Fig 1.
The firefly split luciferase complementation conformational change assay.
(A) Important features of the active “closed” (PDB: 3U1I, magenta) and inactive “open” (2FOM, cyan and blue) conformations of NS2B. NS3s (gray) of 3U1I and 2FOM were best superimposed. Active-site inhibitor (yellow) and T119 and T120 of the 120 loop of NS3 were in stick representation. N-, C-termini of NS2B, and loops 120 (green) of NS3 were colored and labeled. NS3 residue A125 was in sphere representation. The NS2B N-terminal residues 50–66 of 2FOM were in blue and the C-terminal residues 67–95 of 2FOM were in cyan. Blue arrow indicates conformational change of the NS2B C-terminal portion (Cter) upon active-site inhibitor binding. The distances were in dashed lines: (yellow) between the Cter of NS2B in inactive conformation and the NS3 119 loop: 45Å; (blue) between the Cter of NS2B in active conformation and the NS3 119 loop: 11Å. (B) Cartoon representation of firefly luciferase (FLuc) with Nluc (aa. 1–398) in cyan and Cluc (aa. 398–550) in yellow and red. Luciferase inhibitor was shown in stick (magenta). (C) Schematic representation of the SLC strategy. When active site is not occupied, NS2B Nter (45–66) remains associated tightly with NS3, whereas NS2B Cter (67–95) is in the “open” conformation. SLC between NLuc and CLuc will not occur. No luminescence will be generated. (D) Binding of active-site inhibitor triggers conformational change of the NS2B Cter to form the “closed” conformation. SLC occurs. Luminescent light is generated. (E) Binding of allosteric inhibitor to the NS3 allosteric site prevents the NS2B Cter from forming the “closed” conformation, even in the presence of active-site inhibitor. SLC will not occur. No luminescence will be generated.
Fig 2.
The SLC assays for conformational change of NS2B of DENV2.
(A) Design of SLC construct for monitoring the NS2B conformational change (based on DENV2). (B) Active-site inhibitor NPGB triggers the SLC enhancement that can be suppressed by mutations at an allosteric site A125 of NS3. All proteins were at 150 nM. DTNB and NPGB were at 30 μM. n = 3. ***, p<0.01. (C) NPGB and DTNB do not affect the luciferase activity of the full-length luciferase (FLuc) (150 nM). (D) Protease activity of the WT and mutant N2CN3N. All proteins (150 nM) were assayed with the Abz substrate (100 μM) for 1 hour. The activity of WT was set as 100%. The relative activities of the mutants were normalized as percentage of the WT activities.
Fig 3.
A pocket unique to the inactive NS2B-NS3 structure.
(A) Surface representation of the unique pocket of the inactive NS2B-NS3 structure. (B) Surface representation of the active NS2B-NS3 structure at the same area. Surface color was according to atomic color as follows unless otherwise stated: oxygen, red; nitrogen, blue, carbon of residues involved in pocket formation, yellow; other carbon, white. (C) Mapping of sequence conservation to surface of the DENV2 NS3 protease domain (PDB code: 2FOM); the structure is viewed from two angles, showing features in different regions. Left: active site surface (yellow circle). Right: the unique pocket (cyan circle) of the inactive NS2B-NS3 structure. The surface is colored according to sequence conservation, resulting from the multiple sequence alignment. NS2B is displayed as a cartoon representation (green).
Fig 4.
Inhibition of in vitro NS2B-NS3pro activity by compounds identified by VS.
(A) Compounds identified by VS were added to an in vitro protease activity assay at a concentration of 100 μM and protease activity relative to the DMSO control was measured. No inhibitor (No Inh) was set as 100%. The protease activities with compounds were set as percentage of the No Inh control. (B) Schematic formulas of selected compounds showing in vitro anti-protease activities. (C) Concentration curves of in vitro protease activity inhibition by compounds NSC135618, 20594, and 146771. Compounds were incubated with the linked DENV2 NS2B-NS3 protease (150 nM) for 30 min, prior to addition of the Abz substrate (100 μM). Two fold dilutions of 135618 (from 15 μM to 0.23 μM), 260594 (from 100 μM to 1.6 μM), and 146771 (from 100 μM to 1.6 μM) were used. The protease activity of the DMSO control was set as 100%. The protease activities with compounds were set as percentage of the DMSO control. Data was fitted using the Sigmoidal model within the Origin6.0 software suite.
Table 1.
Inhibition of the DENV2 protease activity.
Fig 5.
Suppression of NPGB-induced SLC enhancement by candidate allosteric inhibitors.
(A) Active-site inhibitor NPGB can induce enhanced SLC which can be suppressed by allosteric inhibitors NSC135618, 260594, and 146771 at high concentration. The N2CN3N was at 150 nM. The compounds were at 40 μM. NSC135618 or NPGB was first incubated with the SLC protease construct for 30 min. in different orders, prior to addition of the luciferin substrate, followed by luminescence detection. N = 3. ***, P < 0.01, by one-way ANOVA. (B) The candidate compounds do not significantly affect the luciferase activity of FLuc (150 nM). The experimental procedure was the same as above. (C) Inhibition of the NPGB-induced SLC enhancement by a concentration series of candidate inhibitors, as indicated. In all panels, DMSO only control was set as 100% relative luminescence (RLU). Luciferase activities of the N2CN3N protease construct with NPGB and/or various compounds were set as percentage of the DMSO control. N = 3. ***, P < 0.01, by one-way ANOVA.
Fig 6.
Cell viability and viral reduction activity of candidate compounds in A549 cells.
(A) Inhibition of DENV2 infectivity on A549 cells by compounds NSC135618, 260594, and 146771 at a concentration of 1 μM (dark grey), 10 μM (black), and 100 μM (light grey) measured as titer of virus produced (on a log scale) relative to the DMSO control. (B) Cell viability of A549 cells in the presence of varying concentrations of NSC135618 or 260594. Cell viability for the DMSO control was set as 100%. Cell viability for compound-treated samples were set as percentage of the DMSO control. Curve fitting was performed using the Sigmoidal model within the Origin6.0 software suite. (C) Inhibition of DENV2 infectivity by varying concentrations of NSC135618 and 260594. Left: Sigmoidal curve fitting of experimental data expressed as percentage of virus titers (DMSO control was set as 100%). Right: Virus titers at indicated compound concentrations shown in log orders. (D) qRT-PCR analysis of inhibition of viral RNA from 135618-treated and DENV2-infected samples in A549 cells. Viral RNA copy number of the DMSO control was set as 100% with those of compounds treated samples as percentage of the DMSO control. For all panels, N = 3. ***, p<0.01, by one-way ANOVA.
Table 2.
Cell viabilities and antiviral activities using A549 cells.
Fig 7.
Inhibition of the ZIKV protease activity, viral titer, protein expression, and viral RNA replication by NSC135618.
(A) SDS-page analysis of purified linked and unlinked NS2B/NS3 protease complex of ZIKV. ST, molecular weight (MW) standard. (B) Dose response inhibition of the unlinked ZIKV NS2B-NS3 protease by NSC135618. The NS2b/NS3 protease was at 150 nM. N = 3. (C) Dose-dependent inhibition of ZIKV by NSC135618 in A549 cells. Viral plaque reduction assay was used. N = 3. (D) Immunofluorescence assay (IFA) of inhibition of viral protein expression for ZIKV-infected A549 cells by NSC135618, using pan flavivirus anti-E antibody 4G2. (E) qRT-PCR analyses of inhibition of viral RNA from supernatant of ZIKV infected A549 cells by NSC135618. N = 3, ***, p<0.01, by one-way ANOVA. All dose response curves were fitted with the Sigmoidal model using the Origin6.0 software suite.
Fig 8.
Inhibition of ZIKV in cells relevant to ZIKV by NSC135618.
(A,D) Dose-dependent inhibition of ZIKV by NSC135618 in human placental epithelia cells (HPECs) (A) and human neural progenitor cells (hNPCs) (D). (B,E) qRT-PCR analyses of inhibition of viral RNA from supernatant of ZIKV infected HPECs (B), and hNPCs (E) by NSC135618. N = 3. ***, p<0.001, by one-way ANOVA. (C,F) IFA of inhibition of viral protein expression for ZIKV-infected HPECs (C), and hNPCs (F) by NSC135618, using pan flavivirus anti-E antibody 4G2. For all these primary cells, the experiment was performed as described in A549 cells except that ZIKV with MOI of 2 were used. All dose response curves were fitted with the Sigmoidal model using the Origin6.0 software suite.
Fig 9.
Inhibition of viral polyprotein precursor (PP) processing by NSC135618.
(A) SDS-PAGE analysis of purified linked and unlinked NS2B/NS3 protease complex of DENV2. (B) PTSA for binding of NSC135618 (4.8 μM) to the DENV2 NS2B-NS3 proteins (2.5 μM). Thermal denaturation data was processed using a Derivative model using the Protein Thermal Shift Software v1.0 (ThermalFisher Scientific). ΔTm was defined as Tm-drug-Tm-DMSO. N = 3. (C) Lineweaver–Burk plot of kinetics experimental data for inhibition of the DENV2 NS2B-NS3 protease complex by NSC135618. The DENV2 NS2B-NS3 (150 nM) was mixed with NSC135618 (3 μM) for 30 min. The Abz substrate was added at various concentrations (400 μM to 25 μM in 2-fold dilutions). N = 3. (D) WB analysis of dose-dependent inhibition of ZIKV NS3 expression by NSC135618 using the GTX133309 ZIKV α-NS3 antibody (GeneTex Inc) (left panel). Right panel: NS3 expression (lower bands) normalized to the GAPDH loading control. PP, ZIKV polyprotein precursor. **, p<0.05. (E) MS/MS spectra obtained from the fragmentation of the precursor ion at m/z corresponding to representative ZIKV peptides. Fragment ions corresponding to y- and b-ions were observed (red and orange lines).
Fig 10.
NSC 135618 binding to the unique pocket of the inactive NS2B-NS3 structure.
(A) Predicted poses of compound NSC135618 in the allosteric pocket of the inactive NS2B-NS3 structure. Left: NS2B and NS3 are in surface representation with surface colors the same as in Fig 3. Right: NS3 residues predicted to be in contact with NSC135618 are in stick representation with atomic colors the same as in Fig 3. Potential hydrogen bonds are depicted as green dashed lines. NSC135618 is in stick representation with carbons colored white. (B) Protease activity of each mutant NS2B-NS3 protein relative to that of WT. All proteins were at 150 nM. The protease activity of WT was set as 100%, with that of the mutant protease as percentage of the WT. N = 3. ***, p<0.01, by one-way ANOVA. (C) Relative protease activities of WT and mutant NS2B-NS3 proteins with/without NSC135618 (6 μM). All proteins were at 150 nM. The protease activities of each protein with the DMSO control were set as 100%, with those of compound-treated samples as percentage of the DMSO control. N = 3. ***, p<0.01, by one-way ANOVA. (D) PTSA for binding of NSC1351618 (4.8 μM) to the WT or mutant DENV2 NS2B-NS3 proteins (2.5 μM). ΔTm was defined as Tm-drug-Tm-DMSO. Data was processed as described above. N = 3. ***, p<0.01, by one-way ANOVA.