Non-optimal codon usage preferences of coronaviruses determine their promiscuity for infecting multiple hosts.

Carmi, Gon; Gorohovski, Alessandro; Mukherjee, Sumit; Frenkel-Morgenstern, Milana

Circulating animal coronaviruses occasionally infect humans. The SARS-CoV-2 coronavirus is responsible for the current worldwide outbreak of COVID-19 that has resulted in 2,112,844 deaths as of late January, 2021. We compared genetic code preferences in 496 viruses, including 34 coronaviruses, and 242 corresponding hosts, to uncover patterns that distinguish single and "promiscuous" multiple host-infecting viruses. Based on a codon usage preference score, promiscuous viruses were shown to significantly employ non-optimal codons, namely, codons that involve "wobble" binding to anti-codons, as compared to single host viruses. The codon adaptation index (CAI) and the effective number of codons (ENC) were calculated for all viruses and hosts. Promiscuous viruses were less adapted hosts versus single host viruses (p-value=4.392e-11). All coronaviruses exploit non-optimal codons to infect multiple hosts. We found that non-optimal codon preferences at the beginning of viral coding sequences enhance the translational efficiency of viral proteins within the host. Finally, coronaviruses lack endogenous RNA degradation motifs to a significant degree, thereby increasing viral mRNA burden and infection load. To conclude, we found that promiscuously infecting coronaviruses prefer non-optimal codon usage to remove degradation motifs from their RNAs and to dramatically increase their viral RNA production rates.