Influence of Zika virus 3'-end sequence and nonstructural protein evolution on the viral replication competence and virulence

Abstract

Zika virus (ZIKV) has been circulating in human networks over 70 years since its first appearance in Africa, yet little is known about whether the viral 3'-terminal sequence and nonstructural (NS) protein diverged genetically from ancient ZIKV have different effects on viral replication and virulence in currently prevailing Asian lineage ZIKV. Here we show, by a reverse genetics approach using an infectious cDNA clone for a consensus sequence (Con1) of ZIKV, which represents Asian ZIKV strains, and another clone derived from the MR766 strain isolated in Uganda, Africa in 1947, that the 3'-end sequence -UUUCU-3' homogeneously present in MR766 genome and the -GUCU-3' sequence strictly conserved in Asian ZIKV isolates are functionally equivalent in viral replication and gene expression. By gene swapping experiments using the two infectious cDNA clones, we show that the NS1-5 proteins of MR766 enhance replication competence of ZIKV Con1. The Con1, which was less virulent than MR766, acquired severe bilateral hindlimb paralysis when its NS1-5 genes were replaced by the counterparts of MR766 in type I interferon receptor (IFNAR1)-deficient A129 mice. Moreover, MR766 NS5 RNA-dependent RNA polymerase (RdRp) alone also rendered the Con1 virulent, despite there being no difference in RdRp activity between MR766 and Con1 NS5 proteins. By contrast, the Con1 derivatives expressing MR766 Nsps, like Con1, did not develop severe disease in wild-type mice treated with an IFNAR1 blocking antibody. Together, our findings uncover an unprecedented role for ZIKV NS proteins in determining viral pathogenicity in immunocompromised hosts.

Keywords: 3′-end sequence variants; Zika virus; infectious cDNA clone; nonstructural proteins; virulence determinant.

Figure 1.

Construction of a full-length cDNA clone for a consensus sequence of ZIKV. (A) Schematic illustration of the full-length cDNA clone pBAC-T7-ZK-Con1, which harbours ZIKV cDNA fragments (Con1-1, Con1-2, and Con1-3 representing the ZIKV Con1 genome nt 37–3,445, 3,426–5,870, and 5,851–8,340, respectively) fused to T7 promoter and the cDNA of the last 2,392-nt (nt 8,416–10806) of the Con1 sequence followed by HDV ribozyme (HDVr) and SacII restriction site. The four unique restriction enzyme sites used to assemble the cDNA fragments (Con1-1–3) into a modified pMW119 vector, prior to subcloning the resulting full-length cDNA into a BAC vector, are presented. (B) Infectivity of recombinant ZIKV Con1 rescued from Vero E6 cells transfected with T7 in vitro transcripts of pBAC-T7-ZK-Con1. Infectious viral titres of the Con1 viral stocks with indicated passage numbers were determined by plaque assay. Con1(NS5_GAA), a replication-defective mutant containing a GAA substitution at the active site of NS5 viral RdRp. ND, not detected (limit of detection, 10 PFU/ml). (C) Growth kinetics of PRVABC59 and Con1 in Vero E6 and A549 cells infected by each virus at an MOI of 0.01. (D) Growth attenuation features of Con1 and its derivative Con1(I80T/A2611V) in comparison with PRVABC59 strain in Vero E6 and A549 cells infected with each virus at an MOI of 0.01. (E) Differences in amino acid sequences between PRVABC59 reference sequence (GenBank KU501215.1), the PRVABC59 stock used in infection experiments in this study, and Con1. (F) Relative diameters of plaques from Con1 and the cell culture-adapted PRVABC59 (n = 60 plaques from 3 independent experiments). Plaque diameters were measured digitally using plaque images taken with the sizing bar. Right, representative images showing plaques 4 days after infection. In panels B–D, the results are the mean ± SD from three independent replicates. Statistical analyses were performed using unpaired Student’s t-test (C and F) or one-way ANOVA test (D) on log₁₀-transformed data. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Figure 2.

Conservation of the last 4-nt of ZIKV 3′-end genome sequence in contemporary Asian strains, diverged from the ancient African lineage MR766 strain. (A) Diverged 3′-end viral genome sequences in African and Asian lineage ZIKV isolates. (B and C) Schematic procedure of RACE-PCR. RNA was extracted from ZIKV MR766 strain (GenBank KX830960) and PRVABC59 (GenBank KU501215.1), ligated to the 5′-adenylated DNA adaptor, and reverse-transcribed to cDNA. The cDNA was amplified using a forward primer targeting the 3′-UTR of viral genome (nt 10,607–10,627) and a reverse primer targeting the 3′-adapter (B). The amplicon libraries were sequenced on an Illumina Nextseq platform. Shown in (C) are the proportions of the last 4–5 nucleotides at the 3′-end of PRVABC59 and MR766 genomes.

Figure 3.

Effect of the ZIKV 3′-end sequence variance on viral replication and gene expression. (A) Schematic illustration of the pBAC-CMV-ZK-Con1 subgenomic replicon and its derivative terminating with a heterologous –TTTCT-3′ end sequence originated from MR766. The puromycin resistance gene (Puro) and Renilla luciferase (Rluc) genes, which are separated from each other by FMDV 2A peptide-coding sequence, were in-frame fused downstream of the first 38 amino acids of the viral capsid gene (C38). (B) Inhibition of subgenomic replicon (Con1_sgRep) replication by 2′-C-methyladenosine (2′-CMA), an NS5 inhibitor. Rluc activity was measured at 24 and 36 h post-transfection of Huh7 cells with pBAC-CMV-ZK-Con1_sgRep. (C) Effect of the ZIKV 3′-end sequence variance on subgenomic replicon replication, assessed at the indicated time as described in (B). Con1_sgRep(NS5_GAA), a negative control expressing a defective NS5. (D) Effect of the ZIKV 3′-end sequence variations on viral RNA translation. Rluc activity was measured at 24 h post-transfection of Huh7 cells with the indicated Rluc-expressing in vitro transcribed ZIKV minigenomic RNAs where the Rluc gene was fused to the N-terminal part (38 amino acids) of capsid-coding gene placed in between viral 5′-UTR and 3′-UTR. (E and F) Schematic representation of the secondary structures of ZIKV FL 3′-UTR (E) and 3′-SL (F). SL, stem-loop; DB, dumbbell; sfRNA, subgenomic flavivirus RNA; XRN1, 5′→3′ exoribonuclease 1. Decay kinetics of radioisotope-labeled in vitro transcripts of ZIKV FL 3′-UTR (E) and 3′-SL RNA (3′-SL) (F) with either –GUCU or –UUUCU 3′-terminal sequence in Vero E6 cell lysates (40 ug). The intensity of intact input RNA (mean ± SD from three independent experiments) quantified using a Phosphorimager was presented below representative autoradiography images. Statistical analyses were performed using unpaired Student’s t-test (B and D) or one-way ANOVA test (C) on log₁₀-transformed data. *P < 0.05; ***P < 0.001; ns, not significant.

Figure 4.

Functional equivalence of the evolutionally diverged ZIKV 3′-end sequences in viral genome replication. Comparisons of growth kinetics of Con1 (A) and rMR766 (B) with their derivatives Con1(UUUCU) and rMR766(GUCU) in Vero E6 and A549 cells (MOI = 0.01). ns, not significant (by unpaired Student’s t-test).

Figure 5.

Growth kinetics of Con1, rMR766, and their chimaera expressing heterologous nonstructural viral proteins. (A) Schematics of ZIKV Con1 and MR766 chimeric viruses. (B and C) Growth kinetics of the indicated, rescued recombinant viruses in Vero E6 (MOI = 0.01, using P1 stock of each virus). Shown are the results (mean ± SD) from three independent replicates. Statistical analyses were performed using one-way ANOVA test with multiple comparisons (B) or unpaired Student’s t-test (C) on log₁₀-transformed data. **P < 0.01; ***P < 0.001; ns, not significant.

Figure 6.

Virulence and neuropathogenicity of Con1 and Con1-derived chimeric viruses in A129 mice. (A) Infection and sample collection schedule (top). Schematics of chimeric viruses derived from Con1 and rMR766 (bottom). (B and C) For survival analysis, IFNAR^−/− A129 mice (n = 12 with 4 female and 8 male) were infected with each ZIKV (10² PFU) by sc injection into the footpad. Body weight was monitored daily for 15 days (B). Combined data from two independent experiments are shown as mean ± SEM. Comparisons of Kaplan-Meier survival curves between different groups were performed by log-rank analysis (C). Shown in doughnut charts is the frequency of paralysis observed in infected mice. (D–F) Viral loads in sera on days 3 and 5 (D) and in indicated tissues on day 5 (E). The dotted lines in (D) shows the limit of detection (LOD). Shown in (F) are fold-changes in viral genome copy number in tissues with mean values indicated at the top of each bar. Each dot represents a result from a separate animal. Statistical analyses were performed using unpaired Student’s t-test (B) or one-way ANOVA test with multiple comparisons on log₁₀-transformed data (D and E). *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Figure 7.

African lineage ZIKV structural proteins account for the enhanced pathogenicity of MR766 in wild-type C57BL/6 mice transiently blocked in type I IFN signaling. (A) Schematic of experimental design showing the infection, antibody injection, and sample collection schedule. (B and C) Body weight changes and survival analysis of C57BL/6 mice (six-week-old, n = 10 per each group with 5 female and 5 male mice) treated with an anti-type I IFNR antibody MAb-5A before and after being exposed to 10⁴ PFU (ip injection) of indicated rZIKVs. Statistical analysis methods are identical to those described in Figures 6B and C. (D) Clinical disease in mice infected with indicated rZIKVs on different days of post-infection. Bar charts show the frequency of five different clinical signs of disease. (E–I) Comparison of viral RNA levels (E, F, H, and I) or infectious viral loads (G) in sera (E, F, and G) and tissues (H and I) between mice on 4 and 7 dpi with indicated rZIKVs. Each dot represents a result from a separate animal. (J) Brain tissue samples of infected mice were collected on 7 dpi, stained with H&E for histopathological evaluation. Representative micrographs of the cortex of mock- and rZIKV-infected mice, with higher magnification images (inset) of the area outlined by the white box. Arrows, perivascular cuffing with infiltration of inflammatory cells in the brain tissues (cerebral cortex) of infected mice; asterisk, necrotic cellular debris. Scale bars, 100 µm. In (E–H), statistical analyses were performed using one-way ANOVA test with multiple comparisons on log₁₀-transformed data. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.