Lipid vesicle-cell interactions. III. Introduction of a new antigenic determinant into erythrocyte membranes.

The introduction of a new antigenic determinant, 2,4-dinitrophenyl-aminocaproyl-phosphatidylethanolamine (DNP-Cap-PE), into the surface membranes of intact human erythrocytes is described. Fresh cells were incubated in the presence of liposomes composed of 10% DNP-Cap-PE, 5% stearylamine, 20% lysolecithin, and 65% lecithin. Such liposome-treated erythrocytes are shown to be susceptible to immune lysis by anti-DNP serum in the presence of complement. Uptake of DNP-Cap-PE by erythrocyte membranes is also demonstrated by immunofluorescence using indirect staining with rabbit anti-DNP serum followed by fluroescein-conjugated goat anti-rabbit IgG and by electron microscopy using ferritin-conjugated antibody. Antigen uptake did not occur at low temperatures or from vesicles lacking lysolecithin and stearylamine. Fluorescence microscopy shows that the antigen-antibody complexes are free to diffuse over the cell surface, eventually coalescing into a single area on the cell membrane. Electron microscopy suggests that a substantial proportion of the lipid antigen is incorporated by fusion of vesicles with the cell membrane. There are indications that vesicle treatment causes a small proportion of cells to invaginate.

Lipid vesicle-cell interactions. II. Induction of cell fusion.

The ability of lipid vesicles of simple composition (lecithin, lysolecithin, and stearylamine) to induce cells of various types to fuse has been investigated. One in every three or four cells in monolayer cultures can be induced to fuse with a vesicle dose of about 100 per cell. At such dosages and for exposures of 15 min to 1 h, vesicles have essentially no effect on cell viability. Under anaerobic conditions, these cells lyse rather than fuse. Avian erythrocytes are readily fused with lipid vesicles in the presence of dextran. Fusion indices increase linearly with the zeta potential of the vesicles (increasing stearylamine content), indicating that contact between vesicle and cell membrane is required. Fusion indices increase sublinearly with increasing lysolecithin content. Divalent cations increase fusion indices at high vesicle doses. The data presented are consistent with the hypothesis that cell fusion occurs via simultaneous fusion of a vesicle with two adhering cell membranes.

Lipid vesicle-cell interactions. I. Hemagglutination and hemolysis.

The interaction of lipid vesicles (liposomes) of several different compositions with erythrocytes has been investigated. Lecithin liposomes, rendered positively charged with stearylamine, exhibit potent hemagglutination activity in media containing low concentrations of electrolytes. The hemagglutination titer is found to be a linear function of the zeta potential of the lipid vesicles. Hemagglutination is reduced when the surface potential of the cells is made more positive by pH adjustment or enzyme treatment. Similarly, hemagglutination is reduced by increasing concentrations of electrolytes. Hemagglutination is examined theoretically and is shown to be consistent with vesicle-cell interactions that are due to only electrostatic forces. Vesicles containing lysolecithin in addition to lecithin and stearylamine cause lysis of erythrocytes, provided the lipids of the vesicles are above the crystal-liquid crystal phase transition temperature. In addition, hemolysis requires close juxtaposition of the vesicle to the cell membrane; vesicles precoated with antibodies exhibit severely diminished hemolytic activities, only a small fraction of which can be attributed to a reduction in hemagglutination titer. Evidence is presented indicating that a single vesicle is sufficient to lyse one cell. With regard to hemagglutination and hemolysis, lipid vesicles of simple composition mimic paramyxoviruses such as Sendai virus.

Energetics of permeation of thin lipid membranes by ions.

The energy of an ion in a thin hydrocarbon membrane relative to its energy in a bulk aqueous phase is considered in terms of the electrostatic and surface components that may be expected to be involved. Except when diffusion activation energies are large compared to partition free energies, the latter will control permeation rate and the state of an ion having the lowest partition energy will be critical for its permeability. This minimum is found when an ion is surrounded with a thin layer of water. All ions of the same charge will tend to be at their lowest state in a sphere of water of the same size. It is concluded, therefore, that all ions of a given charge will have about the same permeability in lipid membranes.

Effects of unstirred layers or transport number discontinuities on the transient and steady-state current-voltage relationships of membranes.

The effects of current-induced electrolyte accumulation and depletion on the electrical properties of a two-layered membrane system have been examined. The membrane consisted of a charged, ion permselective layer and an uncharged, non-selective layer. The model was designed to reveal the properties of membranes possessing long pores with ionic charges at one end or of ion-selective membranes bounded by highly unstirred aqueous layers. Electrolyte concentration profiles in the inert layer and their time-dependent changes were obtained from solutions of the diffusion equation under the condition of constant current. The profiles were then used to calculate the voltage developed across the membrane at various times after the current is switched on. The theoretical results are presented in the form of i-V curves with reduced coordinates that can be used to obtain time-current-voltage relationships for membranes of the type considered having any thickness of the non-selective layer and bathed in any concentration of any 1:1 electrolyte. Experimental results on a model composite membrane were in good agreement with calculations that assume that ion transport occurs only under the influence of electrical potential and concentration gradients, suggesting that in such systems, the combined effects of convection, osmosis, electro-osmosis, and concentration-dependence of diffusion coefficients, activity coefficients, and transference numbers are small. Voltage fluctuations in the form of periodic spikes were observed experimentally at the limiting current density (the current density at which the electrolyte concentration at one surface of the selective layer goes to 0). These phenomena were not seen when the current was in the direction leading to accumulation of electrolyte in the non-selective (unstirred) layer. Such composite membranes can exhibit S-shaped and N-shaped i-V curves under ramp-voltage and ramp-current clamps, respectively.

Determination of reflection coefficients of liposomes for some non-electrolytes by osmotic pressure measurement.

The reflection coefficients of bilayer lipid vesicles (liposomes) of various compositions have been determined for a number of non-electrolytes. The solutes were the same and the method of measurement was essentially the same as those which have been used to estimate an equivalent pore radius for erythrocytes. The method involves matching the osmotic pressure of solutions of a permeant test solute with that of a known inpermeant solute. Reflection coefficients for cholesterol-containing liposomes and those of erythrocytes are, when account is taken of those solutes known to permeate the erythrocyte by specialized pathways, not greatly different. Lipid bilayers can thus account for most of the permeability characteristics of the cell originally interpreted as due to aqueous pores. Reflection coefficients are significantly higher for egg phosphatidylcholine membranes that contain cholesterol than those which do not. There is a strong correlation between relative permeabilities derived from reflection coefficients and oil-water partition coefficients. There is also good agreement between these permeabilities and permeabilities measured by others, which exhibit an inverse dependence on molecular size. It is suggested that this tendency of membranes to pass small molecules more readily than large molecules, other properties being equal, is a consequence of the surface pressure of the constituent monolayers of the membrane.

Lipid unsaturation influences melittin-induced leakage of vesicles.

Investigation of vesicles composed of different phosphatidylcholines revealed that the extent of leakage of internal contents induced by the lytic agent melittin can range from practically none to essentially complete, depending upon the fatty acyl chain composition of the phospholipid. The extent of leakage increases with the number of double bonds in the series dioleoylphosphatidylcholine < dilinoleoylphosphatidylcholine < dilinolenoylphosphatidylcholine. It depends on the length of the saturated chain with 1-myristoyl-2-arachidonoylphosphatidylcholine vesicles being more sensitive to melittin induced leakage than 1-palmi-toyl-2-arachidonoylphosphatidylcholine, 1-stearoyl-2-arachidonoylphosphatidylcholine or 1-palmitoyl-2- docosahexaenoylphosphatidylcholine vesicles. The extent of leakage induced by melittin from vesicles composed of 1-palmitoyl-2-oleoylphosphatidylcholine, dioleoylphosphatidylcholine, 1-palmitoyl-2-arachidonoylphosphatidylcholine and 1-palmitoyl-2-docosahexaenoyl-phosphatidylcholine increases with the free volume parameter of these lipids for 1,6-diphenylhexatriene (Straume, M. and Litman, B.J. (1987) Biochemistry 26, 5113-5120). Among the lipid examined here, diphytanoylphosphatidylcholine vesicles were least susceptible to melittin induced leakage. The results indicate that lipid fatty acyl structure may be important in lipid-protein interactions of the kind simulated by melittin.

A metastable state of high surface activity produced by sonication of phospholipids.

Sonication of phosphatidylcholine dispersions generates a metastable high energy assembly of molecules, the existence of which is revealed by its conspicuous surface activity. Freshly sonicated liposome dispersions release molecules to the air/water interface at rates sufficient to produce a close-packed monolayer within minutes. In contrast, monolayers at the surface of multilamellar and extruded vesicles take hours to form. The highly surface active species appears within the first few minutes of sonication, long before a major reduction in turbidity occurs, and accumulates over the course of a few hours of sonication. It disappears upon exhaustive sonication, extrusion, addition of extruded vesicles, or, more slowly, simply on standing. Tests for extraneous substances in the lipids before as well as after sonication revealed amounts of degradation products too small to represent the observed surfactant. Direct evidence that the metastable aggregate releases intact phospholipids was provided by a novel procedure to characterize monolayer composition by comparing surface tension with surface potential, both as a function of surface density. Centrifugation and gel filtration chromatography indicate that the surface activity is associated with a particle of diameter larger than a lysophosphatidylcholine micelle but not larger than limit sonicated vesicles. The metastable material appears to be lipid molecules in other than the normal stable vesicular state, perhaps an incompletely closed vesicle, one in which the inner and outer monolayers have not equilibrated, or possibly a micellar form.

Lipid mixing during freeze-thawing of liposomal membranes as monitored by fluorescence energy transfer.

A new pair of fluorescence-energy-transferring probes, dansylphosphatidylethanolamine and dioctadecylindocarbocyanine, were incorporated separately into phospholipid vesicles to monitor intervesicle lipid mixing under various conditions. The transfer efficiencies of mixtures of sonicated vesicles labeled with 2 wt% donor dansylphosphatidylethanolamine (DnsPE) or with 1 wt% acceptor dioctadecylindocarbocyanine (DiI-C18) were negligible, but increased to about 25% after the vesicles had been frozen in a solid CO2/ethanol bath, thawed and diluted. The freeze-thaw-induced mixing of lipids between vesicles, signified by energy transfer, was dependent on lipid concentration and was promoted by 0.5-1.5 M KCl, 0.5 M potassium trichloroacetate and 5 mM sodium acetate (pH 4) and inhibited by 0.5 M LiCl, 0.5 M glycerol, 0.5 M sucrose, 0.15 M KCl and 0.15-1.5 M NaCl. These results support and complement previously reported measurements of the trapped volumes, turbidities and population size distributions of similarly treated liposomes. Comparison of the responses of paucilamellar vesicles with those of multilamellar vesicles suggests that lipid mixing during freeze-thawing can occur either during interaction of the outermost bilayers of vesicles or during interaction of all bilayers, possibly as a result of breakdown and reformation of bilayer structure.

Preparation of giant liposomes.

A method is described for the preparation of giant unilamellar lipid vesicles that are stable in electrolyte solution. In general, it involves dialysis of lipid and indifferent solute in a water-miscible organic solvent against an aqueous buffer. During dialysis the concentration of organic solvent decreases so that vesicles form under conditions where their internal contents are continuously hyperosmotic. Interlamellar attractive forces are neutralized, even between bilayer membranes with no net charge, and giant vesicles are generated in large numbers. The population is heterogeneous but most large vesicles have diameters between 10 and 100 micron. The method is simple. One procedure involves dialysis for a day or more of a methanol solution of phosphatidylcholine, supersaturated with methylglucoside, against an aqueous phase containing up to 1 M univalent electrolyte. The procedure is effective over a wide range of temperature and pH.

Fragmentation into small vesicles of dioleoylphosphatidylcholine bilayers during freezing and thawing.

Multilayered liposomes of some phosphatidylcholines progressively fragment into small vesicles when the electrolyte solution in which they are suspended is subjected to successive cycles of freezing and thawing. The fragmentation process, routinely monitored by absorbance measurements and verified by electron microscopy and dynamic light scattering, involves bilayer breakage and resealing. After 10 cycles of freezing and thawing in 0.1 M electrolyte solution, the result is a population of vesicles smaller than 200 nm diameter. Sucrose, a common cryoprotectant, completely inhibits fragmentation. Fragmentation is absolutely dependent upon the presence of an electrolyte. Those electrolytes most effective in promoting liposome fragmentation have large freezing point depressions and corresponding high solubilities at the freezing point. This, coupled with the observation that saturating concentrations of electrolyte are less effective than 0.1 M solutions indicates that an essential stage in the fragmentation process is osmotic extraction of water from the vesicles, i.e., ice formation in the external phase leads to a progressive increase in the electrolyte concentration of the residual external solution, which, in turn, dehydrates the vesicle. In addition, for maximal fragmentation, the minimum temperature must be at least as low as the solute eutectic temperature. Particular physical properties of the bilayer are also important, for dioleoyl and diphytanoyl derivatives are much more susceptible to fragmentation than are other phosphatidylcholines, and inclusion of 50 mol% cholesterol in dioleoylphosphatidylcholine completely inhibits membrane breakup. This system provides insight into mechanisms of freezing damage to membranes and may also offer a very simple and rapid assay for biological cryoprotectants.

The organization of n-alkanes in lipid bilayers.

The interaction of n-alkanes (C6--C16) with phosphatidylcholine has been studied by the combined use of differential scanning calorimetry, X-ray diffraction and monolayer techniques. It has been found that the thermal properties and ultrastructure of lipid-alkane vesicles are strongly dependent on the length of the n-alkanes. Long alkanes, such as tetradecane and hexadecane, increase the transition temperature of dimyristoyl phosphatidylcholine and dipalmitoyl phosphatidylcholine, while the X-ray data indicate that these long alkanes align parallel to the lipid acyl chains. In contrast, shorter alkanes, such as hexane and octane, decrease and broaden the thermal transition and electron density profiles show that these alkanes increase bilayer width by partitioning between the apposing monolayers of the bilayer. For lipids in the gel and liquid crystalline states, the short alkanes form an alkane region in the geometric center of the bilayer.

Detergent effects on enzyme activity and solubilization of lipid bilayer membranes.

Over 50 detergents were tested to establish which would be most effective in releasing proteins from membrane-bounded compartments without denaturing them. Various concentrations each of detergent were tested for two activities: (1) solubilization of egg phospholipid liposomes as measured by reduction of turbidity and (2) effect of detergent concentration on the activities of soluble, hydrolytic enzymes. Those detergents most effective in solubilizing 0.2% lipid and least detrimental to enzymes were five pure, synthetic compounds recently introduced: CHAPS, CHAPSO, Zwittergents 310 and 312, and octylglucoside. Industrial detergents were generally much inferior, insofar as they solubilized membranes inefficiently and/or inactivated certain hydrolytic enzymes readily. The five detergents were characterized by (a) an unusually high critical micelle concentration and (b) a preference for forming mixed micelles with lipids instead of forming pure micelles, as indicated by an ability to solubilize lipid at concentrations of detergent significantly below the critical micelle concentration. This characteristic permits solubilization of high concentrations of membrane below the critical micelle concentration of the detergent so that protein denaturation is minimized. A generally applicable guideline that emerged from this study is that detergents should be used at approximately their critical micelle concentration which should not be exceeded by the concentration of membrane. Similar considerations should apply to the use of detergents in purifying and reconstituting intrinsic membrane proteins.

Cholesterol stabilizes hemifused phospholipid bilayer vesicles.

Cholesterol was found to inhibit full fusion of oppositely charged phospholipid bilayer vesicles by stabilizing the contacting membranes at the stage of the hemifused intermediate. Vesicles of opposite charge containing different amounts of cholesterol were prepared using cationic (1,2-dioleoyl-sn-glycero-3-ethylphosphocholine) and anionic (dioleoylphosphatidylglycerol) phospholipids. Pairwise interactions between such vesicles were observed by fluorescence video microscopy in real time after electrophoretically maneuvering the vesicles into contact. Hemifusion accounted for more than 80% of the observed events when the vesicles contained 33-50 mole% cholesterol. In contrast, vesicles containing only a small proportion of cholesterol (

Characteristics of spectrin-induced leakage of extruded, phosphatidylserine vesicles.

At neutral pH spectrin induces modest leakage of trapped calcein from reverse-phase or extruded, but not sonicated, vesicles composed of phosphatidylserine, but not phosphatidylcholine. The extent of leakage from extruded vesicles is not or is only slightly affected by magnesium ions at a physiological concentration or calcium ions at a greater than physiological concentration, respectively. In addition to accounting for several previously discrepant observations on the lytic effects of spectrin, these findings indicate that some proteins like spectrin may destabilize vesicles with low curvature more readily than vesicles of high curvature, in contrast to certain amphiphilic peptides. 60% less leakage is induced from phosphatidylserine vesicles by heat-denatured than by native spectrin. In contrast, both trypsin- and subtilisin-treated spectrins, if sufficiently digested, induce several-fold more leakage than undigested spectrin. Since spectrin prepared either by 1 M Tris dissociation of Triton-extracted cytoskeletons or by low ionic strength extraction of ghosts released the same amounts of calcein from vesicles of various compositions, these effects are unlikely to reflect artifacts of spectrin preparation. Furthermore, spectrin is unlikely to promote leakage in vivo, since vesicles composed of phosphatidylserine, cholesterol and/or phosphatidylethanolamine, which constitute the lipid composition of the inner monolayer of the red cell membrane, did not leak on addition of spectrin, whereas vesicles composed of phosphatidylserine and phosphatidylcholine, did leak in the presence of spectrin.

A calorimetric and monolayer investigation of the influence of ions on the thermodynamic properties of phosphatidylcholine.

The effects of various ions and 2H2O on the thermal properties of phosphatidylcholine dispersions were studied using differential scanning calorimetry and the change in the surface potential of monolayers with temperature. The phosphatidylcholine in 2H2O dispersion exhibits a slightly higher transition temperature and lower enthalpy of melting than a phosphatidylcholine in H2O dispersion. Monovalent (H+, Na+, and Li+) and some divalent cations of chloride salts (Ba2+, Mg2+, and Sr2+) have no effect on the thermal properties of phosphatidylcholine, while halide salts of the di-positive ions Cd2+ and Ca2+ have an effect on both the enthalpy of melting and transition temperature. No effect attributable to the metal ion was observed in non-halide salts of cadmium. The chloride salt of La3+ has no effect on lipid thermal properties whereas that of Fe3+ affects the transition temperature. The enthalpy of melting of phosphatidylcholine in one molar solutions of potassium salts increases in the order: CNS minus greater than acetate greater than I minus. Such large, polarizable anions clearly interact with phosphatidylcholine and must therefore also confer a negative charge on the lipid. The potassium salt of SO4-2 minus has no effect. Possible origins of the observed trends are discussed.

Small-volume extrusion apparatus for preparation of large, unilamellar vesicles.

The design and performance of a filter holder which enables convenient preparation of volumes of up to a milliliter of large, unilamellar vesicles formed by extrusion (LUVETs) from multilamellar vesicles (MLVs) are described. The filter holder provides for back-and-forth passage of the sample between two syringes, a design that minimizes filter blockage, eliminates the need to change filters during LUVET preparation and reduces preparation time to a few minutes. Replicas of slam-frozen LUVETs in the electron microscope are unilamellar and reasonably homogeneous with an average diameter close to the pore size of the filters used to extrude them. Extrusion per se does not destabilize the vesicles, which trapped a fluorescent dye only when they were disrupted on freeze-thawing and during the first extrusion when most of the MLVs were apparently converted to LUVETs.

Metabolism of Transpired Ethanol by Eastern Cottonwood (Populus deltoides Bartr.).

Ethanol has previously been shown to be present in the xylem sap of flooded and nonflooded trees. Because of the constitutive presence of alcohol dehydrogenase in the mature leaves of woody plants, we hypothesized that the leaves and shoots of trees had the ability to metabolize ethanol supplied by the transpiration stream. 1-[14C]Ethanol was supplied to excised leaves and shoots of eastern cottonwood (Populus deltoides Bartr.) in short- and long-term experiments. More than 99% of the radiolabel was incorporated into plant tissue in short-term experiments, with more than 95% of the label remaining in plant tissue after 24 h. In all experiments, less than 5% of the label was transpired as ethanol and less than 1% was emitted as CO2. In excised leaf experiments, less than 0.5% of the radiolabel escaped from the leaf. Fifty percent of the label was incorporated into the petioles of excised leaves; 56% was incorporated into the stems of excised shoots. Very little label reached the leaf mesophyll cells of excised shoots, as revealed by autoradiography. Radiolabel appeared primarily in the water- and chloroform-soluble fractions in short-term experiments, whereas in long-term experiments, label was also incorporated into protein. These results demonstrate that the leaves and stems of trees appear to have substantial ability to scavenge ethanol from the transpiration stream, allowing efficient recovery of ethanol produced elsewhere by hypoxic tissues. When labeled ethanol was supplied to excised petioles in a 5-min pulse, 41% of the label was incorporated into organic acids. Some label was also incorporated into amino acids, protein, and the chloroform-soluble fraction, with very little appearing in neutral sugars, starch, or the insoluble pellet. Labeled organic acids were separated by high performance liquid chromatography and were composed of acetate, isocitrate, [alpha]-ketoglutarate, and succinate. There was no apparent incorporation of label into phosphorylated compounds. We conclude that, in higher plants, ethanol is metabolized to acetaldehyde and then to acetate by alcohol and aldehyde dehydrogenases, and then into general metabolism.

Methanol Emission from Leaves (Enzymatic Detection of Gas-Phase Methanol and Relation of Methanol Fluxes to Stomatal Conductance and Leaf Development).

We recently reported the detection of methanol emissions from leaves (R. MacDonald, R. Fall [1993] Atmos Environ 27A: 1709-1713). This could represent a substantial flux of methanol to the atmosphere. Leaf methanol production and emission have not been investigated in detail, in part because of difficulties in sampling and analyzing methanol. In this study we used an enzymatic method to convert methanol to a fluorescent product and verified that leaves from several species emit methanol. Methanol was emitted almost exclusively from the abaxial surfaces of hypostomatous leaves but from both surfaces of amphistomatous leaves, suggesting that methanol exits leaves via stomates. The role of stomatal conductance was verified in experiments in which stomates were induced to close, resulting in reduced methanol. Free methanol was detected in bean leaf extracts, ranging from 26.8 [mu]g g-1 fresh weight in young leaves to 10.0 [mu]g g-1 fresh weight in older leaves. Methanol emission was related to leaf development, generally declining with increasing leaf age after leaf expansion; this is consistent with volatilization from a cellular pool that declines in older leaves. It is possible that leaf emission could be a major source of methanol found in the atmosphere of forests.

Physical and biological properties of cationic triesters of phosphatidylcholine.

The properties of a new class of phospholipids, alkyl phosphocholine triesters, are described. These compounds were prepared from phosphatidylcholines through substitution of the phosphate oxygen by reaction with alkyl trifluoromethylsulfonates. Their unusual behavior is ascribed to their net positive charge and absence of intermolecular hydrogen bonding. The O-ethyl, unsaturated derivatives hydrated to generate large, unilamellar liposomes. The phase transition temperature of the saturated derivatives is very similar to that of the precursor phosphatidylcholine and quite insensitive to ionic strength. The dissociation of single molecules from bilayers is unusually facile, as revealed by the surface activity of aqueous liposome dispersions. Vesicles of cationic phospholipids fused with vesicles of anionic lipids. Liquid crystalline cationic phospholipids such as 1, 2-dioleoyl-sn-glycero-3-ethylphosphocholine triflate formed normal lipid bilayers in aqueous phases that interacted with short, linear DNA and supercoiled plasmid DNA to form a sandwich-structured complex in which bilayers were separated by strands of DNA. DNA in a 1:1 (mol) complex with cationic lipid was shielded from the aqueous phase, but was released by neutralizing the cationic charge with anionic lipid. DNA-lipid complexes transfected DNA into cells very effectively. Transfection efficiency depended upon the form of the lipid dispersion used to generate DNA-lipid complexes; in the case of the O-ethyl derivative described here, large vesicle preparations in the liquid crystalline phase were most effective.