The molecular basis of virulence of the encephalitogenic flaviviruses.

Summary: Studies of flavivirus virulence have focussed on variants with differing neurovirulence or neuroinvasiveness in animal models, with the majority based on determining the contribution of viral structural proteins, in particular the envelope glycoprotein. Specific virulence determinants on protein E have recently been identified by the application of several experimental techniques such as selection and characterization of neutralization escape variants in the presence of anti-E protein MAbs or by site-directed mutagenesis of the E gene of infectious cDNA clones (see Fig. 2, Tables 2 And 3). However, the association of particular amino acid substitutions in protein E with loss of neurovirulence and/or neuroinvasiveness appears to be dependent on the virus system under study and thus the mechanism by which protein E controls these two virulence phenotypes is unclear. Existing data show that neurovirulence and neuroinvasiveness determinants map to identical regions of protein E, including the lateral surface of domain III (the putative receptor binding site) and the base of domain II (thought to be involved in pH-dependent fusion activity). Attenuation of some domain II and III variants is associated with changes in binding to cell membrane proteins or in pH-dependent fusion activity, suggesting that defects in the early events of virus replication are responsible for virulence attenuation. A small number of amino acid changes in domain I are associated with attenuation of virulence, including mutations in the protein E glycosylation site which result in significant reductions in viral glycoprotein expression. The mechanism of attenuation associated with other mutations in domain I is unknown. Several studies have identified putative neurovirulence determinants within other regions of the genome (e.g. non-structural protein genes, 3′ UTR), and in one example the mutation attenuates neurovirulence without alteration in neuroinvasiveness. Further work is necessary to map neurovirulence determinants within these regions. It is likely that significant advances in this field will come from the application of infectious cDNA clone technology.