Dual and selective inhibitors of pteridine reductase 1 (PTR1) and dihydrofolate reductase-thymidylate synthase (DHFR-TS) from Leishmania chagasi

Abstract Leishmaniasis is a tropical disease found in more than 90 countries. The drugs available to treat this disease have nonspecific action and high toxicity. In order to develop novel therapeutic alternatives to fight this ailment, pteridine reductase 1 (PTR1) and dihydrofolate reductase-thymidylate synthase (DHF-TS) have been targeted, once Leishmania is auxotrophic for folates. Although PTR1 and DHFR-TS from other protozoan parasites have been studied, their homologs in Leishmania chagasi have been poorly characterized. Hence, this work describes the optimal conditions to express the recombinant LcPTR1 and LcDHFR-TS enzymes, as well as balanced assay conditions for screening. Last but not the least, we show that 2,4 diaminopyrimidine derivatives are low-micromolar competitive inhibitors of both enzymes (LcPTR1 Ki = 1.50–2.30 µM and LcDHFR Ki = 0.28–3.00 µM) with poor selectivity index. On the other hand, compound 5 (2,4-diaminoquinazoline derivative) is a selective LcPTR1 inhibitor (Ki = 0.47 µM, selectivity index = 20).


Introduction
Leishmaniasis is an infectious disease caused by several species of the genus Leishmania, which are responsible for its different epidemiological and clinical characteristics 1 . Visceral leishmaniasis (VL), the most severe clinical form of leishmaniasis, affects 300,000 people worldwide, 90% of whom would die if no treatment is provided 2 . The available drugs to fight this disease have been in the market for more than 40 years and for that reason, resistant strains have been selected 3 . On top of that, N-methyl glucamine (Glucantime V R ), the drug of choice to treat patients with leishmaniasis, may cause nephrotoxicity, anorexia, abdominal pain, lethargy and cardiotoxicity 4 and the second-choice drug (amphotericin B) shows cardiac and nephron toxicity 5 . Although two new drug candidates to treat VL entered clinical tests in 2017 6 , the search for new safe, effective, oral, short-course leishmanicidal agents continues to be a priority. In order to achieve this goal, several efforts have been made to develop compounds that inhibit carbonic anhydrase 7-9 , arginase 10 , or enzymes that protect the parasite from the oxidative stress 11,12 .
Trypanosomatids dependence on host-synthesized purine and pteridines has long been targeted to fight Chagas disease 13 , human African trypanosomiasis 14 , and leishmaniasis 15,16 . Among the enzymes that play a key role in the salvage and folate pathway, dihydrofolate reductase is the macromolecular target of several drugs, including those available to fight malaria 17 . However, the selection of pyrimethamine-resistant Plasmodium falciparum strains underscores the risks of blocking a single target from this pathway 18 . Single point mutations 19 , as well as biochemical pathway plasticity 17,20 , can significantly reduce the drug effectiveness. In fact, a bypass mechanism explains why Leishmania parasites have low susceptibility to methotrexate or trimethoprim [21][22][23] , despite the fact that DHFR-TS gene knockout is lethal to the Leishmania parasites 24 . Nare, Luba, and Beverly 22 reported that upon dihydrofolate reductase-thymidylate synthase (DHFR-TS) inhibition, pteridine reductase 1 (PTR1) provides enough folate to guarantee the parasite survival. This finding suggested that both enzymes (DHFR-TS and PTR1) should be targeted to develop useful leishmanicidal drugs 17 . This hypothesis is still a matter of debate once the ptr1 gene knockout proved lethal to the parasite, probably because reduced pterin, which is produced only by PTR1, is essential for parasite growth and metacyclogenesis 22 . Accordingly, some authors consider PTR1 a validated target for drug development and have pursued its inhibition 15,24,25 . Despite the advances in the field, the risk of selecting resistant strains was not addressed in those papers.
Another caveat of previous work is the fact that most of the biological data reported so far was obtained for PTR1 from Leishmania major (LmPTR1), which causes cutaneous leishmaniasis, not the visceral form, as L. chagasi (syn. L. infantum) does. Despite the high overall sequence, similarities between LmPTR1 and LcPTR1 (90% - Figure S1), the active site of L. chagasi PTR1 is expected to be more flexible, once this enzyme is 100% identical to L. donovani PTR1, whose X-ray structure shows a disordered substrate binding site 26 . As a consequence, LmPTR1 inhibitors may not be as effective against LcPTR1.
In order to shed light on this matter, we expressed recombinant DHFR-TS and PTR1 from L. chagasi in Escherichia coli, so that standardized and balanced kinetic assays could be carried out. Next, we investigated 2,4-diaminopyrimidine derivatives as potential dual inhibitors for both enzymes, since derivatives of this chemical class have already been described as dihydrofolate reductase from Schistosoma mansoni (SmDHFR) 27, LmDHFR-TS and LmPTR1 24 inhibitors.

Synthesis of 2,4-diaminopyrimidines
The 2,4-diaminopyrimidine derivatives were synthesized as described by Teles et al. 27 . Briefly, all compounds were obtained in moderate yields by the condensation of suitable ketonitriles and guanidine as described in Figure 1.

Cloning of the enzyme LcPTR1
LcPTR1 cloning was carried out through the ligation independent cloning (LIC) protocol (https://www.helmholtz-muenchen.de/fileadmin/PEPF/Protocols/LIC-cloning.pdf). The genomic sequence that codes for PTR1 from L. chagasi, available in Genbank server (Accession code XM_001465671.1), was employed to primers design. The Phusion system (New England Biolabs V R ) was used to amplify the DNA fragment containing the LcPTR1 sequence using the forward and reverse primers 5 0 -CAGGGCGCCATGGCTGCTCCGACCGTG-3 0 and 5 0 -GACCCGACGCG GTTATCAGGCCCGGGTAAGGCT-3', respectively, which were acquired from Exxtend (https://www.exxtend.com.br/site/). The nucleotide sequence (bold) required for the LIC cloning system use, was inserted in the sequence of the forward and reverse primers. The remaining primer sequences hybridize to LcPTR1 gene. Genomic DNA from Leishmania chagasi cells (MHOM/BR/00/ BA262), provided by Dr. F abio Rocha Formiga (FIOCRUZ -Bahia) was extracted, as suggested in the UltraClean Tissue & Cells DNA Isolation kit protocol (Mobio, Carlsbad, CA, USA), and quantified at 260 nm wavelength (Nanodrop 200, Thermo Scientific, Waltham, MA, USA).
The polymerase chain reaction for LcPTR1 gene amplification was carried out in the ProFlex 3x32-well thermal cycler (Applied Biosystems, Waltham, MA, USA). The following steps and conditions were employed: initial step at 95 C for 3 min, followed by 35 cycles of denaturation step at 95 C for 30 s, annealing at 43 C for 30 s and a final extension step at 72 C for 110 s. The PCR product was loaded on a 1% agarose gel for electrophoretic separation (100 volts for 70 min) and then purified using the kit Wizard V R SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA). The purified product was quantified spectrophotometrically (Nanodrop 200, Thermo Scientific) at 260 nm wavelength.
The pETM11-LIC vector previously linearized by BsaI (NEB) digestion (60 min at 50 C) and the insert were treated with T4 DNA polymerase for cohesive ends formation and then incubated (1:3) for 30 min at 25 C for annealing. Then, E. coli (DH10b) chemically competent cells were transformed by heat shock 28 and the product of the transformation was grown in Luria Bertani (LB) medium (tryptone 1%, sodium chloride 1%, yeast extract 0.5% and bacteriological agar 1.5%), supplemented with 30 lg/mL kanamycin, for 16 h at 37 C.
True-positive clones were confirmed by colony PCR as follows: a colony-forming unit (CFU) was suspended in 100 ll type I sterile-water, which as then boiled for 5 min. The DNA from this suspension was employed as template material for PCR-amplification. The genetic material was PCR amplified using 1X PCR buffer (Invitrogen V R , Madison, WI, USA), MgCl 2 1.5 mM, dNTP 0.2 mM, 0.2 lL Platinum V R Taq DNA polymerase (Thermo Scientific) and water (q.s. to 20 lL) and the T7 primers (forward: 5 0 -TAATACGACTCACTATAGGG-3 0 and reverse: 5 0 -TAGTTATTGCTCAGCGGTGG-3 0 ) and thermocycling parameters previously described for LcPTR1 amplification.
The products of this reaction were analyzed on 1% agarose gel (100 V/1h) to confirm the presence of the gene's amplification band (0.9 kbp) as true-positive construction. Positive clones were cultured for 12 h in LB broth, supplemented with kanamycin (30 lg/ml), at 37 C and 150 RPM and then had their expressionplasmid DNA extracted with the aid of Pureyield Plasmid Miniprep System kit (Promega -A1223). The extracted plasmid was used for heat shock transformation 28 of the chemically competent E. coli BL21 (DE3) cells. The expression-plasmid containing N-terminal His-tag, TEV cleavage site, and LcPTR1 gene was sequenced using T7 primers and ABI 3500 Platform (Applied Biosystem) at Molecular Biology Research Laboratory (Pharmacy college -UFBA, Salvador, Brazil).

Expression and purification of LcPTR1
A CFU previously grown on LB agar was inoculated into 3 ml of sterile LB broth supplemented with kanamycin (30 mg/ml) and kept under constant stirring (250 RPM at 37 C for 12 h). After this time, the cell suspension was diluted (1: 100) in sterile LB broth (10 mg tryptone, 5 mg yeast extract and 5 mg NaCl per ml, pH 7) supplemented with antibiotics and kept under constant shaking (180 RPM) at 37 C until OD 600nm ¼ 0.6, at this point, the temperature was lowered to 25 C, the culture was kept at constant shaking (180 RPM) for 18 h and multiple combinations of IPTG concentrations (0.05-1 mM) were evaluated for the soluble expression of LcPTR1.
The cells were then recovered by centrifugation (16,000 Xg at 4 C for 15 min in a Sigma 1-14 K centrifuge) and resuspended in 50 mM Tris-HCl buffer pH 8.0 containing 125 mM NaCl and 20 mM imidazole supplemented by 1 mM phenylmethanesulphonylfluoride (PMSF P7626 Sigma Aldrich, Chicago, IL EUA) to mechanical lysis by sonication (15 cycles of sonication at 8 Watts for 15 s each, with 30 s intervals), in an ice-cold bath. The soluble fraction from lysate was recovered by centrifugation (14.500 Xg at 4 C for 15 min in a Sigma 1-14 K centrifuge). Next, the soluble fraction was filtered (0.45 lm -Sartorius) and loaded on crude His Trap FF column (GE Healthcare Life Sciences V R , Chicago, IL EUA) previously equilibrated with 10 column volumes (CV) of 50 mM Tris-HCl buffer pH 8.0 containing 200 mM NaCl (buffer A). After the system was washed with 10 CV of buffer A supplemented with 20 mM imidazole, followed by 5 CV of buffer A supplemented with 50 mM imidazole. Finally, the LcPTR1 was eluted with 5 CV of buffer A containing 500 mM imidazole.
All steps of the purification were monitored by 12% polyacrylamide gel electrophoresis (SDS-PAGE). The fractions containing the protein were pooled and the imidazole was removed after successive dilution steps (A 1:10 buffer) and centrifugation at with Amicon 30 kDa (Millipore, Chicago, IL EUA) a 4 C e 3500 Xg. Then, imidazole-free LcPTR1 was incubated with Tobacco Etch Virus protease (TEV), 1:20 (M/M) ratio, at 4 C for 12 h to cleavage of N-terminus His tag. The proteolysis product was loaded on a Nickelsepharose column (GE Healthcare V R ), previously equilibrated with buffer A (10 CV).
The cleaved LcPTR1, collected in the void, had its concentration measured using Bradford reagent 29 (595 nm) ( Figure S2). All purification steps were monitored by 12% SDS-PAGE.

Cloning of LcDHFR coding gene into E. coli ArticExpress (DE3) cells
The coding sequence for Leishmania chagasi strain JPCM5 DHFR-TS available in Genbank server (Accession code: XM_001463132.2) plus 23 nucleotides flanking the gene was employed for primer design and cloning into pET28a vector.
The LcDHFR-TS coding region was PCR-amplified using the genomic DNA, as a template, and the primers described next: The forward primer includes the NdeI restriction site (underlined) 5 0 -CAATACGCATATGTCCAGGGCAGCTG-3 0 and the reverse primer includes the HindIII restriction site (underlined) 5 0 -GCCTCCAAGCTT TCTTAACGGCCATC-3 0 . The restriction site for the NdeI enzyme was positioned in the 5' region of the LcDHFR-TS gene, which allowed the expression of the enzyme containing the polyhistidine tail in the N-terminal domain of the protein.
The PCR mixture included 70 ng (1 mL) of genomic DNA, 1Â HF buffer (Thermo Scientific V R ) buffer, 1.5 mM MgCl 2 , 0.2 mM dNTP, 0.5 lM each primer and water q.s. to 50 lL. The gene amplification reaction was carried out in the ProFlex 3x32-well thermal cycler (Applied Biosystems) using the following parameters: single denaturation step (94 C for 2 min), followed by 35 cycles of denaturation at 94 C for 30 s, annealing at 56.2 C for 30 s and extension at 72 C for 110 s. PCR included a final extension at 72 C for 600 s.
The reaction product was loaded on a 1% agarose gel for electrophoretic separation (100 Volts for 70 min). Next, the amplicon was purified directly from the gel matrix with Gene Jet Gel extract and DNA cleanup kit (Thermo Scientific V R ) and quantified spectrophotometrically (Nanodrop 200, Thermo Scientific) at 260 nm wavelength.
Next, the purified DNA fragment and pET28a vector were individually double-digested with NdeI and HindIII FastDigest restriction enzymes (Thermo Scientific) and then ligated (5:1 insert/ vector ratio) in the presence of T4 DNA ligase (Thermo Scientific) according to manufacturer instructions. E. coli ArticExpress (DE3) cells were transformed with the recombinant plasmid by heat shock protocol 28 . Colonies containing the inserted gene were selected using the resistance markers to kanamycin and gentamycin. True-positive clones were confirmed by colony PCR (using T7 primers: forward: 5 0 -TAATACGACTCACTATAGGG-3 0 and reverse: 5 0 -TAGTTATTGCTCAGCGGTGG-3 0 as described previously. The amplification parameters were set as follows: single initial denaturation step at 94 C for 180 s followed by 35 cycles of denaturation at 94 C for 45 s, annealing at 42.5 C for 30 s and extension at 72 C for 110 s. A final extension step (72 C for 600 s) was carried out complete the reaction. The amplification products were analyzed by 1% agarose gel electrophoresis and those 1718 bp-compatible amplified DNAs (1569 bp from insert plus 149 bp from the vector) was submitted to DNA sequencing using T7 primers on ABI 3730 Platform (Applied Biosystem) at Myleus Biotechnology center (Minas Gerais/Brazil). Colonies transformed with mutation-free plasmid were employed for the following steps.

Expression and purification of LcDHFR-TS
A CFU, previously grown on LB agar supplemented with kanamycin (30 mg/mL) and gentamycin (20 mg/mL) was cultivated overnight in antibiotic-supplemented LB broth (peptone from gelatin 1%, yeast extract 0,5%, sodium chloride 1% adjusted to pH 7,5), at 37 C and 200 RPM. After that, the cellular suspension (pre-inoculum) was employed to inoculate a freshly prepared sterile antibiotic-supplemented LB broth (1:100), which was kept at 37 C and 200 RPM until OD 600nm ¼ 0.6. At this point, the temperature was lowered to 10 C and IPTG was added (0.5 mM). After 48 h, the cells were then recovered by centrifugation (4000 Xg at 4 C for 10 min) and resuspended in 50 mM phosphate buffer pH 7.0, containing 200 mM (NH 4 ) 2 SO 4 supplemented with protease cocktail inhibitors 0,1X (SigmaFAST V R -S8830), for mechanical lysis by sonication (15 cycles of sonication at 8 Watts for 30 s each, with 30 s intervals, using Sonics Vibra-Cell Model VCX130PBt), in an ice bath.
Next, cellular debris and insoluble proteins (sediment) were separated from the supernatant by centrifugation (14.500 Xg at 4 C for 30 min in a Sigma 1-14 K centrifuge) and the soluble fraction was loaded on Ni sepharose His-trap fast flow resin (GE Healthcare) previously equilibrated with 5 CV of buffer A (50 mM phosphate buffer pH 7.0 containing 200 mM (NH 4 ) 2 SO 4 and 50 mM imidazole). 200 mM ammonium sulfate was added to the purification protocol and stock solution to increase the stability of the protein 30 .
After contaminants were eluted with 30 CV of buffer A, LcDHFR-TS was eluted by linear increase of imidazole concentration (50-500 mM) within buffer A. The fractions containing the protein were pooled together and the imidazole was removed by successive dilution and centrifugation steps (3000 Xg at 4 C for 30 min) in 50 mM phosphate buffer pH 7.0 containing 200 mM (NH 4 ) 2 SO 4 . Imidazole-free LcDHFR-TS had the N-terminus polyhistidine tag removed by thrombin (Sigma Cat. T7513) digestion at 25 C for 2.5 h (0.1:5 Thrombin Unit: LcDHFR-TS mg). Cleaved protein was recovered by loading the digested protein solution onto Ni sepharose His-trap resin (GE Healthcare), previously equilibrated with 50 mM phosphate buffer pH 7.0 containing 200 mM of (NH 4 ) 2 SO 4, and eluting it with 20 ml of the same buffer. All purification steps were monitored by 12% SDS-PAGE. The LcDHFR-TS had its concentration measured using Bradford reagent 29 (595 nm) ( Figure S2). The amount of contaminant thrombin present in the final solution resulting from the LcDHFR-TS purification protocol is approximately 0.5% in weight (0.0062 mg of thrombin per LcDHFR-TS mg). Aliquots of the purified enzyme were stored at -80 C in the presence of 30% glycerol (v/v) and thawed immediately before use.

Kinetic studies with LcDHFR-TS and LcPTR1
LcDHFR-TS apparent Km determination Kinetic measurements were carried out using a standard assay 31 , at pH 7.0 (50 mM sodium phosphate). Briefly, NADPH consumption was monitored spectrophotometrically (340 nM) (Shimadzu V R UV-1800, Columbia MD, USA) at constant temperature (25 C) and the absorbance variation within the first 30 s was employed to calculate the initial rate of reaction, by linear regression of the raw data.
The appropriate concentration of enzyme to be employed in the kinetic assays was evaluated by observing the reaction time required for substrate depletion under enzyme concentrations of 1.0 and 2.0 lM in the presence of 50 lM of DHF and NADPH.
The effect of the pH (5.0-8.0) over enzymatic activity was investigated with 0.2 lM of LcDHFR-TS and saturating NADPH and DHF concentrations (100 lM) (50 mM citrate buffer was used for pH 5.0 activity assay. 50 mM sodium phosphate buffer was used for pH range 6.0 to 8.0 activity assay). After the optimal pH was identified the apparent Km of the cofactor was determined by following a decrease in absorbance (340 nm) at growing up the concentration of NADPH (0.39 to 50 lM) at saturating concentration of DHF (100 lM). Likewise, substrate apparent Km was determined at saturating NADPH concentration (50 lM) and varying DHF concentrations (0.20-12 lM). All the measurements were carried out in triplicate and the apparent Km values for the cofactor and substrate were determined under a pseudo first-order kinetics assumption using least-squares non-linear regression method, as available in the GraphPad Prism version 5.0 for Windows (GraphPad V R Software, San Diego, CA, USA, www.graphpad.com).

LcPTR1 apparent Km determination
Kinetic measurements were performed according to the protocol described 32 via spectrophotometer (Shimadzu UV-1800) at pH 4.7 (20 mM sodium acetate). NADPH consumption was monitored spectrophotometrically (340 nm) at constant temperature (30 C). The variation of the absorbance monitored for 60 s was used to calculate the initial rate of the reaction by linear regression.
The apparent Km of NADPH was determined by varying its concentration (6-300 lM), maintaining the biopterin at saturation concentration (100 lM). Likewise, the apparent Km of the substrate was determined by saturation of NADPH concentration (100 lM) and increasing concentrations of biopterin (3.12-100 lM).
All measurements were performed in triplicate and the apparent values of Km for the cofactor and substrate were determined using the non-linear regression method (3-parameter equation), as available in GraphPad Prism version 5.0 for Windows (GraphPad V R Software, San Diego, CA, USA, www.graphpad.com). Comparison of the enzymatic activities in the absence and presence of the compounds (50 lM), which were dissolved in DMSO (5% v/v final concentration), was employed to calculate the percent inhibition (Equation (1))

Single concentration inhibition assays
where V i is the initial velocity in the presence of the inhibitor and V c is the reaction rate in the presence of DMSO (5% (v/v)).

Determination of IC 50 values
The compounds that inhibit either LcPTR1 or LcDHFR activity !30%, in the single concentration assays, had their IC 50 values determined by making rate measurements for at least five inhibitor concentrations. Initial velocity measurements (V i ) were employed to calculate inhibition percentages, whose plot versus log of inhibitor concentration afforded the IC 50 values, by nonlinear regression (assuming Hill constant ¼ 1.0) as available in the GraphPad program Prism version 5.0 (GraphPad V R Software, San Diego, CA, USA, www.graphpad.com).

Determination of the inhibition mechanism
The type of inhibition was determined under the same reaction conditions described above for different inhibitor concentrations Next, the kinetic measurements were carried out in saturating substrate concentrations (biopterin: 100 lM -DHF: 12.5 lM) and varying the concentration of NADPH (6.25, 12.5, 25, 50 and 100 lM). Apparent Km values, calculated by non-linear regression, were compared by T-test, available in GraphPad Prism version 5.0, and considered significantly different for p<.05. The results of this analysis were plotted in double-reciprocal graphs for visual analysis purposes only. The dissociation constant (K i ) of the active derivatives against both enzymes was obtained through Cheng-Prussov equation for competitive inhibitors 33 , as follows: where IC 50 is the inhibitor concentration required to reduce enzyme activity by 50%, [S] is the substrate concentration used in the inhibition assay and Km is the apparent Michaelis constant for the substrate in the absence of the inhibitor, at saturating NADPH concentration.

Results and discussion
Cloning, expression, and purification of LcPTR1 The complete coding sequence for LcPTR1 from L. chagasi plus eight nucleotides flaking the gene was employed for primer design, which began with the identification of restriction enzymes that would not cleave the LcPTR1 gene.
The amplification reaction afforded a single product of 867 bp (Figure 2(A)), which, following purification, was cloned into the pET-M11 vector using the LIC system. This strategy allows the expression LcPTR1 fused to a N-terminal His-tag. A similar approach was employed to obtain the PTR1 from Leishmania donovani in the soluble form 34 . Following the transformation of competent E. coli BL21 (DE3) cells and "colony PCR" screening, a positive and sequence verified ( Figure S3) colony was selected for protein expression studies.
After expression of LcPTR1 was carried out at 25 C, 18 h, 0.5 mM IPTG, a single-step purification, using affinity chromatography, that affords heterologous protein with high purity (Figure 2(B)), once this tag can influence protein stability and/or catalytic properties 35 .
TEV protease was employed to remove it and allow LcPTR1 to be obtained, after another round of affinity chromatography. This protocol affords 5-fold more protein (15 mg.L À1 of bacterial culture) than the one described for LdPTR1 34 and a similar yield as the one described for LmPTR1 36 .

Cloning, expression and purification of LcDHFR-TS
The migration profile of the PCR product shows a band that is consistent with the expected amplicon containing the LcDHFR-TS gene (1563 bp), plus the extension-primers (23 bp) (Figure 3(A)). After purification of the PCR product, the amplified insert (1.2 lg (46 ng/ll)) was cloned into the pET28a expression vector using the restriction site for the NdeI enzyme positioned in the 5' region of the LcDHFR-TS gene. This strategy allows the expression of the recombinant protein fused to His-tag in the N-terminal domain to facilitate its purification through nickel resin affinity chromatography. The same N-terminal His-tag fusion strategy was successfully employed in the cloning, expression, and purification of Leishmania (Vianna) braziliensis 37 , and T. cruzi DHFR-TS 38 .
Finally, heat shock transformation protocol allowed to isolate sequence verified ( Figure S4) colony-forming units containing the gene that codes for LcDHFR-TS (Figure 3(A)). Initial LcDHFR-TS expression tests using E. coli BL21 (DE3) showed that the recombinant protein was observed predominantly in the insoluble fraction of the cell lysate ( Figure S5). Low expression profile was reported for L. major DHFR-TS when BL21 strain was employed 39 . In order to increase the yield of a soluble and correctly folded recombinant protein that might be obtained, the Arctic Express E. coli (DE3) strain was employed 40 . This strategy proved successful and allowed us to obtain soluble LcDHFR-TS when the cells were grown at 10 C, for 48 h, in the presence of IPTG 0.5 mM (Figure 3(B), lanes 2 and 3). Several DHFT-TS are susceptible to protease action (T. brucei DHFR-TS 41,42 , L. tropica DHFR-TS 31 , and L. major DHFR-TS 39 ).
Likewise, proteolytic cleavage of LcDHFR-TS was also observed in the present work (Figure 3(B), lane 7), even upon addition of protease inhibitors cocktail (SigmaFAST) to the cell lysis buffer. Nevertheless, chromatographic purification by affinity made it possible to obtain the partially purified protein (Figure 3(B), lanes 6-7), which has its His-tag cleaved with thrombin. Non-cleaved LcDHFR-TS was removed by a second elution on nickel sepharose resin (Figure 3(B), lanes 8-10), as described for LcPTR1. Hence, the overall yield of LcDHFR-TS is 10.0 mg.L À1 of bacterial culture.

Kinetic characterization of LcDHFR-TS and LcPTR1
Catalytic activity of LcDHFR-TS Absorbance decrease at 340 nm over time (350 s) after the start of the reaction showed a linear profile up to 300 s when used 1.0 lM of the enzyme, indicating the steady-state condition is reached in the assays for at least the initial 30 s of the assay.
In addition, in order to determine optimum assays conditions for LcDHFR-TS kinetic characterization and inhibition, it was necessary to evaluate the effect of pH on the catalytic activity, since medium pH affects protonation states of amino acids related to catalysis and/or intermolecular interactions 43 .
The results obtained showed that catalytic activity of the DHFR domain of LcDHFR-TS is influenced by pH in the range of 5-8. Among the pH ranges evaluated, the maximum catalytic activity was observed at pH 7.0 (Figure 3(C)), that explains why the next steps were carried out on this pH.
Intrategumental pH studies of L. donovani observed that the parasite maintains neutral intracellular pH 44 , even under acidity or basicity medium conditions, thus demonstrating that the assay has biological significance as it approaches the enzyme natural habitat. Studies with DHFR-TS from different organisms agree with this data when describing the optimum pH at 7.0 for L. major 45 and Plasmodium vivax 46 enzymes. Other studies indicate that the optimum pH for the DHFR-TS of Trypanosoma brucei 41 and Plasmodium berghei 47 is in the pH range of 6.5-8.0. On the other hand, some studies show that for vertebrates and certain bacteria, catalytic activity is promoted at lower pH values 48 .
Although the decrease in the pH of the medium tends to promote catalytic activity by facilitating the transfer of the hydride to the nitrogen 5 of the DHF pteridine ring and consequently, increase Vmax 49 , this is not observed for DHFR-TS cases mentioned above, probably due to structural factors.
Once reaction conditions were standardized, we proceeded to LcDHFR-TS kinetic characterization. Determination of kinetic parameters is crucial to the standardization of assay conditions for inhibitors screening 50 since cofactor and substrate concentrations must be close to the Km apparent in the assay to allow identification of different inhibitors modalities: competitive, noncompetitive, and uncompetitive 43 . As an ordered bi-random mechanism is described to the formation of DHFR NADPH DHF complex 51 , the simplifications described by Michaelian kinetics can be applied for the LcDHFR-TS by maintaining saturating conditions of one substrate to allow kinetic characterization of the other 33 .
The values obtained for Km app and Vmax app for DHF were 1.0 ± 0.2 lM and 2.0 ± 0.11 nmol.min À1 .mL À1 , respectively, whereas for NADPH, Km and Vmax values were 2.3 ± 0.5 lM and 1.6 ± 0.1 nmol.min À1 .mL À1 , respectively (Figure 3(D)). The Km app values of LcDHFR-TS are close to the described for DHFR-TS of other species of Leishmania genus such as L. tropica, with 1.5 lM and 2.7 lM 39 , and L. major, 1.6 lM and 0.45 lM 31 for DHF and NADPH, respectively, what are already expected since the differences in its catalytic regions of the sequences are punctual. Both data obtained here for L. chagasi and those reported for L. tropica DHFR-TS shows lower affinity values of cofactor relative to the substrate for the enzyme, such a feature can be exploited by designing inhibitors directed to bind in the cofactor site, competing with NADPH, since its lower affinity would imply in lower concentration of the inhibitor needed to displace it from the enzyme site.
When we consider the number of catalytic turnover events (k cat ) 52 , LcDHFR-TS presents a k cat of 0.033 s À1 . Among DHFR-TS from other organisms, L. major 31 and T. brucei 42 k cat constants, 27 s À1 and 26 s À1 , respectively, are higher than that of the L. chagasi enzyme (0.033 s À1 ).

Catalytic activity of LcPTR1
The enzymatic reaction monitored through the decrease in absorbance (340 nm) showed a linear profile for up to 600 s when using 0.4 lM of LcPTR1 which indicates steady-state conditions in the assays for at least 60 initial seconds of the assay.
Next, in order to determine optimum assays conditions for LcPTR1 activity, its dependence on pH variation was examined. Significant variation of LcPTR1 activity was observed in the pH range between 2.7 and 5.7 (Figure 2(C)). As a marked peak of activity was observed at pH 4.7, this was then considered as the optimum pH for LcPTR1 catalytic activity. Similarly, the pH of 4.7 was determined as optimal for LmPTR1 22 and for LdPTR1, the optimum pH is 4.8 53 .
In addition to the reaction medium pH, the substrate and cofactor concentrations are critical for the establishment of balanced assay conditions in kinetic assays employed for potency and mechanism of inhibition determinations. Such conditions are achieved when the kinetic reaction medium presents substrate and cofactor concentrations close to the Km, since these concentrations can influence the identification of competitive, non-competitive, or uncompetitive inhibitors 52 .
Accordingly, as PTR1 presents an ordered mechanism of PTR1NADPHBPT complex formation 23 , in which the substrate (BPT) only binds the enzyme after formation of PTR1NADPH, it is possible for the use of the Michaelian simplifications for the kinetic characterization of LcPTR1 by maintaining saturating conditions for BPT to enable the determination of cofactor Km, and vice versa 33 .
The values obtained from Km app and Vmax app for BPT were 25 ± 1.2 lM and 13.0 ± 0.4 nmol.min -1 .mL -1 , respectively, whereas for NADPH, the values of Km and Vmax were 37.48 ± 2.1 lM and 15.4 ± 0.6 nmol.min -1 .mL -1 (Figure 2(D)), respectively. The following values were reported for LdPTR1: BPT Km app ¼5.8 lM; NADPH Km app ¼18.5 lM 54 . It is worth mentioning that these values were obtained through the double-reciprocal plot, which according to Copeland 33 can infer errors of up to 20%, in addition, the protein used to determine the kinetic constants were fused with His-tag.
The kinetic parameters found for LcPTR1 substrate and cofactor are not similar to the values reported for other species of Leishmania: for L. tropica PTR1 55 , was found for BPT Km ¼ 3.5 mM and for NADPH Km ¼ 19.0 mM, while L. major PTR1 22 were found Km of 4.6 and 6.7 lM for BPT and NADPH, respectively.
However, the values found for LcPTR1 agree with the crystallographic data of LdPTR1, since a disordered active site must have lower affinity for its ligands 26 . Such PTR1 kinetic data show a lower affinity for the cofactor relative to the substrate, this pattern is also observed for L. tarantoleae PTR1 (BPT/NADPH Km app ¼3.5/ 19) 55 and L. major (BPT/NADPH Km app ¼4.6/6.7) 56 .
Comparing LcPTR1 and LcDHFR kinetic characterization data, both enzymes show lower affinity for the cofactor relative to the substrate, which is evidenced by higher Km app values for NADPH (LcDHFR: DHF/NADPH Km app ¼1.0/2.3; for LcPTR1: BPT/NADPH Km app ¼25.0/37.5). When we consider the number of catalytic turnover events per time (k cat ) 50 , the LcPTR1 presents a k cat of 0.54 s -1 . Among PTR1 from other organisms, L. major PTR1 56 k cat is similar to the L. chagasi enzyme (0.44 versus 0.54 s À1 , respectively), whereas the for T. brucei PTR1 32 k cat is higher than that of L. chagasi (4.3 versus 0.54 s À1 , respectively).

LcPTR1 and LcDHFR inhibition assays
Although PTR1 and DHFR are not related at the primary sequence level, both enzymes are inhibited by compounds containing the 2,4-diaminopyrimidines skeleton 24 which can be explained by the chemical similarity between the natural substrate of PTR1 and DHFR-TS and the 2,4-diaminopyrimidine nucleus.
Once the inhibition of both enzymes may have a synergistic detrimental effect over the parasite survival 17 , and prevent the selection of resistant strains, we decided to assay diaminopyrimidine derivatives, which have already been described as S. mansoni DHFR-TS inhibitors 27 , against LcPTR1 and LcDHFR-TS. Single-dose assays (Figure 4(A) and 5(A)) suggest that compounds 1 and 3 are active against both enzymes (e.g. compound 1 Â LcPTR1 ¼ 90 ± 1% vs. LcDHFR-Ts ¼ 100 ± 1.17%), whereas compounds 2 and 4 are more active against LcPTR1 than LcDHFR (LcPTR1 inhibition ¼ 40 ± 1%, vs. LcDHFR inhibition ¼ 20 ± 2.4%) and compound 5 seems to inhibit LcPTR1 (98 ± 0.8% inhibition at 50 mM) but not LcDHFR (20 ± 2.4% inhibition at 50 mM). However, this preliminary analysis might be misleading once the catalytic efficiency of LcPTR1 and LcDHFR-TS are significantly different. In this case, IC 50 values are suitable only to compare the compounds potency for one specific target, either LcPTR1 or LcDHFR-TS. For instance, compounds 1, 2, 3 show a discrete variation in IC 50 values against LcPTR1 (Figure 4(B)).
This result suggests that structural variations at position 6 of the 2,4-diaminopyrimidine nucleus (Figure 4(B)) appear not to influence the potency of these compounds against this target. On the other hand, the activity profile of compounds 1-3 against LcDHFR-TS ( Figure 5(B)) suggests that the presence of an aromatic ring at least two carbons away from the 2,4 diaminopyrimidine nucleus is crucial for LcDHFR-TS inhibition (see IC 50 of compounds 1 vs. 2).
Although the IC 50 values of compound 1 are lower for LcDHFR-TS than LcPTR1, one cannot say it is more potent against the first, since substrate concentration, reaction pH, temperature and Km of each enzyme can influence the value of IC 50 57 , 58 . In order to overcome this limitation, K i values calculated with the Cheng-Prussov equations might be employed, as long as the compounds mode of inhibition was known. Consequently, we undertake a careful investigation of compounds inhibition mode. Once compounds 1-4 possess high structural similarity, it is reasonable to assume that they share the same mechanism of inhibition. Thus, just results for compound 1 and 5 are discussed next.

Mechanism of inhibition assays
Kinetic parameters like Km app and Vmax app under different inhibitor concentrations were analyzed in order to determine the mechanism of action of compound 1 over LcPTR1. Results obtained with saturating concentration of NADPH (Figure 4(C)), show a set of lines that intercept the Y-axis at the same point (1/Vmax app ) and cross the X-axis (-1/Kmax app ) at different points.
This behavior suggests a competitive mechanism of inhibition, but the fact that PTR has a bi-bi ordered catalytic mechanism 23 requires that compound 1 also shows an uncompetitive mechanism of inhibition to the cofactor.
In order to confirm this hypothesis, enzymatic inhibition experiments were carried out at a saturating concentration of biopterin, while the NADPH concentration was varied (Figure 4(C)). As expected, the values of Km app and Vmax app decrease proportionally, as the concentration of compound 1 increases. Thus, the graph of the reciprocal double shows parallel lines, typical of an inhibitor uncompetitive concerning the cofactor.
As the structure of compound 5 differs substantially from the other compounds, we decided to perform experiments to determine the mechanism of inhibition to this compound as described above.
Thus, when the assays were performed with saturating concentration of NADPH and different concentrations of biopterin, it was observed that the lines intersect the Y-axis at the same point, while the intersection with the X-axis occurs at different points (Figure 4 (D)). Therefore, compound 5 exhibits the same behavior observed for compound 1 (competitive mechanism with biopterin). In fact, when different concentrations of NADPH were used, reduced values of Km app and Vmax app were observed (Figure 4 (D)). These results confirm that compound 5 has an affinity for the PTR1-NADPH complex (mechanism of inhibition uncompetitive with the cofactor).
When the assay was carried out with LcDHFR-TS, it was observed that DHF Km app increases linearly as compound 1 con- Vmax app is not affected by increasing concentrations of compound 1 (Vmax ¼ 1.9 ± 0.1 nmol.min À1 .mL À1 ).
The competitive mechanism related to substrate observed for compound 1 is consistent with the data described for 2,4-diaminopyrimidine inhibition of L. major enzyme 59 . The same pattern was also observed for Bacillus anthracis 60 and Schistosoma  mansoni 27 enzymes. On the other hand, although the noncompetitive inhibition related to cofactor is also observed against S. mansoni 27 and Gallus gallus 59 DHFR, uncompetitive interaction is observed for E. coli DHFR-TS 59 .

Selective and dual inhibitors
The design of dual-acting inhibitors depends on the identification of compounds that have similar affinity for LcPTR1 and LcDHFR-TS, however one must bear in mind that these macromolecules have Km values that differ by more than 15 fold (substrate 25Â (LcPTR1 25 ± 2.1 mM vs. LcDHFR-TS 1.0 ± 0.2 mM); cofactor 16Â (LcPTR1 37.2 ± 3.7 mM vs. LcDHFR-TS 2.3 ± 0.5 mM)). In addition, IC 50 values change according to substrate or enzyme concentration. For instance, when kinetic data is recorded at [S]¼Km, the IC 50 /Ki ratio for noncompetitive inhibitors is equal to 1/2. If higher concentrations of the substrate are employed, the ratio changes.
All kinetic data for LcPTR1 was recorded at S/K M ffi 1.0, but the kinetic assays for LcDHFR-TS were recorded at S/K M ¼ 5.0 for the  substrate (S/K M ¼2.0 for the cofactor). Under these circumstances, the direct comparison of IC 50 values is misleading. In order to circumvent this dilemma, the affinity of the compounds to their targets was assessed by their dissociation constant (K i ), calculated with the Cheng-Prussov equation for competitive inhibitors 33 (Table 1).
This approach shows that compounds 1 and 3 have moderate selectivity for LcDHFR-TS (K i LcPTR1 /K i LcDHFR-TS ratio ¼ 5.4 and 3.3fold selectivity, respectively), whereas compound 2 has similar affinity to both targets (K i LcPTR1 /K i LcDHFR-TS ratio ¼ 0.68). Two out of three 2,4-diaminopyrimidine derivatives described are active against LmDHFR-TS and LmPTR1 have selectivity profile similar to compounds 1 and 3: 6 to 13-fold selectivity for LmDHFR-TS over LmPTR1 24 . On the other hand, compound 5 is highly selective for LcPTR1.
Despite the promising results presented so far, ADMET liabilities, such as high toxicity, might pose a severe hurdle to their advance as lead compounds. In order to evaluate this potential limitation, their predictive cardiac toxicity (hERG receptor inhibition model), mutagenicity (AMES test model) and carcinogenicity (Three-class model) was evaluated through admetSAR server (http://lmmd.ecust.edu.cn/admetsar1). These in silico models underscore no red light alerts and suggests oral acute toxicity doses higher than 500 mg/Kg for in vivo evaluation of all compounds ( Figure S6).

Conclusions
Although DHFR-TS and PTR1 from L. major have long been explored for drug development purposes, the same cannot be said about their counterparts in L. chagasi. This work sheds some light on this subject by reporting not only the optimal conditions to obtain LcDHFR-TS and LcPTR1, but also the balanced conditions for in vitro screening. This assay shows that 2,4 diaminopyrimidine derivatives, substituted at position 6, are competitive inhibitors of both enzymes whereas the 2,4-diaminoquinazoline derivative 5 is a selective inhibitor of LcPTR1. This information shall guide the development of potent inhibitors with dissimilar selectivity profiles.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by Conselho Nacional de Desenvolvimento Cient ıfico e Tecnol ogico; Fundação de Amparo a Pesquisa do Estado da Bahia.