A case for the ancient origin of coronaviruses

Abstract

Coronaviruses are found in a diverse array of bat and bird species, which are believed to act as natural hosts. Molecular clock dating analyses of coronaviruses suggest that the most recent common ancestor of these viruses existed around 10,000 years ago. This relatively young age is in sharp contrast to the ancient evolutionary history of their putative natural hosts, which began diversifying tens of millions of years ago. Here, we attempted to resolve this discrepancy by applying more realistic evolutionary models that have previously revealed the ancient evolutionary history of other RNA viruses. By explicitly modeling variation in the strength of natural selection over time and thereby improving the modeling of substitution saturation, we found that the time to the most recent ancestor common for all coronaviruses is likely far greater (millions of years) than the previously inferred range.

Fig 1

Schematic diagram of the SARS coronavirus genome. Open reading frame 1a (ORF1a) and ORF1b, encoding the nonstructural polyproteins, and those encoding the S, E, M, and N structural proteins are indicated. The approximate locations of the viral sequences used in this study are shown.

Fig 2

Branch length expansion under the BS-REL model compared to that under the GTR + Γ₄ model for six coronavirus sequences: Nsp15⁺, Nsp16, matrix protein, PLP2, RdRp, and the Y domain. Each point represents a branch in the coronavirus phylogeny. For ease of visualization, BS-REL branch lengths experiencing extreme saturation (>50 substitutions per site) are depicted with infinite length.

Fig 3

Substitution saturation curves obtained with the GTR + Γ₄ and BS-REL models on the basis of RdRp simulations. The solid line represents the simulated values, symbols are plotted at median values (100 replicates for each tree length value), and dotted whiskers show the 2.5-to-97.5% range.