Structural Basis and Evolution of Glycan Receptor Specificities within the Polyomavirus Family

Asymptomatic infections with polyomaviruses in humans are common, but these small viruses can cause severe diseases in immunocompromised hosts. New Jersey polyomavirus (NJPyV) was identified via a muscle biopsy in an organ transplant recipient with systemic vasculitis, myositis, and retinal blindness, and human polyomavirus 12 (HPyV12) was detected in human liver tissue. The evolutionary origins and potential diseases are not well understood for either virus. In order to define their receptor engagement strategies, we first used nuclear magnetic resonance (NMR) spectroscopy to establish that the major capsid proteins (VP1) of both viruses bind to sialic acid in solution. We then solved crystal structures of NJPyV and HPyV12 VP1 alone and in complex with sialylated glycans. NJPyV employs a novel binding site for a α2,3-linked sialic acid, whereas HPyV12 engages terminal α2,3- or α2,6-linked sialic acids in an exposed site similar to that found in Trichodysplasia spinulosa-associated polyomavirus (TSPyV). Gangliosides or glycoproteins, featuring in mammals usually terminal sialic acids, are therefore receptor candidates for both viruses. Structural analyses show that the sialic acid-binding site of NJPyV is conserved in chimpanzee polyomavirus (ChPyV) and that the sialic acid-binding site of HPyV12 is widely used across the entire polyomavirus family, including mammalian and avian polyomaviruses. A comparison with other polyomavirus-receptor complex structures shows that their capsids have evolved to generate several physically distinct virus-specific receptor-binding sites that can all specifically engage sialylated glycans through a limited number of contacts. Small changes in each site may have enabled host-switching events during the evolution of polyomaviruses.