The Influenza Hemagglutinin Cytoplasmic Tail Domain Interacts with Phosphatidylinositol 4,5-bisphosphate

Abstract

During the influenza viral life cycle, the viral glycoprotein hemagglutinin (HA) mediates binding, entry, and fusion. Densely packed clusters of HA trimers at the plasma membrane are required to produce infectious virions; however, the mechanism of HA clustering is still unknown. We have shown previously that HA co-clusters with and modulates phosphatidylinositol 4,5-bisphosphate (PIP2) in host cell plasma membranes (PM). Here, we further characterize the relationship between HA and PIP2 using molecular dynamics simulations (MD) and fluorescence photoactivation localization microscopy (FPALM) to elucidate a mechanism of HA-PIP2 interaction. We found that the interaction occurs largely between the PIP2 head group and the cytoplasmic tail domain (CTD) of HA. Mutations of the CTD were made to alter charge (HARE, HARREQ), palmitoylation sites (HAMAY), or a combination thereof (HAREMAY, RREQMAY). MD showed that HARREQ and RREQMAY had the strongest effect on HA-PIP2 interactions through a depletion in the radial distribution function of PIP2 around HA at distances [≤]2.5 nm. FPALM revealed that HA cluster density at the PM was significantly reduced by CTD mutations, with the largest reduction occurring in mutants where the CTD charge and acylation were both altered (HAREMAY). HAREMAY clusters were also found to have larger circularities and perimeters, implying a structural change to the clusters. Mutations in the HA transmembrane domain also caused modest changes to the cluster properties of HA and its co-clustering with PIP2. FPALM showed PIP2 co-clustering with HA was also affected by HA mutations with more free PIP2 localized under HAREMAY clusters. A chemical model of simultaneous HA-PIP2 and PH-PIP2 binding enables interpretation of HA-PIP2 interactions and reveals quantitative differences between PIP2 binding by HA CTD mutants. We conclude that the mechanism of HA-PIP2 interaction consists of at least electrostatic and hydrophobic components. Our insights into the mechanism of HA-PIP2 interaction, and the prevalence of putative PIP2-interacting domains in a number of viral spike proteins suggest it may be fruitful to identify methods of disrupting interactions between phosphoinositides and viral proteins.