Gain without pain: Adaptation and increased virulence of Zika virus in vertebrate host without fitness cost in mosquito vector

Abstract

Zika virus (ZIKV) is now in a post-pandemic period, for which the potential for re-emergence and future spread is unknown. Adding to this uncertainty is the unique capacity of ZIKV to directly transmit between humans via sexual transmission. Recently, we demonstrated that direct transmission of ZIKV between vertebrate hosts leads to rapid adaptation resulting in enhanced virulence in mice and the emergence of three amino acid substitutions (NS2A-A117V, NS2A-A117T, and NS4A-E19G) shared among all vertebrate-passaged lineages. Here, we further characterized these host-adapted viruses and found that vertebrate-passaged viruses also have enhanced transmission potential in mosquitoes. To understand the contribution of genetic changes to the enhanced virulence and transmission phenotype, we engineered these amino acid substitutions, singly and in combination, into a ZIKV infectious clone. We found that NS4A-E19G contributed to the enhanced virulence and mortality phenotype in mice. Further analyses revealed that NS4A-E19G results in increased neurotropism and distinct innate immune signaling patterns in the brain. None of the substitutions contributed to changes in transmission potential in mosquitoes. Together, these findings suggest that direct transmission chains could enable the emergence of more virulent ZIKV strains without compromising mosquito transmission capacity, although the underlying genetics of these adaptations are complex.

Figure 1.. Vector competence of serially passaged Zika virus lineages.

Female Ae. aegypti mosquitoes were exposed to passaged ZIKV strains via an artificial infectious bloodmeal and collected 7, 14, and 21 days post feeding (dpf). Infection (a) dissemination (b) and transmission (c) rates over the three collection time points are shown. Infection rate is the percentage of ZIKV-positive bodies, dissemination rate is the percentage of positive legs, and transmission rate is the percentage of positive saliva samples of total mosquitoes that took a bloodmeal (determined by plaque assay). Data points represent the empirically measured percentages (n=39–80 per data point). The lines represent the logistic regression results and the shaded areas represent the 95% confidence intervals of the logistic regression fits. Infection, dissemination, and transmission rates of VP-A and VP-C were compared to ZIKV-BC at each timepoint (two-tailed fisher’s exact test). *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001; ns, not significant. Infectious virus was quantified via plaque assay from bodies (d) legs (e) and saliva (f) from all positive samples. Mean titers were not significantly different between virus groups in any tissue at any timepoint (one-way ANOVA with Tukey’s multiple comparisons test). ns, not significant. Paired legs and saliva samples from individual mosquitoes exposed to VP-A (g) and VP-C (h) 21 dpf were deep sequenced. Lines represent single nucleotide variant (SNV) frequency percentages between the stock or bloodmeal, legs, and saliva.

Figure 2:. In vitro and in vivo characterization of ZIKV mutant clones.

In vitro growth kinetics of ZIKV clones in Vero (a) and C6/36 (b) cells. Data points represent the mean of three replicates at each time point. Error bars represent standard deviation. (c) Serum viremia 2 and 4 days post infection (dpi) of Ifnar1^−/− mice inoculated with 10³ PFU of different ZIKV clones (n=8 for virus groups, n=4 for PBS control). Serum viremia from mutant clones was compared to previously published viremia data from VP-C infection. Differences in mean serum viremia between virus groups was compared by one-way ANOVA with Tukey’s multiple comparisons test. ns, not significant. The dotted line indicates the assay limit of detection. (d) Survival curves of Ifnar1^−/− mice inoculated with 10³ PFU of virus, or a PBS control. HD-E19G and HD-WTic groups were inoculated with 10⁴ PFU. Survival curves were compared to WTic by Fisher’s exact test. *, p < 0.05. (e) Chromatograms from Sanger sequencing of a subset of E19G 7d and 9d and VP-C 7d brains showing confirmation of the maintained NS4A position 19 substitution (nucleotide substitution at nt 6519 of the polyprotein: GAG → GGG). Sequenced amplicons were aligned to ZIKV-PRVABC59.

Figure 3:. Differential neurotropism and innate immune gene responses to ZIKV mutants.

Viral RNA (a) and infectious virus (b) were quantified from brain tissue collected 3, 7, and 9–10 days after 10³ PFU inoculation of Ifnar1^−/− mice with ZIKV-IC, NS4A-E19G, VP-C or, double mutant clones by qRT-PCR and plaque assay respectively. Transcript abundance of RIG-I (c), MDA-5 (d), MAVS (e), and TLR3 (f) was analyzed from brains collected 3, 7, and 9–10 dpi by qPCR. Expression levels were normalized to ActB and Hprt and the ddC_T value was calculated as log2 fold change expression relative to mock-inoculated mice. Data points represent individual samples. Means with standard deviation are shown. Viral loads, infectious viral titers, and relative fold change of transcript abundance were compared between virus groups at each timepoint by one-way ANOVA with Tukey’s multiple comparisons test. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001; ns, not significant.

Figure 4:. Vector competence of ZIKV mutants.

Female Ae. aegypti mosquitoes were exposed to passaged ZIKV strains via an artificial infectious bloodmeal and collected 7, 14, and 21 days post feeding (dpf). Infection (a) dissemination (b) and transmission (c) rates over the three collection time points are shown. Infection rate is the percentage of ZIKV-positive bodies, dissemination rate is the percentage of positive legs, and transmission rate is the percentage of positive saliva samples (determined by plaque assay screen). Data points represent the empirically measured percentages (n= 20–80 per data point, point size is proportional to group size). The lines represent the logistic regression results and the shaded areas represent the 95% confidence intervals of the logistic regression fits.