Benzoic Acid Derivatives with Trypanocidal Activity: Enzymatic Analysis and Molecular Docking Studies toward Trans-Sialidase

Chagas, or American trypanosomiasis, remains an important public health problem in developing countries. In the last decade, trans-sialidase has become a pharmacological target for new anti-Chagas drugs. In this work, the aims were to design and find a new series of benzoic acid derivatives as trans-sialidase (TS) inhibitors and anti-trypanosomal agents. Three compounds (14, 18, and 19) sharing a para-aminobenzoic acid moiety showed more potent trypanocidal activity than the commercially available drugs nifurtimox and benznidazole in both strains: the lysis concentration of 50% of the population (LC50) was <0.15 µM on the NINOA strain, and LC50 < 0.22 µM on the INC-5 strain. Additionally, compound 18 showed a moderate inhibition (47%) on the trans-sialidase enzyme and a binding model similar to DANA (pattern A).


Introduction
Chagas disease, or American trypanosomiasis, is a chronic disease caused by the kinetoplastid protozoan parasite Trypanosoma cruzi. Although this disease is present mostly in the endemic poor rural areas of America (from Southern California to Argentina), it is also becoming an important health issue in metropolitan areas [1] and other non-endemic areas, such as North America [2] and Europe [3,4], due to population migrations [5][6][7][8][9][10][11]. In fact, the World Health Organization (WHO) has estimated that Chagas disease is nowadays affecting about 7 to 8 million people worldwide [12] and accounting for the death of over 12,500 patients per year [13,14].
Strategies to manage Chagas disease are based fundamentally on the regional elimination of domestic vectors [15] and on the pharmacological treatment of patients. According to chemotherapy  The inhibitors reported for TcTS have complex chemical structures. DANA, (1, Figure 2), a potent inhibitor of influenza neuraminidase, poorly inhibits TcTS with a reported Ki of 12.3 mM [34]. The DANA analogs 2-4 have exhibited improved (Ki of 12.2 mM, IC 50 of 5 mM and 10 µM, respectively) but still weak TcTS inhibition [35,36]. The 2,3-difluorosialic acid analogs, such as 5 (Figure 2), are mechanism-based TcTS inhibitors, which form a covalent bond with the hydroxyl group of Tyr342, but only at high concentrations (mM range) [37]. Mimicking a sialic acid structure has been explored in the search for potential TcTS inhibitors. The benzoic acid derivative compound 6 ( Figure 2) [35] can be highlighted, showing an IC 50 of 0.54 mM. In the predicted docked conformation of 6 ( Figure 2), its carboxylate moiety provides the expected strong interaction with the arginine triad (residues 35, 245 and 314). The hydroxyl group can form hydrogen bonds with both Glu230 and Gln195, and the amide NH group is hydrogen-bonded to Asp96, with the acetyl group located in the hydrophobic pocket. However, the apparently good fit predicted for 6 did not translate into strong TcTS inhibition [35]. The inhibitors reported for TcTS have complex chemical structures. DANA, (1, Figure 2), a potent inhibitor of influenza neuraminidase, poorly inhibits TcTS with a reported Ki of 12.3 mM [34]. The DANA analogs 2-4 have exhibited improved (Ki of 12.2 mM, IC50 of 5 mM and 10 µM, respectively) but still weak TcTS inhibition [35,36]. The 2,3-difluorosialic acid analogs, such as 5 (Figure 2), are mechanism-based TcTS inhibitors, which form a covalent bond with the hydroxyl group of Tyr342, but only at high concentrations (mM range) [37]. Mimicking a sialic acid structure has been explored in the search for potential TcTS inhibitors. The benzoic acid derivative compound 6 ( Figure 2) [35] can be highlighted, showing an IC50 of 0.54 mM. In the predicted docked conformation of 6 ( Figure 2), its carboxylate moiety provides the expected strong interaction with the arginine triad (residues 35, 245 and 314). The hydroxyl group can form hydrogen bonds with both Glu230 and Gln195, and the amide NH group is hydrogen-bonded to Asp96, with the acetyl group located in the hydrophobic pocket. However, the apparently good fit predicted for 6 did not translate into strong TcTS inhibition [35]. Compounds such as lactitol 7 (IC50 = 0.57 mM) ( Figure 2) and other lactose (acceptor substrate) analogues, which target the sialic acid acceptor site, were also weak inhibitors [38,39]. The most potent TcTS inhibitor known, resulting from modifications in the GM3 ganglioside (8, Figure 2) and with IC50 values in the 10-100 µM range [40], fills in both sialic acid acceptor and donor sites. This result shows the importance of designing compounds occupying both binding sites. This can clearly be observed in the docked conformation of compound 9 (MuNANA assay screening at 1 mM: 87% inhibition and IC50 = 0.21 mM) (Figure 2), where the respective carboxylate group interacts well with the arginine triad, and the amide function hydrogen bonds with Arg245 and Asp59, thus providing additional stabilization of the triad. The thiophene ring is seen in the sialic acid acceptor binding site and the benzene ring is positioned in the sialic acid site. The latter does not fill this site, thus the introduction of functional groups could provide further interactions with key amino acid residues (e.g., Asp96, and Gln195) [41].
In view of these previous works, and taking into account that inhibitor design against the sialidases or trans-sialidases of trypanosomes is reduced, our approach involved the synthesis of new non-sugar-based benzoic acid derivative inhibitors of TcTS that may lead to biological tools for or potential leads in drug design. The fact that the negatively charged carboxylate group of sialic acid represents the most important group able to interact with the sialic acid donor site through Compounds such as lactitol 7 (IC 50 = 0.57 mM) ( Figure 2) and other lactose (acceptor substrate) analogues, which target the sialic acid acceptor site, were also weak inhibitors [38,39]. The most potent TcTS inhibitor known, resulting from modifications in the GM3 ganglioside (8, Figure 2) and with IC 50 values in the 10-100 µM range [40], fills in both sialic acid acceptor and donor sites. This result shows the importance of designing compounds occupying both binding sites. This can clearly be observed in the docked conformation of compound 9 (MuNANA assay screening at 1 mM: 87% inhibition and IC 50 = 0.21 mM) (Figure 2), where the respective carboxylate group interacts well with the arginine triad, and the amide function hydrogen bonds with Arg245 and Asp59, thus providing additional stabilization of the triad. The thiophene ring is seen in the sialic acid acceptor binding site and the benzene ring is positioned in the sialic acid site. The latter does not fill this site, thus the introduction of functional groups could provide further interactions with key amino acid residues (e.g., Asp96, and Gln195) [41].
In view of these previous works, and taking into account that inhibitor design against the sialidases or trans-sialidases of trypanosomes is reduced, our approach involved the synthesis of new non-sugar-based benzoic acid derivative inhibitors of TcTS that may lead to biological tools for or potential leads in drug design. The fact that the negatively charged carboxylate group of sialic acid represents the most important group able to interact with the sialic acid donor site through strong interactions with the arginine triad led us to design simple structures containing this group. Neres et al., 2007 synthesized metaand para-substitutions of benzoic acid derivatives providing 4-acetylamino-3-hydroxymethylbenzoic acid (6, Figure 2) as the lead inhibitor, not containing the synthetically complex sialic acid structure (IC 50 = 0.54 mM). However, the weak inhibition prompted us to synthesize new non-sugar-based benzoic acid derivative inhibitors of TcTS with better interaction with both the sialic acid donor and acceptor sites of the TcTS enzyme. Our strategy was based on a molecular simplification of these previous compounds, performing slight, but chemically sensitive changes in the meta and para positions to the carboxylate group, and at the same time, we wondered whether or not an ortho substituent to the carboxylate group in the aromatic ring could establish better interaction with both the sialic acid donor and acceptor sites of the TcTS enzyme. In fact, docking studies were performed to support the design of better TcTS inhibitors and showed that there is a possibility of including an ortho substituent to the carboxylate group in the aromatic ring, which could bind the acceptor substrate binding site, thus offering room for improvement [35]. In this context, and following the interest in our research group to develop new therapeutic options for Chagas disease, we designed new derivatives of benzoic acid with ortho-, meta-, and para-substitutions according to preliminary docking studies and calculated physicochemical properties ( Figure 3).  Figure 2) as the lead inhibitor, not containing the synthetically complex sialic acid structure (IC50 = 0.54 mM). However, the weak inhibition prompted us to synthesize new non-sugar-based benzoic acid derivative inhibitors of TcTS with better interaction with both the sialic acid donor and acceptor sites of the TcTS enzyme. Our strategy was based on a molecular simplification of these previous compounds, performing slight, but chemically sensitive changes in the meta and para positions to the carboxylate group, and at the same time, we wondered whether or not an ortho substituent to the carboxylate group in the aromatic ring could establish better interaction with both the sialic acid donor and acceptor sites of the TcTS enzyme. In fact, docking studies were performed to support the design of better TcTS inhibitors and showed that there is a possibility of including an ortho substituent to the carboxylate group in the aromatic ring, which could bind the acceptor substrate binding site, thus offering room for improvement [35]. In this context, and following the interest in our research group to develop new therapeutic options for Chagas disease, we designed new derivatives of benzoic acid with ortho-, meta-, and para-substitutions according to preliminary docking studies and calculated physicochemical properties ( Figure 3).

Trypanocidal Activity
Initially, the lytic effect of the benzoic acid derivatives and reference drugs (Nfx and Bnz) on mouse blood trypomastigotes was evaluated in vitro, using INC-5 and NINOA T. cruzi strains ( Table  1) [47]. Furthermore, 4-acetamidobenzoic acid (24) previously assayed by Neres et al. [35] for TcTS inhibition has been studied. Biological activity showed that both reference compounds Nfx and Bnz have a similar lysis concentration of 50% of the population (LC50) in each strain: 0.21-0.29 µM against the NINOA strain and 0.62-0.68 µM against the INC-5 strain. As shown in Table 1, amongst the tested compounds, the ethyl benzoate derivatives 14 (benzocaine, a known anesthetic agent), 15, and 18 displayed significant trypanocidal activities in the range of 0.02-0.34 µM against the NINOA strain and 0.10-0.22 µM against the INC-5 strain, comparing to those of the corresponding carboxylic acid precursors 10, 24, and 17, respectively: 0.52-1.39 µM and 0.21-1.24 µM in the NINOA and INC-5 strains, respectively. Thus, the introduction of a hydrophobic group such as ethyl showed an increase of activity. In fact, these compounds presented trypanocidal activity comparable to the reference Bzn and Nfx against the NINOA strain and significantly higher against the INC-5 strain. Biological activity showed that both reference compounds Nfx and Bnz have a similar lysis concentration of 50% of the population (LC 50 ) in each strain: 0.21-0.29 µM against the NINOA strain and 0.62-0.68 µM against the INC-5 strain. As shown in Table 1, amongst the tested compounds, the ethyl benzoate derivatives 14 (benzocaine, a known anesthetic agent), 15, and 18 displayed significant trypanocidal activities in the range of 0.02-0.34 µM against the NINOA strain and 0.10-0.22 µM against the INC-5 strain, comparing to those of the corresponding carboxylic acid precursors 10, 24, and 17, respectively: 0.52-1.39 µM and 0.21-1.24 µM in the NINOA and INC-5 strains, respectively. Thus, the introduction of a hydrophobic group such as ethyl showed an increase of activity. In fact, these compounds presented trypanocidal activity comparable to the reference Bzn and Nfx against the NINOA strain and significantly higher against the INC-5 strain.
Compound 19, showing significant trypanocidal activity against the NINOA strain (0.14 µM) and the highest activity against the INC-5 strain (0.0008 µM), became a lead structure of the benzoic derivative series. The subsequent modification of this structure by replacing the chloride substituent by the acceptor methoxy group 20 did not improve the activity in both the NINOA (0.61 µM) and INC-5 (0.43 µM) strains.
Compound 10 is a simplification of the structures synthesized by Neres et al. [35] displaying moderate trypanocidal activity against the NINOA (0.52 µM) and INC-5 (1.24 µM) strains. The structure-activity relationship (SAR) of these compounds showed that the replacement of the para-amino group by nitro (13), hydrazine (11), or azide (12) groups showed comparable trypanocidal activities against the NINOA strain (0.47-0.66 µM) and an improvement against the INC-5 strain (0.46-0.58 µM). On the other hand, the meta-nitration of compound 10 afforded compound 16 to display a notably reduced activity against the NINOA strain (1.37 µM), whereas it increased the activity against the INC-5 strain (0.63 µM). Alternatively, ortho-hydroxylation afforded compound 21 to display significant trypanocidal activity against the NINOA (0.27 µM) and INC-5 (0.26 µM) strains. Compound 21 presented a higher trypanocidal activity than references Bzn and Nfx. Thus, the incorporation of a hydrophilic group as a hydroxyl in the ortho position of the para-amino benzoic acid moiety may provide a center for H-bond interactions and it might be positively affecting the biological activity. In fact, the elimination of the para-amino group generated compound 23 with lower trypanocidal activities (0.57 µM and 0.72 µM against the NINOA and INC-5 strains, respectively) and in addition, the N-acetylation of this group (compound 22) significantly reduced the biological activity (1.28 µM in both strains). This was partially observed in compounds 24 and 17, which did not reflect an improvement of the activity in relation to the para-aminobenzoic acid precursors 10 and 16 against the NINOA strain (>1.0 µM) but showed better trypanocidal activities against the INC-5 strain (0.87 and 0.21 µM, respectively). It is remarkable that a different trypanocidal activities tendency was observed in both strains. Finally, three compounds (14, 18, and 19) showed more potent trypanocidal activity than the commercially available drugs Nfx and Bnz in both strains: LC 50 < 0.15 µM in the NINOA strain and LC 50 < 0.22 µM in the INC-5 strain, all of them sharing the modified para-aminobenzoic ester moiety except 19. These structures can be used as references for current and future studies for the synthesis of new anti-Chagas compounds. We have demonstrated that the incorporation of ortho-hydroxyl groups in para-aminobenzoic acid derivatives successfully provided a more potent compound (21). The esterification of the carboxylic acid (14, 15, and 18) generating a hydrophobic moiety to increase the activity could be a good strategy in these compounds. However, the introduction of a meta-nitro group did not initially generate a more active compound (16), although subsequent structural modifications finally modulated the activity and improved it. Further assays are required to understand the electronic and steric properties of meta and ortho substituents in the para-aminobenzoic derivative structures.

TcTS Inhibition
TcTS inhibition screening results, including percentage inhibition values for a series of substituted benzoic acid derivatives, are given in Table 2. The percentage inhibition at 1 mM concentration is the average of at least three independent experiments. The enzymatic inhibition assay was performed using a continuous fluorimetric method based on the TcTS-catalyzed hydrolysis of 2-(4-methylumbelliferyl)-a-D-N-acetylneuraminic acid (MuNANA). As a control, the activities of pyridoxal phosphate (Pyr) and compound 24 [35] were measured in the same concentrations of the target compounds due to their respective moderate [48] and weak activities on TcTS.  Pyridoxal has been reported as a non-competitive TcTS inhibitor with a Ki = 7.3 mM [49] showing a moderate inhibition (73% at 1.0 mM in the presence of MuNANA 0.1 mM). In this work, the inhibition percentage of pyridoxal phosphate was 64%, similar to previous studies [50]. On the other hand, contrary to results described by Neres et al. [35], the inhibition of compound 24 was 30% instead of 0%. These differences could be mainly associated with the enzymatic assays' particularities. Nonetheless, this compound shows a weak inhibition, lower than pyridoxal phosphate.
According to Table 2, compounds 10-24 showed a variable influence on TcTS enzyme inhibition. All of the substances displayed a lower inhibitory activity against TcTS than pyridoxal, excepting compounds 11 and 17 which blocked 61 and 66% of the enzyme activity, respectively, with the same inhibitory effect as that of pyridoxal, and compound 16 showed the best inhibition activity (77%). Based on the inhibitory activity of these compounds, the para-amino-meta-nitrobenzoic acid core may have a relevant role in TcTS inhibition. Accordingly, compound 18, obtained from the ethyl esterification of compound 17, also showed moderate results (47%); however, esterification decreases the activity. This behavior also can be observed in compounds 10 and 24, where the activity decreases as compared with their ethyl ester analogs 14 and 15, respectively. The lack of the nitro substituent reduced notably the inhibition activity of these compounds. Consequently, the ortho-hydroxyl benzoic acid derivative compounds 21-23 and the para-amino benzoic acid/benzoate derivative compounds 10, 14 and 15 showed weak inhibition results in the range of 1-34%. However, the replacement of the para-amine group of the benzoic acid by hydrazine (11), azide (12), or nitro (13) groups lead to moderate results, with 61, 40, and 43% of inhibition, respectively. Again, the nitro group observed in compound 13 plays an important role in the interaction with the TcTS active site. Compounds 19-20 were not tested due to a low solubility in DMSO 1.0% in the well.
The significant TcTS inhibition shown by compounds 11, 16 and 17 (>60% inhibition) may direct the development of new derivatives as TcTS inhibitors, pointing out the necessity to maintain benzoic acid/benzoate core and nitrogenous moieties (amine, N-acetyl, nitro, azide, or hydrazine) for an efficient inhibition of this enzyme. Pyridoxal has been reported as a non-competitive TcTS inhibitor with a Ki = 7.3 mM [49] showing a moderate inhibition (73% at 1.0 mM in the presence of MuNANA 0.1 mM). In this work, the inhibition percentage of pyridoxal phosphate was 64%, similar to previous studies [50]. On the other hand, contrary to results described by Neres et al. [35], the inhibition of compound 24 was 30% instead of 0%. These differences could be mainly associated with the enzymatic assays' particularities. Nonetheless, this compound shows a weak inhibition, lower than pyridoxal phosphate.
According to Table 2, compounds 10-24 showed a variable influence on TcTS enzyme inhibition. All of the substances displayed a lower inhibitory activity against TcTS than pyridoxal, excepting compounds 11 and 17 which blocked 61 and 66% of the enzyme activity, respectively, with the same inhibitory effect as that of pyridoxal, and compound 16 showed the best inhibition activity (77%). Based on the inhibitory activity of these compounds, the para-amino-meta-nitrobenzoic acid core may have a relevant role in TcTS inhibition. Accordingly, compound 18, obtained from the ethyl esterification of compound 17, also showed moderate results (47%); however, esterification decreases the activity. This behavior also can be observed in compounds 10 and 24, where the activity decreases as compared with their ethyl ester analogs 14 and 15, respectively. The lack of the nitro substituent reduced notably the inhibition activity of these compounds. Consequently, the ortho-hydroxyl benzoic acid derivative compounds 21-23 and the para-amino benzoic acid/benzoate derivative compounds 10, 14 and 15 showed weak inhibition results in the range of 1-34%. However, the replacement of the para-amine group of the benzoic acid by hydrazine (11), azide (12), or nitro (13) groups lead to moderate results, with 61, 40, and 43% of inhibition, respectively. Again, the nitro group observed in compound 13 plays an important role in the interaction with the TcTS active site. Compounds 19-20 were not tested due to a low solubility in DMSO 1.0% in the well.
The significant TcTS inhibition shown by compounds 11, 16 and 17 (>60% inhibition) may direct the development of new derivatives as TcTS inhibitors, pointing out the necessity to maintain benzoic acid/benzoate core and nitrogenous moieties (amine, N-acetyl, nitro, azide, or hydrazine) for an efficient inhibition of this enzyme.
In summary, we can observe that there is no correlation between trypanocidal activity and TcTS inhibition. For example, of the benzoate derivatives 14-15 and 17-19 showing moderate trypanocidal activities (0.10-0.028 µM in the NINOA strain and 0.0008-0.22 µM in the INC-5 strain), only compound 15 moderately blocked the TcTS enzyme as expected with 47%, whereas the other compounds inhibited less than 7%. Nevertheless, the importance of para-amino carboxylic derivatives acting as TcTS inhibitors or lysing trypomastigotes via other biological mechanisms against T. cruzi has been demonstrated.

Molecular Docking
Docking studies were conducted for compounds 10-24 in order to obtain potential putative interaction of these compounds on the active site of TcTS. The docking analysis was validated using the crystal structure of TcTS against the reference inhibitor DANA [34]. The good superposition between the DANA structures oriented with AutoDock Vina 5.6 as compared with the orientation in the crystal structure (Protein Data Bank (PDB) accession number 1MS8) suggested that the chosen method is appropriate (Figure 4). The potential DANA binding site was also predicted as a deep cavity including a restricted space with two delimited regions, where the blue shift indicates a positive electrostatic potential (carboxylate interaction) and the red shift a negative electrostatic potential (amide interaction) ( Figure 4). In summary, we can observe that there is no correlation between trypanocidal activity and TcTS inhibition. For example, of the benzoate derivatives 14-15 and 17-19 showing moderate trypanocidal activities (0.10-0.028 µM in the NINOA strain and 0.0008-0.22 µM in the INC-5 strain), only compound 15 moderately blocked the TcTS enzyme as expected with 47%, whereas the other compounds inhibited less than 7%. Nevertheless, the importance of para-amino carboxylic derivatives acting as TcTS inhibitors or lysing trypomastigotes via other biological mechanisms against T. cruzi has been demonstrated.

Molecular Docking
Docking studies were conducted for compounds 10-24 in order to obtain potential putative interaction of these compounds on the active site of TcTS. The docking analysis was validated using the crystal structure of TcTS against the reference inhibitor DANA [34]. The good superposition between the DANA structures oriented with AutoDock Vina 5.6 as compared with the orientation in the crystal structure (Protein Data Bank (PDB) accession number 1MS8) suggested that the chosen method is appropriate (Figure 4). The potential DANA binding site was also predicted as a deep cavity including a restricted space with two delimited regions, where the blue shift indicates a positive electrostatic potential (carboxylate interaction) and the red shift a negative electrostatic potential (amide interaction) ( Figure 4). In this study, the proposed interaction modes of the benzoic and benzoate derivatives into the active site of TcTS were determined as the highest-scored conformations (best-fit ligands), which correspond to the structure with the most favorable free energy for binding in TcTS.
According to the results (Table S1, Supplementary Materials), two major binding patterns were found, A and B ( Figure 5A,B). Moreover, some conformations that did not have a particular binding mode were grouped as C ( Figure 5C). The first mode (A) is similar to that reported for DANA in the crystal structure ( Figure 5A). Although this binding mode was found to be the best result for compounds 15, and 24, it does not correspond to the lowest energy conformation in most cases. It is noteworthy that, binding mode A was found to be among the best nine conformations for most of the ligands, having a good calculated affinity (Table S1). Similarly, molecules 12 and 16-19 showed their best-scored binding in an alternative way, designated A2 (Table S1), which is a slight variation of the binding mode A. In this study, the proposed interaction modes of the benzoic and benzoate derivatives into the active site of TcTS were determined as the highest-scored conformations (best-fit ligands), which correspond to the structure with the most favorable free energy for binding in TcTS.
According to the results (Table S1, Supplementary Materials), two major binding patterns were found, A and B ( Figure 5A,B). Moreover, some conformations that did not have a particular binding mode were grouped as C ( Figure 5C). The first mode (A) is similar to that reported for DANA in the crystal structure ( Figure 5A). Although this binding mode was found to be the best result for compounds 15, and 24, it does not correspond to the lowest energy conformation in most cases. It is noteworthy that, binding mode A was found to be among the best nine conformations for most of the ligands, having a good calculated affinity (Table S1). Similarly, molecules 12 and 16-19 showed their best-scored binding in an alternative way, designated A2 (Table S1), which is a slight variation of the binding mode A.  Figure 6A shows the best conformation for compound 24, which corresponds to binding mode A, where the carboxylate group forms hydrogen bound interactions with Arg314 and Arg245 (donor site); these interactions can be slightly different for each ligand in binding mode A. For the molecules in binding mode A2, the nitro group adopts a similar orientation to the carboxylate group in DANA.  Figure 6A shows the best conformation for compound 24, which corresponds to binding mode A, where the carboxylate group forms hydrogen bound interactions with Arg314 and Arg245 (donor site); these interactions can be slightly different for each ligand in binding mode A. For the molecules in binding mode A2, the nitro group adopts a similar orientation to the carboxylate group in DANA.  Figure 6B shows compound 16 in binding mode A2. It is noteworthy that, whereas the nitro group shows interactions with Arg314, Arg245, and Trp312, the carboxylate and amino groups have additional interactions with Gln195, Trp120, and Asp59, respectively. It is worth noting that all nitro benzoic acids and nitro benzoates are classified among the most active compounds in this series. Moreover, compound 16 has a higher score value and good inhibition as compared with its non-nitrated analog 15. Therefore, the results suggest that the nitro group plays an important role in binding TcTS.
The binding mode B observed in this study differs from the DANA crystal structure (opposite binding conformation), and has not been previously described ( Figure 5B). Binding mode B is associated with compounds 10, 11, 13, 14, 21 and 24 in their best-scored conformation; other compounds that exhibit binding mode B also have good scores, but not the best. Arg93 and Trp120 act as hydrogen donors to bind the carboxylate group at the position 1. These interactions can also be found involving a nitro group as a hydrogen acceptor. A variation of binding mode B was found where the amine group or derivatives at the position 4 form an interaction with Arg93 and Asp59. Figure 6C,D shows examples of compounds in binding mode B.
The third binding mode (C) did not show a specific binding conformation, and it was different from modes A and B. It is associated with compounds 20, 22 and 23; other compounds that exhibit binding mode C have good scores, but not the best. These compounds do not involve the characteristic interactions observed in the A and B modes or are binding on the exterior of the cavity ( Figure 5C). Table S1 (Supplementary Materials) shows the binding mode distribution for the best nine hits for each compound. It is noteworthy that some compounds exhibit two and even more binding modes with similar scores; therefore, the biological effect observed can be associated with several binding orientations.  Figure 6B shows compound 16 in binding mode A2. It is noteworthy that, whereas the nitro group shows interactions with Arg314, Arg245, and Trp312, the carboxylate and amino groups have additional interactions with Gln195, Trp120, and Asp59, respectively. It is worth noting that all nitro benzoic acids and nitro benzoates are classified among the most active compounds in this series. Moreover, compound 16 has a higher score value and good inhibition as compared with its non-nitrated analog 15. Therefore, the results suggest that the nitro group plays an important role in binding TcTS.
The binding mode B observed in this study differs from the DANA crystal structure (opposite binding conformation), and has not been previously described ( Figure 5B). Binding mode B is associated with compounds 10, 11, 13, 14, 21 and 24 in their best-scored conformation; other compounds that exhibit binding mode B also have good scores, but not the best. Arg93 and Trp120 act as hydrogen donors to bind the carboxylate group at the position 1. These interactions can also be found involving a nitro group as a hydrogen acceptor. A variation of binding mode B was found where the amine group or derivatives at the position 4 form an interaction with Arg93 and Asp59. Figure 6C,D shows examples of compounds in binding mode B.
The third binding mode (C) did not show a specific binding conformation, and it was different from modes A and B. It is associated with compounds 20, 22 and 23; other compounds that exhibit binding mode C have good scores, but not the best. These compounds do not involve the characteristic interactions observed in the A and B modes or are binding on the exterior of the cavity ( Figure 5C). Table S1 (Supplementary Materials) shows the binding mode distribution for the best nine hits for each compound. It is noteworthy that some compounds exhibit two and even more binding modes with similar scores; therefore, the biological effect observed can be associated with several binding orientations.

Chemistry: General Procedure
Compounds 10 (4-aminobenzoic acid, A9878), 11 (4-hydrazinobenzoic acid, 246395), 13 (4-nitrobenzoic acid, 461091), 21 (4-amino 2-hydroxybenzoic acid, A79604), and 23 (2-hydroxybenzoic acid, 247588) were purchased from Sigma-Aldrich, Mexico city, Mexico, and used without further purification. In all compounds, synthesized melting points were determined on a Mel-Temp capillary apparatus (Electrothermal, Staffordshire, UK) and are uncorrected. Infrared spectra were recorded using a Bruker Alpha FT-IR spectrometer (AXS Inc., Madison, WI, USA). The 1 H-NMR spectra were obtained in CDCl 3 or DMSO-d 6 (14) (1.5 g, 9 mmol) was added to a mixture (1:1) of acetic acid and acetic anhydride (20 mL), stirred, and the reaction mixture was refluxed for 15 min. After completion of the reaction, the mixture was poured into ice-cooled water and a solid residue was obtained after filtration. The crude was washed three times with 100 mL H 2 O to remove excess acid. The crude was recrystallized in EtOH, yielding (15)  For 4-Amino-3-nitrobenzoic acid (16): 4-acetamido-3-nitrobenzoic acid (17) (2 g, 8 mmol) was taken in a reaction flask, H 2 SO 4 (30 mL) was added dropwise with stirring for 15 min, and the mixture was heated for 15 min at 100 • C. After the completion of the reaction, the mixture was poured into ice-cooled water and a solid residue was obtained after filtration. The crude was washed three times with 100 mL H 2 O to remove excess acid. The crude was recrystallized in EtOH, yielding (16)  For 4-Acetamido-3-nitrobenzoic acid (17): 4-acetamidobenzoic acid (24) (2.5 g, 14 mmol) was added slowly to a mixture (1:1) of HNO 3 and conc. H 2 SO 4 (40 mL) with stirring for 10 min at 0 • C. After, the mixture was stirred for 30 min at room temperature. The reaction mixture was poured into ice water and neutralized by adding 15 mL of 5% Na 2 CO 3 . The product was obtained by filtration and washed with an excess of H 2 O. The crude was purified by recrystallization with ethanol, yielding (17  For 4-(4-Methoxybenzamido)-3-nitrobenzoic acid (20): 4-amino-3-nitrobenzoic acid (16) (0.60 g, 3 mmol), 4-methoxybenzoyl chloride (0.90 g, 4 mmol), and Et 3 N (1.0 mL) were dissolved in dry CH 2 Cl 2 (40 mL) and the reaction mixture was stirred for 48 h at room temperature. The crude mixture was purified by column chromatography on silica gel in n-hexane/EtOAc (7:3), yielding (20) (21) (1 g, 6.5 mmol) was added to a mixture (1:1) of acetic acid and acetic anhydride (20 mL), stirred, and the reaction mixture was refluxed for 15 min. After the completion of the reaction, the mixture was poured into ice-cooled water and a solid residue was obtained after filtration. The crude was washed three times with 100 mL H 2 O to remove excess acid. The crude was recrystallized in EtOH, yielding (22)  For 4-Acetamidobenzoic acid (24): 4-aminobenzoic acid (10) (4 g, 29 mmol) was added to a mixture (1:1) of acetic acid and acetic anhydride (20 mL), stirred, and the reaction mixture was refluxed for 15 min. After the completion of the reaction, the mixture was poured into ice-cooled water and a solid residue was obtained after filtration. The crude was washed three times with 100 mL H 2 O to remove excess acid. The crude was recrystallized in MeOH, yielding (24)

Trypanocidal Activity
In vitro studies were carried out using two strains of trypomastigotes of Trypanosoma cruzi: NINOA and INC-5. CD-1 Mice (18-20 g) were inoculated intraperitoneally with 1 × 10 6 /mL of blood trypomastigotes (0.2 mL). Blood was obtained by cardiac puncture of mice infected with trypomastigotes at the peak of parasitemia, using heparin as an anticoagulant. Blood was treated with isotonic saline (NaCl 0.85%) to adjust to a concentration of approximately 2 × 10 6 trypomastigotes/mL and 195 µL of blood and 5 µL of treatment was placed in 96-well plates. Treatments consisted of a negative control, containing dimethyl sulfoxide (DMSO 2.5%) and compounds derived from benzoic acid, and nifurtimox and benznidazole dissolved in DMSO at the following concentrations: 200 µg/mL, 100 µg/mL, 50 µg/mL, 25 µg/mL, and 12.5 µg/mL, to get the lysis concentration of 50% of the population (LC 50 ). LC 50 values were determined using a Probit statistical analysis of the dose-response, and the results are expressed as the mean ± standard deviation. Finally, crystal violet (1 µg/mL) for the lysis of the trypomastigotes and as witness wells containing 200 µL of blood without receiving any treatment were used. Each concentration was tested in triplicate.
Once the compounds derived from benzoic acids after being homogenized with the blood were added, the plates were incubated at 4 • C for 24 h. After the incubation, the plates were kept at room temperature for 30 min and then an aliquot of 5 mL of each well was taken, which was placed between a slide and a coverslip, and viable trypomastigotes were counted using the method of Brener [51] supplemented with Pizzi. The results were later converted to micromolar units.

Enzymatic Inhibition Assays
Inhibition was assessed using the continuous fluorimetric assay described by Douglas and co-workers [52]. The assay was performed in triplicate (and on three different days) in 96-well plates containing phosphate buffer solution at pH 7.4 (25 µL), a recombinant enzyme solution (25 µL), and an inhibitor solution (25 µL of a 4.0 mM solution). This mixture was incubated for 10 min at 26 • C followed by the addition of MuNANA (Km = 0.68 mM; 25 µL of a 0.4 mM solution giving an assay concentration of 0.1 mM). The final concentration of the tested compounds was 1.0 mM, and the positive control was pyridoxal phosphate. The fluorescence of the released product (Mu) was measured after 10 min, with excitation and emission wavelengths of 360 and 460 nm, respectively, and the data were analyzed with GraphPad Prism software version 4.0 (San Diego, CA, USA). Inhibition percentages were calculated by the equation: % I = 100 × [1 − (V i /V 0 )], where V i is the velocity in the presence of the inhibitor, and V 0 is the velocity in the absence of the inhibitor.
Protein preparation: The TcTS were obtained from the Protein Data Bank with the PDB accession number 1MS8 [53]. The structure was prepared using Maestro 9.1 [56], first, the chain A was selected and the ligands, solvent, and other molecules were removed. Missing side chains were added and alternative side chains were defined using Maestro. Then, the .pdb structure was submitted to minimization using the YASARA web server [57]. The optimized structure was exported to AutoDock Tools 1.5.6 in .pdb format to generate a .pdbqt file.
Docking studies: AutoDock Vina 5.6 was used to predict the mode of interaction for each ligand within the active site of TcTS [58]. A grid box of x, y, and z dimensions was set to 60, 60, 60 angstroms centered to the DANA binding site in the original crystal structure. Each compound was set to run 100 dockings, and the best nine conformations were retrieved for the analysis. The results were analyzed employing PyMOL v0.99 and UCSF Chimera 1.11rc [59].

Conclusions
In this work, the benzoic acid derivatives 14, 18 and 19 clearly showed more potent trypanocidal activity than the commercially available drugs beznidazole and nifurtimox towards the NINOA and INC-5 strains of T. cruzi. It is noteworthy that compound 18 showed nanomolar trypanocidal activity against the NINOA strain (20 nM), whereas compounds 11 and 17 displayed similar TcTS inhibition to pyridoxal and compound 16 showed the best inhibitory activity.
The respective docked structures of the compounds showed three different binding patterns according to DANA crystal structure in the active site cavity. Model A is similar to DANA interaction in the cavity, model B represents the opposite binding conformation, and model C is interactions outside the cavity or that do not involve the characteristic interactions observed in the A and B modes.
The benzoic acid derivatives (10-24) evaluated in the present work for the treatment of Chagas disease by the inhibition of TcTS or alternative biological mechanisms (trypanocidal activity) reinforce the development of more effective candidates of this disease. Therefore, we suggest ethyl 4-acetamido-3-nitrobenzoate 18 as a prototype for the development of more effective TcTS inhibitors against Chagas disease, which shows a moderate inhibition (47%), a binding model similar to DANA (pattern A), and significant trypanocidal activity (LC 50 values of 0.02 and 0.22 µM against the NINOA and INC-5 strains, respectively).

Supplementary Materials:
The following are available online. Table S1: Binding modes and docking scores for compounds 10-24.