Is the ADP ribose site of the Chikungunya virus NSP3 Macro domain a target for antiviral approaches?

Chikungunya virus (CHIKV) is a mosquito-transmitted virus of special concern as it causes Chikungunya fever, characterized by an acute febrile illness, rash, and arthralgia that can progress to chronic and debilitating arthritic symptoms. The effects of climate change on the geographic distribution of the mosquito vector has the potential to expose more of the globe to this virus. No antiviral agents or vaccines are currently available against CHIKV infection and the development of novel therapies that may lead to a future treatment is therefore necessary. In this context, the ADP-ribose binding site of the CHIKV nsP3 macro domain has been reported as a potential target for the development of antivirals. Mutations in the ADP-ribose binding site demonstrated decreased viral replication in cell culture and reduced virulence. In this study, 48,750 small molecules were screened in silico for their ability to bind to the ADP-ribose binding site of the CHIKV nsP3 macro domain. From this in silico analysis, 12 molecules were selected for in vitro analysis using a CHIKV subgenomic replicon in Huh-7 cells. Cell viability and CHIKV replication were evaluated and molecules C5 and C13 demonstrated 53 and 66% inhibition of CHIKV replication, respectively. By using a CHIKV-Dual luciferase replicon contain two reporter genes, we also demonstrated that the treatment with either compounds are probably interfering in the early replication rather than after RNA replication has occurred.


Introduction
Chikungunya virus (CHIKV) is an arbovirus that belongs to Alphavirus genus and Togaviridae family (Lumsden, 1955).CHIKV infection can cause fever, skin rash, and arthralgia, and around 15 % of CHIKV infections can be asymptomatic (Miner et al., 2015).The infection can progress to severe symptoms and persist for months or years, leading to an economical burden for many countries (WHO, 2017).
According to the Pan American Health Organization, 349,936 suspected cases were reported in 2016 in the Americas, with 146,914 cases confirmed by laboratory analysis.No commercial antiviral agent or vaccine is available against CHIKV infection.Therefore, the development of an efficient antiviral agent is need to the treatment of infected patients.
Of the CHIKV proteins, nsP3 is the most 'enigmatic' since its functions in the virus life cycle remain uncertain (Götte et al., 2018).nsP3 comprises 3 domains: an N-terminal region macro domain, followed by the alphavirus unique domain (AUD), and the C-terminal hypervariable region (Malet et al., 2009;Shin et al., 2012).The nsP3 macro domain is conserved among alphaviruses and other pathogenic positive single strand RNA viruses (Eckei et al., 2017).Recently, the ADP-ribosylhydrolase activity of CHIKV nsP3 macro domain was shown to be critical for virus replication and virulence (McPherson et al., 2017).
Previously, the macro domain of CHIKV nsP3 has been described as a potential target in the development of anti-alphavirus compounds (Nguyen et al., 2014;Subudhi et al., 2018;Vijayasri and Hopper, 2017).Despite the interest in the macro domain of nsP3, there is a lack of in vitro or in vivo studies on drugs that bind to this region of nsP3.Most are computational studies, with no further in vitro or in vivo validation (Nguyen et al., 2014;Vijayasri and Hopper, 2017).Computational methods can be useful to refine the search for new antiviral drugs, reducing efforts and costs to approve antivirals, however, the validation of these analysis is essential.
In this study, we combined a virtual screening cascade to identify small molecules that putatively bind to ADP-ribose site of the macro domain of nsP3 and further investigated the antiviral activity of the selected molecules in vitro.

Europe PMC Funders Author Manuscripts
Europe PMC Funders Author Manuscripts 2 Material and Methods

In silico screening
An in-house library containing small synthetic molecules from the Medicinal Chemistry/ Chemical Biology (MCCB) group (School of Chemistry, Faculty of Engineering & Physical Sciences -University of Leeds) was docked to the ADP-ribose binding site of the CHIKV nsP3 macro domain (Protein Data Bank -3GPO) (Malet et al., 2009) using the docking algorithm Glide (vHTS mode).Molecules were ranked according to predicted binding affinity and the top compounds were re-docked using Glide XP mode (Maestro, Schrödinger, LLC, New York, NY, 2017).Then the compounds were visually inspected.Based on these parameters, 12 compounds were selected for in vitro activity assays.

Compounds
The 12 compounds selected by in silico screening were kindly provided by MCCB group.

Cell lines and replicons
The human hepatoma (Huh-7) and rhabdomyosarcoma (RD) cell lines were grown in DMEM, supplemented with 10% FBS, 0.5 mM non-essential amino acids and 100 units/mL penicillin-streptomycin. Murine myoblast cells (C2C12) were grown in DMEM supplemented with 20% FBS and 100 units/mL penicillin-streptomycin.All cells were maintained at 37°C with 5% CO 2 in a humidified incubator.

Cell viability assay
Cell viability was measured by MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide] (Sigma-Aldrich) assay.24 hours post treatment, compound-containing media was removed from 96 well plates and MTT at 1 mg/mL solution was added to each well, incubated for 1 hour and replaced with 100 μL of DMSO to solubilize the formazan crystals.The absorbance was measured at 560 nm on Infinite F50 microplate reader (Tecan).Cell viability was calculated according to the equation (T/C) × 100%, which T and C represented the optical density of the treated well and control groups, respectively.DMSO was used as non-treated control.The cytotoxic concentration of 50% (CC 50 ) was calculated using Prism (Graph Pad).

RNA Transcription and electroporation
Plasmids were linearized using restriction enzyme NotI-HF (New England Biolabs), prior to in vitro transcription using the mMESSAGE mMACHINE™ SP6 Transcription Kit (Invitrogen) following the manufacturer's protocol.RNA was purified using the LiCl precipitation protocol.For electroporation, 3×10 6 cells were electroporated with 2 µg of RNA in 4 mm cuvette, 260 V, 25 ms, 1 pulse, using GenePulser Xcell (Bio-Rad) (Roberts et al., 2017).Cells were plated in 48 well plates (6 × 10 4 cells/well) for replication assays and 96 well plates (2 × 10 4 cells/well) for cell viability assays.Four hours post-electroporation (h.p.e.) the media was removed, replaced with compound-containing media in different concentrations and incubated for 24 hours.

Replication assay
To measure CHIKV replication levels, compound-containing media was removed 24 hours after treatment, cells were washed with PBS, and harvested by lysis with Passive Lysis Buffer (Promega).Samples were stored at − 20°C and thawed prior to measurement of luminescence levels on a FLUOStar optima microplate reader (BMG Labtech) using the Luciferase Assay System (Promega) for CHIKV-Fluc SGR and the Dual-Luciferase ® Reporter system (DLR™) (Promega) for CHIKV-D-Luc SGR.The effective concentration of 50% (EC 50 ) was calculated using Prism (Graph Pad).Replication assays were carried out in parallel with cell viability assays.

Statistical Analysis
Individual experiments were performed in triplicate and all assays were performed a minimum of three times in order to confirm the reproducibility of the results.Statistical analysis was performed using GraphPad Prism software.A Kruskal-Wallis test followed by post hoc analysis by Dunn's test was used for multiple comparisons P values of less than 0.05 (indicated by asterisks) were considered statistically significant.

In silico identification of ligands to the Macro domain of CHIKV-nsP3
To identify potential inhibitors against the macro domain of CHIKV nsP3, the structure of CHIKV nsP3 complexed with ADP-ribose (Protein Data Bank -3GPO) was used.The ADP-ribose substrate was manually removed from the ADP-ribose binding site and a library containing 48,750 small synthetic molecules were docked to this region of the protein with Schrodinger's docking tool Glide in vHTS mode using default parameters.The compounds were ranked based on Glide docking scores, a prediction of binding affinity, and the top 1000 scoring compounds were re-docked using Glide XP mode, a higher precision scoring function.After that, the top 50 compounds were visually inspected.And 12 compounds were selected for biological testing based on the docking score (Table 1); inspection of hydrogen bonds between ADP-ribose binding site of CHIKV nsP3 and the compounds; the conformation of each docking pose and the position of the compound inside the binding site relative to the ADP-ribose.

Initial in vitro activity of compounds selected by docking analysis-The
12 compounds selected by docking analysis were evaluated in vitro by their antiviral activity using the CHIKV-Fluc SGR system.Huh-7 cells were electroporated with CHIKV-Fluc Europe PMC Funders Author Manuscripts Europe PMC Funders Author Manuscripts SGR RNA, seed in 48 well plates and treated with each compound at 20 μM, 10 μM and 1 μM at 4 h.p.e. to assess the effect of these compounds on both CHIKV replication and cell viability.Figure 1 presents data obtained from the treatment with the maximum non-toxic concentration of each compound.The results demonstrated that at 20 μM compounds C5 and C13 inhibited CHIKV replication by 53% and 66% respectively, and retained 94% of cell viability (Figure 1, Table 1).Other compounds did not demonstrate significant activity against CHIKV replication.

3.2.2
Compounds with anti-CHIKV activity and related analogues-In silico analysis showed predicted hydrogen bonding interactions of compounds C5 and C13 with residue Tyr114 and van der Waals interactions/hydrophobic effect with residue Val33 and Asn24 (Figure 2A).After initial in vitro screening, compounds C5 and C13 were selected for further assays and structural analogues were purchased from ChemDiv (Figure 2B).
The EC 50 and CC 50 of C5, C13 and their analogues were determined by performing replication and viability assays in cells treated with a range of concentrations of each compound.Huh-7 cells were electroporated with CHIKV-Fluc SGR RNA, seeded in 48 well plates and treated with a two-fold dilution (ranging from 1.5-100 μM) of each compound at 4 h.p.e.Cell viability and CHIKV replication levels were measured 24 h post treatment (Figure 3).The selective index (SI) was calculated by dividing the value of CC 50 by EC 50 (CC 50 /EC 50 ).The EC 50 and CC 50 values and SI of C5, C13 and their analogues are shown in Table 2. Compound C5 showed EC 50 of 12.7 μM and CC 50 of 38.9 μM, and demonstrated lower EC 50 value than its analogues 1C5, 2C5, 3C5 and 4C5.Compound C13 presented a slightly higher EC 50 (11.9μM) and CC 50 (36.8μM).While the C13 analogue 1C13 showed a decrease in EC 50 , the value o6 CC 50 decreased almost by a half.

Investigation of C5 and C13 compounds mechanisms-Compounds C5
and C13 presented the highest SI and their effects on CHIKV replication were further investigated.To this end, the CHIKV-D-Luc replicon containing two reporter genes, the renilla luciferase (Rluc) and firefly luciferase (Fluc) was used, the replicon is described in (Roberts et al., 2017).The reporter Rluc is present in the hypervariable region of nsP3 of CHIKV and its expression indicate both input translation and early replication levels.Alternatively, the reporter Fluc replaces the virus structural genes and it is only expressed from a subgenomic RNA, thus indicating that RNA replication has occurred (Figure 4A).In order to investigate a possible mechanism of C5 and C13 were acting during early or late replication, Huh-7 cells were electroporated with CHIKV-D-Luc and treated with C5 (Figure 4B) or C13 (Figure 4C) at non-toxic concentrations (concentrations above 80% of cell viability were considerated non-toxic).The data obtained demonstrated that the treatment with C5 (Figure 4B) and C13 (Figure 4C) compound significantly decreased Rluc levels, but not Fluc.Therefore, suggesting that these compounds are most likely interfering with early replication events.
Finally, the effects of C5 and C13 were also investigated in muscle derived cell lines.For this, human rhabdomyosarcoma (RD) cells and murine myoblast (C2C12) cells were used.However, the results demonstrated that C5 and C13 were only able to inhibit CHIKV replication at concentrations that also demonstrated some cell toxicity (Figure 5).Ats 100 Europe PMC Funders Author Manuscripts Europe PMC Funders Author Manuscripts µM, C5 inhibited 100% and 70% of CHIKV replication in RD and C2C12, respectively (cell viability 65% and 49%) (Figure 5A and 5B).C13 inhibited CHIKV replication in 95% and 64% in RD and C2C12, respectively (cell viability 70% and 51%) (Figure 5B and 5D).

Discussion
Virtual screening has the advantage of reducing costs/time to identify hit compounds when compared to screening large numbers of compounds in an HTS approach (Piccirillo and Amaral, 2018).In this study, the macro domain of CHIKV nsP3 was investigated since this protein has been reviewed and described to participate in several processes of the viral cycle.
The macro domain of CHIKV nsP3 possesses ADP-ribose binding and ADP-ribose hydrolase activity, removing ADP-ribose from mono(ADP-ribosyl)ated proteins, this activity has been shown to be important for viral replication (McPherson et al., 2017).CHIKV mutants with reduced hydrolase activity showed slower virus replication and decrease of virulence, and mutants with no hydrolase activity were unable to replicate in mammalian or mosquito cells (McPherson et al., 2017).Therefore, a compound that can block CHIKV nsP3 hydrolase activity is proposed to be a target for antiviral therapy.Additionally, the macro domain of nsP3 is a well conserved region in the Alphavirus genus and other positive single strand RNA viruses, demonstrating the potential to be a broad spectrum antiviral (Eckei et al., 2017;Gregor et al., 2016).
A library of 48,750 small molecules was initially screened in silico to predict their ability to bind to ADP-ribose binding site of nsP3.Twelve selected compounds were further investigated in vitro for both their toxicity to human hepatoma cells and antiviral activity on CHIKV replication.Compounds C5 and C13 presented the highest selective index (ratio between the cell viability and the replication inhibition) (5.2 and 31, respectively), and demonstrated to share a similar mode of binding to ADP-ribose binding site, in terms of shape and electrostatic potential.Therefore, analogues of C5 and C13 were also investigated, however, they were less active than the original compounds.Moreover, C5 and C13 demonstrated interference in the early steps of CHIKV replication (for example formation of the replication complex) rather than after RNA replication has occurred.
However, most of the computer-aided virtual screening studies based on CHIKV nsP3 describe only in silico data (Pérez-Pérez et al., 2019).Nguyen and collaborators used the database Diversity Set II chemical library of the National Cancer Institute (NCI) to virtually screen 1541 compounds in conjunction with the structure of the CHIKV nsP3 macro domain from the Protein Data Bank (PDB id: 3GPG) and selected five potential ligands (Nguyen et al., 2014).Vijayasri and collaborators also used the same CHIKV nsP3 macro domain structure to dock 150 phytochemicals from various plant sources, among them fisetin and quercetin (Vijayasri and Hopper, 2017).Later, the phytochemical fisetin and quercetagetin, compound similar to quercetin, showed in vitro anti-CHIKV activity by inhibiting RNA production and viral protein expression (Lani et al., 2016).Baicalein was also considered a potential antiviral candidate against CHIKV nsP3 in a computational study (Seyedi et al., 2016).In 2018, Oo and collaborators showed that baicalein also exhibited binding affinity to CHIKV envelope protein E2.Baicalein present higher inhibition on virus entry than Europe PMC Funders Author Manuscripts Europe PMC Funders Author Manuscripts replication.The authors suggested a possible interaction between the compound and cellular factors (Oo et al., 2018).ADP-ribosylation is a common post-translational modification in cellular proteins (Hottiger et al., 2010).In this regard, the tested compounds showed considerable cytotoxicity, this might be because they were selected to bind to the ADP-ribose binding site of nsP3.By definition this will be similar to the ADP-ribose binding site of cellular macrodomains (Schleicher et al., 2018).ADP-ribosylation is known to regulate cell DNA repair, cell proliferation, transcription and cell death (Barkauskaite et al., 2015;Munnur and Ahel, 2017).Therefore, we suggest that these selected compounds against nsP3 macrodomain may be acting on these other cell pathways affecting cell viability.
In this regard it is interesting to note that compounds C5 and C13 showed higher cytotoxicity in muscle derived cell lines RD and C2C12.Roberts and collaborators, showed that different cell lines present variance in the levels of replication (Roberts et al., 2017).The cytotoxicity observed for RD and C2C12 cell lines could be related to the metabolism of these cells, as well as their ability to support CHIKV replication.
In addition interferon treatment induces the expression of several cellular ADPribosyltransferases which act as inhibitors of cellular translation and virus replication (Atasheva et al., 2014), leading to the hypothesis that de-ADP-ribosylation activity of macro domains can act as a response to cellular antiviral mechanisms.Inhibition of nsP3 macro domain activity would therefore have a two-pronged effect -direct inhibition of viral RNA replication and potential blockade of the innate immune response.However, more studies are needed to precisely understand the role(s) of the nsP3 macro domain during the virus life-cycle.

Conclusion
In this study a virtual screening approach has been used to identify putative CHIKV nsP3 inhibitors that target the ADP-ribose binding site.In vitro analysis showed that two of the selected compounds inhibited CHIKV replication, probably acting in the early replication.Computer-aided drug design (CADD) is a facilitator in the process of searching for new antivirals, however, the selection of compounds by docking does not eliminate other variables in a living system and needs to be combined with biological assays.Docking scores, physicochemical properties and biological activity of selected ligands docked to ADP-ribose binding site at the macro domain of CHIKV nsP3.HBA = hydrogen bond acceptor.HBD = hydrogen bond donor.MW = molecular weight.PSA = polar surface area.

Figure 1 .
Figure 1.In vitro screening of 12 MCCB compounds selected from docking analyses.CHIKV-Fluc SGR replicon scheme (A).Schematic representation of in vitro initial trials of MCCB compounds (B).The results of the treatment with maximum non-toxic concentration of Huh-7 cells harbouring CHIKV-Fluc SGR are shown for cell viability and activity against CHIKV replication (C).DMSO was used as non-treated control.The asterisks indicate statistically significant differences between each compound and DMSO control (*p ≤ 0.05, **p ≤ 0.01), calculated using the Kruskal-Wallis test with the Dunn's post test.

Figure 2 .
Figure 2. Structure of compounds C5, C13 and analogues.(A) Predicted binding conformations of C5 and C13 docked within to the ADP-ribose binding site of CHIKV nsP3 macro domain.Predicted hydrogen bonding interactions between 2D compound structures and CHIKV nsP3 amino acids are shown with a dotted line.(B) Molecular structures of the selected analogues (1C5, 2C5, 3C5, 4C5).

Figure 3 .
Figure 3.Activity of selected compounds and analogues against CHIKV replication.Huh-7 cells were electroporated with CHIKV-Fluc-SGR RNA, seed and treated 4 h later with compounds at concentrations ranging from 100µM to 3µM.Cell viability (■) was measured by MTT and replication (•) by luciferase assay 24 h post treatment.DMSO was used as non-treated control.P < 0.05 was considered significant, the asterisks indicate statistically significant differences between each compound and DMSO control (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 and ****p ≤ 0.0001), calculated using the Kruskal-Wallis test with the Dunn's post test.

Figure 4 .Figure 5 .
Figure 4. Effect of C13 and C5 at different stages in CHIKV replication.CHIKV-D-Luc wild type RNA (A) were electroporated in Huh-7 cells used to analyse the activity of C5 (B) and C13 (C).DMSO was used as non-treated control.The asterisks indicate statistically significant differences between each compound and DMSO control (*p ≤ 0.05, **p ≤ 0.01).P < 0.05 was considered significant calculated using the Kruskal-Wallis test with the Dunn's post test.