Adaptive evolution of West Nile virus facilitated increased transmissibility and prevalence in New York State

Abstract

West Nile virus (WNV; Flavivirus, Flaviviridae) was introduced to New York State (NYS) in 1999 and rapidly expanded its range through the continental United States (US). Apart from the displacement of the introductory NY99 genotype with the WN02 genotype, there has been little evidence of adaptive evolution of WNV in the US. WNV NY10, characterized by shared amino acid substitutions R1331K and I2513M, emerged in 2010 coincident with increased WNV cases in humans and prevalence in mosquitoes. Previous studies demonstrated an increase in frequency of NY10 strains in NYS and evidence of positive selection. Here, we present updated surveillance and sequencing data for WNV in NYS and investigate if NY10 genotype strains are associated with phenotypic change consistent with an adaptive advantage. Results confirm a significant increase in prevalence in mosquitoes though 2018, and updated sequencing demonstrates a continued dominance of NY10. We evaluated NY10 strains in Culex pipiens mosquitoes to assess vector competence and found that the NY10 genotype is associated with both increased infectivity and transmissibility. Experimental infection of American robins (Turdus migratorius) was additionally completed to assess viremia kinetics of NY10 relative to WN02. Modelling the increased infectivity and transmissibility of the NY10 strains together with strain-specific viremia demonstrates a mechanistic basis for selection that has likely contributed to the increased prevalence of WNV in NYS.

Keywords: Flavivirus; West Nile virus; adaptive evolution; arbovirus; virus evolution.

Figure 1.

West Nile virus (WNV) prevalence in Culex mosquitoes in New York State (NYS). Prevalence (WNV positive/1000 tested) was calculated by maximum likelihood estimate (MLE) following pooling and testing of Culex mosquitoes by qRT-PCR. WNV prevalence was significantly greater from 2010–2018 relative to 2000–2009 (chi-squared test, ***p < 0.0001).

Figure 2.

Time tree based on Bayesian analysis of West Nile virus (WNV) isolates from New York State (NYS), ranging from 1999–2018 (BEAST 2). Branch colours reflect the age of taxa, red branches represent the most recent strains. NY10 and NY07 genotype strains are enclosed in the respective boxes on the phylogeny.

Figure 3.

West Nile virus (WNV) genotype displacement in New York State (NYS), 1999-2018. Proportions of sequenced isolates belonging to distinct genotypes are shown. The displacement of the NY99 genotype by WN02 occurred from 2002-2004. NY01, NY07 and NY10 genotypes share a common WN02 backbone.

Figure 4.

Mosquito infectivity of WN02 and NY10 strains in Culex pipiens. A) Proportion of mosquitoes infected at a given dose of WNV WN02 and NY10 strains. The titres at which 50% of mosquitoes are infected (ID50) are extrapolated and indicated by the dotted line. B) The estimated percent of infected Culex pipiens based on experimental results at each indicated range of input titres.

Figure 5.

Vector competence of Culex pipiens for WN02 and NY10 strains. A) At 5 days post-infection (DPI) Culex pipiens infected with NY10 strains showed significantly increased rates of infection and dissemination when compared to WN02. B) At 11 DPI, Culex pipiens infected with NY10 strains showed significantly greater rates of infection, dissemination, and transmission than WN02 infected mosquitoes [p < 0.05 for infection (*), dissemination (α), and transmission (β) respectively, when compared to WN02 by Fisher’s exact test].

Figure 6.

Viremia of American Robins (Turdus migratorius) when infected with WN02 and NY10 strains. (A) Mean WNV viremia values at each timepoint assayed +/- SEM for WN02 relative to NY10 strains (NY10A/C combined) (B) Individual viremia curves for WN02, NY10A and NY10C. The dotted line represents the experimental threshold for infectivity of WN02 to mosquitoes. Mean (vertical line) area under viremia curves (AUC) +/- SEM was calculated for both (C) total and (D) peak (>5.0 log₁₀ pfu/ml). Mean time infectious was calculated using both (E) days with viremia >5.0 log₁₀ pfu/ml and (F) adjusted infectious days based on mosquito studies (Figure 4). Significantly higher mean time infectious was measured for NY10 strains relative to WN02 (p = 0.044, Mann-Whitney test).

Figure 7.

West Nile virus (WNV) infectivity and transmissibility indices. A) Infectivity indices of NY10 strains compared to those of WN02. Points represent individual birds and vertical lines represent means. The average infectivity of NY10 strains was 2.7 times greater than that of WN02 (student’s t-test, p < 0.01). B) Transmissibility indices of NY10 strains when compared to WN02. Enhanced transmissibility of NY10 strains further increased the mean difference between genotypes (student’s t-test, p < 0.01).