The coronavirus (CoV) RNA genome is the largest among single-stranded positive sense RNA viruses. CoVs encode a proofreading 3′-to-5′ exoribonuclease within nonstructural protein 14 (nsp14-ExoN) that is responsible for CoV high-fidelity replication. Alanine substitution of ExoN catalytic residues [ExoN(-)] in SARS-CoV and murine hepatitis virus (MHV) disrupts ExoN activity, yielding viable mutant viruses with defective replication, up to 20-fold decreased fidelity, and increased susceptibility to nucleoside analogs. To test the stability of the ExoN(-) genotype and phenotype, we passaged MHV-ExoN(-) 250 times in cultured cells (P250), in parallel with WT-MHV. Compared to MHV-ExoN(-) P3, MHV-ExoN(-) P250 demonstrated enhanced replication, reduced susceptibility to nucleoside analogs, and increased competitive fitness. However, passage did not select for complete or partial reversion at the ExoN-inactivating mutations. We identified novel amino acid changes within the RNA-dependent RNA polymerase (nsp12-RdRp) and nsp14 of MHV-ExoN(-) P250 that partially account for the observed changes in replication, susceptibility to nucleoside analogs, and competitive fitness observed in the passaged virus population, indicating that additional determinants can compensate for the activities of nsp14-ExoN. Our results suggest that while selection favors restoration of replication fidelity in ExoN(-) CoVs, there may be a significant barrier to ExoN(-) reversion. These results also support the hypothesis that high-fidelity replication is linked to CoV fitness and identify additional candidate proteins that may regulate CoV replication fidelity.

Proofreading-deficient coronaviruses adapt over long-term passage for increased fidelity and fitness without reversion of exoribonuclease-inactivating mutations