Studies on membrane fusion. II. Induction of fusion in pure phospholipid membranes by calcium ions and other divalent metals

The effect of divalent metals on the interaction and mixing of membrane components in vesicles prepared from acidic phospholipids has been examined using freeze-fracture electron microscopy and differential scanning calorimetry. Ca²⁺, and to a certain extent Mg²⁺, induce extensive mixing of vesicle membrane components and drastic structural rearrangements to form new membranous structures. In contrast to the mixing of vesicle membrane components in the absence of Ca²⁺ described in the accompanying paper which occurs via diffusion of lipid molecules between vesicles, mixing of membrane components induced by Ca²⁺ or Mg²⁺ results from true fusion of entire vesicles. There appears to be a “threshold” concentration at which Ca²⁺ and Mg²⁺ become effective in inducing vesicle fusion and the threshold concentration varies for different acidic phospholipid species. Different phospholipids also vary markedly in their relative responsiveness to Ca²⁺ and Mg²⁺, with certain phospholipids being much more susceptible to fusion by Ca²⁺ than Mg²⁺. Vesicle fusion induced by divalent cations also requires that the lipids of the interacting membranes be in a “fluid” state ($\text{T > T}{}_{\mathrm{c}}$). Fusion of vesicle membranes by Ca²⁺ and Mg²⁺ does not appear to be due to simple electrostatic charge neutralization. Rather the action of these cations in inducing fusion is related to their ability to induce isothermal phase transitions and phase separations in phospholipid membranes. It is suggested that under these conditions membranes become transiently susceptible to fusion as a result of changes in molecular packing and creation of new phase boundaries induced by Ca²⁺ (or Mg²⁺).     

Effects of proteins on the thermotropic phase transitions of phospholipid membranes

A variety of proteins have been studied for their ability to interact and alter the thermotropic properties of phospholipid bilayer membranes as detected by differential scanning calorimeter. The proteins studied included: basic myelin protein (A1 protein), cytochrome c, major apoprotein of myelin proteolipid (N-2 apoprotein), gramicidin A, polylysine, ribonuclease and hemoglobin. The lipids used for the interactions were dipalmitoylphosphatidylcholine and dipalmitoylphosphatidylglycerol. The interactions were grouped in three categories each having very different effects on the phospholipid phase transition from solid to liquid crystalline. The calorimetric studies were also correlated with data from vesicle permeability and monolayer expansion.Ribonuclease and polylysine which exemplify group 1 interactions, show strong dependence on electrostatic binding. Their effects on lipid bilayers include an increase in the enthalpy of transition (ΔH) accompanied by either an increase or no change in the temperature of transition (T_c). In addition, they show minimal effects on vesicle permeability and monolayer expansion. It was concluded that these interactions represent simple surface binding of the protein on the lipid bilayer without penetration into the hydrocarbon region.Cytochrome c and Al protein, which exemplify group 2 interactions, also show a strong dependence on the presence of net negative charges on the lipid bilayers for their binding. In contrast to the first group, however, they induce a drastic decrease in both T_c and ΔH of the lipid phase transition. Furthermore, they induce a large increase in the permeability of vesicles and a substantial expansion in area of closely packed monolayers at the air-water interface. It was concluded that group 2 interactions represent surface binding followed by partial penetration and/or deformation of the bilayer.Group 3 interactions, shown by proteolipid apoprotein and gramicidin A, were primarily non-polar in character, not requiring electrostatic charges and not inhibited by salt and pH changes. They had no appreciable effect on the T_c but did induce a linear decrease in the magnitude of the ΔH, proportional to the percentage of protein by weight. Membranes containing 50% proteolipid protein still exhibited a thermotropic transition with a ΔH one half that of the pure lipid, and only a small diminution of the size of the cooperative unit. It was concluded that in this case the protein was embedded within the bilayer, associating with a limited number of molecules via non-polar interactions, while the rest of the bilayer was largely unperturbed.     

Annexin-mediated membrane fusion of human neutrophil plasma membranes and phospholipid vesicles.

Membrane fusion was studied using human neutrophil plasma membrane preparations and phospholipid vesicles approximately 0.15 microns in diameter and composed of phosphatidylserine and phosphatidylethanolamine in a ratio of 1 to 3. Liposomes were labeled with N-(7-nitrobenzo-2-oxa-1,3-diazol-4-yl (NBD) and lissamine rhodamine B derivatives of phospholipids. Apparent fusion was detected as an increase in fluorescence of the resonance energy transfer donor, NBD, after dilution of the probes into unlabeled membranes. 0.5 mM Ca2+ alone was sufficient to cause substantial fusion of liposomes with a plasma membrane preparation but not with other liposomes. Both annexin I and des(1-9)annexin I caused a substantial increase in the rate of fusion under these conditions while annexin V inhibited fusion. Fusion mediated by des(1-9)annexin I was observed at Ca2+ concentrations as low as approximately 5 microM, suggesting that the truncated form of this protein may be active at physiologically low Ca2+ concentrations. Trypsin treated plasma membranes were incapable of fusion with liposomes, suggesting that plasma membrane proteins may mediate fusion. Liposomes did not fuse with whole cells at any Ca2+ concentration, indicating that the cytoplasmic side of the membrane is involved. These results suggest that annexin I and unidentified plasma membrane proteins may play a role in Ca(2+)-dependent degranulation of human neutrophils.     

Reconstitution of membrane proteins. Spontaneous incorporation of integral membrane proteins into preformed bilayers of pure phospholipid

The spontaneous reconstitution of lipid-protein complexes was examined by mixing bacteriorhodopsin or UDP-glucuronosyltransferase with preformed, unilamellar bilayers of pure dimyristoylphosphatidylcholine. Spontaneous insertion of these proteins into vesicles of dimyristoylphosphatidylcholine was facilitated by resonicating the vesicles at 4 degrees C. The property of resonicated vesicles that led to spontaneous reconstitution could be annealed by melting the bilayers, which slowed down reconstitution. The overall process of reconstitution consisted, however, of two steps. There was an initial insertion of proteins into a small portion of vesicles followed by subsequent fusion between protein-free vesicles and vesicles containing lipid-protein complexes. The first step appeared to proceed rapidly in all vesicles in a gel phase, whether or not they were resonicated or whether or not resonicated vesicles were annealed. The rate of the second step was sensitive to these treatments. The membrane proteins also inserted into preformed vesicles in a liquid crystalline phase, but this step was slower than for vesicles in a gel phase. Fusion between protein-free and protein-containing vesicles in a liquid crystalline phase was extremely slow. The data show that the spontaneous insertion of pure membrane proteins into preformed vesicles can be a facile event and that the overall reconstitution of membrane proteins into preformed unilamellar vesicles may be simpler to achieve than has been appreciated.     

Fusion in phospholipid spherical membranes. I. Effect of temperature and lysolecithin.

A study concerning membrane contact and fusion phenomena was made for phospholipid spherical bilayer systems with respect to temperature. Specific temperatures were obtained for the spherical bilayer membranes of phosphatidyl choline (PC) and phosphatidyl serine (PS) which indicated a greater degree of membrane fusion and were designated Tf (the fusion temperature -- PC: 43 degrees C, PS: 38 degrees C). These temperatures were reduced by about 10 degrees C for the membranes incorporated with 20% lysophosphatidyl choline. The results of the contact and fusion observed in the spherical membranes are compared and discussed with the conductance characteristics of the PC and PS planar bilayer membranes as well as dissolution study on the phospholipid monolayers formed at the air/water interface with respect to temperature. Also, a possible molecular mechanism of membrane fusion is discussed in terms of the fluidity and instability of the membrane.     

Effects of local anesthetics on membrane properties I. Changes in the fluidity of phospholipid bilayers

The effect of the local anesthetic dibucaine on the solid to liquid-crystalline phase transition in phospholipid vesicles was studied by calorimetry and fluorescence polarization. The partition coefficient (> 3000) of dibucaine in the membranes of vesicles prepared from acidic phospholipids was more than 20 times higher than in neutral phospholipid membranes under the same conditions. Calorimetric measurements on vesicles prepared form acidic phospholipids (bovine brain phosphatidylserine; dipalmitoylphosphatidylglycerol) showed that dibucaine (1 · 10⁻⁴M) produced a significant reduction in the gel-liquid crystalline transition temperature (T_c). This fluidizing effect of dibucaine on acidic phospholipid membranes was even more marked in the presence of Ca²⁺. In contrast, dibucaine at the same concentration did not alter the T_c of neutral phospholipids (dipalmitoylphosphatidylcholine). Significant increase in the fluidity of neutral phospholipid membranes occurred only at higher dibucaine concentrations (2 · 10⁻³M. Measurements of the fluorescence polarization and lifetime of the probe, 1,6-diphenylhexatriene, in acidic phospholipid vesicles revealed that dibucaine (1 · 10⁻⁴M caused an increase in the probe rotation rate indicating an increase in the fluidity of the phospholipid membranes. A good correlation was obtained between fluorescence polarization data on dibucaine-induced changes in membrane fluidity and calorimetric measurements on vesicles of the same type.     

Perturbations of phospholipid head groups by membrane proteins in biological membranes and recombinants.

P-31 nuclear magnetic resonance (NMR) spin-lattice relaxation times (T1) have been used to probe the behavior of phospholipid head groups in the presence of membrane proteins. Measurements have been made on rabbit muscle sarcoplasmic reticulum and recombinants of the Ca2+ Mg2+ ATPase, rod outer segment disk membranes and recombinants of rhodopsin, and human erythrocyte ghosts and recombinants of human erythrocyte glycophorin. Recombined membranes with lipid/protein ratios greater than or equal to that found in biological membranes showed T1 behavior similar to the biological membranes and pure phosphatidylcholine. However, recombined membranes with a low lipid/protein ratio exhibited a T1 that was dramatically shorter than any of the other systems. Analysis of the relaxation mechanism and the factors contributing to it implicate a phospholipid head group conformation change at high protein content. It is suggested that this is due to trapping of phospholipid between proteins and is not the same phenomenon as motional restriction at the lipid-protein interface at higher lipid contents.     

Endotoxic activity of complexes of myristic acid and proteins.

Complexes of myristic acid and bovine serum albumin, myristic acid and concanavalin A, beta-hydroxymyristic acid and concanavalin A, or dimethyl myristamide and concanavalin A are lethal for male BALB/c mice treated with mithramycin. Prior treatment of mice with myristic acid-protein complexes renders the animals resistant to a dose of bacterial endotoxin that is lethal for untreated animals. Prior treatment of mice with bacterial endotoxin renders them resistant to a combination of mithramycin and a complex of myristic acid and bovine serum albumin or dimethyl myristamide and concanvalin A that is lethal for untreated animals. These data indicate that a fatty acid is an important functional component of the endotoxin toxophore.     

Applications of phospholipid bilayer nanodiscs in the study of membranes and membrane proteins

Phospholipid bilayer Nanodiscs are novel model membranes derived from high-density lipoprotein particles and have proven to be useful in studies of membrane proteins. Membrane protein enzymology has been hampered by the inherent insolubility of membrane proteins in aqueous environments and has necessitated the use of model membranes such as liposomes and detergent-stabilized micelles. Current model membranes display a polydisperse particle size distribution and can suffer from problems of inconsistency and instability. It is also unclear how well they mimic biological lipid bilayers. In contrast, Nanodiscs, the particle size of which is constrained by a coat of scaffold proteins, are relatively monodisperse, stable model membranes with a "nativelike" lipid bilayer. Nanodiscs have already been used to study a variety of membrane proteins, including cytochrome P450s, seven-transmembrane proteins, and bacterial chemoreceptors. These proteins are simultaneously monomerized, solubilized, and incorporated into the well-defined membrane environment provided by Nanodiscs. Nanodiscs may also provide useful insights into the thermodynamics and biophysics of biological membranes and binding of small molecules to membranes.     

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

We have investigated the contribution of various phospholipids to membrane fusion induced by divalent cations. Fusion was followed by means of a new fluorescence assay monitoring the mixing of internal aqueous contents of large (0.1 μm diameter) unilamellar liposomes. The rate and extent of fusion induced by Ca²⁺ in mixed phosphatidylserine/phosphatidylcholine vesicles were lower compared to those in pure phosphatidylserine vesicles. The presence of 50% phosphatidylcholine completely inhibited fusion, although the vesicles aggregated upon Ca²⁺ addition. When phosphatidylserine was mixed with phosphatidylethanolamine, however, rapid fusion could be induced by Ca²⁺ even in mixtures that contained only 25% phosphatidylserine. Phosphatidylethanolamine also facilitated fusion by Mg²⁺ which could not fuse pure phosphatidylserine vesicles. In phosphatidylserine/phosphatidylethanolamine/phosphatidylcholine mixtures, in which the phosphatidylcholine content was kept at 25%, phosphatidylethanolamine could not substitute for phosphatidylserine, and the fusogenic capacity of Mg²⁺ was abolished by the presence of merely 10% phosphatidylcholine. The initial rate of release of vesicle contents was slower than the rate of fusion in all the mixtures used. The presence of phosphate effected a considerable decrease in the threshold concentration of Ca²⁺ and also enhanced     

Promotion of acid-induced membrane fusion by basic peptides. Amino acid and phospholipid specificities

The ability of oligo- and polymers of the basic amino acids L-lysine, L-arginine, L-histidine and L-ornithine to induce lipid intermixing and membrane fusion among vesicles containing various anionic phospholipids has been investigated. Among vesicle consisting of either phosphatidylinositol or mixtures of phosphatidic acid and phosphatidylethanolamine rapid and extensive lipid intermixing, but not complete fusion, was induced at neutral pH by poly-L-ornithine or L-lysine peptides of five or more residues. When phosphatidylcholine was included in the vesicles, the lipid intermixing was severely inhibited. Such lipid intermixing was also much less pronounced among phosphatidylserine vesicles. Poly-L-arginine provoked considerable leakage from the various anionic vesicles and caused significantly less lipid intermixing than L-lysine peptides at neutral pH. When the addition of basic amino acid polymer was followed by acidification to pH 5-6, vesicle fusion was induced. Fusion was more pronounced among vesicles containing phosphatidylserine or phosphatidic acid than among those containing phosphatidylinositol, and occurred also with vesicles whose composition resembles that of cellular membranes (i.e., phosphatidylcholine/phosphatidylethanolamine/phosphatidylserine, 50:30:20, by mol). Liposomes with this composition are resistant to fusion by Ca2+ or by acidification after lectin-mediated contact. The tight interaction among vesicles at neutral pH, resulting in lipid intermixing, does not seem to be necessary for the fusion occurring after acidification, but the basic peptides nevertheless appear to play a more active role in the fusion process than simply bringing the vesicles in contact. However, protonation of the polymer side chains and transformation of the polymer into a polycation does not explain the need for acidification, since the pH-dependence was quite similar for poly(L-histidine)- and poly(L-lysine)-mediated fusion.     

Physiological levels of diacylglycerols in phospholipid membranes induce membrane fusion and stabilize inverted phases

In the preceding paper (Ellens et al., 1989), it was shown that liposome fusion rates are substantially enhanced under the same conditions which induce isotropic 31P NMR resonances in multilamellar dispersions of the same lipid. Both of these phenomena occur within the same temperature interval, delta TI, below the L alpha/HII phase transition temperature, TH. TH and delta TI can be extremely sensitive to the lipid composition. The present work shows that 2 mol% of diacylglycerols like those produced by the phosphatidylinositol cycle in vivo can lower TH, delta TI, and the temperature for fast membrane fusion by 15-20 degrees C. N-Monomethylated dioleoylphosphatidylethanolamine is used as a model system. These results show that physiological levels of diacylglycerols can substantially increase the susceptibility of phospholipid membranes to fusion. This suggests that, in addition to their role in protein kinase C activation, diacylglycerols could play a more direct role in the fusion event during stimulus-exocytosis coupling in vivo.     

Lipopeptide daptomycin: Interactions with bacterial and phospholipid membranes,stability of membrane aggregates and micellation in solution

Daptomycin is a cyclic lipopeptide effective against multidrug Gram-positive bacteria. Despite having a net negative charge, it is selective against negatively charged bacterial membranes. It has been established that daptomycin's antibiotic activity is based on directly targeting the bacterial membranes and that this antibacterial activity depends on calcium ions. Importantly, however, both the precise role of ions and the physical mechanisms responsible for daptomycin's action remain poorly understood. We investigate these issues using three types of molecular dynamics simulations: umbrella sampling free energy calculations for a single daptomycin, unbiased simulations for daptomycin tetramers, and unbiased simulations of micellation of daptomycin both in the absence and presence of calcium ions. The simulations are in the excess of 4 μs. As the most important finding, we establish that binding of the calcium ions on the aspartic acid residues is the key to stabilizing daptomycin tetramers inside the model material membrane. These complexes are vital for daptomycin's antibacterial activity. In the absence of binding, the tetramer is not stable and moves slowly out of the membrane. We also demonstrate that in solution, micellation of daptomycin occurs both in the presence and absence of calcium ions, and discuss the similarities between the behaviors of daptomycin and amyloid peptides in membranes.     

Effects of proteins on thermotropic phase transitions of phospholipid membranes.

A variety of proteins have been studied for their ability to interact and alter the thermotropic properties of phospholipid bilayer membranes as detected by differential scanning calorimeter. The proteins studied included: basic myelin protein (A1 protein), cytochrome c, major apoprotein of myelin proteolipid (N-2 apoprotein), gramicidin A, polylysine, ribonuclease and hemoglobin. The lipids used for the interactions were dipalmitoylphosphatidylcholine and dipalmitoylphosphatidylglycerol. The interactions were grouped in three catagories each having very different effects on the phospholipid phase transition from solid to liquid crystalline. The calorimetric studies were also correlated with data from vesicle permeability and monolayer expansion. Ribonuclease and polylysine which exemplify group 1 interactions, show strong dependence on electrostatic binding. Their effects on lipid bilayers include an increase in the enthalpy of transition (deltaH) accompanied by either an increase or no change in the temperature of transition (Tc). In addition, they show minimal effects on vesicle permeability and monolayer expansion. It was concluded that these interactions represent simple surface binding of the protein on the lipid bilayer without penetration into the hydrocarbon region. Cytochrome c and A1 protein, which exemplify group 2 interactions, also show a strong dependence on the presence of net negative charges on the lipid bilayers for their binding. In contrast to the first group, however, they induce a drastic decrease in both Tc and deltaH of the lipid phase transition. Furthermore, they induce a large increase in the permeability of vesicles and a substantial expansion in area of closely packed monolayers at the air-water interface. It was concluded that group 2 interactions represent surface binding followed by partial penetration and/or deformation of the bilayer. Group 3 interactions, shown by proteolipid apoprotein and gramicidin A, were primarily non-polar in character, not requiring electrostatic charges and not inhibited by salt and pH changes. They had no appreciable effect on the Tc but did induce a linear decrease in the magnitude of the deltaH, proportional to the percentage of protein by weight. Membranes containing 50% proteolipid protein still exhibited a thermotropic transition with a deltaH one half that of the pure lipid, and only a small diminution of the size of the cooperative unit. It was concluded that in this case the protein was embedded within the bilayer, associating with a limited number of molecules via non-polar interactions, while the rest of the bilayer was largely unperturbed.     

Reconstitution of model membranes from phospholipid and outer membrane proteins of Proteus mirabilis. Role of proteins in the formation of hydrophilic pores and protection of membranes against detergents.

Outer membrane proteins extracted from isolated cell walls of Proteus mirabilis were able to combine the cell wall phospholipids in a model membrane system. The presence of outer membrane proteins in vesicular model membranes mediated the release of previously entrapped [14C]sucrose while [3H]inulin was retained. Incorporation of lipopolysaccharide from the same cell walls was not required for the formation of such selectively permeable membranes. Three major outer membrane proteins of apparent molecular weights 39000, 36000 and 17000 were isolated using acetic acid and sodium deoxycholate solution as solvents and avoiding the strongly denaturing sodium dodecyl sulfate. The isolated proteins were assayed for their ability to form hydrophilic pores in reconstituted membranes. The trypsin-sensitive 39000-Mr protein and the peptidoglycan-associated 36000-Mr protein were equally effective in this function whereas the 17000-Mr protein mediated little penetration of low molecular weight solute. The 39000-Mr and 36000-Mr proteins also protected reconstituted membrane vesicles from disruption by detergent while 17000-Mr protein was ineffective in this regard.