Back

Bovine coronavirus evolution preserves spike architecture while remodeling key functional regions

Tzarum, N.; Tarabih, H.; Asiku, J.; Levi-Kalisman, Y.; Slavin, M.; Zalk, R.; Shahar, A.; Weisz, J.; Fraenkel, R.; David, D.; Kalisman, N.; Sol, A.

2026-08-05 microbiology
10.64898/2026.08.05.743018 bioRxiv
Show abstract

Bovine coronavirus (BCoV) is a major pathogen of cattle and the closest known ancestor of the human coronavirus OC43, yet how viral evolution reshapes spike protein structure and function remains poorly understood. Here, we combined comparative genomics, glycoproteomics, cryo-EM, and antigenic characterization to define the structural mechanisms underlying spike evolution across representative BCoV lineages. We identify a previously unrecognized lineage-specific N-glycosylation site in contemporary European viruses and validate its occupancy by glycoproteomics. High-resolution cryo- EM structures reveal that BCoV evolution preserves the overall prefusion architecture of the spike glycoprotein while selectively remodeling key functional regions involved in receptor recognition, conformational dynamics, and antigenicity. Comparative analysis with OC43 demonstrates increased conformational heterogeneity of the receptor-binding loop, whereas the Mebus vaccine strain exhibits enhanced membrane-proximal stalk flexibility despite maintaining thermal stability. Finally, structural modeling together with antibody-binding experiments reveals substantial antigenic remodeling despite >90% spike sequence identity between BCoV and OC43. Together, these findings establish a mechanistic framework for Embecovirus spike evolution and provide structural insights that may inform the development of vaccines based on contemporary circulating strains.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.