Mass spectrometry data showing adduct formation between AC2P20 and N-acetylcysteine 582 at pH 7.0. Spectra were analyzed in negative ESI mode. 20 +NAC 7.0 583 S4C. Mass spectrometry data showing adduct formation between AC2P20 and N-acetylcysteine 584 at pH 8.5. Spectra were analyzed in negative ESI mode. 585 S4D. AC2P20 is still able to form an adduct with GSH in the presence of the oxidant, H2O2. Spectra 586 were analyzed in negative ESI mode. 587