Synthesis and Anti-Chikungunya Virus (CHIKV) Activity of Novel 1,4-Naphthoquinone Sulfonamide and Sulfonate Ester Derivatives

Chikungunya virus (CHIKV) is a re-emerging disease caused by an alphavirus of the Togaviridae family. Since its first description in 1952, the disease has spread worldwide, affecting populations in both tropical and temperate countries. To date, there is no licensed vaccine or specific pharmacological treatment. Therefore, there is an increasing urgency in developing new antiviral drugs capable of specifically inhibiting viral replication. In the present work, we report the synthesis and antiviral activity evaluation of nineteen naphthoquinone derivatives, containing a sulfonamide or sulfonate group. Cell viability assays indicated a low toxic potential for all tested compounds and inhibitory assays against CHIKV identified five compounds with potent activity. The compounds were also evaluated for their virucidal potential, and the results demonstrated that compound 11a exhibited a virucidal effect higher than 70% in the treatment with 20 μM. Furthermore, in silico studies were performed to predict the antiviral drug targets.


Introduction
Quinones constitute a class of natural products with wide distribution in nature and are structurally characterized by two conjugated carbonyl groups in sixmembered unsaturated rings. 1 Naphthoquinone presents an aromatic ring fused to the quinonoid nucleus, and this chemical characteristic is identified as responsible for its redox potential. 2 Naphthoquinones are considered privileged structures in Medicinal Chemistry and have been extensively studied as synthetic platforms due to their broad range of biological properties, such as anticancer, antibacterial, anti-inflammatory, antiviral, and others. [3][4][5][6][7][8][9] Chikungunya virus (CHIKV) is a mosquito-transmitted alphavirus of the Togaviridae family that causes a febrile syndrome (Chikungunya fever) usually associated with intense and debilitating arthralgia. [10][11][12] It was first described in 1952 during an outbreak in southern Tanzania. 13 Since then, it has spread, and infection cases have been reported in all continents, except Antarctica and approximately 3-5 million cases of CHIKV are reported each year worldwide. 14 The disease is transmitted during blood repast of peridomestic mosquitoes, such as Aedes aegypti and Aedes albopictus. 15,16 Although research has brought compounds with activity against CHIKV, most of them have not yet been validated in vivo and in clinical trials. Thus, until now, neither specific antivirals drugs nor a licensed vaccine are available. Chikungunya fever therapy is based on supportive measures and the treatment of symptoms through non-steroidal anti-inflammatory drugs (NSAIDs) and fluid therapy. 17 This shows that the effort to find a specific treatment is extremely important. The sulfonamide group (-SO 2 NH 2 ) is found in the structure of several drugs with high clinical relevance, such as sulfamethoxazole (1, antibacterial), acetazolamide (2, carbonic anhydrase inhibitors), celecoxib (3, cyclooxygenase-2 inhibitor), and glibenclamide (4, hypoglycemic agent) ( Figure 1). 7, [18][19][20] Other biological activities have also been reported by compounds containing this functionality, such as anticancer, antimicrobial, anti-inflammatory, antioxidant, and others. [21][22][23][24] Regarding antiviral activity, sulfonamide moiety is present in the compound structure used clinically to treat viral infections, such as amprenavir (5) and darunavir (6). 25,26 Herein, we describe the synthesis of two series of naphthoquinones derivatives containing a sulfonamide group (8a-8e, 9a-9e, and 10a-10d) and sulfonate esters group (11a-11e). Additionally, all synthesized compounds were also screened in vitro for their cytotoxicity and anti-CHIKV activity.

Materials and methods
The solvents and reagents were obtained from Merck (São Paulo, Brazil) and used without further purification. For qualitative monitoring of the reaction progress, analytical thin-layer chromatography (TLC) was produced with silica gel plates (Merck, São Paulo, Brazil, TLC silica gel 60 F254), and the plots were visualized using UV light (Sigma-Aldrich, São Paulo, Brazil). For purification of the final compounds, column chromatography was performed with silica gel 60 (Merck 70-230 mesh, São Paulo, Brazil). Melting points, when necessary, were obtained on a Thermo Scientific 9100 apparatus (Waltham, USA) and were uncorrected. All final products were submitted to complete spectroscopic characterization. Infrared spectra were measured with KBr pellets on a PerkinElmer model 1420 Fourier transform infrared (FTIR) spectrophotometer (São Paulo, Brazil), and the spectra were calibrated using the 1601.8 cm -1 polystyrene absorbances. Nuclear magnetic resonance (NMR) spectra were obtained with a Varian Unity Plus VXR 500 MHz (Palo Alto, USA) instrument in dimethyl sulfoxide (DMSO-d 6 ) solutions. The chemical shift data were reported in units of d (ppm) downfield from tetramethylsilane (TMS) or the solvent, either of which was used as an internal standard, coupling constants (J) are reported in hertz and refer to apparent peak multiplicities. High-resolution mass spectra (HRMS) were recorded on a MICROMASS Q-TOF mass spectrometer (Waters, Milford, USA).
General procedure for 8a-8e, 9a-9e, and 10a-10d preparation The corresponding amino-naphthoquinone (7a-7c) (1 mmol) was dissolved with pyridine (4 mL) under magnetic stirring in a round-bottom flask. Next, 4-dimethylaminopyridine (DMAP) (0.1 mmol) and the appropriate sulfonyl chloride (1.5 mmol) were added to the reaction flask, and the mixture was stirred for 24 h at room temperature. After the start material total consumption, the aqueous solution of HCl 1 M was added to neutralize the medium, and the resulting mixture was extracted with ethyl acetate. The organic phase was washed successively with saturated NaHCO 3 solution, water, and brine, dried over sodium sulfate anhydrous, and evaporated under reduced pressure. The resulting crude product was purified in silicagel column chromatography using a gradient mixture of hexane/ethyl acetate as eluent to obtain the pure derivatives.  General procedure for 11a-11e preparation A stirred solution of 7d (100 mg; 0.540 mmol) in 5 mL of acetonitrile was cooled in an ice bath, and 2.16 mmol of the appropriate sulfonyl chloride was added. Subsequently, trimethylamine (1.6 mmol) was added to the reaction mixture, and the ice bath was removed. The reaction occurred at room temperature for an additional 1 h. After the starting material total consumption, the reaction mixture was poured into 25 mL of water and extracted using ethyl acetate (3 × 50 mL). The organic phase was washed with brine (3 × 50 mL), dried with anhydrous sodium sulfate, filtered, and evaporated under reduced pressure. The mixture was purified by column chromatography on silicagel using a gradient mixture of hexane/ethyl acetate as eluent to obtain the pure derivatives.

Effect of the compounds on the cellular viability
The naphthoquinone derivatives cytotoxicity in Vero cells was evaluated using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide tetrazolium (MTT) reduction colorimetric assay in 96-well plate and incubated for 24 h at 37 °C. 28 After, naphthoquinones derivatives were added in increasing concentrations, ranging from 50 to 1000 µM, diluted in DMEM. After this period, the cells were incubated with the MTT method and evaluated in a microplate reader to determine cell viability. DMSOtreated cells (1.0%) in DMEM were used as control of the evaluation. Determination of cell viability was calculated by percentage comparing treated cells with untreated cells and cells in DMSO.

TCID 50 (median tissue culture infectious dose) assay for infectious virus
The effect of naphthoquinone derivatives on CHIKV replication was determined with a fixed dose of 20 µM, added to 2 × 10 5 infected Vero cells. This first step was to determine which key compounds were able to inhibit viral replication at this concentration and thus initiate studies with compounds that have a high inhibitory potential. Then, 2 × 10 5 Vero cells were plated in 24-well plate in DMEM medium with 5% fetal bovine serum (FBS) and incubated at 37 °C at 5% CO 2 . After the cells reach above 90% of confluence, the antiviral evaluation is initiated. The virus inoculum was diluted to a multiplicity of infection (MOI) of 0.1 in serum-free DMEM and added for 1 h for adsorption of viral particles at 37 °C. Afterwards, the viral inoculum was derived, the cells were washed with phosphate buffer solution (PBS) to remove non-adsorbed viral residue and fresh medium was added at the concentration of compounds. Infected cells were also treated with culture medium containing 1% DMSO to determine residual antiviral effect and ribavirin was used as a positive control for antiviral activity. After 24 h of incubation at 37 °C, the culture medium was collected and the yield of infectious virus in the cell supernatant was determined by titration using the TCID 50 assay on Vero cells and it was also evaluated by assay to produce viral plaques.

Virucidal profile
For the virucidal potential analysis, a mixture of 200 PFU (plaque-forming unit) containing CHIKV plus an equivalent compounds volume was prepared, varying concentrations of 5, 10, or 20 μM mL -1 and incubated in microtiter plates for 2 h at 37 °C. The virus was added to Vero cells in 24-well plates for 2 h, washed, and incubated under plate assay conditions. A virucidal effect was defined as the compounds ability to inactivate the particles, blocking infection and any subsequent cytopathic effect. All treatment conditions were compared to a virus-only control.

Time-of-addition assay
To analyze the possible viral multiplication cycle stage affected by the compounds, we conducted a study with different addition times. Vero cell monolayers were grown in 24-well plates with 90% confluence. A triplicate was treated with 10 µM one hour before infection, called time -1. After one hour, the plate was fully infected with CHIKV at an MOI of 0.1, and immediately after infection, another triplicate was treated with the compounds, called time zero or concomitant. The remaining triplicates were treated according to times 1, 2, 4, 6, 8 h. After incubation for a further 12 h, the culture supernatants were collected, and the cells in the culture medium were supplemented with 3% carboxymethylcellulose. After 72 h of incubation at 37 °C, the cells were fixed with 20% formaldehyde for 2 h and stained with crystal violet for 5 min. The plaques formed after each treatment were counted, and inhibition of viral replication was determined.

Molecular docking studies
We obtained the three-dimensional structures of the potential targets of 10c and 11a from the Protein Data Bank (PDB) for molecular docking purposes. The structures of CHIKV proteins were obtained under the following codes: the envelope proteins E1-E2-E3 complex (PDB 3N42), 29 the capsid protein (PDB 5H23) 30 protease domain, the protease (PDB 3TRK), 31 and the nsP2 helicase (PDB 6JIM) 32 domains, and the nsP3 (PDB 3GPO) 33 macro domain. The three-dimensional structure of 10c, 11a, D9, a known nsP2 protease inhibitor, and heparitin, a known envelope proteinligand, was constructed and optimized using the Spartan'10 software (Wavefunction Inc. Irvine, CA). 34 Initially, the compounds were submitted to a conformational analysis using the MMFF (Merck Molecular Force Field) force field. The lowest-energy conformer was subjected to a geometry optimization using the semi-empirical RM1 method, followed by an energy calculation using the density functional theory method with B3LYP/6-311G* basis set. The molecular docking studies were performed using Autodock Tools 1.5.7 and Autodock 4.2.6. 35 Dockings of 10c and 11a into the potential targets were conducted considering the results of the experimental assays. Solvent and other artifact molecules were removed from the protein structures. Autodock Tools was used to add polar hydrogens and Gasteiger charges. Proteins were considered rigid, while total flexibility of ligands was allowed. The grid box for each protein was defined according to their binding sites described in the literature: the heparan-sulfate-binding and the predicted binding pocket 1 sites of the envelope protein, the active (comprising the catalytic triad) and the hydrophobic (where E2 protein interacts with the capsid protein) sites, the active site of nsP2 protease, the adenosine triphosphate (ATP) binding site of nsP2 helicase, and the ADP-ribose binding site of nsP3. [36][37][38] The Lamarckian genetic algorithm was employed, and a total of 50 binding poses were calculated for each binding site. The pose with the lowest binding energy was selected for evaluation. The potential target of 10c and 11a was elected based on the binding energy and visual inspection of the binding mode and interaction analysis compared with known ligands. These analyses were realized using the Discovery Studio Visualizer 2019 (Dassault Systèmes BIOVIA, San Diego, 2019) 39 and Pymol version 2.5 (The PyMOL Molecular Graphics System, version 2.5, Schrödinger, LLC). 40

Statistical analysis
The data were analyzed by one-way analysis of variance (ANOVA) followed by Tukey's post-test using the GraphPad Instat version 3 program. 41 A p-value of < 0.05 was considered statistically significant.

Chemistry
The amino-naphthoquinones (7a-7d) used as starting material for the preparation of sulfonamide-based naphthoquinones derivatives were prepared according to the reported methodology. 42,43 Subsequently, the desired sulfonamide derivatives (8a-8e, 9a-9e, and 10a-10d) were obtained in one step by choosing the appropriate amino-naphthoquinones (7a-7c) with various aryl sulfonyl chlorides catalyzed by DMAP in pyridine. The overall yield calculated after column chromatography purification ranged from 45-99%. Spectroscopic techniques confirmed the structures of all final compounds (8)(9)(10). In general, the FTIR analysis of sulfonamide derivatives exhibited absorption bands at 1600-1680 cm -1 , referring to the stretch vibrations of C=O bonds. It was possible to observe absorption bands of medium intensity in the region between 3100-3300 cm -1 , corresponding to the stretch vibrations of N-H bonds. The analysis also showed absorptions at 1150-1330 cm -1 , referring to asymmetric and symmetrical stretching of the SO 2 group bounds. In the 1  The sulfonated naphthoquinones derivatives (11a-11e) were also prepared in a single step by treating a solution of 7d in acetonitrile with triethylamine (TEA) and the appropriate sulfonyl chlorides. The products were obtained with excellent yields ranging from 95-99% (Scheme 1). Spectroscopic techniques also confirmed the chemical structures of these compounds. For example, in the 1 H NMR of the compound 11b, it was possible to show that the sulfonylation reaction proceeded in the hydroxyl group by the absence of a characteristic signal referring to hydroxyl hydrogens. Another evidence is a singlet at 7.05 ppm (s, 2H) corresponding to amino hydrogens. Besides, it is possible to observe the presence of two sets of signals referring to the aromatic hydrogens of the sulfonic ester group: two doublet signals at 7.74 ppm (d, 2H, J 8.4 Hz, 2H) and 7.46 ppm (d, 2H, J 8.5 Hz, 2H); and a singlet at 2.42 ppm (s, 3H), corresponding to a methyl group.

Cytotoxicity of naphthoquinone derivatives and antiviral effect
The MTT assay was used to determine the cytotoxicity of each naphthoquinone derivatives to Vero cells. The 50% cytotoxic concentration (CC 50 ) value of each compound was calculated (Table 1). According to our results, no high cytotoxicity was observed for all tested compounds and no cytotoxicity in cells treated with 0.1% DMSO (final solvent concentration used to dissolve naphthoquinone in the cell culture medium). After determining cytotoxicity, the compounds were used in a fixed dose (20 µM) to determine whether they would be able to inhibit CHIKV replication in Vero cells. In the initial evaluation, the  compounds 9b, 9e, 10c, 11a, and 11b showed an inhibition effect above 98% of CHIKV replication at this fixed concentration ( Table 1).

Determination of the effect of naphthoquinone derivatives on CHIKV replication
The antiviral effect evaluation was performed in Vero cells for all substances that affected greater than 90%. Vero cells were infected with CHIKV using a multiplicity of infection (MOI) of 0.1 and treated for 24 h with increasing substances concentrations. The TCID 50 method was used to titrate the virus yield, and the concentration capable of reducing the virus yield by 50% (EC 50 ) was obtained from a dose-response curve, and the values are shown in Table 2 and Figure 2. The relationship between cytotoxicity and the compound antiviral effect was assessed using the selectivity index (SI ratio, CC 50 /EC 50 ) and determined for each molecule, obtaining unit values. As observed in Table 2, five naphthoquinone derivatives, 9b, 9e, 10c, 11a, and 11b, showed an expressive SI for CHIKV, with 289, 457, 418, 400, and 382, respectively. In the same way, in Figure 2, we can observe a dose-dependent response on the inhibitory effects obtained by the substances assessed. Ribavirin was used as a control (Table 2).  Virucidal effect Given all the tested compounds, at least five derivatives exhibited a significant inhibitory profile for CHIKV replication with low cytotoxicity. A crucial step in discovering the mechanism of action is evaluating the compound's effect on the viral particle. For this reason, we determined the virucidal potential of the compounds. Our results showed that of the five compounds tested, only 11a presented some virucidal effect in the tested concentrations. As shown in Figure 3, the compound 11a was able to inhibit above 70% in the treatment with 20 µM. Interestingly, ribavirin at the same concentrations did not exhibit virucidal potential.

Time-of-addition studies
A time-of-addition experiment was conducted to identify at which point of the CHIKV replication cycle the substances exert their specific antiviral effect. All compounds already have an antiviral effect at time 0 (concomitant) to infection, inhibiting above 90% of viral replication at the concentration used (10 µM).
The substance 11a showed a differentiated profile, since, at point -1, it presented an inhibitory potential of approximately 70%, increasing its effect for up to 2 h post-infection, reaching inhibition above 90% at time 0, maintaining inhibition above 80% until the second hour, declining sharply after 4 h post-treatment, falling below 40% inhibition. The other substances showed a greater inhibitory potential in the treatment times of 0, 1, and 2 h, decreasing progressively after the other times, reaching an inhibitory potential below 40%. Conversely, compound 10c showed low inhibitory potential at time -1, although 0 inhibiting replication above 90%, maintaining potent inhibition until 8 h post-infection (Figure 4). Time dependence on the inhibitory effect of compound 11a is consistent with the hypothesis that its anti-CHIKV appears to be in events before entry. The compound 10c showed an inhibitory effect with more consistent post-infection effects.

In silico prediction of the antiviral targets
In silico target prediction of 10c According to the antiviral assays, we observed that 10c had no significant effect on the CHIKV particle infectivity but inhibited virus replication when added during attachment and later, which, in turn, indicates that this compound acts on post-entry steps. To evaluate the potential targets of this compound, we docked 10c with the available crystallographic structures of CHIKV proteins Concentration that reduced 50% cytotoxic concentration when compared to untreated controls; b concentration that reduced 50% of CHIKV replication when compared to infected controls; c selectivity index was defined as the ratio between CC 50 and EC 50 and represents the safety for in vitro assays. The mean values ± standard deviations are representative of three independent experiments. CHIKV: Chikungunya virus; ND: not detected.  Among the proteins evaluated, 10c exhibited the highest theoretical binding affinity with nsP3 (-9.02 kcal mol -1 ). 10c showed a slightly bent conformation bound to nsP3, which is very similar to the co-crystallized ligand ADP-ribose ( Figure 5). The naphthoquinone ring of 10c occupied the same region of the phosphates and distal ribose groups and was hydrogen-bonded to V33, C34, and Y114. Also, we observed that this ring was involved in a T-shaped π-π stacking interaction with Y114 and van der Waals contacts with N24 and V113. The sulfonamide and the central phenyl groups of 10c were positioned at the binding region of the co-crystallized ligand adenosine ribose. They established a double hydrogen-bond with L108 and a van der Waals contact with G112. The naphthalene moiety bound in the same region as observed for the adenine group of ADP-ribose and established van der Waals interactions with I11, C143, and R144.
Furthermore, the extended effect observed in the time-of-addition assay suggested that this compound may have a multi-target mechanism of action. 10c showed the second strongest binding affinity with nsP2 protease (-7.08 kcal mol -1 ). Also, we docked the known inhibitor D9 18 with this protease for comparison purposes. 10c presented an extended binding mode as observed for the inhibitor (Figure 5). The naphthoquinone ring was positioned in the S2 subsite. This molecule was anchored by several interactions, such as a hydrogen bond with W1084, a π-sulfur interaction with the catalytic C1013 residue, and van der Waals contacts with A1046, N1082, and L1205. The central phenyl group of 10c bound at the S3 subsite of the protease and interacted with M1242. Like the distal phenyl and cyclopropyl rings of D9, the sulfonamide and naphthalene groups of 10c explored the S4 subsite. The sulfonamide group was hydrogen-bonded with D1246. This residue was also anchoring the naphthalene ring by an anion-π interaction in addition to the van der Waals contact with Q1241.
Moreover, 10c was docked to the capsid protein and nsP2 helicase. However, it had the lowest theoretical affinity (-6.25 and -5.50 kcal mol -1 for the active and capsid protein hydrophobic sites, respectively, and -5.06 kcal mol -1 nsP2 helicase). This prototype did not exhibit a binding manner similar to known ligands of these proteins (data not shown).

In silico target prediction of 11a
Unlike 10c, 11a showed a direct effect on virus particles and acted in the early stages of the CHIKV lifecycle. So, we investigated in silico whether this compound could bind Data are presented as percentage of virus titer, when compared to control cells and are expressed as the mean of three experiments ± standard error. Statistical analysis was performed using the Tukey's test to compare naphthoquinone derivatives with ribavirin in each time: **p < 0.01. Figure 5. Docking of 10c with its potential anti-CHIKV targets. Binding mode and molecular interactions of (a) 10c (green) and the co-crystallized inhibitor adenosine diphosphate ribose (ADP-ribose, orange) with the macro domain of nsP3, and (b) 10c (green) and the known inhibitor D9 (orange) with the nsP2 protease. Hydrogen bonds are described as black dashed lines, anion-π or π-π stacking interactions are shown as red dashed lines, and π-sulfur interaction is represented as orange dashed line.
to the envelope proteins and the nsP2 protease and helicase domains and nsP3 macro domain.
11a showed the highest theoretical affinity with the macro domain of nsP3 (-5.10 kcal mol -1 ). The naphthoquinone moiety was aligned with the adenosine ribose group of ADP-ribose and was involved in a hydrogen bond interaction with L108 and van der Waals contacts with A22, V33, T111, and W148 (Figure 6a). On the other hand, the sulfone and phenyl groups were positioned at the same binding region as observed for the known ligand phosphates and distal ribose. Consequently, 11a was allowed to establish similar interactions with the protein. The sulfone group interacted with V33 and V113 by hydrogen bonding. Besides, the phenyl ring was π-stacked with Y114 and interacted with V113 via van der Waals contact.
Following nsP3, 11a showed the second strongest theoretical affinity with the CHIKV envelope proteins. Two binding sites were explored, and 11a presented a similar affinity towards pocket 1 (-4.57 kcal mol -1 ) compared to the heparan-sulfate-binding site (-4.46 kcal mol -1 ). At pocket 1, the naphthoquinone nucleus was π-stacked with Y51 and hydrogen-bonded to Y233 and S238 of E1 (Figure 6b). Also, the sulfone group established hydrogen bond interactions with the side chain of R36 of E2 and with the principal and side chains of T53 of E1, while the phenyl ring was involved in van der Waals contacts with I55 from E1. For the heparan-sulfate binding site, the naphthoquinone ring was sandwiched between R104 and F141 by cation-π and π-π stacking interactions, respectively, while the amino group was hydrogen-bonded to H142 (Figure 6b). The sulfone group established hydrogen bond interactions with V135 and F141, whereas the phenyl ring was involved in a van der Waals interaction with I136. Interestingly, we observed that 11a explored the same binding region as one of the sulfated saccharide nuclei of heparitin (a heparan sulfate disaccharide used as a control herein) and shared interactions R104, F141, and H142. Additionally, 11a showed the lowest theoretical binding affinity with both protease and helicase domains of CHIKV nsP2 (-4.35 and -3.61 kcal mol -1 , respectively) (data not shown).
The cell viability assay revealed a low toxic potential for all compounds. Remarkably, compounds 9b, 9e, 10c, 11a, and 11b exhibited potent inhibitory activity with a highly selective index. We highlight compounds 9e and 10c that presented EC 50 equal to 1.18 and 0.77 µM, generating SI of 457 and 418, respectively. In contrast, ribavirin, which is an escort drug for clinical use and used as a control, shows a lower inhibitory potential with 2.42 µM of EC 50 and SI of 122.
Compound 11a showed a virucidal effect above 40% at 5 µM, reaching above 70% at 20 µM, demonstrating that possibly most of its antiviral effects are at the earliest moments of the viral infection process. These results are in accordance with the time-of-addition assays showed that all compounds already had an antiviral effect at time 0 to infection. The addition time tests showed that although all compounds already had an antiviral effect at time 0 of infection, compound 11a already has an early inhibitory potential (-1) that is, still in pre-treatment, maintaining its inhibitory potential until 2 h post-infection, with its inhibitory potential declining rapidly, reaching an inhibitory effect of less than 40% at 4 h post-infection ( Figure 4).
We highlight compounds 9e (CC 50 = 540) and 10c (CC 50 = 322) and they presented EC 50 equal to 1.18 and 0.77 µM, generating SI of 457 and 418, respectively. These results prove to be promising, requiring further studies that Figure 6. Docking of 11a with its potential anti-CHIKV targets. Binding mode and molecular interactions of (a) 11a (cyan) and the co-crystallized inhibitor ADP-ribose (orange) with the macro domain of nsP3, and (b) 11a (cyan) and the heparan sulfate disaccharide, heparitin (orange), within the heparan-sulfate binding site (HSBS) and pocket 1 of the CHIKV envelope protein complex. E1, E2, and E3 are colored in violet, green and brown, respectively. Hydrogen bonds are exhibited as black dashed lines, cation-π or π-π stacking interactions are shown as red dashed lines, and ionic interactions are represented as orange dashed lines.
can describe in greater depth the mechanism of action of these compounds. Taking into account that there are no specific medicines for the treatment of infections caused by arboviruses, it is important to search for substances that can be used in the control, treatment and prevention of these infections in order to reduce morbidity and mortality rates caused by CHIKV.
In contrast, compound 11a presented a differentiated profile consistent with action on events before entry. Only the compound 10c showed an inhibitory effect consistent with post-infection effects. In silico studies, some molecular targets may explain these differences in these compound mechanisms of action. Therefore, our results indicate that these compounds can serve as promising scaffolds for developing specific inhibitors for the proliferation of CHIKV.
According to the experimental assay, we conducted computational studies to get more insights into the antiviral targets of the most potent compounds 10c and 11a. Both compounds showed the strongest binding affinity with the macro domain of nsP3. Many contacts observed for both compounds (e.g., V33, T111, V113, and Y114) were proved to be essential for ligand binding into the macro domain of CHIKV nsP3 by structural and computational methods, which reinforces the potential of these naphthoquinone derivatives to inhibit this enzyme. Interestingly, the druggability of CHIKV nsP3 has been recently demonstrated. 44,45 The precise role of this protein in CHIKV replication remains to be fully understood, but it is essential for several steps in virus replication like genome replication, viral assembly, and host cell shutoff. 46,47 This protein is also involved in the virulence and pathogenesis of the alphaviruses. 48,49 Although these derivatives likely share a similar target at early steps of virus replication, the mechanistic studies indicated that these compounds might act at different targets as well. The activity of 10c was prolonged and observed only at post-entry steps. Our in silico strategy also suggested that it could also inhibit the CHIKV nsP2 protease. This protein is responsible for the polyprotein processing of CHIKV, which is crucial for virus replication. 50,51 Despite the lack of experimental data regarding the structural basis for CHIKV nsP2 protease inhibition, mutation and computational studies have provided valuable information to date. For instance, C1013, N1082, H1083, and W1084 are required for the maintenance of the protease activity of alphavirus nsP2. 52,53 In this context, the hydrogen bond with W1084 seems to be a critical interaction for its inhibition. 54,55 As well, interactions with residues C1013, A1046, N1082, W1084, and L1205 are commonly observed in the design of novel inhibitors. 56 It is important to note that 10c interacted with all these residues like the known inhibitor D9.
Additionally, inhibitors of CHIKV nsP2 protease have shown their antiviral activity at the early stages of viral replication, which agrees with the experimental data obtained in this work. 57,58 On the other hand, 11a was also shown to act directly on virus particles. Interestingly, this compound showed the second-highest theoretical binding affinity with two sites of the CHIKV envelope protein. Pocket 1 was predicted by computational tools and is found between the domain II of E1 and β-ribbon of E2. Targeting this site could stabilize the envelope complex, prevent E1-E2 dissociation, and, consequently, block fusion process. 59 Indeed, 11a presented similar interactions as potential ligands of this site. 35,54 Additionally, 11a bound to the heparan-sulfate binding site with a similar affinity as heparitin. It could establish important interactions like the one with the key residue R104, which, in turn, could impair the virus-cell interaction and inhibit the virus entry. In either case, the binding towards envelope proteins may contribute to inhibition of the viral particle production observed when the compound is added during or a little after virus attachment. 34 Our results showed the promising potential of 10c and 11a, and further experimental assays with the putative targets should be performed to confirm our predictions.

Conclusions
In this work, we report the synthesis of nineteen compounds containing a sulfonamide or sulfonate group attached to a naphthoquinone framework. The compounds were assessed in vitro for their anti-viral activity towards CHIKV. The cell viability assay revealed low toxic potential for all compounds. Remarkably, compounds 9b, 9e, 10c, 11a and 11b exhibited potent inhibitory activity with high selective index. Among them, the compound 11a also showed strong virucide effect. Time-of-addition assays showed that all compounds already had an antiviral effect at time 0 to infection, whereas the compound 11a presented a differentiated profile consistent with post entry inhibitory action with a virucidal effect above 40% at 5 µM, reaching above 70% at 20 µM, demonstrating that possibly most of its antiviral effects are at the earliest moments of the viral infection process. Collectively, our findings also suggested that 10c and 11a share the macro domain of nsP3 as a common antiviral target. In addition, 10c likely inhibits the nsP2 protease, while 11a seems to bind to the envelope protein to inhibit virus entry and fusion processes as well.

Supplementary Information
Supplementary data are available free of charge at http://jbcs.sbq.org.br as PDF file.