Synthesis and Biophysical and Biological Studies of N-Phenylbenzamide Derivatives Targeting Kinetoplastid Parasites

The AT-rich mitochondrial DNA (kDNA) of trypanosomatid parasites is a target of DNA minor groove binders. We report the synthesis, antiprotozoal screening, and SAR studies of three series of analogues of the known antiprotozoal kDNA binder 2-((4-(4-((4,5-dihydro-1H-imidazol-3-ium-2-yl)amino)benzamido)phenyl)amino)-4,5-dihydro-1H-imidazol-3-ium (1a). Bis(2-aminoimidazolines) (1) and bis(2-aminobenzimidazoles) (2) showed micromolar range activity against Trypanosoma brucei, whereas bisarylimidamides (3) were submicromolar inhibitors of T. brucei, Trypanosoma cruzi, and Leishmania donovani. None of the compounds showed relevant activity against the urogenital, nonkinetoplastid parasite Trichomonas vaginalis. We show that series 1 and 3 bind strongly and selectively to the minor groove of AT DNA, whereas series 2 also binds by intercalation. The measured pKa indicated different ionization states at pH 7.4, which correlated with the DNA binding affinities (ΔTm) for series 2 and 3. Compound 3a, which was active and selective against the three parasites and displayed adequate metabolic stability, is a fine candidate for in vivo studies.


INTRODUCTION
Neglected tropical diseases caused by kinetoplastid parasites (i.e., Trypanosoma cruzi, Trypanosoma brucei, and Leishmania) are a great cause of suffering around the world. American and African trypanosomiases, as well as leishmaniasis, threaten millions of people mainly in the least developed countries. 1−4 Available therapies to treat these illnesses are not satisfactory as they often present poor efficacy against drug-resistant parasite strains or a particular stage of the disease or patient condition (e.g., chronic Chagas disease, late-stage rhodesiense sleeping sickness, and HIV/leishmaniasis coinfection), and habitually require prolonged treatment regimens and high doses, with associated severe side effects. 4−6 In addition, T. brucei rhodesiense, T. cruzi, and Leishmania spp. are all zoonotic and cause disease in various domestic animals as well, for which the treatment options are even more limited. 7−9 Hence, new antiprotozoal drugs are needed to improve this far from ideal therapeutic arsenal.
Several classes of promising new chemical entities (NCE) are currently in clinical development under the aegis of the Drugs for Neglected Diseases initiative (DNDi). 10 For instance, the polyadenylation specificity factor 3 (CPSF3) inhibitor acoziborole has reached phase IIb/III for the singledose oral treatment of gambiense human African trypanoso-miasis (HAT), 11 and its derivative DNDI-6148 is a lead candidate in phase I for leishmaniasis 12 and Chagas disease. 13 Other NCEs being developed for visceral leishmaniasis include the nitroimidazole DNDI-0690, the proteasome inhibitors LXE408 and GSK3494245/DDD01305143, and the cdc2related kinase 12 (CRK12) inhibitor GSK899/DDD853651. 13 Bis(2-aminoimidazolines) are a class of AT-rich DNA minor groove binders (MGBs) with established in vitro and in vivo efficacy against T. brucei. The N-phenylbenzamide derivative 1a, which was curative by oral administration in an acute mouse model of African trypanosomiasis, is the prototype of this series (Chart 1). 14−18 Strong experimental evidence suggested that this compound can displace High Mobility Group (HMG)-box-containing proteins that are essential for kinetoplast DNA (kDNA) function from their DNA binding sites. This interaction led to the disruption of kDNA and eventually to the death of the parasite. 19 Hence, these compounds were chosen as starting points to develop new molecules against all three main kinetoplastid parasite species, and potentially also against parasites containing AT-rich nuclear DNA (e.g., Trichomonas vaginalis).
In spite of their action on kDNA and their excellent activity against T. brucei, bis(2-aminoimidazolines) are significantly less active against Leishmania or T. cruzi. 16,17 For instance, lead compound 1a displays a 5-fold lower activity against promastigotes of L. donovani (EC 50 = 4.29 μM) than against bloodstream forms of T. b. brucei (EC 50 = 0.83 μM) 20 and was inactive against T. cruzi (Table 1). Since a drug targeting the kDNA of intracellular amastigotes of Leishmania or T. cruzi will have to cross 4 membranes (host cell membrane, parasitophorous vacuole membrane, parasite cell membrane, and mitochondrial membranes) to reach its target, optimum membrane permeability (or efficient, mediated drug uptake) will be required to observe antiparasitic activity against these intracellular forms. We reported before that the chemical modification of bis(2-aminoimidazoline) compounds to lower their pK a can lead to a notable improvement in membrane permeability and antitrypanosomal activity of these dicationic compounds. 17,20 In this study, we decided to combine different chemical strategies to modify the physicochemical properties of lead compound 1a in order to improve its activity against the intracellular parasites L. donovani and T. cruzi and conserve its binding affinity toward kDNA. We have shown earlier that the three NH groups of each 2-aminoimidazolinium group of 1a are crucial for binding to AT-rich DNA because they form hydrogen bonds with thymine and adenine from both DNA strands. 18,19 In addition, the nitrogen atoms that are not hydrogen bonded to DNA are associated with water molecules. 19 Thus, new derivatives replacing the 2-aminoimidazoline rings of 1a by the more acidic 1H-benzimidazol-2ylamine (series 2) or pyridine-2-carboxamidine (series 3) heterocycles were synthesized. The central scaffold was also modified by incorporating lipophilic and/or electronegative substituents (F and Cl) and large electron-donating substituents (O i Pr) (Chart 1). Hence, three families of compounds with structural characteristics related to 1a were prepared, their physicochemical properties (pK a , log P, and solubility) were determined, and the compounds were assayed in vitro against four protozoan parasites (T. brucei, T. cruzi, L. donovani, and T. vaginalis) for SAR studies. To evaluate the potential of the new compounds to target kDNA, we also assessed their DNA binding affinity and mode of binding with AT-rich and GC-rich DNA using different biophysical techniques. Finally, the metabolic stability of one selected candidate, active against all three trypanosomatid parasites, was measured in vitro to assess its potential for further in vivo studies. heating at 60°C was required to reduce the reaction time. The progress and completion of the reaction were monitored by thin-layer chromatography (TLC) and by HPLC−MS.
In most cases, the reaction did not go to completion, and a mixture of mono-and disubstituted compounds was obtained. The crude Boc-protected products were purified by centrifugal PTLC on silica plates previously deactivated with a 9:1 hexane/Et 3 N mixture (6c−h) or by column chromatography using neutral aluminum oxide (6i) (Scheme 1). Removal of the Boc-protecting groups was accomplished with TFA at 0°C to yield the corresponding bis(imidazolidin-2-imine) compounds 1c−i as trifluoroacetate salts. In this reaction, working at low temperature in the absence of the protic solvent (e.g., using CH 2 Cl 2 ) was important to avoid the rupture of the PhN�C bond, resulting in the loss of the imidazoline ring and the generation of the amine starting material. Compound 1i was obtained in low yield because several recrystallizations were necessary to remove traces of the monosubstituted byproduct.
For the synthesis of halogen-containing bis(pyridine-2carboxamidines) 3c−l, dianiline precursors 5c−l were reacted with N-(tert-butoxycarbonyl)pyridine-2-carbimidothioate (21) in the presence of HgCl 2 and Et 3 N (Scheme 3). 26 We have shown recently that this protocol is useful for the synthesis of bis(arylimidamides) from electron-poor dianilines, where reagent 17 failed to provide the expected products. 26 Using an excess of the reagents and microwave irradiation as the heating source (50°C, 1 h), 24c−l were obtained smoothly by silica and/or reverse phase chromatography. Boc group removal was performed with TFA or with 4 M HCl g −dioxane solution. This protocol also works satisfactorily with different anilines such as 5a, 5h, 8, 26 9, and 14.
Eventually, we noticed that the yields of the synthesis of bis(pyridine-2-carboxamidines) using reagent 17 were greatly improved when DMF was used as solvent instead of the "classical" mixture of EtOH/CH 3 CN. In fact, even the fluorine-containing compounds 3i−l could be obtained in 61−78% yield using this solvent and 4.5 equiv of 17 (Scheme 3, see method C2 in the Experimental Part).
2.2. Biology. 2.2.1. Antiprotozoal Activity. The activity of the three series of compounds was tested in vitro against three trypanosomatid parasites (T. brucei, T. cruzi, and L. donovani) and against the urogenital human parasite T. vaginalis (Tables  1, 2, and 3). Unspecific cytotoxicity against different mammalian cell lines (i.e., HEK, L929 fibroblasts, THP-1, and Vero CCL-80) was also tested to determine the selectivity indexes (SI) against the parasites.

SAR Studies. Bis(2-aminoimidazoline) Series 1
The anti-T. cruzi activity of series 3 was outstanding, with EC 50 values on epimastigotes in the midnanomolar range and SI (L929/epi) from 159 to 3454 for the chlorine-containing derivatives 3e (R 1 /R 4 = Cl) < 3d (R 1 /R 3 = Cl) < 3g (R 2 /R 4 = Cl). These were 9-, 7-, and 5-fold more active than the fluorine-containing counterparts 3j, 3i, and 3l, respectively. In contrast, the fluorine-containing analogue 3k (R 2 /R 3 = F) was 69 times more potent than its chloro analogue 3f. This difference could be due to the capacity of fluorine atoms in position ortho to the amide bridge to form intramolecular hydrogen bonds (IMHB) with the NH group, thus shielding polarity and improving membrane permeability. 35 Other compounds displayed activities in the submicromolar [i.e., 3a (R 1 −R 4 = H) > fluorene 23 > 3c (X/Y = N)] and low micromolar range (18 > 3h). The anti-T. cruzi activity against intracellular amastigotes was maintained for most series 3 compounds, although a loss of potency of ∼6 to 20-fold was observed for 3a, 3d, 3e, and 3g; even in those cases, however, the activity remained very promising (amastigote EC 50 = 0.26− 1.28 μM). Remarkably, the activity of compound 3h, which has two large hydrophobic isopropoxy groups (R 2 /R 3 = O i Pr), was twice as active against intracellular amastigotes as against epimastigotes, probably indicating that these substituents are favorable for crossing the host cell membrane (i.e., to reach the intracellular parasites), although the mechanism by which this occurs has not yet been studied and the compound was nontoxic to HEK cells. This trend was confirmed with intracellular amastigotes of L. donovani, which were equally susceptible to 3h as promastigotes, and it could be noted that the addition of such groups would be deleterious to recognition by at least some protozoan drug transporters. 36  All of the bis(arylimidamides), except 3f, inhibited the growth of promastigotes of L. donovani with EC 50 values ranging from 0.26 (3a) to 3.0 μM (3h). This represented a clear improvement in antileishmanial activity versus the bis(2aminoimidazoline) lead compound 1a (EC 50 = 4.29 μM; P < 0.05, n = 3). In general, the antileishmanial activity was maintained against intracellular amastigotes, although with a small loss of potency (approximately 1.5 to 3-fold) for 3a, 3g, 3h, 18, and 23 (amastigote EC 50 = 0.65−4.19 μM). The best activities were in the same range as those of the reference drug amphotericin B (EC 50 = 0.19 μM). In contrast, the presence of chlorine atoms ortho to the 4-picolinimidamido groups (3d: R 1 /R 3 = Cl; 0.91 μM and 3e: R 1 /R 4 = Cl; 0.78 μM) or two nitrogen atoms next to the amide linker (3c, X/Y = N) promoted the antileishmanial activity against intracellular amastigotes, as shown by the nearly 2-fold lower EC 50 values versus promastigotes [(P = 0.0424, P = 0.0404, and P = 0.0026, respectively, n = 3)] (Table 3). Remarkably, the closely related analogue 3b (X = CH, Y = N) was ineffective against amastigotes of L. donovani and epimastigotes of T. cruzi, probably accounting for its inability to form an IMHB with the NH of the amide group (vide supra).
Moreover, the compounds displayed a good selectivity toward intracellular forms of L. donovani, with SI (THP-1/amast) from >11.9 for the least active compound 3h to >76.9 for the most effective one (3a). Thus, compound 3a, which displayed submicromolar efficacy against amastigotes and excellent selectivity, emerged as a new antileishmanial hit compound. The lack of antiprotozoal activity of 20 versus 3a indicates that the amidine structure is crucial for the activity of this series. Moreover, two pyridine-2-carboxamidine groups are needed for the compounds to display potent antitrypanosomatid activity, as shown by the lack of activity of the monosubstituted compounds 3b_b and 3c_c versus 3b and 3c, respectively. This is consistent with previous reports showing that monoarylimidamide derivatives are generally less potent than bis(arylimidamides), even though submicromolar in vitro activities against L. donovani have been reported for several compounds. 38, 39 2.2.5. DNA Binding Affinity. Thermal Melting Assays. Since kDNA is a potential target of these compounds, we assessed their binding properties with AT-containing DNA using thermal melting (T m ) assays. Thermal denaturation can be used to measure the stabilization effect (i.e., the binding affinity) produced by a molecule on binding to duplex DNA. Hence, the T m increase (ΔT m ) for compound complexes relative to uncomplexed DNA provides an estimation of the DNA binding affinity of a compound. ΔT m values were determined by circular dichroism spectroscopy with a DNA hairpin duplex containing CGATATATATCG ["(AT) 4 "] ( Table 4). With the exception of 1e (ΔT m = 0.9°C), the T m increases measured for 1c−i ranged from 3.1°C (1h) to 8.2°C (1c) and were larger than ΔT m of lead 1a (1.6°C). This showed that the introduction of Cl, O i Pr or nitrogen atoms in the scaffold positively affected the binding affinity of bis(2-aminoimidazoline) derivatives to (AT) 4 DNA. No binding to the (CG) 4 -containing hairpin was observed, as illustrated by 1g (ΔT m = 0°C), which indicated selective binding to AT-DNA. Of note, the big difference between the value reported for lead 1a with poly(dA·dT) 2 (ΔT m = 47.1°C ), 18 and our value (ΔT m = 1.6°C) may be attributed to the different oligonucleotides used in the two experiments. DNA containing an A-tract has a narrow minor groove and can bind minor-groove agents without a significant average change in groove width or local helix axis bends in solution. In contrast, straight alternating AT sequences (i.e., (AT) 4 used in our assays) require the minor groove to narrow with bending of the helix upon binding minor-groove agents. 40 Bis(2-aminobenzimidazoles) 2a−g, 12, 13, and 16 showed ΔT m values in the range 4.0−7.0°C, indicating that this series also binds to the (AT) 4 -containing hairpin DNA (Table 4). With the exception of 2a and 2d, the binding affinities of the bis(2-aminobenzimidazoles) were lower than those of their bis(2-aminoimidazoline) counterparts, which may be related to the mostly uncharged nature of the benzimidazole series (pK a < 6.75) at physiological pH (Table 4). Indeed, positive charge(s) and a crescent-shaped molecule are known to promote kDNA minor groove binding. 14,41 The replacement of the amide bond linker of 2a−g by a more flexible ethylene chain (12) produced a T m increase of 1.2°C with respect to 2a (ΔT m = 4.9°C). In contrast, replacing the N-phenylbenzamide skeleton with a rigid and planar fluorene scaffold (16) resulted in an insignificant (within the experimental error) increase in T m versus 2a.
The T m increases for 3a−h were in the range 1.6°C (3e) to 11°C (23), confirming that all the synthesized bis-(arylimidamides) bind to AT-containing DNA, albeit with various affinities ( Table 4). Most of the compounds except 3e showed larger T m increases than the bis(2-aminoimidazoline) lead 1a upon binding to the (AT) 4 DNA hairpin. According to these experiments, the strongest binder was the rigid fused-ring tricyclic fluorene derivative 23 (ΔT m = 11.0°C). This data agreed with the value reported by Boykin and co-workers for 23 complexed with poly(d(A-T) 2 ) (ΔT m = 15.2°C). 25 The chlorinated derivatives 3d, 3f, and 3g also showed good affinity with ΔT m values in the range 5.1−8.9°C. In general, except for 3g (R 2 /R 4 = Cl), which displayed one of the strongest binding affinities (ΔT m = 8.9°C) of the series, the introduction of chlorine (3d−f) or pyridine rings into the scaffold (3b and 3c) led to a drop in binding affinity with respect to the unsubstituted compound 3a.  4 ]. i Taken from ref 42. j Unspecific binding. k Fitting to a one-site binding model. l No binding at low concentration and unspecific binding to the chip dextran matrix at high concentration. m Aggregate or precipitate. n Not measured due to the lack of sample. o pK a value corresponding to the pyridine N. p pK a value corresponding to the heterocyclic NH.

Surface Plasmon Resonance-Biosensor Assays.
Surface plasmon resonance (SPR) was used to measure the binding affinities of the compounds to two DNA hairpin duplexes containing CGAATTCG ["A 2 T 2 "] and CGCGCGCG ["(CG) 4 "] sequences. 42 All of the N-phenylbenzamide-based bis(2-aminoimidazoline) derivatives (1c−i) bound to the A 2 T 2 sequence with K D values in the submicromolar to the micromolar range (Table 4, Figures 1 and S1), in agreement with the T m experiments and with previous results obtained with similar compounds. 19,42 The stoichiometry of binding was 1:1 for compounds 1c, 1f, 1h, and 1i, and >1:1 for 1d, 1e, and 1g, as shown by r values (mole of bound compound per mole of DNA hairpin duplex) ≥ 1, respectively (Figures 1 and S1). Bis(2-aminoimidazoline) analogues with substituents on the N-phenylbenzamide rings were >2-fold weaker binders than 1a (K D = 0.17 × 10 −6 M), 42 as shown by primary binding constants (i.e., high affinity site) ranging from 0.38 × 10 −6 M for 1c (X/Y = N) to 45 × 10 −6 M for the dichloro analogue 1d (R 1 /R 3 = Cl) ( Table 4). Secondary binding constants were 80-to 30-times weaker, respectively. In these experiments, secondary binding generally accounts for nonspecific binding interactions with the DNA hairpin loop, as previously reported. 19 The relative position and size of the substituents (i.e., Cl, O i Pr, and F) had a clear influence on the binding strength to the minor groove of A 2 T 2 -DNA. The compounds with chlorine (1d, 1f, and 1g) or O i Pr (1h) groups in positions ortho to the amide bond (R 2 /R 3 ) had 4-to >40-times weaker binding affinity than the molecules with the Cl atoms next to the 2aminoimidazoline group (1e, R 1 /R 4 = Cl). This could be due to the conformational restriction and/or steric clash imposed by the presence of these groups near the amide bridge of the molecule. The fact that the fluorine atom ortho to the amide bond (1i, R 3 = F) hardly affected the binding to DNA (K D = 0.55 μM) compared to 1a is consistent with the hypothesis that small size groups are preferred in this position. These compounds were AT-specific, as no significant binding to (CG) 4 was observed up to 100 μM (Figure 1 and S1; Table 4).
For the bis(2-aminobenzimidazole) derivatives (2a−g, 12, 13, and 16) no binding to A 2 T 2 or (CG) 4 sequences was observed at the highest concentration tested (100 μM). Unspecific binding to the chip dextran matrix (compounds 2a−g) or aggregation/precipitation (12, 13, and 16) was observed instead, indicating that SPR-biosensor assays are not adequate to determine the binding affinity of this class of poorly water-soluble compounds.
Bis(arylimidamides) 3a and 18 (amide and ethylene linkers, respectively) bound selectively to A 2 T 2 -containing DNA versus CG-containing sites with primary binding constants in the submicromolar range comparable to those of lead compound 1a (Table 4). For compound 18, this represents a 10-fold increase in affinity compared with its bis(2-aminoimidazoline) analogue reported earlier. 42 In contrast, the fluorene derivative 23 showed little sequence selectivity with similar binding constants for A 2 T 2 and (CG) 4 oligonucleotides, which is consistent with a mixed intercalative/groove binding mode of interaction. 43 The binding stoichiometry for compounds 3a, 18, and 23 was 2 mol of bound compound per mole of ATcontaining DNA hairpin duplex ( Figure S2), which is similar to lead 1a and other related bis(imidazolin-2-imine) deriva- tives. 19,42,44 From these SPR experiments, we did not detect significant binding to A 2 T 2 and (CG) 4 -containing sequences at the highest concentration tested (100 μM) for the chlorinecontaining compounds 3d−g, whereas 3b, 3c, and 3h showed nonspecific binding to the chip dextran matrix. This apparent discrepancy with the thermal melting assays, which showed binding of 3d−g to (AT) 4 DNA (vide supra) might indicate strong sequence selectivity for these derivatives. The A 2 T 2 sequence used in the SPR assays has a very narrow minor groove of 3.5−4.0 Å in the center of the sequence (vs 5.16− 6.79 Å for (AT) 4 ), 45 which tends to accommodate a more planar conformation of the bound ligand and favors van der Waals interactions with the walls of the groove. 42 Hence, the presence of large chlorine substituents may be unfavorable for such interactions with this minor groove.
2.2.7. DNA Binding Mode. Once the DNA binding affinity and AT-selectivity of the compounds were established, we explored the binding modes of series 1 and 3 using flow linear dichroism (LD). 46 Positive-induced LD signals are indicative of minor groove binding, whereas intercalating molecules induce a negative LD signal. LD spectra were recorded for natural DNA (salmon testes) titrated with representative compounds 1g and 3a in phosphate buffer at 25°C working with a base pair/drug (Bp/D) ratio of 0/1, 1/1, and 1/5. 47 A positive-induced LD signal was observed at ≈310 nm (where the DNA does not absorb) upon addition of increasing concentrations of 1g and 3a ( Figure 2). Altogether, the results of LD, SPR, and CD experiments indicate that 1g and 3a bind specifically to the minor groove of AT-containing DNA, like lead 1a.

Fluorescent Intercalator Displacement (FID) Assay.
We sought to determine whether the bis(2-aminobenzimidazole) series 2 compounds could bind DNA by intercalation. A FID assay utilizing the displacement of the DNA-bound fluorescent intercalator ethidium bromide was accomplished with compounds 2a, 2c, 2d, 12, 13, 16, and 3a as controls (see Supporting Information). 48 The addition of compounds 2a, 2c, and 12 to (AT) 4 and (CG) 4 hairpins resulted in >50 and >87% decreases in fluorescence, respectively, indicating strong intercalation (Table S2B). Binding of 2d, 13, and 16 to (AT) 4 and (CG) 4 reduced the fluorescence (F) by 25 and ≥43%, respectively, indicating that the compounds intercalate, although to a lower extent. In contrast, the addition of 3a to (AT) 4 and (CG) 4 did not decrease the % F, consistent with its specific minor groove mode of binding.
2.2.9. Determination of Physicochemical Parameters (pK a , log P, and Solubility). The protonation state of drugs in the body has a profound influence on their ADMET properties and their binding to biological targets. 49 For DNA MGBs, the (di)cationic nature of the ligand, in addition to a crescent shape matching the curve of the groove, is crucial to allowing the best fitting between the MGB and the groove through van der Waals and hydrogen bonding interactions. 50 Therefore, the pK a values of the compounds were measured by UV spectrophotometry using the 96-well plate methodology developed in our group. 51,52 This allowed the determination of the average pK a value of these dibasic compounds (Table 4). For several compounds, the pK a was also determined with the Sirius T3 apparatus, which allowed distinguishing the pK a values of both heterocyclic moieties within the same molecule (i.e., pyridine-2-carboxamidine, 2-aminoimidazoline, and 2aminobenzimidazole) ( Table 4).
In contrast to the predicted values (6.54−7.70, calculated with the Chemicalize ChemAxon LLC software), the experimental pK a (H 2 O, 25°C) of the bisbenzimidazole derivatives 2a−g was in the range 3.38−6.75 (i.e., % ionization at pH 7.4 ranging from 0.01 to 18.3), indicating that the compounds are mostly neutral at physiological pH (Table 4). These measured pK a values were much lower (i.e., approximately 1.5−2 pK a units) than the predicted ones, highlighting the limits of pK a prediction tools for compounds containing a large number of possible tautomeric states. 53,54 In contrast, the bis(2-aminoimidazolines) 1a−i are mostly dicationic, with experimental pK a values in the range 8.05− 10.26 (i.e., > 85.5% ionized) that is approximately >1 pK a unit higher than the predicted ones. Replacement of the 2aminoimidazoline heterocycles by pyridine-2-carboxamidine substituents led to a reduction of the pK a of the molecules by >2 pK a units, giving mean values in the range 4.19 (3h)−8.44 (19). Accordingly, the derivatives showed disparate percentages of ionization at pH 7.4 ranging from 0.2% (3h) up to 71% (3a) and 91.6% (19). The pK a SAR for these series followed the trends of previous series of bis(2-aminoimidazolines) containing similar scaffolds. 20,51,52,54 Interestingly, a positive correlation between the high pK a values of each compound and ΔT m was observed for series 2 (2c, 2e, 12, 13, and 16) and 3 (3a−d, 3f, 3h, and 18) ( Figure S3; r 2 = 0.76), indicating that an increased percentage of ionization favored binding to DNA. This is consistent with the well-established central role played by the positively charged ends of DNA MGBs in enhancing electrostatic interactions with DNA. 55 The absence of a similar correlation with series 1 is probably due to the high basicity of the bis(2-aminoimidazoline) compounds, which are almost totally dicationic at pH 7.4.
The log P of 11 representative compounds from the different series was measured, and the calculated values were also determined for all the compounds. Other physicochemical parameters are listed in Table S2. All of the new compounds display higher log P values than lead 1a, with most of them being <5. It is noteworthy that calculations tend to overestimate log P values from 0.3−0.4 (series 1) or 0.2−2.4  (Table 4). Based on these estimated values, compounds 2e (6.86), 12 (7.24), 16 (6.49), or 18 (5.15) would break the "rule of 5" (i.e., log P < 5) for this parameter, whereas the experimental log P is much lower than predicted (3.21, 4.32, 0.84, and 2.75, respectively). This is especially relevant for series 2, whose most predicted values were >5. Of note, there was no apparent correlation between the activity against T. brucei and the experimental pK a or log P of the molecules. The aqueous kinetic solubility of compounds 3a, 3d, and 18 was determined by UV spectroscopy in 96 multiwell plates ( Table 5). As expected, the solubility of these basic compounds increased substantially (2-, 16-, and 7-fold) at low pH (1.2). Altogether, these compounds fulfill the R05 for oral bioavailability and are potential candidates for in vivo studies.

Microsomal and Plasma Stability.
The microsomal stability toward metabolism by cytochrome P450 (Phase-I metabolism) and uridine glucuronosyl-transferase (UGT) (Phase-II metabolism) of the selected hit compound 3a was studied in the presence of NADPH and UDPGA. Compound 3a was apparently not metabolized by human and mouse (CD-1) liver microsomes, as we did not observe the formation of any significant metabolite during the course of the reaction, i.e., up to 120 min, indicating a half-life >2 h. In comparison, human liver microsomes rapidly metabolized (high intrinsic clearance) diclofenac under the same conditions with a half-life <30 min (Table S1). Compound 3a was not apparently modified in human serum, and it was stable for 1 h of incubation. Thus, under the experimental conditions used in this study, compound 3a is metabolically stable in human and mouse microsomal fractions and in human serum.

DISCUSSION
AT-specific DNA MGBs have a record of success as chemotherapeutic agents against human and animal infections caused by trypanosomatid parasites. 41,56,57 In particular, the bis(2-aminoimidazoline) class of DNA MGBs, such as lead 1a, are very effective in vitro and in vivo against T. brucei. 14,18,28 However, like diamidines, 58,59 these compounds are much less active against the intracellular parasite Leishmania. The high polarity (i.e., dicationic nature) of the compounds is probably not the sole reason for reduced cidal activity against intracellular parasites. In fact, Leishmania does not express particularly good transporters for diamidines, including pentamidine, 60,61 but expression of T. brucei AQP2 in Leishmania renders them just as susceptible to pentamidine as trypanosomes. 62 The same hypersensitization occurs with expression of TbAQP2 or TbAT1 in Trypanosoma congolense, which are otherwise relatively insensitive to diamidines. 63 Hence, the different parasites express transporters with different substrate preferences and affinities, which will influence the SAR from cell to cell.
In this work, we studied the effect of scaffold modifications that lower the basicity and/or increase the lipophilicity of 1a on the antiparasitic activity and DNA binding properties. For the bis(2-aminoimidazolines) 1c−i, which displayed EC 50 values in the micromolar range against T. brucei but were mostly inactive against T. cruzi and Leishmania (except 1h), the introduction of Cl, O i Pr, pyridine rings, 1,2-diphenylethane, 18 or 1,3-diphenylurea 18 did not improve the antikinetoplastid activity versus 1a. From this series, the dichloro derivative 1e, which was the most active and selective compound against both WT and pentamidine-resistant (B48) T. brucei strains, was approximately 7-times less potent than 1a against this parasite. In previous studies, we observed that the introduction of one chlorine atom in position ortho to the imidazoline rings (R 4 = Cl) increased the anti-T. brucei activity by nearly 4-fold. 20 Hence, the presence of two halogen atoms in the bis(2-aminoimidazoline) scaffold appears to be counterproductive for antiprotozoal activity.
In contrast, the same scaffold modifications for the benzimidazole series 2 (2c−g, 12, 13, and 16), led to an up to 50-fold increase in anti-T. brucei activity with respect to the unmodified N-phenylbenzamide analogue 2a. The replacement of the 2-aminoimidazoline groups by 2-aminobenzimidazole heterocycles afforded compounds with the same hydrogen bond-forming capacity as series 1, but with much lower pK a values (3.38−6.75) and higher lipophilicity (calcd. log P = 4.30−7.24). However, these structural modifications did not improve the in vitro antiparasitic activity against extracellular (T. brucei) or intracellular (T. cruzi, Leishmania) kinetoplastid parasites with respect to lead 1a or the monochlorinated analogues reported earlier. 20 The moderate selectivity indexes and the nondifferentiation against cells lacking the high affinity transporters TbAT1 and TbAQP2 for diamidines, 61,64 melaminophenyl arsenicals, and other trypanocidal drugs suggest a lack of high affinity carriers for this class of molecules in T. cruzi and Leishmania spp., consistent with earlier reports. 61 The absence of specific protozoal transporters is exacerbated by the lack of a strong accumulative driving force for the neutral bis(2-aminobenzimidazole) compounds, potentially impacting the translocation across both plasma membrane and mitochondrial membranes. A large body of evidence suggests that pentamidine and other dicationic MGBs derive their antiprotozoal selectivity from selective uptake by the parasites. 29,32,62,63,65 This concerns probably mostly uptake over the plasma membrane, but accumulation of dicationic drugs in kinetoplastid parasites is ultimately driven by the mitochondrial membrane potential, driving a buildup of the dications in the mitochondria, 61,66−68 where MGBs interfere with the functioning and replication of the kinetoplast. 69−71 On the other hand, the neutral compounds might be expected to remain cytosolic rather than accumulate disproportionally in the mitochondria and therefore act on AT-rich nuclear rather than kinetoplast DNA, albeit at a much lower level of Interestingly, the main SAR results (e.g., effect of Cl atoms) seemed to apply to both series 1 and 2. However, none of the series 2 compounds showed activity against T. cruzi or L. donovani, except 12, which had marginal activity against promastigotes of L. donovani (EC 50 = 8.7 μM), albeit with low selectivity (SI (THP-1/promast.) = 1.6). As a whole, the chemical modification 2-aminoimidazoline ↔ 2-aminobenzimidazole provided no improvement in antiprotozoal activity or DNA binding affinity, which is probably related to the low pK a of the bis(2-aminobenzimidazoles).
Consistent with earlier results from Boykin and co-workers showing that the bis(arylimidamide) analogues of furamidinebased diamidines are especially active against the intracellular parasites Leishmania and T. cruzi, 72−77 series 3 analogues were very potent inhibitors of T. cruzi and Leishmania, but also of T. brucei growth, with adequate selectivity. In particular, 3a, 3c−e, 18, and 23, which display submicromolar EC 50 values and SI (THP-1/amast.) > 20 against intracellular amastigotes of L. donovani in the same range as amphotericin B and present druglike properties, can be considered as new antileishmanial hit compounds. These values are in the same range as furamidine's bis(arylimidamide) derivative DB766 73 (i.e., N,N″-(furan-2,5-diylbis(3-isopropoxy-4,1-phenylene))dipicolinimidamide), and superior to analogues modified on the terminal group. 72 Regarding the SAR of this series, the presence of nitrogen atoms in the central scaffold had diverse effects on the activity. The introduction of two pyridine rings (3c) increased the antileishmanial activity very efficiently, especially against amastigote forms (EC 50 = 0.55 μM). In contrast, the closely related N-(pyridin-2-yl)benzamide analogue (3b), which was almost as active as 3c against promastigotes (EC 50 = 1.46 μM), was ineffective against intracellular amastigotes of L. donovani. Since 3b and 3c have similar DNA binding affinities (ΔT m = 1.8 and 2.3°C, respectively), this difference is probably related to dissimilar uptake of the compounds resulting from different membrane permeabilities or drug transport. The introduction of two chlorine atoms adjacent to the arylimidamide group (3d, 3e) or O i Pr groups (3g) on the N-phenylbenzamide scaffold also appeared as a good strategy to improve the activity against intracellular amastigotes of T. cruzi and, to a lesser extent, L. donovani. Of note, the positive effect of large O-alkyl groups was consistent with the SAR reported for furamidine-based bis(arylimidamides) series (e.g., DB766) despite being based on quite different central scaffolds. 73,78,79 The 1,2-diphenylethane (18) is another interesting scaffold that displayed excellent activities and selectivity against intracellular amastigotes of L. donovani. The latter also worked as an N-phenylbenzamide surrogate for the bis(2-aminoimidazoline) series active against T. brucei. 15,18 All of the compounds were also tested against the protozoan parasite T. vaginalis because its genome has a high content of AT base pairs, including an unusually high number of genomic repeats, 80 which may potentially be targeted by DNA MGBs.
In previous studies, Crowell et al. showed that dicationic DNA MGBs at AT sites, such as aromatic diamidines, have potential as antitrichomonal agents with activities in the micromolar range. 81 Among the compounds studied here, only seven showed a weak trichomonacidal effect (EC 50 = 11.5−32.5 μM) displaying a cytotoxic effect on mammalian cells, with the exception of 3l. We do not have an evidence-based explanation for the lack of trichomonacidal activity of these compounds, which are AT-specific DNA MGBs, but this is likely to be related to a poor uptake of these dicationic compounds in Trichomonas, as is the case for T. congolense, which is much less susceptible to diamidines than T. brucei. 63 In the present study, we did not observe a correlation between the antiparasite activity of the new compounds and the DNA binding affinity (ΔT m ). On the one hand, such a correlation between the cellular activity and the binding affinity for the target depends on several factors (e.g., cell uptake differences, intracellular distribution, and distinct cellular targets) and is not always easy to demonstrate. 69,82 On the other hand, we were able to confirm that kDNA is the cellular target of lead compound 1a in T. brucei using a combination of flow cytometry, imaging techniques, and biophysical experiments. 19 Notwithstanding, in the present work, we showed that the three series of compounds stabilize AT-rich DNA duplexes upon binding, with affinity constants to A 2 T 2 in the submicromolar to micromolar range similar to previously reported bis(2-aminoimidazolines) 19,42 and diamidines (e.g., pentamidine, DB829). 41 Most of series 1−3 compounds hold the basic structural features of MGBs, including a crescentshaped N-phenylbenzamide scaffold and positive charge(s) that can promote binding to kDNA for series 1 (pK a = 8.05− 10.26) and 3 (pK a = 6.46−8.44), contrary to series 2 compounds (pK a = 3.38−6.75), which are mostly neutral at physiological pH. Linear dichroism experiments with selected molecules from series 1 and 3 (1g, 3a) confirmed unequivocally that both compounds are minor groove binders like lead 1a. FID assays ruled out an intercalative mode of binding for 3a. These results were consistent with the SPR data showing AT base-pair specificity versus CG and, hence, a minor groove preference for these ligands. 14,19,42,47,83−87 In contrast, the fluorene derivative 23, which bound to AT and GC-containing oligonucleotides with similar K D values (9.8 and 13 μM, respectively), displayed no base-pair sequence preference, which is consistent with previous findings reporting both intercalation and minor groove binding for the bis(2aminoimidazoline) analogue of 23. 43 According to the FID assay, bis(2-aminobenzimidazole) compounds 2a, 2c, 2d, 12, 13, and 16 appeared to have an intercalative mode of binding, although a mixed "groove binding/intercalation" mode cannot be ruled out.
Despite their strong binding affinity and selectivity for ATrich DNA, the compounds reported here may have additional cellular target(s) besides kDNA. This is the case, for instance, for the dicationic DNA MGB pentamidine, whose effects against the nonkinetoplastid pathogens Plasmodium falciparum and Pneumocystis carinii point toward other cellular targets in those organisms, 88 including RNA splicing in P. carinii 89 and binding to ferriprotoporphyrin IX in P. falciparum. 90 For some compounds, the level of human cell line toxicity supports this hypothesis. Daliry et al. showed that the trypanocidal activity of bis(arylimidamides) related to DB766 did not correlate with their binding affinity to T. cruzi kinetoplast DNA. They observed that a strong affinity with kDNA per se was not sufficient to trigger the trypanocidal activity of the studied diamidines and speculated that "cell uptake differences and possibly distinct cellular targets need to be considered". 91 As far as bis(arylimidamides) are concerned, the recently discovered cytochrome P450, CYP5122A1 (sterol 14-demethylase, an antifungal azole target), which is essential for L. donovani survival, 92 could also be involved in their mode of action, as proposed by Werbovetz and co-workers. 93

CONCLUSIONS
Efforts to improve the activity of the DNA MGB lead compound 1a against intracellular kinetoplastid parasites were successful with the bis(arylimidamide) series 3. The bis-(pyridine-2-carboxamidine) 3a was definitively the best in class, showing excellent activity and selectivity against T. brucei, T. cruzi, and L. donovani, as well as druglike properties. Hence, compound 3a is a candidate for further in vivo studies. Nonetheless, other compounds from this series were also highly efficacious in terms of activity and selectivity against intracellular amastigotes of T. cruzi (3a, 3d, 3g, and 23) and L. donovani (3c−e, 3g, 18, and 23).
We showed that compounds from series 1 and 3 are DNA MGBs, whereas compounds from series 2 appear to intercalate, even though a mixed mode of binding cannot be ruled out. Therefore, kDNA is a probable target of the new compounds reported here, although their exact MoA remains to be determined. The correlation observed between the pK a and the DNA binding affinities (ΔT m ) for series 2 and 3 showed that the ionization state of these molecules is a driving force that influences both the uptake into the parasites and the extent and mode of binding to DNA.

Chemistry.
All the commercial chemicals were obtained from Sigma-Aldrich, Fluorochem, Acros Organics, or Alfa Aesar and were used without further purification. Deuterated solvents for NMR use were purchased from Merck (Sigma-Aldrich). Dry solvents were either obtained from Acros Organics and Sigma-Aldrich in SureSeal bottles or were distilled using standard procedures, according to Vogel's Textbook of Practical Organic Chemistry. Solvents for synthesis purposes were used at GPR grade. Reactions heated by microwaves were realized in a Biotage Initiator microwave oven reactor (frequency: 2045 GHz). Chromatographic columns were run using Silica gel 60 (230−400 mesh ASTM) or Aluminum Oxide (activated, Neutral Brockman I STD grade 150 mesh). Analytical TLC was performed using Merck Kieselgel 60 F 254 silica gel plates or Alugram© Alox N/UV 254 aluminum oxide plates. TLC plates were visualized under UV light (254 and 365 nm) and/or revealed with staining reagents (i.e., iodine, phosphomolybdic acid, and ninhydrin). Flash chromatography was performed in an Isolera One (Biotage) with Isolera 3.3.0 version, using Biotage Sfar columns (silica D, duo 60 μm) or SiliaSep Flash cartridges (SiliCycle) 40−63 μm, 60 Å. Reverse-phase chromatography was performed using LiChroprep RP-18 (25−40 μm), Merck, and Claricep Screw-on Flash C18 columns (spherical, 20−35 μm, 100 Å, 122 g). Centrifugal TLC was carried out with the Chromatotron using circular glass plates prepared with silica gel 60 PF 254 containing gypsum. 1 H-and 13 C NMR spectra were recorded on a Bruker Avance-300, a Bruker DPX-400 Advance spectrometer, and a Varian-500. Chemical shifts of the 1 H NMR spectra were referenced to the residual peak of the deuterated solvent: CDCl 3 (δ 7.26 ppm), methanol-d 4 (δ 3.31 ppm), and DMSO-d 6 (δ 2.5 ppm). Chemical shifts of the 13 C NMR spectra were referenced to CDCl 3 (δ 77.16 ppm), methanol-d 4 (δ 49.0 ppm), and DMSO-d 6 (δ 39.52 ppm). Signal multiplicity for 1 H NMR is defined as singlet (s), doublet (d), triplet (t), quartet (q), multiplet (m), and br (broad signal). Coupling constants J are expressed in hertz (Hz). LC−MS spectra were recorded on a WATERS apparatus integrated with a HPLC separation module (2695), PDA detector (2996), and a Micromass ZQ spectrometer using electrospray ionization (ESI + ). Analytical HPLC was performed with a SunFire C18−3.5 μm column (4.6 mm × 50 mm). Mobile phase A: CH 3 CN + 0.08% formic acid and B: H 2 O + 0.05% formic acid. UV detection was carried out over 190 to 440 nm. Melting points were determined using a Mettler Toledo MP70 digital melting point apparatus and are uncorrected. Some of them show a broad melting range due to the low degree of crystallinity. Elemental analysis was carried out at the Microanalysis Laboratory, "Manuel Lora Tamayo" Organic Chemistry Centre� CSIC. High-resolution mass spectra were recorded at the Elemental Microanalysis Unity at the Pharmacy College, Complutense University of Madrid. All of the biologically tested compounds were ≥95% pure by HPLC except 2d (93%), 3b (94%), 3b_b (91%), 3g (91%), and 20 (93%). (1c−i). 5.1.1.1. Method A. Trifluoroacetic acid (2 mL) was added to a cooled solution (ice−water bath) of the Boc-protected compound 6c−i (1 mmol, 60−220 mg scale) dissolved in CH 2 Cl 2 (2−6 mL). The resulting solution was stirred for 4 h at the same temperature. Trifluoroacetic acid and dichloromethane excesses were evaporated under vacuum. This process was repeated by dissolving the crude residue in CH 2 Cl 2 and evaporating the solvents under a high vacuum. The sticky solid was crushed with diethyl ether to precipitate the product as a powder.      2a, 2c−g, 12, 13, and  16). 5 .1.2.1. Method B. The reaction was performed in a KIMAX tube. A solution of isothiocyanate (7c−g, 10, 11, 13; 1 equiv) and ophenylendiamine (2.2 equiv) in anhydrous DMF (4 mL) was stirred at 0°C (ice−water bath) until complete formation of the thiourea intermediate (checked by TLC and HPLC−MS). Then, EDC hydrochloride (2.5 equiv) solid was added in one step, and the reaction mixture was allowed to stir at 60°C for 24 h. Ice was added to the crude reaction mixture, and the tube was shaken vigorously. The solid precipitate was collected and purified by centrifugal PTLC using silica plates.

4-(Picolinimidamido)-N-(4-(picolinimidamido)phenyl)benzamide (3a).
The reaction was performed following method C1 with 5a (0.5 g, 2.2 mmol) and 17 (1.97 g, 5.5 mmol), in 6 mL of EtOH/CH 3 CN (3:1). The resulting reaction mixture was allowed to stir at room temperature for 48 h. The solvents were removed under vacuum and the crude product was purified by silica gel column chromatography using n-hexane:EtOAc (20:80 → 0:100). The free base of 3a was obtained as a yellowish solid (220 mg, 23%). 1 Salt (3b). Method C1. Diamine 5b (86 mg, 0.38 mmol) was suspended in anhydrous acetonitrile (2 mL) and anhydrous EtOH (6 mL). The flask was cooled with an ice−water bath, and solid 17 (284 mg, 0.79 mmol) was added at once to the suspension. The thick, yellowish slurry was stirred at 0°C and allowed to warm up to room temperature. After 5 days of stirring at room temperature, Et 2 O was added to the reaction mixture. The precipitate was collected on a fritted plate and rinsed with Et 2 O. The precipitate was dissolved in EtOH (20 mL), and the cooled solution (ice−water bath) was basified with 1 N NaOH until pH ≈ 10. The solution was concentrated under vacuum, and the product was partitioned between water (25 mL) and EtOAc (40 mL). The aqueous phase was extracted with EtOAc (2 × 50 mL), and the combined organic extracts were washed with brine, dried (Na 2 SO 4 ), and evaporated to give a crude yellow residue. Silica chromatography (5 g SI cartridge) with CHCl 3 /NH 3(sat.) -MeOH: 0 → 5% yielded a mixture (≈75/25, 38 mg) of the expected product 3b (M = 436) and the monosubstituted product 3b_b (M = 332). The mixture was dissolved in CH 2 Cl 2 / MeOH (2 mL), cooled with an ice bath, and treated with HCl(sat.)dioxane solution for 1 h with gentle stirring. The monosubstituted product 3b_b that precipitated from the reaction mixture was collected by filtration and rinsed with Et 2 O to give a colorless solid (20 mg, 12%). The filtrate was evaporated under a vacuum, and product 3b was recrystallized from CH 2 Cl 2 /MeOH at −20°C. The product was rinsed with Et 2 O to yield 3b as a yellowish solid (10 mg, 6%). 1 165.3, 160.2, 151.8, 149.9, 148.5,  145.9, 144.3, 138.5, 138.4, 138.3, 136.5, 130.0, 129.8, 129.0, 128.7,  125.7, 124.1, 123.9, 122.6, 119.8, 119 anhydrous EtOH (9 mL). The flask was cooled with an ice bath and solid 17 (428 mg, 1.19 mmol) was added at once to the suspension. The thick, brownish slurry was stirred at 0°C and allowed to warm up to room temperature. After 10 days of stirring at room temperature, the reaction was concentrated under vacuum, and Et 2 O (70 mL) was added to the reaction mixture. The precipitate was collected on a fritted plate and rinsed with Et 2 O to give a beige solid (257 mg). The solid was suspended in EtOH (10 mL), and the cooled solution (ice bath) was basified with 1 N NaOH until pH ≈ 10. The mixture was partitioned between water (20 mL) and EtOAc (50 mL). The aqueous phase was extracted with EtOAc (2 × 50 mL), and the combined organic extracts were washed with brine, dried (Na 2 SO 4 ), and evaporated to give an orangish residue. Silica chromatography (5 g SI cartridge) with CHCl 3 /NH 3(sat.) -MeOH: 0 → 5% yielded the pure monosubstituted product 3c_c as a brownish solid ( 5.1.3.7. Method C2. Diamine 5i−l (1 equiv) and 17 (4.5 equiv) were dissolved in anhydrous DMF (2.5 mL) in a KIMAX tube. The reaction mixture was stirred under an argon atmosphere for several days. Water (3 mL) was added, and the mixture was extracted with EtOAc and CH 2 Cl 2 to remove organic byproducts. The aqueous layer was concentrated in vacuo to yield a brownish crude oil. Then, hexane was added, and the precipitate was collected by filtration over a Buchner funnel. The precipitate was redissolved in water, and a saturated aqueous NaHCO 3 solution was added to precipitate the product. The precipitate was filtered over a filter plate, giving 3i−l as white powder.  -1,2-diylbis(4,1-phenylene))dipicolinimidamide (18). The reaction was performed following the general Method C1 with commercial 4,4′-diaminobibenzyl (106 mg, 0.5 mmol) and 17 (449 mg, 1.25 mmol) suspended in a 3:1 mixture of anhydrous EtOH/CH 3 CN (16 mL). The mixture was allowed to stir at room temperature overnight. The solvent was eliminated under a vacuum, and the solid crude was purified by silica column chromatography using hexane/EtOAc (100:0 → 10:90) as the elution system. Product 18 was obtained as a yellowish solid (51 mg, 24%).  (20). A suspension of 5a (150 mg, 0.7 mmol) and picolinyl chloride (294 mg, 1.7 mmol) in dry THF (8 mL) was stirred at room temperature. Et 3 N (140 mg, 1.4 mmol) was added dropwise, turning the brown into a whitish solution. After 2 h, the reaction mixture was diluted with CH 2 Cl 2 whereupon a whitish precipitate appeared. The product was collected by filtration as a whitish solid (271 mg, 94%).  (3 equiv) was added to a cooled solution (ice−water bath) of di-tertbutyl 2-thioxoimidazolidine-1,3-dicarboxylate 23 (3 equiv), diamine 5c−i (1 equiv., 100−300 mg scale), and anhydrous triethylamine (7 equiv) in dry DMF (3 mL/0.7 mmol) under an argon atmosphere. The reaction mixture was stirred for 1 h at 0°C. Then, the ice−water bath was removed, and the reaction mixture was stirred at 60°C for the time specified in each case. The reaction mixture was diluted with CH 2 Cl 2 and filtered over celite using a mixture of CH 2 Cl 2 : MeOH  (1:1, 200 mL). The solvent was removed under vacuum, and the crude product was diluted with EtOAc (100 mL) and extracted with water (3 × 150 mL). The organic phase was washed with brine, dried over MgSO 4 , and evaporated under vacuum. The pure compounds were obtained by crystallization or by silica centrifugal thin-layer chromatography, as specified in each case.

butoxycarbonyl)picolinimidamido)-3-chlorophenyl)carbamoyl)-3-chlorophenyl)amino)(pyridin-2-yl)methylene)carbamate (24g).
The reaction was performed following the general Method F with diamine 5g (109 mg, 0.37 mmol), 21 (235 mg, 0.93 mmol), HgCl 2 (242 mg, 0.89 mmol), and Et 3 N (150 mg, 1.5 mmol). The crude solid (539 mg) was purified by circular chromatography using a 2 mm silica plate previously neutralized with n-hexane (235 mL) and Et 3 N (15 mL); CH 2 Cl 2 / EtOAC (95:5 → 90:10) was used as the elution system. The product was obtained as a yellowish solid (32 mg, 12%  (3 mL) was stirred at room temperature, followed by the dropwise addition of Et 3 N (61 mg, 0.6 mmol). The reaction mixture was stirred 6 days at room temperature. The crude was filtered through a pad of florisil rinsing with CH 2 Cl 2 and MeOH. The filtrate was extracted with H 2 O, washed with brine, and dried over Na 2 SO 4 . The organic phase was concentrated and purified by circular chromatography using a 2 mm silica plate previously neutralized with n-hexane (235 mL) and Et 3 N (15 mL); hexane/EtOAc (2:3) was used as the elution system. The product was obtained as a yellowish solid (56 mg, 50%). 1 13   5.1.7. Spectrophotometric pK a Measurements. The pK a was measured using the 96-well microtiter plate method reported earlier. 51,52 Briefly, the compounds were dissolved in DMSO to a concentration of 5 mM (stock solution) or less, ensuring that the maximum absorbance of the compound was below 1.5 AU during the assay. Each line of the UV-transparent 96-well microplate (Thermo Scientific Nunc) was loaded with 196 μL of buffer solutions of increasing pH. Then, 4 μL of the compound stock solutions were added to each well with a micropipette (the resulting analyte solution was premixed with the micropipette). One blank solution was prepared for each buffer by adding 4 μL of DMSO to 196 μL of the corresponding buffer solution (i.e., free of analyte compounds) in the well. The 96-well plate was loaded into the UV spectrophotometer (CLARIOStar Plus), incubated at 25°C, and shaken at 700 rpm for 5 min in a double orbital mode before the reading was performed. UVspectra scans were recorded between 200 and 600 nm at 2 nm resolution. The raw UV-spectra scans were imported to the Excel template provided in Dardonville et al., 51 and the data were processed, as reported. 52,95 Buffer solutions (pH from 1 to 12) of constant ionic strength (0.1 M KCl) were prepared according to the online buffer calculator: https://www.biomol.net/en/tools/ buffercalculator.htm. The number and range of buffer solutions (e.g., every 0.2, 0.5, or 1 pH unit) needed to determine pK a values were adjusted depending on the compound tested. In general, a first screening with 12 buffers ranging from 3 to 12 should give an approximate pK a value, which can be refined when repeating the experiment using buffers within ±2 pH units of the pK a value. p s K a values were measured with the Sirius T3 apparatus at 25°C in a 0.15 M aqueous KCl solution under a nitrogen atmosphere using the UV-metric method, DMSO stock solutions of the samples, and methanol as a cosolvent. The aqueous pK a (at 0% co-solvent) was worked out using the Yasuda-Shedlovski extrapolation.
Log P values were determined with the Sirius T3 apparatus at 25°C using the pH-metric method. 5.1.8. Kinetic Solubility Measurements. Ten μL of 10 mM stock solution (DMSO) of the compound (3a, 3d, and 18) was added to 990 μL of buffer solution (pH 1.2, 5.5, or 7.4) in an Eppendorf tube (DMSO final concentration was 1% (v/v). Blank samples (1% DMSO in buffer) were also prepared. The samples (prepared in triplicate) were shaken at room temperature for 2 h and centrifuged at 135 rpm for 15 min. 170 μL of the supernatant were transferred to a 96-well plate and diluted by adding 30 μL of CH 3 CN. The absorption of the compounds and blank samples was read at 269 nm with a Multiscan Spectrum (Thermo Electron) UV spectrophotometer. The blank contribution was subtracted, and the solubility was worked out from a calibration curve obtained between 0 and 500 μM for each sample. Results are expressed as mean ± SD. Frozen samples of lyophilized DNA oligonucleotides were removed from the fridge and warmed to room temperature over 5 min. These oligonucleotides were dissolved in 1 mL of 10 mM sodium phosphate buffer (pH 7.0) containing NaCl (100 mM) (oligonucleotide stock solution). The DNA concentration was measured using a Nanodrop 2000 spectrophotometer. CD spectra were recorded on a Jasco 8 J-810 Spectropolarimeter using a 1 mm path-length quartz cuvette. The oligonucleotide (AT) 4 or (CG) 4 was diluted at 60 μM concentration in a volume of 200 μL of oligonucleotide stock solution. Test compounds were prepared as 120 μM stock solutions in DMSO. Scans from 320 to 220 nm were performed with a 50 nm/min scanning speed. For each spectrum, an average of three spectra was taken, and the spectrum of the corresponding buffer was subtracted for baseline correction. The melting curves were obtained by recording the change of the molar ellipticity at 270 nm in a range of temperatures from 5 to 85°C. The temperature was controlled using a Jasco Peltier, with the rate of temperature rising at 40°C/h. The resulting melting temperatures were calculated by fitting the denaturing curves with the program Origin Pro 6.0.

SPR-Biosensor
Assays. The compounds were dissolved in DMSO and diluted to the required concentrations with the buffer containing 10 mM MES, 100 mM NaCl, 1mM EDTA, and surfactant P-20 at 0.005% (v/v) at pH 6.25. SPR binding experiments were performed at 25°C with a Biacore X100 apparatus (Biacore GE) using the filtered buffer described above. The DNA hairpins were immobilized on a streptavidin-derivatized gold chip (SA chip from Biacore) by injection of a 25 nM hairpin DNA solution with a flow rate of 1 μL/min until ∼400 RU were reached. Flow cell 1 was used as a reference, while flow cell 2 was immobilized with the hairpins in different chips. Direct binding was measured by the injection of increasing concentrations of each compound over the immobilized DNA surfaces at a flow rate of 50 μL/min for a period of 60 s, followed by a dissociation period of 120 s. Regeneration of the surface was made with NaCl 200 mM/NaOH 10 mM using a flow rate of 10 μL/min for 30 s. The binding affinity was determined by fitting the results to a two-site or one-site (K 2 = 0) binding model according to the equation where r is the mole of bound compound per mole of DNA hairpin duplex, C f is the free concentration at equilibrium, and K 1 and K 2 the microscopic binding constants.

LD Experiments.
The spectra were recorded for natural DNA (salmon testes) titrated with compounds 1a in phosphate buffer at 25°C , following the methodology previously reported by Rozas' group. 47 Titrations were carried out with a DNA concentration of 378.8 μM working with a Bp/D ratio of 0, 1, and 5, varying the Bp/D ratio from 5 to 1 over 2 additions (Figure 2). Since the experiment was done in a very small volume (70 μL), the concentrations of DNA and the compounds were kept constant to avoid the dilution effect. 47 Each solution was then prepared individually, and the corresponding spectra were recorded.

5.3.2.
In Vitro Activity against L. donovani. Leishmania donovani (MHOM/ET/67/HU3) was employed in the screening of synthesized compounds. This line was grown at 28°C in RPMI 1640modified medium (Invitrogen) supplemented with 10% heatinactivated fetal bovine serum (hiFBS) (Invitrogen), as described. 96 The sensitivity of Leishmania promastigotes to the different compounds was determined after incubation for 72 h at 28°C in the presence of increasing concentrations of the compounds. The concentration of compound required to inhibit 50% of parasite growth (EC 50 ) was calculated using the MTT colorimetric assay, as described previously. 97 For the compound susceptibility analysis of the intracellular amastigote forms of L. donovani clinical isolates, stationary-phase promastigotes were used to infect macrophagedifferentiated THP-1 cells at a macrophage/parasite ratio of 1:10. After overnight infection at 35°C with 5% CO 2 in RPMI 1640 medium plus 5% hiFBS, extracellular parasites were removed by washing twice with PBS buffer. Infected macrophages were incubated with different concentrations of compounds in RPMI 1640 medium plus 10% hiFBS at 37°C in a 5% CO 2 atmosphere for 72 h. Following incubation, the medium of the samples was removed and SDS 0.05% was added for around 10 min until lysis of macrophages occurred, liberating intact and viable intracellular amastigotes. We diluted the samples 1/10 with RPMI modified medium supplemented with 10% hiFBS and incubated the samples for 4−7 days at 28°C in order to let the intracellular amastigotes transform into promastigote forms. Finally, the resazurin colorimetric assay was used to determine the EC 50 in a similar way, as previously described. 98 Data are means ± standard deviations from three independent experiments (n = 3).

Human Cell Lines Culture and Determination of Cellular
Toxicity. The human myelomonocytic cell line THP-1 was grown in RPMI-1640 supplemented with 10% hiFBS, 2 mM glutamate, 100 U/ mL penicillin, and 100 μg/mL streptomycin at 37°C and 5% CO 2 . Five ×10 5 THP-1 cells per well in 24-well plates were differentiated to macrophages with 20 ng/mL of PMA treatment for 48 h followed by 24 h of culture in complete fresh medium. The cellular toxicity of all compounds was determined using the colorimetric MTT-based assay, 99 as described for Leishmania promastigotes, with the exception of the incubation temperature, which was 37°C in this case. 5.3.4. In Vitro Activity against T. brucei. EC 50 values were determined for bloodstream forms of both strain Lister 427 100 and the multidrug-resistant strain B48, which is derived from Lister 427 by knockout of the TbAT1 aminopurine transporter 101 and in vitro adaptation to high concentrations of pentamidine. 30 Both strains were cultured in full HMI-9 media (Gibco) supplemented with 10% hiFBS at 37°C/5% CO 2 , as described. 102 The drug sensitivity was determined using a resazurin-based assay exactly as described previously in 96 wells, with 23 doubling dilutions and no drug control for each compound. Each well was seeded with 2000 cells and incubated for 70 h with the drug before the addition of the resazurin indicator dye and a further incubation of 24 h (n = 3). 20 Results are expressed as the mean value of EC 50 ± SEM (standard error of the mean). 5.3.5. Unspecific Cytotoxicity Assays. Cytotoxicity against human embryonic kidney (HEK) cells was determined exactly as described previously. 103 Briefly, cells were grown in Dulbecco's modified Eagle's medium (DMEM; Sigma) supplemented with 10% New-born Calf Serum (Gibco), 1% of a penicillin/streptomycin solution (Gibco), and 1% (vv) of 200 mM glutamax (Gibco) at 37°C/5% CO 2 . 96-well plates were seeded with 30,000 cells/well in 100 μL medium and incubated for 24 h to allow adhesion, after which an equal value of a drug serial dilution was added and the plates incubated for 30 h prior to the addition of resazurin solution, followed by a final 24 h incubation. Plates were read on a FLUOstar Optima fluorimeter (BMG Labteach, Durham, NC), and the data was analyzed using Prism 8.0 (GraphPad).
5.3.6. T. cruzi Epimastigote Susceptibility Assays. The activity profile of the compounds was explored on the T. cruzi CL-B5 lacZ strain (DTU TcVI) by applying the screening procedure previously described. 104,105 Briefly, axenic cultures of log-phase epimastigotes in the Liver Infusion Tryptose (LIT) medium were seeded in 96-well microplates at a density of 2.5 × 10 5 parasites/mL and treated with compounds for 72 h at 28°C. Stock solutions of either the studied compounds or the reference drug benznidazole were prepared in DMSO and added to the parasite cultures at a final concentration of the solvent lower than 0.2% v/v. Afterward, 50 μL/well of the chromogenic substrate CPRG prepared in 0.9% Triton X-100 (pH 7.4, final concentration 200 μM) was added. After 3 h of incubation at 37°C, absorbance was read at 595 nm (ELx808 ELISA reader, Biotek Instruments Inc.). Each assay was run in similar conditions three times separately (n = 3).
5.3.7. Unspecific Cytotoxicity Assays. The potential toxic effect induced by the studied compounds was investigated on cultures of L929 cells maintained in MEM (Sigma-Aldrich) medium at 37°C and 5% CO 2. Accordingly, 10,000 cells/well were distributed in 96well plates and incubated for 72 h (37°C, 5% CO 2 ) within the studied compounds previously dissolved in DMSO; the final concentration of the solvent in cell cultures was up to 1% v/v, which is not toxic to the cells. 106 Then, 20 μL of a resazurin solution prepared in PBS (pH 7.0, 2 mM) was added per well. After 3 h of incubation at 37°C with 5% CO 2 , fluorescence intensity was read at 535 nm (excitation) and 590 nm (emission) (Infinite 200 multifunctional microplate reader, Tecan). Each assay was performed separately three times (n = 3). 104,105 5.3.8. T. cruzi Amastigotes Susceptibility Assays. Only those compounds with an activity profile on epimastigotes similar to that of benznidazole were moved to a more specific assay on the intracellular form of the parasite. 59 First, 10,000 L929 cells/well were seeded in 48-well plates and, after attachment, infected with CL-B5 lacZ tissue culture-derived trypomastigotes at a ratio of 1:6 (cell/parasite). After 24 h of incubation at 33°C with 5% CO 2 , nonpenetrated parasites were rinsed with PBS, and then infected cultures were treated with compounds diluted in fresh MEM for 7 days at the same conditions of temperature and humidity. Finally, 50 μL of CPRG prepared in 3% Triton X-100 (pH 7.4, final concentration 400 μM) was added, and after 3 h of incubation at 37°C, absorbance was read at 595 nm (Infinite 200 multifunctional microplate reader, Tecan). Each assay was run similarly for three (n = 3). Results are expressed as the mean value of EC 50 ± SD (standard deviation) (SPSS, v20, IBM). 5.3.9. Trichomonacidal Assays. The antiparasitic effect of the compounds was evaluated against the T. vaginalis isolate JH31A#4 from the American Type Culture Collection (ATCC). The parasite was cultured in TYM (trypticase-yeast extract-maltose) medium and supplemented with 10% hiFBS and antibiotics. The in vitro screening process was executed following the sequential procedure described previously. 27, 107 The compounds, diluted in DMSO at different concentrations, were added to 1 × 10 5 T. vaginalis cells/mL cultures in the exponential growth phase. After 24 h in contact with the parasites at 37°C and 5% CO 2 , 200 μL of each tube were seeded in sterile 96-well flat-bottomed microplates. Then, culture media was discarded from the plates by centrifugation, and the parasites were subsequently resuspended in 200 μL of sterile phosphate buffered saline (PBS). The antiparasitic activity of each compound was measured by a fluorometric method using the redox dye resazurin. For that, each well was incubated for 1 h in contact with 20 μL of the redox dye (3 mM stock solution in PBS) at 37°C and 5% CO 2 . Finally, the fluorescence was read in an Infinite 200 Tecan fluorometer at λ excitation 535 nm and λ emission 590 nm. Each plate included a growth control and a positive control in which the reference drug metronidazole was evaluated at its MIC 100 concentration (24 μM). Each experiment was performed in triplicate and repeated at least two times (n ≥ 2). 5.3.10. Unspecific Cytotoxic Assays in the T. vaginalis Model. Compounds with trichomonacidal effects (i.e., 3c, 3e, 3g, 3i, 3j, 3l, and 16) were evaluated at the same concentrations against African green monkey kidney epithelial cells (Vero CCL-81, ATCC). Cells were previously cultured in RPMI-1640 medium supplemented with 10% of hiFBS and antibiotics in a humidified atmosphere at 37°C and 5% CO 2 . The cytotoxicity experiments were executed as reported previously. 27,104,108 Briefly, 5 × 10 4 cells/well were incubated in 96well flat-bottom microplates for 6 h at 37°C and 5% CO 2 . Then, the compound was added at the same concentrations used for the susceptibility assays with T. vaginalis and incubated with mammalian cells for 24 h at 37°C and 5% CO 2 . Then, 20 μL of resazurin (1 mM stock solution in PBS) was added per well. After 3 h of incubation, the plates were read in the fluorometer at λ excitation 535 nm and λ emission 590 nm (Infinite 200, Tecan). Each assay was performed at least two times (n > 2). 63 5.3.11. Statistical Analysis. Statistical significance was calculated using the Student's unpaired-t test. Differences were considered significant at a P value of <0.05.