Rates of evolutionary change in viruses: patterns and determinants

Key Points Viral mutation rates vary over five orders of magnitude, whereas viral substitution rates vary over six orders of magnitude. Instead of simplifying the differences by stating that RNA viruses mutate faster than DNA viruses owing to differences in polymerase fidelity, it seems more likely that small viruses mutate faster than large viruses irrespective of the nucleic acid of their genome. Single-stranded DNA (ssDNA) and double-stranded DNA (dsDNA) viruses have non-overlapping ranges of mutation and substitution rates, with ssDNA viruses behaving more like RNA viruses. However, there are currently no good estimates of substitution rate for dsRNA viruses. There are several sources of mutation in addition to polymerase error. These include host antiviral enzymes, spontaneous chemical reactions and environmental mutagens such as ultraviolet irradiation. Various processes shape the evolution of mutation rates in viruses, although more research is needed to determine their precise contribution, and whether and how natural selection has acted to optimize these rates. Coalescent methods that use serially sampled data represent a powerful way to estimate substitution rates from rapidly evolving RNA and ssDNA viruses. It is conventionally thought that there is a simple relationship between viral mutation rates and polymerase fidelity. This article argues that the pattern of virus evolution is also shaped by other aspects of viral biology. Understanding the factors that determine the rate at which genomes generate and fix mutations provides important insights into key evolutionary mechanisms. We review our current knowledge of the rates of mutation and substitution, as well as their determinants, in RNA viruses, DNA viruses and retroviruses. We show that the high rate of nucleotide substitution in RNA viruses is matched by some DNA viruses, suggesting that evolutionary rates in viruses are explained by diverse aspects of viral biology, such as genomic architecture and replication speed, and not simply by polymerase fidelity.