Bayesian estimates of the evolutionary rate and age of hepatitis B virus

Accurately estimating the evolutionary rate and age of hepatitis B virus (HBV) has proven to be one of the most difficult problems in studies of viral evolution. To help resolve these issues we employed a recently developed Bayesian coalescent approach to globally sampled human and avian hepadnavirus genome sequences, accounting for lineage-specific rate variation, the presence of overlapping reading frames, and the potential impact of recombination. Our analysis revealed an unexpectedly high rate of evolutionary change--up to 10(-4) nucleotide substitutions (subs) per site per year and always more than approximately 10(-6) subs/site/year. These rates suggested a time to the most recent common ancestor (tMRCA) of the sampled isolates of consistently less than approximately 1500 years ago for human HBV and less than 6000 years ago for the avian hepadnaviruses. Notably, the evolutionary rate of nonoverlapping regions of the viral genome was approximately 2-fold greater than that of overlapping genome regions, reflecting the complex patterns of selective constraint inherent in the former. We also reveal that most recombination events in both human and avian HBV tend to fall in a specific region of the viral genome, which contains all four viral open reading frames and which may therefore represent a "hot spot" for recombination. However, while recombination affects estimates of both evolutionary rate and tMRCA, in no case was this sufficient to challenge the hypothesis that the dominant mode of HBV evolution is by recent cross-species transmission. We conclude that HBV exhibits rapid evolutionary dynamics, typical of other viruses dependent on reverse transcriptase-mediated replication.