Interaction of influenza virus with gangliosides and liposomes containing gangliosides

Mimicking influenza virus fusion using supported lipid bilayers

Influenza virus infection is a serious public health problem in the world, and understanding the molecular mechanisms involved in viral replication is crucial. In this paper, we used a minimalist approach based on a lipid bilayer supported on mica, which we imaged by atomic force microscopy (AFM) in a physiological buffer, to analyze the different steps of influenza fusion, from the interaction of intact viruses with the supported bilayer to their complete fusion. Our results show that sialic acid recognition and priming upon acidification are sufficient for a complete fusion with the host cell membrane. After fusion, a flat and continuous membrane was observed. Because of the fragility of the viral membrane that was removed by the tip, most probably due to the disorganization of the matrix layer at acidic pH, fine structural details of ribonucleoproteins (RNP) were obtained. In addition, AFM topography of intact virus in interaction with the supported lipid bilayer confirms that hemeagglutinin and neuraminidase can form isolated clusters within the viral membrane.

Influence of calcium on lipid mixing mediated by influenza hemagglutinin

We studied the influence of calcium on lipid mixing mediated by influenza hemagglutinin (HA). Lipid mixing between HA-expressing cells and liposomes containing disialoganglioside, influenza virus receptor, was studied at 37 degrees C and neutral pH after a low-pH pulse at 4 degrees C. With DSPC/cholesterol liposomes, calcium present after raising the temperature significantly promoted lipid mixing only when it was triggered by a short low-pH application. In case of DOPC/cholesterol liposomes, calcium promotion was observed regardless of the duration of the low-pH pulse. Calcium present during a short, but not long, low-pH application to HA-expressing cells with bound DSPC/cholesterol liposomes at 4 degrees C inhibited subsequent lipid mixing. We hypothesize that calcium influences lipid mixing because it binds to a vestigial esterase domain of hemagglutinin or causes expulsion of the fusion peptide from an electronegative cavity. We suggest that calcium promotes the transition from early and reversible conformation(s) of low pH-activated HA towards an irreversible conformation that underlies both HA-mediated lipid mixing and HA inactivation.

Heterogeneity of early intermediates in cell-liposome fusion mediated by influenza hemagglutinin

To explore early intermediates in membrane fusion mediated by influenza virus hemagglutinin (HA) and their dependence on the composition of the target membrane, we studied lipid mixing between HA-expressing cells and liposomes containing phosphatidylcholine (PC) with different hydrocarbon chains. For all tested compositions, our results indicate the existence of at least two types of intermediates, which differ in their lifetimes. The composition of the target membrane affects the stability of fusion intermediates at a stage before lipid mixing. For less fusogenic distearoyl PC-containing liposomes at 4 degrees C, some of the intermediates inactivate, and no intermediates advance to lipid mixing. Fusion intermediates that formed for the more fusogenic dioleoyl PC-containing liposomes did not inactivate and even yielded partial lipid mixing at 4 degrees C. Thus, a more fusogenic target membrane effectively blocks nonproductive release of the conformational energy of HA. Even for the same liposome composition, HA forms two types of fusion intermediates, dissimilar in their stability and propensity to fuse. This diversity of fusion intermediates emphasizes the importance of local membrane composition and local protein concentration in fusion of heterogeneous biological membranes.

Targeting efficiency of galactosylated liposomes to hepatocytes in vivo: effect of lipid composition

PURPOSE

To investigate the effects of the lipid composition of galactosylated liposomes on their targeted delivery to hepatocytes.

METHODS

Several types of liposomes with a particle size of about 90 nm were prepared using distearoyl-L-phosphatidylcholine (DSPC), cholesterol (Chol) and cholesten-5-yloxy-N-(4-((1-imino-2-D- thiogalactosylethyl)amino)butyl)formamide (Gal-C4-Chol), and labeled with [3H]cholesterol hexadecyl ether. Their tissue disposition was investigated in mice following intravenous injection. The binding and internalization characteristics were also studied in HepG2 cells.

RESULTS

Compared with [H]DSPC/Chol (60:40) liposomes, [3H]DSPC/Chol/Gal-C4-Chol (60:35:5) liposomes exhibit extensive hepatic uptake. Separation of the liver cells showed that galactosylated liposomes are preferentially taken up by hepatocytes, whereas those lacking Gal-C4-Chol distribute equally to hepatocytes and nonparenchymal cells (NPC). Increasing the molar ratio of DSPC to 90% resulted in enhanced NPC uptake of both liposomes, suggesting their uptake via a mechanism other than asialoglycoprotein receptors. DSPC Chol/Gal-C4-Chol (60:35:5) and DSPC/Chol/Gal-C4-Chol (90:5:5) liposomes exhibited similar binding to the surface of HepG2 cells, but the former were taken up faster by the cells.

CONCLUSIONS

The recognition of galactosylated liposomes by the asialoglycoprotein receptors is dependent on the lipid composition. Cholesterol-rich galactosylated liposomes, exhibiting less non-specific interaction and greater receptor-mediated uptake, are better for targeting drugs to hepatocytes in vivo.

New fluorescent cholesterol analogs as membrane probes

New fluorescent cholesterol analogs, (22E, 20R)-3beta-hydroxy-23-(9-anthryl)-24-norchola-5,22-die ne (R-AV-Ch), and the 20S-isomer (S-AV-Ch) were synthesized, their spectral and membrane properties were characterized. The probes bear a 9-anthrylvinyl (AV) group instead of C22-C27 segment of the cholesterol alkyl chain. Computer simulations show that both of the probes have bulkier tail regions than cholesterol and predict some perturbation in the packing of membranes, particularly for R-AV-Ch. In monolayer experiments, the force-area behavior of the probes was compared with that of cholesterol, pure and in mixtures with palmitoyloleoyl phosphatidylcholine (POPC) and N-stearoyl sphingomyelin (SSM). The results show that pure R-AV-Ch occupies 35-40% more cross-sectional area than cholesterol at surface pressures below film collapse (0-22 mN/m); whereas S-AV-Ch occupies nearly the same molecular area as cholesterol. Isotherms of POPC or SSM mixed with 0.1 mol fraction of either probe are similar to isotherms of the corresponding mixtures of POPC or SSM with cholesterol. The probes show typical AV absorption (lambda 386, 368, 350 and 256 nm) and fluorescence (lambda 412-435 nm) spectra. Steady-state anisotropies of R-AV-Ch and S-AV-Ch in isotropic medium or liquid-crystalline bilayers are higher than the values obtained for other AV probes reflecting hindered intramolecular mobility of the fluorophore and decreased overall rotational rate of the rigid cholesterol derivatives. This suggestion is confirmed by time-resolved fluorescence experiments which show also, in accordance with monolayer data, that S-AV-Ch is better accommodated in POPC-cholesterol bilayers than R-AV-Ch. Model and natural membranes can be labeled by either injecting the probes via a water-soluble organic solvent or by co-lyophilizing probe and phospholipid prior to vesicle production. Detergent-solubilization studies involving 'raft' lipids showed that S-AV-Ch almost identically mimicked the behavior of cholesterol and that of R-AV-Ch was only slightly inferior. Overall, the data suggest that the AV-labeled cholesterol analogs mimic cholesterol behavior in membrane systems and will be useful in related studies.

New fluorescent lysolipids: preparation and selective labeling of inner liposome leaflet

Two new fluorescent lysophosphatidylcholine probes have been synthesized for use as a donor-acceptor pair in fluorescence resonance energy transfer (FRET): 9-anthrylvinyl (LAPC) as donor and 3-perylenoyl (LPPC) as acceptor. The partition coefficients between membrane and aqueous phases were 8.3 x 10(5) and 10.5 x 10(5) for LAPC and LPPC, respectively. The inner leaflets of unilamellar lipid vesicles were labeled with these probes to assess conservation of membrane sidedness after membrane fusion. After medium-sized unilamellar vesicles (MUV) were prepared with a probe in both leaflets, probe in the outer leaflet was removed by repeatedly washing with an excess of unlabeled giant unilamellar vesicles (GUV). MUV and GUV were separated by centrifugation. The probes did not flip-flop across bilayers at 25 degrees C for at least 12 h. MUV containing the ganglioside GT1b were labeled with the LAPC/LPPC pair in the inner leaflet and incubated for 30 min at neutral pH with influenza virus. Fusion was triggered by acidification to pH 5.0 and was monitored by an increase in donor fluorescence in a FRET assay. When the inner leaflets of MUV were labeled by LAPC only, its fluorescence did not change after fusion. However, the fluorescence decreased by 60% when the LAPC was removed from the outer leaflets of the fused membranes by repeated washings with GUV. We conclude that the lipids of the inner and outer leaflets of the fused MUV/virus complexes intermixed.

Target cell membrane sialic acid modulates both binding and fusion activity of influenza virus

Influenza virus binds to cell surface sialic acid receptors, and following endocytosis fuses with the endosome membrane at low pH. Whether sialic acid plays a role in the virus-cell membrane fusion step is not known. We investigated the effect of the removal of cell membrane sialic acid on the fusion activity of influenza virus (A/PR/8/34 strain) toward human T lymphocytic leukemia (CEM) cells at low pH. Fusion was monitored by fluorescence dequenching of octadecylrhodamine incorporated in the virus membrane. Removal of sialic acid by neuraminidase resulted in a drastic reduction in both viral binding and fusion. The association of the virus with neuraminidase-treated cells was enhanced at pH 5, compared to that at neutral pH, probably due to the unfolding of the hemagglutinin and the resulting increase in viral surface hydrophobicity, but the fusion capacity of the virus was reduced significantly. The results were analysed with a mass-action kinetic model which could explain and predict the kinetics of fusion. Our results indicate that binding of influenza virus to sialic acid residues on the cell surface leads to rapid and extensive fusion and partially inhibits the low pH-induced viral inactivation.

Ganglioside GD1a generates domains of high curvature in phosphatidylethanolamine liposomes as determined by solid state 31P-NMR spectroscopy

We have studied the effects of gangliosides on the polymorphic behaviour of phosphatidylethanolamines. The ganglioside GD1a promotes the formation of phases which give rise to isotropic 31P-NMR resonance lines, particularly in the temperature range of the L alpha to HII transition. In addition, higher mol fractions of ganglioside raise the L alpha to HII phase transition temperature. Our results demonstrate that small mol fractions of gangliosides can have profound effects on the molecular organization of phosphatidylethanolamines.

Transient domains induced by influenza haemagglutinin during membrane fusion

During low pH-induced fusion of influenza virus with erythrocytes we have observed differential dispersion of viral lipid and haemagglutinin (HA) into the erythrocyte membrane, and viral RNA into the erythrocyte using fluorescence video microscopy. The movement of both viral lipid and HA from virus to cell was restricted during the initial stages of fusion relative to free diffusion. This indicates the existence of relatively long-lived barriers to diffusion subsequent to fusion pore formation. Fluorescence anisotropy of phospholipid analogues incorporated into the viral membrane decreased when the pH was lowered to levels required for optimum fusion. This indicates that the restricted motion of viral membrane components was not due to rigidification of membrane lipids. The movement of HA from the fusion site was also assessed by photosensitized labelling by means of a fluorescent substrate (NBD-taurine) passing through the band 3 sialoglycoprotein (the erythrocyte anion transporter). We also examined the flow of lipid and aqueous markers during fusion of HA-expressing cells with labelled erythrocytes. During this cell-cell fusion, movement of lipid between fusing membranes begins before the fusion pore is wide enough to allow diffusion of aqueous molecules (M(r) > 500). The data indicate that HA is capable of creating domains in the membrane and controlling continuity of aqueous compartments which are bounded by such domains.

Dynamic lipid heterogeneity and receptor events

Receptor occupation by specific ligands induces changes in the dynamic domain organization of surrounding lipids. Such changes were observed by measuring changes in the fluorescence parameters of fluorescent-labelled lipids incorporated into plasma membranes of intact cells, membrane vesicles or lipoprotein particles in response to specific binding of a broad range of biologically active agents, including drugs, prostaglandins, neuropeptides, antibodies and viruses. The high sensitivity of the fluorescence response allowed us to register changes in lipid heterogeneity induced in a multitude of discrete targets by transient weak binding of a single rapidly translocating molecule. To explain these observations a non-equilibrium model of ligand-receptor interaction based on low relaxation phenomena in heterogeneous lipid matrixes is proposed.

Anthrylvinyl-labeled phospholipids as fluorescent membrane probes. The action of melittin on multilipid systems

The interaction of melittin with multicomponent lipid mixtures composed of phosphatidylcholine, sphingomyelin and phosphatidylserine or phosphatidylglycerol was investigated by measuring the intrinsic fluorescence of the peptide, steady state fluorescence anisotropy of, and Trp-fluorescence energy transfer to fluorescent analogs of the same phospholipids bearing the anthrylvinyl fluorophore in one of the aliphatic chains at various distances from the polar head group. Based on the finding that at high lipid/peptide ratio the peptide induces unequal changes in the fluorescence parameters of phospholipid probes differing structurally only in their polar head groups, it is concluded that melittin induces lipid demixing in its nearest environment. Comparison of the fluorescence energy transfer from Trp to different lipid probes indicates that the depth of penetration of melittin into the bilayer depends on the polar head group composition of the phospholipid matrix and that certain segments of the melittin chain display a specific affinity for a given lipid head group.

The interaction of prostaglandin E1 with serum lipoproteins. Possible role in cholesterol homeostasis

Prostaglandin (PG) E1 significantly stimulates the rate of cholesterol esterification in plasma. This effect could be attributed to an enhancement by PGE1 of the interlipoprotein transfer of phosphatidylcholine and cholesteryl esters, i.e., the substrate and product of lecithin-cholesterol acyltransferase (LCAT). The enhancement effect appears to be due to a rearrangement of the lipoprotein surface induced by specific interaction of PGE1 with some apolipoproteins, although the binding capacity of serum lipoproteins for PGE1 was found to be rather weak. To explain these findings, an hypothetical non-equilibrium model was put forward. The purpose of the present article is to summarize available data on the PGE1-lipoprotein interaction.