Repurposing of Inhibitors of Plasmodial Aspartate Transcarbamoylase Toward Trypanosoma Cruzi

TL;DR

Scientists found that some compounds originally designed to target a specific enzyme in malaria parasites can also inhibit a similar enzyme in a parasite that causes Chagas disease, suggesting a new path for drug development.

Contribution

A library of Plasmodium falciparum ATC inhibitors was repurposed to identify potent and selective inhibitors of Trypanosoma cruzi ATC.

Findings

34 compounds from a 70-member library inhibited Trypanosoma cruzi ATC by over 90%.

Five compounds showed IC50 values below 250 nM, indicating strong inhibitory activity.

Two compounds demonstrated low micromolar cellular activity but exhibited co-toxicity.

Abstract

Aspartate transcarbamoylase (ATC) catalyzes the committed and rate‐limiting step in the pyrimidine de novo biosynthesis pathway. While previously suggested to be a potential target for antimalarial, antitubercular, and antioncologic drug discovery, we hypothesized that an existing compound library of ATC inhibitors designed from one scaffold by fragment screening against Plamodium falciparum ATC (PfATC) may also contain inhibitors of Trypanosoma cruzi ATC (TcATC). In this manuscript, we screened the 70‐member library at 35 μM against 50 nM TcATC, and in these initial experiments, 34 compounds showed over 90% inhibition. Of the 34 compounds, 5 compounds demonstrated IC50 values of lower than 250 nM in a follow‐up enzymatic half inhibition concentration analysis. Kinetic studies on one of these compounds indicate that they inhibit TcATC in a noncompetitive manner, and a druggable allosteric pocket is seen in the available crystal structure. While cocrystallization and soaking experiments were unsuccessful, molecular modeling was performed to assess potential binding modes. Two of the best‐performing compounds were selected for a cellular assay, showing EC50s of 7.4 µM and 6.5 µM. However, significant cotoxicity was also observed, demonstrating that further elaboration of the compounds is necessary. These results suggest that this compound library might be a starting point for antitrypanosomatid drug discovery.

Aspartate transcarbamoylase (ATC) catalyzes the committed, rate‐limiting step in pyrimidine de novo biosynthesis and has been proposed as a drug target in several pathogenic contexts. Here, screening a 70‐member ATC inhibitor library originally developed against Plasmodium falciparum ATC identified 34 compounds with ⟩90% inhibition of Trypanosoma cruzi ATC, including five with IC50 values below 250 nM, which likely act via non‐competitive inhibition consistent with a druggable allosteric site. Although selected compounds showed low‐micromolar cellular activity, associated co‐toxicity indicates that further optimization is required, positioning this library as a promising starting point for anti‐trypanosomatid drug discovery.

Figures

(A) Initial BDA series potency screening. 35 μM of each compound was tested against 50 nM TcATC with 30 mM ASP and 2 mM CP. All experiments were performed in triplicate and data are presented as mean ± standard deviation (SD). Compounds displaying over 90% inhibition compared to the control were selected for further analysis. 34 compounds out of 70 BDA series compounds were selected for further inhibition tests. (B) BDA series IC50 values. Selected compounds were tested for the determination of IC50 inhibition test against 50 nM TcATC with 30 mM ASP and 2 mM CP. 17 compounds in the BDA library presented IC50 values lower than 1 μM IC50 values. Overall, the BDA series showed a good inhibition against TcATC. Out of the 17 compounds, BDA‐06, −07, −14, −33, −41 (highlighted in red) were the most potent inhibitors, with IC50 values of 160, 291, 280, 244, and 207 nM, respectively. (C) Structure of the BDA scaffold.

Subsequent IC50 analysis of the PALA and the best 5 inhibitors. Selected compounds were tested for the determination of IC50 inhibition test against 50 nM TcATC with 30 mM ASP and 2 mM CP. PALA demonstrate an IC50 of 11,6 μM, while compounds BDA‐6, BDA‐7, BDA‐14, BDA‐33, and BDA‐41 are shown to have IC50 values of 160, 291, 280, 244 and 198 nM, respectively and their molecular structures are shown in the inset.

Comparison of BDA series effectiveness across different species. The average logarithm value of each species is taken as the baseline. The values were calculated as the logarithm of the ratio between each individual value and the average value, where blue indicates stronger than average inhibition and red indicates worse than average inhibition. Gray indicates either IC50 is not measurable or not measured because of poor inhibition during the compound screening. BDA‐05 and BDA‐46 seem to show more specificity towards T. cruzi.

Enzyme kinetics with fixed substrate concentrations. Panel (A) shows the effect of BDA‐41 on enzyme activity at a fixed CP concentration of 2 mM, while Panel (B) shows the effect at a fixed ASP concentration of 15 mM. In both cases, BDA‐41 reduces the reaction rate without altering substrate concentration, consistent with a noncompetitive mode of inhibition.

(A) Molecular blind docking against the homology model of TcATC and protein–ligand interactions of potential hit BDAcompounds. Chains A, B, and C of the proteins shown in cartoon representation are colored in yellow, pale green, and pale cyan, respectively. The ligands are shown in sticks, the 120’s loop in red, the active sites are in spheres, and the allosteric pockets are shown in blue surface representation. The 2D protein‐ligand interactions of each ligand are further shown around the structure. The dashed lines represent the interactions between the ligand and the protein, while the colored circles indicate the specific protein residues involved in these interactions. Each color corresponds to a different type of interaction. Green = Conventional Hydrogen Bond, Orange = Pi‐Cation/Pi‐Anion interaction, Purple = Pi‐Sigma interaction, Pink = Alkyl/Pi‐alkyl interaction. The colored arrows link each ligand from its binding site in the protein to its interaction diagram using matching color coding. (B) Sequence alignment of TcATC (Uniprot ID: O15636) and HsATC (Uniprot ID: P27708) sequence. All red residues indicate conserved residues. The highlighted red residues are 100% conserved.

Full text

Introduction

1

Pyrimidine nucleotides act as essential molecules for all living organisms for their critical role in cellular proliferation, DNA, and RNA production, and so on. [1]. Pyrimidine nucleotides can be synthesized through the salvage pathway [2] or generated via a de novo biosynthesis pathway [3]. The distribution and reliance of organisms upon the two pathways vary depending on both the type and stage of cells [4]. Generally, the salvage pathway is sufficient for resting or fully differentiated cells, while de novo biosynthesis is required for proliferating cells [5]. However, some parasites, such as Plasmodium species, lack the salvage pathway [6]. Therefore, inhibiting the de novo biosynthesis pathway has been proposed to provide a potential treatment for various diseases [7, 8, 9].

The de novo biosynthesis of pyrimidine nucleotides begins with ATP, glutamine, and bicarbonate and involves six enzymes. Aspartate transcarbamoylase (ATC), the second enzyme in this pathway, catalyzes the reaction between L‐aspartate (ASP) and carbamoyl phosphate (CP) to form carbamoyl‐aspartate. ATC has been a popular target for drug discovery by inhibiting the de novo pyrimidine biosynthesis, as it is well‐characterized on a structural and enzymatic level and often considered as both the first committed and rate‐limiting enzyme in the pathway (recently reviewed in [10]). Taking Escherichia coli ATC (EcATC) as an example, the enzyme is a heterododecamer composed of two catalytic trimers and three regulatory dimers. Each catalytic monomer contains a single substrate‐binding site located at the interface between adjacent subunits within the trimer [11]. Naturally, EcATC exists in two distinct states: a high‐affinity relaxed (R) state and a low‐affinity tight (T) state. The binding of substrates induces a conformational change, shifting the enzyme from the ‘T’ state to the ‘R’ state, whereas the binding of cytidine triphosphate (CTP), the final product of pyrimidine de novo biosynthesis, stabilize the protein in its T state [11]. N‐phosphonacetyl‐L‐aspartate (PALA), the transition state intermediate analog, was reported as a potent competitive inhibitor against EcATC and human ATC [12] but failed as an antitumor drug during clinical trial due to its inefficiency, toxicity, and rapid development of resistance [13]. In eukaryotes such as human, yeast, ATC along with carbamoyl phosphate synthetase II (CPS II) and dihydroorotase (DHO) form a large multifunctional complex where CPS II is associated with the regulation [14]. Although the regulatory subunit is not present in plant ATC, Bellin et al. determined the structure of plant ATC bound with UMP, demonstrating that Arabidopsis ATC can be regulated by uridine 5‐monophosphate (UMP) in a feedback inhibition mechanism [15, 16]. It is worth mentioning that bacterial ATC is divided into three distinct classes: class B, represented by EcATC, is the most common one where the two copies of the catalytic homotrimers and three copies of the regulatory dimeric subunits form a complex of around 300 kDa [11]; class C including Bacillus subtilis ATC is the simplest form that only has three copies of the catalytic chains; class A is the least understood and the biggest complex (about 500 kDa) formed by two copies of the catalytic homotrimers and three copies of dimers which resembles the downstream enzyme DHO but lack of DHO activity [17]. Despite the differences in the regulation of ATC in different species, the catalytic subunit of ATC shares a large degree of similarity [10].

Our laboratory has performed a structure‐based fragment screening to identify inhibitors of Plasmodium falciparum ATC (PfATC), which resulted in a 70‐compound library (referred to as the BDA series, (2‐amino‐Boc substituted 5‐aroyl thiophene‐3‐carboxamide)) and the discovery of an allosteric pocket which is close to the active sites [18]. The core scaffold consists of a thiophene aryl‐acetamide linker connecting two aromatic moieties (R^1^ and R^2^ positions) that can accommodate diverse hydrophobic or polar substituents (Figure 1C). This modularity allowed systematic tuning of shape and polarity to match the hydrophobic cleft adjacent to the carbamoyl phosphate–binding loop (‘120s loop’). Compounds with small polar groups at R^1^ and extended aromatic or cycloalkyl substituents at R^2^ were found to stabilize the enzyme's low‐activity (structurally similar to the observed EcATC T‐state) conformation through π‐cation and hydrogen‐bond interactions with residues near Ser63, Arg64, and Glu94. Thus, the BDA scaffold serves as a privileged allosteric chemotype capable of inducing conformational stabilization across ATC homologs. The best performing compound, BDA‐04, shows an IC_50_ of 45.6 nM and a K d of 66.3 nM in in vitro experiments against PfATC, while also displaying an EC_50_ against blood stage of Plasmodium 3D7 cultures of 2 μM. Additionally, it shows selectivity between malarial and normal human cells. Subsequently, we tested the library on tuberculosis and human ATC, demonstrating the potential for inhibition of the de novo pyrimidine biosynthesis as an avenue for antiinfective and oncolytic drug development [19, 20]. In both cases, the compounds showed low nM to μM inhibition and the associated kinetic data support the noncompetitive inhibition mechanism of the compound series displayed against plasmodial ATC. These findings lead naturally to the potential for ATC inhibition in the treatment of further diseases. The potential for ATC inhibition in the treatment of other diseases has been recently reviewed [10], and this study confirms that the BDA series indeed contains hit molecules that may be used to address Chagas disease.

(A) Initial BDA series potency screening. 35 μM of each compound was tested against 50 nM TcATC with 30 mM ASP and 2 mM CP. All experiments were performed in triplicate and data are presented as mean ± standard deviation (SD). Compounds displaying over 90% inhibition compared to the control were selected for further analysis. 34 compounds out of 70 BDA series compounds were selected for further inhibition tests. (B) BDA series IC50 values. Selected compounds were tested for the determination of IC50 inhibition test against 50 nM TcATC with 30 mM ASP and 2 mM CP. 17 compounds in the BDA library presented IC50 values lower than 1 μM IC50 values. Overall, the BDA series showed a good inhibition against TcATC. Out of the 17 compounds, BDA‐06, −07, −14, −33, −41 (highlighted in red) were the most potent inhibitors, with IC50 values of 160, 291, 280, 244, and 207 nM, respectively. (C) Structure of the BDA scaffold.

Chagas disease, one of the neglected tropical diseases, is caused by the parasite Trypanosoma cruzi and is primarily endemic to the Americas. Currently, it is estimated that between 6–7 million people are infected with Chagas disease, leading to ≈12,000 deaths annually [21]. While animal reservoirs contribute to the potential spread of infection, factors such as global warming [22] and globalization are increasing the risk of Chagas disease becoming a worldwide concern. The current treatments, benznidazole and nifurtimox, are effective during the initial acute phase of Chagas disease, which is often asymptomatic or presents with mild flu‐like symptoms. However, these treatments show limited efficacy in the chronic stage of the disease. As a result, exploring new therapeutic options for Chagas disease is crucial. Similar to malarial parasites, the replication of T. cruzi after infection requires a substantial amount of pyrimidine nucleotides. Studies have shown that T. cruzi primarily relies on the de novo pyrimidine biosynthesis for proliferation [23], and the knockdown of CPS II significantly impairs parasite growth [24]. Moreover, Matoba et al. resolved the structure of T. cruzi ATC (TcATC) in it's ligand‐free (PDB: 6JKQ) and substrate bound form (6JKS), structures that are a crucial step towards structure‐based drug design against Chagas disease [25].

Given these facts, we hypothesized that the BDA series library may present allosteric inhibitors for TcATC. In this manuscript, we evaluated the inhibition of the BDA series against TcATC in vivo and in vitro and described a model for the binding mode of the most potent molecule that provides a basis for further compound elaboration. We have also performed in vitro validation of the molecule that demonstrates inhibitory effects of the molecule on T. cruzi proliferation in human cell cultures, along with a strong cytotoxic impact on the host cell system. In summary, this report provides a starting point for the development of more potent TcATC inhibitors, which could then potentially have lower off‐target impact on the host cell and provide a lead towards developing a treatment of Chagas disease.

Results and Discussion

2

BDA Series Inhibits TcATC

2.1

Following the hypothesis that the BDA series could potentially inhibit TcATC, we performed an in vitro enzymatic activity screening using 35 μM BDA compounds against 50 nM TcATC (Figure 1). 34 of 70 compounds tested displayed an inhibition of over 90%, again demonstrating that the BDA series contains a range of high‐potency compounds against ATC homologs from different species. These 34 compounds were selected for further analysis, specifically the determination of the half maximal inhibitory concentration (IC_50_). Of the 34 compounds tested, 17 showed an IC_50_ of less than 1 μM IC_50_ with the best‐performing compounds BDA‐6, BDA‐7, BDA‐14, BDA‐33, and BDA‐41, providing IC_50_ values of 160, 291, 280, 244 and 207 nM, respectively (Figure 2). Under the same conditions, PALA was also tested and showed an IC_50_ of 10 μM (Figure 2). The comparison of the effectiveness of the BDA‐series across malaria, human, tuberculosis, and T.cruzi ATC indicates that, while less potent, BDA‐05 and BDA‐46 might have a stronger selectivity towards TcATC (Figure 3). Compared to the core structure, both BDA‐5 and BDA‐46 have a cyclohexyl ring at the R^1^ position and a phenethyl group at the R^2^ position, providing a potential core scaffold specific to the TcATC homolog and providing a starting point for further optimization. As the BDA series is known to target an allosteric pocket in other organisms, these data lead us to hypothesize that the BDAs identified above would target a similar low sequence conservation pocket in TcATC.

Subsequent IC50 analysis of the PALA and the best 5 inhibitors. Selected compounds were tested for the determination of IC50 inhibition test against 50 nM TcATC with 30 mM ASP and 2 mM CP. PALA demonstrate an IC50 of 11,6 μM, while compounds BDA‐6, BDA‐7, BDA‐14, BDA‐33, and BDA‐41 are shown to have IC50 values of 160, 291, 280, 244 and 198 nM, respectively and their molecular structures are shown in the inset.

Comparison of BDA series effectiveness across different species. The average logarithm value of each species is taken as the baseline. The values were calculated as the logarithm of the ratio between each individual value and the average value, where blue indicates stronger than average inhibition and red indicates worse than average inhibition. Gray indicates either IC50 is not measurable or not measured because of poor inhibition during the compound screening. BDA‐05 and BDA‐46 seem to show more specificity towards T. cruzi.

BDA‐41 Likely Inhibits TcATC Noncompetitively

2.2

Kinetic assays were performed to confirm the hypothesis that the identified BDA series compounds inhibit TcATC in a noncompetitive manner. BDA‐41 was selected to illustrate the inhibition mechanism, as it is one of the most potent inhibitors, with an IC_50_ value of 198 nM. While the data indicate that BDA‐41 is a noncompetitive in line with BDA inhibition of ATCs from other organisms, further kinetic analysis would be required to demonstrate this unambiguously (Figure 4).

Enzyme kinetics with fixed substrate concentrations. Panel (A) shows the effect of BDA‐41 on enzyme activity at a fixed CP concentration of 2 mM, while Panel (B) shows the effect at a fixed ASP concentration of 15 mM. In both cases, BDA‐41 reduces the reaction rate without altering substrate concentration, consistent with a noncompetitive mode of inhibition.

Molecular Docking Suggests the BDA Series Preferably Binds between the Active Site and Allosteric Pocket of TcATC

2.3

In the absence of an experimentally derived structure for the BDA scaffold, we decided to investigate the binding modes of the best‐performing compounds (BDA‐6, 7, 14, 33, 41) and those with potentially stronger selectivity for the T. cruzi protein (BDA‐5 and 46) against both the parasite protein and its human homolog using molecular blind docking. For each compound, 20 potential docking poses were considered. This approach identified four regions in which the BDA compounds dock in both TcATC and HsATC. These regions are: the active site, the area between the active site and the allosteric site, (a middle pocket between the trimers) and less commonly, an area adjacent to the allosteric site. Although the compounds bind to similar regions in both Tc and human proteins, they show distinct binding preferences: in Tc, mainly between the active site and the allosteric pocket or in the middle pocket of the trimer, and in Hs, in the active site or the middle pocket. Overall, in most docking poses, BDA‐6, 7, 14, 33, and 41 tended to bind between the active site and allosteric pocket of TcATC, whereas BDA‐5 and 46 bound to a pocket located in the middle of the trimer (Tables S1–S7). While these five compounds (BDA‐6, 7, 14, 33, and 41) did not fully occupy either the allosteric pocket or the active site in TcATC in their preferential binding pose, their positioning may stabilize the protein in its T‐like state, thereby interfering with its catalytic activity. Although it is not yet clear how the compounds that were preferentially bound within the trimeric cavity affect the function of protein, taken together, these results suggest a model for inhibition of enzymatic activity.

In HsATC, BDA‐5, 33, 41, and 46 preferentially occupied the middle pocket of the trimer, BDA‐6 and 7 were located in the active site, and BDA‐14 was positioned adjacent to the allosteric pocket (Tables S1–S7). Notably, in all cases, the compounds showed slightly stronger predicted binding energy scores toward TcATC than toward HsATC.

In Figure 5A, we present the docked compound positions in TcATC corresponding to their highest binding scores, which do not necessarily coincide with their preferred binding positions among the 20 potential poses; interestingly, while BDA‐14, BDA‐33, and BDA‐41 bound between the allosteric pocket and active site, with BDA‐33 and BDA‐41 displaying stronger binding energies than BDA‐14, probably due to larger compound size, BDA‐5, BDA‐6, BDA‐7 and BDA‐46 may bind to a pocket located in the middle of the trimer. BDA‐33, in its preferred position, engaged with residues Ser62, Ser63, Arg64, Lys92, and Glu94 through conventional hydrogen bonds and hydrophobic interactions such as π‐sigma, π‐alkyl, and π‐anion (Figure 5). A similar binding profile was also observed with BDA‐41 (Figure 5). We also present all predicted binding poses of BDA‐41 for Tc and Hs in the Figure S2, which again shows that although the compound bind to similar regions in both Tc and human proteins, it has distinct binding preferences, likely due to sequence differences.

Figure 5B shows the sequence conservation between the Tc and HsATC proteins, which may help explain the differences in compound binding preferences. For example, the two structurally similar ligands BDA‐33 and BDA‐41 preferentially bind between the active site and the allosteric pocket in TcATC but bind to the middle pocket in HsATC. This shift may be influenced by residue differences: in the active‐site region, Leu288 in Tc is replaced by Met2185 in the human homolog. Additionally, three residues differ in the allosteric pocket (Ser63, Thr99, and Leu103 in Tc versus Thr1974, Ser2000, and Met2014 in Hs).

(A) Molecular blind docking against the homology model of TcATC and protein–ligand interactions of potential hit BDAcompounds. Chains A, B, and C of the proteins shown in cartoon representation are colored in yellow, pale green, and pale cyan, respectively. The ligands are shown in sticks, the 120’s loop in red, the active sites are in spheres, and the allosteric pockets are shown in blue surface representation. The 2D protein‐ligand interactions of each ligand are further shown around the structure. The dashed lines represent the interactions between the ligand and the protein, while the colored circles indicate the specific protein residues involved in these interactions. Each color corresponds to a different type of interaction. Green = Conventional Hydrogen Bond, Orange = Pi‐Cation/Pi‐Anion interaction, Purple = Pi‐Sigma interaction, Pink = Alkyl/Pi‐alkyl interaction. The colored arrows link each ligand from its binding site in the protein to its interaction diagram using matching color coding. (B) Sequence alignment of TcATC (Uniprot ID: O15636) and HsATC (Uniprot ID: P27708) sequence. All red residues indicate conserved residues. The highlighted red residues are 100% conserved.

BDA‐33 and BDA‐41 Inhibit the Proliferation of T. cruzi

2.4

Two compounds, BDA‐33 and BDA‐41, were further selected for efficacy evaluation in cell culture on their potency in enzymatic assays and their availability in our laboratory. The molecules were tested against intracellular T. cruzi amastigotes in a standard in vitro model using rat L6 cardiomyoblasts as host cells [26]. BDA‐33 and BDA‐41 had 50% inhibitory concentrations (IC_50_) against T. cruzi of 7.4 and 6.5 µM, respectively, not significantly higher than the reference drug benznidazole (IC_50_ of 2.1 µM) (Figure S2). However, the BDA molecules also showed cytotoxicity against uninfected L6 cells, with an IC_50_ of around 11 µM (Table S8). Therefore, the efficacy data against intracellular T. cruzi are not conclusive, and further optimization of the BDA series towards higher selectivity against TcATC over its mammalian orthologues is required.

Materials and Methods

3

TcATC Cloning, Expression and Purification

3.1

The TcATC gene (Trypanosoma cruzi tcact2 gene for aspartate carbamoyltransferase, Gene ID: AB074139.2) was ordered from Eurofins with the modification of the N‐terminus to convert the translation product from MLEL to MVEL required by the cloning strategy. It was then cloned into pET‐M11 with an N‐terminal His6‐tag from the restriction sites HindIII and NcoI. The recombinant plasmid was introduced into E. coli BL21 Star (DE3). A single colony from transformation was inoculated into 10 mL LB supplemented with 50 μg/mL Kanamycin at 310 K, incubated overnight, then grown at 1 L TB under the same conditions until OD_600_ reached 0.8. 0.5 mM IPTG was added to induce the expression at 291 K for 16 h. Cells were harvested by centrifuging at 5 K rpm for 30 min, followed by resuspension in 90 mL lysis buffer (50 mM Tris‐HCl, pH 8, 50 mM imidazole, 300 mM NaCl, 5% (v/v) glycerol, 3 mM β‐Mercaptoethanol (BME), 30 μg/ mL lysozyme). After lysis with sonication, the lysate was obtained by centrifuging at 16,000 g for 60 min.

The supernatant was filtered using a 0.45 μm membrane and loaded onto a 5 mL His‐trap High Performance column (Fischer Scientific) after equilibration with buffer A (50 mM Tris‐HCl, pH 8, 300 mM NaCl, 5% (v/v) glycerol, 3 mM BME). The His6‐tagged protein was eluted with a gradient of buffer B (buffer A containing 500 mM imidazole). Fractions were pooled and dialyzed at 4°C in dialysis buffer (50 mM Tris‐HCl, pH 8, 300 mM NaCl, 5% (v/v) glycerol, 3 mM BME) to cleave the His‐tag through the addition of a 5% molar ratio of TEV (tobacco etch virus) protease. Untagged TcATC was purified using the His‐trap column to remove the TEV‐protease and uncleaned TcATC. TcATC was then concentrated to 1 mL using a centrifugal concentrator with a 10 kDa cutoff (Millipore). The concentrated protein was then purified with a size‐exclusion Superdex 75 16/60 HiLoad column (GE Healthcare) pre‐equilibrated with 50 mM Tris‐HCl, pH 8, 50 mM NaCl, 5% (v/v) glycerol, 3 mM BME. The equate was pooled from a single peak and concentrated to 10 mg/mL, determined by the absorbance at 280 nM [ABS 0.1% (1 mg/mL) = 0.267]. Finally, the purified protein was flash‐frozen in liquid nitrogen and stored at −70 °C. Protein purity was assessed by SDS‐PAGE. The specific enzyme activity of the purified protein, determinedusing the method described below, was 72.5 U/mg.

Activity Assays

3.2

Enzymatic activity was measured according to a colorimetric method in 96‐well plates adapted from [27] with modifications. Briefly, 50 nM TcATC was tested in 50 mM Tris‐Acetate (pH 8) in a final volume of 150 μL. Aspartate (ASP) and carbamoyl phosphate (CP) saturation and kinetic curves were monitored with a fixed concentration of 2 mM CP and 15 mM ASP. TcATC was first incubated with ASP and compounds for 10 min in the shaker at room temperature. The reaction was started by adding CP into the mixture and stopped by adding 100 μL colometric mixture, which is a mixture of two volumes of 26.5 mM 2,3‐dimethyl‐1‐phenyl‐3‐pyrazolin‐5‐one in 50% (v/v) H_2_SO_4_ and one volume of 80 mM 2,3‐butanedione monoxime in 5% (v/v) acetic acid. After overnight incubation in the dark, the plate was heated to 50° for 20 min and cooled at room temperature for 40 min before reading the absorbance at 466 nM. Data was analyzed using Microsoft Excel, R, and GraphPad Prism.

Homology Modeling

3.3

The 3D structures of ATC of T. cruzi (PDB ID: 6JKQ) and H. sapiens (PDB ID: 5G1O) were retrieved from the Protein Data Bank (PDB) [28]. Missing residues in the 120 s loop were modeled using MODELLER version 10.5 [29], with templates 8BPS and 1EZZ. From the 100 generated models, the best ones based on z‐DOPE scores of −1.47 for TcATC and −1.31 for HsATC, were selected. The models were further validated using Verify3D [30], ProSA‐Web [31], and ProCheck [32] to assess modeling quality.

Ligand Docking

3.4

Blind docking of seven BDA compounds (5, 6, 7, 14, 33, 41, 46) to the T‐state of TcATCase and HsATCase was done using AutoDock Vina [33] with an exhaustiveness of 300, an energy range of 4 and 20 binding modes at the Center for High Performance Computing (CHPC, Cape Town, South Africa). The docking parameters were validated through the redocking of PALA to the E. coli ATCase crystal structure retrieved from the PDB [28] with PDB ID 1Q95. The protein was protonated using H++ [34] at a pH of 8.0 and converted to pdbqt using pre‐pare_receptor4.py script from AutoDock tools [35]. The ligands were retrieved in SMILE format and converted to 3D using an in‐house Python script that uses RDKit (RDKit: Open‐source cheminformatics. https://www.rdkit.org) for the conversion, addition of explicit hydrogens, and optimization of the ligands (using the Merck Molecular Force Field). The ligands were further converted to pdbqt using Babel [36]. The grid box parameters of the proteins were retrieved from the UCSF Chimera‐integrated Autodock Vina plugin [37] with the following dimensions: TcATCase (Center x, y, z : ‐5.94303, 7.22703, 2.42273 and Size x, y, z: 95.0529, 108.48, 93.9907), HsATCase (Center x, y, z: 15.9427, −6.06953, 25.235 and Size x, y, z: 99.9618, 81.2261, 100.112). The results were further analysed through binding energies, binding pockets, and protein–ligand interactions using Discovery Studio [38].

Cytotoxicity Assay

3.5

L6 rat skeletal myoblasts were cultivated in RPMI 1640 medium supplemented with 1% L‐glutamine (200 mM) and 10% fetal bovine serum. The cultures were maintained at 37 °C in an atmosphere of 5% CO_2_. Assays were performed in 96‐well microtiter plates with 2 × 10^3^ L6 cells per well. After 24 h, the test compounds (10 mM stock solutions in DMSO) were added in eleven 3‐fold dilution steps covering a range from 100 to 0.002 μM. After further incubation for 70 h, 10 µL per well of resazurin solution (12.5 mg in 100 mL PBS) were added and the plates incubated for another 2 h [39, 40]. Then they were read with a Spectramax Gemini EM microplate fluorometer (Molecular Devices) at 536 nm excitation and 588 nm emission. IC_50_ values were calculated from the dose–response curves by linear regression using Microsoft Excel [41]. Two independent replicates of the assay were performed, each in duplicate. Podophylotoxin was used as a positive reference compound.

Cruzi Assay

3.6

Mammalian stages of T.cruzi Tulahuen strain C2C4 containing a β‐galactosidase reporter (LacZ gene [42]) were routinely maintained in Mouse Embryo Fibroblasts (MEF) at 37 °C, 5% CO_2_, in RPMI 1640 medium supplemented with 1% L‐glutamine (200 mM) and 10% fetal bovine serum. Drug testing was done with L6 rat skeletal myoblasts as host cells. The assays were performed in 96‐well microtiter plates with 10^3^ L6 cells/well. After 24 h, trypomastigote T. cruzi were added (5 × 10^3^/well). 48 h after infection, the medium was replaced with 100 μL of fresh medium containing test compounds in serial dilution from 100 to 0.002 μM (prepared in a separate plate from 10 mM stock solutions in DMSO). After further incubation for 96 h, the substrate CPRG/IGEPAL CA 630 (50 μL/well) was added. A color reaction developed within 2–6 h and was read photometrically at 540 nm using the TECAN Sparks (Tecan). IC_50_ values were calculated from the dose–response curves by linear regression using Microsoft Excel [30]. Two independent replicates of the assay were performed, each in duplicate. Benznidazole was used as a positive reference compound.

Conclusion

4

This manuscript describes the screening of a 70‐member compound library developed for use against PfATCase against the homologous enzyme in T. cruzi, TcATC (Figure 1). As predicted, the library developed in the search for a noncompetitive inhibitor of PfATCase also contained potent inhibitors of TcATCase (Figure 1), with in vitro IC_50_s of low triple‐digit nanomolar being seen (Figure 2). As the BDA library is poised at an allosteric pocket of the plasmodial enzyme, which displays a greater degree of sequence variation than the residues of the active site pocket, it was also expected that the best‐performing BDA molecules showed a degree of selectivity between species (Figure 3). However, the best compounds display in vitro IC_50_ values of 160, 291, 280, 244, and 198 nM, respectively, making them suitable to be assessed as hit molecules in subsequent experiments. The subsequent kinetic assay suggests that BDA‐41 is a noncompetitive inhibitor (Figure 4).

Cellular assays demonstrated that BDA‐33 and ‐41 inhibit T. cruzi cultures, albeit with significant toxicological impacts. Thus, reusing the BDA series in drug development targeting Chagas disease is feasible with further optimization to reduce the toxicity and increase the potency and specificity. It should be stressed that the current molecules do not show sufficient selectivity between the parasite and human host, and significant improvement in both potency and selectivity is required.

Subsequent elaboration of the compound would greatly benefit from the availability of structural information and would greatly aid in elucidating the binding mode and guiding future hit‐to‐lead optimization. TcATC was crystallized and diffracted to 1.8 Å (data not shown). However, attempts to obtain costructures of TcATC with BDA series members are currently unsuccessful. Compound BDA‐03 was cocrystallised with TcATC, because of its relative hydrophilicity and the structure was solved at 2.5 Å. However, the electron density of the ligand is weak. We are hesitant to draw conclusions on the binding mode from this data, and further optimization of crystallisation conditions is underway to obtain a clearer structure. Further, subsequent examination of the docking models could not provide a rationale for the failure to obtain cocrystal structures.

However, molecular modeling approaches have provided models of BDA‐binding to the proposed binding site and can be used both to design confirmatory experiments, as well as provide the basis for further molecular elaboration. Future expansion of the BDA scaffold should exploit sequence and structural differences between T. cruzi and human ATC around the allosteric cleft rather than the conserved catalytic core. Comparative modeling identifies residues 60–95 (the CP‐binding loop) and 160–175 (helix α6) as variable regions that define pocket volume and polarity. Substitutions introducing bulkier or more polar R^1^ substituents could enhance shape complementarity with the deeper TcATC pocket while disfavoring the flatter human site. Likewise, restricting overall molecular flexibility through bicyclic or conformationally locked linkers could improve selectivity and reduce nonspecific binding. Combining structure‐guided design with focused virtual screening of BDA analogs ‐ varying amide orientation, ring electronics, and heteroatom donors ‐ should therefore yield TcATC‐selective derivatives with lower host toxicity. Intriguingly, the additional data provide a basis for the design of BDA‐compounds that may be bioactive in a variety of diseases whilst having a limited impact on the human homolog. Development of such pan‐inhibitors may provide significant benefits in the development of treatments for diseases that predominantly impact the Global South.

Supporting Information

Additional supporting information can be found online in the Supporting Information section. Supporting Fig. S1: IC_50_ values of BDA series against TcATC. The chemical structures of the compounds have been reported previously (18). Supporting Fig. S2: BDA‐41 binding poses in (A) TcATC and (B) HsATC. Chains A, B, and C of the proteins shown in cartoon representation are colored in yellow, pale green, and pale cyan, respectively. The ligand poses are shown in sticks, the 120's loop in red, the active sites are in spheres, and the allosteric pockets are shown in blue surface representation. Supporting Fig. S3: Dose‐response curves of compounds BDA‐33 and BDA‐41 against T. cruzi intracellular amastigotes. Dose response curve of BDA‐33 and BDA‐41 on T. cruzi . Panel A shows the effect of BDA‐33, while Panel B shows the effect of BDA‐41. Supporting Table S1: Binding positions and binding energies (BEs) of BDA‐05 in TcATC and HsATC for the top 20 docking poses. The positions highlighted in red indicate the binding pockets most commonly occupied by the docking poses for each protein. Supporting Table S2: Binding positions and binding energies (BEs) of BDA‐06 in TcATC and HsATC for the top 20 docking poses. The positions highlighted in red indicate the binding pockets most commonly occupied by the docking poses for each protein. Supporting Table S3: Binding positions and binding energies (BEs) of BDA‐07 in TcATC and HsATC for the top 20 docking poses. The positions highlighted in red indicate the binding pockets most commonly occupied by the docking poses for each protein. Supporting Table S4: Binding positions and binding energies (BEs) of BDA‐14 in TcATC and HsATC for the top 20 docking poses. The positions highlighted in red indicate the binding pockets most commonly occupied by the docking poses for each protein. Supporting Table S5: Binding positions and binding energies (BEs) of BDA‐33 in TcATC and HsATC for the top 20 docking poses. The positions highlighted in red indicate the binding pockets most commonly occupied by the docking poses for each protein. Supporting Table S6: Binding positions and binding energies (BEs) of BDA‐41 in TcATC and HsATC for the top 20 docking poses. The positions highlighted in red indicate the binding pockets most commonly occupied by the docking poses for each protein. Supporting Table S7: Binding positions and binding energies (BEs) of BDA‐46 in TcATC and HsATC for the top 20 docking poses. The positions highlighted in red indicate the binding pockets most commonly occupied by the docking poses for each protein. Supporting Table S8: In vitro activity and cytotoxicity of compounds against T. cruzi Tulahuen C4 strain

Funding

This study was supported by H2020 Marie Skłodowska‐Curie Actions (860816), China Sponsorship Council.

Patents

The authors have submitted a patent application on the molecules described in the manuscript (21211903.6), which has now lapsed.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary Material