Influenza A virus enters host cells when the viral envelope fuses with the late endosomal membrane at low pH, a process driven by the haemagglutinin (HA) glycoprotein through hemifusion intermediates in which only the outer lipid leaflets of the two membranes have merged. Cryo-electron tomography of HA-bearing virus-like particles fusing with liposomes (Chlanda et al., 2016) identified two distinct hemifusion structures — a classic hemifusion diaphragm and a ruptured 'lipidic junction' — and showed, by continuum membrane-elasticity analysis, that target-membrane cholesterol acts as a pathway switch through its negative spontaneous curvature: above roughly 31 mol% cholesterol the hemifusion stalk becomes energetically favored, shifting fusion from a rupture–insertion route to a stalk-expansion route that the authors propose is the productive route in infection. The unresolved problem stated by that work is whether this in vitro switch describes fusion in cells: the cholesterol concentration and the complement of curvature-active lipids (such as the late-endosome-enriched BMP/LBPA) of the membrane at the viral fusion site are not known precisely enough to map the in vitro thresholds onto in vivo conditions, and the fate of the hemifusion products — productive fusion pore versus dead-end or leak-prone intermediate — has not been determined in cellular infection. The problem is to establish quantitatively how the physical lipid composition of the late endosomal membrane determines which HA-induced hemifusion pathway executes productive entry, and whether the composition–pathway–outcome mapping predicted by membrane mechanics holds in cells. Progress on each ingredient — single-virus cholesterol assays, in situ cryo-electron tomography of hemifusion in cells, and continuum-theory refinements — has not yet closed this mapping; the problem remains open.

Background

Influenza A virus is an enveloped virus: its genome is enclosed in a membrane derived from the host plasma membrane, so infection requires merging (fusion) of the viral envelope with a cellular membrane. After endocytosis, acidification of endosomes triggers a large, irreversible conformational change in haemagglutinin (HA), the trimeric viral surface glycoprotein, whose spring-like refolding pulls the viral and endosomal membranes together. Bilayer fusion is believed to proceed through hemifusion: a stage in which only the proximal (outer) leaflets of the two membranes have merged into a stalk — an hourglass-shaped lipidic connection — while the distal leaflets still separate the two aqueous compartments. A stalk can expand into a hemifusion diaphragm (HD), a flat bilayer patch made of the two distal leaflets, whose rupture opens a fusion pore. That hemifusion is a true intermediate of HA-mediated fusion was established by Kemble, Danieli and White (Cell, 1994), who showed that HA anchored to the viral membrane only through a lipid tail produces lipid mixing between membranes without mixing of aqueous contents, i.e. fusion arrested at hemifusion. The energy landscape of the stalk was made quantitative by the 'stalk model' of Kozlovsky and Kozlov (Biophysical Journal, 2002), and membrane fusion mechanics is reviewed by Chernomordik and Kozlov (Nature Structural & Molecular Biology, 2008). Chlanda et al. (Nature Microbiology, 2016) studied influenza virus-like particles (VLPs — non-infectious particles assembled from HA, neuraminidase, the M1 matrix protein and the M2 ion channel) fusing with liposomes of defined composition at low pH, using Volta phase plate cryo-electron tomography (VPP–cET): tomography of vitrified (flash-frozen) specimens in which a Volta phase plate provides in-focus phase contrast, improving the signal-to-noise ratio for thick biological samples. They resolved two hemifusion structures: the classical HD, and a novel 'lipidic junction' — a Y-shaped three-bilayer junction arising when the target liposome ruptures and its membrane edges are stabilized by HA — and proposed two independent merger pathways: a rupture–insertion pathway leading to the lipidic junction, and a hemifusion-stalk pathway leading to a fusion pore. Neither the rupture frequency nor the HD diameter was affected by the G1S mutation (a glycine-to-serine substitution at position 1 of the HA fusion peptide, the membrane-inserted N-terminal segment of HA2, which classically arrests fusion at hemifusion), but both decreased when liposomal cholesterol approached physiological levels. A continuum analysis — minimizing the Helfrich elastic energy of the lipid monolayers, which depends on bending rigidity, tilt modulus, area modulus and the spontaneous curvature k0 (the intrinsic curvature a monolayer adopts when unconstrained; cholesterol and other cone-shaped lipids make k0 more negative) — reproduced the measured HD diameter distributions and showed that the energy of forming a hemifusion stalk falls with cholesterol and becomes favorable at about 31 mol%, identifying cholesterol's negative spontaneous curvature as the pathway switch. The authors argue the stalk pathway is the one relevant to infection. Whether this in vitro switch describes the in-cell situation is, however, open, because the target in infection is the late endosomal membrane, whose physical composition is poorly constrained. Late endosomes (the acidic, multivesicular pre-lysosomal compartment where influenza fusion occurs) are enriched in bis(monoacylglycero)phosphate (BMP, historically called lysobisphosphatidic acid or LBPA) — a cone-shaped, negatively charged phospholipid associated with intraluminal vesicles and cholesterol transport (Kobayashi et al., Nature Cell Biology, 1999) — and their cholesterol content was estimated only roughly (~30 mol%, by analogy with the endosomal recycling compartment). BMP, like cholesterol, has negative spontaneous curvature and could substitute for it in switching the pathway. The source paper thus left a coherent cluster of unanswered questions: the precise cholesterol concentration and curvature-active lipid complement at the actual fusion site; the fate of the rupture–insertion (lipidic junction) products in cells — productive fusion, dead end, or leak; the compensatory mechanisms (lipid or protein composition, the M1 matrix layer) that may modulate pathway probabilities across host cell types; and the extension of the continuum description from equilibrium energetics with illustrative parameter values to physiologically accurate compositions and to pathway kinetics.

Problem Statement

For influenza A virus entry, determine quantitatively how the physical lipid composition of the late endosomal target membrane selects the hemifusion pathway executed by haemagglutinin during infection of cells. This comprises three jointly linked parts serving one objective. First, establish the cholesterol concentration and the complement of curvature-active lipids (including BMP/LBPA and other late-endosome-enriched species) of the late endosomal membrane at the site of HA-mediated fusion, with accuracy sufficient for comparison with the in vitro pathway-switch thresholds (stalk energy favorable above ~31 mol% cholesterol; leaky-versus-non-leaky poration switching near a target-membrane spontaneous curvature of -0.2 nm^-1, below which fusion is predominantly non-leaky and above which it is predominantly leaky (Haldar et al., 2019)). Second, determine which hemifusion structures form and their fate during productive cellular infection — in particular, whether the rupture-insertion (lipidic-junction) pathway observed in vitro proceeds to complete fusion or is a dead-end or leak-prone intermediate — together with the relative probabilities of the two pathways under native conditions and how compensatory factors (lipid or protein composition, the M1 matrix layer) modulate them across host cell types. Third, assess whether the measured composition-pathway-outcome mapping is consistent with membrane-mechanical modeling (continuum elasticity or equivalent) using physiologically accurate, composition-dependent parameters, extended from equilibrium energetics to the kinetics and transition states of the pathways. The overarching completion boundary, which all three parts must jointly address, is a validated account of how target-membrane physical properties select the productive fusion pathway during influenza infection of cells, replacing the in vitro liposome-based switch with one established for the native late endosomal membrane; 'validated' here means the account substantiates each of the three parts against the named measurements, thresholds and models — reporting the best available in-cell values or estimates with their uncertainties, and explicitly flagging any part that no existing measurement constrains.

References

  1. Chlanda P, Mekhedov E, Waters H, Schwartz CL, Fischer ER, Ryham RJ, Cohen FS, Blank PS, Zimmerberg J. The hemifusion structure induced by influenza virus haemagglutinin is determined by physical properties of the target membranes. Nature Microbiology 1:16050 (2016). DOI 10.1038/nmicrobiol.2016.50, https://doi.org/10.1038/nmicrobiol.2016.50
  2. Kemble GW, Danieli T, White JM. Lipid-anchored influenza hemagglutinin promotes hemifusion, not complete fusion. Cell 76:383-391 (1994). DOI 10.1016/0092-8674(94)90344-1, https://doi.org/10.1016/0092-8674(94)90344-1
  3. Kozlovsky Y, Kozlov MM. Stalk model of membrane fusion: solution of energy crisis. Biophysical Journal 82:882-895 (2002). DOI 10.1016/S0006-3495(02)75450-7, https://doi.org/10.1016/S0006-3495(02)75450-7
  4. Chernomordik LV, Kozlov MM. Mechanics of membrane fusion. Nature Structural & Molecular Biology 15:675-683 (2008). DOI 10.1038/nsmb.1455, https://doi.org/10.1038/nsmb.1455
  5. Kobayashi T, Beuchat M-H, Lindsay M, Frias S, Palmiter RD, Sakuraba H, Parton RG, Gruenberg J. Late endosomal membranes rich in lysobisphosphatidic acid regulate cholesterol transport. Nature Cell Biology 1:113-118 (1999). DOI 10.1038/10084, https://doi.org/10.1038/10084
  6. Zawada KE, Wrona D, Rawle RJ, Kasson PM. Influenza viral membrane fusion is sensitive to sterol concentration but surprisingly robust to sterol chemical identity. Scientific Reports 6:29842 (2016). DOI 10.1038/srep29842, https://doi.org/10.1038/srep29842
  7. Haldar S, Mekhedov E, McCormick C, Blank PS, Zimmerberg J. Lipid-dependence of target membrane stability during influenza viral fusion. Journal of Cell Science 132:jcs218321 (2019). DOI 10.1242/jcs.218321, https://doi.org/10.1242/jcs.218321
  8. Liu K, Boxer SG. Target membrane cholesterol modulates single influenza virus membrane fusion efficiency but not rate. Biophysical Journal 118:2426-2433 (2020). DOI 10.1016/j.bpj.2020.03.021, https://doi.org/10.1016/j.bpj.2020.03.021
  9. Liu K, Boxer SG. Single-virus content-mixing assay reveals cholesterol-enhanced influenza membrane fusion efficiency. Biophysical Journal 120:4832-4841 (2021). DOI 10.1016/j.bpj.2021.09.023, https://doi.org/10.1016/j.bpj.2021.09.023
  10. Rice A, Haldar S, Wang E, Blank PS, Akimov SA, Galimzyanov TR, Pastor RW, Zimmerberg J. Planar aggregation of the influenza viral fusion peptide alters membrane structure and hydration, promoting poration. Nature Communications 13:7336 (2022). DOI 10.1038/s41467-022-34576-z, https://doi.org/10.1038/s41467-022-34576-z
  11. Pabis A, Rawle RJ, Kasson PM. Influenza hemagglutinin drives viral entry via two sequential intramembrane mechanisms. Proceedings of the National Academy of Sciences USA 117:7200-7207 (2020). DOI 10.1073/pnas.1914188117, https://doi.org/10.1073/pnas.1914188117
  12. Klein S, Golani G, Lolicato F, Lahr C, Beyer D, Herrmann A, Wachsmuth-Melm M, Reddmann N, Brecht R, Hosseinzadeh M, Kolovou A, Makroczyova J, Peterl S, Schorb M, Schwab Y, Brügger B, Nickel W, Schwarz US, Chlanda P. IFITM3 blocks influenza virus entry by sorting lipids and stabilizing hemifusion. Cell Host & Microbe 31:616-633.e20 (2023). DOI 10.1016/j.chom.2023.03.005, https://doi.org/10.1016/j.chom.2023.03.005
  13. Mannsverk S, Villamil Giraldo A, Kasson PM. Influenza viral infection at the plasma membrane is restricted by lipid composition. Journal of Virology 99:e01105-25 (2025). DOI 10.1128/jvi.01105-25, https://doi.org/10.1128/jvi.01105-25
  14. Tavakoli A, Hu S, Ebrahim S, Kachar B. Hemifusomes and interacting proteolipid nanodroplets mediate multi-vesicular body formation. Nature Communications 16:4609 (2025). DOI 10.1038/s41467-025-59887-9, https://doi.org/10.1038/s41467-025-59887-9
  15. Akıl C, Xu J, Shen J, Zhang P. Unveiling the structural spectrum of SARS-CoV-2 fusion by in situ cryo-ET. Nature Communications 16:5150 (2025). DOI 10.1038/s41467-025-60406-z, https://doi.org/10.1038/s41467-025-60406-z
  16. Song Y, Yao H, Wu N, Xu J, Zhang Z, Peng C, Li S, Kong W, Chen Y, Zhu M, Wang J, Shi D, Zhao C, Lu X, Echavarría Galindo M, Li S. In situ architecture and membrane fusion of SARS-CoV-2 Delta variant. Proceedings of the National Academy of Sciences USA 120:e2213332120 (2023). DOI 10.1073/pnas.2213332120, https://doi.org/10.1073/pnas.2213332120
  17. Molotkovsky RJ, Galimzyanov TR, Jiménez-Munguía I, Pavlov KE, Batishchev OV, Akimov SA. Switching between successful and dead-end intermediates in membrane fusion. International Journal of Molecular Sciences 18:2598 (2017). DOI 10.3390/ijms18122598, https://doi.org/10.3390/ijms18122598
  18. Gardner J, Abrams CF. Rate of hemifusion diaphragm dissipation and ability to form three-junction bound HD determined by lipid composition. The Journal of Chemical Physics 147:134903 (2017). DOI 10.1063/1.4994320, https://doi.org/10.1063/1.4994320
  19. Warner JM, An D, Stratton BS, O'Shaughnessy B. A hemifused complex is the hub in a network of pathways to membrane fusion. Biophysical Journal 122:374-385 (2023). DOI 10.1016/j.bpj.2022.12.003, https://doi.org/10.1016/j.bpj.2022.12.003
  20. Su R, Wang S, McDargh ZA, O'Shaughnessy B. Three membrane fusion pore families determine the pathway to pore dilation. Biophysical Journal 122:3986-3998 (2023). DOI 10.1016/j.bpj.2023.08.021, https://doi.org/10.1016/j.bpj.2023.08.021
  21. Golani G, Schwarz US. High curvature promotes fusion of lipid membranes: predictions from continuum elastic theory. Biophysical Journal 122:1868-1882 (2023). DOI 10.1016/j.bpj.2023.04.018, https://doi.org/10.1016/j.bpj.2023.04.018
  22. Boonstra S, Blijleven JS, Roos WH, Onck PR, van der Giessen E, van Oijen AM. Hemagglutinin-mediated membrane fusion: a biophysical perspective. Annual Review of Biophysics 47:153-173 (2018). DOI 10.1146/annurev-biophys-070317-033018, https://doi.org/10.1146/annurev-biophys-070317-033018
  23. Kephart S, Hom N, Lee KK. Visualizing intermediate stages of viral membrane fusion by cryo-electron tomography. Trends in Biochemical Sciences 49:916-931 (2024). DOI 10.1016/j.tibs.2024.06.012, https://doi.org/10.1016/j.tibs.2024.06.012
  24. Lira RB, Hammond J, Cavalcanti R, Rous M, Riske KA, Roos WH. The underlying mechanical properties of membranes tune their ability to fuse. Journal of Biological Chemistry 299:105430 (2023). DOI 10.1016/j.jbc.2023.105430, https://doi.org/10.1016/j.jbc.2023.105430
  25. Peterl S, Lahr CM, Schneider CN, Meyer J, Podlipensky X, Lechner V, Villiou M, Eis L, Klein S, Funaya C, Cavalcanti-Adam EA, Graw F, Selhuber-Unkel C, Rohr K, Chlanda P. Morphology-dependent entry kinetics and spread of influenza A virus. The EMBO Journal 44:3959-3982 (2025). DOI 10.1038/s44318-025-00481-6, https://doi.org/10.1038/s44318-025-00481-6