A supramolecular system mimicking the infection process of an enveloped virus through membrane fusion

Membrane fusion is an essential step for the entry of enveloped viruses, such as human immunodeficiency virus and influenza virus, into the host cell, often triggered by the binding of membrane proteins on the viral envelope to host cell membrane. Recently, external stimuli was shown to trigger membrane fusion in an artificial system. Direct observation of artificial membrane fusion using a giant unilamellar vesicle (GUV), which is similar in size to a cell, is useful as a biological model system. However, there are no model systems for studying membrane fusion of enveloped viruses with host cells. Here, we report a supramolecular model system for viral entry into a GUV or cell through membrane fusion. The system was constructed by complexing a cationic lipid bilayer on an anionic artificial viral capsid, self-assembled from viral β-annulus peptides. We demonstrate that the cationic enveloped artificial viral capsid electrostatically interacts with the anionic GUV or cell, and the capsid enters the GUV or cell through membrane fusion. The model system established in this study will be important for analyzing membrane fusion during infection of a natural virus.


A supramolecular system mimicking the infection process of an enveloped virus through membrane fusion
Hiroto Furukawa 1 , Yuuna Kimura 1 , Hiroshi Inaba 1,2 & Kazunori Matsuura 1,2* Membrane fusion is an essential step for the entry of enveloped viruses, such as human immunodeficiency virus and influenza virus, into the host cell, often triggered by the binding of membrane proteins on the viral envelope to host cell membrane.Recently, external stimuli was shown to trigger membrane fusion in an artificial system.Direct observation of artificial membrane fusion using a giant unilamellar vesicle (GUV), which is similar in size to a cell, is useful as a biological model system.However, there are no model systems for studying membrane fusion of enveloped viruses with host cells.Here, we report a supramolecular model system for viral entry into a GUV or cell through membrane fusion.The system was constructed by complexing a cationic lipid bilayer on an anionic artificial viral capsid, self-assembled from viral β-annulus peptides.We demonstrate that the cationic enveloped artificial viral capsid electrostatically interacts with the anionic GUV or cell, and the capsid enters the GUV or cell through membrane fusion.The model system established in this study will be important for analyzing membrane fusion during infection of a natural virus.
Membrane fusion is fundamental to biological processes, such as intracellular compartmentalization, cell growth, hormone secretion, neurotransmission, and oocyte fertilization 1,2 .Membrane fusion proceeds with the gradual merging of two opposite membranes into a continuous lipid bilayer, causing the mixing of lipids and contents 3 .Membrane fusion is an essential step during infection of host cells by enveloped viruses, and is often triggered by the binding of viral membrane proteins to host cell membrane 4 .For example, glycoprotein gp41 on the envelope of human immunodeficiency virus (HIV) promotes membrane fusion for entry into the host cell 5,6 .Influenza virus is taken up through endocytosis, which is triggered by the binding of hemagglutinin on the envelope membrane to sialyl oligosaccharides on host cell surface; membrane fusion proceeds by partial cleavage in the trans-Golgi network, and the capsid enters the cell 7 .Thus, membrane fusion induced by molecular recognition between membrane proteins on the envelope and the host cell membrane is important for enveloped virus infection.
Membrane fusion between cells can be triggered by external stimuli, such as laser irradiation 8 and nanoheating 9 .Artificial membrane fusion systems have been developed between liposomes triggered by reconstituted protein 10,11 , fusogenic peptide [12][13][14] , macromolecule 15,16 , and fusogenic DNA 17 .Direct observation of artificial membrane fusion using a giant unilamellar vesicle (GUV), which is similar in size to a cell, is useful as a model system to examine the physicochemical mechanism of membrane fusion.Riske et al. analyzed electrostatic interactions between a cationic large unilamellar vesicle (LUV, ~ 140 nm) containing 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP) and an anionic GUV (> 20 μm) containing 1-palmitoyl-2-oleoylsn-glycero-3-phospho-(1′-rac-glycerol) (sodium salt) (POPG), and found that the more anionic lipid POPG (in GUV) induced faster and more efficient membrane fusion 18 .Other artificial membrane fusion processes using GUVs have been reported to be induced by an electrical pulse 19,20 , external stimulus-responsive lipids 21,22 , fusogenic peptides 23,24 , fusogenic DNA 25 , and optically heated gold nanoparticles 26 .However, to our knowledge, there are no model systems for studying membrane fusion of enveloped virus with host cell.Membrane fusion of an enveloped viral model into a GUV or cell is a promising artificial system to understand the infection mechanism of enveloped viruses.

Interaction between enveloped artificial viral capsids and GUVs
We evaluated how multifluorescence-labeled enveloped artificial viral capsids interact with GUVs by fluorescence imaging.GUVs comprising only DOPC or 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) and mixed GUVs containing 40 mol% anionic lipid 1,2-dioleoyl-sn-glycero-3-phospho-rac-(1-glycerol) (DOPG) of total lipids (40% DOPG/DOPC or 40% DOPG/POPC) were prepared by the hydration method.Since the ratio of anionic lipids in typical mammalian cell membranes is approximately 30-50% 65 , we prepared GUVs containing 40% DOPG/DOPC or DOPG/POPC, which are commonly used for direct observation of membrane fusion 18,[21][22][23][24][25][26] .GUV suspensions in 10 mM phosphate buffer (pH 7.0) were incubated with a solution of TMR/NBD-labeled enveloped capsid in the same buffer at a volume ratio of 1:1 for 1 h and observed by CLSM (Fig. 3A).TMRderived fluorescence was strong inside the GUV containing anionic DOPG and weak inside the GUV comprising DOPC or POPC (Fig. 3B).Interestingly, the GUV containing anionic DOPG showed NBD-derived fluorescence on the GUV membrane and increased TMR-derived fluorescence inside the GUV (Fig. 3C).These results indicate that because the cationic enveloped capsid fused with the anionic GUV, the envelope lipids on the GUV membrane internalized the capsid only.By contrast, no increase in TMR-derived fluorescence was observed inside DOPC or POPC GUVs.It is probable that the membrane fusion induced the electrostatic relaxation between the anionic capsid and the cationic envelope within the enveloped capsid.As the result, a part of the anionic capsid should be exposed inside the GUV.We envisage that electrostatic repulsion between the exposed anionic surface of capsid and the anionic lipids on the GUV membrane might promote internalization of the capsid inside the GUV.And the lipids derived from the envelope should remain in the GUV membrane due to the significant hydrophobic interaction in the membrane.
GUVs without green or red internalized fluorescence, and that with green fluorescence in the membrane and no internal fluorescence might be caused by diversity in the interaction between the enveloped capsid and each GUV due to differences in each GUV particle size and membrane composition (Fig. 3B, Figs. S3-S6).For GUVs without green fluorescence in the membrane but with red fluorescence inside, the NBD fluorescence might be quenched due to fluorescence resonance energy transfer (FRET) between the TMR labeled on capsid inside the GUV and the NBD on the membrane.The fluorescence intensity ratio of TMR inside/outside (I TMR [in]/I TMR [out]) of each GUV significantly increased in the presence of anionic DOPG (Fig. 3D, Figs. S3-S6).For GUVs with anionic DOPG, fluorescence intensity ratio of TMR inside/outside increased with decreasing GUV surface area, whereas for GUVs without anionic DOPG, the ratio did not increase significantly with decreasing GUV surface area.The correlation between GUV surface area and inside/outside fluorescence intensity ratio were roughly approximated by a power law equation (Fig. S7).These results suggest that anionic GUV membranes possessing higher curvature promote membrane fusion with the cationic enveloped capsid.
To directly confirm the entry of envelope components into the GUV by membrane fusion, we observed CLSM images of fluorescence-labeled GUVs before and after the membrane fusion based on lipid-mixing assay 66 (Fig. 4A).We prepared NBD/rhodamine-labeled 40% DOPG/DOPC GUVs, in which the NBD-derived fluorescence quenched by FRET from NBD to rhodamine (Fig. 4B).When the enveloped capsid was added to the NBD/rhodamine-labeled 40% DOPG/DOPC GUV, the quenched NBD-derived fluorescence was recovered by decrease in FRET efficiency through the entry of envelope-derived lipids into the GUV membrane (Fig. 4C).In contrast, when the enveloped capsid was added to NBD/rhodamine-labeled 40% DOPC GUV, fluorescence recovery was hardly observed (Fig. 4D).These results strongly support entry of the cationic enveloped capsids into the anionic GUVs by membrane fusion.In addition, we evaluated the interaction between the cationic enveloped capsid and a GUV containing the anionic lipid ganglioside monosialate 3 (GM3), which is present on cell membrane surface.When the TMR/ NBD-labeled enveloped capsid was added to a GUV containing 40% GM3, TMR-derived fluorescence inside the GUV was qualitatively comparable to that of the DOPG-containing GUV (Fig. 5A,B).Inside/outside fluorescence intensity ratio was significantly higher for GUVs with GM3 than GUVs without GM3 (Fig. 5C, Figs. S8, S9).These results suggest that enveloped capsids interact with anionic glycolipids on the plasma membrane and get internalized.

Interaction between enveloped artificial viral capsids and cells
We evaluated whether enveloped capsids are taken up by cells through membrane fusion (Fig. 6A).Human hepatocellular carcinoma-derived HepG2 cells have an anionic plasma membrane surface containing the anionic lipids DOPS, DOPG, and DOPI, which comprise approximately 40% of the total phospholipids 67 .TMR/NBDlabeled enveloped capsids were incubated with HepG2 cells for 5 min, 1 h, and 3 h, and intracellular fluorescence was observed by CLSM (Fig. 6B,C).After 5 min, no fluorescence was observed inside the cell or on the cell surface, whereas after 1 h, yellow fluorescent dots colocalizing TMR and NBD were observed on the cell surface.Remarkably, after 3 h, strong TMR fluorescence was observed throughout the cell.The fluorescence intensity profile of TMR and NBD after 5 min showed no significant increase in fluorescence intensity on the cell surface or inside the cell.After 1 h, TMR-and NBD-derived fluorescence intensity increased on the cell surface.After 3 h, NBD-derived fluorescence was visible on the cell surface, and TMR-derived fluorescence intensity dramatically increased inside the cell.By contrast, when anionic capsids or enveloped capsids coated with DOPC (without DOTAP) were added to HepG2 cells, only weak TMR fluorescence was observed inside the cell (Fig. S10).Thus, the enveloped capsids entered the cells through membrane fusion mediated by electrostatic interactions.To rule out cellular uptake of the enveloped capsids by endocytosis, we analyzed the effect of endocytosis inhibitors on cellular uptake (Fig. 7A).In this experiment, TMR-labeled enveloped capsids with diameter of 121 ± 58 nm (Fig. S11) were used instead of the TMR/NBD-labeled enveloped capsids since it does not have to track membrane components.HepG2 cells were pretreated with ethylisopropyl amiloride 3-amino-N-(aminoiminomethyl)-6-chloro-5-[ethyl(1-methylethyl)amino]-2-pyrazinecarboxamide) (EIPA), inhibitor of micropinocytosis; genistein, inhibitor of caveolae-mediated endocytosis; and Pitstop 2, inhibitor of clathrinmediated endocytosis 68 , incubated with TMR-labeled enveloped capsids for 3 h, and analyzed using CLSM.Strong TMR fluorescence was observed on the surface and inside of genistein-and Pitstop 2-treated cells, as well as untreated cells, whereas fluorescence intensity was significantly decreased in EIPA-treated cells (Fig. 7B).After addition of the inhibitors, TMR-derived fluorescence was observed mainly on the cell membrane surface.Probably, the inhibitors might also inhibited membrane fusion.Compared to untreated cells, genistein-and Pitstop 2-treated cells showed minimum decrease and EIPA-treated cells showed significant decrease (to 69.3%) in fluorescence intensity on the cell surface and inside the cell (Fig. 7C).These results indicate that enveloped capsids are taken up into cells through not only membrane fusion, but also micropinocytosis.Because these pathways are dependent on interactions with the cell membrane, the interaction between a cationic lipid-containing enveloped capsid and an anionic cell membrane may trigger uptake.

Conclusion
In conclusion, we have developed a supramolecular model system that mimics the entry of natural enveloped viruses into the host cell.Electrostatic interactions between cationic enveloped artificial viral capsids and GUVs containing anionic lipids or glycolipids induced membrane fusion to internalize the capsid only.Although GUVs were prepared by simple and convenient hydration method in this study, in the future, we would like to prepare more homogeneous GUVs by other methods (droplet transfer, electroformation, microfluidic device methods) to clarify the correlation between GUV structure and membrane fusion properties of the enveloped virus capsids.In addition, since anionic lipid composition may affect membrane fusion, we plan to evaluate in detail the correlation between anionic lipid ratio and membrane fusion activity in the future.
HepG2 cells internalized the cationic enveloped artificial viral capsids through membrane fusion and micropinocytosis.These results indicate that membrane fusion could be induced by interactions between the enveloped artificial viral capsid and cell membrane, without external stimuli.This study could provide insights into the infection mechanism of natural enveloped viruses into the host cell.We aim to construct an artificial infection model for a GUV or host cell of enveloped artificial viral replica equipped with membrane proteins, such as influenza virus-derived hemagglutinin and SARS-CoV-2-derived spike protein.Such artificial virus models can be used as a cell-specific carrier in drug delivery systems.

General
Reversed-phase HPLC was performed at ambient temperature using a Shimadzu LC-6AD Liquid Chromatography system equipped with a UV/vis detector (220 and 551 nm, Shimadzu SPD-10AVvp) and Inertsil WP300 C18 (GL Science) column (250 × 4.6 mm and 250 × 20 mm).MALDI-TOF MS spectra were obtained using the ultrafleXtreme instrument (Bruker Daltonics) in linear/positive mode with α-cyano-4-hydroxy cinnamic acid (α-CHCA) as matrix.Deionized water of high resistivity (> 18 MΩ cm) was purified using Millipore Purification System (Milli-Q water) and used as solvent for the peptides.All reagents were obtained from commercial sources and used without further purification.

DLS and ζ-potential
DLS values of samples were measured at 25 °C using Zetasizer Nano ZS (MALVERN) with an incident He-Ne laser (633 nm) and ZEN2112-Low volume glass cuvette cell.During measurements, count rates (sample scattering intensities) were provided.Correlation time for scattered light intensity G(t) was measured several times, and the averaged results were fitted to Eq. ( 1), where B is the baseline, A is the amplitude, q is the scattering vector, t is the delay time, and D is the diffusion coefficient: The hydrodynamic radius (R H ) of the scattering peptides was calculated using Stokes-Einstein equation (Eq.( 2)), where η is solvent viscosity, k B is Boltzmann's constant, and T is absolute temperature.
Zeta potentials of each sample were measured at 25 °C using a Zetasizer Nano ZS (MALVERN) with a DTS1070 clear disposable zeta cell.

TEM
Cu-grids (thin carbon film TEM grids, ALLIANCE Biosystems) were hydrophilized by plasma treatment using JEOL HD Treatment apparatus for 40 s (60 Hz, 500 VA) at 25 °C.Aliquots (5 μL) of aqueous sample solutions were applied to the hydrophilized Cu-grids for 1 min.Subsequently, the TEM grids were instilled in the staining solution (5 μL) EM stainer (Nisshin-EM) for 10 min.The sample-loaded Cu-grids were dried in vacuo and observed by TEM (JEOL JEM 1400 Plus) at an accelerating voltage of 80 kV.