The proton-conducting domain of the influenza A M2 homotetrameric channel (M2TM-AH; residues 22-62), consisting of four transmembrane (TM; residues 22-46) and four amphipathic helices (AHs; residues 47-62), promotes the release of viral RNA via acidification. Previous studies have also proposed the formation of clusters of M2 channels in the budding neck areas in raft-like domains of the plasma membrane,1,2which are rich in cholesterol, resulting in cell membrane scission and viral release. Experiments showed that cholesterol has a significant contribution to lipid bilayer undulations in viral buds suggesting a significant role for cholesterol in the budding process. However, a clear explanation of membrane curvature effect based on the distribution of cholesterol around M2TM-AH clusters is lacking. Using coarse-grained molecular dynamics simulations of M2TM-AH in bilayers, we observed that M2 channels form specific clusters with conical shapes, driven by attraction of their amphipathic helices (AHs). We showed that cholesterol stabilized the formation of M2 channel clusters, by filling and bridging the conical gap between M2 channels at specific sites in the N-terminals of adjacent channels or via the C-terminal region of TM and AHs, the latter sites displaying longer interaction time and higher stability. Potential of mean force calculations showed that when cholesterols occupy the identified interfacial binding sites between two M2 channels, the dimer is stabilized by 11 kJ/mol. This translates to the cholesterol-bound dimer being populated by almost two orders of magnitude compared to a dimer lacking cholesterol. We demonstrated that the cholesterol bridged M2 channels can exert lateral force on the surrounding membrane to induce the necessary negative Gaussian curvature profile which permits the spontaneous scission of the catenoid membrane neck and leads to viral buds and scission.Significance StatementA key role of influenza A M2 channel, a prototype viroporin, is the generation of curvature in the host cell membrane to form budding necks leading to scission and release of the virions. It has been shown experimentally that this process is mediated through the amphipathic helices of M2 and that cholesterol enhanced lipid bilayer undulations in viral budding. Here, using coarse-grained molecular dynamics simulations we revealed that M2 channels form clusters in membranes, induced in an additive manner by both the AHs and cholesterol, which in turn increase local membrane curvature. We showed that both cluster formation and the induced membrane curvature are enhanced by cholesterol by bridging the conical area between M2 channels.
