Design, Synthesis and Biological Evaluation of N-phenylindole Derivatives as Pks13 Inhibitors againstMycobacterium tuberculosis

Polyketide synthase 13 (Pks13), an essential enzyme for the survival of Mycobacterium tuberculosis (Mtb), is an attractive target for new anti-TB agents. In our previous work, we have identified 2-phenylindole derivatives against Mtb. The crystallography studies demonstrated that the two-position phenol was solvent-exposed in the Pks13-TE crystal structure and a crucial hydrogen bond was lost while introducing bulkier hydrophobic groups at indole N moieties. Thirty-six N-phenylindole derivatives were synthesized and evaluated for antitubercular activity using a structure-guided approach. The structure–activity relationship (SAR) studies resulted in the discovery of the potent Compounds 45 and 58 against Mtb H37Rv, with an MIC value of 0.0625 μg/mL and 0.125 μg/mL, respectively. The thermal stability analysis showed that they bind with high affinity to the Pks13-TE domain. Preliminary ADME evaluation showed that Compound 58 displayed modest human microsomal stability. This report further validates that targeting Pks13 is a valid strategy for the inhibition of Mtb and provides a novel scaffold for developing leading anti-TB compounds.


Introduction
Tuberculosis (TB) is a chronic infectious disease caused by Mtb, the second leading infectious killer after COVID-19. According to the 2021 World Health Organization (WHO) report, there were 1.5 million people who died from TB worldwide in 2020, including over 0.2 million co-infected with human immunodeficiency viruses (HIV) [1]. Multidrug-resistant (MDR) and extensively drug-resistant (XDR)-TB remain a public health crisis globally. In more than half a century, only three new anti-TB drugs, bedaquiline (1, Figure 1) [2], delamanid (2) [3], and pretomanid (3) [4], have entered the market. Recently, an inspiring outcome showed that a novel regimen of bedaquiline, pretomanid, and linezolid for treating highly drug-resistant pulmonary TB cured 90% of patients and cut the treatment period to 6 months [5]. Therefore, it is an urgent development of novel anti-TB drugs. Blocking the biosynthesis pathway of the mycobacterial cell wall is an effective strategy for anti-TB. In Mtb, mycolic acids are C 60 -C 90 long-chain fatty acids used to structure the unique cell wall, crucial for its persistence and pathogenesis [6]. Pks13 has been proved as the key enzyme that catalyzes the final step of mycolic acid synthesis [7]. Pks13, belonging to the type-I PKS family, comprises 1733 amino acid residues encoded by the Pks13 gene [8]. Its topological structure is ACP (N-terminus acyl carrier protein), KS (ketoacyl synthase), AT (acyl transferase), ACP (C-terminus acyl carrier protein), and TE (thioesterase). The TE domain of Pks13 firstly exerts hydrolase activity, and ester bonds form after hydrolysis thioester between the mycolic β-ketoester and the hydroxyl group of Ser1533 [9]. Then, the domain acts as an acyltransferase to transfer mycolic β-ketoester onto trehalose to form the trehalose monomycolate precursor. Therefore, inhibiting Pks13 disturbs the biosynthesis of mycolic acids and kills Mtb. In recent years, as an emerging and attractive target, its inhibitors have been reported in succession: thiophene-based Pks13 inhibitors targeting the ACP domain (4) [10], benzofuran derivatives inhibitors Pks13-TE (5, 6) [11], β-lactonebased compounds [12], and 4H-chromen-4-one derivatives [13], and our group reported conformational restricted tetracyclic compounds (7,8) [14][15][16].
Indole represents a privileged scaffold in various marine or terrestrial natural products with pharmacological and medical potential for developing novel and effective medications [17]. Indole derivatives have attracted considerable interest in medicinal scientists due to their broad, interesting bioactivities, including being anticancer [18], anti-inflammatory [19], antiviral [20], antimicrobial [21], and antitubercular [22][23][24], etc. In our previous work [14], adopting a scaffold hopping strategy, we have replaced the benzofuran core with the indole to identify novel anti-TB compounds. Unfortunately, 2-phenylindole-based derivatives ( Figure 2) were deleterious for the activity compared with the corresponding benzofuran derivatives. The co-crystal structure between ligand TAM16 and the Pks13-thioesterase (TE) domain have been reported previously [11]. Similar to the binding mode of TAM16 reported previously ( Figure 3A), the indole derivative 18 formed various hydrogen bonds interactions with key residues D1644 and N1640. The van der Waals and stacking interactions were observed between the piperidine ring and the side chain of the residue Y1674. However, a crucial hydrogen bond was lost between the hydroxyl group of the para position of 2-phenyl and Q1633 residue ( Figure 3B). Herein, based on the structure-guided strategy, our continuous efforts developed N-phenylindole-based derivatives and evaluated their activity against Mtb.

Chemistry
The compounds were synthesized in three to five steps by utilizing the synthetic routes shown in Scheme 1. Ethyl (Z)-3-amino-3-(4-methoxyphenyl) acrylate (11) was obtained from ethyl 3-(4-methoxyphenyl)-3-oxopropanoate (10) with ammonium formate in refluxing ethanol. The indole derivative (12) was formed from Compound 11 through a Nenitzescu reaction with 1,4-benzoquinone catalyzed by ZnBr2 and the subsequent Mannich reaction with 37% aqueous formaldehyde and piperidine to yield Compound 13 [25,26]. Compound 14 was obtained from 13 via demethylation by boron tribromide. Similarly, Compounds 15-20 were prepared in similar methods described above. The key intermediate (21) was synthesized following similar procedures as Compound 12. The hydrolysis of Compound 21 was subsequently subjected to amide coupling and the Mannich reaction to afford Compounds 22 and 23, and then demethylation by boron tribromide to give Compounds 24 and 25.
The synthesis route is shown in Scheme 2. Starting Material 26 with aniline and 1,4benzoquinone three-component catalyzed by montmorillonite one-pot directly synthesized the 5-hydroxy indole derivative (27) [27] and the subsequent via formaldehyde and piperidine to give Compound 28. Compounds 29-39 were prepared in a similar manner to Compound 28.

Chemistry
The compounds were synthesized in three to five steps by utilizing the s routes shown in Scheme 1. Ethyl (Z)-3-amino-3-(4-methoxyphenyl) acrylate (11) tained from ethyl 3-(4-methoxyphenyl)-3-oxopropanoate (10) with ammonium for refluxing ethanol. The indole derivative (12) was formed from Compound 11 th Nenitzescu reaction with 1,4-benzoquinone catalyzed by ZnBr2 and the subseque nich reaction with 37% aqueous formaldehyde and piperidine to yield Compo [25,26]. Compound 14 was obtained from 13 via demethylation by boron tribromi ilarly, Compounds 15-20 were prepared in similar methods described above. The termediate (21) was synthesized following similar procedures as Compound 12. drolysis of Compound 21 was subsequently subjected to amide coupling and the M reaction to afford Compounds 22 and 23, and then demethylation by boron tribro give Compounds 24 and 25.
The synthesis route is shown in Scheme 2. Starting Material 26 with aniline benzoquinone three-component catalyzed by montmorillonite one-pot directly sized the 5-hydroxy indole derivative (27) [27] and the subsequent via formaldeh piperidine to give Compound 28. Compounds 29-39 were prepared in a similar to Compound 28.
The synthesis of Compounds 45 and 48-58 was shown in Scheme 3. Pent dione (40) and 4-bromoaniline (41) were refluxed in EtOH for 8 h to give Compo The indole derivative (43) was formed from Compound 42 through the Nenitzes tion, then, Suzuki coupling afforded 44, which was subjected to the Mannich rea give the desired Compound 45. Following a similar procedure, Compound 48 was from 4-(piperidin-1-yl) aniline (47), which was prepared via sequential nucleoph stitution and reduction using piperidine and Pd/C, respectively, from 1-bromo

Chemistry
The compounds were synthesized in three to five steps by utilizing the synthetic routes shown in Scheme 1. Ethyl (Z)-3-amino-3-(4-methoxyphenyl) acrylate (11) was obtained from ethyl 3-(4-methoxyphenyl)-3-oxopropanoate (10) with ammonium formate in refluxing ethanol. The indole derivative (12) was formed from Compound 11 through a Nenitzescu reaction with 1,4-benzoquinone catalyzed by ZnBr 2 and the subsequent Mannich reaction with 37% aqueous formaldehyde and piperidine to yield Compound 13 [25,26]. Compound 14 was obtained from 13 via demethylation by boron tribromide. Similarly, Compounds 15-20 were prepared in similar methods described above. The key intermediate (21) was synthesized following similar procedures as Compound 12. The hydrolysis of Compound 21 was subsequently subjected to amide coupling and the Mannich reaction to afford Compounds 22 and 23, and then demethylation by boron tribromide to give Compounds 24 and 25.

Results and Discussion
All final compounds for anti-TB activity were initially evaluated for their minimal concentration of the 90% growth inhibitory against the Mtb H37Rv strain as their MIC values in a microplate alamar blue assay (MABA) [23]. Firstly, we explored substitutions at the nitrogen of indole with H, methyl, isopropyl, and phenyl, which resulted in Compounds 14-16 and 18, with MIC values from 32 µ g/mL to 2 µ g/mL (Table 1). Surprisingly a bulkier aromatic derivative (18) (MIC = 2 µ g/mL) was favorable in activity against Mtb Next, the ester group of 3-position on indole, which may be prone to metabolic liability was assessed for the effects of replacement with acetyl and amides to give Compounds 20, 24, and 25. Compared with Compound 18 (MIC = 2 µ g/mL), the replacement with a methyl amide (24, MIC = 4 µ g/mL) or ethyl amide (25, MIC = 8 µ g/mL) at R 3 resulted in a slight drop in activity against Mtb, whereas acetyl substituent (20) had an eight-fold decrease in activity. The antitubercular activity of methoxyl substitution on the two-position phenyl derivatives slightly decreased when compared to the corresponding hydroxyl substitution derivatives, e.g., 13 (MIC = 32 µ g/mL) vs. 14 (MIC = 16 µ g/mL), 17 (MIC = 4 µ g/mL) vs. Subsequently, we investigated 2-methyl-N-phenylindole derivatives. As shown in Table 2, the deletion of the two-position benzene ring resulted in Compound 28, with an MIC value of 8 µ g/mL. Among the series, the amide of three-position on indole resulted in Compounds 29 (MIC > 64 µ g/mL) and 30 (MIC = 16 µ g/mL), which were not favorable for activity. The N-phenyl of the indole was substituted with different groups, such as methoxyl, methyl, fluorine, N, N-dimethyl, and t-butyl to form Compounds 31-39, 49, and 57, with MIC values ranging from 0.5 to 8 µ g/mL. However, whether the substituent groups were electron-withdrawing or electron-donating, the para-position substituents of

Results and Discussion
All final compounds for anti-TB activity were initially evaluated for their minimal concentration of the 90% growth inhibitory against the Mtb H37Rv strain as their MIC values in a microplate alamar blue assay (MABA) [23]. Firstly, we explored substitutions at the nitrogen of indole with H, methyl, isopropyl, and phenyl, which resulted in Compounds 14-16 and 18, with MIC values from 32 µg/mL to 2 µg/mL (Table 1). Surprisingly, a bulkier aromatic derivative (18) (MIC = 2 µg/mL) was favorable in activity against Mtb. Next, the ester group of 3-position on indole, which may be prone to metabolic liability, was assessed for the effects of replacement with acetyl and amides to give Compounds 20, 24, and 25.  Therefore, the loss of hydrogen bonds with Gln1633 seems to have little effect on tivity. Next, different Mannich substructures with N, N-dimethyl (51), pyrrolidin and 56), and 4-methylpiperidyl (53) were placed at the four-position of the indole pounds. The substituent by the piperidine was the most potent, such as 55 (MIC = µ g/mL) vs. 56 (MIC = 0.25 µ g/mL), 50 (MIC = 0.5 µ g/mL) vs. 51(MIC = 2 µ g/mL), a (MIC = 1 µ g/mL), which were consistent with the SAR trend that we reported prev [11].   [14].
Subsequently, we investigated 2-methyl-N-phenylindole derivatives. As shown in Table 2, the deletion of the two-position benzene ring resulted in Compound 28, with an MIC value of 8 µg/mL. Among the series, the amide of three-position on indole resulted in Compounds 29 (MIC > 64 µg/mL) and 30 (MIC = 16 µg/mL), which were not favorable for activity. The N-phenyl of the indole was substituted with different groups, such as methoxyl, methyl, fluorine, N, N-dimethyl, and t-butyl to form Compounds 31-39, 49, and 57, with MIC values ranging from 0.5 to 8 µg/mL. However, whether the substituent groups were electron-withdrawing or electron-donating, the para-position substituents of N-phenylindole were optimal. The activity of acetyl group substitution on the three-positions derivatives had a slight increase when compared to the corresponding ester group substitutions, e.g., 50 (MIC = 0.5 µg/mL) vs. 49 (MIC = 1 µg/mL) and 55 (MIC = 0.125 µg/mL) vs. 57 (MIC = 0.5 µg/mL). We kept investigating the effect of the para-position substituent at the N-phenyl of the indole. Gratifyingly, introducing hydrophobic groups (phenyl, piperidyl, i-propyl, t-butyl, and Br) resulted in compounds 45, 48, 54, 55, and 58 having considerable anti-TB activity against the Mtb strain, with MIC values of 0.125-0.0625 µg/mL. Compound 45 was the most potent, with an MIC value of 0.0625 µg/mL. Therefore, the loss of hydrogen bonds with Gln1633 seems to have little effect on the activity. Next, different Mannich substructures with N, Ndimethyl (51), pyrrolidinyl (52 and 56), and 4-methylpiperidyl (53) were placed at the four-position of the indole compounds. The substituent by the piperidine was the most potent, such as 55 (MIC = 0.125 µg/mL) vs. 56 (MIC = 0.25 µg/mL), 50 (MIC = 0.5 µg/mL) vs. 51(MIC = 2 µg/mL), and 53 (MIC = 1 µg/mL), which were consistent with the SAR trend that we reported previously [11].  a The lowest concentration of compounds leading to at least 90% inhibition of bacterial growth by the MABA. MIC values are reported as an average of three individual measurements. b was calculated using ChemBioDraw Ultra 16.0. c Compound 9 was a reference compound fo parison [14]. To demonstrate whether the indole compounds were binding to the Pks13-TE protein, selected representative compounds were evaluated for thermal stability in the presence of Pks13-TE by utilizing the nano-differential scanning fluorimetry (nano DSF) method [14]. The stabilization of the Pks13-TE protein at a 30 µM concentration upon binding of highaffinity ligands (at 150 µM and 300 µM concentration, respectively) was evaluated, while the T m value of apo-Pks13-TE was 56.2 ± 0.03 • C. TAM16 and Compound 7 were used as a positive control. As shown in Table 3, it was observed that following the addition of the five-fold or 10-fold compounds (∆T m > 4.5 • C), there was a significant increase in the thermal stability of Pks13-TE, which indicated a high-affinity binding of the compounds to the Pks13-TE. Meanwhile, we observed that the extent of the ∆T m values tends to be accompanied by the antitubercular activity of the corresponding compounds. These data implied that the antitubercular effect of N-phenylindole derivatives might be due to the targeting of the Pks13-TE.  [11,16].
Two compounds, 48 and 58, were selected for the initial assessment of microsomal stability based on their potency. As shown in Table 4

General
All the solvents, starting materials, and chemical reagents were purchased from commercial sources and used without further purification, unless otherwise stated. Anhydrous tetrahydrofuran, dichloromethane, and 1,2-dichloroethane (DCE) were obtained by distillation over sodium wire or calcium hydride, respectively. TLC was performed on silica gel plates (GF254) to visualize components by UV light (254 nm). Column chromatography was carried out on silica gel (200-300 mesh). All non-aqueous reactions are carried out under a nitrogen atmosphere, the reagents do not contain water, and all reaction vessels are dried. 1 H NMR spectra were obtained on Bruker at 400 MHz. 13 C NMR spectra were obtained at 101 MHz (see Supplementary Materials). High-resolution mass spectra (HRMS) were performed using a Bruker ESI-TOF high-resolution mass spectrometer. NMR chemical shifts were reported in δ (ppm) using the δ 0 signal of tetramethylsilane, or the residual non-deuterated solvent signal (δ 7.26 signal for CDCl 3 , δ 3.31 signal for CD 3 OD, or δ 2.50 signal for (CD 3 ) 2 SO, or in case of a mixed solvent, the CD 3 OD signal, as internal standards).

3-(4-methoxyphenyl)-3-oxopropanoate (method A)
A commercially available mixture of 10 (1.0 mmol), ammonium formate (5.0 mmol), and molecular sieves (4 Å, 0.2 g) in 10 mL ethanol was refluxed for 8 h under N 2 and then cooled to room temperature. The reaction mixture was filtered through celite. The filtrate was evaporated. The residue was extracted with ethyl acetate (3 × 20 mL), washed with NaCl aqueous solution, the combined organic layer was dried with anhydrous Na 2 SO 4, and the residues were purified by flash chromatography (petroleum ether: ethyl acetate = 20:1) to give 11.

General procedure for the preparation of indole construction (method B)
To a solution of 11 (1.0 mmol) and ZnBr 2 (1.0 mmol) in 3 mL, anhydrous THF has added a solution of benzoquinone (1.0 mmol) in THF (2 mL) under N 2 . After stirring for 6 h at room temperature. The reaction mixture was quenched with saturated NH 4 Cl aqueous solution and extracted with EtOAc (3 × 10mL). The combined organic phases were dried over anhydrous Na 2 SO 4 and evaporated. The residue was purified by flash chromatography (petroleum ether: ethyl acetate = 10:1) to give 12. General procedure for the Mannich reaction of 5-hydroxy indole derivatives with amine and formaldehyde (method C) To a solution of indole analogues (1 mmol) in ethanol (3 mL) were added formaldehyde (37% in water, 4 mmol) and the appropriate amine (4 mmol) at room temperature under N 2 . The reaction mixture was allowed to reflux for 8-12 h and then cooled to rt. The reaction mixture was evaporated, and the residue was purified by flash chromatography (dichloromethane: methanol = 60:1). General procedure for the One-pot reaction of 5-hydroxy indole derivatives (method D) Under the catalysis of montmorillonite (5 mmol), the corresponding amine (1 mmol) and ethyl acetoacetate (1 mmol) was dissolved in refluxing anhydrous DCE under N 2 for 30 min and then to it was added benzoquinone (1 mmol) DCE solution dropwise. After reacting for 8 h, it was cooled to room temperature. The reaction mixture was filtered through celite; the filtrate was evaporated; the residue was purified by flash chromatography (petroleum ether: ethyl acetate = 5:1) to give the product. Ethyl (Z)-3-amino-3-(4-methoxyphenyl) acrylate (11).
To a solution of Compound 13 (1.0 mmol) in anhydrous CH 2 Cl 2 (4 mL), BBr 3 (1 M in CH 2 Cl 2 , 4.0 mmol) was added at room temperature under N 2 . After being stirred overnight, the reaction mixture was quenched with EtOH. The residue was purified by flash chromatography (dichloromethane: methanol = 50:1). Yield 89%; pale yellow solid. 1 99 (s, 3H), and 1.86-1.39 (m, 6H). 13   This compound was obtained from N-methyl-3-oxobutanamide and 4-methoxyaniline, employing Methods D and C. Overall yield 35%; pale brown solid. 1 To a solution of Compound 43 (1mmol) in toluene: EtOH: H 2 O (3:2:1) (12 mL) under N 2 , were added 3 M aqueous solution of K 2 CO 3 (18.88 mmol), phenylboronic acid (11.62 mmol), and Pd(PPh 3 ) 4 (0.29 mmol). The mixture was stirred at 100 • C for 4 h. After cooling to r.t. The resulting suspension was diluted with EtOAc (20 mL), the aqueous layer was extracted with EtOAc (30mL), and the combined organic layers washed with NaCl aqueous solution. The residue was purified by flash chromatography (dichloromethane: methanol = 40:1) to afford 44. Yield 42%; pale yellow solid. 1  To a solution of indole P-nitroaniline (1 mol) in DMSO (30 mL) was added K 2 CO 3 and piperidine (4 mol) at room temperature under N 2 . The reaction mixture was allowed to 90 • C for 8-12 h and then cooled to rt. The reaction mixture was extracted with ethyl acetate (3 × 20 mL), washed with NaCl aqueous solution, and the combined organic layer was dried with anhydrous Na 2 SO 4 . The residue was evaporated and purified by flash chromatography (petroleum ether: ethyl acetate = 20:1). Then, performed catalytic hydrogenation without further purification. Overall yield 90%; yellow solid. 1 H NMR intensity ratio (or the first derivative of the ratio) as nano DSF thermogram. The thermal transition temperature (T m ) is obtained in the post-run data analysis. Liver microsome stability assay. The assay was performed with liver microsomes from human. The incubation was performed as follows: microsomes in 0.01 M phosphate buffer pH 7.4 (0.56 mg/mL microsomal protein), tested compounds (final concentration 0.1 µM, cosolvent (DMSO)), and then NADPH (1 mM) at 37 • C with constant shaking for 60 min. The reaction can be started by adding NADPH or the same volume buffer. Aliquots were sampled at 5, 15, 30, 45, and 60 min incubation, and enzymatic reaction was quenched by addition of acetonitrile. After centrifugation, samples were then analyzed by LC-MS/MS. The assay evaluated the metabolic stability of compounds by measuring the substrate remaining with or without NADPH cofactor.

Conclusions
Pks13 is a promising target for the development of novel anti-TB drugs. This report has synthesized a series of N-phenylindole derivatives targeting Pks13-TE based on a structureguided strategy. The SAR studies showed that the nitrogen of indole substituted by phenyl was favorable for activity. Further exploration demonstrated that introducing hydrophobic groups at the para position of the benzene ring resulted in a significant improvement for antitubercular activity against Mtb. At the three-position of N-phenylindole derivatives, acetyl substituents are better than the ester group and amide substituents for activity. The rational drug design on the N-phenylindole series resulted in the discovery of the potent Compounds 45 and 58, with an MIC value of 0.0625 µg/mL and 0.125µg/mL, respectively. We further verified that N-phenylindole derivatives are bound to the Pks13-TE domain using the nano DSF method, consistent with the observed MIC trends. The preliminary metabolism evaluation of Compound 58 revealed moderate human microsomal stability. Taken together, N-phenylindole derivatives as a novel anti-TB scaffold have the promise to warrant the further development of lead compounds.