Studies on the mechanism of membrane fusion. Role of head-group composition in calcium- and magnesium-induced fusion of mixed phospholipid vesicles

Do hydrophobic sequences cleaved from cellular polypeptides induce membrane fusion reactions in vivo?

The concept that a direct interaction between Ca2+ and phospholipids is a major factor in membrane fusion reactions is questioned. Attention is drawn to a number of findings on associations between fusion and the proteolysis of membrane proteins. It is proposed that hydrophobic polypeptides, which are functionally comparable to the fusogenic proteins of certain viruses but which are produced in cells by the endogenous proteolysis of membrane and cellular proteins, may induce membrane fusion reactions in vivo.

GM1 micelles modify the transport properties of the ionophore gramicidin D in artificial planar bilayers

We have analyzed the effects induced in different phospholipid planar bilayers by monosialoganglioside micelles containing the ionophore gramicidin D. The membrane conductance increases after the addition of GM1 micelles at various ionophore/ganglioside ratios. We believe this fact may be ascribed to gramicidin molecules that incorporate into the bilayer together with gangliosides. In the presence of micelles the mean lifetime and the amplitude of the gramicidin single channel did not present relevant modifications when dioleoylphosphatidylcholine or phosphatidylserine were used to form the bilayer. Calcium proved to trigger the interaction between phosphatidylethanolamine membranes and GM1 micelles containing gramicidin. In this case the ionic pore presents a longer lifetime and a lower amplitude with respect to pure gramicidin. We suggest that different properties developed by gramicidin may depend on structural organization of gangliosides when incorporated into the phospholipid bilayer.

Membrane interaction and deformation

Membrane interaction and membrane deformation should be considered two aspects of the same phenomenon. Because membrane interaction depends on structural features such as curvature, chemical composition, and surface organization, one must expect that these features will change as membranes are brought together. Our observations on model phospholipid membranes have allowed us to measure and to distinguish electrostatic (coulombic or double-layer) forces, electrodynamic (van der Waals, dispersion) interactions, and hydration forces due to the solvation of polar groups that are stabilizing the membrane surface. As a consequence of these interactions, approaching membranes may flatten against each other, may change the packing density of their constituent molecules, may rearrange the composition of components in regions close to the apposing membrane, or may even show conformational changes in the arrangement of the lipid hydrocarbon chains. Abundant examples of such deformation accompanying membrane interaction are evident from x-ray diffraction and electron microscopic observations.

Binding of monovalent cations to phosphatidylserine and modulation of Ca2+- and Mg2+-induced vesicle fusion

The effect of several monovalent cations on the Ca2+-induced aggregation and fusion of sonicated phosphatidylserine (PS) vesicles is studied by monitoring the mixing of internal compartments of the fusing vesicles using the Tb/dipicolinic acid assay. The dissociation of the fluorescent Tb-dipicolinate complex which accompanies Ca2+-induced vesicle fusion is determined directly and is due to leakage of contents and entry of medium into vesicles. PS vesicles do not fuse when the medium contains only monovalent cations (at pH 7.4), regardless of the cation concentration or whether there is aggregation of the vesicles. A mass-action kinetic analysis of the data provides estimates for the rate of aggregation, C11, and for the rate of fusion per se, f11. Values of f11 increase dramatically with reduction in monovalent cation concentration and are primarily determined by binding ratios of Ca2+ or Mg2+ per PS. With 300 mM of monovalent cations, the fusion per se is essentially rate-limiting to the overall fusion process and values of f11 are significantly larger with the monovalent cations which bind the least, i.e., according to the sequence tetramethylammonium greater than K+ greater than Na+ greater than Li+. With monovalent cations in concentrations of 100 mM or less, the aggregation is rate-limiting to the fusion and the overall initial fusion rates are determined by an interplay between aggregation and fusion rates. Under conditions of fast aggregation, the Ca2+-induced fusion of small PS vesicles can occur within milliseconds or less.

Comparison of two liposome fusion assays monitoring the intermixing of aqueous contents and of membrane components

Divalent cation-induced fusion of large unilamellar vesicles (approx. 0.1 micron diameter) made of phosphatidylserine (PS) or phosphatidylglycerol (PG) has been studied. Intermixing of aqueous contents during fusion was followed by the Tb/dipicolinic acid fluorescence assay, and intermixing of membrane components by resonance energy transfer between fluorescent lipid probes. Both assays gave identical threshold concentrations for Ca2+, which were 2 mM for PS and 15 mM for PG. The dependencies of the initial rate of fusion on the concentration of PG vesicles determined by either assay were identical, the order of this dependence being 1.2 in the concentration range of 5-200 microM lipid. For PS liposomes, this order was found to be 1.5 in the fluorescent lipid assay. No leakage of contents was detected during the fusion of PG vesicles. Mg2+ inhibited the Ca2+-induced fusion of PS vesicles, but did not cause any fusion by itself, consistent with previous results with the Tb/dipicolinic acid assay.

Physicochemical characterization of large unilamellar phospholipid vesicles prepared by reverse-phase evaporation

Properties of large unilamellar vesicles (LUV), composed of phosphatidylcholine and prepared by reverse-phase evaporation and subsequent extrusion through Unipore polycarbonate membranes, have been investigated and compared with those of small unilamellar vesicles (SUV) and of multilamellar vesicles (MLV). The unilamellar nature of the LUV is shown by 1H-NMR using Pr3+ as a shift reagent. The gel to liquid-crystalline phase transition of LUV composed of dipalmitoylphosphatidylcholine (DPPC) monitored by differential scanning calorimetry, fluorescence polarization of diphenylhexatriene and 90 degrees light scattering, occurs at a slight lower temperature (40.8 degrees C) than that of MLV (42 degrees C) and is broadened by about 50%. The phase transition of SUV is shifted to considerably lower temperatures (mid-point, 38 degrees C) and extends over a wide temperature range. In LUV a well-defined pretransition is not observed. The permeability of LUV (DPPC) monitored by leakage of carboxyfluorescein, increases sharply at the phase transition temperature, and the extent of release is greater than that from MLV. Leakage from SUV occurs in a wide temperature range. Freeze-fracture electron microscopy of LUV (DPPC) reveals vesicles of 0.1-0.2 micron diameter with mostly smooth fracture faces. At temperatures below the phase transition, the larger vesicles in the population have angled faces, as do extruded MLV. A banded pattern, seen in MLV at temperatures between the pretransition and the main transition, is not observed in the smaller LUV, although the larger vesicles reveal a dimpled appearance.

Role of calcium in triggering the release of transmitters at the neuromuscular junction

Calcium ions have a key role in triggering the release of packaged transmitter at the amphibian neuromuscular junction and of the chromaffin granules at the adrenal medulla. It is suggested that (i) proteins on the vesicle and plasma membranes are of particular importance in promoting membrane fusion and exocytosis (ii) they may be divalent cation-stimulated ATPases, which form the calcium-binding sites or have a specific calcium-binding protein in close molecular apposition (iii) these ATPases in synaptic vesicles and chromaffin granules also generate a protonmotive force which is associated with the uptake of transmitter (iv) the osmotic properties of the vesicle may be important during fission, but it is not suggested that chemiosmotic effects are involved in Ca2+-triggered fusion (v) the action of calcium is markedly co-operative (vi) the adrenal medullary cell and the n.m.j. may differ in the Ca2+-binding site; there is evidence for the involvement of calmodulin in granule-plasmalemma fusion in the chromaffin cells, but not at present (surprisingly) for a role of this Ca2+-binding protein at the n.m.j. (vii) exocytosis requires MgATP (viii) phosphorylation of the ATPase may well be involved; phosphorylation via cAMP does not seem to be involved in fusion in either system (ix) the ATPase may undergo configurational changes during exocytosis and is markedly sensitive to the physical state of its phospholipid environment and to the oxidation of its -SH groups.

Interaction of phosphatidylserine-phosphatidylcholine liposomes with sickle erythrocytes. Evidence for altered membrane surface properties

The sickle erythrocyte (RBC) is a pathologic RBC that contains multiple membrane abnormalities. Some of these abnormalities have been implicated in the pathophysiology of vasoocclusive crises characteristic of sickle cell disease; others have yet to be defined in terms of their clinical significance. Recent information has shown that sickle RBC adhere abnormally to cultured endothelial cells yet little is known about the ways in which sickle cells interact with model membranes of defined size and lipid composition. We investigated this phenomenon by interacting sickle RBC with artificial lipid vesicles (liposomes) containing acidic phospholipids. Our results demonstrate that sickle disease (hemoglobin SS) RBC bind more of these liposomes than do normal or sickle trait (hemoglobin AS) RBC and that these differences are accentuated by hypoxia-induced sickling. Binding of liposome phospholipid to sickled RBC was not attributable to phospholipid exchange between liposomes and RBC and was consistent with a mechanism involving both membrane fusion and a stable reversible adhesion of liposomes to the RBC membrane.Investigations into the mechanism(s) underlying increased liposome binding to sickled RBC suggested that the known reversible translocation of aminophospholipids, phosphatidylserine (PS) and phosphatidyl-ethanolamine (PE), from the inner to the outer leaflet of the reversibly sickled RBC (RSC) plasma membrane during sickling may be a component of increased liposome binding to RSC. This idea was supported from results of experiments in which normal RBC were treated with diamide resulting in the expression of outer leaflet PE and PS and a stimulation of liposome binding to these cells. However, sickle RBC separated according to cell density on stractan gradients showed that irreversibly sickled RBC (ISC) were less capable of liposome binding than were discoid RSC. Since ISC are known to contain elevated levels of outer leaflet aminophospholipids, such a result suggests that other changes in the plasma membrane of sickle cells, in addition to phospholipid reorganization, are probably involved in enhanced liposome binding to these cells. In other experiments, we showed that liposomes containing l-phenylalanine were capable of delivering this antisickling agent into intact sickle RBC as demonstrated by the partial inhibition of hypoxia-induced sickling in vitro. Our results suggest that liposomes can be used as sensitive probes for investigating changes in RBC membrane properties, especially those that affect intermembrane interactions, and that liposomal transport systems may have significant implications in the therapy of sickle cell disease.

Raman spectroscopic and X-ray diffraction studies of the effect of temperature and Ca2+ on phosphatidylethanolamine dispersions

Raman spectroscopy and X-ray diffraction are used to study the effect of heat and Ca2+ on dimyristoylphosphatidylethanolamine dispersions. Unlike phosphatidylcholine dispersions, dimyristoylphosphatidylethanolamine bilayers (at pH 8) require heating above Tm in order for hydration to occur and apparently bind Ca2+ at very low levels. These results are related to models for membrane fusion.

Differentiation-related differences in the plasma membrane phospholipid asymmetry of myogenic and fibrogenic cells

We have determined the asymmetric distribution of two aminophospholipids phosphatidylethanolamine and phosphatidylserine in the plasma membrane of chick embryo fibroblast and myoblasts. Right-side-out membrane preparations were incubated with two different amidating reagents, trinitrobenzenesulfonate and isethionylacetimidate, under nonpenetrating conditions. Inside-out membranes were incubated with trinitrobenzenesulfonate. In fibroblasts, the similar plateau values suggested that 35% of the phosphatidylethanolamine and 20% of the phosphatidylserine is externally disposed. These values agree with previous measurements on fibroblast plasma membranes. In myoblasts, however, labelling plateaux were achieved which suggested that 65% of the phosphatidylethanolamine and 45% of the phosphatidylserine is externally disposed. This represents a 2-3-fold increase in potentially fusogenic lipids on the external leaflet of the plasma membrane. This unique distribution of aminophospholipids in myoblasts extends through the stage of development during which myoblasts become competent to fuse and form myotubes in culture. Two inferences may be drawn from these results. First, the external concentration of aminophospholipids in myoblasts is enriched significantly over that of fibroblasts or erythrocytes. This orientation may contribute to its fusion competence. Second, although large amounts of externally disposed aminophospholipid may be necessary for myoblast fusion, they do not confer fusion competence.

Lipid domain structures in biological membranes

Phase separation represents a possibility for segregation of lipidic membrane components into structurally distinct domains. Freeze-fracture electronmicroscopy is a useful method for detection of lipid domains. Indications of a possible domain-nature of structures are a regular pattern within a separated area, a regular outline of such an area and a local modulation of curvature (evagination or invagination). Candidates for domain structures in biological membranes are smooth particle-free areas and arrays of regularly arranged particles. The interpretation of the particle-free areas is more reliable than that of the arrays with regularly arranged particles. Phase separation in biological membranes can be induced experimentally by lowering the temperature, but physiologically the isothermically induced domains are more important. Factors in control of isothermic domain formation are divalent cations, proteins, cholesterol etc. Suggestions on the biological relevance of domain formation concern mainly their role in the mechanism of membrane fusion, but domains in form of transient or stable membrane structures seem to occur also otherwise and disturbances in domain formation or artificially induced domains can be suitable for pathological alterations.

Fusion of phospholipid vesicles arrested by quick-freezing. The question of lipidic particles as intermediates in membrane fusion

We have examined the early events in Ca2+-induced fusion of large (0.2 microns diameter) unilamellar cardiolipin/phosphatidylcholine and phosphatidylserine/phosphatidylethanolamine vesicles by quick-freezing freeze-fracture electron microscopy, eliminating the necessity of using glycerol as a cryoprotectant. Freeze-fracture replicas of vesicle suspensions frozen after 1-2 s of stimulation revealed that the majority of vesicles had already undergone membrane fusion, as evidenced by dumbbell-shaped structures and large vesicles. In the absence of glycerol, lipidic particles or the hexagonal HII phase, which have been proposed to be intermediate structures in membrane fusion, were not observed at the sites of fusion. Lipidic particles were evident in less than 5% of the cardiolipin/phosphatidylcholine vesicles after long-term incubation with Ca2+, and the addition of glycerol produced more vesicles displaying the particles. We have also shown that rapid fusion occurred within seconds of Ca2+ addition by the time-course of fluorescence emission produced by the intermixing of aqueous contents of two separate vesicle populations. These studies therefore have produced no evidence that lipidic particles are necessary intermediates for membrane fusion. On the contrary, they indicate that lipidic particles are structures obtained at equilibrium long after fusion has occurred and they become particularly prevalent in the presence of glycerol.

Ca2+-induced fusion of cardiolipin/phosphatidylcholine vesicles monitored by mixing of aqueous contents

The kinetics of Ca2+-induced fusion of large (0.1 micrometer) unilamellar cardiolipin/phosphatidylcholine (1:1) vesicles have been investigated by continuous monitoring of the mixing of the aqueous vesicle contents. In parallel, release of vesicle contents to the external medium has been followed. Initial fusion of the vesicles is non-leaky, release of vesicle contents to the external medium have been followed. Initial fusion of the vesicles is non-leaky, release of vesicle contents being largely a secondary phenomenon. The minimal Ca2+ concentration required for fusion in this system is approx. 9 mM. At higher Ca2+ concentrations fusion is extremely fast, occurring on the time scale of seconds.

Synexin facilitates fusion of specific phospholipid membranes at divalent cation concentrations found intracellularly

The effect of synexin (an adrenal medullary protein) on the kinetics of Ca2+- and Mg2+-mediated membrane fusion was examined. Membrane fusion was studied by monitoring intermixing of the aqueous contents of phospholipid vesicles. Synexin facilitated Ca2+-mediated, but not Mg2+-mediated, fusion of phosphatidate/phosphatidylethanolamine (1:3) and phosphatidate/phosphatidylserine/phosphatidylethanolamine/cholesterol (1:2:3:2) vesicles. The threshold concentration of Ca2+ for fusion was decreased to approximately equal to 10 microM in the presence of synexin at 6 micrograms/ml and 1.5 mM Mg2+ in vesicle suspensions containing 50 microM lipid. This effect of synexin was drastically inhibited by including 25% phosphatidylcholine (mol/mol) in the vesicle membrane. It is proposed that the Ca2+-dependent lipid-specific enhancement of membrane fusion by synexin contributes to an increase in the sensitivity of specific intracellular membranes to Ca2+ with respect to fusion.

Phospholipid transfer between vesicles. Dependence on presence of cytochrome P-450 and phosphatidylcholine-phosphatidylethanolamine ratio

The rate of transfer of spin-labeled phospholipid from donor vesicles of sonicated 1-acyl-2-(10-doxylstearoyl)-sn-glycero-3-phosphocholine to other vesicle was determined as a function of content of cytochrome P-450 and the phosphatidylcholine/phosphatidylethanolamine ratio in the acceptor vesicles. The transfer rate was measured as an increase in intensity that resulted from a decrease in the line width in the EPR spectrum of the spin-labeled phospholipids as they was transferred to the nonspin-labeled acceptor vesicles. A lower transfer rate was observed for acceptor vesicles of pure egg phosphatidylcholine vesicles than for vesicles for a mixture of phosphatidylcholine and phosphatidylethanolamine. The presence of cytochrome P-450 in the acceptor vesicles further increased the transfer rate. Those alterations in the mole ratios of the protein and the two phospholipids that made the bilayer of the reconstituted vesicles more like the membrane of the endoplasmic reticulum resulted in an increase in phospholipid-transfer rate. The mole ratios of components that produce high phospholipid-transfer rates were similar to those that in an earlier study produced a 31P-NMR spectrum characteristic of a nonbilayer phase. These findings suggest that, in the membrane of the endoplasmic reticulum, phospholipid exchange may be an important element in function and interaction with other intracellular organelles.

Early events in calcium-induced liposome fusion

Calcium-induced interaction of liposomes composed of pure phosphatidylserine (PS) has been studied using a rapid-mixing, rapid-freeze device. Freeze-fracture electron microscopy of this material revealed that liposomes react very rapidly after addition of calcium ions. After only 10 ms (the resolution of the technique) vesicle fusion was apparent. At the same time, however, vesicles also collapsed, and appeared as aggregates of flattened membranes. This may explain controversies which have arisen over vesicle fusion studied with more indirect methods.

Intermembrane contact affects calcium binding to phospholipid vesicles

Binding of Ca2+ to liposomes composed of phosphatidylserine (PtdSer) was analyzed by potentiometric titrations. Ca2+ binding to large unilamellar PtdSer vesicles was saturable at a stoichiometry of 1:2 (Ca2+/PtdSer). At approximately 6 X 10(-4) M [Ca2+]free, the binding curve exhibited a discontinuity that can be attributed to the formation of a Ca2+/PtdSer complex with a higher affinity. When both Ca2+ and Mg2+ are present, depending on the relative concentrations, Mg2+ can either complete or can enhance Ca2+ binding. Concomitant to the enhanced binding, the vesicle suspension was found to aggregate, suggesting that close contact of membranes is a prerequisite for the abrupt change in affinity. This concept was tested by binding studies with liposomes of mixed composition. It was found that the incorporation of 50 mol% phosphatidylethanolamine (PtdEtn) into PtdSer liposomes produced a similar binding pattern to that of pure PtdSer with a saturable stoichiometry of 1:2 (Ca2+/PtdSer). However, incorporation of 50 mol% phosphatidylcholine (PtdCho) completely abolished the discontinuous shift in affinity and apparent saturation was reached at a stoichiometry of 1:4 (Ca2+/PtdSer). In addition, Ca2+ binding to PtdSer liposomes with 10 mol% galactosylcerebroside was not altered when compared to pure PtdSer, whereas 10 mol% of the glycolipid GL-4 abolished the increased binding. The results are closely correlated with recent findings on the role of the membrane composition in Ca2+-induced fusion of liposomes and argue in favor of a specific Ca2+/PtdSer complex (with 1:2 stoichiometry) forming only at points of close contact between membranes and serving as the trigger for membrane fusion.

Modulation of membrane fusion by calcium-binding proteins

The effects of several Ca2+-binding proteins (calmodulin, prothrombin, and synexin) on the kinetics of Ca2+-induced membrane fusion were examined. Membrane fusion was assayed by following the mixing of aqueous contents of phospholipid vesicles. Calmodulin inhibited slightly the fusion of phospholipid vesicles. Bovine prothrombin and its proteolytic fragment 1 had a strong inhibitory effect on fusion. Depending on the phospholipid composition, synexin could either facilitate or inhibit Ca2+-induced fusion of vesicles. The effects of synexin were Ca2+ specific. 10 microM Ca2+ was sufficient to induce fusion of vesicles composed of phosphatidic acid/phosphatidylethanolamine (1:3) in the presence of synexin and 1 mM Mg2+. We propose that synexin may be involved in intracellular membrane fusion events mediated by Ca2+, such as exocytosis, and discuss possible mechanisms facilitating fusion.

Control of membrane fusion by phospholipid head groups. I. Phosphatidate/phosphatidylinositol specificity

We have studied the characteristics of fusion of large unilamellar vesicles composed of phosphatidate and phosphatidylinositol alone and in mixtures with other naturally occurring phospholipids. Fusion was induced by the addition of Ca2+ or Mg2+ and was monitored by detecting the mixing of aqueous vesicle contents. Release of vesicle contents was measured by dequenching of carboxyfluorescein fluorescence. Aggregation was monitored by 90 degrees light scattering. The results indicated striking differences with respect to the fusion capacity of the different vesicles. Phosphatidate vesicles fuse in the presence of both Ca2+ and Mg2+ at threshold concentration ranges of 0.03-0.1 mM (Ca2+) and 0.07-0.15 mM (Mg2+) depending on the pH of the medium, 8.5-6.0, respectively. In contrast, phosphatidylinositol vesicles do not fuse with either Ca2+ or Mg2+ even at 50 mM concentrations, in spite of aggregation induced by both cations in the range of 5-10 mM. A large difference in terms of fusion capacity is retained even when these two phospholipids are mixed with phosphatidylserine, phosphatidylethanolamine and phosphatidylcholine in 2 : 2 : 4 : 2 molar ratios. The results are discussed in terms of the molecular mechanism of membrane fusion and the possible role of the metabolic interconversion of phosphatidylinositol to phosphatidate as an on-off control system for membrane fusion phenomena involved in secretion.

Calcium- and magnesium-induced fusion of mixed phosphatidylserine/phosphatidylcholine vesicles: effect of ion binding

The aggregation, leakage, and fusion of pure PS (phosphatidylserine) and mixed PS/PC (phosphatidylcholine) sonicated vesicles were studied by light scattering, the release of encapsulated carboxyfluorescein, and a new fusion assay which monitors the mixing of the internal compartments of fusing vesicles. On a time scale of 1 min the extent of fusion was considerably greater than leakage. The Ca2+ and Mg2+ concentrations required to induce fusion increased when the PS content of the vesicles was decreased, and/or when the NaCl concentration was increased. Calculations employing a modified Gouy-Chapman equation and experimentally determined intrinsic binding constants of Na+ and Ca2+ to PS were shown to predict correctly the amount of Ca2+ bound in mixed PS/PC vesicles. For vesicles composed of either pure PS or of mixtures with PC in 100 mM NaCl (4:1 and 2:1 PS/PC); the induction of fusion (on a time scale of minutes) occurred when the amount of Ca or Mg bound/PS molecule exceeded 0.35-0.39. The induction of fusion for both pure PS and PS/PC mixed vesicles (with PS exceeding 50%) can be explained by assuming that destabilization of these vesicles requires a critical binding ratio of divalent cations to PS.