UC Berkeley

Flaviviruses are arthropod-borne, positive-sense RNA viruses of significant medical and public health importance. While safe and effective vaccines have been developed against some members of the Flavivirus genus, limited progress has been made against others, such as dengue (DENV), West Nile (WNV), and Zika (ZIKV) viruses – which cause distinct human diseases and affect different tissues. Importantly, there are currently no flavivirus-specific therapeutics. Flaviviruses encode a secreted non-structural protein 1 (NS1) that directly interacts with endothelial cells, resulting in barrier dysfunction. The clinical manifestations of these interactions are thought to include DENV NS1-induced vascular leak in the lungs, WNV NS1 enhancing viral infection of the brain, and ZIKV NS1 causing hyperpermeability in placental tissues. We previously found that flavivirus NS1 binds to endothelial cells and disrupts barrier functions in a tissue-specific manner consistent with the disease tropism of the respective viruses. Understanding the underlying molecular mechanism of tissue-specific NS1-endothelial cell interactions will help guide better interventions to prevent and control flaviviral diseases. To elucidate the distinct role(s) that the domains of NS1 (β-roll, wing, and β-ladder) play in NS1 interactions with endothelial cells, we constructed flavivirus NS1 chimeras that exchanged the wing and β-ladder domains in a pair-wise manner between DENV, WNV, and ZIKV NS1. We found that both the NS1 wing and β-ladder domains conferred NS1 tissue-specific endothelial dysfunction, with the wing conferring cell binding and the β-ladder involved in inducing endothelial hyperpermeability as measured by transendothelial electrical resistance assay. Then, we utilized the DENV-WNV NS1 chimera and identified residues 91 to 93 (GDI) of DENV NS1 as a molecular motif determining binding specificity. Using a mouse model of localized leak, we corroborated that the NS1 wing domain and GDI molecular determinant motif triggered NS1-induced vascular leak in vivo. Further, we report other molecular determinants of interest, including a highly conserved motif in the NS1 wing domain that mediates pan-flavivirus NS1 binding to pulmonary endothelial cells and identified the wing domain to mediate interactions with endothelial cell surface glycans. Finally, we explored the protective effects of the iminosugar UV-4B against DENV NS1-induced endothelial dysfunctions by inhibiting proper host-dependent glycosylation processing of NS1. Taken together, these findings contribute to improved understanding of flavivirus NS1 and the molecular details conferring its tissue-specific functional patterns, that could help guide future vaccine design and therapeutic developments.