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

Characterization of Ca2(+)-dependent phospholipid binding, vesicle aggregation and membrane fusion by annexins

The annexins are a family of structurally similar, Ca2(+)-dependent, phospholipid-binding proteins. We compared six members of this family (calpactin I heavy chain, lipocortins I and III, endonexin II, p68 and protein II) to determine their phospholipid-binding specificities, as well as their ability to promote aggregation and fusion of phospholipid vesicles. The Ca2+ requirement for all of the proteins was lowest for binding to vesicles composed of phosphatidic acid, followed by phosphatidylserine and then phosphatidylinositol. Only protein II, p68, lipocortin III and endonexin II bound to vesicles composed of phosphatidylethanolamine, and none bound to phosphatidylcholine. Both calpactin I heavy chain and lipocortin I promoted aggregation of phosphatidylserine- or phosphatidylinositol-containing vesicles in the presence of less than 10 microM-Ca2+. Lipocortin I promoted fusion of liposome membranes by lowering threshold Ca2+ concentrations. Although calpactin I heavy chain did not affect threshold Ca2+ concentrations, it did increase the rate and extent of spontaneous fusion. In contrast, p68 inhibited fusion at threshold Ca2+ concentrations. Whereas previous reports have emphasized properties that the annexins have in common, these findings reveal quantitative and qualitative differences among the annexins which may relate to distinct intracellular functions.

Cholesterol affects divalent cation-induced fusion and isothermal phase transitions of phospholipid membranes

The influence of cholesterol on divalent cation-induced fusion and isothermal phase transitions of large unilamellar vesicles composed of phosphatidylserine (PS) was investigated. Vesicle fusion was monitored by the terbium/dipicolinic acid assay for the intermixing of internal aqueous contents, in the temperature range 10-40 degrees C. The fusogenic activity of the cations decreases in the sequence Ca2+ greater than Ba2+ greater than Sr2+ much greater than Mg2+ for cholesterol concentrations in the range 20-40 mol%, and at all temperatures. Increasing the cholesterol concentration decreases the initial rate of fusion in the presence of Ca2+ and Ba2+ at 25 degrees C, reaching about 50% of the rate for pure PS at a mole fraction of 0.4. From 10 to 25 degrees C, Mg2+ is ineffective in causing fusion at all cholesterol concentrations. However, at 30 degrees C, Mg2+-induced fusion is observed with vesicles containing cholesterol. At 40 degrees C, Mg2+ induces slow fusion of pure PS vesicles, which is enhanced by the presence of cholesterol. Increasing the temperature also causes a monotonic increase in the rate of fusion induced by Ca2+, Ba2+ and Sr2+. The enhancement of the effect of cholesterol at high temperatures suggests that changes in hydrogen bonding and interbilayer hydration forces may be involved in the modulation of fusion by cholesterol. The phase behavior of PS/cholesterol membranes in the presence of Na+ and divalent cations was studied by differential scanning calorimetry. The temperature of the gel-liquid crystalline transition (Tm) in Na+ is lowered as the cholesterol content is increased, and the endotherm is broadened. Addition of divalent cations shifts the Tm upward, with a sequence of effectiveness Ba2+ greater than Sr2+ greater than Mg2+. The Tm of these complexes decreases as the cholesterol content is increased. Although the transition is not detectable for cholesterol concentrations of 40 and 50 mol% in the presence of Na+, Sr2+ or Mg2+, the addition of Ba2+ reveals endotherms with Tm progressively lower than that observed at 30 mol%. Although the presence of cholesterol appears to induce an isothermal gel-liquid crystalline transition by decreasing the Tm, this change in membrane fluidity does not enhance the rate of fusion, but rather decreases it. The effect of cholesterol on the fusion of PS/phosphatidylethanolamine (PE) vesicles was investigated by utilizing a resonance energy transfer assay for lipid mixing. The initial rate of fusion of PS/PE and PS/PE/cholesterol vesicles is saturated at high Mg2+ concentrations. With Ca2+, saturation is not observed for cholesterol-containing vesicles.(ABSTRACT TRUNCATED AT 400 WORDS)

Effects of platelet activating factor on calcium-lipid interactions and lateral phase separations in phospholipid vesicles

Recent studies localizing the inflammatory mediator, platelet activating factor (PAF, 1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine), to the membranes of stimulated neutrophils, raise the possibility that PAF may, in addition to its activities as a mediator, alter the physical properties of membranes. This, and the increasing evidence that calcium-lipid interactions may have central importance in membrane organizational structure and in functions of cell homeostasis and stimulus-response coupling, prompted us to study the effects of PAF on calcium-lipid interactions in lipid vesicles. Using fluorescence polarization of dansylated probes located in the glycerol portion of the membrane bilayer, PAF (at a concentration as low as 1 mol%) was shown to reduce membrane rigidification significantly during calcium-induced lateral phase separations. This effect of PAF was structurally dependent on both the 1-position alkyl linkage and the 2-position acetyl group as shown by studies of related lipid analogs. Furthermore, using a self-quenching probe, it was shown that inhibition of lateral phase separation did not account for this reduction in the calcium-induced membrane rigidification attributed to PAF. Data suggest that PAF at low concentrations may alter phospholipid head packing and, thereby, change membrane surface features during calcium-lipid interactions, effects which may ultimately explain some of its biological actions.

Protein-mediated fusion of liposomes with microsomal membranes of Aspergillus niger: evidence for a complex mechanism dealing with membranous and cytosolic fusogenic proteins

Membrane fusion is a fundamental and wide-spread phenomenon in the functioning of cells. Many studies were carried out concerning fusion of plasma membranes as for example cell-cell fusions or uptake by cells of lipid-enveloped viruses. The present study deals with the interaction of intracellular membranes of Aspergillus niger with artificial membranes (liposomes). Association is monitored by the uptake of radioactive liposomes by fungal microsomal membranes. The discrimination between aggregation and pure fusion is done by layering the liposomes-microsomes mixture on a continuous sucrose gradient. The accurate quantitation of the fusion phenomenon is monitored with a fluorescent assay based on resonance energy transfer (Struck, D.K. et al. (1981) Biochemistry 20, 4093-4099). Both methods show that, at physiological pH, there is a spontaneous fusion of microsomes with cholesterol-free liposomes. This phenomenon is protein dependent as trypsinized microsomal membranes are no longer able to fuse with liposomes. Biological significance of the fusion process has been demonstrated using microsomal intrinsic protein mannosylation assay; the enhancement of the lipid to protein ratio due to the fusion of liposomes with microsomes of A. niger results in an increase in the rate of endogenous proteins mannosylation. Moreover, cytosolic proteins of A. niger promote the fusion of any kind of liposomes with microsomes.

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.

Calcium-induced aggregation of phosphatidylcholine vesicles containing free oleic acid

Small unilamellar vesicles (SUV) formed by egg yolk phosphatidylcholine (PC) and free oleic acid (OA) undergo aggregation induced by Ca2+ at pH greater than 7.0. The rate of the process, as monitored by turbidity changes, presents a linear dependence on phospholipid concentration and a hyperbolic dependence on Ca2+ concentration. The aggregation curves show a lag period which is tentatively attributed to an activation step induced by Ca2+. The incorporation of either cholesterol, alpha-tocopherol or egg yolk phosphatidylethanolamine (PE) produced a decrease in the aggregation rate and an increase in the lag period. Fusion was not detected in any of the different experiments, either by the assay of mixing the membrane phospholipids or by the assay of mixing the aqueous contents. A possible mechanism, explaining the aggregation process is proposed.

Lipid mixing during membrane aggregation and fusion: why fusion assays disagree

The kinetics of lipid mixing during membrane aggregation and fusion was monitored by two assays employing resonance energy transfer between N-(7-nitro-2,1,3-benzoxadiazol-4-yl)phosphatidylethanolamine (NBD-PE) and N-(lissamine Rhodamine B sulfonyl)phosphatidylethanolamine (Rh-PE). For the "probe mixing" assay, NBD-PE and Rh-PE were incorporated into separate populations of phospholipid vesicles. For the "probe dilution" assay, both probes were incorporated into one population of vesicles, and the assay monitored the dilution of the molecules into the membrane of unlabeled vesicles. The former assay was found to be very sensitive to aggregation, even when the internal aqueous contents of the vesicles did not intermix. Examples of this case were large unilamellar vesicles (LUV) composed of phosphatidylserine (PS) in the presence of Mg2+ and small unilamellar vesicles (SUV) composed of phosphatidylserine in the presence of high concentrations of Na+. No lipid mixing was detected in these cases by the probe dilution assay. Under conditions where membrane fusion (defined as the intermixing of aqueous contents with concomitant membrane mixing) was observed, such as LUV (PS) in the presence of Ca2+, the rate of probe mixing was faster than that of probe dilution, which in turn was faster than the rate of contents mixing. Two assays monitoring the intermixing of aqueous contents were also compared. The Tb/dipicolinic acid assay reported slower fusion rates than the 1-aminonaphthalene-3,6,8-trisulfonic acid/N,N'-p-xylylene-bis(pyridinium bromide) assay for PS LUV undergoing fusion in the presence of Ca2+. These observations point to the importance of utilizing contents mixing assays in conjunction with lipid mixing assays to obtain the rates of membrane destabilization and fusion.(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of cholesterol on the divalent cation-mediated interactions of vesicles containing amino and choline phospholipids

We have used assays of lipid probe mixing, contents mixing and contents leakage to monitor the divalent cation-mediated interactions between lipid vesicles containing phosphatidylserine (PS) as a minority component together with mixtures of phosphatidylethanolamine (PE), phosphatidylcholine (PC) or sphingomyelin, and cholesterol in varying proportions. The initial rates of calcium- and magnesium-induced lipid probe quenching between vesicles, which reflect primarily the rates of vesicle aggregation, are strongly reduced as progressively higher proportions of PC or sphingomyelin are incorporated into PE/PS vesicles. The initial rates of divalent cation-induced contents mixing and contents leakage for PE/PS vesicles are also strongly reduced when choline phospholipids are incorporated into the vesicles in even low molar proportions. Sphingomyelin has a more potent inhibitory effect on these processes than does PC at an equal level in the vesicle membranes. The inclusion of cholesterol in these vesicles, at levels up to 1:2 moles sterol/mole phospholipid, has little effect on the rates of calcium- or magnesium-induced vesicle aggregation. However, cholesterol significantly enhances the initial rates of vesicle contents mixing and contents leakage in the presence of divalent cations when the vesicles contain choline as well as amino phospholipids. This effect is substantial only when the level of cholesterol exceeds the level of choline phospholipids in the vesicles. These results may have significance for the fusion of certain cellular membranes in mammalian cells, whose cytoplasmic faces have lipid compositions very similar to those of the vesicles examined in this study.

Fusion of Sindbis virus with model membranes containing phosphatidylethanolamine: implications for protein-induced membrane fusion

The pH-induced fusion of Sindbis virus with model lipid membranes containing phosphatidylethanolamine has been studied using a quantitative fluorescence technique. The headgroup and acyl chain domains of the lipids have been altered systematically to determine their effect on fusion. Unsaturated phosphatidylethanolamines (PE) have been found to promote fusion, either by themselves, or in combination with phosphatidylcholines (PC). Cholesterol added to a mixture of unsaturated PE and PC was also shown to increase the extent of viral fusion. The results of these studies have been interpreted in terms of a tentative model for the molecular aspects of the target membrane which are necessary for viral fusion. In this model, the target membrane must have a sufficiently-sized domain containing poorly hydrated lipids which are capable of existing in a non-bilayer arrangement.

Molecular mechanisms of membrane fusion: steps during phospholipid and exocytotic membrane fusion

Exocytosis is considered as four separate steps: adhesion, fusion/pore formation, pore widening, and content discharge. Experiments on both synthetic and natural membranes are presented to show each of these steps. Major differences are seen in the two fusing systems. These differences are discussed in terms of molecular mechanisms of fusion.

Divalent cation induced fusion and lipid lateral segregation in phosphatidylcholine-phosphatidic acid vesicles

The interactions of unilamellar vesicles containing phosphatidylcholine (PC) and phosphatidic acid (PA) in the presence of calcium and magnesium were examined by fluorometric assays of vesicle lipid mixing, contents mixing, and contents leakage and by spray-freezing freeze-fracture electron microscopy. These results were correlated with calorimetric and fluorometric measurements of divalent cation induced lateral segregation of lipids in these vesicles under comparable conditions. PA-PC vesicles in the presence of calcium show a rapid but limited intermixing of vesicle lipids and contents, the extent of which increases as the vesicle size decreases or the PA content increases. Calcium produces massive aggregation and efficient mixing of the contents of vesicles containing high proportions of dioleoyl-PA or egg PA, but vesicle coalescence in the latter case is followed rapidly by vesicle collapse and massive leakage of contents. The effects of magnesium are similar for vesicles of very high PA content. However, in the presence of magnesium, vesicles containing lower amounts of PA exhibit "hemifusion", a mode of interaction in which vesicles aggregate and mix approximately 50% of their lipids, apparently representing the lipids of the outer monolayer of each vesicle, without significant mixing of vesicle contents or collapse of the vesicles. Fluorometric measurements of lipid lateral segregation demonstrate that lateral redistribution of lipids in PA-PC vesicles begins at submillimolar concentrations of divalent cations and shows no abrupt change at the "threshold" divalent cation concentration, above which coalescence of vesicles is observed. By correlating calorimetric and fluorometric measurements of lipid lateral segregation and mixing of vesicle components, we can demonstrate that lipid segregation is at least strongly correlated with calcium-promoted coalescence of PA-PC vesicles and is essential to the magnesium-promoted interactions of vesicles of low PA contents.

Morphological responses to calcium-induced interaction of phosphatidylserine-containing vesicles

Structural changes in phospholipid vesicles made of dioleylphosphatidylethanolamine (DOPE)/bovine phosphatidylserine (PS) (1/1, 3/1, 10/1) or of egg phosphatidylcholine (PC)/PS (3/1) and exposed to calcium chloride for various times have been observed by means of video-enhanced light microscopy and freeze-fracture electron microscopy. Calcium induces the formation of large, smooth double-bilayer diaphragms as the spherical vesicles adhere to and deform each other. No subsequent changes are seen with PC/PS vesicles. DOPE/PS vesicles respond to the resultant stress, with about equal probability, by either fusing, through diaphragm rupture, or deflating, by way of volume loss through intact bilayers, even when they contain up to 400 mM sucrose. The diaphragm areas only rarely show the structural destabilization necessary for fusion. The final state is lipid segregated into DOPE hexagonal and Ca-PS lamellar bulk phases with the exclusion of most of the vesicle contents. Results with these and pure PS vesicles studied earlier indicate that the early response of vesicles to calcium chloride is determined by the competing rates at which mechanical stress (bilayer tension and intravesicular pressure) builds up as the vesicles adhere and flatten against each other, and is relieved by vesicle fusion or by volume loss. We attribute the qualitatively different responses of these three lipid systems to their measured differences in adhesion energies and consequent rate of build-up of mechanical stress. Yield to that stress for any one of these lipid systems is not a unique sequence of morphological changes, and so it remains obscure how such a stochastic process could be used in the controlled process of cellular fusion.

Dolichyl phosphate induces non-bilayer structures, vesicle fusion and transbilayer movement of lipids: a model membrane study

The effect of dolichol and dolichyl phosphate on fusion between large unilamellar vesicles comprised of 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) was studied using a fluorescence resonance energy transfer assay. The influence of dolichyl phosphate on the transbilayer movement of DOPC in multilamellar vesicles (MLV) and large unilamellar vesicles (LUV) composed of DOPC and DOPE (1:2) was investigated by using the phosphatidylcholine-specific transfer protein. 31P-NMR and freeze-fracture electron microscopy were employed to study the macroscopic organization of DOPC and DOPE containing model membranes in the absence or presence of dolichyl phosphate. The results indicate that both dolichol and dolichyl phosphate enhance vesicle fusion in a comparable and concentration-dependent way; the amount of exchangeable PC from MLVs is increased by dolichyl phosphate, probably as a result of fusion processes; dolichyl phosphate destabilizes the bilayer organization in MLVs comprised of DOPE and DOPC, resulting in the formation of hexagonal (HII) phase and 'lipidic' particles.

Solubility of palmitoyl-coenzyme A in acyltransferase assay buffers containing magnesium ions

The solubility of palmitoyl-CoA is strongly affected by Mg2+ concentrations commonly used in acyltransferase reactions. In 0.10 M Tris-HCl buffer at pH 7.4 or 8.5, all of the palmitoyl-CoA in 10 microM solutions and 90% of the palmitoyl-CoA in 100 microM solutions are precipitated by 1 mM Mg2+. In 0.05 M phosphate at pH 7.4, and in 0.10 M Tris-HCl containing 0.4 M KCl, the substrate remains soluble at Mg2+ concentrations below 4-5 mM. Above 5 mM Mg2+, palmitoyl-CoA is insoluble in all of these buffers. Substrate solubility could therefore be a limiting factor when free Mg2+ and fatty acyl-CoAs are present together during acyltransferase assays.

Area per molecule and distribution of water in fully hydrated dilauroylphosphatidylethanolamine bilayers

The area per lipid molecule for fully hydrated dilauroylphosphatidylethanolamine (DLPE) has been obtained in both the gel and liquid-crystalline states by combining wide-angle X-ray diffraction, electron density profiles, and previously published dilatometry results [Wilkinson, D. A., & Nagle, J. F. (1981) Biochemistry 20, 187-192]. The molecular area increases from 41.0 +/- 0.2 to 49.1 +/- 1.2 A2 upon melting from the gel to liquid-crystalline phase. The thickness of the bilayer, as measured from the electron density profiles, decreases about 4 A upon melting, from 45.2 +/- 0.3 to 41.0 +/- 0.6 A. A somewhat unexpected result is that the fluid layer between fully hydrated bilayers is the same in both gel and liquid-crystalline phases and is only about 5 A thick. From these data, plus the volume of the anhydrous DLPE molecule, it is possible to determine the number of water molecules per lipid and their approximate distribution relative to the lipid molecule. Our analysis shows that there are about 7 and 9 waters per DLPE molecule in the gel and liquid-crystalline phases, respectively. About half of the water is located in the fluid space between adjacent bilayers, and the remaining waters are intercalated into the bilayer, presumably in the head group region. There are significantly fewer water molecules in the fluid spaces between DLPE bilayers than in the fluid spaces in gel- or liquid-crystalline-phase phosphatidylcholine bilayers. This small fluid space in PE bilayers could arise from interbilayer hydrogen bond formation through the water molecules or electrostatic interactions between the amine and phosphate groups on apposing bilayers.

Epidermal permeability barrier: transformation of lamellar granule-disks into intercellular sheets by a membrane-fusion process, a freeze-fracture study

Freeze-fracture replication of lamellar granules and intercellular sheets of the horny layer in mouse, chicken, and snake epidermis reveals a pattern of serial fracture faces which is highly suggestive of polar lipids in a bilayer configuration. The occurrence of alternating wide and narrow fracture faces separated by intervening steps supports the view that epidermal barrier bilayers display lipid asymmetry similar to membranes. Within the lamellar granules, bilayers arrange to form disks which in fact are equivalent to flattened unilamellar liposomes. Stacking of the disks in turn gives rise to the lamellar pattern. After exocytosis into the intercellular space, the disks are arranged parallel to the cell membranes. In tangentially fractured specimens, the cleavage plane jumps back and forth from the plasma membrane to a disk-bilayer, thereby giving rise to the known phenomenon of EF-ridges (on the extracellular fracture face) and PF-grooves (in the plasmatic fracture face) which both represent the level of the plasma membrane sur- or subjacent to the aisles between disks. Concomitantly with the upward movement of the keratinocytes, the ridges and grooves become narrower until they fade away by the second or third cell layer of the stratum corneum. This phenomenon is explained by the fusion of adjacent disks at their highly curved brims due to a mechanism similar to the process of membrane fusion which causes the formation of wide, uninterrupted sheets.

Membrane fusion: lipid vesicles as a model system

In many cellular functions the process of membrane fusion is of vital importance. It occurs in a highly specific and strictly controlled fashion. Proteins are likely to play a key role in the induction and modulation of membrane fusion reactions. Aimed at providing insight into the molecular mechanisms of membrane fusion, numerous studies have been carried out on model membrane systems. For example, the divalent-cation induced aggregation and fusion of vesicles consisting of negatively charged phospholipids, such as phosphatidylserine (PS) or cardiolipin (CL), have been characterized in detail. It is important to note that these systems largely lack specificity and control. Therefore conclusions derived from their investigation can not be extrapolated directly to a seemingly comparable counterpart in biology. Yet, the study of model membrane systems does reveal the general requirements of lipid bilayer fusion. The most prominent barrier to molecular contact between two apposing bilayers appears to be due to the hydration of the polar groups of the lipid molecules. Thus, dehydration of the bilayer surface and fluctuations in lipid packing, allowing direct hydrophobic interactions, are critical to the induction of membrane fusion. These membrane alterations are likely to occur only locally, at the site of intermembrane contact. Current views on the way membrane proteins may induce fusion under physiological conditions also emphasize the notion of local surface dehydration and perturbation of lipid packing, possibly through penetration of apolar amino acid segments into the hydrophobic membrane interior.

Inverted micellar intermediates and the transitions between lamellar, cubic, and inverted hexagonal lipid phases. I. Mechanism of the L alpha----HII phase transitions

A model for the thermotropic transitions between lamellar (L alpha) and inverted hexagonal (HII) phases is developed. According to this model, the first structures to form during the L alpha----HII transition are inverted micellar intermediates (IMI). The structure, formation rates, and half-lives of IMI ("lipidic particles") were described previously. IMI coalesce in the planes between apposed bilayers to form two types of HII phase precursors. The first is a monolayer-encapsulated HII tube (RMI), which forms via coalescence of IMI in pearl-string fashion. These structures have been proposed previously based on electron microscopic evidence. I show that if only RMI form, L alpha in equilibrium HII transitions cannot occur on observed time scales (faster than seconds). I propose that a second type of intermediate, a line defect (LD), forms as well. LD should form via IMI-IMI coalescence in significant numbers, and elongate rapidly into structures consisting of two apposed halves of HII tubes. Transitions via LD can occur in less than seconds, the time depending on the fraction of IMI-IMI coalescence events producing LD and the number of IMI per unit of bilayer area. Hysteresis in the phase transition temperature may be due to the difference in water content of the two phases and their low water permeabilities. The model is in qualitative agreement with morphological, NMR, and x-ray diffraction data on phospholipid systems. The results are relevant to IMI-mediated interactions between unilamellar bilayer vesicles, and to the structure of inverted cubic phases observed in some phospholipid systems. These will be discussed in subsequent publications (D. P. Siegel, manuscript in preparation).

Aliphatic aldehydes promote myelin basic protein-induced fusion of phospholipid vesicles

Myelin basic protein induces slow and limited fusion of phospholipid vesicles composed of a mixture of phosphatidylcholine and phosphatidylethanolamine. Addition of palmitoyl aldehyde to these vesicles dramatically increases their ability to fuse in the presence of myelin basic protein. Compared to aliphatic aldehydes, fatty acids are much less potent promoters of myelin basic protein-induced membrane fusion. The ability of aliphatic aldehydes to promote myelin basic protein-induced membrane fusion may be of relevance to myelin structure and function and, particularly, to the pathology of demyelinating diseases such as multiple sclerosis.

Promotion and inhibition of vesicle fusion by polylysine

Polylysine induced rapid aggregation of large unilamellar vesicles composed of phosphatidylcholine-cardiolipin (1:1 molar ratio) but not their fusion. Application of the terbium-dipicolinic acid fusion assay showed that addition of polylysine at nanomolar concentrations enabled a significant lowering of the Ca2+ threshold concentration for vesicle fusion from 9 to 1 mM. Analysis of the kinetics of fusion with a mass-action kinetic model showed that polylysine enhanced significantly the rate of aggregation but affected only slightly the rate of fusion per se. Maximal enhancement of overall fusion rates occurred at a charge ratio (polylysine/cardiolipin) of about 0.5. At larger polylysine concentrations, e.g., at charge ratios greater than 3, polylysine inhibited vesicle fusion.

Interaction of clathrin with liposomes: pH-dependent fusion of phospholipid membranes induced by clathrin

Clathrin, dissociated from coated vesicles of bovine brain and purified by gel chromatography, was found to interact with the lipid bilayer as shown by the spontaneous release of encapsulated fluorescent dye in liposome. Clathrin-induced dye release was enhanced at acidic pH in phosphatidylserine-containing vesicles. A strong correlation between dye release and fusion of liposomes was observed. In general, when there was a fast release of encapsulated dye induced by clathrin, a pH-dependent, clathrin-induced fusion was observed. Clathrin did not induce either dye release or fusion of egg phosphatidylcholine liposomes. The self-association of clathrin at low pH diminished the fusogenic activity. Fusion induced by clathrin at low pH could be stopped at pH above 5.0 and resumed by lowering the pH below 5.0. This suggests that the interaction of clathrin with phospholipid membranes can be regulated by pH.

Fusogenic capacities of divalent cations and effect of liposome size

The initial kinetics of divalent cation (Ca2+, Ba2+, Sr2+) induced fusion of phosphatidylserine (PS) liposomes, LUV, is examined to obtain the fusion rate constant, f11, for two apposed liposomes as a function of bound divalent cation. The aggregation of dimers is rendered very rapid by having Mg2+ in the electrolyte, so that their subsequent fusion is rate limiting to the overall reaction. In this way the fusion kinetics are observed directly. The bound Mg2+, which by itself is unable to induce the PS LUV to fuse, is shown to affect only the aggregation kinetics when the other divalent cations are present. There is a threshold amount of bound divalent cation below which the fusion rate constant f11 is small and above which it rapidly increases with bound divalent cation. These threshold amounts increase in the sequence Ca2+ less than Ba2+ less than Sr2+, which is the same as found previously for sonicated PS liposomes, SUV. While Mg2+ cannot induce fusion of the LUV and much more bound Sr2+ is required to reach the fusion threshold, for Ca2+ and Ba2+ the threshold is the same for PS SUV and LUV. The fusion rate constant for PS liposomes clearly depends upon the amount and identity of bound divalent cation and the size of the liposomes. However, for Ca2+ and Ba2+, this size dependence manifests itself only in the rate of increase of f11 with bound divalent cation, rather than in any greater intrinsic instability of the PS SUV. The destabilization of PS LUV by Mn2+ and Ni2+ is shown to be qualitatively distinct from that induced by the alkaline earth metals.

On the correlation between HII phase and the contact-induced destabilization of phosphatidylethanolamine-containing membranes

The abundance of phosphatidylethanolamine (PtdEtn) in biological membranes and the capacity of this lipid to sustain nonbilayer structures have been promoted as evidence for a role of PtdEtn in biological fusion processes. To date there has been no direct evidence of a connection between the kinetics of bilayer destabilization and the polymorphism accessible to PtdEtn. We have developed a model system to examine this point directly using the proton-induced destabilization of PtdEtn/cholesterylhemisuccinate unilamellar liposomes. We find that the initial rate of bilayer mixing rapidly increases with temperature and reaches a maximal level just below the HII-phase transition temperature. The leakage from these liposomes rapidly increases, both in rate and extent, within the HII-phase transition temperature range. Of an even greater significance is that at no temperature is there any mixing of aqueous contents within the liposomes. Thus, these lipids can begin to undergo the lamellar- to HII-phase transition at the stage of two apposed liposomes. However, the nonbilayer structures formed do not cause fusion--i.e., the concomitant mixing of aqueous contents.

Proton-induced fusion of oleic acid-phosphatidylethanolamine liposomes

Liposomes composed of oleic acid and phosphatidylethanolamine (3:7 mole ratio) aggregate, become destabilized, and fuse below pH 6.5 in 150 mM NaCl. Fusion is monitored by (i) the intermixing of internal aqueous contents of liposomes, utilizing the quenching of aminonaphthalene-3,6,8-trisulfonic acid (ANTS) by N,N'-p-xylylenebis(pyridinium bromide) (DPX) encapsulated in two separate populations of vesicles, (ii) a resonance energy transfer assay for the dilution of fluorescent phospholipids from labeled to unlabeled liposomes, (iii) irreversible changes in turbidity, and (iv) quick-freezing freeze-fracture electron microscopy. Destabilization is followed by the fluorescence increase caused by the leakage of coencapsulated ANTS/DPX or of calcein. Ca2+ and Mg2+ also induce fusion of these vesicles at 3 and 4 mM, respectively. The threshold for fusion is at a higher pH in the presence of low (subfusogenic) concentrations of these divalent cations. Vesicles composed of phosphatidylserine/phosphatidylethanolamine or of oleic acid/phosphatidylcholine (3:7 mole ratio) do not aggregate, destabilize, or fuse in the pH range 7-4, indicating that phosphatidylserine and phosphatidylcholine cannot be substituted for oleic acid and phosphatidylethanolamine, respectively, for proton-induced membrane fusion. Freeze-fracture replicas of oleic acid/phosphatidylethanolamine liposomes frozen within 1 s of stimulation with pH 5.3 display larger vesicles and vesicles undergoing fusion, with membrane ridges and areas of bilayer continuity between them. The construction of pH-sensitive liposomes is useful as a model for studying the molecular requirements for proton-induced membrane fusion in biological systems and for the cytoplasmic delivery of macromolecules.

H+- and Ca2+-induced fusion and destabilization of liposomes

A new liposome fusion assay has been developed that monitors the mixing of aqueous contents at neutral and low pH. With this assay we have investigated the ability of H+ to induce membrane destabilization and fusion. The assay involves the fluorophore 1-aminonaphthalene-3,6,8-trisulfonic acid (ANTS) and its quencher N,N'-p-xylylenebis(pyridinium bromide) (DPX). ANTS is encapsulated in one population of liposomes and DPX in another, and fusion results in the quenching of ANTS fluorescence. The results obtained with the ANTS/DPX assay at neutral pH give kinetics for the Ca2+-induced fusion of phosphatidylserine large unilamellar vesicles (PS LUV) that are very similar to those obtained with the Tb3+/dipicolinic acid (DPA) assay [Wilschut, J., & Papahadjopoulos, D. (1979) Nature (London) 281, 690-692]. ANTS fluorescence is relatively insensitive to pH between 7.5 and 4.0. Below pH 4.0 the assay can be used semiquantitatively by correcting for quenching of ANTS due to protonation. For PS LUV it was found that, at pH 2.0, H+ by itself causes mixing of aqueous contents, which makes H+ unique among the monovalent cations. We have shown previously that H+ causes a contact-induced leakage from liposomes composed of phosphatidylethanolamine and the charged cholesteryl ester cholesteryl hemisuccinate (CHEMS) at pH 5.0 or below, where CHEMS becomes protonated. Here we show that H+ causes lipid mixing in this pH range but not mixing of aqueous contents. This result affirms the necessity of using both aqueous space and lipid bilayer assays to comprehend the fusion event between two liposomes.

Sodium-induced aggregation of phosphatidic acid and mixed phospholipid vesicles

Sodium-induced aggregations of sonicated vesicles prepared from synthetic phosphatidic acid and from its 1:1 mixtures with synthetic phosphatidylethanolamine and phosphatidylcholine were studied by turbidimetric measurements. The aggregation reactions were almost completely reversible on change in the Na+ concentration, pH or temperature. The threshold concentrations of Na+ for aggregations of pure dipalmitoylphosphatidic acid vesicles and mixed dipalmitoylphosphatidylethanolamine- and dimyristoylphosphatidylcholine-dipalmitoylphosphatidic acid vesicles were found to be 200, 310 and 550 mM, respectively, at 25 degrees C and pH 7.2. The hydrocarbon chain lengths of phosphatidic acid and phosphatidylethanolamine had little effect on the threshold concentrations. The threshold concentrations for phospholipid vesicles composed of phosphatidic acid alone or its 1:1 mixture with phosphatidylethanolamine were changed by varying either the pH or temperature, while that for phosphatidylcholine-phosphatidic acid vesicles was almost independent of the pH and temperature, implying that aggregation of the latter vesicles is induced by a somewhat different mechanism.

Temperature dependence of divalent cation induced fusion of phosphatidylserine liposomes: evaluation of the kinetic rate constants

The effect of temperature and divalent cation binding (Ca2+, Sr2+, Ba2+) on the kinetic rate constants of aggregation and fusion of large phosphatidylserine liposomes is measured for the first time. Fusion is monitored by the Tb3+/dipicolinate assay. Fusion rate constants increase with temperature (15-35 degrees C) in a roughly linear fashion. These rate constants are not otherwise sensitive to whether the temperature is above or below the phase transition temperature of the Ba2+ or Sr2+ complex of phosphatidylserine, as measured by differential scanning calorimetry. Hence, the isothermal transition of the acyl chains from liquid-crystalline to gel phase induced by the cations is not the driving force of the initial fusion event. The aggregation rate constants increase with temperature, and it is the temperature dependence of the energetics of close approach of the liposomes which underlies this increase. On the other hand, the aggregation becomes more reversible at higher temperatures, which has also been observed with monovalent cation induced liposome aggregation where there is no fusion. Calculations on several cases show that the potential energy minimum holding the liposome dimer aggregates together is approximately 5-6 kT deep. This result implies that the aggregation step is highly reversible; i.e., if fusion were not occurring, no stable aggregates would form.

Modulation of membrane fusion by membrane fluidity: temperature dependence of divalent cation induced fusion of phosphatidylserine vesicles

We have investigated the temperature dependence of the fusion of phospholipid vesicles composed of pure bovine brain phosphatidylserine (PS) induced by Ca2+ or Mg2+. Aggregation of the vesicles was monitored by 90 degrees light-scattering measurements, fusion by the terbium/dipicolinic acid assay for mixing of internal aqueous volumes, and release of vesicle contents by carboxyfluorescein fluorescence. Membrane fluidity was determined by diphenylhexatriene fluorescence polarization measurements. Small unilamellar vesicles (SUV, diameter 250 A) or large unilamellar vesicles (LUV, diameter 1000 A) were used, and the measurements were done in 0.1 M NaCl at pH 7.4. The following results were obtained: (1) At temperatures (0-5 degrees C) below the phase transition temperature (Tc) of the lipid, LUV (PS) show very little fusion in the presence of Ca2+, although vesicle aggregation is rapid and extensive. With increasing temperature, the initial rate of fusion increases dramatically. Leakage of contents at the higher temperatures remains limited initially, but subsequently complete release occurs as a result of collapse of the internal aqueous space of the fusion products. (2) SUV (PS) are still in the fluid state down to 0 degree C, due to the effect of bilayer curvature, and fuse rapidly in the entire temperature range from 0 to 35 degrees C in the presence of Ca2+. The initial rate of leakage is low relative to the rate of fusion. At higher temperatures (15 degrees C and above), subsequent collapse of the vesicles' internal space causes complete release.(ABSTRACT TRUNCATED AT 250 WORDS)

pH-dependent fusion of liposomes using titratable polycations

Polylysine promoted extensive membrane mixing of liposomes only if the buffer pH was below the pKa of the lysyl residues. This observation suggested that fusion could be regulated in a physiological pH range if the homopolymer of L-histidine was substituted as fusogen. Microgram quantities of polyhistidine were added to liposomes composed of soybean phospholipids, or to defined phospholipid-cholesterol mixtures which simulate the lipid composition of plasma membranes. A quantitative resonance energy transfer assay determined the extent of lipid phase mixing related to fusion. No fusion was detected at pH 7.4, but when the pH was lowered to 6.5 or below, fusion was rapid and substantial. The extent of membrane mixing increased with progressive acidification of the vesicle-fusogen suspension. The charge density of each polyhistidine molecule, not the total cationic charge per vesicle, influenced the extent of fusion. The kinetics of the fusion reaction were rapid, as membrane mixing was completed within 1 min. If the vesicle suspension was acidified before fusogen addition, the rate of membrane mixing slowed 4-fold. This, as well as a slight increase in light scattering noted whenever polyhistidine was added at pH 7.4, suggests an enhancement of fusion kinetics by preaggregation of vesicles at neutral pH. The lipid composition, regulation of membrane mixing by pH in a physiological range, and rapid kinetics suggest that this model of liposome fusion may be pertinent to understanding some biological fusion events.

pH-sensitive liposomes mediate cytoplasmic delivery of encapsulated macromolecules

Negatively charged liposomes are endocytosed by the coated vesicle system and accumulate in acidic intracellular vesicles. Liposomes that become unstable at acidic pH improve cytoplasmic delivery of membrane-impermeant macromolecules such as calcein (CAL) and FITC dextran (18 or 40 kDa). Oleic acid (OA): phosphatidylethanolamine (PE) (3:7 mole ratio) liposomes become permeable to CAL at pH less than 7.0. Control liposomes of phosphatidylserine:PE or OA:phosphatidylcholine are stable at pH 4-8. OA:PE liposomes promote cytoplasmic delivery of encapsulated CAL to CV-1 cells, as evidenced by the emergence of diffuse, cytoplasmic CAL fluorescence. Delivery requires metabolic energy and is partially inhibited by chloroquine or monensin, which raise the pH of intracellular vesicles.

Effect of lipid structure on the capacity of myelin basic protein to alter vesicle properties: potent effects of aliphatic aldehydes in promoting basic protein-induced vesicle aggregation

The capacity of myelin basic protein or of poly-L-lysine to promote leakage of carboxyfluorescein from vesicles or the aggregation of vesicles was studied. The vesicles were composed of phosphatidylcholine as the sole or major lipid component. Addition of 10% sphingomyelin, 10% phosphatidylglycerol, 10% egg or bovine brain phosphatidylethanolamine, or 30% dodecanal had relatively little effect on the extent of carboxyfluorescein release in the presence of either myelin basic protein or poly-L-lysine. In contrast with these results, the extent of vesicle aggregation was very sensitive to lipid composition. Addition of 10% phosphatidylglycerol induced more aggregation than the other phospholipids tested. Admixing 10% of a partially degraded sample of bovine brain phosphatidylethanolamine also led to a large amount of aggregation induced by the myelin basic protein. This latter aggregation appeared more specific for the basic protein, as it occurred to a much smaller extent with poly-L-lysine. In general, the effects of the myelin basic protein on either carboxyfluorescein release or vesicle aggregation were similar to, although somewhat greater than, that of poly-L-lysine. The aggregation of vesicles containing degradation products of phosphatidylethanolamine can be ascribed largely to the presence of aliphatic aldehydes. The effect of aliphatic aldehydes was specific in that the aliphatic alcohol, hexadecanol, or the short-chain aldehydes, acetaldehyde or butyraldehyde, did not promote myelin basic protein-induced vesicle aggregation. In addition, poly-L-lysine was less effective than the basic protein in aggregating vesicles containing aliphatic aldehydes. (ABSTRACT TRUNCATED AT 250 WORDS)

Calcium interactions in pulmonary surfactant

The surfactant properties of natural bovine pulmonary surfactant, its lipid extracts and acetone precipitates of lipid extracts have been examined with an artificial alveolus model, the pulsating-bubble surfactometer. At bulk concentrations of 0.4% (wt./vol.) phospholipid in saline, all three preparations exhibited surfactant activity, i.e., were capable of reducing the surface tension of the pulsating bubble to approx. 27 dynes/cm at maximum bubble radius and to near zero at minimum bubble radius. At a concentration of 0.1% (wt./vol.) in saline, only natural surfactant was effective. Acetone-precipitated surfactant at 0.1% (wt./vol.) achieved these criteria in the presence of 5 mM calcium, but 15-20 mM calcium was required to restore the surfactant activity of lipid extract surfactant. Chemical analysis revealed that lipid extraction decreases the protein content but does not alter the endogenous calcium levels. A calcium requirement for natural surfactant could only be demonstrated after repeated treatment with chelators for divalent cations. Surfactant activity was restored by low levels of calcium or high levels of magnesium. Paradoxically, a calcium requirement could not be demonstrated by treating acetone-precipitated lipid extract with chelators. The subtle differences noted between natural, lipid extract and acetone-precipitated lipid extract surfactant with the pulsating-bubble assay show that the latter preparations do not represent simplified model systems for the natural product.

Proton-induced membrane fusion. Role of phospholipid composition and protein-mediated intermembrane contact

Glycolipid-phospholipid vesicles containing phosphatidate and phosphatidylethanolamine were found to undergo proton-induced fusion upon acidification of the suspending medium from pH 7.4 to pH 6.5 or lower, as determined by an assay for lipid intermixing based on fluorescence resonance energy transfer. Lectin-mediated contact between the vesicles was required for fusion. Incorporation of phosphatidylcholine in the vesicles inhibited proton-induced fusion. Vesicles in which phosphatidate was replaced by phosphatidylserine underwent fusion only when pH was reduced below 4.5, while no significant fusion occurred (pH greater than or equal to 3.5) when the anionic phospholipid was phosphatidylinositol. It is suggested that partial protonation of the polar headgroup of phosphatidate and phosphatidylserine, respectively, causes a sufficient reduction in the polarity and hydration of the vesicle surface to trigger fusion at sites of intermembrane contact.

Lectins facilitate calcium-induced fusion of phospholipid vesicles containing glycosphingolipids

Ca2+-induced fusion of phospholipid vesicles containing globoside (GL-4) or disialoganglioside (GDla) is several-fold slower than the fusion of the pure phospholipid vesicles. Lectins specific for these glycosphingolipids, soybean agglutinin and wheat germ agglutinin, respectively, enhance the rate of fusion when added to the vesicle suspension before the introduction of Ca2+. The enhancement depends on the lectin concentration and the time of preincubation with the lectin. We propose that lectins facilitate membrane fusion by inducing intermembrane contact, which is the first step in the overall process of membrane fusion, or by laterally phase separating the inhibitory glycolipids.

Free energy potential for aggregation of mixed phosphatidylcholine/phosphatidylserine lipid vesicles in glucose polymer (dextran) solutions

The energetics of lipid vesicle-vesicle aggregation in dextran (36,000 mol wt) solutions have been studied with the use of micromechanical experiments. The affinities (free energy reduction per unit area of contact) for vesicle-vesicle aggregation were determined from measurements of the tension induced in an initially flaccid vesicle membrane as it adhered to another vesicle. The experiments involved controlled aggregation of single vesicles by the following procedure: two giant (approximately 20 micron diam) vesicles were selected from a chamber on the microscope stage that contained the vesicle suspension and transferred to a second chamber that contained a dextran (36,000 mol wt) salt solution (120 mM); the vesicles were then maneuvered into position for contact. One vesicle was aspirated with sufficient suction pressure to create a rigid sphere outside the pipette; the other vesicle was allowed to spread over the rigid vesicle surface. The aggregation potential (affinity) was derived from the membrane tension vs. contact area. Vesicles were formed from mixture of egg lecithin (PC) and phosphatidylserine (PS). For vesicles with a PC/PS ratio of 10:1, the affinity showed a linear increase with concentration of dextran; the values were on the order of 10(-1) ergs/cm2 at 10% by weight in grams. Similarly, pure PC vesicle aggregation was characterized by an affinity value of 1.5 X 10(-1) ergs/cm2 in 10% dextran by weight in grams. In 10% by weight in grams solutions of dextran, the free energy potential for vesicle aggregation decreased as the surface charge (PS) was increased; the affinity extrapolated to zero at a PC/PS ratio of 2:1. When adherent vesicle pairs were transferred into a dextran-free buffer, the vesicles did not spontaneously separate. They maintained adhesive contact until forceably separated, after which they would not read here. Thus, it appears that dextran forms a "cross-bridge" between the vesicle surfaces.

pH-induced destabilization of phosphatidylethanolamine-containing liposomes: role of bilayer contact

The mechanism of pH-induced destabilization of liposomes composed of phosphatidylethanolamine and a charged cholesteryl ester was studied by following the release of encapsulated aqueous contents. The kinetics of release were measured continuously by using the water-soluble fluorophore 8-aminonaphthalene-1,3,6-trisulfonic acid in combination with the water-soluble quencher p- xylylenebis (pyridinium) bromide. With this fluorescence assay, release of contents from liposomes composed of phosphatidylethanolamine and cholesteryl hemisuccinate was shown to be a function of pH, ratio of phosphatidylethanolamine to cholesteryl hemisuccinate, and acyl chain composition of the phosphatidylethanolamine. Leakage was very slow at pH 5.5 and increased dramatically with decreasing pH down to 4.0. Replacing phosphatidylethanolamine by phosphatidylcholine eliminated the effect of pH on leakage. Analysis of the kinetics of release by a mass action model demonstrated that bilayer destabilization and leakage occur subsequent to aggregation. The requirement of bilayer contact for destabilization has been found previously for acidic phospholipid bilayers in the presence of divalent cation and for saturated phosphatidylcholine bilayers below the isothermal phase transition temperature. The phosphatidylethanolamine-containing bilayers examined here satisfy the same requirement.

Inverted micellar structures in bilayer membranes. Formation rates and half-lives

Two sorts of inverted micellar structures have previously been proposed to explain morphological and 31P-NMR observations of bilayer systems. These structures only form in systems with components that can adopt the inverse hexagonal (HII) phase. LIP (lipidic particles) are intrabilayer structures, whereas IMI (inverted micellar intermediates) are structures that form between apposed bilayers. Here, we calculate the formation rates and half-lives of these structures to determine which (or if either) of these proposed structures is a likely explanation of the data. Calculations for the egg phosphatidylethanolamine and the Ca+-cardiolipin systems show that IMI form orders of magnitude faster than LIP, which should form slowly, if at all. This result is probably true in general, and indicates that "lipidic particle" electron micrograph images probably represent interbilayer structures, as some have previously proposed. It is shown here that IMI are likely intermediates in the lamellar----HII phase transitions and in the process of membrane fusion in some systems. The calculated formation rates, half-lives, and vesicle-vesicle fusion rates are in agreement with this observation.