Inferring nonneutral evolution from contrasting patterns of polymorphisms and divergences in different protein coding regions of enterovirus 71 circulating in Taiwan during 1998-2003

Background. Enterovirus (EV) 71 is one of the common causative agents for hand, foot, and, mouth disease (HFMD). In recent years, the virus caused several outbreaks with high numbers of deaths and severe neurological complications. Despite the importance of these epidemics, several aspects of the evolutionary and epidemiological dynamics, including viral nucleotide variations within and between different outbreaks, rates of change in immune-related structural regions vs. non-structural regions, and forces driving the evolution of EV71, are still not clear. Results. We sequenced four genomic segments, i.e., the 5' untranslated region (UTR), VP1, 2A, and 3C, of 395 EV71 viral strains collected from 1998 to 2003 in Taiwan. The phylogenies derived from different genomic segments revealed different relationships, indicating frequent sequence recombinations as previously noted. In addition to simple recombinations, exchanges of the P1 domain between different species/genotypes of human enterovirus species (HEV)-A were repeatedly observed. Contrasting patterns of polymorphisms and divergences were found between structural (VP1) and non-structural segments (2A and 3C), i.e., the former was less polymorphic within an outbreak but more divergent between different HEV-A species than the latter two. Our computer simulation demonstrated a significant excess of amino acid replacements in the VP1 region implying its possible role in adaptive evolution. Between different epidemic seasons, we observed high viral diversity in the epidemic peaks followed by severe reductions in diversity. Viruses sampled in successive epidemic seasons were not sister to each other, indicating that the annual outbreaks of EV71 were due to genetically distinct lineages. Conclusions. Based on observations of accelerated amino acid changes and frequent exchanges of the P1 domain, we propose that positive selection and subsequent frequent domain shuffling are two important mechanisms for generating new genotypes of HEV-A. Our viral dynamics analysis suggested that the importation of EV71 from surrounding areas likely contributes to local EV71 outbreaks.