Characterising alphacoronavirus phenotypic traits through diversity-driven selection of spike

The capacity for viruses to spillover from one host to another is dependent on their ability to bind to and enter cells from a new host. Using a computational approach that maximises phylogenetic diversity, we selected an optimal subset of 40 alphacoronavirus spike proteins, including the two human viruses NL63 and 229E, and characterised their host-range using broad mammalian APN and ACE2 receptor libraries. Based on this data, we were able to determine molecular genotypes that contribute to receptor tropism and identify alphacoronaviruses with broad (generalist) or restricted (specialist) receptor usage. Strikingly, we observed that generalism and specialism can vary significantly between closely related viruses. Using structural information, we identified key residues that determine the bat tropism of 229E-like viruses, as well as residues in the ACE2 receptor that likely restrict NL63 infection of certain mammals. Furthermore, we observed that the host range of certain bat alphacoronaviruses is expanded by TMPRSS2 priming of spike. All APN- and ACE2-using alphaCoVs in our study interacted with the receptor of at least one animal species within ecological proximity to humans, suggesting new routes for spillover. However, most bat alphacoronaviruses did not use any of receptors in our screen, refining our understanding of coronavirus entry. We propose a new approach to investigating receptor usage for viral orders/families/genera, which is not constrained by focused analysis of a limited number of sequences from human-tropic viruses. Understanding phenotypic traits, such as entry, at this genus-wide level can revolutionise our ability to predict zoonotic potential.