Mutations abrogating VP35 interaction with double-stranded RNA render Ebola virus avirulent in guinea pigs

Abstract

Ebola virus (EBOV) protein VP35 is a double-stranded RNA (dsRNA) binding inhibitor of host interferon (IFN)-alpha/beta responses that also functions as a viral polymerase cofactor. Recent structural studies identified key features, including a central basic patch, required for VP35 dsRNA binding activity. To address the functional significance of these VP35 structural features for EBOV replication and pathogenesis, two point mutations, K319A/R322A, that abrogate VP35 dsRNA binding activity and severely impair its suppression of IFN-alpha/beta production were identified. Solution nuclear magnetic resonance (NMR) spectroscopy and X-ray crystallography reveal minimal structural perturbations in the K319A/R322A VP35 double mutant and suggest that loss of basic charge leads to altered function. Recombinant EBOVs encoding the mutant VP35 exhibit, relative to wild-type VP35 viruses, minimal growth attenuation in IFN-defective Vero cells but severe impairment in IFN-competent cells. In guinea pigs, the VP35 mutant virus revealed a complete loss of virulence. Strikingly, the VP35 mutant virus effectively immunized animals against subsequent wild-type EBOV challenge. These in vivo studies, using recombinant EBOV viruses, combined with the accompanying biochemical and structural analyses directly correlate VP35 dsRNA binding and IFN inhibition functions with viral pathogenesis. Moreover, these studies provide a framework for the development of antivirals targeting this critical EBOV virulence factor.

FIG. 1.

Alignment of the amino acid sequences of the carboxy-terminal IIDs of Ebola and Marburg virus VP35s. Comparison of the filoviral VP35 IID sequences for Zaire EBOV, Reston EBOV, Sudan EBOV, and MARV. Residues that are completely conserved across all four sequences are highlighted in black, and residues that are identical or show high sequence similarity are highlighted in gray. Basic residues are highlighted with an asterisk, and K319 and R322 are indicated by arrowheads.

FIG. 2.

KRA mutant VP35 lacks dsRNA binding and IFN antagonist activity but retains polymerase cofactor function. (A) Wild-type but not KRA mutant VP35 binds to poly(I·C)-Sepharose beads. WT or KRA mutant VP35 transfected cell lysates were incubated with poly(I·C) beads (+) or control Sepharose beads (−). The precipitated material was analyzed by Western blotting for VP35 (upper panel). Whole-cell extracts (WCX) were probed for VP35 (lower panel). (B) Polymerase cofactor activity of WT and KRA mutant VP35 was assessed with an EBOV minigenome assay. Wedges represent relative amounts of transfected WT or mutant VP35 expression plasmid. The value obtained with the higher concentration of WT VP35 was set to 100%. (C) WT and KRA mutant VP35 inhibition of SeV-induced ISG54 promoter activation. Cells were transfected with empty vector (EV) or increasing amounts (indicated by wedges) of WT VP35 or KRA mutant VP35 expression plasmid, an ISG54 promoter-firefly luciferase reporter plasmid, and a constitutively expressed Renilla luciferase reporter plasmid. Firefly luciferase activity was normalized to Renilla luciferase activity, and the activities of the EV-transfected and SeV-infected samples were set to 100% (top panel). Western blots show expression levels of WT and KRA mutant VP35 (lower panel). (D) WT and KRA mutant VP35 inhibition of RIG-I-induced IFN-β-promoter activation. Cells were transfected with empty vector (EV) or increasing amounts (indicated by wedges) of WT or KRA mutant VP35 expression plasmid, an IFN-β-promoter-firefly luciferase reporter plasmid, and a constitutively expressed Renilla luciferase reporter plasmid. Uninduced samples received additional empty vector DNA, and induced samples (RIG-I) received HA-tagged RIG-I expression plasmid. Firefly luciferase activity was normalized to Renilla luciferase activity, and the activities of the EV, RIG-I-transfected samples were set to 100% (top panel). Western blots (WB) show expression levels of WT and KRA mutant VP35 (lower panel). (E) IFN bioassay to assess suppression of endogenous IFN-α/β production by WT and KRA mutant VP35. Cells were transfected with 4 μg of empty vector (○ and •), VP35 (▴), or KRA mutant VP35 (▪) expression plasmids. Cells were mock infected (open symbol) or infected with SeV (solid symbols). Two-fold dilutions of UV-irradiated supernatants were applied to Vero cells. Direct treatment of Vero cells with dilutions of human IFN-β were included to generate a standard curve (×). Treated Vero cells were then infected with NDV-GFP, and virus replication was quantified with a fluorescence plate reader. (F) Inhibition of SeV-induced IRF-3 (S396) phosphorylation by WT and KRA mutant VP35. Cells were transfected with indicated amounts of empty vector (EV), WT, or KRA mutant VP35. Cells were mock infected or infected with SeV. Lysates were analyzed by Western blotting using anti-phospho-S396 IRF-3 (P-IRF-3) antibody (top panel), total IRF-3 antibody (middle panel), and an anti-VP35 antibody (lower panel). Error bars represent one standard deviation of results of at least three independent experiments.

FIG. 3.

Mutation of K319 and R322 does not significantly alter the structure of VP35 IID. (A) Alignment of WT VP35 IID (magenta, PDB 3FKE) and KRA mutant VP35 IID (cyan) crystal structures. (B) 1H/15N HSQC NMR spectrum overlay of WT and KRA mutant VP35 IID proteins showing localized chemical shift changes. Chemical shifts corresponding to WT, K319A, R322A, and KRA are colored black, red, green, and blue, respectively. Electrostatic surface potential (scale of −10 kT e⁻¹ to +10 kT e⁻¹) of VP35 IID structures for wild-type (C) and KRA mutant (D) proteins. The dsRNA binding “footprint” is shown in black outline, and residues that are mutated are identified. Electrostatic potentials: blue, positive; white, neutral; red, negative.

FIG. 4.

Growth of EBOV/VP35KRA is strongly attenuated in 293T cells. Vero E6 and 293T cells were infected at an MOI of 0.01 with EBOVwt, EBOVwt/GFP, EBOV/VP35KRA, or EBOV/VP35KRA/GFP. (A) The replication of recombinant viruses expressing GFP was monitored using fluorescence microscopy on the indicated days postinfection (p.i.). Single GFP-expressing 293T cells observed after infection with EBOV/VP35KRA/GFP are indicated by arrows. (B) The culture medium from cells infected with either EBOVwt or EBOV/VP35KRA was harvested at different intervals postinfection and analyzed by Western blotting using polyclonal horse anti-EBOV antibodies. (C) Relative virus infectivity of virus stocks was estimated by normalization of the infectious titers (TCID₅₀/ml) determined by virus titration on Vero E6 cells to the amount of viral proteins as quantified by Western blot analysis using rabbit anti-VP24 antibody and goat anti-rabbit antibodies. EBOV/VP35KRA is less infectious than wild-type EBOV.

FIG. 5.

EBOV/VP35KRA is highly attenuated in guinea pigs and protects against challenge with EBOVwt. (A) For the initial inoculation (on day 0), guinea pigs were mock infected (administered DMEM) (top panel) or inoculated intraperitoneally with 500 TCID₅₀ of either recombinant EBOVwt (middle panel) or recombinant EBOV/VP35KRA (bottom panel). On day 17 after the initial inoculation, animals were challenged with 500 TCID₅₀ of recombinant EBOVwt. Infections are indicated by labeled arrows. Animal weight was monitored for each animal on the indicated days postinfection (p.i.), and each animal is represented by an individual line. Legends under the graphs indicate a number designating individual animals. (B) Immunofluorescence analysis using sera from animals to stain Vero E6 cells transfected with plasmids expressing VP40 and NP of EBOV. Sera were collected from EBOVwt-infected animals on day 6 postinfection and from mock- and EBOV/VP35KRA-infected animals on day 17 postinfection. At 24 h posttransfection the cells were fixed and permeabilized using methanol/acetone. The cells were incubated with the sera from the animals as indicated and with secondary polyclonal donkey anti-guinea pig antibody coupled with Alexa 555. No specific anti-EBOV antibodies were detected in mock- or EBOVwt-infected animals. Animals infected with EBOV/VP35KRA showed specific anti-EBOV antibodies on day 17 postinfection. (C) Immunohistochemistry analysis of liver and spleen. Slides containing formalin-fixed tissues were processed and then stained using anti-EBOV VP40 mouse monoclonal antibody. EBOVwt-infected animals showed massive virus replication in liver and spleen. No specific staining was observed in tissues from EBOV/VP35KRA-infected animals.