Synthesis and biological evaluation of orally active prodrugs and analogs of para-Aminosalicylic Acid (PAS)

Abstract

Tuberculosis (TB) is one of the world’s most deadly infectious diseases resulting in nearly 1.3 million deaths annually and infecting nearly one-quarter of the population, para-Aminosalicylic acid (PAS), an important second-line agent for treating drug-resistant Mycobacterium tuberculosis, has moderate bioavailability and rapid clearance that necessitate high daily doses of up to 12 grams per day, which in turn causes severe gastrointestinal disturbances presumably by disruption of gut microbiota and host epithelial cells. We first synthesized a series of alkyl, acyloxy and alkyloxycarbonyloxyalkyl ester prodrugs to increase the oral bioavailability and thereby prevent intestinal accumulation as well as undesirable bioactivation by the gut microbiome to non-natural folate species that exhibit cytotoxicity. The pivoxyl prodrug of PAS was superior to all of the prodrugs examined and showed nearly quantitative absorption. While the conceptually simple prodrug approach improved the oral bioavailability of PAS, it did not address the intrinsic rapid clearance of PAS mediated by N-acetyltransferase-1 (NAT-1). Thus, we next modified the PAS scaffold to reduce NAT-1 catalyzed inactivation by introduction of groups to sterically block N-acetylation and fluorination of the aryl ring of PAS to attenuate N-acetylation by electronically deactivating the para-amino group. Among the mono-fluorinated analogs prepared, 5-fluoro-PAS, exhibited the best activity and an 11-fold decreased rate of inactivation by NAT-1 that translated to a 5-fold improved exposure as measured by area-under-the-curve (AUC) following oral dosing to CD-1 mice. The pivoxyl prodrug and fluorination at the 5-position of PAS address the primary limitations of PAS and have the potential to revitalize this second-line TB drug.

Graphical abstract

1. Introduction

Mycobacterium tuberculosis (Mtb), the etiological agent of tuberculosis (TB), is a major and enduring global health threat that caused 5.8 million new infections and 1.3 million deaths in 2020.[1–3] Treatment of drug susceptible strains require two months of intensive phase of therapy with the first-line agents isoniazid, rifampicin, ethambutol and pyrazinamide, followed by a four month continuation phase with isoniazid and rifampicin.[4] Multidrug-resistant (MDR) and extensively-drug resistant (XDR) TB strains require a modified regimen with an arsenal of drugs, which are generally less efficacious and considerably more toxic.[5–7] The development of more potent, less toxic, faster sterilizing drug regimens for treatment of both drug-sensitive and drug-resistant Mtb strains is needed to control the global TB pandemic.[8]

para-Aminosalicylic acid (PAS) was reported by Jörgen Lehmann and co-workers in 1943 for the treatment of TB based on the observation that mycobacterial respiration is stimulated by salicylic acid suggesting related analogs may interfere with respiration.[9, 10] Dr. Lehmann decided to incorporate a para-amino substituent on salicylic acid inspired by the sulfonamide antibacterials, discovered a decade earlier by Domagk and co-workers, which have a strict requirement for a para-amino group on the phenylsulfonamide pharmacophore.[11–14] The near concurrent discoveries of PAS and streptomycin for the treatment of TB were considered monumental achievements since TB had been the leading cause of death through the Industrial Revolution in the United States and Europe.[15–18] When drug-monotherapy for TB was used in early clinical trials, relapse of the disease was observed due to the development of resistance,[19–21] thus combination therapy was introduced in the early 1950s employing PAS, streptomycin and isoniazid as the first curative treatment regimen for TB.[22–25] However, approval of newer more potent and better tolerated agents including rifampicin, ethambutol and pyrazinamide—relegated PAS to a second-line TB drug. In 1992, the increased cases of multi drug-resistant TB led to the reintroduction of PAS for the treatment of drug-resistant TB.[26]

Following the elucidation of the folate pathway and mechanism of action of the structurally related sulfonamide antibiotics in the 1960s (Figure 1),[27–29] PAS was surmised to act as a competitive inhibitor of dihydropteroate synthase (DHPS) because its activity is antagonized by para-aminobenzoic acid (PABA).[30–32] In 2013, Rhee and co-workers–using a metabolomics platform, along with an international consortium led by Camacho at the Novartis Institute of Tropical Diseases (NITD), employing a complimentary genetic approach–shattered this long-held dogma. These research groups showed PAS, rather than inhibiting DHPS–is a competent substrate that is transformed into hydroxydihydropteroate–which is subsequently converted to hydroxydihydrofolate by dihydrofolate synthetase (DHFS).[33, 34] Mycobacteria containing mutant DHFS unable to convert hydroxydihydropteroate to hydroxydihydrofolate were resistant to PAS while mycobacterial strains overexpressing DHFR were also resistant to PAS. Taken together, these results suggested hydroxydihydrofolate inhibits dihydrofolate reductase (DHFR).[35] This hypothesis was subsequently biochemically validated using recombinant DHFR from Mtb by Kordus, Dawadi and co-workers employing an authentic synthetic standard of hydroxydihydrofolate.[36] In another study, Wright and co-workers showed the PAS metabolite hydroxydihydrofolate also inhibits the flavin-dependent thymidylate synthase encoded by thyX in the folate pathway.[37] Mtb is uniquely susceptible to PAS, whereas enteric bacteria such as E. coli can utilize PAS as a PABA surrogate in folate biosynthesis and are not inhibited by the hydroxyfolate metabolites.[38]

Figure 1.

Pathway for synthesis of precursors for nucleic acids, proteins, cofactors and other essential metabolites and mechanism of action of PAS. Figure modified from Camacho et al.^[33]

Given the recent revelation into PAS’ complex multitarget mechanism of action, lack of significant clinical resistance as a result of its limited use, activity against most drug-resistant Mtb strains, and ability to potentiate other TB drugs,[22, 39–41] we sought to reinvestigate PAS and address some of the key liabilities of this valuable TB drug. PAS is noted for its rapid clearance and moderate bioavailability that necessitate high daily doses of up to 12 grams per day,[42] which in turn causes severe gastrointestinal disturbances likely by disruption of gut microbiota and host epithelial cells. The bioavailability of PAS is approximately 60% and the large amounts of unabsorbed drug remains in the gastrointestinal tract and likely contribute to the gastrointestinal symptoms.[43] The absorbed PAS is rapidly inactivated through N-acetylation of the critical p-amino group by the human N-acetyl transferase-1 (NAT-1) and the metabolite is then renally cleared.[36, 44–46]

Herein, we describe two complimentary approaches to address PAS’ shortcomings. We first synthesized a series of prodrugs designed to increase the oral bioavailability of PAS, thereby preventing intestinal accumulation and undesirable bioactivation by the gut microbiome to non-natural potentially cytotoxic folate species (Figure 2). We explored both simple esters such as methyl, ethyl and isopropyl (2–5) as well as the labile pivoxyl,[47] proxetil,[48, 49] and 5-methyl-1,3-dioxol-2-one[50–52] (6–8) that have been used to enhance pharmacokinetic properties of other antibiotics and drugs.[53, 54] We also introduced an α-methyl substitution on the proxetil and pivoxyl prodrugs (9–10) to further sterically modulate the rate of cleavage of these more labile pro-moieties. The conceptually simple prodrug approach does not address the intrinsic rapid clearance of PAS by human N-acetyltransferase-1 (NAT-1), and thus we modified PAS in attempt to decrease its metabolic clearance. Given the reported rigid structure-activity relationships (SAR) of PAS[31, 55] showing the phenol is essential, the carboxylic acid must be strictly maintained, and substitution of the p-amino group is poorly tolerated, except for a methyl, we synthesized a small series of analogs containing relatively conservative modifications including an isomeric series of mono-fluoro-PAS analogs 11–13 through fluorination at the 3, 5 and 6 positions in an attempt to electronically deactivate the para-amino group. We also modified the scaffold to sterically prevent N-acetylation through synthesis of the N-methyl derivative 14 as well as azide analog 15 that we hypothesized could be bioreductively activated within mycobacteria by mycothiol based on the reported thiol-promoted reductive activation of other aryl-azide drugs.[56]

Figure 2.

Proposed PAS prodrugs and analogs.

2. Results and Discussion

2.1. Chemistry

Methylation of PAS using dimethyl sulfate and sodium carbonate in acetone led to methyl ester 2 as well as the N-methyl methyl ester 3 (Scheme 1A). The other prodrugs (4–9) were synthesized from PAS by alkylation with the respective alkyl iodides using cesium carbonate (Scheme 1A). Cesium carbonate and methyl iodide were not used for the synthesis of the methyl ester since multiple side reactions were observed over a very short time and the major product isolated was N,N-dimethyl-PAS methyl ester. The requisite alkyl iodides for the prodrugs were synthesized from the corresponding alkyl chlorides through a Finkelstein reaction using sodium iodide in acetone (Scheme 1B). The 1-chloroethyl pivalate 20 was conveniently synthesized by condensation of acetaldehyde and pivaloyl chloride (Scheme 1B). The reagent stoichiometry and reaction times were optimized for each esterification reaction in the preparation of 4–9 in order to avoid the need to protect the para-amino group. However, the synthesis of prodrug 10 proved more challenging and required amine protection due to competitive N-alkylation with 4-(iodomethyl)-5-methyl-1,3-dioxol-2-one 21. This was accomplished by protection of PAS as the tert-butyloxycarbonyl (Boc) carbamate 22 followed by alkylation with 21 to afford 23, and deprotection with TFA to afford the final prodrug 10 (Scheme 1C).

Scheme 1.

Synthesis of PAS prodrugs.

PAS analogs designed to improve the pharmacokinetic profile of PAS through reduced clearance by attenuating NAT-1 catalyzed N-acetylation were also synthesized (Scheme 2). The first analog N-methyl derivative 14 was designed to sterically block N-acetylation while azide 15 was conceived to bypass NAT-1 altogether and be bioreductively activated within Mtb. Saponification of 3 afforded N-methyl 14 (Scheme 2A) while azide 15 was prepared by diazotization of PAS in the presence of sodium azide (Scheme 2B). We additionally prepared a series of regioisomeric mono-fluoro analogs substituted at the 3, 5 or 6 positions of the aryl ring (11–13) to electronically deactivate the para-amino group in order to reduce the rate of N-acetylation by human NAT-1. 5-Fluoro-PAS 11 was synthesized according to the reported method by sequential fluorination of methyl 4-acetamido-2-hydroxybenzoate (24) using Selectfluor to afford 25 followed by hydrolysis to furnish 11 (Scheme 2C).[57] The ethyl ester of 11 was prepared by alkylation with ethyl iodide based on its pharmacokinetic potential and observation that ethyl esters are selectively metabolized within mycobacteria (vide infra). The regioisomeric 3- and 6-fluoro PAS analogs were synthesized from the respective fluorinated 4-bromosalicylic acid derivatives by palladium-catalyzed carbamoylation (Scheme 2D). This was accomplished by para-methoxybenzyl (PMB) protection of the carboxylic acid and phenol functional groups of 26a-b to yield 27a-b. Subsequent Buchwald-Hartwig carbamoylation of 27a-b with tert-butylcarbamate employing palladium(0) bis(dibenzylideneacetone) as the catalyst and Xanthphos as the ligand afforded 28a-b in excellent yields.[58–62] Deprotection of 28a-b using HCl in 1,4-dioxane provided the final target compounds 12–13 as hydrochloride salts.

Scheme 2.

Synthesis of PAS analogs.

2.2. Plasma Stability

A common strategy to improve the oral bioavailability of drugs is to synthesize ester prodrugs that mask the negatively charged carboxylate and facilitate gastrointestinal absorption. The ester prodrug is then cleaved by serum or tissue esterases to release the parent drug. To study the release properties of the prodrugs 2–10, we performed plasma stability assays using human and mouse plasma at 37 °C. The prodrugs (1 μg/mL) were incubated with human or mouse plasma containing 5% DMSO and amount of both hydrolyzed product (i.e. the parent drug PAS) and the prodrug remaining as a function of time were monitored by liquid chromatography tandem mass spectrometry (LC-MS/MS). The simple ester prodrugs 2–5 were extremely stable in both human and mouse plasma with 84–96% of the prodrug remaining at 2 hours (Table 1). While the simple alkyl ester prodrugs were observed to be highly stable, the acyloxy ester and alkyloxycarbonyloxyalkyl ester prodrugs 6–10 were rapidly cleaved in mouse and human plasma with half-lives under 30 minutes (Table 1). Taken together, these data indicate esters 6–10 have desirable release characteristics, which may lead to improved drug absorption while plasma esterases are unable to release the parent drug PAS from the simple alkyl esters 2–5 suggesting they cannot be used to deliver PAS in vivo unless released by a serum esterase-independent mechanism.

Table 1:

Stability of PAS prodrugs in human and mouse plasma

The % of prodrug remaining after incubation for 120 min is reported;

ND: Not determined

2.3. Pharmacokinetic Experiments

We next evaluated the prodrugs in mice following oral dosing to ascertain if any of the prodrugs led to increased levels of PAS in vivo through enhanced oral absorption and prodrug hydrolysis. The prodrugs were administered at a dose of 10 mg/kg (2–8) or 25 mg/kg (9–10) in a solution formulation composed of 5% dimethylacetamide:60% polyethylene glycol 300 (PEG 300): 35% (5% dextrose in water) to 2 female CD-1 mice by oral gavage (p.o.) and plasma samples were obtained from the lateral vein of each mouse at 0.5, 1, 3, and 5 hours. The prodrug and released PAS were both quantified by high-performance liquid chromatography tandem mass spectrometry (LC-MS/MS) and the resulting snapshot concentration time-profiles were fit by non-compartmental analysis to determine the total integrated area under the plasma-concentration time curve (AUC), which was used as an overall proxy for oral exposure in vivo. To rank order the prodrugs, we compared the oral exposure as measured by the oral AUC values compared to the AUC value of PAS at an equivalent dose. Although the simple alkyl ester prodrugs 2–5 were not cleaved by serum carboxyesterases, we decided to evaluate these in vivo nonetheless as there are many other potential enzymes[63, 64] in vivo capable of hydrolyzing simple esters besides the abundant carboxyesterase-1 (CES-1) found in blood. All of the prodrugs 2–5 could be observed in plasma with AUCs ranging from 0.04–1.7 h·μg/mL (Table 2) indicating they were absorbed, but not rapidly cleaved, consistent with the plasma stability studies. However, only the methyl ester 2 released PAS in vivo. The AUC of the released PAS from methyl ester 2 was 0.05 h·μg/mL, which is 7-fold lower than PAS at an equivalent dose indicating the simple methyl ester did not enhance oral delivery of PAS. On the other hand, the more labile acyloxy and alkyloxycarbonyloxyalkyl esters 6–10 were not observed in plasma suggesting they are rapidly cleaved to liberate PAS, also consistent with the plasma stability studies. Indeed 6–10 released PAS in vivo with AUCs for PAS ranging from 0.019 for proxetil-PAS 7 to 0.46 h·μg/mL for pivoxyl-PAS 6. Since the prodrugs were administered at an equivalent weight as PAS (i.e. 10 mg/kg), when corrected for the mass of PAS contained in the pivoxyl-PAS 6 prodrug (the prodrug delivers only 5.7 mg PAS per 10 mg of the prodrug), we estimate an approximate 200% increase in PAS levels in vivo suggesting the pivoxyl-PAS 6 prodrug was nearly quantitively absorbed since the oral bioavailability of PAS, in humans, is reported as 50–60%.^[43]

Table 2:

Exposure of PAS prodrugs after oral administration to CD-1 female mice.

The prodrugs were administered at a dose of 10 mg/kg* (PAS, 2-8) or 25 mg/kg^# (9-10) to 2 female CD-1 mice by oral gavage (p.o.) and plasma samples were obtained from the lateral vein of each mouse at 0.5, 1, 3, and 5 hours. AUC reported for 0 to 5 h.

NA: Not applicable; n=2

The superior oral bioavailability of pivoxyl-PAS 6 prodrug as compared to PAS when formulated as a solution composed of 5% dimethylacetamide:60% polyethylene glycol 300 (PEG 300): 35% (5% dextrose in water) motivated us to further investigate formulations eliminating the use of a strong solvent vehicle such as PEG. Requirement of solvent vehicles such as PEG create hurdles to development and therefore, water-based suspensions are preferred. Suspensions of PAS are bioavailable, thus we evaluated the full PK of pivoxyl-PAS 6 prodrug as a suspension in 5% DMSO: 95% carboxymethylcellulose (CMC)/Tween80. Unfortunately, the oral bioavailability of PAS released from the suspension of prodrug 6 was significantly lower than the parent PAS. We report that in a more aqueous formulation, the pivoxyl-PAS 6 prodrug does not provide a substantial advantage over the parent compound (Table 3), whereas, in the presence of a solvent vehicle such as PEG, the pivoxyl-PAS 6 prodrug shows some promise.

Table 3:

Exposure of PAS prodrugs after oral administration of CMC/Tween suspension to CD-1 female mice.

The compounds were administered at a dose 25 mg/kg to 2 female CD-1 mice by oral gavage (p.o.) and plasma samples were obtained from the lateral vein of each mouse at 0.5, 1, 3, 5, 7 and 24 hours. AUC reported for 0 to 24 h; n=3

While PAS prodrugs can improve the bioavailability, they cannot alter the intrinsic rapid clearance of PAS due to NAT-1 catalyzed acetylation that results in low AUC values. Therefore, we prepared a series of PAS analogs containing conservative modifications designed to maintain antitubercular activity, but ameliorate the undesirable rapid clearance. The compounds were evaluated analogously to the prodrugs in snapshot pharmacokinetic (PK) experiments dosing 25 mg/kg in CD-1 female mice and the compounds were quantified by LC-MS/MS to determine their AUC values. N-Methyl PAS 14 was reported to be nearly equipotent to PAS,[31, 55] but its pharmacokinetic profile was never reported. We observed the AUC was an astonishing 130 h·μg/mL, which is approximately 59 times greater than PAS (Table 4). Azide analog 15 designed to be bioreductively activated by M. tuberculosis performed even better with an AUC of 900 h·μg/mL—or more than 429-fold better than PAS—at an equivalent dose. We observed no formation of PAS in vivo from azide 15 demonstrating the azide functional group is not metabolized by mice. Amongst the mono-fluoro-substituted analogs, 5-fluoro-PAS 11 showed the best exposure with an AUC of 10 h·μg/mL that is 5-times greater than PAS. The 3- and 6-fluoro PAS analogs 12 and 13 also had improved AUCs relative to PAS of 4.0 and 2.4 h·μg/mL, respectively (Table 4).

Table 4.

Area under the curve for PAS analogs

^aplasma concentration of prodrug following a 25 mg/kg p.o; AUC reported for 5 h, n=2

2.4. NAT-1 catalyzed inactivation of PAS and N-methyl-PAS (14) and 5-fluoro-PAS (11)

To determine if the tremendously enhanced pharmacokinetic profiles of N-methyl-PAS 14 and 5-fluoro-PAS 11 were due to decreased metabolism by human N-acetyltransferase-1 (NAT-1) we characterized their kinetic parameters with recombinant NAT-1 (Table 5). The kinetic parameters for PAS were nearly identical to the reported values.[65] By contrast, the specificity constant (k_cat/K_M), which is a measure of substrate preference, for 11 and 14 were attenuated by 11- and 30- fold, respectively, primarily due to an increase in the K_M values. The biochemicals results are largely congruent with the observed PK profiles. Thus, N-methyl PAS 14 and 5-fluoro-PAS 11 exhibit a 30-fold and 11-fold decrease in specificity constants for NAT-1 indicating they are significantly poorer substrates for NAT-1 enzyme inactivation and correspondingly possesses a 59-fold and 5-fold enhanced exposure as measured by the AUC value following intravenous administration. Taken together, these biochemical results validate our hypothesis that steric and electronic modulation of PAS can ameliorate the poor pharmacokinetic profile of PAS through attenuation of N-acetyl ation by NAT-1.

Table 5:

Substrate specificity of NAT-1^a

Compound	PAS	11	14
KM (μM)	12±3	(1.1±0.3)×102	42±16
kcat (s−1)	(9.5±0.5)×10−1	(7.9±0.4)×10−1	(1.1±0.3)×10−1
kcat/KM (M−1·s−1)	79.2±0.1	6.9±0.9	2.6±0.2
(kcat/KM)PAS/(kcat/KM)cmpd	1	11	30
AUCcmpd/AUCPAS	1	5	59

^aemployed a F125S mutant because of substantially higher expression levels of the recombinant enzyme. The mutant has comparable K_M values for PAS compared to the wild-type enzyme, but the k_cat is reduced. n = 2.

2.5. Microbiological Activity

We determined the minimum inhibitory concentration (MIC) of the PAS analogues against Mycobacterium tuberculosis H37Rv in 7H9 liquid medium. The MICs were determined by measuring the optical density (OD₆₀₀) of each culture and MIC was reported as the minimum concentration required to inhibit at least 90% of growth compared to negative controls (Table 6). We observed the mono-fluoro-substituted PAS analogs 11–13 lost considerable activity with MIC values ranging from 8–64 μM. The 5-fluoro-PAS 11 showed the best activity (8 μM) in this series, whereas the 3-fluoro-PAS 12 and 6-fluoro-PAS 13 had diminished activity of 32 and 64 μM, respectively. To our dismay, the N-methyl analog 14, which had the best PK profile, was inactive. Our findings contradict the previously reported literature that showed the N-methyl analog was equipotent to PAS.[31, 55] The azide analog 15 was also weakly active with a nearly 200-fold loss of activity compared to PAS. We hypothesized putative mycobacterial reductases potentially induced under anaerobic conditions may be able to reduce azide analog 15 to PAS. Unfortunately, we observed that 15 was inactive against Mtb under both aerobic and anaerobic conditions, indicating no conversion to PAS.

Table 6:

Antimycobacterial activity and relative exposure of PAS prodrugs and analogs with respect to PAS

We then evaluated the anti-mycobacterial potential of the simple prodrugs using the same assay conditions described for the PAS analogs. The simple alkyl esters were not cleaved by human or mouse esterases to release free PAS and remained intact in the circulation, hence, we felt it was necessary to explore their antimicrobial activity. We hypothesized the simple alkyl prodrugs could potentially be cleaved by mycobacterial esterases.[66–68] Indeed, we observed some inhibitory activity with the PAS methyl ester 2 and the PAS ethyl ester 4, where the MIC decreased from 64 μM to 2 μM by increasing the size of the alkyl group from a methyl to an ethyl group (Table 6). However, further increasing the size of the ester, from an ethyl (4) to isopropyl (5), abolished activity. Amongst the PAS analogs, 5-fluoro-PAS 11 was the most active; hence we designed the ethyl ester of 5-fluoro-PAS 16 to determine whether it may be activated by mycobacterial esterases as observed for 4. Unfortunately, 16 was inactive highlighting the high substrate specificity of the putative mycobacterial esterases.

As the last part of our investigation, we evaluated the promising PAS analog 5-fluoro-PAS 11 for its antimycobacterial potential against drug susceptible, multidrug resistant (MDR), extensively drug resistant (XDR) and PAS-resistant Mtb strains (Table 7). The PAS-resistant strains contained mutations in the folC gene encoding for DHPS involved in folate biosynthesis, which is essential for bioactivation of PAS. PAS, isoniazid and linezolid were used as positive controls. 5-Fluoro-PAS 11 displayed slightly reduced activity against the drug susceptible and MDR strains with MICs ranging from 6.3–50 μg/mL and was largely inactive against most XDR and PAS resistant strains.

Table 7:

MIC of 5-fluoro PAS (11) against various strains of Mtb

Compound	PAS	11	Isoniazid	Linezolid
	MIC (μM)a
Drug susceptible (2)	0.2–0.6	6.3–25	0.3	0.7–1.9
Multidrug resistant (4)	0.2–3.1	6.3–50	0.4–19	0.8–1.4
Extensively drug resistant (4)	≥50	≥50	6.3–50	0.9–50
PAS resistant: FolC mutants (3)	12.5–50	≥50	0.3–0.4	1.9

^aMIC₉₀ was determined using a microbroth dilution assay in Middlebrook 7H9/ADC/Tween medium with a DMSO negative control and standard isoniazid and linezolid positive controls included in each plate. Number of strains tested are represented in parenthesis.

3. Conclusion

This study was performed to improve the pharmacokinetic properties of PAS to address its adverse effects associated with high doses. We synthesized and evaluated PAS prodrugs and analogs for their antimycobacterial activity, plasma stability and oral exposure. We observed that the simple PAS alkyl esters are not readily cleaved by mouse and human plasma carboxyesterases, whereas the labile pivoxyl, proxetil and 5-methyl-1,3-dioxol-2-one esters are cleaved almost instantly to release PAS. The simple alkyl esters did not release PAS and thus were further evaluated for their antitubercular activity. Most of the simple prodrugs had no significant antimycobacterial activity. However, the ethyl ester 4, possessed excellent activity, which may be attributed to its selective cleavage by mycobacterial esterases. Unfortunately, preliminary pharmacokinetic studies on these prodrugs revealed that the simple alkyl ester prodrugs (2–5) did not show improved oral exposure as measured by the AUC following oral administration. Among the labile prodrugs (6–10) only pivoxyl-PAS prodrug 6 showed an improved oral exposure relative to PAS, in the presence of a solvent vehicle such as PEG. However, the formulation of the prodrug plays a crucial role in this improvement of oral exposure, and water-based formulations provide no significant advantage in the oral bioavailability. Thus, our observations necessitate auxiliary evaluation of additional formulations of PAS 6. Overall, the pivoxyl-PAS 6, warrants further investigation to determine if the improved oral absorption of this prodrug translates to reduced gastrointestinal toxicity.

We hypothesized the structural modifications of the PAS scaffold to sterically hinder or electronically deactivate the para-amino group would reduce the rate of metabolism by NAT-1 leading to decreased clearance and enhanced AUC values. In our preliminary pharmacokinetic experiments, we observed all the analogs showed a higher oral exposure as compared to the parent PAS. The N-methyl analog 14, which was reported to be equipotent to PAS,[31, 55] showed a 59-fold improvement in exposure, as observed by the enhancement in AUC after oral administration, coincident with a 30-fold reduced clearance by NAT-1, validating our design strategy. However, we showed 14 is inactive and the reported activity may have been due to residual PAS in the sample. The azide derivative 15 also showed over 429-fold enhancement in exposure and we hypothesized it may be bioreductively activated[56] within mycobacteria to form PAS by mycothiol under aerobic or anaerobic conditions; unfortunately, 15 was inactive under both aerobic and anaerobic conditions. We also prepared three regioisomeric mono-fluoro-PAS analogs by introducing fluorine at the 3-, 5- and 6-positions to electronically deactivate the para-amino group. While the 3- and 6-fluoro-PAS analogs were microbiologically inactive, we discovered 5- fluoro-PAS 11 retained useful antimycobacterial activity with a MIC of 8 μM against M. tuberculosis H37Rv and displayed superior pharmacokinetic properties with a 5-fold enhanced oral exposure over PAS. We showed the enhanced oral exposure directly correlated to the reduced clearance by NAT-1 in vitro. While the modifications such as mono-fluoro substitution of the ring, methyl substitution on the amine or conversion of the amine to an azide, reduced the intrinsic clearance of the molecules in comparison to PAS; they resulted in the loss of antibacterial activity. Thus, we conclude these PAS analogs have reduced potential as antimycobacterial agents.

4. Experimental Section

4.1. Chemistry

All glassware was dried in a 150 °C oven overnight. All chemicals, solvents, and glassware were purchased from either Fisher Scientific (Pittsburg, Pennsylvania), Ambeed Inc (Arlington Heights, Illinois) or Sigma Aldrich (St. Louis, Missouri). All reactions were performed under an inert atmosphere of argon. The chemical reactions were tracked using fluorescent silica gel-coated TLC plates and the separated components were visualized using UV light (254 nm) or staining with Ninhydrin. All final products were characterized by ¹H NMR, ¹³C NMR and MS analyses. All new compounds were also characterized and confirmed by HRMS. Mass spectra were acquired either on an Agilent 1200/AB Sciex® API 5500 QTrap LC/MS/MS, using electrospray ionization (ESI) or a single quadrupole analyzer or on an Agilent 7200/ Accurate-Mass Q-TOF GC/MS, using electron impact (EI) or chemical ionization (CI). All ¹H and ¹³C NMR spectra were obtained on Varian 7600-AS spectrometer or on an Ascend™ 600 MHz Bruker spectrometer. All NMR spectra were recorded at 600 and 400 MHz for ¹H, 151 and 101 MHz for ¹³C, and 376 MHz for ¹⁹F. ¹H NMR spectra were referenced to residual CDCl₃ (7.27 ppm), DMSO-d₆ (2.50 ppm), or CD₃OD (3.31 ppm); ¹³C NMR spectra were referenced to CDCl₃ (77.23 ppm) DMSO-d₆ (39.51 ppm), or CD₃OD (49.15 ppm). NMR chemical shift data are reported as follows: chemical shift, multiplicity (br = broad, s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, ABq = AB quartet, dm = doublet of multiplets), coupling constant, integration. Coupling constants are given in Hertz (Hz). Melting points for solid final compounds were determined using a Thomas Hoover capillary melting point apparatus. All final compounds were tested for purity using an analytical high-performance liquid chromatography (HPLC), at 270 nm, using either: A) reverse phase C18 column (150 × 4.6 mm, Phenomenex, Torrance, United States) with the following gradient (mobile phase contains 0.1% formic acid) from 5% acetonitrile and 95% water to 95% acetonitrile and 5% water over 20 min or B) normal phase silica column (250 × 19 mm, Waters, Ireland) with the following gradient from 5% isopropanol and 95% heptane to 95% isopropanol and 5% heptane over 13 min. Compounds 17[53], 18[53], 19[69], 20[70], 21[71] and 25[57] were prepared using the reported procedures and characterization data matched the reported values.

4.1.1. Methyl 4-amino-2-hydroxybenzoate (2) and Methyl 2-hydroxy-4-(methylamino)benzoate (3).

Dimethyl sulfate (150 mg, 1.20 mmol, 1.20 equiv) was added to a mixture of 4-aminosalicylic acid (153 mg, 1.00 mmol, 1.00 equiv), anhydrous sodium carbonate (130 mg, 1.20 mmol, 1.20 equiv) and acetone (5 mL/mmol) at 23 °C. After stirring for 24 h at 23 °C, the reaction mixture was filtered, the filtrate was concentrated under vacuum, and the residue purified by column chromatography with a 0–70% gradient of ethyl acetate in hexanes to obtain 2 (62 mg, 37%) and 3 (64 mg, 35%) as white solids.

Data for Methyl 4-amino-2-hydroxybenzoate (2).

Mp = 114–115 °C; R_f = 0.42 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, DMSO-d₆) δ 10.72 (s, 1H), 7.41 (d, J = 8.7 Hz, 1H), 6.09–6.06 (m, 3H), 5.96 (d, J = 2.1 Hz, 1H), 3.75 (s, 3H); ¹³C NMR (101 MHz, DMSO-d₆) δ 170.3, 163.3, 156.5, 131.5, 107.1, 99.9, 98.9, 52.0; HRMS (EI) calcd for C₈H₉NO₃ [M + H]⁺ 168.0655, found 168.0655 (error 0 ppm); Purity (Method A) = 97%, t_R = 11.3 min, k′ = 3.8.

Data for Methyl 2-hydroxy-4-(methylamino)benzoate (3).

Mp = 99–100 °C; R_f = 0.54 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, DMSO-d₆) δ 10.83 (s, 1H), 7.44 (d, J = 8.8 Hz, 1H), 6.65 (q, J = 5.1 Hz, 1H), 6.11 (dd, J = 8.8, 2.3 Hz, 1H), 5.89 (d, J = 2.2 Hz, 1H), 3.76 (s, 3H), 2.66 (d, J = 4.9 Hz, 3H); ¹³C NMR(101 MHz, DMSO-d₆) δ 170.4, 163.5, 156.5, 131.0, 106.0, 99.7, 96.2, 52.0, 29.5; HRMS (EI) calcd for C₉H₁₁NO₃ [M + H]⁺ 182.0812, found 182.0807 (error 2.4 ppm); Purity (Method A) = 99%, t_R = 13.4 min, k′ = 2.4.

4.1.2. General procedure for alkylation.

Iodoalkane (1.50 equiv) was added to a mixture of 4-aminosalicylic acid (1.00 equiv) and anhydrous cesium carbonate (1.00 equiv) in acetone (10 mL/mmol of 4-aminosalicylic acid) at 23 °C. After stirring for 2 h at 23 °C, the reaction mixture was filtered, and the filtrate was concentrated under vacuum. The product was purified using column chromatography on silica gel with a 20–60% gradient of ethyl acetate in hexanes.

4.1.2.1. Ethyl 4-amino-2-hydroxybenzoate (4).

The title compound was prepared from ethyl iodide using the general procedure for alkylation and isolated (146 mg, 81%) as a brown solid. Mp = 105–106 °C; R_f = 0.44 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, DMSO-d₆) δ 10.80 (s, 1H), 7.41 (d, J = 8.7 Hz, 1H), 6.11–6.06 (m, 3H), 5.95 (d, J = 2.1 Hz, 1H), 4.22 (q, J = 7.1 Hz, 2H), 1.25 (t, J = 7.1 Hz, 3H); ¹³C NMR (101 MHz, DMSO-d₆) δ 170.0, 163.4, 156.4, 131.4, 107.0, 100.1, 99.0, 60.6, 14.6; HRMS (EI) calcd for C₉H₁₁NO₃ [M + H]⁺ 182.0812, found 182.0810 (error 1.0 ppm); Purity (Method A) = 98%, t_R = 16.9 min, k′ = 6.3.

4.1.2.2. Isopropyl 4-amino-2-hydroxybenzoate (5).

The title compound was prepared from isopropyl iodide using the general procedure for alkylation and isolated (166 mg, 85%) as an off- white solid. Mp = 75–76 °C; R_f = 0.5 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, DMSO-d₆) δ 10.87 (s, 1H), 7.39 (d, J = 8.7 Hz, 1H), 6.07–6.06 (m, 3H), 5.94 (d, J = 2.1 Hz, 1H), 5.06 (hept, J = 6.3 Hz, 1H), 1.25 (d, J = 6.2 Hz, 6H); ¹³C NMR (101 MHz, DMSO-d₆) δ 169.7, 163.4, 156.4, 131.4, 106.9, 100.3, 98.9, 68.0, 22.2; HRMS (EI) calcd for C₁₀H₁₃NO₃ [M + H]⁺ 196.0968, found 198.0966 (error 1.2 ppm); Purity (Method A) = 98%, t_R = 13.4 min, k′ = 2.3.

4.1.2.3. (Pivaloyloxy)methyl 4-amino-2-hydroxybenzoate (6).

The title compound was prepared from iodomethyl pivalate (17) using the general procedure for alkylation and isolated (195 mg, 73%) as a colorless oil. R_f = 0.41 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, CDCl₃) δ 10.61 (s, 1H), 7.61 (d, J = 9.2 Hz, 1H), 6.15–6.12 (m, 2H), 5.94 (s, 2H), 4.25 (br s, 2H), 1.21 (s, 9H); ¹³C NMR (101 MHZ CDCl₃) δ 177.3, 168.6, 164.1, 154.2, 132.1, 107.0, 101.8, 100.5, 79.3, 38.8, 26.9; HRMS (EI) calcd for C₁₃H₁₈NO₅ [M + H]⁺ 268.1179, found 268.1177 (error 0.9 ppm); Purity (Method A) = 99%, t_R = 19.6 min, k′ = 9.9.

4.1.2.4. [(Isopropoxycarbonyl)oxy]methyl 4-amino-2-hydroxybenzoate (7).

The title compound was prepared from 1-iodomethylisopropyl carbonate (18) using the general procedure for alkylation and isolated (210 mg, 78%) as a brown oil. R_f = 0.27 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, CDCl₃) δ 10.54 (s, 1H), 7.64 (d, J = 9.0 Hz, 1H), 6.14–6.11 (m, 2H), 5.94 (s, 2H), 4.91 (hept, J = 6.2 Hz, 1H), 4.21 (br s, 2H), 1.30 (d, J = 6.2 Hz, 6H); ¹³C NMR (101 MHz, CDCl₃) δ 168.4, 164.2, 154.2, 153.4, 132.1, 107.0, 101.7, 100.5, 81.6, 73.1, 21.6; HRMS (EI) calcd for C₁₂H₁₆NO₆ [M + H]⁺ 270.0972, found 270.0976 (error 1.4 ppm); Purity (Method A) = 96%, t_R = 14.1 min, k′ = 5.6.

4.1.2.5. 1-[(Isopropoxycarbonyl)oxy]ethyl 4-amino-2-hydroxybenzoate (8).

The title compound was prepared from 1-iodoethylisopropyl carbonate (19) using the general procedure for alkylation and isolated (198 mg, 70%) as a brown oil. R_f = 0.34 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, CDCl₃) δ 10.63 (s, 1H), 7.58 (d, J = 9.3 Hz, 1H), 6.97 (q, J = 5.4 Hz, 1H), 6.12–5.98 (m, 2H), 4.88 (hept, J = 6.3 Hz, 1H), 4.22 (br s, 2H), 1.59 (d, J = 5.5 Hz, 3H), 1.27 (d, J = 6.4 Hz, 6H); ¹³C NMR (101 MHz, CDCl₃) δ 168.0, 164.4, 154.1, 152.5, 131.9, 106.9, 101.9, 100.5, 90.9, 72.8, 21.6, 19.7; HRMS (ESI) calcd for C₁₃H₁₆NO₆ [M − H]⁻ 282.0983, found 282.0992 (error 3.2 ppm); Purity (Method A) = 99%, t_R = 14.8 min, k′ = 6.0.

4.1.2.6. 1-(Pivaloyloxy)ethyl 4-amino-2-hydroxybenzoate (9).

The title compound was prepared from 1-chloroethyl pivalate (20) using the general procedure for alkylation and isolated (152 mg, 54%) as a brown oil. R_f = 0.4 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, CDCl₃) δ 10.71 (s, 1H), 7.61 (d, J = 9.2 Hz, 1H), 7.07 (q, J = 5.4 Hz, 1H), 6.16–6.13 (m, 2H), 4.22–4.02 (br s, 2H), 1.58 (d, J = 5.4 Hz, 3H), 1.20 (s, 9H); ¹³C NMR (101 MHz, CDCl₃) δ 176.5, 168.1, 164.1, 153.8, 131.8, 106.8, 102.2, 100.6, 88.6, 38.7, 26.8, 19.6; HRMS (ESI) calcd for C₁₄H₁₈NO₅ [M − H]⁻ 280.1190, found 280.1205 (error 5.3 ppm); Purity (Method A) = 98%, t_R = 15.0 min, k′ = 2.2.

4.1.3. (5-Methyl-2-oxo-1,3-dioxol-4-yl)methyl 4-amino-2-hydroxybenzoate (10).

To a solution of 23 (365 mg, 1.00 mmol, 1.00 equiv) in CH₂Cl₂ (2.5 mL) at 23 °C was added trifluoroacetic acid (2.5 mL). After stirring for 2 h, the acid was neutralized with a solution of saturated aqueous sodium bicarbonate until the pH was ~7. The compound was extracted in CH₂Cl₂ (3 × 15 mL) and the combined extracts were washed with brine, dried (MgSO₄), and concentrated to afford the title compound (231 mg, 87%) as an off-white solid. Mp = 118–120 °C; R_f 0.43 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, CDCl₃) δ 10.61 (s, 1H), 7.61 (d, J = 9.3 Hz, 1H), 6.19 – 6.12 (m, 2H), 5.04 (s, 2H), 4.26–4.04 (br s, 2H), 2.22 (s, 3H); ¹³C NMR (101 MHz, CDCl₃) δ 169.2, 163.89, 154.0, 152.1 140.3, 133.5, 131.81, 107.1, 102.0, 100.6, 53.6, 9.5; HRMS (ESI) calcd for C₁₂H₁₀NO₆ [M − H]⁻ 264.0514, found 264.0518 (error 1.5 ppm); Purity (Method A) = 99%, t_R = 12.0 min, k′ = 1.6.

4.1.4. 4-Amino-5-fluoro-2-hydroxybenzoic acid (11).

Compound 25 (230 mg, 1.00 mmol, 1.00 equiv) in 20% w/v aqueous sodium hydroxide (4 mL) was heated at reflux for 2 h. The reaction was cooled to 23 °C, then concentrated aqueous HCl was added until the pH was ~2. The resulting precipitate was collected by filtration and dried under vacuum overnight to afford the title compound (140 mg, 82%) as a brown solid. Mp = 170–173°C. R_f = 0.16 (95:5 CH₂Cl₂–MeOH); ¹H NMR (400 MHz, DMSO-d₆,) δ 13.11 (br s, 1H), 11.18 (br s, 1H), 7.23 (d, ³J_H-F = 11.9 Hz, 1H), 6.13 (s, 1H), 6.11 (s, 2H); ¹⁹F NMR (376 MHz, DMSO-d₆,) δ −145.1; ¹³C NMR (101 MHz, DMSO-d₆) δ 172.0, 160.1, 144.9 (d, ²J_C-F = 41.7 Hz), 144.8 (d, ¹J_C-F = 206.0 Hz), 114.9 (d, ²J_C-F = 19.8 Hz), 100.9 (d, ³J_C-F = 4.1 Hz), 98.8 (d, ³J_C-F = 6.4 Hz); HRMS (EI) calcd for C₇H₅FNO₃ [M − H]⁻ 170.0259, found 170.0266 (error 4.1 ppm); Purity (Method B) = 99%, t_R = 4.3 min.

4.1.5. General method for deprotection of PMB and Boc groups.

A solution of 4 N HCl in 1,4-dioxane (2 mL) was added to the compound (1.0 equiv) and the reaction stirred for 8 h at 23 °C. The reaction was sparged with nitrogen for 30–45 min to remove the dissolved HCl and then concentrated in vacuo under reduced pressure. The product was washed with 4:1 hexane–CH₂Cl₂ and dried under vacuum to obtain the title compound as the hydrochloride salt.

4.1.5.1. 4-Amino-3-fluoro-2-hydroxybenzoic acid hydrochloride salt (12).

The title compound was synthesized from 28a using the general method for deprotection of PMB and Boc groups and isolated (135 mg, 65%) as a yellow solid. Mp = 157–160 °C; R_f = 0.35 (90:10 CH₂Cl₂–MeOH); 1H NMR (400 MHz, DMSO-d₆) δ 13.46–12.70 (br s, 1H), 11.64–11.18 (br s, 1H), 7.26 (ovlp dd, J = 8.8 Hz, 1.5 Hz, 1H), 7.11 (ovlp br t, ¹J_N-H = 50.9 Hz, observed 1H corresponding to 3H from symmetric ammonium protons, lower signal due to broadening), 6.22 (t, ³J_H-H = ⁴J_H-F = 8.3 Hz, 1H); ¹³C NMR (101 MHz, DMSO-d₆) δ 172.6, 150.7 (d, ²J_C-F = 9.9 Hz), 143.6 (d, ²J_C-F = 9.8 Hz), 138.0 (d, ¹J_C-F = 232.8 Hz), 126.4 (d, ⁴J_C-F = 3.1 Hz), 106.7 (d, ³J_C-F = 3.6 Hz), 101.8 (d, ³J_C-F = 2.5 Hz); ¹⁹F NMR (376 MHz, DMSO-d₆) δ −163.2; HRMS (ESI) calcd for C₇H₅FNO₃ [M − H]⁻ 170.0259, found 170.0267 (error 4.7 ppm); Purity (Method A) = 97%, t_R = 16.1 min, k′ = 12.2.

4.1.5.2. 4-Amino-2-fluoro-6-hydroxybenzoic acid hydrochloride salt (13).

The title compound was synthesized from 28b by the general method for deprotection of PMB and Boc groups and isolated (168 mg, 81%) as a yellow solid. Mp = 133–136 °C; R_f = 0.75 (90:10 CH₂Cl₂–MeOH); ¹H NMR (400 MHz, DMSO-d₆) δ 13.14 (s, 1H), 12.05 (s, 1H), 6.30 (s, 2H), 5.86 (d, ³J_H-F = 14.1 Hz, 1H), 5.82 (s, 1H); ¹³C NMR (151 MHz, DMSO-d₆) δ 171.36, 164.54 (d, ³J_C-F = 6.8 Hz), 162.82 (d, ¹J_C-F = 247.5 Hz), 155.84 (d, ³J_C-F = 11.6 Hz), 95.48, 93.34 (d, ²J_C-F = 26.2 Hz), (missing carbon at C-5 due to low sample concentration); ¹⁹F NMR (376 MHz, DMSO-d₆) δ −106.3; HRMS (CI) calcd for C₇H₅FNO₃ [M − H]⁻ 170.0259, found 170.0263 (error 2.4 ppm); Purity (Method A) = 94%, t_R = 14.4 min, k′ = 12.1.

4.1.6. 2-Hydroxy-4-(methylamino)benzoic acid (14).

A solution of methyl 3 (181 mg, 1.00 mmol, 1.00 equiv) in 10% w/v aqueous sodium hydroxide (2 mL) was heated at reflux for 2 h. The reaction was cooled to 23 °C, then quenched with concentrated aqueous HCl until the pH was ~2. The reaction mixture was extracted with CH₂Cl₂ (3 ×15 mL) and the combined organic extracts were washed with brine (15 mL), dried (MgSO₄) and concentrated under reduced pressure to afford the title compound (137 mg, 82%) as a brown solid. Mp = 132–134 °C; R_f = 0.56 (90:10 CH₂Cl₂–MeOH); ¹H NMR (600 MHz, DMSO-d₆) δ 12.80 (br s, 1H), 11.51 (s, 1H), 7.46 (d, J = 8.8 Hz, 1H), 6.63 – 6.58 (m, 1H), 6.11 (dd, J = 8.8, 2.2 Hz, 1H), 5.91 (d, J = 2.2 Hz, 1H), 2.70 (d, J = 3.2 Hz, 3H); ¹³C NMR (101 MHz, DMSO-d₆) δ 172.6, 164.1, 156.4, 131.4, 105.5, 100.2, 96.2, 29.5; HRMS (ESI) calcd for C₈H₈NO₃ [M − H]⁻ 166.0510, found 166.0510 (error 0 ppm); Purity (Method A) = 99%, t_R = 11.2 min, k′ = 3.9.

4.1.7. 4-Azido-2-hydroxybenzoic acid (15).

Concentrated sulfuric acid (2.5 mL) was added to a suspension of 4-amino-2-hydroxybenzoic acid (500 mg, 3.30 mmol, 1.00 equiv) in H₂O (13 mL) at 23 °C. The suspension was cooled to 0 °C, then a solution of NaNO₂ (280 mg, 4.00 mmol, 1.20 equiv) in H₂O (2 mL) was added dropwise and the reaction mixture was stirred for 1 h at 0 °C, during which time the solution become homogenous. Next, a solution of NaN₃ (360 mg, 5.60 mmol, 1.70 equiv) in H₂O (2.5 mL) was added dropwise at 0 °C during which time strong evolution of nitrogen was observed along with formation of a precipitate. Once addition was complete, the reaction mixture was stirred at 0 °C for another 1 h, then EtOAc (20 mL) was added to the suspension. The organic layer was separated, washed with brine (5 mL), dried (Na₂SO₄), and concentrated under reduced pressure to obtain the title compound (565 mg, 97%) as a brown solid. Mp = Decomposes over 130 °C; R_f = 0.46 (90:10 CH₂Cl₂–MeOH); ¹H NMR (600 MHz, DMSO-d₆) δ 14.41–13.61 (br s, 1H), 11.95–11.25 (br s, 1H), 7.87 (d, J = 8.5 Hz, 1H), 6.74 (dd, J = 8.5, 2.3 Hz, 1H), 6.72 (d, J = 2.2 Hz, 1H); ¹³C NMR (101 MHz, DMSO-d₆) δ 176.5, 167.6, 151.6, 137.3, 115.6, 115.1, 112.0; HRMS (ESI) calcd for C₇H₄N₃O₃ [M − H]⁻ 178.0258, found 178.0260 (error 1.1 ppm); Purity (Method A) = 94%, t_R = 12.4 min, k′ = 5.2.

4.1.8. Ethyl 4-amino-5-fluoro-2-hydroxybenzoate (16).

The title compound was prepared from 4-amino-5-fluoro-2-hydroxybenzoic acid and ethyl iodide using the general procedure for alkylation and isolated (167 mg, 84%) as a brown solid. Mp = 132–134 °C; R_f = 0.52 (2:1 Hexane–EtOAc); ¹H NMR (600 MHz, DMSO-d₆) δ 10.64 (s, 1H), 7.29 (d, ³J_C-F = 11.9 Hz, 1H), 6.26 (s, 2H), 6.18 (d, ⁴J_C-F = 7.7 Hz, 1H), 4.27 (q, J = 7.1 Hz, 2H), 1.30 (t, J = 7.1 Hz, 3H); ¹³C NMR (151 MHz, DMSO-d₆) δ 169.5, 159.6, 145.0 (d, ²J_C-F = 15.0 Hz), 144.1 (d, ¹J_C-F = 232.0 Hz), 114.5 (d, ²J_C-F = 20.4 Hz), 101.0 (d, ³J_C-F = 4.2 Hz), 98.3 (d, ³J_C-F = 6.5 Hz), 61.1, 14.6; ¹⁹F NMR (376 MHz, DMSO-d₆) δ −146.2; HRMS (CI) calcd for C₉H₁₁FNO₃ [M+H]⁺ 200.0717, found 200.0719 (error 1.0 ppm); Purity (Method A) = 99%, t_R = 13.4 min, k′ = 1.9.

4.1.9. 4-[(tert-Butoxycarbonyl)amino]-2-hydroxybenzoic acid (22).

An aqueous solution of sodium hydroxide (525 mg, 13.0 mmol, 2.00 equiv in 5 mL H₂O) and di-tert-butyl dicarbonate (2.85 g, 13.0 mmol, 2.00 equiv in 10 mL 95% ethanol) were added sequentially to a solution of 4-aminosalicylic acid (1.0 g, 6.5 mmol, 1.0 equiv) in 95% ethanol (15 mL) and the reaction was stirred overnight at 23 °C. The reaction was concentrated in vacuo, then 1.0 M aqueous hydrochloric acid (~3-5 mL) was added dropwise until the pH was ~3. The reaction was filtered, and the precipitate was washed with H₂O (30 mL), then dried under reduced pressure to yield the title compound (1.47 g, 89%) as a brown powder. R_f = 0.34 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, DMSO-d₆) δ 13.84–13.00 (br s, 1H), 11.29 (br s, 1H), 9.67 (s, 1H), 7.62 (d, J = 8.7 Hz, 1H), 7.09 (d, J = 2.1 Hz, 1H), 6.95 (dd, J = 8.7, 2.1 Hz, 1H), 1.44 (s, 9H); ¹³C NMR (101 MHz, DMSO-d₆) δ 176.9, 166.7, 157.6, 151.4, 136.1, 114.4, 111.7, 109.7, 85.0, 33.2; HRMS (ESI) calcd for C₁₂H₁₄NO₅ [M − H]⁻ 252.0877, found 252.0892 (error 5.9 ppm).

4.1.10. (5-Methyl-2-oxo-1,3-dioxol-4-yl)methyl 4-[(tert-butoxycarbonyl)amino]-2-hydroxybenzoate (23).

To a solution of of 22 (250 mg, 1.00 mmol, 1.00 equiv) and cesium carbonate (325 mg, 1.00 mmol, 1.00 equiv) in acetone (10 mL) at 23 °C was added 21 (360 mg, 1.50 mmol, 1.50 equiv). After stirring for 1 h at 23 °C, the reaction mixture was filtered, and the filtrate was concentrated under vacuum. Purification by flash chromatography with a 20-60% EtOAc/hexane afforded the title compound (317 mg, 87%) as an off-white solid. R_f = 0.59 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, DMSO-d₆) δ 10.36 (s, 1H), 9.74 (s, 1H), 7.64 (d, J = 8.9 Hz, 1H), 7.15 (s, 1H), 6.98 (d, J = 8.9 Hz, 1H), 5.17 (s, 2H), 2.17 (s, 3H), 1.44 (s, 9H); ¹³C NMR (101 MHz, CDCl₃) δ 169.0, 163.1, 152.0, 151.8, 145.8, 140.5, 133.2, 131.1, 109.5, 106.1, 105.3, 81.5, 54.1, 28.2, 9.5; HRMS (ESI) calcd for C₁₇H₁₈NO₈ [M − H]⁻ 364.1038, found 364.1058 (error 5.5 ppm).

4.1.11. General method for p-methoxybenzyl (PMB) protection.

4-Methoxybenzyl bromide (2.1 equiv) was added to a solution of the salicylic acid derivative (1.0 equiv), anhydrous Cs₂CO₃ (2.5 equiv) and catalytic tetrabutylammonium iodide (TBAI) (0.1 equiv) in anhydrous DMF (2mL/mmol of reactant) at 23 °C. The reaction was monitored by TLC and was complete in 16-18 h. The reaction mixture was partitioned between H₂O and EtOAc, then the EtOAc layer was washed with H₂O (3×) to remove DMF and the organic layer was concentrated under reduced pressure. The crude material was purified by flash chromatography on silica gel using hexanes-EtOAc to afford the title compound.

4.1.11.1. 4-Methoxybenzyl 4-bromo-3-fluoro-2-[(4-methoxybenzyl)oxy]benzoate (27a).

The title compound was synthesized from 4-bromo-3-fluoro-2-hydroxybenzoic acid using the general method for PMB protection and isolated (466 mg, 98%) as a white solid. R_f = 0.6 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, CDCl₃) δ 7.46 (dd, ³J_H-H = 8.5, ⁴J_H-F = 1.8 Hz, 1H), 7.33 (d, J = 8.5 Hz, 2H), 7.28 (d, J = 8.5 Hz, 2H), 6.91–6.87 (m, 3H), 6.85 (d, J = 8.6 Hz, 2H), 5.25 (s, 2H), 5.02 (s, 2H), 3.812 (ovlp s, 3H), 3.807 (ovlp s, 3H); ¹³C NMR (151 MHz, CDCl₃) δ 164.7, 159.79, 159.76, 153.8 (d, ¹J_C-F = 248.5 Hz), 147.3 (d, ²J_C-F = 14.1 Hz), 130.4, 130.3, 128.7, 128.4, 127.4, 126.6 (d, ³J_C-F = 4.3 Hz), 114.5 (d, ²J_C-F = 19.6 Hz), 114.2, 114.0, 113.8, 71.5, 67.2, 55.30, 55.28; ¹⁹F NMR (376 MHz, CDCl₃) δ −120.5; HRMS (ESI) calcd for C₁₅H₁₁BrFO₄ [M − PMB]⁻ 352.9830, found 352.9830 (error 0 ppm).

4.1.11.2. 4-Methoxybenzyl 4-bromo-2-fluoro-6-[(4-methoxybenzyl)oxy] benzoate (27b).

The title compound was synthesized from 4-bromo-2-fluoro-6-hydroxybenzoic acid using the general method for PMB protection and isolated (465 mg, 98%) as a white solid. R_f = 0.56 (2:1 Hexane–EtOAc); ¹H NMR (600 MHz, CDCl₃) δ 7.20–7.16 (m, 4H), 6.86–6.82 (m, 2H), 6.82–6.79 (m, 2H), 6.76–6.72 (m, 2H), 5.18 (s, 2H), 4.92 (s, 2H), 3.75 (s, 3H), 3.72 (s, 3H); ¹³C NMR (151 MHz, CDCl₃) δ 163.1, 160.2 (d, ¹J_C-F = 254.1 Hz), 159.6 (d, ³J_C-F = 5.3 Hz), 157.67, 157.62, 130.1, 129.1, 127.5, 127.4, 124.7 (d, ³J_C-F = 12.1 Hz), 114.0, 113.9, 112.4 (d, ⁴J_C-F = 3.3 Hz), 112.3 (d, ²J_C-F = 25.1 Hz), 111.8 (d, ²J_C-F = 19.3 Hz), 71.1, 67.4, 55.29, 55.26; ¹⁹F NMR (376 MHz, CDCl₃) δ [−111.9; HRMS (ESI) calcd for C₁₅H₁₁BrFO₄ [M − PMB]⁻ 352.9830, found 352.9842 (error 3.4 ppm).

4.1.12. General method for Buchwald-Hartwig coupling:

A flame-dried Schlenk tube with a side-arm was charged with Pd(OAc)₂ (4.60 mg, 0.005 mmol, 0.01 equiv), Xantphos (8.7 mg, 0.015 mmol, 0.015 equiv), benzyl carbamate (180 mg, 1.20 mmol, 1.20 equiv ), Cs₂CO₃ (460 mg, 1.40 mmol, 1.40 equiv), and the PMB protected compound (400 mg, 1.00 mmol, 1.00 equiv). The Schlenk tube was sealed with a septum, evacuated and backfilled with argon (3×). Next, 1,4-dioxane (2 mL) was added through the septum, then the septum was replaced with a Teflon screwcap while flushing with argon. The Schlenk tube was sealed, and the mixture was stirred at 60 °C for 16 h. The reaction mixture was cooled to 23 °C, diluted with CH₂Cl₂ (10 mL), filtered through Celite, and concentrated in vacuo under reduced pressure. The crude material was purified by flash chromatography on silica gel using hexanes–EtOAc to afford the title compound.

4.1.12.1. 4-Methoxybenzyl 4-[(tert-butoxycarbonyl)amino]-3-fluoro-2-[(4-methoxybenzyl)oxy]benzoate (28a).

The title compound was synthesized from 27a using the general method for Buchwald-Hartwig coupling and isolated (480 mg, 94%) as a white solid. R_f = 0.58 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, CDCl₃) δ 7.90 (t, J = 8.0 Hz, 1H), 7.63 (dd, J = 8.0, 1.9 Hz, 1H), 7.34 (d, J = 8.6 Hz, 2H), 7.28 (d, J = 8.6 Hz, 2H), 6.88 (d, J = 8.6 Hz, 2H), 6.84 (ovlp d, J = 8.6 Hz, 2H), 6.82 (ovlp s, 1H), 5.25 (s, 2H), 4.98 (s, 2H), 3.811 (ovlp s, 3H), 3.805 (ovlp s, 3H), 1.53 (s, 9H); ¹³C NMR (151 MHz, CDCl₃) δ 164.9, 159.66, 159.64, 151.8, 146.7 (d, ²J_C-F = 11.6 Hz), 145.7 (d, ¹J_C-F = 242.9 Hz), 132.1 (d, ³J_C-F = 8.7 Hz) 130.26, 130.21, 128.8, 128.0, 127.1 (d, ⁴J_C-F = 4.2 Hz), 119.5, 114.0, 113.7, 113.3, 81.8, 76.6, 66.7, 55.30, 55.28, 28.2; ¹⁹F NMR (376 MHz, CDCl₃) δ −148.4; HRMS (ESI) calcd for C₂₈H₂₉FNO₇ [M − H]⁻ 510.1934, found 510.1933 (error 0.2 ppm).

4.1.12.2. 4-Methoxybenzyl 4-[(tert-butoxycarbonyl)amino]-2-fluoro-6-[(4-methoxybenzyl)oxy]benzoate (28b).

The title compound was synthesized from 27b using the general method for Buchwald-Hartwig coupling and isolated (430 mg, 84%) as a yellow solid. R_f =0.49 (2:1 Hexane–EtOAc); ¹H NMR (400 MHz, CDCl₃) δ 7.29–7.24 (m, 4H), 6.98 (s, 1H), 6.86 (d, J = 8.6 Hz, 2H), 6.81 (d, J = 8.7 Hz, 2H), 6.70 (dd, ³J_H-F = 11.4, ⁴J_H-H = 1.8 Hz, 1H), 6.60 (s, 1H), 5.24 (s, 2H), 4.99 (s, 2H), 3.81 (s, 3H), 3.79 (s, 3H), 1.51 (s, 9H); ¹³C NMR (151 MHz, CDCl₃) δ 163.7, 161.3 (d, ¹J_C-F = 249.5 Hz), 159.5, 159.4, 158.3 (d, ³J_C-F = 8.6 Hz), 152.0, 142.1 (d, ³J_C-F = 14.3 Hz), 130.0, 129.1, 128.1, 127.9, 113.9, 113.8, 106.5 (d, ²J_C-F = 18.4 Hz), 98.2, 98.0 (d, ²J_C-F = 27.2 Hz), 81.4, 70.7, 67.0, 55.27, 55.25, 28.3; ¹⁹F NMR (376 MHz, CDCl₃) δ −111.8; HRMS (ESI) calcd for C₂₈H₂₉FNO₇ [M − H]⁻ 510.1934, found 510.1952 (error 3.5 ppm).

4.2. Microbiological assays

4.2.1. Bacterial Strains and Media.

M. tuberculosis H37Rv was a gift from W. R. Jacobs, Jr. of the Albert Einstein College of Medicine. For determination of MICs, M. tuberculosis H37Rv was grown in Middlebrook 7H9 liquid medium supplemented with 10% oleate-albumin-dextrose-catalase, 0.2% glycerol and 0.05% tyloxapol. Other drug susceptible and drug resistant M. tuberculosis strains used were K20b00MR.NIH 57, NIH_G21R, K33b00MR, NIH_G10R, NIH_G11R, K18b01MR.NIH 79, 028K111NH_G269DR, H37Rv ATCC 27294, H37Rv folC: E153A, H37Rv folC: S150G, H37Rv folC: R49W, K04b00DS, 0K116, NIH_G367DR, K37b00XR, K32b00MR and 001K113a; which are a subset of those described in Jeon et al (2008)[72] and Shamputa et al (2010).[73] These strains were grown in standard Middlebrook 7H9 broth (BD Difco) supplemented with 0.5% albumin, 0.2% glucose, 0.085% sodium chloride, 0.2% glycerol, and 0.05% Tween 80.

4.2.2. MIC Determination.

Determination of liquid culture minimum inhibitory concentration (MIC) was performed as previously described.[74] Mycobacterium tuberculosis H37Rv was grown exponentially to mid-log phase to an OD₆₀₀ of ~0.5-0.8 and then inoculated into 30-mL TriForest square-inkwell bottles with 5 mL of supplemented 7H9 liquid medium containing PAS compounds to a starting OD₆₀₀ of 0.01. PAS and PAS-analogs were added to a range of concentrations as indicated in the figures or text. Cultures were incubated with shaking at 100 rpm at 37 °C for 10 days total and following incubation, MICs were determined by measuring the optical density (OD₆₀₀) of each culture. The liquid MIC₉₀ was defined as the minimum inhibitory concentration required to inhibit at least 90% of growth compared to no drug control bottles. These experiments were completed twice with biological replicates for each drug assayed.

4.2.3. MIC determination for susceptible and resistant mycobacterial strains.

M. tuberculosis strains were grown in the respective medium of choice to an OD₆₅₀ of 0.2. Middlebrook 7H9 medium was supplemented with 0.5% BSA fraction V, 0.08% NaCl, 0.2% glucose, 0.2% glycerol and 0.05% Tween 80 (7H9/ADC/Tw); The cultures were diluted 1000-fold in fresh medium and 50 μL dispensed per well in round-bottom clear 96-well plate (Nunclon) containing 50 μL of the respective medium per well, with or without 2-fold serial dilutions of compound. Plates were sealed in ziplock bags and incubated at 37 °C for 2 weeks. Growth was recorded after 1- and 2-weeks of growth by monitoring growth using an inverted enlarging mirror. The MIC was determined as the lowest concentration that completely inhibited growth. Each compound was tested in duplicate. DMSO was used as negative control which resulted in no growth inhibition up to 2%. Isoniazid and linezolid served as positive control drugs

4.3. Enzymatic assays

4.3.1. hNAT1 Enzyme Assay for Substrate Specificity Studies.

All reactions were conducted in triplicate under initial velocity conditions in a final volume of 1.00 mL using a 1 cm pathlength quartz cuvette on a Varian Cary 50 UV-Vis spectrophotometer. The plasmid for hNAT1 F125S was obtained from Addgene, and expressed and purified using standard protocols reported by Hanna et al.[65] The activity of NAT-1 was assayed using a continuous assay measuring the increase in absorbance at 400 nm due to release of 4-nitrophenol (δ_400nm = 9400 M⁻¹ cm⁻¹) from 4-nitrophenyl acetate (PNPA). The assay mixture contained human NAT-1 (20 nM), PNPA (2 mM), various concentration of substrates PABA, PAS, 11 or 14 in MOPS buffer (100 mM, pH 7.0; 150 mM NaCl; 0.1 mM DTT). Incubations were conducted at 25 °C and initiated by addition of PNPA dissolved in DMSO (10 μL providing a final concentration of 1% DMSO in the assay) and mixed by immediately inverting the cuvette a couple of times. Serial two-fold dilutions of substrate concentrations between 12.5–2000 μM were used. The results were corrected for nonenzymatic hydrolysis of PNPA by conducting the reaction in the absence of protein. The initial velocities were normalized with respect to enzyme concentrations and are expressed as μmol of product formed per μmol of protein per min to provide velocity in min⁻¹. The data were fit to the Michaelis-Menten equation 1:

$$v=\frac{k_{cat}[S]}{K_M+[S]}$$ (Eq. 1)

using GraphPad Prism 8.0 where v is the initial velocity of the reaction, K_M is the Michaelis-Menten constant, k_cat is the turnover number, and S is the substrate concentration.

4.4. Pharmacokinetics studies

CD-1 female mice (22-25 g) were used in oral pharmacokinetic studies. All animal studies were performed in biosafety level 2 (BSL2) facilities and approved by the Institutional Animal Care and Use Committee of the New Jersey Medical School, Rutgers University, Newark, NJ.

4.4.1. Snapshot PK profiling.

PAS analogs/prodrugs were administered as a single dose by oral gavage at 10 mg/kg (2–8) or 25 mg/kg (9–15) in a solution formulation composed of 5% dimethylacetamide:60% polyethylene glycol 300 (PEG 300): 35% (5% dextrose in water). Aliquots of 50 μL of blood were taken by puncture of the lateral tail vein from each mouse (n = 2 per route and dose) at 0.5, 1-, 3-, and 5-h post-dose and captured in CB300 blood collection tubes containing K2EDTA and stored on ice. Plasma was recovered after centrifugation and stored at −80 °C until analyzed by high pressure liquid chromatography coupled to tandem mass spectrometry.

4.4.2. Full PK profiling.

PAS and PAS 6 were administered as a single dose by oral gavage at 25 mg/kg for Full PK (n = 3) profiling using a suspension formulation composed of 0.5% CMC/0.5% Tween 80. Aliquots of blood were taken at 30 minutes, 1, 3, 5, 7, and 24 h for full PK profiling. Blood was captured in CB300 blood collection tubes containing K2EDTA and stored on ice. Plasma was recovered after centrifugation and stored at −80°C until analyzed by high pressure liquid chromatography coupled to tandem mass spectrometry. Only the active PAS was monitored in plasma study samples due to the high rate of metabolism of PAS 6 into PAS in mouse plasma.

4.4. LC-MS/MS analytical methods

Neat 1 mg/mL DMSO stocks of PAS analogs/prodrugs were serial diluted in 50/50 MeCN:H₂O to create standard curve solutions. Standards were created by adding 10 μL of spiking solutions to 90 μL of drug free plasma (CD-1 K2EDTA Mouse, Bioreclamation IVT). 5 μL of control, standard, or study sample were added to 100 μL of MeCN protein precipitation solvent containing 100 ng/mL of the internal standard Labetalol (Sigma Aldrich) or 10 ng/mL of the internal standard Verapamil (Sigma Aldrich). Extracts were vortexed for 5 min and centrifuged at 4000 RPM for 5 min. 75 μL of supernatant was transferred for HPLC-MS/MS analysis and diluted with 75 μL of Milli-Q deionized H₂O. LC-MS/MS analysis was performed on a Sciex Applied Biosystems Qtrap 6500+ triple-quadrupole mass spectrometer coupled to a Shimadzu Nexera X2 UHPLC system to quantify each drug in plasma. Chromatography was performed on a Waters HSS Cyano column (2.1×100 mm; particle size, 5 μm) using a reverse phase gradient. Milli-Q deionized H₂O with 0.1% formic acid was used for the aqueous mobile phase and 0.1% formic acid in MeCN for the organic mobile phase. Multiple-reaction monitoring of parent/daughter transitions in electrospray negative-ionization mode or electrospray positive-ionization mode was used to quantify compounds. The following MRM transitions were used for PAS (154.0/136.0), 2 (168.0/ 62.9), 3 (182.0/150.0), 4 (182.2/136.0), 5 (196.0/154.0), 6-10 (154.0/136.0), 11 (169.9/104.8), 12 (169.9/125.8), 13 (169.9/63.9), 14 (168.18/150.10), 15 (177.9/105.8), 16 (200.0/136.9), Labetalol (327.2/176.0), and Verapamil (455.40/165.00). Sample analysis was accepted if the concentrations of the quality control samples were within 20% of the nominal concentration. Data processing was performed using Analyst software (version 1.6.2; Applied Biosystems Sciex).

4.5. Plasma Stability

The stability assays were carried out using plasma from female CD-1 mice and human with K 2 EDTA anticoagulant (Bioreclamation). Stability samples consisted of 5 μL of stock compound (20 μg/mL ) solution in 100% DMSO and 95 μL of plasma to a final concentration of 1 μg/mL. The samples were incubated at 37 °C with shaking; 10 μL were removed at each time point and combined with 100 μL of 1:1 MeCN/MeOH, 10 ng/mL of verapamil, and 10 μL of MeCN/H₂O. Samples were analyzed by LC-MS/MS using the Q-Exactive high-resolution mass spectrometer (Thermo Fisher Scientific) using 5 ppm mass accuracy. Stability was calculated as the percent remaining of the parent compound compared to the initial concentration otherwise.

Highlights.

• p-Aminosalicylic acid (PAS), a potent anti-tubercular agent, has limiting oral bioavailability

• Prodrugs and structural analogues improve pharmacokinetic properties of drugs

• Pivoxyl PAS prodrug quantitatively improves oral bioavailability of PAS

• PAS analogues show a 5 to 430-fold improvement in oral bioavailability, but result in a loss of potency

• Increased exposures of 5-Fluoro PAS and N-Methyl PAS are due to poor substrate specificities towards N-acetyl transferase 1