Calibration of Multiple Poliovirus Molecular Clocks Covering an Extended Evolutionary Range

We have calibrated five different molecular clocks for circulating poliovirus based upon the rates of fixation of total substitutions ( K t ), synonymous substitutions ( K s ), synonymous transitions ( A s ), synonymous transversions ( B s ), and nonsynonymous substitutions ( K a ) into the P1/capsid region (2,643 nucleotides). Rates were determined over a 10-year period by analysis of sequences of 31 wild poliovirus type 1 isolates representing a well-defined phylogeny derived from a common imported ancestor. Similar rates were obtained by linear regression, the maximum likelihood/single-rate dated-tip method, and Bayesian inference. The very rapid K t [(1.03 ± 0.10) × 10 −2 substitutions/site/year] and K s [(1.00 ± 0.08) × 10 −2 ] clocks were driven primarily by the A s clock [(0.96 ± 0.09) × 10 −2 ], the B s clock was ∼10-fold slower [(0.10 ± 0.03) × 10 −2 ], and the more stochastic K a clock was ∼30-fold slower [(0.03 ± 0.01) × 10 −2 ]. Nonsynonymous substitutions at all P1/capsid sites, including the neutralizing antigenic sites, appeared to be constrained by purifying selection. Simulation of the evolution of third-codon positions suggested that saturation of synonymous transitions would be evident at 10 years and complete at ∼65 years of independent transmission. Saturation of synonymous transversions was predicted to be minimal at 20 years and incomplete at 100 years. The rapid evolution of the K t , K s , and A s clocks can be used to estimate the dates of divergence of closely related viruses, whereas the slower B s and K a clocks may be used to explore deeper evolutionary relationships within and across poliovirus genotypes.