Predicting antigenic sites on the FMDV capsid from cross-reactivity data

Abstract

Conventional FMD vaccines consist of chemically inactivated virus preparations that are applied parenterally. The selection of FMD vaccine candidates to be used in emergency and control settings is complicated by the genetic variability characteristic of this virus and the selection of mutants escaping the host immune response. Identifying epitopes for vaccine strains is therefore essential to determine whether the vaccine will protect against newly circulating outbreak strains from the same geographic location. Antigenically significant sites have previously been identified for O, A and C serotypes as well as an epitope for a Southern African Territories type 2 (SAT2) virus by sequencing of monoclonal escape mutants. We took a different approach by combining amino acid variation and in vitro cross-protection titres from virus neutralisation tests, together with crystallographic structural data, to derive similar information indirectly. Examining SAT1 and SAT2 strains, we identified a general negative correlation between genetic distance and serological relatedness (r-values), but more significantly we could identify areas on the surface of the capsid where mutations were strong predictors of antigenic distance. These were consistent within, but not between serotypes and were found to match some of the independently-identified antigenic sites in other serotypes, which reinforces our belief that this technique is effective at identifying epitopes. We predict that these sites should be under positive selection, and plan to verify this by examining non-synonymous to synonymous mutation rates, and further propose a reverse genetics approach to verify the predictions.