Zika Virus NS3 Protease Pharmacophore Anchor Model and Drug Discovery

Abstract

Zika virus (ZIKV) of the flaviviridae family, is the cause of emerging infections characterized by fever, Guillain-Barré syndrome (GBS) in adults and microcephaly in newborns. There exists an urgent unmet clinical need for anti-ZIKV drugs for the treatment of infected individuals. In the current work, we aimed at the promising virus drug target, ZIKV NS3 protease and constructed a Pharmacophore Anchor (PA) model for the active site. The PA model reveals a total of 12 anchors (E, H, V) mapped across the active site subpockets. We further identified five of these anchors to be critical core anchors (CEH1, CH3, CH7, CV1, CV3) conserved across flaviviral proteases. The ZIKV protease PA model was then applied in anchor-enhanced virtual screening yielding 14 potential antiviral candidates, which were tested by in vitro assays. We discovered FDA drugs Asunaprevir and Simeprevir to have potent anti-ZIKV activities with EC₅₀ values 4.7 µM and 0.4 µM, inhibiting the viral protease with IC₅₀ values 6.0 µM and 2.6 µM respectively. Additionally, the PA model anchors aided in the exploration of inhibitor binding mechanisms. In conclusion, our PA model serves as a promising guide map for ZIKV protease targeted drug discovery and the identified 'previr' FDA drugs are promising for anti-ZIKV treatments.

Figure 1

Overview of the workflow.

Figure 2

The ZIKV NS3 protease Pharmacophore Anchor (PA) model - anchors and their features. (A) The ZIKV NS3 protease active site (5GJ4) with anchors (mesh spheres) at subpockets S1′, S1, S2, S3 (blue dotted curves) with anchor residues (sticks). The anchors are colored as E-red, H-green, V-grey, EV-purple, HV-orange. (B) The anchor features, interaction-types E-H-V or mixed, anchor residues with interacting side chain S & main chain M atoms (NS2B cofactor residues in italics), anchor moiety preferences showing top five prefered functional groups (with % among all interacting moieties).

Figure 3

ZIKV NS3 protease PA model anchors with inhibitors and substrate peptides. (A) The crystal structures of eight inhibitor-bound ZIKV NS3 proteases overlapped with the protease PA model depicting active site subpockets (surface), anchors (mesh spheres), residues and inhibitors (both shown in sticks). The anchor occupancy profiles of inhibitors are shown as heat-map (green – anchor interaction, black - no interaction). (B) Summary of ZIKV NS3 protease substrate cleavage motifs in the genome polyprotein showing P4-P3-P2-P1↓P1′ residues; Web logo showing consensus residues at cleavage motifs. (C) ZIKV NS3 protease PA model with P4-P1 substrate peptides, crystal structure for bound peptide TGKR (5GJ4) and docked poses for substrate peptides VTRR, SGKR, AGKR, VKRR and Bez-VKKR-H are displayed; the anchor profiles of the substrate peptides are displayed as heat-map. (D) At each subpocket, the binding substrate peptide residues P1, P2, P3 and P4 are seen occupying the anchors.

Figure 4

Anchor-enhanced virtual screening and inhibitor candidates. (A) Stepwise virtual screening using ZIKV PA model anchors to obtain inhibitor candidates. (B) 2D structures of 14 inhibitor candidates including 10 FDA drugs (blue outline), one Maybridge compound (orange outline), one natural product (green outline) and two NCI compounds (purple outline).

Figure 5

Analysis of anti-ZIKV activities of inhibitor candidates. (A) MTT assay for evaluation of the cytotoxicities of the inhibitor candidates at 5 µM concentration, post-treatment and incubation for 24 hours. (B) Candidate compounds tested for activity against ZIKV by Fluorescence reporter assay, where the infected cells were incubated with 5 µM of compounds for 24 hours. The fluorescence intensities measured correspond to the amount of ZIKV.

Figure 6

Quantification of inhibitor activities. (A) In MTT assay for cytotoxicities, Asunaprevir and Simeprevir were tested at various concentrations by measuring the cell viabilities (by Absorbance at O.D., 570 nM) after inhibitor treatment and 24-hour incubation; further curves were plotted to obtain CC₅₀ values. (B) ZIKV production assay measuring inhibitor antiviral activities, in which U-87 MG cells were infected with ZIKV (PRVABC-59, MOI = 1), treated with various inhibitor doses and incubated for 24 hours. The residual virus in the titer was quantified by measuring the viral focus forming units (ffu/ml); the statistical significance versus DMSO (**p < 0.0021, ***p < 0.0002, ****p < 0.0001) by Dunnett’s multiple comparison test was represented. The results plotted as dose-inhibition curves and inhibitor EC₅₀ values were calculated. (C) ZIKV NS2B/NS3 protease inhibition assay to assess the protease activity inhibition by Asunaprevir and Simeprevir at multiple doses. The residual protease activities in the presence of inhibitors were measured, sigmoidal curves were plotted and the IC₅₀ values were calculated.

Figure 7

Analysis of inhibitor binding mechanisms for Asunaprevir and Simeprevir. (A) Anchor occupancy profiles for Asunaprevir, Simeprevir and known inhibitor 6T8 (green – interaction, black - no interaction). Inhibitor binding models with ZIKV NS3 protease for Asunaprevir, Simeprevir and 6T8 (in 5LC0). (B) 2D structures of, i. Asunaprevir, ii. Simeprevir and iii. 6T8 showing inhibitor substructures (colored magenta, brown, blue and green) occupying the anchors. Comparison of interaction energies by corresponding substructures from three inhibitors. ^ainteraction energy scoring function not applicable (due to covalent bonding of 6T8 moiety to Ser135).