Synthesis, SAR, and Docking Studies Disclose 2-Arylfuran-1,4-naphthoquinones as In Vitro Antiplasmodial Hits

A total of 28 lapachol-related naphthoquinones with four different scaffolds were synthesized and spectroscopically characterized. In vitro antiplasmodial activity was assayed against the chloroquine-resistant Plasmodium falciparum W2 strain by the parasite lactate dehydrogenase (pLDH) method. Cytotoxicity against Hep G2A16 cell was determined by the MTT assay. All compounds disclosed higher in vitro antiplasmodial activity than lapachol. Ortho- and para-naphthoquinones with a furan ring fused to the quinonoid moiety were more potent than 2-hydroxy-3-(1′-alkenyl)-1,4-naphthoquinones, while ortho-furanonaphthoquinones were more cytotoxic. Molecular docking to Plasmodium targets Pfcyt bc1 complex and PfDHOD enzyme showed that five out of the 28 naphthoquinones disclosed favorable binding energies. Furanonaphthoquinones endowed with an aryl moiety linked to the furan ring are highlighted as new in vitro antiplasmodial lead compounds and warrant further investigation.


Introduction
Malaria, a tropical disease caused by protozoa Plasmodium spp., is still a serious public health problem, affecting mostly Africa (88%) [1].P. falciparum is the main agent responsible for mortality and morbidity in endemic tropical regions.Drug-resistance to available agents, including artemisinin ( , Figure 1), is a serious limitation to chemotherapeutics [2], the main strategy for combating malaria.Aiming to control selection of artemisinin resistant parasites, ACT (Artemisinin-Based Combination Therapy) was developed and is the first-line treatment recommended by WHO [3].
Lapachol ( , Figure 1), a prenylnaphthoquinone occurring in Central and South American species of Handroanthus spp.(synon.Tabebuia, Bignoniaceae family) [4] and first isolated from Tabebuia avellanedae, in 1882 by Paternó [5], was used as an antiplasmodial drug during World War II, when there was a shortage of quinine, the only available antimalarial agent at that time [6].Since the pioneering work of Wendel (1946), who demonstrated that lapachol ( , Figure 1) was active against P. vivax [7], numerous publications reported on the antiplasmodial activity of naphthoquinones leading to the development of atovaquone ( , Figure 1), a totally synthetic hydroxy-1,4-naphthoquinone [4].Atovaquone, in combination with proguanil ( , Figure 1) (Malarone5), is a recommended regimen for prophylaxis in areas of chloroquine resistance and for treatment of uncomplicated falciparum malaria [8].
Previous studies demonstrated that atovaquone ( , Figure 1) is a potent and specific inhibitor of the mitochondrial electron transport at the level of cytochrome bc1 complex (Pf cyt bc1) [9][10][11][12].This inhibition results in the collapse of the mitochondrial membrane potential that causes shutdown of mitochondrial metabolism and blockage of the dihydroorotate dehydrogenase (Pf DHOD) enzyme, which participates in the fourth step of the de novo pyrimidine biosynthesis, which is essential to parasite nucleic acids synthesis, resulting in parasite death [12,13].
Atovaquone has a broad spectrum of activity against various apicomplexan parasites and mutations in the Pf cyt bc1 gene were shown to be associated with resistance to atovaquone in various organisms, such as TM90C2B P. falciparum strain, Toxoplasma, and Pneumocystis jirovecii [12][13][14][15].Therefore, atovaquone resistant P. falciparum strains represent limitations to its clinical use in malaria endemic areas and are a motivation for the synthesis of atovaquone analogs [16,17].
The present paper reports the results of an work aiming to investigate the in vitro antiplasmodial activity against the resistant strain of P. falciparum W2 and the cytotoxicity to Hep G2A16 cell cultures of naphthoquinones with different scaffolds, including ortho-and para-derivatives represented by ortho-FNQs ( ) para-FNQs ( ) supporting alkyl and aryl groups at the furan ring that were synthesized from HNQs ( ) obtained from lawsone ( ), as shown in the retrosynthetic scheme (Figure 2).Nine aryl derivatives ( , , f-h) were designed as atovaquone analogs.Another group of naphthoquinones, the aminopropoxynaphthoquinones (ANQs) a-c, derived from (Figure 2), was synthesized aiming to evaluate the influence of the amino group in the antiplasmodial activity of simple naphthoquinones.The proposed synthetic work was carried out leading to 28 naphthoquinones of four different scaffolds that were assayed for in vitro antiplasmodial activity against chloroquineresistant P. falciparum (W2) and for cytotoxicity to Hep G2A16 cells.Molecular docking simulations for interaction of the synthesized compounds were performed aiming to evaluate the correlation of binding energies with structure activity relationship (SAR) data.Besides P. falciparum cytochrome bc1 complex (Pf cyt bc1) that has been proven to be the target for atovaquone [9,11,12], docking simulations were also performed against P. falciparum dihydroorotate dehydrogenase enzyme (Pf DHOD) that is considered an attractive antimalarial chemotherapeutic target [16].

Materials and Methods
Melting points were determined using a MQAPF-307 melting point apparatus (Microquimica, Brazil) and remained uncorrected.The progress of the reactions was monitored by thin layer chromatography (TLC) in silica gel plates.Column chromatography was performed over silica gel (60-230 mesh).Infrared spectra were recorded on a Spectra One Perkin-Elmer spectrophotometer, fitted with a Paragon ATR accessory.Mass spectra (HRMS) were recorded on a Shimadzu GC MS-QP5050A instrument using direct insertion together with the electrospray ionization method and a quadrupole analyzer.The 1 H, 13 C, and DEPT 135 nuclear magnetic resonance (NMR) spectra and two dimensional NMR spectra-correlations spectroscopy (COSY), heteronuclear single-quantum correlation (HSQC), and heteronuclear multiple bound correlation (HMBC)-were recorded on a Bruker Avance DPX 200 and a DPX 400 spectrometer at 200 and 400 MHz using CDCl 3 and DMSOd 6 as the solvent and Tetramethylsylane (TMS) as the internal standard, unless otherwise stated.The NMR data are presented as follows: chemical shift in ppm, multiplicity, number of protons,  in Hz, and proton assignments.Multiplicities are indicated by the following abbreviations: s (singlet), d (doublet), dd (double doublet), t (triplet), q (quartet), m (multiplet), qn (quintet), st (sextet), ht (heptet), and bs (broad singlet).( a-h).The appropriated aldehyde (86 mmol) was added to a solution of lawsone (17 mmol) in glacial acetic acid (20 mL), followed by concentrated HCL (15 mL).The reaction mixture was kept under magnetic stirring and reflux for 40 min, at which time the TLC showed the consumption of lawsone.Subsequently, water was added to the reaction mixture (50 mL), followed by extraction with dichloromethane.The organic layer was extracted with aqueous sodium carbonate (2.5%), and the water layers were combined, neutralized with HCl 2N, and extracted with dichloromethane.The organic layer was dried with anhydrous Na 2 SO 4 and then filtered.The solvent was removed under lower pressure, and the residue was fractionated in a silica gel column chromatography.Fractions eluted with a mixture of hexane/dichloromethane (7 : 3) afforded 2-hydroxy-3-(1 alkenyl)-1,4-naphthoquinones ( a-h, Scheme 1).

Furanonaphthoquinones: FNQs ( a-h and a-h).
Hg(OAc) 2 (4 mmol) in glacial acetic acid (15 mL) was added to a solution of HNQ (2 mmol) in glacial acetic acid (15 mL).The reaction mixture was kept under magnetic stirring at room temperature for 30 min and at 80 ∘ C for 15 min.Subsequently, diethyl ether was added to the reaction mixture.The solid was separated by filtration, and the solution solvent was removed, resulting in a crude product.To obtain the ortho-FNQs a-h, the crude product was dissolved in ethanol (20 mL), and HCl 2N (20 mL) was added.The mixture was kept under magnetic stirring and reflux for 15 min.Subsequently, water (30 mL) was added to the reaction mixture, followed by extraction with ethyl acetate.The organic layer was extracted with aqueous sodium bicarbonate, rinsed with water, and dried with anhydrous Na 2 SO 4 .The solvent was removed, and the residue was chromatographed through a silica gel column.Elution with a mixture of hexane/dichloromethane (8 : 2) led to the orthofuranonaphthoquinones ( a-h, Scheme 1).To obtain the para-FNQs a-h, the crude product was dissolved in ethanol (20 mL), and concentrated HCl (20 mL) was added in this solution.The mixture was kept under magnetic stirring and reflux for 3 h.Water (30 mL) was then added to the reaction mixture, followed by extraction with ethyl acetate.The organic layer was extracted using an aqueous sodium bicarbonate, rinsed with water, and dried with anhydrous Na 2 SO 4 .The solvent was removed, and the residue was fractionated in a silica gel column chromatography.Elution with a mixture of hexane/dichloromethane (9 : 1) led to the para-FNQs a-h (Scheme 1).

2-(Oxiran-2-ylmethoxy)naphthalene-1,4-dione ( ).
Lawsone (6 mmol) was added to a round bottom flask containing sodium hydride (11 mmol) and dimethylformamide (5 mL).The reaction mixture was kept under stirring at room temperature for 30 min, after which time epichlorohydrin (30 mmol) was added to the mixture, which was kept under stirring and reflux for 3 h, when TLC showed the consumption of lawsone.Ethyl acetate (50 mL) was added to the reaction mixture, followed by the extraction with a solution of NaCl.The organic layer was dried with anhydrous Na 2 SO 4 .The solvent was removed, and the residue was purified by silica gel column chromatography.The product was eluted with a mixture of hexane/ethyl acetate (9 : 1), leading to compound (Scheme 2).

Aminopropoxynaphthoquinones: ANQS ( a-c
).An appropriated amine (0.2 mol) was added to a round bottom flask containing compound (4 mmol) and MeOH.The reaction solution was kept under stirring and reflux for 3 h, when TLC showed the consumption of starting materials.Ethyl acetate (20 mL) was then added to the mixture, followed by extraction with HCl 2N.removed and the residue was chromatographed through a silica gel column chromatography.The product was eluted with a mixture of hexane/ethyl acetate (9.5 : 0.5), leading to the ANQs a-c (Scheme 2).

In Vitro Antiplasmodial Activity.
There are several in vitro methods commonly used for the assessment of parasite growth inhibition [38].The parasite protein lactate dehydrogenase (pLDH) method [39,40] was chosen to evaluate the in vitro antiplasmodial activity of HNQs a-h, ortho-FNQs a-h, a-h, , and ANQs a-c.P. falciparum (W2 strain) was maintained in a continuous culture, as described by Trager and Jensen (1976) [41].Ringstage parasites in sorbitol-synchronized blood cultures were added to 96-well culture plates, at 2% parasitemia and 1% hematocrit.Stock solutions at 50 mg⋅mL −1 of each compound in DMSO were diluted in a complete medium, in serial concentrations of 2 : 1 to final concentrations of 0.002% (v/v), and 20 l of each one was added to the 96-well culture plates containing the parasite culture.After 48 h incubation, the plates were frozen (−20 ∘ C for 24 h) and thawed before starting the pLDH assay [40].
The hemolyzed cultures were transferred to another 96well culture plate to which Malstat5 and NBT/PES reagents were added.After 1 h of incubation at 37 ∘ C in the dark, absorbance was read at 540 nm in a spectrophotometer (Infinite5200 PRO, Tecan).The experiment was performed three times, each one in triplicate.The results were evaluated by the Microcal Origin 8.5 software to determine the doseresponse curves plotted with sigmoidal fit.The IC 50 was determined by comparison with controls for standard drug and without drugs.

Cytotoxicity to Hep G2A16
Cells.Cytotoxicity determination was carried out in human hepatoma cell culture (Hep G2A16), which was kept at 37 ∘ C in RPMI medium supplemented with 5% fetal calf serum (complete medium) in a 5% CO 2 environment.Cells from confluent monolayers were trypsinized, washed, counted, diluted in complete medium, distributed in 96-well microtiter plates (4 × 10 5 cells/well), and then incubated for 18 h at 37 ∘ C. The compounds to be tested were diluted in DMSO (final concentration of 0.02%).After 24 h incubation at 37 ∘ C, 28 L of MTT solution (2 mg⋅mL −1 in PBS) was added to each well.After 1.5 h of incubation at 37 ∘ C, the supernatant was removed and 130 L of DMSO was added to each well [40,42].The experiment was performed three times, each one in triplicate.The culture plates were read by a spectrophotometer with a 510 nm [42].
The selective indexes (SI = CC 50 /IC 50 ) of compounds a-h, a-h, and a-h were calculated.
2.3.Computational Methods: Molecular Docking.Initially, the molecular targets were obtained from a Protein Data Bank (PDB) [43] under codes 4PD4 (P.falciparum cytochrome bc1 complex) [8] and 5FI8 (P.falciparum dihydroorotate dehydrogenase) [44].Next, the molecular target structures were represented according to the Swiss Model software [45] for building the missing loops, and the resulting structures were submitted to H++3.0 [46], a program that automates the key steps in the preparation of biomolecular structures for molecular modeling and simulations, as well as by the MGLTools software [47]  for 5FI8, using the MGLTools software [47].Subsequently, the HNQs and the FNQs structures were generated by the MarvinSketch program [48].The protonation states and the tautomers, when possible, were adjusted according to pH 7.4 [49].Next, all compound structures were refined by the Octopus software [50].Octopus is an automated workflow management tool that integrates MOPAC2016, MGLTools, PyMOL, and AutoDock Vina to perform molecular docking through a user-friendly interface [45,[51][52][53][54][55].HNQs and FNQs were refined by the parametric method 7 (PM7) [56] and the Eigenvector Following (EF) [56] routine implemented in MOPAC2016 [52] by the Octopus Run mopac routine, which automatically assigns the total molecular charge for each ligand and checks for errors in the structures.The AutoDock Vina methodology was evaluated by redocking, and its exhaustiveness parameter was set to 24 to achieve more accurate results [57].The figures were generated by academic versions of the Discovery Studio 4.5 program [58].
Some of the HNQs, ortho-and para-FNQs described in this work, had also been synthesized by our research group in prior studies, but these were only evaluated against Toxoplasma gondii [30][31][32].
A fourth series of naphthoquinones, the ANQs ( a-c), was prepared from lawsone ( ) by etherification with epichlorohydrin to produce compound (58% yield), followed by a nucleophilic opening of the epoxide ring with appropriated amines affording a-c with low yields (20% to 30%), Scheme 2. All compounds of this series represent new chemical entities.

Spectroscopic Characterization of Compounds.
All the synthesized compounds were chemically characterized by their spectra in the IR (infrared), 1 H NMR, and 13 C NMR and some of them by HRMS and two dimensional NMR spectra.Spectra for the new chemical entities are shown in the Supplementary Data in Supplementary Material available online at https://doi.org/10.1155/2017/7496934.

Biological Evaluation.
Initially, the percentage of parasite growth inhibition was determined for each compound in two different concentrations (50 and 25 g⋅mL −1 ), in triplicate, and the results are shown in Table 1.Compounds showing a parasitemia reduction greater than 50% at 25 g⋅mL −1 were classified as active.
All of the HNQs ( a-h), ortho-FNQs ( a-h), and para-FNQs ( a-h) were classified as in vitro antiplasmodial active compounds (parasitemia reduction > 50% at 25 g⋅mL −1 ), and their IC 50 values were experimentally determined by the pLDH methodology (Table 2) [40,42].Lawsone glycidyl ether ( ) and ANQs ( a-c) were inactive, with parasitemia reduction ranging from 9% to 34% at 25 g⋅mL −1 (Table 1); therefore, their IC 50 and CC 50 were not determined.It is important to highlight that the in vitro assays were performed in 48 h, because that is the duration of the asexual erythrocytic stage of P. falciparum, and it is within the time range recommended by Makler and collaborators (1998) [39] whose methodology was followed in the present in vitro antiplasmodial evaluations.
All of the naphthoquinones of the first three series, namely, HNQs ( a-h), ortho-FNQs ( a-h), and para-FNQs ( a-h), were more active than lapachol (IC 50 206.38mol⋅L −1 ) (Table 2).The HNQs series ( a-h) was the least effective, with only f and h disclosing moderate activity (IC 50 < 50 mol⋅L −1 ) (Table 2).In general, ortho-FNQs ( a-h) and para-FNQs ( a-h) showed higher activity (lower IC 50 ) than HNQs.In total, twelve FNQs showed IC 50 < 30 mol⋅L −1 : six ortho- FNQs ( a, b, c, e, f, and h) and six para- FNQs ( a, b, c, d, f, and g), highlighting c (IC 50 11.65 mol⋅L −1 ), f (IC 50 18.78 mol⋅L −1 ), and h (IC 50 19.91mol⋅L −1 ), which were the most active.Another interesting aspect to be noted is related to compound pairs with a phenyl and a chlorophenyl substituent in each of these three structural series.The presence of a chlorine atom was planned by analogy with atovaquone, and a surprising result was observed.For the HNQs g and h and the ortho-FNQs g and h, the chlorinated derivatives were about three times more active in HNQs and two times more active in ortho-FNQs, while for the para-FNQs series, g and h, an inverse effect was observed, which could, possibly, be related to the low solubility of these compounds in the conditions of the biological assays.
To investigate the selectivity of the active compounds, CC 50 against Hep G2A16 cells was determined, and the SI values (SI = CC 50 / IC 50 ) were calculated (Table 2).
Six of the 13 less cytotoxic compounds (CC 50 > 90 mol⋅L −1 ) are HNQs ( a, b, c, e, g, and h), while seven are para-FNQs ( a, c, d, e, f, g, and h).As expected, ortho-FNQs were more cytotoxic because they can be easier reduced and, therefore, more reactive oxygen species (ROS) are generated, increasing thus cell damage [61].Nine compounds showed SI > 5: four HNQs ( a, e, g, and h) and five para- FNQs ( a, c, d, f,  and g).Compounds c and f are highlighted here because they exhibit the best IC 50 values ( c 11.65 mol⋅L −1 ; f 18.78 mol⋅L −1 ), presented low cytotoxicity ( c 115.90 mol⋅L −1 ; f 97.89 mol⋅L −1 ), and showed SI > 5 ( c 9.94 and f 5.21).Despite disclosing a moderate activity, compound a (IC 50 26.57mol⋅L −1 ) showed low cytotoxicity (CC 50 869.10mol⋅L −1 ), resulting in the highest SI value (32.71), and might be highlighted as a lead compound, when its SI is considered; that is, a can be selected for in vivo antimalarial efficacy evaluation and early safety and toxicity studies.Finally, compounds h, a, c, d, f, and g showing moderate antiplasmodial activity (IC 50 < 40 mol⋅L −1 ) against the chloroquine-resistant P. falciparum W2 strain, low cytotoxicity (CC 50 > 90 mol⋅L −1 ), and SI > 5 are thus selective to resistant parasite and deserve also further investigations.
The in vitro antiplasmodial activity of some naphthoquinones described in this work was previously described [27,28] as causing completely growth inhibition of an unidentified strain of P. falciparum.Recently, Duran-Lengua et al. (2015) [34] reported enhanced effects of , , -a, , , -b, and , , -d against sensitive and moderately resistant strains of P. falciparum, 3D7, and Dd2, respectively.With respect to the IC 50 values, it should be taken into account that, in the present case, the higher IC 50 values might be explained by the fact that the in vitro assays were performed upon P. falciparum W2, a chloroquine-resistant strain, which represents a challenge to the discovery of new antiplasmodial drugs.
3.3.Computational Methods: Molecular Docking.Intermolecular interactions between HNQs and FNQs against the molecular targets 4PD4 (Pf cyt bc1) and 5FI8 (Pf DHOD) were simulated by the AutoDock Vina software [8,37].The methodology was evaluated by redocking, which consists of the docking of crystallographic structures onto the targets binding sites.This simulation can show if the docking parameters and the methodology are adequately adjusted to the system.As a result, the root mean square deviation (RMSD) of atovaquone ( ) and DSM422 (Figure 3) redocking were 0.75 Å and 0.86 Å, respectively.These results were below the threshold value of 2.0 Å [52] and binding energies of −11.3 Kcal⋅mol −1 to atovaquone with 4PD4 and −12.1 Kcal⋅mol −1 to DSM422 with 5FI8 (Figure 3).These data suggest that the AutoDock Vina methodology is appropriate for simulating the intermolecular interaction and predicting the binding energy of HNQs and FNQs to targets (4PD4 and 5FI8).
Binding energies of docking simulations ranged from −8.2 Kcal⋅mol −1 to −10.6 Kcal⋅mol −1 for 4PD4 and from −8.2 Kcal⋅mol −1 to −11.5 Kcal⋅mol −1 for 5FI8 (Table 2).These values are relevant in the interpretation of docking results for the HNQs and FNQs, as active compounds are expected to disclose binding energies close to the models (atovaquone and DMS422) and should promote good activities.
Molecular docking simulations disclosed favorable binding energies (BE ≤ −10 Kcal⋅mol −1 ) for interactions of twelve compounds with the targets 4PD4 ( g, h, g, h, g, h) and  5FI8 ( f, g, f, f, g, h) which are coherent with SI > 5 for g, h, f, and g, highlighting FNQs.
For HNQ d, one of the less active compounds (SI 0.24), a weak correlation with the binding energies (4PD4 −8.5 and 5FI8 −9.0 Kcal⋅mol −1 ) for both of the targets (Table 2), was observed.On the other hand, for the HNQ h (SI = 8.49), the binding energy is −10.6 Kcal⋅mol −1 for 4PD4, while the value of BE for para-FNQ f (SI = 5.21) is −11.5 Kcal⋅mol −1 for 5FI8 (Table 2); these are the most active compounds of these series and their complexed structures into the hydrophobic cavities of the targets, as shown in Figure 4.
Figure 5(a) shows that HNQ h represents van der Waals intermolecular interactions with 4PD4 through amino acids Met139, Gly143, Val146, Trp142, Ile269, Pro271, Leu275, and Phe278.Similarly, for compound f, van der Waals interactions are depicted with Leu34, Cys37, Leu38, Leu59, Cys46, Phe89, Ile99, Leu363, and Leu364 amino acids of 5FI8 (Figure 5(b)).In addition, the molecular recognition for compound f is improved by two weak noncovalent intermolecular interactions: His47 and the furan oxygen with a distance of 2.69 Å and a -sulfur interaction between Cys95 and the benzenoid ring in the naphthoquinone moiety with a distance of 2.47 Å.These intermolecular interactions might certainly contribute for the biological activity of compounds h and f.
Previous studies have shown the inhibition of Pf cyt bc1 and Pf DHOD by naphthoquinones [8,10,11], and the molecular docking results described here are in accordance with this mode of action, although further enzymatic assays are necessary for confirmation.

Conclusions
The present study reported on the synthesis of 28 compounds belonging to four structural series-HNQs, ortho-FNQs, para-FNQs, and ANQs-eight of which represented new chemical entities.All of the HNQs ( a-h) and FNQs ( a-h and a-h) were classified as active compounds with in vitro parasitemia reduction ranging from 45% to 90% at 25 g⋅mL −1 and proved to be more potent than lapachol (IC 50 206.38mol⋅L −1 ) against the chloroquine-resistant P. falciparum W2 strain.The insertion of a furan ring in the naphthoquinone moiety increased the biological activity, and the FNQs were, in general, more active than the HNQs.As expected, ortho-FNQs ( a-h) were more cytotoxic to Hep G2A16 cells and less selective to the parasite (SI < 5).The ortho-FNQs might be highlighted for their potential antitumoral rather than antiplasmodial effect.On the other hand, ANQs, although containing an amino group in their structures, disclosed low biological activity and were classified as inactive compounds (parasitemia reduction < 50% at 25 g⋅mL −1 ).Considering the experimental values of IC50 < 40 g⋅mL −1 , CC50 > 90 g⋅mL −1 , and SI > 5, the para-FNQ a is highlighted for its high value of SI (32.71).However, the following naphthoquinones h, c, d, f, and g deserve also further investigations.Molecular docking simulations disclosed most favorable binding energies (< −10 Kcal⋅mol −1 ) for h and f, disclosing good correlations with the present biological data for naphthoquinones that support aryl moieties.This study points out to arylnaphthoquinones as new potential inhibitors of the Pf cyt bc1 and Pf DHOD enzymes since, for both of these targets, favorable binding energies were calculated.However enzymatic assays are necessary for confirming this hypothesis.

Scheme 2 :
Scheme 2: Synthesis of naphthoquinones and a-c and their yields.

3 C 1 RMSDFigure 3 :
Figure 3: Values of the root mean square deviation (RMSD) of redocking and binding energy (BE) for the interaction of 4PD4 with atovaquone (a) and of 5FI8 with DSM422 (b).

Figure 4 :Figure 5 :
Figure 4: (a) Compound h complexed into the hydrophobic surface of 4PD4.(b) Compound f complexed into the hydrophobic surface of 5FI8.