Exploring the Antitubercular Activity of Anthranilic Acid Derivatives: From MabA (FabG1) Inhibition to Intrabacterial Acidification

Mycobacterium tuberculosis, the pathogen that causes tuberculosis, is responsible for the death of 1.5 million people each year and the number of bacteria resistant to the standard regimen is constantly increasing. This highlights the need to discover molecules that act on new M. tuberculosis targets. Mycolic acids, which are very long-chain fatty acids essential for M. tuberculosis viability, are synthesized by two types of fatty acid synthase (FAS) systems. MabA (FabG1) is an essential enzyme belonging to the FAS-II cycle. We have recently reported the discovery of anthranilic acids as MabA inhibitors. Here, the structure–activity relationships around the anthranilic acid core, the binding of a fluorinated analog to MabA by NMR experiments, the physico-chemical properties and the antimycobacterial activity of these inhibitors were explored. Further investigation of the mechanism of action in bacterio showed that these compounds affect other targets than MabA in mycobacterial cells and that their antituberculous activity is due to the carboxylic acid moiety which induces intrabacterial acidification.


Introduction
Tuberculosis (TB), an infectious disease caused by Mycobacterium tuberculosis, is one of the leading causes of death from a single infectious agent [1]. The World Health Organization (WHO) estimates that in 2021, 10.6 million people developed the disease and 1.6 million died from TB including 187,000 deaths among HIV-positive people [2]. In addition, the number of bacteria resistant to anti-TB drugs is constantly increasing and in 2021, around 450,000 people developed rifampicin-resistant TB, of which 78% had multidrug-resistant TB (MDR-TB) [2].
In the last 10 years, three new drugs, bedaquiline [3], delamanid [4] and pretomanid [5] have been approved for the treatment of resistant TB. Nevertheless, treatment success rates remain insufficient, particularly in patients with MDR-and XDR-TB (extensively drugresistant) [6], and bacteria resistance to these new drugs has already been reported in many countries, warning that new compounds acting on unexploited targets are needed.
In the last 10 years, three new drugs, bedaquiline [3], delamanid [4] and pretomanid [5] have been approved for the treatment of resistant TB. Nevertheless, treatment success rates remain insufficient, particularly in patients with MDR-and XDR-TB (extensively drug-resistant) [6], and bacteria resistance to these new drugs has already been reported in many countries, warning that new compounds acting on unexploited targets are needed.
Mycolic acids, a key component of the bacterial cell wall, are essential for M. tuberculosis viability [7]. Characterized as very long-chain α-alkyl-β-hydroxy fatty acids (C70-C90) and composed of a C24-C26 saturated α-chain and a meromycolate chain up to C56, their biosynthesis is carried out by two types of fatty acid synthase (FAS) systems, FAS-I and FAS-II [8,9]. FAS-I synthesizes short-chain fatty acids with two different chain lengths: C24-C26 which correspond to the α-branch in mycolic acids and C16-C18 which are then elongated by FAS-II to produce the meromycolate chain [10]. FAS-II is composed of four types of enzymes, which act successively to ensure fatty acid elongation. Several cycles catalyzed by MabA, HadAB/BC, InhA and KasA/B lead to very long meromycolic acid chains (C42-C56) which will produce mycolic acids [11]. These enzymes have been shown to be essential for bacterial growth [12][13][14][15] and several antibiotics target InhA, HadAB and KasA/B [16][17][18][19]. However, MabA (FabG1) was the only enzyme of the FAS-II cycle without a specific inhibitor. We have recently reported the discovery of the first MabA inhibitors through a fragment-based screening approach [20].
MabA catalyzes the NADPH-specific reduction of β-ketoacyl derivatives into β-hydroxyacyl derivatives [21]. In order to discover MabA inhibitors, we developed a new LC-MS/MS-based enzymatic assay in microplates [20]. In this assay, MabA catalyzes the reduction of acetoacetyl-CoA into hydroxybutyryl-CoA (HBCoA) using NADPH as a cofactor and we detect the formation of HBCoA by mass spectrometry. The screening of a 1280fragment library led to the selection of an anthranilic acid series and early structure-activity relationships studies led to the identification of compound 1. This inhibitor, composed of a 3,4-dichlorobenzoyl motif linked to the nitrogen of the anthranilic core and an iodine atom in position 5 of the phenyl ring, inhibits MabA with an IC50 of 38 ± 6 µ M (Figure 1) [20]. The direct brominated analog (compound 2) displayed a similar potency (IC50 = 45 ± 6 µ M).  [20].
With the aim of increasing the potency of our compounds, the structure-activity relationships (SAR) around the anthranilic acid core were further explored. The SAR study includes the bioisosteric replacement of the amide and carboxylic acid functions as well as the modification of the iodophenyl ring. The direct binding of a fluorinated anthranilic  [20].
With the aim of increasing the potency of our compounds, the structure-activity relationships (SAR) around the anthranilic acid core were further explored. The SAR study includes the bioisosteric replacement of the amide and carboxylic acid functions as well as the modification of the iodophenyl ring. The direct binding of a fluorinated anthranilic acid analog to MabA was confirmed by 19 F ligand-observed NMR experiments. Physicochemical properties and antimycobacterial activity of five chemically diverse compounds were then evaluated. Further study on how these compounds affect bacterial growth revealed that they impact targets other than MabA in mycobacterial cells and that their antituberculous activity is due to the carboxylic acid moiety which causes acidification inside the bacteria. In the iodine series, the first modifications introduced were the methylation of the nitrogen of the amide linker leading to compound 3 and the replacement of the amide bond by a sulfonamide leading to compound 4. In the bromine series, the retro amide (compound 5) and thioamide (compound 6) analogs were synthesized. The bioisosteric replacement of the amide by a 1,2,4-oxadiazole ring (compound 7) was also tested [22].
Following the methodology described previously [20], the coupling between 2-amino-5-bromo benzoic acid and isopropoxycarbonyl 3,4-dichlorobenzoate led to compound 2 (Scheme 1). Compound 3 was obtained by alkylation of compound 1 with an excess of iodomethane, which led to the alkylation of both the amide and the carboxylic acid. The hydrolysis of the methyl ester was then performed using LiOH. Compound 4 was synthesized by coupling 2-amino-5-iodobenzoic acid with 3,4-dichlorobenzenesulfonyl chloride. The reaction of 4-bromo-phthalic anhydride with 3-4-dichloroaniline led to compound 5 and its regioisomer which were separated by preparative HPLC. Thioamide 6 was synthesized following a three-step procedure: firstly, methyl-2-amino-5-bromobenzoate was coupled with 3,4-dichlorobenzoyl chloride to give the corresponding amide, followed by thionation using Lawesson's reagent, and finally, the desired compound 6 was obtained by hydrolysis of the methyl ester using NaOH. The 1,2,4-oxadiazole analog (compound 7) was also obtained after three chemical steps. Firstly, 4-bromo-2-methoxycarbonyl-benzoic acid was functionalized in the corresponding acid chloride using oxalyl chloride, which was then coupled with 3,4-dichloro-N -hydroxy-benzamidine. The 1,2,4-oxadiazole ring formation was obtained at 100 • C and the hydrolysis of the methyl ester using NaOH led to the corresponding acid (compound 7). acid analog to MabA was confirmed by 19 F ligand-observed NMR experiments. Physicochemical properties and antimycobacterial activity of five chemically diverse compounds were then evaluated. Further study on how these compounds affect bacterial growth revealed that they impact targets other than MabA in mycobacterial cells and that their antituberculous activity is due to the carboxylic acid moiety which causes acidification inside the bacteria.

Study of the Modification of the Amide Bond
In the iodine series, the first modifications introduced were the methylation of the nitrogen of the amide linker leading to compound 3 and the replacement of the amide bond by a sulfonamide leading to compound 4. In the bromine series, the retro amide (compound 5) and thioamide (compound 6) analogs were synthesized. The bioisosteric replacement of the amide by a 1,2,4-oxadiazole ring (compound 7) was also tested [22].

Study of the Modification of the Phenyl Ring
The substitution of the phenyl ring and the introduction of new aromatic rings were explored. Since an increase in activity correlated with an increase in halogen size has been previously observed in the Boc-anthranilic acid series [20], we assumed that a bulky and hydrophobic substituent was preferable for activity. Firstly, we moved the bromine atom to position 4 on the phenyl ring of the anthranilic acid core to obtain compound 14. We then replaced the bromine atom with a hydrogen atom (compound 15) to confirm the loss of activity. The introduction of alkyl substituents was then explored. As the ethynyl group was described as a bioisostere of the iodine atom due to their similar molecular electrostatic potentials [24], an ethynyl group was introduced in position 4 (compound 17) as well as a cyclopropylethynyl in position 5 (compound 16). To complete this exploration, a pyrazole was introduced in position 4 (compound 18). Finally, the introduction of a naphthalene (compound 19) and a quinoline (compound 20) as a replacement for the phenyl core was also investigated.
Amides 14 and 15 were obtained as described above from commercial anilines (Scheme 3). Alkynes 16 and 17 were obtained using a Sonogashira coupling from iodo or bromo derivatives. Compounds 18 and 19 were isolated after the coupling of the appropriate anilines with isopropoxycarbonyl 3,4-dichlorobenzoate. Finally, 2-nitrobenzaldehyde was used in a Knoevenagel condensation with ethyl 2-cyanoacetate, then reduction of the nitro group followed by an intramolecular cyclization led to ethyl 2-aminoquinoline-3-carboxylate. Coupling with 3,4-dichlorobenzoyl chloride and hydrolysis of the ethyl ester led to the desired compound 20.

Study of the Modification of the Phenyl Ring
The substitution of the phenyl ring and the introduction of new aromatic rings were explored. Since an increase in activity correlated with an increase in halogen size has been previously observed in the Boc-anthranilic acid series [20], we assumed that a bulky and hydrophobic substituent was preferable for activity. Firstly, we moved the bromine atom to position 4 on the phenyl ring of the anthranilic acid core to obtain compound 14. We then replaced the bromine atom with a hydrogen atom (compound 15) to confirm the loss of activity. The introduction of alkyl substituents was then explored. As the ethynyl group was described as a bioisostere of the iodine atom due to their similar molecular electrostatic potentials [24], an ethynyl group was introduced in position 4 (compound 17) as well as a cyclopropylethynyl in position 5 (compound 16). To complete this exploration, a pyrazole was introduced in position 4 (compound 18). Finally, the introduction of a naphthalene (compound 19) and a quinoline (compound 20) as a replacement for the phenyl core was also investigated.

Biological Results
All compounds were evaluated for their inhibitory activity on MabA using our LC-MS/MS-based enzymatic assay. The potency of the compounds is reported as the concentration inhibiting 50% of the transformation of acetoacetyl-CoA in HBCoA in the presence of NADPH. In the iodine series, the replacement of the amide function with a sulfonamide function (compound 4) led to a compound as potent as the reference compound 1 while methylation of the amide function (compound 3) decreased the potency by more than five times (Table 1). In the bromine series, the replacement of the amide function with a thioamide function was tolerated (compound 6) while the introduction of a reverse amide or a 1,2,4-oxadiazole (compounds 5 and 7) was less favorable.

Biological Results
All compounds were evaluated for their inhibitory activity on MabA using our LC-MS/MS-based enzymatic assay. The potency of the compounds is reported as the concentration inhibiting 50% of the transformation of acetoacetyl-CoA in HBCoA in the presence of NADPH.

Modifications of the Amide Bond
In the iodine series, the replacement of the amide function with a sulfonamide function (compound 4) led to a compound as potent as the reference compound 1 while methylation of the amide function (compound 3) decreased the potency by more than five times (Table 1). In the bromine series, the replacement of the amide function with a thioamide function was tolerated (compound 6) while the introduction of a reverse amide or a 1,2,4oxadiazole (compounds 5 and 7) was less favorable.

7
Br 102 The exploration of the replacement of the amide function did not lead to more potent analogs; therefore, all the following modifications were performed in the amide series.

Modifications of the Carboxylic Acid Moiety
The replacement of the carboxylic acid with alcohol (compound 8), a carboxamide (compound 9) or a nitrile (compound 10) was deleterious for activity (IC50 > 333 µ M). On the contrary, its replacement by bioisosteres such as a tetrazole (compound 11) or acylsulfonamides (compounds 12-13) was well tolerated and led to compounds with IC50 between 24 and 35 µ M (Table 2). Overall, these SARs show that the presence of an acid function is required to inhibit MabA in vitro.

Biological Results
All compounds were evaluated for their inhibitory activity on MabA using our LC-MS/MS-based enzymatic assay. The potency of the compounds is reported as the concentration inhibiting 50% of the transformation of acetoacetyl-CoA in HBCoA in the presence of NADPH.

Modifications of the Amide Bond
In the iodine series, the replacement of the amide function with a sulfonamide function (compound 4) led to a compound as potent as the reference compound 1 while methylation of the amide function (compound 3) decreased the potency by more than five times (Table 1). In the bromine series, the replacement of the amide function with a thioamide function was tolerated (compound 6) while the introduction of a reverse amide or a 1,2,4oxadiazole (compounds 5 and 7) was less favorable.

7
Br 102 The exploration of the replacement of the amide function did not lead to more potent analogs; therefore, all the following modifications were performed in the amide series.

Modifications of the Carboxylic Acid Moiety
The replacement of the carboxylic acid with alcohol (compound 8), a carboxamide (compound 9) or a nitrile (compound 10) was deleterious for activity (IC50 > 333 µ M). On the contrary, its replacement by bioisosteres such as a tetrazole (compound 11) or acylsulfonamides (compounds 12-13) was well tolerated and led to compounds with IC50 between 24 and 35 µ M (Table 2). Overall, these SARs show that the presence of an acid function is required to inhibit MabA in vitro. The exploration of the replacement of the amide function did not lead to more potent analogs; therefore, all the following modifications were performed in the amide series.

Modifications of the Carboxylic Acid Moiety
The replacement of the carboxylic acid with alcohol (compound 8), a carboxamide (compound 9) or a nitrile (compound 10) was deleterious for activity (IC 50 > 333 µM). On the contrary, its replacement by bioisosteres such as a tetrazole (compound 11) or acylsulfonamides (compounds 12-13) was well tolerated and led to compounds with IC 50 between 24 and 35 µM (Table 2). Overall, these SARs show that the presence of an acid function is required to inhibit MabA in vitro.

Modifications of the Phenyl Ring
Removal of the halogen atom on the phenyl ring (compound 15) led to an inactive compound (IC 50 > 1000 µM). The introduction of a bromine atom in position 4 of the phenyl ring (compound 14) instead of position 5 (compound 2) led to a compound with similar activity suggesting that both positions are tolerated for the introduction of substituents (Table 3). This observation was confirmed by the introduction of a fused phenyl ring (compound 19), as a similar activity to that of the halogenated compounds was observed for this naphthyl analog. Replacing the naphthyl ring with a quinoline ring (compound 20) resulted in a 2-fold reduction in the potency (IC 50 = 76 µM). The replacement of the iodine atom in position 5 of the phenyl ring with a cyclopropyl-substituted alkyne (compound 16) led to a one-third reduction in potency (IC 50 = 60 µM) compared to the reference compound whereas the introduction of an unsubstituted alkyne in position 4 (compound 17) led to a 2-fold decrease in potency (IC 50 = 77 µM). Interestingly, the introduction of a pyrazole in position 4 of the phenyl ring (compound 18) improved the activity by a factor of (1.6 IC 50 = 23 µM).

Modifications of the Carboxylic Acid Moiety
The replacement of the carboxylic acid with alcohol (compound 8), a carboxamide (compound 9) or a nitrile (compound 10) was deleterious for activity (IC50 > 333 µ M). On the contrary, its replacement by bioisosteres such as a tetrazole (compound 11) or acylsulfonamides (compounds 12-13) was well tolerated and led to compounds with IC50 between 24 and 35 µ M (Table 2). Overall, these SARs show that the presence of an acid function is required to inhibit MabA in vitro.

Modifications of the Phenyl Ring
Removal of the halogen atom on the phenyl ring (compound 15) led to an inactive compound (IC50 > 1000 µ M). The introduction of a bromine atom in position 4 of the phenyl ring (compound 14) instead of position 5 (compound 2) led to a compound with similar activity suggesting that both positions are tolerated for the introduction of substituents (Table 3). This observation was confirmed by the introduction of a fused phenyl ring (compound 19), as a similar activity to that of the halogenated compounds was observed for this naphthyl analog. Replacing the naphthyl ring with a quinoline ring (compound 20) resulted in a 2-fold reduction in the potency (IC50 = 76 µ M). The replacement of the iodine atom in position 5 of the phenyl ring with a cyclopropyl-substituted alkyne (compound 16) led to a one-third reduction in potency (IC50 = 60 µ M) compared to the reference compound whereas the introduction of an unsubstituted alkyne in position 4 (compound 17) led to a 2-fold decrease in potency (IC50 = 77 µ M). Interestingly, the introduction of a pyrazole in position 4 of the phenyl ring (compound 18) improved the activity by a factor of 1.6 (IC50 = 23 µ M). Table 3. Biological activities of compounds 1, 2, 14-20.

13
-CO-NH-SO 2 -CH 3 28 These SARs show that the introduction of a bulky substituent in position 4 or 5 of the phenyl ring is important for activity. This suggests that a new hydrophobic contact with the binding pocket of MabA may have been created. The replacement of the halogen atom by ethynyl groups, which are described as bioisosteres of the iodine atom [24], is well tolerated and supports this hypothesis. 19 F-NMR experiments have already been used to confirm the binding of fluorinated inhibitors to MabA [20]. Ligand-observed NMR techniques can be applied to detect the weak binding of small molecules as measured by the reduction in fluorine or proton signal intensity [25,26]. Compound 12 bearing a trifluoromethyl group was selected for NMR experiments using a 600 MHz spectrometer equipped with a 19 F cryogenic probe. The disappearance of 19  These SARs show that the introduction of a bulky substituent in position 4 or 5 of the phenyl ring is important for activity. This suggests that a new hydrophobic contact with the binding pocket of MabA may have been created. The replacement of the halogen atom by ethynyl groups, which are described as bioisosteres of the iodine atom [24], is well tolerated and supports this hypothesis. 19 F-NMR experiments have already been used to confirm the binding of fluorinated inhibitors to MabA [20]. Ligand-observed NMR techniques can be applied to detect the weak binding of small molecules as measured by the reduction in fluorine or proton signal intensity [25,26]. Compound 12 bearing a trifluoromethyl group was selected for NMR experiments using a 600 MHz spectrometer equipped with a 19 F cryogenic probe. The disappearance of 19 F-and 1 H-NMR signals of compound 12 in the presence of MabA confirmed the direct binding of this inhibitor to the protein ( Figure 2). Interestingly, a comparison of the 19 F-NMR signal observed with compound 12 and three previously published MabA inhibitors [20] shows a correlation between the signal decrease and the inhibitory activity of the compounds. Experiments carried out with compound 12, which displays the best activity, show a total disappearance of the signal in contrast to the other three compounds where the signal is still visible (Supplementary materials Figure S1).

Evaluation of Physico-Chemical Properties and Antimycobacterial Activity
The two reference compounds (1 and 2) and three analogs (12, 16 and 18), chemically Interestingly, a comparison of the 19 F-NMR signal observed with compound 12 and three previously published MabA inhibitors [20] shows a correlation between the signal decrease and the inhibitory activity of the compounds. Experiments carried out with compound 12, which displays the best activity, show a total disappearance of the signal in contrast to the other three compounds where the signal is still visible (Supplementary materials Figure S1).

20)
resulted in a 2-fold reduction in the potency (IC50 = 76 µ M). The replacement of the iodine atom in position 5 of the phenyl ring with a cyclopropyl-substituted alkyne (compound 16) led to a one-third reduction in potency (IC50 = 60 µ M) compared to the reference compound whereas the introduction of an unsubstituted alkyne in position 4 (compound 17) led to a 2-fold decrease in potency (IC50 = 77 µ M). Interestingly, the introduction of a pyrazole in position 4 of the phenyl ring (compound 18) improved the activity by a factor of 1.6 (IC50 = 23 µ M).

20)
resulted in a 2-fold reduction in the potency (IC50 = 76 µ M). The replacement of the iodine atom in position 5 of the phenyl ring with a cyclopropyl-substituted alkyne (compound 16) led to a one-third reduction in potency (IC50 = 60 µ M) compared to the reference compound whereas the introduction of an unsubstituted alkyne in position 4 (compound 17) led to a 2-fold decrease in potency (IC50 = 77 µ M). Interestingly, the introduction of a pyrazole in position 4 of the phenyl ring (compound 18) improved the activity by a factor of 1.6 (IC50 = 23 µ M).

20)
resulted in a 2-fold reduction in the potency (IC50 = 76 µ M). The replacement of the iodine atom in position 5 of the phenyl ring with a cyclopropyl-substituted alkyne (compound 16) led to a one-third reduction in potency (IC50 = 60 µ M) compared to the reference compound whereas the introduction of an unsubstituted alkyne in position 4 (compound 17) led to a 2-fold decrease in potency (IC50 = 77 µ M). Interestingly, the introduction of a pyrazole in position 4 of the phenyl ring (compound 18) improved the activity by a factor of 1.6 (IC50 = 23 µ M).

20)
resulted in a 2-fold reduction in the potency (IC50 = 76 µ M). The replacement of the iodine atom in position 5 of the phenyl ring with a cyclopropyl-substituted alkyne (compound 16) led to a one-third reduction in potency (IC50 = 60 µ M) compared to the reference compound whereas the introduction of an unsubstituted alkyne in position 4 (compound 17) led to a 2-fold decrease in potency (IC50 = 77 µ M). Interestingly, the introduction of a pyrazole in position 4 of the phenyl ring (compound 18) improved the activity by a factor of 1.6 (IC50 = 23 µ M). 20) resulted in a 2-fold reduction in the potency (IC50 = 76 µ M). The replacement of the iodine atom in position 5 of the phenyl ring with a cyclopropyl-substituted alkyne (compound 16) led to a one-third reduction in potency (IC50 = 60 µ M) compared to the reference compound whereas the introduction of an unsubstituted alkyne in position 4 (compound 17) led to a 2-fold decrease in potency (IC50 = 77 µ M). Interestingly, the introduction of a pyrazole in position 4 of the phenyl ring (compound 18) improved the activity by a factor of 1.6 (IC50 = 23 µ M).     These SARs show that the introduction of a bulky substituent in position 4 or 5 of the phenyl ring is important for activity. This suggests that a new hydrophobic contact with the binding pocket of MabA may have been created. The replacement of the halogen atom by ethynyl groups, which are described as bioisosteres of the iodine atom [24], is well tolerated and supports this hypothesis. 19 F-NMR experiments have already been used to confirm the binding of fluorinated inhibitors to MabA [20]. Ligand-observed NMR techniques can be applied to detect the weak binding of small molecules as measured by the reduction in fluorine or proton signal intensity [25,26]. Compound 12 bearing a trifluoromethyl group was selected for NMR experiments using a 600 MHz spectrometer equipped with a 19 F cryogenic probe. The disappearance of 19 F-and 1 H-NMR signals of compound 12 in the presence of MabA confirmed the direct binding of this inhibitor to the protein ( Figure 2). These SARs show that the introduction of a bulky substituent in position 4 or 5 of the phenyl ring is important for activity. This suggests that a new hydrophobic contact with the binding pocket of MabA may have been created. The replacement of the halogen atom by ethynyl groups, which are described as bioisosteres of the iodine atom [24], is well tolerated and supports this hypothesis. 19 F-NMR experiments have already been used to confirm the binding of fluorinated inhibitors to MabA [20]. Ligand-observed NMR techniques can be applied to detect the weak binding of small molecules as measured by the reduction in fluorine or proton signal intensity [25,26]. Compound 12 bearing a trifluoromethyl group was selected for NMR experiments using a 600 MHz spectrometer equipped with a 19 F cryogenic probe. The disappearance of 19 F-and 1 H-NMR signals of compound 12 in the presence of MabA confirmed the direct binding of this inhibitor to the protein ( Figure 2). These SARs show that the introduction of a bulky substituent in position 4 or 5 of the phenyl ring is important for activity. This suggests that a new hydrophobic contact with the binding pocket of MabA may have been created. The replacement of the halogen atom by ethynyl groups, which are described as bioisosteres of the iodine atom [24], is well tolerated and supports this hypothesis. 19 F-NMR experiments have already been used to confirm the binding of fluorinated inhibitors to MabA [20]. Ligand-observed NMR techniques can be applied to detect the weak binding of small molecules as measured by the reduction in fluorine or proton signal intensity [25,26]. Compound 12 bearing a trifluoromethyl group was selected for NMR experiments using a 600 MHz spectrometer equipped with a 19 F cryogenic probe. The disappearance of 19 F-and 1 H-NMR signals of compound 12 in the presence of MabA confirmed the direct binding of this inhibitor to the protein ( Figure 2). 76

Evaluation of Physico-Chemical Properties and Antimycobacterial Activity
The two reference compounds (1 and 2) and three analogs (12, 16 and 18), chemically diverse and of similar potency to the reference compounds, were selected to further evaluate their physico-chemical properties ( Table 4). All compounds displayed a solubility greater than 140 µM and a logD between 2.54 and 4.10. Plasma protein binding was measured and these compounds, which are hydrophobic and bear a carboxylic acid function, were highly bound to plasma proteins. The introduction of a pyrazole ring (compound 18) expectedly decreased lipophilicity but did not reduce plasma protein binding.
The antibacterial activity of compounds 1 and 2 reported previously was assessed using green fluorescent protein (GFP)-expressing M. tuberculosis H37Rv grown in Middlebrook 7H9 liquid medium supplemented with Oleic Albumin Dextrose Catalase (OADC). After 5 days of incubation, the GFP signal was measured and showed that both compounds had an equivalent potency to inhibit 90% of the bacterial growth (MIC 90 = 300 µM). The bacterial culture medium used in this previous study contained albumin, which we know now may limit the quantity of unbound compounds 1 and 2. In the present study, the cul-Pharmaceuticals 2023, 16, 335 9 of 27 ture medium was therefore switched to Sauton, which does not contain albumin. In these conditions, compound 1 displayed a MIC 90 of 100 µM while the addition of albumin to the Sauton medium reversed its MIC 90 to 300 µM ( Figure 3A). A similar pattern was observed in the presence of albumin for all compounds tested such as compound 18 ( Figure 3B). were highly bound to plasma proteins. The introduction of a pyrazole ring (compound 18) expectedly decreased lipophilicity but did not reduce plasma protein binding. The antibacterial activity of compounds 1 and 2 reported previously was assessed using green fluorescent protein (GFP)-expressing M. tuberculosis H37Rv grown in Middlebrook 7H9 liquid medium supplemented with Oleic Albumin Dextrose Catalase (OADC). After 5 days of incubation, the GFP signal was measured and showed that both compounds had an equivalent potency to inhibit 90% of the bacterial growth (MIC90 = 300 µ M). The bacterial culture medium used in this previous study contained albumin, which we know now may limit the quantity of unbound compounds 1 and 2. In the present study, the culture medium was therefore switched to Sauton, which does not contain albumin. In these conditions, compound 1 displayed a MIC90 of 100 µ M while the addition of albumin to the Sauton medium reversed its MIC90 to 300 μM ( Figure 3A). A similar pattern was observed in the presence of albumin for all compounds tested such as compound 18 ( Figure 3B). were highly bound to plasma proteins. The introduction of a pyrazole ring (compound 18) expectedly decreased lipophilicity but did not reduce plasma protein binding. The antibacterial activity of compounds 1 and 2 reported previously was assessed using green fluorescent protein (GFP)-expressing M. tuberculosis H37Rv grown in Middlebrook 7H9 liquid medium supplemented with Oleic Albumin Dextrose Catalase (OADC). After 5 days of incubation, the GFP signal was measured and showed that both compounds had an equivalent potency to inhibit 90% of the bacterial growth (MIC90 = 300 µ M). The bacterial culture medium used in this previous study contained albumin, which we know now may limit the quantity of unbound compounds 1 and 2. In the present study, the culture medium was therefore switched to Sauton, which does not contain albumin. In these conditions, compound 1 displayed a MIC90 of 100 µ M while the addition of albumin to the Sauton medium reversed its MIC90 to 300 μM ( Figure 3A). A similar pattern was observed in the presence of albumin for all compounds tested such as compound 18 ( Figure 3B). were highly bound to plasma proteins. The introduction of a pyrazole ring (compound 18) expectedly decreased lipophilicity but did not reduce plasma protein binding. The antibacterial activity of compounds 1 and 2 reported previously was assessed using green fluorescent protein (GFP)-expressing M. tuberculosis H37Rv grown in Middlebrook 7H9 liquid medium supplemented with Oleic Albumin Dextrose Catalase (OADC). After 5 days of incubation, the GFP signal was measured and showed that both compounds had an equivalent potency to inhibit 90% of the bacterial growth (MIC90 = 300 µ M). The bacterial culture medium used in this previous study contained albumin, which we know now may limit the quantity of unbound compounds 1 and 2. In the present study, the culture medium was therefore switched to Sauton, which does not contain albumin. In these conditions, compound 1 displayed a MIC90 of 100 µ M while the addition of albumin to the Sauton medium reversed its MIC90 to 300 μM ( Figure 3A). A similar pattern was observed in the presence of albumin for all compounds tested such as compound 18 ( Figure 3B). were highly bound to plasma proteins. The introduction of a pyrazole ring (compound 18) expectedly decreased lipophilicity but did not reduce plasma protein binding. The antibacterial activity of compounds 1 and 2 reported previously was assessed using green fluorescent protein (GFP)-expressing M. tuberculosis H37Rv grown in Middlebrook 7H9 liquid medium supplemented with Oleic Albumin Dextrose Catalase (OADC). After 5 days of incubation, the GFP signal was measured and showed that both compounds had an equivalent potency to inhibit 90% of the bacterial growth (MIC90 = 300 µ M). The bacterial culture medium used in this previous study contained albumin, which we know now may limit the quantity of unbound compounds 1 and 2. In the present study, the culture medium was therefore switched to Sauton, which does not contain albumin. In these conditions, compound 1 displayed a MIC90 of 100 µ M while the addition of albumin to the Sauton medium reversed its MIC90 to 300 μM ( Figure 3A). A similar pattern was observed in the presence of albumin for all compounds tested such as compound 18 ( Figure 3B). were highly bound to plasma proteins. The introduction of a pyrazole ring (compound 18) expectedly decreased lipophilicity but did not reduce plasma protein binding. The antibacterial activity of compounds 1 and 2 reported previously was assessed using green fluorescent protein (GFP)-expressing M. tuberculosis H37Rv grown in Middlebrook 7H9 liquid medium supplemented with Oleic Albumin Dextrose Catalase (OADC). After 5 days of incubation, the GFP signal was measured and showed that both compounds had an equivalent potency to inhibit 90% of the bacterial growth (MIC90 = 300 µ M). The bacterial culture medium used in this previous study contained albumin, which we know now may limit the quantity of unbound compounds 1 and 2. In the present study, the culture medium was therefore switched to Sauton, which does not contain albumin. In these conditions, compound 1 displayed a MIC90 of 100 µ M while the addition of albumin to the Sauton medium reversed its MIC90 to 300 μM ( Figure 3A). A similar pattern was observed in the presence of albumin for all compounds tested such as compound 18 ( Figure 3B).  The antimycobacterial activity of our five best compounds was then evaluated on GFP-expressing M. tuberculosis H37Rv grown in Sauton medium (Table 4). All compounds inhibited bacterial growth with MIC90 values between 100 and 300 µ M. However, there was no apparent correlation between enzymatic and antibacterial activities.  The antimycobacterial activity of our five best compounds was then evaluated on GFP-expressing M. tuberculosis H37Rv grown in Sauton medium (Table 4). All compounds inhibited bacterial growth with MIC 90 values between 100 and 300 µM. However, there was no apparent correlation between enzymatic and antibacterial activities.

Analysis of Mycolic Acids Inhibition in M. tuberculosis
To assess whether the studied compounds were able to inhibit the FAS-II system in bacterio, 14 C metabolic labeling of M. tuberculosis H37Rv treated with selected compounds was performed, mycolic acids were isolated, derivatized to corresponding methyl esters and analyzed by TLC (Figure 4). Mycolic acids are synthesized by two fatty acid synthase systems FAS-I and FAS-II. In this assay, M. tuberculosis was grown in Sauton medium without BSA, and radiolabeled acetate was used to produce radiolabeled mycolic acids. The fatty acid profiles of the cells treated with 2% DMSO as a control showed the presence of the three types of mycolic acids (alpha, methoxy and keto), as well as standard fatty acid methyl esters (FAME) coming from FAS-I and used as precursors for the mycolic acid's synthesis. In this assay, a treatment with isoniazid (INH), a FAS-II inhibitor, led to a decrease in mycolic acids synthesis and an accumulation of the FAMEs which are no longer used for the biosynthesis of mycolic acids. The antimycobacterial activity of our five best compounds was then evaluated on GFP-expressing M. tuberculosis H37Rv grown in Sauton medium (Table 4). All compounds inhibited bacterial growth with MIC90 values between 100 and 300 µ M. However, there was no apparent correlation between enzymatic and antibacterial activities.

Analysis of Mycolic Acids Inhibition in M. tuberculosis
To assess whether the studied compounds were able to inhibit the FAS-II system in bacterio, 14 C metabolic labeling of M. tuberculosis H37Rv treated with selected compounds was performed, mycolic acids were isolated, derivatized to corresponding methyl esters and analyzed by TLC (Figure 4). Mycolic acids are synthesized by two fatty acid synthase systems FAS-I and FAS-II. In this assay, M. tuberculosis was grown in Sauton medium without BSA, and radiolabeled acetate was used to produce radiolabeled mycolic acids. The fatty acid profiles of the cells treated with 2% DMSO as a control showed the presence of the three types of mycolic acids (alpha, methoxy and keto), as well as standard fatty acid methyl esters (FAME) coming from FAS-I and used as precursors for the mycolic acid's synthesis. In this assay, a treatment with isoniazid (INH), a FAS-II inhibitor, led to a decrease in mycolic acids synthesis and an accumulation of the FAMEs which are no longer used for the biosynthesis of mycolic acids.  Two MabA inhibitors (compounds 16 and 18) were selected and tested at 100 and 300 µM in this assay. The solubility of these compounds in Sauton medium was measured and they displayed a solubility greater than 300 µM. TLC analysis showed no specific inhibition of mycolic acids in cells treated with these compounds. It revealed a general dose-dependent decrease in all 14

Exploration of the Mechanism of Action of Inhibitors in Bacteria
To evaluate the contribution of MabA inhibition in the mechanism of action of the compounds, MabA (Rv1483) was cloned into pMV261 constitutive overexpression plasmid under the control of the strong hsp60 promoter. The construct was transformed in H37Rv and the overexpression of MabA was verified by Western blot. For an equivalent loading material, as exemplified by the housekeeping protein Hsp65 band, MabA was undetectable in the parental H37Rv, while H37Rv_pMV261-MabA displayed a strong MabA production ( Figure 5, left). No difference in activity for the inhibitors was observed between the parental and mabA overexpressing H37Rv, suggesting that the inhibition of MabA plays a minor role, if any, in the mechanism of action of these inhibitors ( Figure 5, right). and they displayed a solubility greater than 300 µ M. TLC analysis showed no specific inhibition of mycolic acids in cells treated with these compounds. It revealed a general dose-dependent decrease in all 14 C-labeled molecules, correlated to the activity of the compounds measured on M. tuberculosis (compound 16 MIC90 = 100 µ M and compound 18 MIC90 = 300 µ M). This result indicates that compounds 16 and 18 affect other targets than MabA in mycobacterial cells.

Exploration of the Mechanism of Action of Inhibitors in Bacteria
To evaluate the contribution of MabA inhibition in the mechanism of action of the compounds, MabA (Rv1483) was cloned into pMV261 constitutive overexpression plasmid under the control of the strong hsp60 promoter. The construct was transformed in H37Rv and the overexpression of MabA was verified by Western blot. For an equivalent loading material, as exemplified by the housekeeping protein Hsp65 band, MabA was undetectable in the parental H37Rv, while H37Rv_pMV261-MabA displayed a strong MabA production ( Figure 5, left). No difference in activity for the inhibitors was observed between the parental and mabA overexpressing H37Rv, suggesting that the inhibition of MabA plays a minor role, if any, in the mechanism of action of these inhibitors ( Figure 5, right). Several rounds of selection of spontaneous resistance mutants were undertaken at 2× and 3× MIC of compound 16 and this resulted in a bacterial lawn or lack of resistant colonies, respectively. The selection of resistant mutants was also attempted in liquid. Cultures were treated with increasing concentrations of compound 16 in Sauton medium, up to 45 µ M, where growth inhibition was apparent by optical density. After 7 days of treatment, the bacteria were pelleted and plated on 7H11 agar plates carrying 2× MIC and 3× MIC of compound 16. Again, this resulted in the growth of a lawn after 3 months of incubation at 37 °C or in the absence of colonies, respectively.
Transcriptomic profiling can be a useful tool to inform on the mechanism of action of a drug [27]. Therefore, H37Rv was treated with 50 µ M of compound 16 in Sauton medium for 2 h after which the bacteria were pelleted, and their RNA extracted for library preparation and RNA sequencing.
Transcriptional analysis revealed the downregulation of genes implicated in DNA replication, transcription, translation and growth, in line with the growth-inhibitory effect of the compound ( Figure 6). Furthermore, overexpression of drug efflux operons (Rv1687c-1686c, Rv1219c-1216c) was particularly significant suggesting a rapid induction Several rounds of selection of spontaneous resistance mutants were undertaken at 2× and 3× MIC of compound 16 and this resulted in a bacterial lawn or lack of resistant colonies, respectively. The selection of resistant mutants was also attempted in liquid. Cultures were treated with increasing concentrations of compound 16 in Sauton medium, up to 45 µM, where growth inhibition was apparent by optical density. After 7 days of treatment, the bacteria were pelleted and plated on 7H11 agar plates carrying 2× MIC and 3× MIC of compound 16. Again, this resulted in the growth of a lawn after 3 months of incubation at 37 • C or in the absence of colonies, respectively.
Transcriptomic profiling can be a useful tool to inform on the mechanism of action of a drug [27]. Therefore, H37Rv was treated with 50 µM of compound 16 in Sauton medium for 2 h after which the bacteria were pelleted, and their RNA extracted for library preparation and RNA sequencing.
Transcriptional analysis revealed the downregulation of genes implicated in DNA replication, transcription, translation and growth, in line with the growth-inhibitory effect of the compound ( Figure 6). Furthermore, overexpression of drug efflux operons (Rv1687c-1686c, Rv1219c-1216c) was particularly significant suggesting a rapid induction of export of the compound. Lastly, Rv0560c was strongly upregulated. Rv0560c is described as an S-adenosylmethionine-dependent methyltransferase that was shown to be induced by salicylate stress. In fact, our transcriptional results agreed well with described gene expression profile observed after M. tuberculosis treatment with salicylate [28].
The anthranilic acid derivatives are close analogs of salicylic acid and transcriptional profiling reveals that this feature plays a major role in the bacterial response to compound 16 treatment. M. tuberculosis is unusually sensitive to weak acids such as salicylic acid. Weak acids display higher potency in decreasing pH conditions, and this was correlated to their ability to act as protonophores, inducing acidification of intrabacterial pH and decreased membrane potential [29]. To determine whether this is a trait of our MabA inhibitors, the activity of compound 16 was tested under a range of pH conditions on H37Rv (Figure 7). of export of the compound. Lastly, Rv0560c was strongly upregulated. Rv0560c is described as an S-adenosylmethionine-dependent methyltransferase that was shown to be induced by salicylate stress. In fact, our transcriptional results agreed well with described gene expression profile observed after M. tuberculosis treatment with salicylate [28]. The anthranilic acid derivatives are close analogs of salicylic acid and transcriptional profiling reveals that this feature plays a major role in the bacterial response to compound 16 treatment. M. tuberculosis is unusually sensitive to weak acids such as salicylic acid. Weak acids display higher potency in decreasing pH conditions, and this was correlated to their ability to act as protonophores, inducing acidification of intrabacterial pH and decreased membrane potential [29]. To determine whether this is a trait of our MabA inhibitors, the activity of compound 16 was tested under a range of pH conditions on H37Rv (Figure 7). Indeed, decreasing pH improved the potency of compound 16 ( Figure 7A,B). Moreover, we used an H37Rv expressing a ratiometric pH-sensitive GFP to measure the effect of compound 16 treatment in the maintenance of the pH homeostasis [30]. Consistently, we observed a dose-dependent decrease in the intrabacterial pH in the presence of compound 16 ( Figure 7C). These observations are consistent with our inability to select resistance mutants to compound 16 as the peculiar mechanism of action of these weak acids does not rely on a single target, but rather on a general effect on membrane potential.
Overall, these results suggest that the anthranilic acid core of MabA inhibitors is the main driver of bacterial response, inducing intrabacterial acidification and arrest of replication and fatty acid synthesis [28]. These effects are shared with pyrazinamide (PZA), a prodrug bioactivated by the enzyme PncA into pyrazinoic acid. This compound, which is  The anthranilic acid derivatives are close analogs of salicylic acid and transcriptional profiling reveals that this feature plays a major role in the bacterial response to compound 16 treatment. M. tuberculosis is unusually sensitive to weak acids such as salicylic acid. Weak acids display higher potency in decreasing pH conditions, and this was correlated to their ability to act as protonophores, inducing acidification of intrabacterial pH and decreased membrane potential [29]. To determine whether this is a trait of our MabA inhibitors, the activity of compound 16 was tested under a range of pH conditions on H37Rv (Figure 7). Indeed, decreasing pH improved the potency of compound 16 ( Figure 7A,B). Moreover, we used an H37Rv expressing a ratiometric pH-sensitive GFP to measure the effect of compound 16 treatment in the maintenance of the pH homeostasis [30]. Consistently, we observed a dose-dependent decrease in the intrabacterial pH in the presence of compound 16 ( Figure 7C). These observations are consistent with our inability to select resistance mutants to compound 16 as the peculiar mechanism of action of these weak acids does not rely on a single target, but rather on a general effect on membrane potential.
Overall, these results suggest that the anthranilic acid core of MabA inhibitors is the main driver of bacterial response, inducing intrabacterial acidification and arrest of replication and fatty acid synthesis [28]. These effects are shared with pyrazinamide (PZA), a prodrug bioactivated by the enzyme PncA into pyrazinoic acid. This compound, which is Indeed, decreasing pH improved the potency of compound 16 ( Figure 7A,B). Moreover, we used an H37Rv expressing a ratiometric pH-sensitive GFP to measure the effect of compound 16 treatment in the maintenance of the pH homeostasis [30]. Consistently, we observed a dose-dependent decrease in the intrabacterial pH in the presence of compound 16 ( Figure 7C). These observations are consistent with our inability to select resistance mutants to compound 16 as the peculiar mechanism of action of these weak acids does not rely on a single target, but rather on a general effect on membrane potential.
Overall, these results suggest that the anthranilic acid core of MabA inhibitors is the main driver of bacterial response, inducing intrabacterial acidification and arrest of replication and fatty acid synthesis [28]. These effects are shared with pyrazinamide (PZA), a prodrug bioactivated by the enzyme PncA into pyrazinoic acid. This compound, which is also poorly active in neutral pH conditions, is nevertheless a cornerstone drug in current TB regimens.

Chemistry
The solvents used for synthesis, analysis and purification were obtained as analytical grade from commercial suppliers and used without further purification. Chemical reagents were obtained from Fisher Scientific, Merck, Fluorochem, Enamine or TCI as reagent grade and used without further purification.
The LC-MS Waters system used included a 2747 sample manager, a 2695 separations module, a 2996 photodiode array detector (200-400 nm), and a Waters Micromass ZQ2000 detector (scan 100-800). The XBridge C18 column (50 mm × 4.6 mm, 3.5 µm, Waters) was used with an injection volume of 20 µL and a mobile phase mixture of water and acetonitrile in gradient-elution. The pH of the mobile phase was adjusted to 3.8 with HCOOH and NH 4 OH to form a buffer solution. Analysis time was 5 min (at a flow rate of 2 mL/min), 10 min (at a flow rate of 1 mL/min) or 30 min (at a flow rate of 1 mL/min). Purity was determined by reversed-phase HPLC with UV detection (215 nm) and all isolated compounds had a purity greater than 95%.
HRMS analysis was conducted on a LC-MS system utilizing a LCT Premier XE mass spectrometer from Waters and a XBridge C18 column (50 mm × 4.6 mm, 3.5 µm, Waters). The mobile phase gradient started with 98% H 2 O 5 mM Ammonium Formate pH 3.8 and reached 100% CH 3 CN 5 mM Ammonium Formate pH 3.8 within 3 min at a flow rate of 1 mL/min. NMR spectra were recorded on a Bruker DRX-300 spectrometer using the solvent as an internal reference [e.g., 2.50 (residual DMSO-d 6 ) and 39.52 (DMSO-d 6 ) ppm for 1 H and 13 C NMR spectra, respectively]. Chemical shifts (δ) were in parts per million (ppm) and assignments were made using one-dimensional (1D) 1 H and 13 C spectra and two-dimensional (2D) HSQC-DEPT, COSY, and HMBC spectra. NMR coupling constants (J) were reported in Hertz (Hz) and splitting patterns were indicated as follows: m for multiplet, s for singlet, brs for broad singlet, d for doublet, t for triplet, q for quartet, dd for doublet of doublet, ddd for doublet of doublet of doublet.
Reverse flash chromatography was performed using a CombiFlash ® C18 Rf200, which was equipped with C 18 silica gel cartridges and UV detection (215 and 254 nm) was used to isolate the desired product.
Preparative HPLC was performed using a Varian PRoStar system, which was equipped with an Omni-Sphere 10 µm column C18 Dynamax (250 mm × 41.4 mm) from Agilent Technologies. A gradient starting from 20% MeCN/80% H 2 O/0.1% formic acid and reaching 100% MeCN/0.1% formic acid at a flow rate of 80 mL/min was used with UV detection (215 and 254 nm) to isolate the desired product.
The synthesis process utilizing microwave irradiation was carried out using Biotage ® Initiator+.

Synthesis of 2-[(3,4-dichlorobenzoyl)-methyl-amino]-5-iodo-benzoic Acid (3)
To a solution of 2-[(3,4-dichlorobenzoyl)amino]-5-iodo-benzoic acid (1, 0.05 mmol) in a mixture of anhydrous THF and DMF (2 mL, v/v, 3:1) was added cesium carbonate (0.19 mmol) and iodomethane (0.16 mmol). The mixture was stirred at RT overnight. Lithium hydroxide (0.15 mmol) in water (2 mL) was then added and the reaction was stirred at RT for 5 h. The solvent was removed under reduced pressure and to the white residue was added HCl 1 N (5 mL). The obtained precipitated was filtered to afford the desired product as a white powder. Yield = 88%. In a round bottom flask charged with 2-amino-5-iodo-benzoic acid (0.5 mmol) and Na 2 CO 3 (0.5 mmol) dissolved in water (1.5 mL) was slowly added 3,4-dichlorobenzenesulfonyl chloride (0.5 mmol). During the addition of sulfonyl chloride and the progression of reaction the pH was maintained at a value around 8 (sodium carbonate was added portion-wise if necessary). The resulting mixture was stirred at RT for 30 min. The product precipitated as a yellow solid which was filtered and washed with water, cyclohexane and DCM to afford the desired product as a white powder. Yield = 45%. In the round bottom flask, methyl 2-amino-5-bromo-benzoate (2 mmol) and 1 mL of pyridine were added. 3,4-dichlorobenzoyl chloride (2 mmol) dissolved in 2 mL of pyridine was added dropwise to the solution. The mixture was stirred at RT for 1 h. The solvent was removed under reduced pressure, EtOAc was added and a precipitated was formed. It was filtrated to afford 744 mg of white powder leading to a 92% yield.  (7) 3,4-dichloro-N -hydroxy-benzamidine (Intermediate 7a) In a 25 mL round bottom flask, 3,4-dichlorobenzonitrile (3 mmol, 516 mg, 1 eq), hydroxyammonium chloride (4.5 mmol, 312 mg, 1.5 eq) and DIEA (6.4 mmol, 1.115 mL, 1.6 eq) dissolved in 4 mL of EtOH were added. The mixture was stirred under reflux at of AcOEt was added. The reaction was allowed to reach RT and stirred overnight. The precipitate of DCU was filtered off and washed with EtOAc. The filtrate was washed successively with saturated aqueous sodium hydrogen carbonate and brine, dried over MgSO 4 , and evaporated under reduced pressure to give the desired product as a white powder. Yield = 64%. (8) To a stirred solution of (2,5-dioxopyrrolidin-1-yl)-2-[(3,4-dichlorobenzoyl)amino]-5iodo-benzoate (0.22 mmol) dissolved in THF (0.5 mL) a solution of NaBH 4 (0.69 mmol) in THF (1.5 mmol) and water (0.2 mL) was added dropwise. The reaction was stirred for 1.5 h at RT. Saturated aqueous ammonium chloride (10 mL) was added to quench the reaction. The aqueous layer was then extracted twice with EtOAc. Then, the organic layer was washed with brine, dried over MgSO 4 , and evaporated under reduced pressure. Concentrated sulfuric acid (2 mL) was carefully added dropwise at 0 • C to 2-amino-5-iodobenzonitrile (1 mmol). After completion of the addition, the reaction mixture was allowed to warm from 0 • C to RT and stirred for 72 h. The reaction mixture was poured onto ice and then brought to a pH of 9−10 by addition of concentrated aqueous ammonium hydroxide solution. The resulting solid was collected by filtration, washed three times with water (3 × 1 mL), and three times with a 1:1 mixture of diethyl ether cyclohexane (3 × 1 mL) to afford a pale brown solid which was used in the next step without further purification. Yield = 83%. To a solution of 2-amino-5-iodo-benzonitrile (1 mmol) in pyridine (3 mL), was added portion wise 3,4-dichlorobenzoyl chloride (1.5 mmol) and the reaction mixture was stirred at RT for 1 h. The solvent was removed under reduced pressure, water (5 mL) and EtOAc (5 mL) were added and the mixture was stirred at RT for 15 min. The solid formed was filtered, washed with water, HCl, MeOH, AcOEt and acetonitrile, and then dried in a desiccator. Yield = 64%.  13 (11) To a solution of 3,4-dichloro-N-(2-cyano-4-iodo-phenyl)benzamide (10, 0.24 mmol) dissolved in DMF (0.5 mL), zinc bromide (0.24 mmol) and sodium azide (0.24 mmol) were added. The resulting mixture was heated at 130 • C for 1h30. Supplementary sodium azide (7.8 mg, 0.12 mmol, 0.5 eq) was then added and the mixture was stirred for 30 min. The mixture was cooled to RT and a few drops of NaOH in water (50% w/v) were added. The resulting emulsion was filtered and HCl 1 N was added to the solution until a pH of around 4 was reached. The formed solid was filtered, washed with water and dried in a desiccator to obtain a pale-yellow powder as the desired product. Yield = 86%.  (12) A solution of 2-[(3,4-dichlorobenzoyl)amino]-5-iodo-benzoic acid (1, 0.25 mmol) and CDI (0.5 mmol) was prepared in dry THF (4 mL) and refluxed under argon for 1 h. Trifluoromethanesulfonamide (0.5 mmol) and DBU (0.75 mmol) were dissolved in anhydrous THF (0.5 mL) and the solution was added to the initial reaction mixture. The reaction was then stirred at RT for 2 h. The solvent was removed under reduced pressure and DCM (4 mL) was added. The organic layer was washed with 5% citric acid aqueous solution and brine, dried over MgSO 4 , and then the solvent was removed under reduced pressure. The crude was purified by flash chromatography (from DCM/MeOH 100/0 to 90/10) to afford the desired product as a white power. Yield = 52%.  (13) A solution of 2-[(3,4-dichlorobenzoyl)amino]-5-iodo-benzoic acid (1, 0.25 mmol) and CDI (0.5 mmol) was prepared in dry THF (4 mL) and refluxed under argon for 1 h. Methanesulfonamide (0.5 mmol) and DBU (0.75 mmol) were dissolved in anhydrous THF (0.5 mL) and the solution was added to the initial reaction mixture. The reaction was then stirred at RT for 2 h. THF was then removed under reduced pressure and DCM (4 mL) was added. The organic layer was washed with 5% citric acid and brine, dried over MgSO 4 , and then the solvent was evaporated under reduced pressure. The crude was purified by flash chromatography (from DCM/MeOH 100/0 to 90/10). The product precipitated into collected fractions and the solid was filtered and dried in a desiccator to give the desired product as a white powder. Yield = 41%. In a round bottom flask charged with 2-[(3,4-dichlorobenzoyl)amino]-5-iodo-benzoic acid (1, 0.57 mmol), Pd(PPh 3 ) 2 Cl 2 (0.034 mmol), tetrabutylammonium iodide (0.86 mmol) and copper iodide (0.1 mmol) dissolved in anhydrous acetonitrile (6 mL) was added dropwise TEA (9.7 mmol). The resulting solution was stirred at RT for 10 min, and then ethynylcyclopropane (1.26 mmol) was added dropwise. The reaction mixture was stirred at RT for 1h30, and then the solution was heated at 60 • C for 30 min. The solvent was removed under reduced pressure, the residue was dissolved in EtOAc (20 mL) and the organic layer was washed with an aqueous solution of HCl 1 N, water, brine, dried over MgSO 4 , filtered and the solvent was removed under reduced pressure to afford a brown oil. The crude was purified by reverse flash chromatography to afford a dark beige powder. DCM was added, a precipitate was formed and filtrated to afford the desired product as a white powder. Yield = 33%.  13  In a round bottom flask, 2-amino-4-bromo-benzoic acid (1.15 mmol), triphenyl phosphine (0.11 mmol), copper iodide (0.11 mmol), bis(triphenylphosphine)palladium(II) chloride (0.11 mmol) and anhydrous THF (7.5 mL) were added. The round bottom flask was purged with argon for 30 min, and then TEA (54 mmol) and ethynyl(trimethyl)silane (6.9 mmol) were added dropwise to the solution. The mixture was stirred for 2 h at 80 • C. The resulting mixture was filtered, and the filtrate was evaporated under reduced pressure. with the DMSO volume being adjusted to keep the concentration across all wells at 1% (200 nL). A reference inhibitor INH-NADP [31,32] was used (100% inhibition at 200 µM) and prepared as described in the literature [32]. We added manually, using a 16-channel VIAFLO II electronic pipette (Integra Biosciences), 3.8 µL of Hepes buffer (100 mM pH 7) and 6 µL of MabA (1.33 µM in Hepes buffer). Compounds were incubated with the MabA enzyme for 30 min at 22 • C. We then added manually 10 µL of a mixture of the substrate AcAcCoA and the cofactor NADPH (100 µM in Hepes buffer) to start the enzymatic reaction. The final concentrations were as follows: MabA at 400 nM, AcAcCoA at 50 µM, NADPH at 50 µM. After 15 min of incubation at 22 • C we stopped the enzymatic reaction by adding manually 10 µL of a solution of trifluoroacetic acid in water at a final concentration of 1%, containing BCoA at a final concentration of 5 µM. BCoA was used as an internal standard for MS/MS analysis. The inhibition percentage of each compound at various concentrations was determined by averaging the results of negative controls (1% DMSO) and positive controls (INH-NADP adduct at 200 µM). The data was then analyzed using GraphPad Prism software and a nonlinear regression method to create dose-response curves. The IC 50 values are the average of at least two experiments.

LC-MS/MS Analysis
AcAcCoA, HBCoA and BCoA were separated using an Acquity UPLC BEH C18 column (2.1 × 50 mm, 1.7 µm, Waters) and detected by multiple-reaction monitoring (MRM) [20]. We used a UPLC Acquity I-class coupled with a Xevo TQD mass spectrometer (Waters). The mobile phases A and B were composed of ammonium acetate 10 mM in water and MeOH, respectively. The mass spectrometer was set to specific parameters (polarity ES+, capillary 1200 V, desolvation temperature 600 • C, cone voltage 46 V, source temperature 150 • C, cone gaz flow 50 L/h, desolvation gaz flow 1200 L/h) and the specific transitions being monitored were AcAcCoA 852.1-342.5 (collision energy: 38 eV), HBCoA 854.1-347.2 (collision energy: 36 eV), BCoA 838.1-331.2 (collision energy: 34 eV). Data was acquired and processed using MassLynx and Target-Lynx software (Waters, France). The amount of HBCoA in each well was determined by dividing the peak area of HBCoA by the peak area of BCoA which was used as an internal standard.

Ligand-Observed NMR Experiments
19 F NMR measurements were performed at 298 K using 3 mm tubes containing 160 µL of samples on a Bruker 600 MHz Avance III HD spectrometer equipped with a CP-QCI-F cryoprobe specifically designed for 19 F detection. Both the 19 F and 1 H NMR reference spectra were acquired on samples containing compound 12 at a concentration of 100 µM, in an NMR buffer containing 50 mM Tris-Cl pH 7.5, 250 mM NaCl, 1% DMSO-d 6 , and 5% D 2 O. Then, 19 F and 1 H NMR spectra of the compounds were obtained in the presence of MabA at a concentration of 20 µM in the same NMR buffer. The parameters used for the 1 H experiments were TD = 16,384 points, NS = 64 scans, relaxation delay D1 = 1 s, carrier frequency O1P = 4.698 ppm, spectral window sw = 16 ppm, and with 19 F decoupling. The acquisition time was about 2.5 min per sample. The parameters used for the 19 F experiments were TD = 8192 points, NS = 512 scans, relaxation delay D1 = 2 s, carrier frequency O1P = −65.0 ppm, spectral window sw = 20 ppm, and with 1 H decoupling. The acquisition time was about 22 min per sample. The NMR data were processed using Bruker Topspin 4.06 software.

MIC Determination
Compounds were dissolved in DMSO at a concentration of 100 mM and transferred to a 384-well low-dead-volume polypropylene plate (LP-0200, labcyte), which was used to prepare assay plates. These compounds were then tested at concentrations ranging from A total of 5 µL of a 10 mM solution of the compound in DMSO was diluted in 245 µL of a 1/1 octanol/PBS mixture at pH 7.4. The mixture was gently shaken for 2 h at room temperature. A total of 10 µL of each phase were diluted in 490 µL of MeOH and analyzed by LC-MS/MS. Each compound was tested in triplicate. Log D was determined as the logarithm of the ratio of the concentration of product in octanol and PBS, determined by mass signals. The test was validated if:

Plasma Protein Binding
As it was anticipated that the compounds were likely to be highly bound, a 10% plasma (plasma beforehand diluted in a 1 to 10 ratio in a phosphate buffer pH 7.2) was used for this experiment. The 10% plasma, spiked with the tested compound at 10 µM final (0.1% DMSO), was added to the red-ring chamber of a RED (Rapid Equilibrium Dialysis) device. Blank phosphate buffer pH 7.2 was added to the outer chamber of the RED device and the plate was placed at 37 • C with shaking. After 4 h of incubation, aliquots of the buffer and plasma chambers were removed and the concentration in each compartment was determined by LC-MS/MS analysis.
The calculation of the unbound fraction (fu) is performed according to the formula: The following equation is used to convert from the fraction unbound at 10% to a fraction unbound at 100%: f u 100% = f u 10% /(10 − 9f u 10% ) The percentage of recovery was calculated according to: where A is the area corresponding to the compound and V, the volume of solution presents in each compartment (i.e., V PBS = 350 µL and V Plasma = 200 µL).

Conclusions
Our work led to the discovery and optimization of the first MabA inhibitors. Twenty compounds sharing an anthranilic acid scaffold were synthesized and tested as MabA inhibitors to delineate structure-activity relationships. These SARs showed that the replacement of the amide function with a sulfonamide or a thioamide function was tolerated and that an acidic function was necessary for the activity. Ligand-observed NMR experiments confirmed the direct binding of a fluorinated anthranilic acid to MabA. The antimycobacterial activity of 5 MabA inhibitors, selected based on their potency and chemical diversity, was evaluated on M. tuberculosis H37Rv grown in Sauton medium and although all the compounds were able to inhibit bacterial growth, there was no apparent correlation between enzymatic and antibacterial activities.
Metabolic labeling of M. tuberculosis H37Rv grown in the presence of two studied compounds did not reveal any changes in mycolic acids profiles suggesting that these inhibitors do not affect the FAS-II system. Further investigation of the mechanism of action of these compounds was carried out by overexpressing MabA, attempting to select resistance mutants and using transcriptional analysis. These experiments led us to conclude that, although anthranilic acid derivatives inhibit MabA in vitro, the activity observed in bacteria was partly due to the presence of the carboxylic acid moiety, which induces internal pH changes. These effects are shared with pyrazinamide, which historically allowed to reduce TB treatment duration from 9 to 6 months. This major outcome is attributed to the ability of pyrazinamide to target intracellular bacilli in acidified phagosomes. Therefore, compound 16 or its derivatives could be tested in combination with anti-TB drugs to assess whether a synergistic effect is observed as with pyrazinamide.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ph16030335/s1, Figure S1: 1D 19 F NMR spectra of 4 compounds alone and in the presence of MabA; Figure S2: Western blot showing overexpression of MabA in H37Rv_pMV261-MabA, as compared to the parental control; Figures S3-S40: 1 H NMR, 13 C NMR and HRMS data of compounds 2-20.