Structure of West Nile Virus

See allHide authors and affiliations

Science  10 Oct 2003:
Vol. 302, Issue 5643, pp. 248
DOI: 10.1126/science.1089316

West Nile virus (WNV) is found throughout Africa, Europe, Central Asia, and, most recently, in North America. The first outbreak in the United States was in New York City during the summer of 1999, and the virus subsequently spread across the United States (1). In 2002, there were 4156 reported cases of human illness associated with WNV and 284 deaths. WNV is a flavivirus transmitted primarily by Culex mosquitoes to vertebrate hosts. Flaviviruses include members such as dengue virus, yellow fever virus, and tick-borne encephalitis virus and are positive-strand RNA viruses that contain three structural proteins and a host-derived lipid bilayer. Two lineages of WNV (I and II) have been identified (2). All isolates that cause severe human disease fall into lineage I.

With the use of cryoelectron microscopy (cryo-EM) and image reconstruction techniques, we have determined a 17 Å resolution structure of WNV New York 99, the strain responsible for the outbreak in the United States (3). The virus has icosahedral symmetry and is ∼500 Å in diameter with no surface projections or spikes (Fig. 1A), which are prominent on other envelope-containing viruses such as influenza, HIV, and measles virus. A cross-sectional view (Fig. 1B) shows concentric layers of density indicating the multilayer organization of the virus. The outermost layer contains the highest density and corresponds to the E and M transmembrane proteins. The 35 to 40 Å thick lipid bilayer appears nonspherical with the transmembrane components of the M and E proteins visible. The nucleocapsid core contains multiple copies of the capsid protein and the genome RNA.

Fig. 1.

A 17 Å structure of West Nile virus determined by cryo-EM. (A) A surface shaded view of the virus with one asymmetric unit of the icosahedron shown by the triangle. The 5-fold and 3-fold icosahedral symmetry axes are labeled. (B) A central cross section showing the concentric layers of density. (C) The arrangement of the homology modeled WNV E protein (Cα-backbone is shown in blue). Residues 307 and 330, which bind neutralizing monoclonal antibodies, are shown in green. Residue Asn154, which is glycosylated, is shown in pink. (D) A difference density mapbetween dengue virus and WNV shown at a radius 248 Å. Positive density (from dengue virus) is shown in black and negative density (from WNV) is shown in white. The outlines for a set of three E homodimers are indicated by the blue lines.

The major surface protein E is responsible for receptor binding, host membrane fusion, and can elicit a neutralizing antibody response. Although the atomic structure of the WNV E protein is not known, structures of the ectodomains of the closely related dengue and tick-borne encephalitis virus E proteins have been determined by x-ray crystallography (4, 5). The long, thin molecule is arranged as a homodimer with each monomer having three domains (I, II, III). Domain III consists of an immunoglobulin (Ig)-like fold and is thought to contain the putative receptor binding site. Domain II has been proposed to promote the merging of the host and viral membranes. Based on the similarity between WNV and dengue virus E proteins (49% identity), a homology model of the WNV E protein was constructed and fitted into the outer layer of the cryo-EM density using the program EMFit (Fig. 1C) (6). A series of three dimers is aligned back to back to give a herringbone arrangement that forms a protein shell around the lipid bilayer. A large proportion of the surface of the E protein is accessible to neutralizing monoclonal antibodies (7).

The general arrangement of WNV and the fitted E glycoprotein is similar to that of dengue virus (6); however, several features unique to each virus can be seen in the difference density map (Fig. 1D). Positive difference density corresponds to the carbohydrate moieties of the glycosylation sites at E residues 67 and 153 in dengue virus. Negative density corresponds to the carbohydrate moiety found at E residue 154 in WNV. There is a significant positional shift between the densities contributed by the glycosylation sites of dengue E 153 and WNV E 154. This is likely due to the difference in the glycosylation site and because of an addition of five amino acids around this site in WNV. Whether these visible differences in the structure contribute to transmission and pathogenesis remain to be determined.

Supporting Online Material

Materials and Methods

Fig. S1

References and Notes

Stay Connected to Science

Navigate This Article