Membrane fusion through point defects in bilayers.

Fusion between bilayers of mixed egg phosphatidylcholine and soybean phosphatidylethanolamine was induced by freezing and thawing. Contact points between bilayers were observed by freeze fracture electron microscopy, and isotropic molecular motional averaging was detected by phosphorus-31 nuclear magnetic resonance under fusion conditions. A molecular model of point defect structure is proposed as an intermediate stage of fusion.

Polymorphic phase behaviour of dilinoleoylphosphatidylethanolamine and palmitoyloleoylphosphatidylcholine mixtures. Structural changes between hexagonal, cubic and bilayer phases.

The polymorphic phase behaviour of dilinoleoylphosphatidyethanolamine (DLPE) and 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) is investigated by freeze-fracture electron microscopy, X-ray diffraction and 31P-NMR. The structures at 5% or less POPC are predominantly inverted hexagonal (HII), whereas at 15% or more POPC, the structure is mostly bilayer (L), interrupted by defects (lipidic particles). A cubic phase structure is observed in the transition range between H and L phases; the cubic arrangement deteriorates at higher temperatures into an amorphous aggregate of spherical units. Both cubic and amorphous structures contribute to the isotropic 31P resonance, with no preference for PC or PE partitioning in the isotropic motion as observed by high resolution NMR. The existence of the cubic phase seems to depend cirtically on the homogeneity and the degree unsaturation of the phospholipids.

Interactions of liposome bilayers composed of 1,2-diacyl-3-succinylglycerol with protons and divalent cations.

Bilayer liposomes were prepared by using pure DOSG (1,2-dioleoyl-3-succinylglycerol) or DPSG (1,2-dipalmitoyl-3-succinylglycerol) at pH 7.4 or above. These liposomes undergo destabilization upon incubation with acid. When calcein was used as an entrapped aqueous marker, half maximal content leakage was observed between pH 5.8-6.3. Differential scanning calorimetry showed that at pH 7.4, the chain-melting temperature (Tm) of DPSG was 60.4 degrees C, and increased with decreasing pH (Tm = 57.0 degrees C and 62.7 degrees C at pH 8.9 and 6.7, respectively). Below pH 6.7, extensive phase separation occurred as the major chain melting peak split into three peaks. These three peaks coalesced into one peak below pH 5. Freeze fracture electron micrographs of DOSG liposomes at pH 4 showed the formation of non-bilayer as well as hexagonal phase structures. The effects of divalent cations, such as Ca2+ and Mg2+, on the destabilization of DASG bilayers have also been studied. Differential scanning calorimetry studies of bilayers composed of DPSG showed that both Ca2+ and Mg2+ could increase the Tm of DPSG with increasing concentrations. However, under identical conditions Mg2+ was more effective than Ca2+ in increasing the Tm of DPSG. X-ray diffraction indicated that both Ca2+ and Mg2+ could induce DPSG bilayers to undergo a complete lamellar to hexagonal phase transition. There was a size-dependency on the plasma stability of DOSG liposomes. DOSG liposomes that were smaller in size were more stable in plasma than the larger ones. After incubation with plasma, DOSG liposomes became less acid-sensitive. DOSG immunoliposomes entrapping diphtheria toxin A chain were used as a model for cytoplasmic delivery of the novel pH-sensitive liposomes. The delivery activity was comparable to that of the conventional pH-sensitive liposomes containing unsaturated phosphatidylethanolamine. Our data indicate that the mechanism of liposome destabilization involves extensive bilayer phase separation as well as the formation of non-bilayer structures.

Kinetic measurements of fusion of phosphatidylserine-containing vesicles by electron microscopy and fluorometry.

Large unilamellar vesicles (REV) containing phosphatidylserine and phosphatidylethanolamine at a ratio of 1:3 were induced to fuse by adding calcium (4 mM). The kinetics of fusion was monitored by fluorometry using terbium or dipicolinic acid-containing vesicles. The morphology and the states of vesicle aggregation and fusion were examined at approx. 2, 30, 60, 150 and 900 s after calcium addition, by rapid quenching and freeze-fracture electron microscopy. The size and the state of aggregation of vesicles are quantitated from 4000 randomly selected vesicles. The aggregation and fusion kinetics as assayed by fluorescence volume mixing is very well simulated and predicted by the mass action model. The model essentially predicts the time course of the distribution of the aggregates and the increase in size of fused particles as measured by electron microscopy, although in some cases the predicted fusion rate exceeds that by morphometric measurement. No morphological features can be defined as fusion intermediates, although bead-like and rim-like materials may be attributed to the remnants of broken diaphragms between fusion partners.

Aggregation and fusion of unilamellar vesicles by poly(ethylene glycol).

Various aspects of the interaction between the fusogen, poly(ethylene glycol) and phospholipids were examined. The aggregation and fusion of small unilamellar vesicles of egg phosphatidylcholine (PC), bovine brain phosphatidylserine (PS) and dimyristoylphosphatidylcholine (DMPC) were studied by dynamic light scattering, electron microscopy and NMR. The fusion efficiency of Dextran, glycerol, sucrose and poly(ethylene glycol) of different molecular weights were compared. Lower molecular weight poly(ethylene glycol) are less efficient with respect to both aggregation and fusion. The purity of poly(ethylene glycol) does not affect its fusion efficiency. Dehydrating agents, such as Dextran, glycerol and sucrose, do not induce fusion. 31P-NMR results revealed a restriction in the phospholipid motion by poly(ethylene glycol) greater than that by glycerol and Dextran of similar viscosity and dehydrating capacity. This may be associated with the binding of poly(ethylene glycol) to egg PC, with a binding capacity of 1 mol of poly(ethylene glycol) to 12 mol of lipid. Fusion is greatly enhanced below the phase transition for DMPC, with extensive fusion occurring below 6% poly(ethylene glycol). Fusion of PS small unilamellar vesicles depends critically on the presence of cations. Large unilamellar vesicles were found to fuse less readily than small unilamellar vesicles. The results suggest that defects in the bilayer plays an important role in membrane fusion, and the 'rigidization' of the phospholipid molecules facilitates fusion possibly through the creation of defects along domain boundaries. Vesicle aggregation caused by dehydration and surface charge neutralization is a necessary but not a sufficient condition for fusion.

On the use of N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)phosphatidylethanolamine in the study of lipid polymorphism.

The change in the fluorescence properties of dioleoyl-N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)phosphatidylethanola mine (N-NBD-PE) as an indicator of the (liquid-crystalline) bilayer-to-non-bilayer hexagonalII (HII) phase transition has been investigated. Lipid bilayer systems which are known to undergo the bilayer-to-HII phase transition on addition of Ca2+ were compared with systems which can undergo aggregation and fusion but not HII phase formation. The former included Ca2+-triggered non-bilayer transitions in cardiolipin and in phosphatidylethanolamine mixed with phosphatidylserine. The latter type of system investigated included the addition of polylysine to cardiolipin and Ca2+ to phosphatidylserine. Freeze-fracture electron microscopy was used to confirm that under the experimental conditions used, the formation of HII phase was occurring in the first type of system, but not in the second, which was stable in the bilayer state. It was found that the fluorescence intensity of N-NBD-PE (at 1 mol% of the phospholipids) increased in both types of system, irrespective of the formation of the HII phase. A dehydration at the phospholipid head group is a common feature of the formation of the HII phase, the interaction of divalent cations with phosphatidylserine and the interaction of polylysine with lipid bilayers, suggesting that this may be the feature which affects the fluorescence properties of the NBD. The finding of a fluorescence intensity increase in systems lacking HII phase involvement clearly indicates that the effect is not unique to the formation of the HII phase. Thus, while offering high sensitivity and the opportunity to follow kinetics of lipid structural changes, changes in the N-NBD-PE fluorescence properties should be interpreted with caution in the study of the bilayer-to-HII phase transition.

Curvature dependent induction of the interdigitated gel phase in DPPC vesicles.

Ethanol causes biphasic melting behavior in saturated lecithins (Rowe (1983) Biochemistry 22, 3299-3305), a consequence of the formation of the stable interdigitated phase (Simon, S.A. and McIntosh, T.J. (1984) Biochim. Biophys. Acta 773, 169-172). The membrane systems studied to date have been large vesicle systems in which the membrane surface can be assumed to be locally planar. An immediate question arises as to whether surfaces of higher curvature interdigitate. To address this question we have prepared DPPC vesicles of varying diameters which we employed to determine the limiting size at which interdigitation occurs using ethanol as the inducer. We find that with decreasing vesicle size the concentration of ethanol necessary for the onset of interdigitation increases. Small isolated vesicles, at inducing concentrations of ethanol, do not stably interdigitate but rupture and coalesce into a viscous gel comprised of interdigitated lipid sheets. As discussed elsewhere (Ahl et al. (1992) Biophys. J. 243a) these sheets can be used as precursors for producing liposomes of large size and high internal volumes useful in drug delivery or modeling applications.

Characterization of cholesterol hemisuccinate and alpha-tocopherol hemisuccinate vesicles.

Cholesterol hemisuccinate (CHS) and alpha-tocopherol hemisuccinate (alpha-THS) were found to be capable of forming liposomes of multi- or single lamellar character. Such vesicles formed spontaneously, did not require the use of organic solvents and yielded high trapping efficiencies and captured volumes. Both CHS and alpha-THS systems greatly restricted the motion of intercalated spin labelled fatty acids, yet were more osmotically responsive than similar vesicle types comprised of phosphatidylcholine. Small angle X-ray diffraction measurements were consistent with vesicles possessing extremely weak interlamellar forces. CHS vesicles were found to remain intact in vivo, yet followed a pattern of distribution dissimilar to phosphatidylcholine vesicles.

Amphotericin B-phospholipid interactions responsible for reduced mammalian cell toxicity.

When interacting with phospholipid in an aqueous environment, amphotericin B forms unusual structures of markedly reduced toxicity (Janoff et al. (1988) Proc. Natl. Acad. Sci. USA 85, 6122-6126). These structures, which appear ribbon-like by freeze-fracture electron microscopy (EM), are found exclusively at amphotericin B to lipid mole ratios of 1:3 to 1:1. At lower mole ratios they occur in combination with liposomes. Circular dichroism (CD) spectra revealed two distinct modes of lipid-amphotericin B interaction, one for liposomes and one for the ribbon-like structures. In isolated liposomes, amphotericin B which comprised 3-4 mole percent of the bulk lipid was monomeric and exhibited a hemolytic activity comparable to amphotericin B suspended in deoxycholate. Above 3-4 mole percent amphotericin B, ribbon-like structures emerged and CD spectra indicated drug-lipid complexation. Minimal inhibitory concentrations for Candida albicans of liposomal and complexed amphotericin B were comparable and could be attributed to amphotericin a release as a result of lipid breakdown within the ribbon-like material by a heat labile extracellular yeast product (lipase). Negative stain EM of the ribbon-like structures indicated that the ribbon-like appearance seen by freeze-fracture EM arises as a consequence of the cross-fracturing of what are aggregated, collapsed single lamellar, presumably interdigitated, membranes. Studies examining complexation of amphotericin B with either DMPC or DMPG demonstrated that headgroup interactions played little role in the formation of the ribbon-like structures. With these results we propose that ribbon-like structures result from phase separation of amphotericin B-phospholipid complexes within the phospholipid matrix such that amphotericin B release, and thus acute toxicity, is curtailed. Formation of amphotericin B-lipid structures such as those described here indicates a possible new role for lipid as a stabilizing matrix for drug delivery of lipophilic substances, specifically where a highly ordered packing arrangement between lipid and compound can be achieved.

Interdigitation-fusion: a new method for producing lipid vesicles of high internal volume.

Previously we demonstrated that fused phospholipid sheets can be formed from small unilamellar vesicles (SUVs) comprised of saturated symmetric chain lipids by exposing them to concentrations of ethanol sufficient to cause bilayer interdigitation (Boni et al. (1993) Biochim. Biophys. Acta 1146, 247-257). Here we report that these sheets spontaneously form large, predominately unilamellar vesicles, when exposed to temperatures above their main phase transition temperature (Tm). These vesicles, termed interdigitation-fusion vesicles (IFVs), have mean diameters between 1 and 6 microns, and, once produced, are stable both above and below the Tm of the lipid. The average captured volume of IFVs is dependent upon lipid chain length, the concentration of ethanol used to induce interdigitation-fusion, and size of the precursor liposomes. IFVs comprised of DPPC and DSPC had averaged captured volumes of 20-25 microliters/mumol lipid. IFVs produced from SUVs containing only DPPG or DPPC/DPPG mixtures had captured volumes equivalent to those made from pure DPPC SUVs indicating that charge can be introduced without consequence to the IFV process. Inclusion of cholesterol in precursor vesicles reduced IFV captured volume in a concentration dependent fashion by interfering with interdigitation. Cholesterol could be incorporated, however, into IFVs through admixture with the already formed phospholipid sheets producing far less comprise to captured volume. IFVs are useful as model systems or drug carriers, since their large internal volume allows for efficient encapsulation particularly with regard to compounds such as iodinated radiocontrast agents which otherwise interfere with vesicularization.

Interleukin-2-induced small unilamellar vesicle coalescence.

Recombinant human interleukin-2 (rhIL-2) was incorporated in liposomes for potential therapeutic applications using a novel process. In this process, rhIL-2 caused the formation of large, unique multilamellar vesicles (MLVs) from small unilamellar vesicles (SUVs) of dimyristoylphosphatidylcholine (DMPC). Vesicle coalescence occurred most rapidly at 19 degrees C, between the pre- and main phase transition temperatures of DMPC, and showed a dependence upon pH (pH <5.5), ionic strength (>50 mM) and the initial size of the unilamellar vesicles (

The effect of cholesterol in a liposomal Muc1 vaccine.

A liposomal Muc1 mucin vaccine for treatment of adenocarcinomas was formulated by incorporating a synthetic Muc1 mucin-based lipopeptide and Lipid A into a DPPC/cholesterol bilayer. Vaccination of mice with the liposomal formulation produced a peptide-specific immune response dependent on the cholesterol content. The response occurred at a threshold of 20-23 mol% cholesterol, and was optimal at cholesterol levels of > or =30 mol%. To understand this cholesterol dependency, we studied the effect of cholesterol on the liposomal bilayer and surface properties. Freeze-fracture electron microscopy showed a unique surface texture that was codependent upon cholesterol (> or =20 mol%) and lipopeptide content. Fluorescence anisotropy measurements exhibited a significant decrease in the rotational motion of 1,6-diphenyl-1,3,5-hexatriene in formulations containing >20 mol% cholesterol and only in the presence of the lipopeptide. At 20 mol% cholesterol and with lipopeptide, DSC showed a significant increase in the main phase transition of the DPPC bilayers, while Raman spectroscopy indicated a more ordered arrangement of DPPC molecules compared to control liposomes containing DPPC/cholesterol alone. Taken together, the data suggest the presence of lipopeptide-rich microdomains at and above a threshold of 20 mol% cholesterol that may play a role in the induction of a peptide-specific immunological response.

The nature of lipidic particles and their roles in polymorphic transitions.

The structural transition between bilayer (L alpha), inverted hexagonal (HII) and inverted cubic (CII) phases in mixtures of unsaturated phosphatidylethanolamines (PE) and phosphatidylcholines (PC) were investigated. Freeze fracture electron micrographs of intermediate stages of phase transitions showed that CII was a stable intermediate form between the L alpha--HII transition. The electron microscopic observation was supported by X-ray diffraction and 31P-NMR results. Detailed morphology revealed that during the L alpha--CII transition, interlamellae attachment points (conical lipidic particles) connect adjacent bilayers to form arrays of entrapped water pockets (inverted micelles). These water-containing spherical units were packed in a cubic lattice. In the CII to HII transition, these spherical units were linearly connected to form tubes. During the L alpha--HII transition, a ripple pattern was observed across the otherwise smooth lamellar. The troughs of the ripples were transformed into linear connections between adjacent bilayers, thereby converting multilayer structures into parallel tubes. No lipidic particles were involved in this type of transition. We show that there are different mechanisms involved in the L alpha, HII, CII polymorphic transitions, and that different types of 'lipidic particles' representing different molecular organizations may be involved in each case. Models of these transitions are proposed.

The nature of protein kinase C activation by physically defined phospholipid vesicles and diacylglycerols.

Protein kinase C is activated by a 1,2-sn-diacylglycerol and phospholipid at low calcium concentrations. Of the various phospholipids studied, phosphatidylserine has been shown to be the most effective one and is usually used in assaying the enzyme (Kaibuchi, K., Takai, Y., and Nishizuka, Y. (1981) J. Biol. Chem. 256, 7146-7149). It is shown here that under the conditions of the enzymatic assay, phosphatidylserine does not form typical fluid bilayer structures as seen by electron microscopy and fluorescence polarization. On the other hand, 1:4 phosphatidylserine/phosphatidylcholine bilayer vesicles can be formed which support protein kinase C activation. They have the advantage in that they are characterizable, form physiologically relevant bilayer structures, and are readily and reproducibly formed. In addition, they do not support protein kinase C activation in the absence of added diacylglycerol, a property that makes them invaluable in studying the role of diacylglycerol structure in protein kinase C activation. It is further demonstrated that the rat brain enzyme is activated by 1,2-sn-diolein but not by 2,3-sn-diolein nor 1,3-diolein, demonstrating the high specificity of the kinase toward the glycerol backbone. 1,2-rac-Dielaidin, 1,2-rac-distearin, and 1,2-sn-dipalmitin are all active, which is consistent with the idea that the specificity of protein kinase C is not directed toward the fatty acid side chain of the diacylglycerols.

Alterations in phospholipid polymorphism by polyethylene glycol.

The fusogen polyethylene glycol is shown to alter the polymorphism of dimyristoyl phosphatidylcholine, soybean phosphatidylethanolamine, bovine phosphatidylserine, egg phosphatidylcholine/cholesterol mixture, dilinoleoylphosphatidylethanolamine/palmitoyl-oleoylphosphatidy lcholine mixture, and egg lysolecithin. Suspension of these lipids in 50% polyethylene glycol (mol wt = 6000) reduces both the lamellar and the hexagonal II repeat spacings as measured by X-ray diffraction. An increase in the gel to liquid crystalline and bilayer to hexagonal transition temperatures are observed by freeze-fracture, X-ray diffraction, differential scanning calorimetry and 31P NMR. Freeze-fracture electron micrographs revealed different bilayer defects depending on the physical states of the lipid. Lipidic particles in mixtures containing unsaturated phosphatidylethanolamine is eliminated. Some of the influences of polyethylene glycol on lipids may be explained by its dehydrating effect. However, other nonfusogenic dehydrating agents failed to produce similar results. These findings are consistent with the proposal that close bilayer contact and the formation of bilayer defects are associated with the fusogenic properties of polyethylene glycol.

Lipid-polyethylene glycol interactions: I. Induction of fusion between liposomes.

Fusion between unilamellar vesicles of both egg phosphatidylcholine and bovine phosphatidylserine was induced by polyethylene glycol. Aggregation and fusion events were monitored by electron microscopy and turbidity measurements. The threshold concentration of polyethylene glycol for aggregation and fusion is found to be independent of lipid concentration. Typically, aggregation of phosphatidylcholine vesicles starts at 2.5% (wt/wt) polyethylene glycol, but fusion is not significant until the polyethylene glycol concentration reaches 35%. Multilamellar vesicles were formed as a result of fusion.

Lipid-polyethylene glycol interactions: II. Formation of defects in bilayers.

The kinetic properties of L-leucine transport across the human red blood cell membrane was analyzed according to the simple pore and carrier theory of Lieb and Stein (Biochim. Biophys. Acta, 1974, 373: 165-177 and 178-196) at 25 degrees C, pH 7.4. Several methods were used in order to obtain a thorough kinetic description of L-leucine transport. A rejection of the simple pore model was suggested from the result of zero-trans influx and zero-trans and equilibrium-exchange efflux experiments. Several predictions from the simple carrier model, based on the requirement of consistency among different kinetic parameters, were tested in infinite experiments, i.e. experiments performed at a high concentration of substrate at one of the faces of the membrane. The simple pore model was rejected, but no crucial evidence against a simple carrier model, which displays symmetric properties at 25 degrees C, was found in the concentration range considered (0.002-68 mM). The relative magnitudes of the rate constants of the translocation process are discussed, and it is concluded (a) that both the dissociation and translocation of carrier-complex is faster than the translocation of the empty carrier, (b) that no translocation step is rate determining, and (c) that the carrier-complex is equally distributed across the membrane at equilibrium. The present work provides a unique example of a carrier-mediated transport mechanism which displays symmetric properties. L-leucine transport in red blood cells may be a convenient system for studying molecular mechanisms of facilitated transport.