Phylogeography of West Nile virus: from the cradle of evolution in Africa to Eurasia, Australia, and the Americas

Abstract

West Nile virus (WNV) is the most widely distributed of the encephalitic flaviviruses and is a major cause of encephalitis, with isolates obtained from all continents, apart from Antarctica. Subsequent to its divergence from the other members of the Japanese encephalitis virus complex, presumably in Africa, WNV has diverged into individual lineages that mostly correspond with geographic distribution. Here we elucidate the phylogeography and evolutionary history of isolates from lineage 1 of WNV. Interestingly, there are many examples of the same amino acid having evolved independently on multiple occasions. In Africa, WNV exists in an endemic cycle, whereas it is epidemic in Europe, being reintroduced regularly from Africa either directly (in western Europe) or via the Middle East (in eastern Europe). Significantly, introduction into other geographic areas has occurred on one occasion only in each region, leading to subsequent establishment and expansion of the virus in these areas. Only one endemic genotype each is present in India and Australia, suggesting that WNV was successfully introduced into these locations once only. Each introduction occurred many centuries ago, probably due to trade and exploration during the 19th century. Likewise, in the Americas, WNV was successfully introduced in 1999 and subsequently became endemic across most temperate regions of North America (NA). In contrast to previous suggestions, an isolate from the epidemic in Israel in 1998 was not the direct progenitor of the NA epidemic; rather, both epidemics originated from the same (unknown) location.

FIG. 1.

Worldwide distribution of lineage 1 WNV isolates with known complete genome sequences. The number of isolates from each country is shown in parentheses. Dates of divergence and predicted direction of movement are shown for some isolates.

FIG. 2.

Maximum clade credibility (MCC) tree of WNV genomes constructed using BEAST version 1.5.3 software. Colors of branches indicate geographic locations per the color key. Branch lengths correspond to lengths of time, as measured by the scale bar, and the 95% HPD range of divergence dates is shown as a bar. The letters at the nodes correspond to data in Table 1. (A) All isolates (with the branch of isolates from the Americas collapsed). (B) Cluster 4 isolates only.

FIG. 3.

MCC tree and amino acid changes defining cluster 2. (A) Partial phylogenetic tree showing cluster 2 isolates only. (B) Summary of amino acid changes. Only changes occurring in more than one isolate are shown. For designations marked with an asterisk, isolates from the same outbreak (same year and location) have been combined for simplicity.

FIG. 4.

MCC tree and amino acid changes defining cluster 4. (A) Partial phylogenetic tree showing cluster 4 only. The established genotypes of isolates from the Americas have been collapsed for simplicity. (B) Summary of amino acid changes. Only changes occurring in two or more isolates are shown.