A Small-Plaque Isolate of the Zika Virus with Envelope Domain III Mutations Affect Viral Entry and Replication in Mammalian but Not Mosquito Cells

Abstract

An Asian Zika virus (ZIKV) isolated from a Thai patient that was serially passaged in Primary Dog Kidney (PDK) cells for attenuation displayed both big and small plaque-forming viruses by the 7th passage. Two small-plaque isolates were selected and purified for characterization as attenuated ZIKV candidates. In vitro growth kinetics showed significantly reduced titers for small-plaque isolates in Vero cells early post-infection compared to the parental ZIKV and a big-plaque isolate, but no significant difference was observed in C6/36 cells. Viral entry experiments elucidate that titer reduction likely occurred due to the diminished entry capabilities of a small-plaque isolate. Additionally, a small-plaque isolate displayed lowered neurovirulence in newborn mice compared to 100% lethality from infection with the parental ZIKV. Genomic analysis revealed the same three unique non-synonymous mutations for both small-plaque isolates: two on the envelope (E) protein at residues 310, alanine to glutamic acid (A310E), and 393, glutamic acid to lysine (E393K), and one on residue 355 of NS3, histidine to tyrosine (H355Y). Three-dimensional (3D) mapping suggests that the E protein mutations located on the receptor-binding and fusion domain III likely affect cell entry, tropism, and virulence. These ZIKV isolates and genotypic markers will be beneficial for vaccine development.

Keywords: E protein; NS3; Zika virus; attenuation; cell tropism; vaccine.

Figure 1

Plaque morphology of wild-type, ZIKV-PDK7, ZIKV-PDK7-B5, ZIKV-PDK7-S3, and ZIKV-PDK7-S1.4 variants. Plaque morphology of ZIKV variants were observed by avicel-based plaque assays in Vero cells at 5 dpi. ZIKV-CVD_06-020, an Asian wild-type ZIKV, contains mixed plaque variants with plaques diameter ranging from 0.1 to 0.5 mm. ZIKV-PDK7—a wild-type ZIKV serially passaged in PDK cells for seven passages—shows an increase in small-plaque-forming virus population. ZIKV-PDK7-B5—a big plaque isolated from ZIKV-PDK7—presented homogeneous big-plaque-forming viruses with diameter ≥ 0.4 mm. ZIKV-PDK7-S3 and ZIKVV-PDK7-S1.4—small-plaque variants purified from ZIKV-PDK7—showed major populations of small-plaque-forming viruses with diameter < 0.2 mm.

Figure 2

(a) Two E mutations, A310E and E393K of ZIKV-PDK7-S3 and ZIKV-PDK7-S1.4, are mapped on the ZIKV E protein dimer using PBD: 5LBV as a template. Both mutations are located at domain III (DIII) of the E protein as red indicates. (b) NS3 mutation H355Y of ZIKV-PDK7-S3 and ZIKV-PDK7-S1.4 is mapped onto the ZIKV NS3 protein using PBD: 5GJB as a template. The mutation is located at domain II (DII) of the helicase–RNA complex. The dotted square presents an ATP/Mn2+ binding site in a cleft of domain I and II (DI and DII), which is composed of residues G197, K200, T201, R202 (P loop), E286, N330, R459, and R462.

Figure 3

Effects of mutation on E and NS3 helicase protein stability and interactomics in the mutation region. Dynamut normal mode analysis (EnCoM) was used to analyze and visualize the potential impact of the mutations dynamics and stability resulting from vibrational entropy changes. PBD: 5LBV was used as the model of E protein. A310E (a) and E393K (b) are predicted to be stabilizing and decrease the flexibility of the protein. PBD: 5GJB was used as the model of NS3 helicase protein. The H355Y mutation is predicted to be a slightly destabilizing mutation and slightly increases the flexibility of the protein (c). Change in interactomics of the mutation residue (shown in green) with surrounding residues was also predicted for A310E (a), E393K (b), and H355Y (c).

Figure 4

Effects of mutations on polarity and charge of protein in the mutation region. SWISS-MODEL was used to model and visualize the changes in charge (left) and polarity (right) in the mutation residues of E:DIII and NS3:DII, using PBD: 5LBV and PBD: 5GJB, respectively. The A310E mutation on E:DIII (a) shows a change from an uncharged, nonpolar residue to a negatively charged, polar residue, while the E393K mutation (b) goes from a negatively charged, polar residue to a positively charged polar residue. The H355Y mutation on NS3 (c) shows a change from a positively charged, polar residue at pH 7.2 to an uncharged, polar residue.

Figure 5

Growth kinetics of wild-type and ZIKV variants. Viral growth kinetics were determined in C6/36 (a) and Vero (b) cells. The cells were infected with ZIKVs; wild-type (circle), ZIKV-PDK7-B5 (triangle), ZIKV-PDK7-S3 (inverted triangle), and ZIKV-PDK7-S1.4 (square) at an MOI of 0.01. The experiment was performed in triplicate with duplicate plaque counting. The mean viral titers at each time-point were plotted with the error bar representing the standard deviation (SD). Two-way ANOVA was used to statistically analyze the data against the wild-type strain. Asterisks (*) indicate a statistically significant difference (p < 0.05). If not displayed, the data did not show significant differences.

Figure 6

Ratios of replication rate per day of ZIKV isolates. The fold-changes of replication rate were calculated from viral growth kinetics in C6/36 (a) and Vero cells (b). Bar graphs represent the ratio of increasing titers of ZIKV isolates per day between day 1 and day 5, and the error bars represent the SD. Fold-changes of viral yield per day are shown in the figure. A significant difference is calculated by using two-way ANOVA with Tukey’s multiple comparisons test. p-values: 0.1234 = ns; 0.0021 **; <0.0002 ***; <0.0001 ****.

Figure 7

Viral entry into C6/36 (a) and Vero (b) cells was determined by real-time PCR. The viral RNA copies number per cell was determined by real-time PCR. The cells were infected with ZIKVs, wild-type, ZIKV-PDK7-B5, and ZIKV-PDK7-S1.4, at an MOI of 10. After viral adsorption, the viruses were removed and treated with acid glycine to inactivate the remaining un-internalized viruses. After 24 h post-infection, the infected cells were harvested for RNA extraction and subjected to one-step real-time PCR. The experiment was performed in triplicate with triplicate real-time PCR quantification. The bar graph represents mean values, with the error bar representing SD. One-way ANOVA was used, and asterisks (*) indicate a statistically significant difference (p < 0.05).

Figure 8

Mouse neurovirulence. Two-day-old ICR mice were intracranially injected with 2 × 10⁴ pfu of the wild-type ZIKV (dotted line), ZIKV-PDK7-S1.4 (dash line), or NSS (solid line). The mice were observed daily for morbidity and mortality for 21 days. The survival rate analyzed by Log-rank (Mantel–Cox) tests shows significant differences between the mice infected with the ZIKV-PDK7-S1.4 isolate and the wild-type ZIKV infection (p < 0.0001).