Benzimidazoles are recognized as an excellent platform for the development of new molecules with potential biological activity. Taking into account that the new molecules should have a simple, easily-prepared chemically structure, the test compounds were herein synthesized by environmentally friendly methods based on short and efficient synthesis procedures.

Among the reported methods for obtaining benzimidazoles, some utilize harsh (Gobis et al., 2015; Aghapoor et al., 2019) and others soft reaction conditions (Leutbecher et al., 2011; Sriramoju et al., 2018). Whereas, the former is carried out at a high temperature and with strong acids, the second is performed at or near room temperature (rt) in the presence of soft acids, mainly through heterogeneous catalysis. The first series of benzimidazoles herein synthesized was elaborated in two stages with a short, efficient, and sustainable method. The N-alkylation reaction used to form the second series required a longer synthetic procedure and led to low yields.

The procedure published by Abdelhamid et al. (1988) was modified to synthesize the new benzimidazoles, 5a and 5b. In order to substitute a nitro group at the 5-position of 3a and 3b (containing Cl and Br, respectively), the two 2-thiophenehydroxymoyl halides were treated with 4-nitro-1,2-phenylenediamine [4] in dry DMF at 80°C for 6 h (Scheme 1). Then the 2,5-disubstituted benzimidazoles were N-alkylated with hexadecyl methanesulfonate in basic medium at 120°C for 4 h to provide the 1,2,5-trisubstituted benzimidazoles (7a and 7b) (Scheme 1).

2-Thiophenecarboxaldehyde oxime 2 was synthesized following the procedure described by Iwakura et al. (1968) to give the desired product as a white solid in 93% yield (mp 137–138°C, reported as 136°C) (Iwakura et al., 1968). In its FT-IR spectrum, oxime 2 showed a strong and broad band at 3,025 cm⁻¹ attributed to the NO-H bond, and a band at 1632 due to the C=NOH double bond (Figure S1). In its ¹H NMR spectrum, the compound exhibited two singlets at 11.68 and 7.68, corresponding to the hydroxyl and oxime protons, respectively, and three doublets of doublets at 7.54, 7.35, and 7.05 ppm assigned to H₅, H₃, and H₄, respectively (Figure S2). In its ¹³C NMR spectrum, it displayed signals at 139.4, 131.0, 130.2, 130.1, and 125.6 ppm (Figure S3). The ¹H NMR data were similar to those in the literature (Langa et al., 2000). On the other hand, 5-chlorothiophene-2-carbohydroximoyl chloride (3a) and 5-bromothiophene-2-carbohydroxamoyl bromide (3b) were synthesized with a modified protocol of Kanemasa (Kanemasa et al., 2000) by using 3 molar equivalents of N-chlorosuccinimide (NCS) or N-bromosuccinimide (NBS) relative to the oxime. Compound 3a was afforded as a yellow solid in 95% yield (mp 129.5°C, similar to the 127–128°C reported by Iwakura et al.) (Iwakura et al., 1968), and 3b was furnished as a brown oil in 93% yield.

One of the new benzimidazoles, 5a, was obtained as an amorphous beige solid in 60% yield (mp 262–263°C). In the FT-IR spectrum of 5a, there were strong bands at 3,253, 1,576, 1,504, and 1,338 cm⁻¹, corresponding to the H-N, C=N, and NO₂ groups, respectively (Figure S4). In its ¹H NMR spectrum, the benzimidazole ring generated a doublet at (δ) 8.44 (J = 2.4 Hz for H₄), a doublet of doublets at 8.17 (J = 8.7 and 2.4 Hz for H₆), and a doublet at 7.72 (J = 9.7 Hz for H₇), while the thiophene system produced two doublets at 7.76 and 7.19 (J = 3.9 Hz) for H3′ and H4′, respectively (Figure S5). In the ¹³C NMR spectrum, resonances for 11 aromatic carbons were observed at δ (ppm): 150.5, 144.5, 132.6, 128.9, 128.2, 119.7, 119.5, 118.8, 115.8, 112.1, and 108.5 (Figure S6). The LR-MS spectrum of benzimidazole 5a revealed the expected pattern of molecular ions at 279 M and 281 M+2. The HR-MS spectrum allowed for the determination of an exact experimental mass of 279.9811 (M+H) (Figure S7) and an exact calculated mass of 279.9947 (M+H), thus confirming the structure of benzimidazole 5a (Figure S8).

The other new benzimidazole, 5b, was provided as an amorphous brown solid in 98% yield (mp 254–255°C). In the FT-IR spectrum of 5b, the strong and broad band at [υ (cm⁻¹)] 3,385 cm⁻¹ corresponded to the N-H bond, and the strong bands at 1,625, 1,573, and 1,337 cm⁻¹ to the C=N and NO₂ groups (Figure S9). The ¹H NMR spectrum showed a singlet [δ (ppm)] at 8.38 (for H₄), a doublet at 8.09 (J = 7 Hz for H₆), a singlet at 7.70 (for H₇ and H3′), and another singlet at 7.38 (for H4′) (Figure S10). In the ¹³C NMR spectrum, only nine signals were found [δ (ppm)], at 150.1, 142.8, 133.9, 132.0, 128.9, 118.2, 116.1, 114.4, and 111.8] (Figure S11). In the LR-MS spectrum, 5b displayed molecular ions at 323 and 325 due to the presence of a bromine atom (Figure S12). Based on the HR-MS spectrum, an exact experimental mass of 323.0410 was established, as was the exact calculated mass of 322.9364 (Figure S13).

The 1,2,5-trisubstituted benzimidazoles 7a and 7b were obtained by N-alkylation with the hexadecyl methanesulfonate in basic medium by heating at 124°C for 3 h, giving the regioisomer 7a as a beige amorphous solid in 35% yield (704 mg) (mp 94–95°C). Its IR (KBr) spectrum exhibited the following absorption bands [υ(cm⁻¹)]: two very strong bands at 2,916 and 2,848 corresponding to the aliphatic chain, a weak band at 1,520 owing to C=N, and bands at 1,472 and 1,462 attributed to the NO₂ group. The ¹H NMR spectrum showed the following signals at δ (ppm): a singlet signal at 8.31 for H₄, a broad doublet of doublets at 8.21, the coupling constant (J = 10 Hz) for H₅, and three doublets at 7.79 (J = 10 Hz), 7.41, and 7.05 (J = 5 Hz) assigned to H₇, H₃, and H₄, respectively. Additionally, the triplet at 3.62 (J = 5.0 Hz) was ascribed to the methylene protons next to the benzimidazole nitrogen, a very broad singlet at 1.33–1.24 associated with protons of the aliphatic chain, and finally a triplet signal at 0.86 (J = 5 Hz) corresponding to protons of the methyl group (Figure S14).

The regioisomer 7b was furnished as a beige amorphous solid in 30% yield (202.9 mg) (mp 93–94°C). In the IR (KBr) spectrum, the absorptions bands [υ (cm-1)] consisted of: two very strong bands at 2,917 and 2,850 due to C-H of the aliphatic chain, a weak band at 1,614 owing to C=N, and two bands at 1,470 and 1,331 caused by the NO₂ group. The ¹H NMR spectrum displayed the following signals at δ (ppm): a broad singlet at 8.39 attributed to H₄, a broad triplet of doublets at 8.28 (J = 5 and 10 Hz) assigned to H₅, a doublet at 7.88 (J = 10 Hz) corresponding to H₇, two doublets at 7.73 and 7.69 (J = 5 Hz) ascribed to H3′ and H4′, a triplet at 3.98 (J = 5 Hz) owing to H₂CN, a very broad singlet from 1.41–1.31 due to aliphatic chain protons, and finally, a triplet at 0.93 (J = 5 Hz) produced by methyl group protons (Figure S15).

Benzimidazoles 11 and 12 were synthesized by adopting recently published environmentally friendly protocols (Cardozo et al., 2015) with 9 as a key intermediate. Thus, nitro compound 8 was reduced with sodium metabisulfite in aqueous medium and heated to reflux to give 4-chloro-1,2-phenylenediamine 9. Compound 9 was obtained as an amorphous beige solid (mp 173–174°C) (Park et al., 2014) in quantitative yield. The spectroscopic data were similar to results documented in the literature (Cantillo et al., 2013). 4-Chloro-1,2-phenylenediamine 9 (without purification) was reacted with 3-nitrobenzaldehyde 10 and bentonite clay, the latter as a heterogeneous catalyst, under constant stirring at rt (Scheme 2) (Mckinney et al., 1962; Naeimi and Alishahi, 2014). 2-(3′-Nitrophenyl)-5-chlorobenzimidazole (11) was produced as an amorphous beige solid in 64% yield (mp 243–244°C) as described by Keurulainen et al. (2010) (mp 243°C).

Its FT-IR spectrum showed strong bands at υ: 3,322, 1,513, and 1,347 cm⁻¹, corresponding to the N-H and NO₂ groups. In the ¹H NMR spectrum of benzimidazole 11, the phenyl group exhibited triplets (δ) at 8.96 (J = 1.2 Hz for H2′) and 7.82 (J = 7.8 Hz for H5′), and doublets of triplets at 8.55 (J = 7.8 and 1.2 Hz for H6′) and 8.31 (J = 7.8 and 1.2 Hz for H4′). On the other hand, the protons of the benzimidazole ring generated a triplet at (δ) 7.64 (J = 9.0 Hz for H₄ and H₇) and a doublet of doublets at 7.24 (J = 8.4 Hz for H₆). In the ¹³C NMR spectrum, there were eight resonances for 13 carbons. The LR-MS spectrum of benzimidazole 11 displayed the expected pattern with molecular ions at 273 (M) and 275 (M+2). Spectroscopic and spectrometric properties were similar to those reported previously (Keurulainen et al., 2010).

The trisubstituted benzimidazole 12 was prepared by N-alkylation of 11 in hexadecyl methanesulfonate 6 basic medium and heating at 120°C for 4 h to provide an amorphous beige solid in 18% yield (36 mg) (mp 95–96°C). In the ¹H NMR spectrum, the following signals can be appreciated at δ (ppm): a broad singlet at 8.59 assigned to H2′, a broad doublet of doublets at 8.39, (J = 5 Hz) attributed to H4′, two broad doublets at 8.13 (J = 5 Hz) and 7.81 (J = 5 Hz) ascribed to H6′ and H₄, respectively, a triplet at 7.75 (J = 5 Hz) corresponding to H5′, a triplet at 7.36 (J = 10 Hz) and 7.34 (10 Hz) owing to H5′ and H₇, a triplet at 3.64 (J = 5 Hz) associated with the methylene protons next to the benzimidazole nitrogen, a quintuplet at 1.57 (J = 5 Hz) due to β-methylene protons close to the benzimidazole nitrogen, a very broad singlet at 1.26 caused by protons of the aliphatic chain, and finally, a triplet at 0.88 (J = 5 Hz) generated by methyl group protons (Figure S16).

The benzimidazole scaffold is attractive for drug design because of the diverse activities of its derivatives. Benzimidazoles substituted with electron-withdrawing groups, such as nitro (Pytela and Klimešová, 2011; Roh et al., 2017) and halogen groups (Cl, F, and Br), have shown enhanced biological activity (Klimesová et al., 2002). The disubstituted and trisubstituted benzimidazole molecules herein synthesized were evaluated for their in vitro antimicrobial effect on M. tuberculosis H37Rv. The in vitro data revealed that the 2,5-disubstituted benzimidazoles were more efficient than the 1,2,5-trisubstituted compounds in their capacity to kill the mycobacterium. The benzimidazole derivatives capable of effectively inhibiting the growth of M. tuberculosis were 5a, 5b, and 11, with MIC values of 89.6, 19.4, and 22.9 nM, respectively (Table 1). Contrarily, low antimycobacterial activity was detected for the molecules trisubstituted at the 1-position with the aliphatic chain, 7b, 7a, and 12, with MIC values of >182, >198, and >201 nM, respectively.

Based on the in vitro results, the toxicity of 5a, 5b, and 11 was tested on murine macrophages infected with M. tuberculosis (Figure 1). This intracellular bacillus mainly resides in alveolar macrophages and type II pneumocytes. Although only moderate antimycobacterial activity was found for 5a in vitro (25 μg/mL MIC), it displayed a good intracellular effect. After 48 h of treatment, only 10% of mycobacteria were viable, demonstrating the inhibition of mycobacterial replication, similar to the type of activity reported for isoniazid. Compounds 5b and 11 exhibited a moderate intracellular effect. After exposure to 5b and 11, only 30% of the mycobacteria could recover. Hence, these compounds also impeded bacterial replication, and their toxicity toward mycobacteria was lower than that of 5a. The data published for 11 on the inhibition of the intracellular C. pneumoniae are similar to the current findings (Keurulainen et al., 2010).

For the cytotoxicity assays, macrophages were exposed to the benzimidazole derivatives at concentrations higher than the MIC for 24 h. Subsequently, the viability of the cells was determined with Alamar Blue. None of the compounds showed any significant cytotoxic effect on macrophages (Figure 2). Even 5a, with the lowest viability (86%), is not precluded from being used as an antimycobacterial agent.

Since nitro and halogen groups are known to enhance the biological activity of pharmacophores (Salar et al., 2017), examination was presently made of the effect of different electron-withdrawing groups (2-haloheteroaryl or nitroaryl, 5-chloro or nitro) at positions two and five of the benzimidazole core. The disubstituted (5a, 5b, and 11) but not the trisubstituted (7a, 7b, and 12) benzimidazoles produced good antimycobacterial activity against M. tuberculosis. Compounds 5b and 11 induced a better effect than 5a (19.4 nM and 22.9 nM, respectively, versus 89.6 nM). The former two molecules each contain two electron-withdrawing groups, and a cooperative effect between them may explain the robust antimycobacterial activity. In relation to the two electron-withdrawing groups of 5b, there is one bromo at the C-5′ position of the 2-(thien-2′-yl) and a nitro group at the C-5 position of the benzimidazole ring. On the other hand, 11 bears a nitro group at the C-3′ position of 2-(p-3'-nitrophenyl-1'-yl) and a chloro group at the C-5 position of the benzimidazole ring. Compound 5a, an analog of 5b, has a chloro group at the C-5'position of 2-(thien-2'-yl) instead of a bromo group, and a nitro group at the C-5 position of benzimidazole ring. Its lower antimycobacterial potency possibly owes itself to a less efficient cooperative effect of the substituents against M. tuberculosis. The charge of the cell wall of M. tuberculosis is negative (Hoque et al., 2012; Phillips et al., 2017), while that of the 2,5-disubstituted benzimidazoles is likely to be positive, due to the acceptor ability of these compounds for nitrogen atoms. Hence, another plausible hypothesis about the efficient antimycobacterial activity of the 2,5-disubstituted benzimidazoles is their potential ability to interact with the cell wall of mycobacteria through electrostatic charges.

In the case of 1,2,5-trisubstituted benzimidazoles, we hypothesized that the long aliphatic chain of sixteen carbon atoms bonded on the 1-position of the benzimidazole ring might improve the non-covalent intermolecular interaction with the highly lipophilic cell wall of the mycobacterium (Daffé, 2015), leading to an increase in antimycobacterial activity. According to the current data, however, the introduction of the aliphatic chain was detrimental to such activity. Similarly, the alkyl chains in novel aminopyrimidinyl benzimidazole were reported to reduce antibacterial potency (Liu et al., 2018).

The difference in the biological activity between the two series of benzimidazoles may stem from their distinct solubility properties. The 2,5-disubstituted benzimidazoles were soluble in dimethylsulfoxide and the 1,2,5-trisubstituted benzimidazoles in chloroform, owing to the strong influence in the latter compounds of a hydrophilic/hydrophobic imbalance, being predominant the hydrophobic effect. In this sense, a shorter aliphatic chain for the 1,2,5-trisubstituted benzimidazoles could possibly generate better antimicrobial activity against M. tuberculosis, considering greater equilibrium in the hydrophobic/hydrophilic effect.

In order to explore the target of the two series of compounds and provide insights into why they showed disparate antimycobacterial effects, molecular docking was carried out on the FtsZ protein of M. tuberculosis. This docking study was based on the binding site previously reported by Li et al. (2017). The protein subunit consists of two globular N- and C-terminal subdomains linked by a central H7 helix having an adjacent T7 loop. FtsZ, a prokaryote tubulin homolog protein, plays a major role in the cytokinesis of mycobacteria. In the presence of guanosine triphosphate (GTP), FtsZ polymerizes to form a dynamic Z-ring inside the cell on the inner membrane and then recruits other cell division proteins (Hong et al., 2013). There is evidence that the disturbance of proper FtsZ assembly would block septum formation and consequently prevent cell division, which should lead to the inhibition of bacterial growth and the promotion of cell death (Slayden et al., 2006; Respicio et al., 2008; Schaffner-Barbero et al., 2012).

Molecular docking was performed on the six test compounds, which had been found to exhibit different antimycobacterial activity. The compounds were docked to get the best molecular confirmations in the interdomain cleft, which is crucial to GTP-dependent polymerization of the FtsZ protein. The binding energy and H-bonding were taken as two important parameters of the interactions (Figure 3).

Based on the molecular docking study, the three 2,5-disubstituted benzimidazoles, 5a, 5b, and 11, adopt a similar arrangement. The halogens are exposed to FtsZ and the hydrophobic groups interact with the T7 loop of the protein. Two hydrogen bonds were detected for 11 and only one for 5a and 5b. Thus, the former showed a better binding energy (−6.2 kcal/mol). The main amino acids in the protein that established hydrogen bonds with the benzimidazoles were I 288 and D 290 (Figure 3). The molecular docking results for 5a, 5b, and 11 are consistent with previous reports. The benzimidazole scaffold and cyclohexyl group were observed inside the hydrophobic environment, with the halogen groups exposed to the cleft of FtsZ (Ding et al., 2016).

Contrarily, the three 1,2,5-trisubstituted derivatives containing hydrocarbon chains, 7a, 7b, and 12, do not dock at the same domain nor show strong affinity. Rather, they dock to the N-terminal side of the protein close to the GTP binding domain. Structurally, they interact with the domain formed by helices four and five (H4 and H5). Because of their size, they are not able to position themselves in any cavity of the protein. Both binding affinity and docking results correlated with the antimicrobial effective dose.

4-Nitro-1,2-phenylenediamine, 4-chloro-2-nitroaniline, 2-thiophenecarboxaldehyde, 3-nitrobenzaldehyde, 2-pyridinecarboxaldehyde, N-bromosuccinimide, and N-chlorosuccinimide (Sigma-Aldrich, St. Louis, MO, USA), were used for syntheses without additional purification. The organic solvents were purified by the standard methods. The progress of each reaction was followed by thin-layer chromatography (TLC) on silica plates (silica gel 60 F₂₅₄, Merck) and with UV light, iodine vapor, or an acidic aqueous solution of KMnO4 for visualization. Column chromatography was performed on silica gel (70–230 mesh, Sigma-Aldrich) with freshly distilled hexane/ethyl acetate as the eluent. Dimethylformamide (DMF) was dried with CaH₂ at reflux. Acetonitrile was dried with phosphorous pentoxide under constant stirring at rt. Melting points were determined on an electrothermal melting point apparatus. Infrared spectra (FT-IR) were recorded on a double-beam Perkin-Elmer Model 1605 FT-IR spectrometer. NMR spectra were recorded in acetone-d₆, DMSO-d₆, or CDCl₃ solution on Gemini 200, Eclipse 300, or Agilent 500 spectrometers operating at 200, 300, and 500 MHz for ¹H NMR and 50, 75, and 125 MHz for ¹³C NMR, respectively. Chemical shifts are described in parts per million (ppm), employing Me₄Si as the internal standard. Coupling constants (J) are expressed in Hz. Mass spectra were recorded on a double-beam JEOL JMS AX505HA spectrometer in the electron impact mode and high-resolution mass spectrometry (HR-MS) with electrospray ionization (Bruker Daltonics, Billerica, MA, USA). All experiments were carried out at rt.

The benzimidazole derivatives were synthesized by two different chemical pathways, one of which was a modification of the method proposed by Abdelhamid et al. (1988), who used phenylenediamine or phenylenediamine with electron-donating substituents. Briefly, 2-thiophenebenzaldoxime (2) was synthesized by stirring a mixture of 2-thiophencarboxaldehyde and hydroxylamine hydrochloride in basic medium for 2 h. Subsequently, 2 was treated with NCS or NBS in dry DMF at 15–20°C for 2 h to form 3a or 3b.

For the synthesis of the new benzimidazoles, the key intermediate (2-thiophenhydroxymoyl halides 3) was reacted without purification with 4-nitro-1,2-phenylenediamine (4) in dry DMF at 80°C for 6 h to obtain 5a and 5b. The novel benzimidazoles 7a, 7b, and 12 were synthesized by N-alkylation of 5a, 5b, and 11 with hexadecyl methanesulfonate 6 and sodium bicarbonate in dry acetonitrile and heating at reflux for 4 h.

The second method was only for the synthesis of 11 and 12. 4-Chloro-1,2-phenylenediamine 9 served as a key intermediate, prepared from the reduction of 4-chloro-2-nitroaniline 8 with sodium metabisulfite (Mckinney et al., 1962; Naeimi and Alishahi, 2014). Compound 9 was mixed with 3-nitrobenzaldehyde 10 in dry acetonitrile at rt, using bentonite clay as a heterogeneous catalyst, to give 11 (Cardozo et al., 2015). The latter was N-alkylated with hexadecyl methanesulfonate 6 basic medium and heated at 120°C for 4 h to afford the trisubstituted benzimidazole 12.

The mycobacterial effect of benzimidazole derivatives was evaluated on the M. tuberculosis H37Rv strain by performing microcolorimetric assays with Alamar Blue dye, according to Collins and Franzblau (1997). Briefly, the four benzimidazoles at concentrations in the range of 200–3.125 μg/mL were tested in 96-well sterile microplates. Isoniazid was included as the reference drug. Stock solutions of the benzimidazoles were prepared in dimethylsulfoxide at 20 mg/mL in sterile conditions and stored at −70°C to await further use. Standard solutions were obtained by consecutive twofold dilutions of the stock solutions in 7H9 Middlebrook medium until reaching 800 μg/mL, followed by the addition of 100 μL of Middlebrook 7H9 broth, 100 μL of the benzimidazole solution and 100 μL of mycobacterial suspension to each well. A 1:10 dilution was made for the control from the bacterial suspension, representing the growth of 10% of the bacterial population assayed. After incubating the plates at 37°C for 5 days, 20 μL of Alamar Blue solution (AbD Serotec) and 12 μL of sterile 10% Tween 80 were added. The plates were incubated at 37°C for another 24 h before the determination was made, considering wells with a pink color darker than that of the 10% control well to be positive for growth. The minimal inhibitory concentration (MIC) was defined as the lowest at which benzimidazole impeded development of the pink color.

In order to examine the intracellular antimycobacterial activity, monolayers of mouse macrophages (J774A.1) were infected and then treated in independent assays with each benzimidazole or isoniazid. Macrophage monolayers were prepared by utilizing 2.5 × 10⁵ cells/mL in RPMI-1640 medium (Sigma-Aldrich) supplemented with 10% fetal bovine serum (HyClone®, Thermo SCIENTIFIC) and placed in 24-well plates. The plates were incubated at 37°C in a humidified 5% CO₂ atmosphere until the monolayers reached 90% cell confluency. Subsequently, the monolayers were washed with Hanks' balanced salt solution (HBSS) to remove the non-adhered cells. To infect the cells, the bacterial suspensions were adjusted to a multiplicity of infection (MOI) ratio of 1:10, and 1 mL of the suspension was added to each well. After infected monolayers were incubated at 37°C under 5% CO₂ for 2 h, monolayers were washed three times with HBSS to eliminate the bacteria not internalized, followed by the addition of a 1-mL solution of one of the compounds or isoniazid diluted in RPMI-1640 medium. Whereas, all test compounds were evaluated at MIC, +MIC, and –MIC, isoniazid was assayed at 0.5, 0.25, and 0.125 μg/mL. Untreated infected cells were included as controls. The cells were incubated at 37°C under 5% CO₂, and the kinetics of the activity was monitored at 24, 48, 72, and 96 h. At each time point, cells were lysed by the adding of 500 μL of 0.25% SDS and neutralized with 500 μL of 5% bovine serum albumin (BSA) solution. Serial dilutions of the cell lysates were plated on 7H11 Middlebrook agar supplemented with 10% OADC (oleic acid, albumin, dextrose and catalase) and incubated at 37°C for 3 weeks to determine the number of colony-forming units (CFUs).

In order to assess the possible toxicity of benzimidazole derivatives to macrophages, a viability assay was performed. Macrophage monolayers were prepared as aforementioned (in subsection intracellular activity of the benzimidazoles) up to washing with HBSS to remove the non-adhered cells. Subsequently, 1 mL of the 50 μg/mL solution of one of the benzimidazoles was added. Some wells were maintained without treatment (viability control), and to others were added 40 μg/mL ursolic acid (death control). The cells were incubated for 24 h at 37°C under 5% CO₂. Upon completion of this time, 100 μL of Alamar Blue solution was added, the monolayers were incubated again until the viability control wells turned pink. The relative fluorescence units were quantified in a Fluoroskan Ascent FL Microplate fluorometer, and the RFUs of the control wells without treatment were taken as 100% viability.

The molecular docking simulations were carried out on Autodock Vina 4.2.6, which has been calibrated and tested with molecular dynamics by using a set of small-protein ligands of medicinal interest (Perryman et al., 2015; Forli et al., 2016). This software maximizes the accuracy of predicting the binding of small molecules to proteins, assessing the veracity of the predictions by comparing them to the experimental structure. Evaluation of the difference between the experimental and predicted structures considers symmetry, partial symmetry, and near-symmetry (Trott and Olson, 2010).

The X-ray structure of M. tuberculosis eukaryotic tubulin homolog protein FtsZ (2.08Å resolution, PDBID 1RLU) (Leung et al., 2004) was obtained from the Protein Data Bank database (http://www.pdb.org). The three-dimensional structures of the ligands were built and their energy minimized with MarvinSketch 18.24.0 ChemAxon software (http://www.chemaxon.com). The ligands and protein were then prepared for molecular docking by adding hydrogens and Gasteiger charges with Autodock tools 1.5.6 (Sanner, 1999). A grid box of 30 Å on each side was placed into the dimer interdomain with grid points separated by 1.0 Å centered at the middle of the protein at x = 6.0, y = 29, and z = 27. The binding energies of the docked complexes were provided by the output Autodock Vina file and the ligand-protein binding interactions at the binding sites were analyzed on UCFS Chimera software (Pettersen et al., 2004).