Formation of unilamellar vesicles by repetitive freeze-thaw cycles: characterization by electron microscopy and 31P-nuclear magnetic resonance

 It has been reported that repetitive freeze-thaw cycles of aqueous suspensions of dioleoylphosphatidylcholine form vesicles with a diameter smaller than 200 nm. We have applied the same treatment to a series of phospholipid suspensions with particular emphasis on dioleoylphosphatidylcholine/dioleoylphosphatidic acid (DOPC/DOPA) mixtures. Freeze-fracture electron microscopy revealed that these unsaturated lipids form unilamellar vesicles after 10 cycles of freeze-thawing. Both electron microscopy and broad-band ³¹P NMR spectra indicated a disparity of the vesicle sizes with a highest frequency for small unilamellar vesicles (diameters ≤30 nm) and a population of larger vesicles with a frequency decreasing exponentially as the diameter increases. From ³¹P NMR investigations we inferred that the average diameter of DOPC/DOPA vesicles calculated on the basis of an exponential size distribution was of the order of 100 nm after 10 freeze-thaw cycles and only 60 nm after 50 cycles. Fragmentation by repeated freeze-thawing does not have the same efficiency for all lipid mixtures. As found already by others, fragmentation into small vesicles requires the presence of salt and does not take place in pure water. Repetitive freeze-thawing is also efficient to fragment large unilamellar vesicles obtained by filtration. If applied to sonicated DOPC vesicles, freeze-thawing treatment causes fusion of sonicated unilamellar vesicles into larger vesicles only in pure water. These experiments show the usefulness of NMR as a complementary technique to electron microscopy for size determination of lipid vesicles. The applicability of the freeze-thaw technique to different lipid mixtures confirms that this procedure is a simple way to obtain unilamellar vesicles. Received: 2 September 1999 / Revised version: 27 February 2000 / Accepted: 27 February 2000     

Phase behaviour of mixtures of lipid X with phosphatidylcholine and phosphatidylethanolamine

The effect of increasing concentrations of lipid X (2,3-bis(3-hydroxymyristoyl)-alpha-D-glucosamine 1-phosphate) on the phase behaviour of EPC (egg phosphatidylcholine) and EPE (egg phosphatidylethanolamine) is studied at a pH greater than or equal to 7 where lipid X carries one to two negative charges. Small amounts of lipid X (molar ratio approximately 0.01) induce continuous swelling of EPC and EPE bilayers and consequently the formation of large unilamellar vesicles in excess water. In many respects, the effect of lipid X on EPC and EPE bilayers is similar to that of phosphatidic acid. However, lipid X/EPC mixtures form micelles in excess lipid X whereas mixtures of phosphatidic acid/EPC vesiculate at all ratios. The same is true for lipid X/EPE mixtures. Small unilamellar vesicles of an average diameter of 40 nm form spontaneously upon dispersion of a dry lipid X/EPE film (molar ratio = 10). Unsonicated dispersions of lipid X/EPC (molar ratio = 1) are subjected to pH-jump treatment which involves raising of the pH to 11-12 and subsequent lowering of the pH to between 7.5 and 8.5. Such a treatment has little effect on the vesicle size and size distribution as compared to a control dispersion at pH 8.2. The mean size is determined to be 92 +/- 60 nm. Electron micrographs of freeze-fractured samples of lipid X/EPC (molar ratio = 1) reveal the presence of mainly micelles at pH 12. Upon lowering the pH to neutrality these micelles become unstable and aggregate/fuse rapidly to unilamellar vesicles (average diameter 95 +/- 40 nm). Sonication of equimolar mixtures of lipid X and EPC at pH 7 yields small unilamellar vesicles of a diameter of 20-25 nm as well as mixed micelles of a size between 15 and 17 nm. This behaviour is again different from that of mixed EPC/phosphatidic acid dispersions which form small unilamellar vesicles. The presence of lipid X in such mixtures does not prevent the aggregation/fusion to larger vesicles during freezing of the dispersion. As with pure EPC bilayers, stabilization is, however, achieved in the presence of 10% sucrose. This indicates that the covalently bonded glucosamine group of lipid X cannot substitute water of hydration in neighbouring EPC molecules.     

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.     

Stabilization of small unilamellar phospholipid vesicles during spray-drying

In the presence of 10% (0.3 M) sucrose in the aqueous medium, small unilamellar phospholipid vesicles are preserved during freeze-drying and spray-drying. Moreover, the bilayer integrity and permeability barrier are maintained during these processes.     

pH-dependent fusion of phosphatidylcholine small vesicles. Induction by a synthetic amphipathic peptide

A synthetic, amphipathic 30-amino acid peptide with the major repeat unit Glu-Ala-Leu-Ala (GALA) was designed to mimic the behavior of the fusogenic sequences of viral fusion proteins. GALA is a water-soluble peptide with an aperiodic conformation at neutral pH and becomes an amphipathic alpha-helix as the pH is lowered to 5.0 where it interacts with bilayers. Fluorescence energy transfer measurements indicated that GALA induced lipid mixing between phosphatidylcholine small unilamellar vesicles but not large unilamellar vesicles. This lipid mixing occurred only at pH 5.0 and not at neutral pH. Concomitant with lipid mixing, the vesicles increased in diameter from 500 to 750 to 1000 A as measured by dynamic light scattering and internal volume determination. GALA induced leakage of small molecules (Mr 450) at pH 5.0 was too rapid to permit detection of contents mixing. However, retention of larger molecules (Mr 4100) under the same conditions suggests that vesicle fusion is occurring. For a 100/1 lipid/peptide ratio all vesicles fused just once, whereas for a 50/1 ratio higher order fusion products formed. A mass action model gives good simulation of the kinetics of increase in fluorescence intensity and yields rate constants of aggregation and fusion. As the lipid to peptide ratio decreases from 100/1 to 50/1 both rate constants of aggregation and fusion increase, indicating that GALA is a genuine inducer of vesicle fusion. The presence of divalent cations which can alter GALAs conformation at pH 7.5 had little effect on its lipid mixing activity. GALA was modified by altering the sequence while keeping the amino acid composition constant or by shortening the sequence. These peptides did not have any lipid mixing activity nor did they induce an increase in vesicle size. Together, these results indicate that fusion of phosphatidylcholine small unilamellar vesicles induced by GALA requires both a peptide length greater than 16 amino acids as well as a defined topology of the hydrophobic residues.     

Effects of temperature and lipid composition on the serum albumin-induced aggregation and fusion of small unilamellar vesicles

Small unilamellar vesicles of egg phosphatidylcholine (PC) or dimyristoylphosphatidylcholine, mixed with small unilamellar vesicles labelled with 2-(10-(1-pyrene)decanoyl)phosphatidylcholine, exhibit a constant average size and excimer to monomer (E/M) ratio for several hours when incubated at pH 3.6 at a temperature higher than the phase transition temperature (T_c) of the lipids. Addition of bovine serum albumin to this system produces a transient turbidity increase, a fast decrease in the E/M ratio, a partial loss of vesicle-entrapped [¹⁴C]sucrose and a measurable leak-in of externally added sucrose. Sepharose 4B filtration of the system demonstrates that the E/M ratio decrease is strictly paralleled by the formation of liposomes which exhibit a low E/M ratio and a hydrodynamic radius larger than that of small unilamellar vesicles. These data demonstrate that the E/M ratio decrease can be unequivocally ascribed to a vesicle-vesicle fusion process induced by serum albumin. The rate of serum-albumin induced fusion of small unilamellar vesicles is: (a) maximal at a stoichiometric ratio of approx. 2 albumins per vesicle: (b) sensitive to the nature of the lipid and; (c) not altered when human serum albumin replaces bovine serum albumin. The rate of albumin-induced fusion of dimyristoylphosphatidylcholine small unilamellar vesicles is higher below the T_c of the lipid and increases with temperature above the T_c. The formation of protein-bound aggregates with defined stoichiometries and a high local vesicle concentration, as well as changes in the local degree of hydration, are proposed to be the driving forces for the protein-induced vesicle fusion in this system.     

Fusogenic activity of δ-haemolysin from Staphylococcus aureus in phospholipid vesicles in the liquid-crystalline phase

A study has been conducted of the interaction of the lytic toxin δ-haemolysin with vesicles of phospholipid, using electron microscopy, fluorescence depolarisation and excimer fluorescence. The peptide is shown to be a fusogen towards phosphatidylcholine vesicles in fluid phases. In the presence of gel phase lipid, fusion between fluid and gel phases is not seen. Fluid phase lipid vesicles are fused together to form large multilamellar structures, and initial vesicle size does not appear to be important since small unilamellar vesicles and large unilamellar vesicles are similarly affected. Fusogenic activity of δ-haemolysin is compared to that of melittin. The former is a progressive fusogen for fluid phase lipid, while the latter causes vesicle fusion in a manner related to occurrence of a lipid phase transition.     

Fusogenic activity of delta-haemolysin from Staphylococcus aureus in phospholipid vesicles in the liquid-crystalline phase

A study has been conducted of the interaction of the lytic toxin delta-haemolysin with vesicles of phospholipid, using electron microscopy, fluorescence depolarisation and excimer fluorescence. The peptide is shown to be a fusogen towards phosphatidylcholine vesicles in fluid phases. In the presence of gel phase lipid, fusion between fluid and gel phases is not seen. Fluid phase lipid vesicles are fused together to form large multilamellar structures, and initial vesicle size does not appear to be important since small unilamellar vesicles and large unilamellar vesicles are similarly affected. Fusogenic activity of delta-haemolysin is compared to that of melittin. The former is a progressive fusogen for fluid phase lipid, while the latter causes vesicle fusion in a manner related to occurrence of a lipid phase transition.     

Temperature-induced fusion of small unilamellar vesicles formed from saturated long-chain lecithins and diheptanoylphosphatidylcholine

Small unilamellar vesicles which form when gel-state long-chain phosphatidylcholines are mixed with micellar short-chain lecithins undergo an increase in size as the long-chain species melts to its liquid-crystalline form. Analysis of the vesicle population with quasi-elastic light scattering shows that the particle size increases from 90-A radius to greater than 5000-A radius. Resonance energy transfer experiments show total mixing of lipid probes with unlabeled vesicles only when the Tm of the long-chain phosphatidylcholine is exceeded. This implies that the large size change represents a fusion process. Aqueous compartments are also mixed during this transition. 31P NMR analysis of the vesicle mixtures above the phase transition shows a great degree of heterogeneity with large unilamellar particles coexisting with oligo- and multilamellar structures. Upon cooling the vesicles below the Tm, the original size distribution (e.g., small unilamellar vesicles) is obtained, as monitored by both quasi-elastic light scattering and 31P NMR spectroscopy. This temperature-induced fusion of unilamellar vesicles is concentration dependent and can be abolished at lower total phospholipid concentrations. It occurs over a wide range of long-chain to short-chain ratios and occurs with 1-palmitoyl-2-stearoylphosphatidylcholine and dimyristoylphosphatidylcholine as well. Characterization of this fusion event is used to understand the anomalous kinetics of water-soluble phospholipases toward these unusual vesicles.     

Characterization of the size distribution of unilamellar vesicles by gel filtration, quasi-elastic light scattering and electron microscopy

The size and size distribution of unilamellar phospholipid vesicles present in unsonicated phosphatidic acid and mixed phosphatidic acid/phosphatidylcholine dispersions were determined by gel filtration, quasi-elastic light scattering and freeze-fracture electron microscopy. The vesiculation in these dispersions was induced by a transient increase in pH as described previously (Hauser, H. and Gains, N. (1982) Proc. Natl. Acad. Sci. USA 79, 1683–1687). The resulting phospholipid dispersions are heterogeneous consisting of small unilamellar vesicles (average radius r < 50 nm) and large unilamellar vesicles (average r ranging from about 50 to 500 nm). The smallest vesicles with r = 11 ± 2 nm are observed with dispersions of pure phosphatidic acid, the population of these vesicles amounting to about 80% of the total lipid. With increasing phosphatidylcholine content the radius of the small unilamellar vesicles increases and at the same time the population of small unilamellar vesicles decreases. The average radius of small unilamellar vesicles present in phosphatidic acid/phosphatidylcholine dispersions (mole ratio, 1:1) is 17.5 ± 2 nm, the population of these vesicles amounting to about 70% of the total lipid. By a combination of gel filtration, quasi-elastic light scattering and freeze-fracture electron microscopy it was possible to characterize the large unilamellar vesicles. This population is heterogeneous with its mean radius also increasing with increasing phosphatidylcholine content. After separating the large unilamellar vesicles from small unilamellar vesicles on Sepharose 4B it can be shown by quasi-elastic light scattering that in pure phosphatidic acid dispersions 80–90% of the large unilamellar vesicle population consist of vesicles with a mean radius of 170 nm. In mixed phosphatidic acid/phosphatidylcholine dispersions this radius increases to about 265 nm as the phosphatidylcholine content is raised to 90 mol%.     

Membrane proteins exposed on the external side of the intestinal brush-border membrane have fusogenic properties

The intestinal brush-border membrane contains one or several membrane proteins that mediate fusion and/or aggregation of small unilamellar egg phosphatidylcholine vesicles. The fusion is accompanied by a partial loss of vesicle contents. Proteolytic treatment of the brush-border membrane with proteinase K abolishes the fusogenic property. This finding suggests that the fusogenic activity is associated with a membrane protein exposed on the external or luminal side of the brush-border membrane. Activation of intrinsic proteinases of the brush-border membrane liberates water-soluble proteins (supernate proteins). These proteins behave in an analogous way to intact brush-border membrane vesicles; they induce fusion of egg phosphatidylcholine vesicles and render the egg phosphatidylcholine bilayer permeable to ions and small molecules (Mr less than or equal to 5000). Furthermore, supernate proteins mediate phosphatidylcholine and cholesterol exchange between two populations of small, unilamellar phospholipid vesicles. Supernate proteins are fractionated on Sephadex G-75 SF yielding three protein peaks of apparent Mr greater than or equal to 70,000, Mr = 22,000 and Mr = 11,500. All three protein fractions show similar phosphatidylcholine-exchange activity, but they differ in their effects on the stability of egg phosphatidylcholine vesicles. The protein fraction with an apparent Mr greater than or equal to 70,000 has the highest fusogenic activity while the protein fraction of apparent Mr = 11,500 appears to be most effective in rendering the egg phosphatidylcholine bilayer permeable.     

Effect of Phosphatidylcholine on Acrosome Breakdown and Fertilizing Capacity of Amphibian Spermatozoa

The effect of vesicles of purified egg yolk phosphatidylcholine on the fertilizing capacity and acrosome breakdown of amphibian spermatozoa was studied. When Bufo arenarum spermatozoa were incubated with either small unilamellar vesicles (prepared by sonication) or with large unilamellar vesicles (prepared by reverse-phase evaporation) a decrease in the fertilizing capacity of spermatozoa was found. At the same phosphatidylcholine concentration, large unilamellar vesicles were more inhibitory than small unilamellar vesicles. The inhibition was dependent upon the phospholipid concentration and the length of the incubation period. Small unilamellar vesicles did not modify the time course of acrosome breakdown in Leptodactylus chaquensis , while large unilamellar vesicles markedly accelerated the rate of acrosome breakdown. In both biossays, the charge of the vesicles (made either positive or negative by the addition of 5% stearylamine or 5% phosphatidic acid) did not influence their biological effect. Multilamellar vesicles did not alter the fertilizing capacity nor the acrosome breakdown. We conclude that the size and the structure of the vesicles are important parameters in determining the inhibitory capacity of phosphatidyl choline on amphibian fertilization.     

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

Summary 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.Abbreviations PEG Polyethylene glycol - IMP Intramembranous particle - PC Phosphatidylcholine - PS Phosphatidylserine - SUV Small unilamellar vesicles - MLV Multilamellar vesicles - DPPC Dipalmitoyl phosphatidylcholine - DSC Differential scanning calorimetry     

Aggregation and Phospholipid Intermixing of Oleic-Acid-Containing Phosphatidylcholine Vesicles Induced by Polylysine

Abstract

Polylysine induced aggregation and phospholipid intermixing between small unilamellar vesicles of egg yolk phosphatidylcholine containing free oleic acid. the process was dependent on pH, being attributed to the presence of oleic acid. Neither intermixing nor leakage of the encapsulated aqueous contents was detected, nor did the size of such vesicles increase after treatment with polylysine. the maximum value of phospholipid intermixing was about 50%. these results are interpreted as representing reversible hemifusion between vesicles, without total membrane fusion.     

Interaction of short-chain lecithin with long-chain phospholipids: characterization of vesicles that form spontaneously

Stable unilamellar vesicles formed spontaneously upon mixing aqueous suspensions of long-chain phospholipid (synthetic, saturated, and naturally occurring phosphatidylcholine, phosphatidylethanolamine, and sphingomyelin) with small amounts of short-chain lecithin (fatty acid chain lengths of 6-8 carbons) have been characterized by using NMR spectroscopy, negative staining electron microscopy, differential scanning calorimetry, and Fourier transform infrared (FTIR) spectroscopy. This method of vesicle preparation can produce bilayer vesicles spanning the size range 100 to greater than 1000 A. The combination of short-chain lecithin and long-chain lecithin in its gel state at room temperature produces relatively small unilamellar vesicles, while using long-chain lecithin in its liquid-crystalline state produces large unilamellar vesicles. The length of the short-chain lecithin does not affect the size distribution of the vesicles as much as the ratio of short-chain to long-chain components. In general, additional short-chain decreases the average vesicle size. Incorporation of cholesterol can affect vesicle size, with the solubility limit of cholesterol in short-chain lecithin micelles governing any size change. If the amount of cholesterol is below the solubility limit of micellar short-chain lecithin, then the addition of cholesterol to the vesicle bilayer has no effect on the vesicle size; if more cholesterol is added, particle growth is observed. Vesicles formed with a saturated long-chain lecithin and short-chain species exhibit similar phase transition behavior and enthalpy values to small unilamellar vesicles of the pure long-chain lecithin prepared by sonication. As the size of the short-chain/long-chain vesicles decreases, the phase transition temperature decreases to temperatures observed for sonicated unilamellar vesicles. FTIR spectroscopy confirms that the incorporation of the short-chain lipid in the vesicle bilayer does not drastically alter the gauche bond conformation of the long-chain lipids (i.e., their transness in the gel state and the presence of multiple gauche bonds in the liquid-crystalline state).     

Fusion of phospholipid vesicles in association with the appearance of lipidic particles as visualized by freeze fracturing

The addition of Ca2+ to small unilamellar vesicles of an equimolar mixture of egg phosphatidylcholine and cardiolipin induces fusion of these vesicles in association with the appearance of lipidic particles on the fusion sites.     

Vaccine entrapment in liposomes.

The use of liposomes as carriers of peptide, protein, and DNA vaccines requires simple, easy-to-scale-up technology capable of high-yield vaccine entrapment. Work from this laboratory has led to the development of techniques that can generate liposomes of various sizes, containing soluble antigens such as proteins and particulate antigens (e.g., killed or attenuated bacteria or viruses), as well as antigen-encoding DNA vaccines. Entrapment of vaccines is carried out by the dehydration-rehydration procedure which entails freeze-drying of a mixture of "empty" small unilamellar vesicles and free vaccines. On rehydration, the large multilamellar vesicles formed incorporate up to 90% or more of the vaccine used. When such liposomes are microfluidized in the presence of nonentrapped material, their size is reduced to about 100 nm in diameter, with much of the originally entrapped vaccine still associated with the vesicles. A similar technique applied for the entrapment of particulate antigens (e.g., Bacillus subtilis spores) consists of freeze-drying giant vesicles (4-5 microm in diameter) in the presence of spores. On rehydration and sucrose gradient fractionation of the suspension, up to 30% or more of the spores used are associated with generated giant liposomes of similar mean size.     

Entrapment of proteins in phosphatidylcholine vesicles.

The trapping efficiency of globular proteins in four different types of phosphatidylcholine vesicles was systematically studied. Vesicles were generated in a mixture of 125I-labeled proteins of various molecular weights. The trapped proteins were separated from untrapped proteins by gel filtration and ultrafiltration and subsequently analyzed by gel electrophoresis and autoradiography. Entrapment of proteins was demonstrated by their resistance to trypsin digestion. The relative amount of each entrapped protein species was then compared to that of the original protein solution. In multilamellar vesicles and large unilamellar vesicles, proteins of molecular weight up to 97 000 had the same trapping efficiency as sucrose. In small unilamellar vesicles generated by either sonication or ethanol injection, however, the relative trapping efficiency of protein decreased progressively as the molecular weight of the protein became greater. For example, the trapping efficiency of alpha-amylase (Mr 97 000) was only half of that for sucrose. The apparent decrease in trapping efficiency with the protein's molecular weight in small unilamellar vesicles canbe accounted for by the combination of the bound water layer at the vesicle's internal surface and the steric hindrance when protein is captured during vesicle formation.     

Liposomes as Immunoadjuvants and Vaccine Carriers: Antigen Entrapment

Successful use of liposomes as immunological adjuvants in vaccines requires simple, easy to scale up technology capable of high-yield antigen entrapment. Recent work from this laboratory has led to the development of techniques that can generate liposomes of various sizes containing soluble antigens such as proteins or particulate antigens such as whole, live, or attenuated bacteria or viruses. Entrapment of proteins is carried out by the dehydration-rehydration procedure, which entails freeze-drying of a mixture of "empty" small unilamellar vesicles and free antigens. Upon rehydration, the large multilamellar vesicles that are formed incorporate up to 80% of the antigen used. When such liposomes are microfluidized in the presence of nonentrapped material, their size is reduced to about 100 nm in diameter, with much of the originally entrapped antigen still associated with the vesicles. A similar technique applied to the entrapment of particulate antigens (e.g., Bacillus subtilis spores) consists of freeze-drying giant vesicles (4-5 μm in diameter) in the presence of spores. On rehydration and sucrose gradient fractionation of the suspension, up to 27% of the spores used are associated with generated giant liposomes of similar mean size.     

Effect of DNA complexation and freeze-drying on the physicochemical characteristics of cationic liposomes

We describe the use of saccharides, such as sorbitol, mannitol, sucrose, maltodextrin, and dextran, as cyoprotectants for freeze-drying cationic liposomes. Saccharides can protect liposomes either by interacting with phospholipid headgroups or by forming an amorphous glass surrounding the vesicles, thus preventing aggregation, mechanical rupture of membrane, fusion of liposomes, and drug leakage. We have particularly considered liposome characteristics, such as size, zeta potential, and ability in complexing DNA, before and after freeze-drying. Our study indicates that cationic liposomes are able to maintain liposome characteristics after lyophilization and rehydration and maintain the ability to complex DNA even if the strength of the interaction forces was of lower intensity with respect to liposomes before lyophilization.