Effects of arsenolipids on in vitro blood-brain barrier model.

Arsenic-containing hydrocarbons (AsHCs), a subgroup of arsenolipids (AsLs) occurring in fish and edible algae, possess a substantial neurotoxic potential in fully differentiated human brain cells. Previous in vivo studies indicating that AsHCs cross the blood-brain barrier of the fruit fly Drosophila melanogaster raised the question whether AsLs could also cross the vertebrate blood-brain barrier (BBB). In the present study, we investigated the impact of several representatives of AsLs (AsHC 332, AsHC 360, AsHC 444, and two arsenic-containing fatty acids, AsFA 362 and AsFA 388) as well as of their metabolites (thio/oxo-dimethylpropionic acid, dimethylarsinic acid) on porcine brain capillary endothelial cells (PBCECs, in vitro model for the blood-brain barrier). AsHCs exerted the strongest cytotoxic effects of all investigated arsenicals as they were up to fivefold more potent than the toxic reference species arsenite (iAsIII). In our in vitro BBB-model, we observed a slight transfer of AsHC 332 across the BBB after 6 h at concentrations that do not affect the barrier integrity. Furthermore, incubation with AsHCs for 72 h led to a disruption of the barrier at sub-cytotoxic concentrations. The subsequent immunocytochemical staining of three tight junction proteins revealed a significant impact on the cell membrane. Because AsHCs enhance the permeability of the in vitro blood-brain barrier, a similar behavior in an in vivo system cannot be excluded. Consequently, AsHCs might facilitate the transfer of accompanying foodborne toxicants into the brain.

Influence of lipid saturation grade and headgroup charge: a refined lung surfactant adsorption model.

Rapid adsorption of surfactant material to the air/liquid interface of the lung is essential for maintaining normal lung function. The detailed mechanism of this process, however, remains unclear. In this study, we elucidate the influence of lipid saturation grade and headgroup charge of surface layer lipids on surfactant protein (SP)-induced vesicle insertion into monolayers spread at the air/water interface of a film balance. We used dipalmitoylphosphatidlycholine (DPPC),1,2-dipalmitoyl-sn-glycero-3-phosphoglycerol (DPPG), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG) as monolayer lipids doped with either hydrophobic surfactant-specific protein SP-B or SP-C (0.2 and 0.4 mol %, respectively). Vesicles consisting of DPPC/DPPG (4:1, mol ratio) were injected into a stirred subphase to quantify adsorption kinetics. Based on kinetic film balance and fluorescence measurements, a refined model describing distinct steps of vesicle adsorption to surfactant monolayers is presented. First, in a protein-independent step, lipids from vesicles bridged to the interfacial film by Ca(2+) ions are inserted into defects of a disordered monolayer at low surface pressures. Second, in a SP-facilitated step, active material insertion involving an SP-B- or SP-C-induced flip-flop of lipids occurs at higher surface pressures. Negatively charged lipids obviously influence the threshold pressures at which this second protein-mediated adsorption mechanism takes place.

Arsenic-containing hydrocarbons disrupt a model in vitro blood-cerebrospinal fluid barrier.

Lipid-soluble arsenicals, so-called arsenolipids, have gained a lot of attention in the last few years because of their presence in many seafoods and reports showing substantial cytotoxicity emanating from arsenic-containing hydrocarbons (AsHCs), a prominent subgroup of the arsenolipids. More recent in vivo and in vitro studies indicate that some arsenolipids might have adverse effects on brain health. In the present study, we focused on the effects of selected arsenolipids and three representative metabolites on the blood-cerebrospinal fluid barrier (B-CSF-B), a brain-regulating interface. For this purpose, we incubated an in vitro model of the B-CSF-B composed of porcine choroid plexus epithelial cells (PCPECs) with three AsHCs, two arsenic-containing fatty acids (AsFAs) and three representative arsenolipid metabolites (dimethylarsinic acid, thio/oxo-dimethylpropanoic acid) to examine their cytotoxic potential and impact on barrier integrity. The toxic arsenic species arsenite was also tested in this way and served as a reference substance. While AsFAs and the metabolites showed no cytotoxic effects in the conducted assays, AsHCs showed a strong cytotoxicity, being up to 1.5-fold more cytotoxic than arsenite. Analysis of the in vitro B-CSF-B integrity showed a concentration-dependent disruption of the barrier within 72 h. The correlation with the decreased plasma membrane surface area (measured as capacitance) indicates cytotoxic effects. These findings suggest exposure to elevated levels of certain arsenolipids may have detrimental consequences for the central nervous system.

Partition of chlorpromazine into lipid bilayer membranes: the effect of membrane structure and composition.

Partition coefficients, kp, of chlorpromazine between the aqueous phase and lipid bilayer vesicles were determined as function of drug concentration, lipid chain length, cholesterol content and temperature encompassing the range of the lipid phase transition. Radioactivity and absorption measurements were performed to determine the kp values. Up to a concentration of 3 . 10(-5) M, the partition coefficient is independent of chlorpromazine concentration, whereas it decreases drastically at higher chlorpromazine concentrations, at which membrane lysis is observed. Membrane structure is not disturbed at less than 3 . 10(-5) M chlorpromazine, as was concluded from electron paramagnetic resonance studies measuring TEMPO partitioning and order degree. However, the lipid phase-transition temperature decreases and is broadened at higher chlorpromazine concentrations. From fluorescence measurements, we conclude the formation of chlorpromazine micelles at concentrations higher than 5 . 10(-5) M in chlorpromazine in the absence of lipids and the formation of mixed micelles in the presence of lipids. The effect of lipid chain length on kp values was investigated. The partition coefficient decreases from 8100 in dilauroyl- to 3400 in dipalmitoylphosphatidylcholine vesicles, both at 50 degrees C, that is, above their corresponding phase-transition temperature tt. At t less than tt the kp values are strongly reduced, by at least a factor of 10, depending on lipid chain length and membrane composition. It is possible to establish a lipid phase-transition curve from the temperature-dependent measurements of the kp values. Cholesterol within the lipid membrane strongly decreases kp. At 20 mol% cholesterol in dipalmitoylphosphatidylcholine membranes, the partition coefficient is reduced from 3400 to 2300. This value is well comparable to the kp value obtained in erythrocyte ghosts. In contradiction to earlier experiments by Conrad and Singer (Biochemistry 20 (1981) 808-818), this value in a biological membrane could be obtained by the hygroscopic desorption as well as the centrifugation method. From our experiments we are justified in further considering artificial bilayer membranes as models for biological membranes.

The effect of gangliosides on the lamellar phase behaviour of phosphatidylethanolamines.

The thermotropic properties of aqueous phosphatidylethanolamine dispersions vary with hydration. Measured by EPR-spectroscopy freshly hydrated dimyristoylphosphatidylethanolamine dispersions exhibit a gel to liquid-crystalline phase transition at Tml = 48 degrees C. Dehydration could be induced by prolonged incubation of a hydrated sample at 4 degrees C. The phase transition temperature of the dehydrated phase was determined to be Tmh = 54 degrees C. From the measured phase transition curves we followed the dehydration with time and found a cooperative nucleation process. A 50% dehydration was reached after 5 days. This dehydration process could be prevented by gangliosides: 1.5 mol% of GT1b, 4 mol% of GM1 or 7 mol% of GD1a or GM3 but also 7 mol% of phosphatidic acid were able to stabilize the hydrated phase completely. The effect of gangliosides GM1, GM3, GD1a, GT1b and of the negatively charged phosphatidic acid on the phase behaviour of dimyristoylphosphatidylethanolamine (DMPE) dispersions were investigated. The phase transition temperature of freshly hydrated DMPE samples was successively decreased from 48 to 43 degrees C with increasing amounts of GD1a up to 10 mol% whereby the phase transition was significantly broadened. Gangliosides GM1, GM3 and GT1b as well as phosphatidic acid had minor effects. Dispersions of pure DMPE prepared below the transition temperature Tml form the dehydrated phase again with a melting temperature of Tmh = 54 degrees C. In the presence of 10 mol% GD1a or GT1b this value is reduced to Tml, the phase transition temperature of the hydrated phase. The reduction induced by GM3 is less pronounced. With GM1 or phosphatidic acid the samples remain partially dehydrated and the phase transition curves become biphasic up to 7 mol% ganglioside or phosphatidic acid.

Pressure variation of the lateral diffusion in lipid bilayer membranes.

A pressure-induced decrease of the lateral diffusion in pure and cholesterol containing phosphatidylcholine bilayer membranes has been determined by the excimer formation technique using pyrene as probe molecule. The experimental results at pressures up to 150 bars are described satisfactorily by the free volume theory of a molecular transport in liquids. A pressure increase of extrapolated 575 bars decreases the lateral diffusion of lipids by a factor of two in pure dipalmitoylphosphatidylcholine membranes. Higher pressures are necessary to induce the same effect in cholesterol containing membranes. This result is interpreted by the condensing effect of cholesterol in fluid bilayer membranes.

Cooperative lipid-protein interaction. Effect of pH and ionic strength on polymyxin binding to phosphatidic acid membranes.

The binding of polymyxin-B to charged dipalmitoyl phosphatidic acid membranes has been studied as function of the external pH and of the ionic strength of the buffer solution. The phase transition curves were obtained by measuring the fluorescence depolarization of diphenyl hexatriene incorporated into the membrane with temperature. The molecular process of polymyxin binding was elucidated: 1. At an ionic strength of I greater than or equal to 0.1 mol/l a three step phase transition curve is found. A high-temperature step corresponds to the non-bound lipid. A lowered phase transition concerns to protein-bound lipid domains. This again is splitted into two steps. An inner core of the domain is characterized by a lipid-protein complex which is stabilized through hydrophobic and electrostatic interactions between polymyxin and the charged lipid. This core is surrounded by an outer belt of only hydrophobically bound molecules. This part shows a lower phase transition temperature than the inner core. 2. The binding curves of polymyxin to phosphatidic acid membranes depend strongly on the ionic strength of the water phase. The cooperativity of the binding process increases with increasing ionic strength and reaches a constant value at I greater than 0.2 mol/l. The maximum fraction of bound lipid decreases with increasing ionic strength. 3. The pH of the water phase strongly influences the cooperative binding process. At pH 6 a loss of cooperativity is observed at low ionic strength. Increasing the ion concentration to I = 0.3 mol/l recuperates the cooperativity of the binding process. At pH 3.0 no cooperative binding is obtained even at high ionic strength.

Pressure-induced changes in the molecular organization of a lipid-peptide complex: polymyxin binding to phosphatidic acid membranes.

The effect of 100 atm pressure on the organization of the lipid-peptide complex formed between polymyxin and dipalmitoyl phosphatidic acid has been investigated. Phase transition curves were obtained by electron paramagnetic resonance by measuring the partition coefficient of the spin label, 2, 2, 5, 5-tetramethylpiperidine-N-oxyl. The three-step phase transition curve previously obtained with fluorescence polarization measurements was confirmed, demonstrating three distinct phosphatidic acid domains in the bilayer. Pressure increases binding of polymyxin to phosphatidic acid bilayers and alters the proportions of the two domains that differ in the mode of binding between phosphatidic acid and polymyxin. The binding curves of polymyxin to phosphatidic acid bilayers wre determined and it was shown that application of pressure reduces the cooperativity of the binding curve.

Asymmetric antagonistic effects of an inhalation anesthetic and high pressure on the phase transition temperature of dipalmitoyl phosphatidic acid bilayers.

The phase transition temperature (Tt) of dipalmitoyl phosphatidic acid multilamellar liposomes is depressed 10 degrees C by the inhalation anesthetic methoxyflurane at a concentration of 100 mmol/mol lipid. Application of 100 atm of helium pressure to pure phosphatidic acid liposomes increased Tt only 1.5 degrees C. However, application of 100 atm helium pressure to dipalmitoyl phosphatidic acid lipsomes containing 100 mmol methoxyflurane/mol lipid almost completely antagonized the effect of the anesthetic. A non-linear pressure effect is observed. In a previous study, a concentration of 60 mmol methoxyflurane/mol dipalmitoyl phosphatidylcholine depressed Tt only 1.5 degrees C, exhibiting a linear pressure effect. The completely different behavior in the charged membrane is best explained by extrusion of the anesthetic from the lipid phase.

Binding of polylysine to charged bilayer membranes: molecular organization of a lipid.peptide complex.

The interaction between a positively charged peptide (poly-L-lysine) and model membranes containing charged lipids has been investigated. Conformational changes of the polypeptide as well as changes in the membrane lipid distribution were observed upon lipid-protein agglutination: 1. The strong binding of polylysine is shown directly by the use of spinlabelled polypeptide. Upon binding to phosphatidic acid a shift in the hyperfine coupling constant from 16.5 to 14.6 Oe is observed. The spectrum of the lipid-bound peptide is superimposed on the spectrum of polylysine in solution. Half of the lysine groups are bound to the charged membranes. A change in the conformation of polylysine from a random coil to a partially ordered configuration is suggested. 2. Spin labelling of the lipid component gives evidence concerning the molecular organization of a lipid mixture containing charged phosphatitid acid. Addition of polylysine induces the formation of crystalline patches of bound phosphatidic acid. 3. Excimer forming pyrene decanoic acid has been employed. Addition of positively charged polylysine (pH 9.0) to phosphatidic acid membranes increases the transition temperature of the lipid from Tt = 50 to Tt = 62 degrees C. Thus, a lipid segregation of lipid into regions of phosphatidic acid bound to the peptide which differ in their microviscosity from the surrounding membrane is induced. One lysine group binds one phosphatidic acid molecule, but only half of the phosphatidic acid is bound. 4. Direct evidence for charge induced domain formation in lipid mixtures containing phosphatidic acid is given by electron microscopy. Addition of polylysine leads to a change in the surface curvature of the bound charged lipid. The domain size is estimated from the electron micrographs. The number of domains present is dependent on both the ratio of charged to uncharged lipids as well as on the amount of polylysine added to the vesicles. The size of the domains is not dependent on membrane composition. However, the size seems to increase in a stepwise manner that is correlated with a multiple of the area covered by one polylysine molecule.

Calorimetric investigation of polymyxin binding to phosphatidic acid bilayers.

The cooperative binding process between the antibiotic peptide polymyxin-B and negatively-charged phosphatidic acid bilayers was investigated by differential thermal analysis and completed by fluorescence polarization measurements. The sigmoidal binding curves were analyzed in terms of the interaction energy within a domain formed by polymyxin and phosphatidic acid molecules. The formation of such a heterogeneous domain structure was favoured by high concentration of external monovalent ions. The cooperativity of the binding increased while a charge-induced decrease in the phase transition temperature of the pure lipid phase was observed with increasing ion concentration at a given pH. The reduced lateral coupling within the lipid bilayer in the presence of salt ions, as demonstrated by an increase in the lipid phase transition enthalpy, was considered to facilitate the cooperative domain formation. Moreover, an increase in the cooperativity of the polymyxin binding could be observed if phosphatidic acids of smaller chain length and thus of a lowered phase transition temperature were used. By the use of chemically-modified polymyxin we were able to demonstrate the effect of electrostatic and hydrophobic interaction. Acetylated polymyxin with a reduced positive charge was used to demonstrate the pure hydrophobic effect of polymyxin binding leading to a decrease in the phosphatidic acid phase transition temperature by about 20 degrees C. The cooperativity of the binding was strongly reduced. Cleavage of the hydrophobic polymyxin tail yielded a colistinnonapeptide which caused an electrostatically-induced increase in the phosphatidic acid phase transition temperature. With unmodified polymyxin we observed the combined effects of electrostatic as well as hydrophobic interaction making this model system interesting for the understanding of lipid-protein interactions. Evidence is presented that the formation of the polymyxin-phosphatidic acid complex is a lateral phase separation phenomenon.

Lateral and transversal diffusion and phase transitions in erythrocyte membranes. An excimer fluorescence study.

The lateral diffusion of the excimer-forming probe pyrene decanoic acid has been determined in erythrocyte membranes and in vesicles of the lipid extracts. The random walk of the probe molecules is characterized by their jump frequency, nu j, within the lipid matrix. At T = 35 degrees C a value of nu i = 1.6 . 10(8) s-1 is found in erythrocyte membranes. A somewhat slower mobility is determined in vesicles prepared from lipid extracts of the erythrocyte membrane. Depending on structure and change of the lipids we obtain jump frequencies between 0.8 . 10(8) s-1 and 1.5 . 10(8) s-1 at T - 35 degrees C. The results are compared with jump frequencies yielded in model membranes. The mobility of molecules perpendicular to the membrane surface (transversal diffusion) is investigated. Erythrocyte ghosts doped with pyrene phosphatidylcholine were mixed with undoped ghosts in order to study the exchange kinetics of the probe molecule. A fast transfer between the outer layers of the ghost cells tau 1/2 = 1.6 min at T = 37 degrees C) is found. The exchange process between the inner and the outer layer of one erythrocyte ghost (flip-flop process) following this fast transfer occurs with a half-life time value of t 1/2 = 100 min at T = 37 C. The application of excimer-forming probes presumes a fluid state of the membrane. Therefore we investigated the phase transition behaviour using the excimer technique. Beside a thermotropic phase transition at T = 38 degrees C and T = 33 degrees C we observed an additional fluidity change at T = 38 degrees C in erythrocyte ghost. This transition is attached to a separation of the boundary lipid layer from the intrinsic proteins. No lipid phase transition is observed in liposomes from isolated extracts of the erythrocyte membrane with our methods.

Polymyxin interaction with negatively charged lipid bilayer membranes and the competitive effect of Ca2+.

The binding of cationic polymyxin-B to negatively charged phosphatidic acid and phosphatidylglycerol membranes has been investigated by fluorescence polarization study. Competition experiments with Ca2+ were performed. 1. Binding of polymyxin-B to mixed dipalmitoylphosphatidic acid/distearoylphosphatidylcholine membranes leads to a phase separation. Domains of polymyxin-bound phosphatidic acid are formed. 2. Ca2+ is found to be a strong competitor in displacing polymyxin from the complex in the mixed membrane system. Complete displacement is obtained at pH 9.0. With decreasing pH value, Ca2+ becomes a less strong competitor and is ineffective at pH 5.0. 3. Binding of polymyxin to dipalmitoylphosphatidylglycerol membranes is observed. Incorporation of polymyxin lowers the lipid phase transition by 10 degrees C. One polymyxin is found to bind five phosphatidylglycerol molecules. The binding curve is determined and in contrast to phosphatidic acid membranes, a noncooperative binding could be established. 4. Addition of Ca2+ decreases the amount of phosphatidylglycerol bound to polymyxin by about 20%. No complete displacement is achieved even at 10-fold excess of Ca2+ with respect to phosphatidylglycerol.

Chemically induced lipid phase separation in model membranes containing charged lipids: a spin label study.

The lipid distribution in binary mixed membranes containing charged and uncharged lipids and the effect of Ca2+ and polylysine on the lipid organization was studied by the spin label technique. Dipalmitoyl phosphatidic acid was the charged, and spin labelled dipalmitoyl lecithin was the uncharged (zwitterionic) component. The ESR spectra were analyzed in terms of the spin exchange frequency, Wex. By measuring Wex as a function of the molar percentage of labelled lecithin a distinction between a random and a heterogeneous lipid distribution could be made. It is established that mixed lecithin-phosphatidic acid membranes exhibit lipid segregation (or a miscibility gap) in the fluid state. Comparative experiments with bilayer and monolayer membranes strongly suggest a lateral lipid segregation. At low lecithin concentration, aggregates containing between 25% and 40% lecithin are formed in the fluid phosphatidic acid membrane. This phase separation in membranes containing charged lipids is understandable on the basis of the Gouy-Chapman theory of electric double layers. In dipalmitoyl lecithin and in dimyristoyl phosphatidylethanolamine membranes the labelled lecithin is randomly distributed above the phase transition and has a coefficient of lateral diffusion of D = 2.8-10(-8) cm2/s at 59 degrees C. Addition of Ca2+ dramatically increases the extent of phase separation in lecithin-phosphatidic acid membranes. This chemically (and isothermally) induced phase separation is caused by the formation of crystalline patches of the Ca2+-bound phosphatidic acid. Lecithin is squeezed out from these patches of rigid lipid. The observed dependence of Wex on the Ca2+ concentration could be interpreted quantitatively on the basis of a two-cluster model. At low lecithin and Ca2+ concentration clusters containing about 30 mol % lecithin are formed. At high lecithin or Ca2+ concentrations a second type of precipitation containing 100% lecithin starts to form in addition. A one-to-one binding of divalent ions and phosphatidic acid at pH 9 was assumed. Such a one-to-one binding at pH 9 was established for the case of Mn2+ using ESR spectroscopy. Polylysine leads to the same strong increase in the lecithin segregation as Ca2+. The transition of the phosphatidic acid bound by the polypeptide is shifted from Tt = 47.5 degrees to Tt = 62 degrees C. This finding suggests the possibility of cooperative conformational changes in the lipid matrix and in the surface proteins in biological membranes.

Lateral diffusion of small solutes and partition of amphipaths in defect structures of lipid bilayers.

The excimer formation technique has been applied to investigate the mechanism of the lateral diffusion in the crystalline P beta' phase of dipalmitoylphosphatidylcholine vesicles. This became possible at low pyrene concentrations. From the shape of phase-transition curves, from the effect of pressure up to 150 bar and from the perturbation induced by cholesterol, we conclude that the free volume model could also be applied to the P beta' phase. However, the diffusion is thought to occur in defect structures that are considered to form fluid pathways between domains of crystalline lipid. Partition coefficients of amphipaths provide a basis for testing for the role of defects. The amphipath chlorpromazine partitions into fluid membranes with a partition coefficient, kp, of 3200, into the crystalline phase with kp = 200 but into the P beta' phase with a value of kp = 800. This again gives rise to the assumption that the P beta' phase contains fluid domains that behave like the fluid L alpha phase and make up about 10-20% of the total amount of lipid in the bilayer. Cholesterol is known to interfere especially with defect structures in the P beta' phase. It fills up the gaps, and therefore reduces the partition coefficient to almost zero in the P beta' phase.

Interaction between glycophorin and a spin-labeled cholesterol analogue in reconstituted dimyristoylphosphatidylcholine bilayer vesicles.

The interaction between glycophorin and a spin-labeled cholesterol analogue has been investigated by EPR spectroscopy. In vesicles which were reconstituted by the freeze and thaw technique, direct evidence was obtained for a reorganisation of the membrane at low protein content (protein/lipid ratio less than 1:300). From the spin exchange interaction we were able to show a protein-induced clustering of the steroid in fluid and in gel state membranes. Tryptic cleavage of the complete N-terminus of glycophorin vanishes the effect. Whereas the removal of the sialic acid residues by neuraminidase digest had no influence on the EPR spectra. The interaction seems to be cholestane spin label specific since it was not observed with an androstane spin-label.

Monitoring the location profile of fluorophores in phosphatidylcholine bilayers by the use or paramagnetic quenching.

Spin probes differing in the position of their paramagnetic centre are used to quench the fluorescence of pyrene derivatives and chlorophylls incorporated into dimyristoyl phosphatidylcholine membranes. Pyrene butyric acid and pyrene decanoic acid with known orientation relative to the membrane surface are investigated. The quenching efficiency of fatty acid spin probes is dependent on the position of the nitroxide radical group in the fatty acid chain. Using this short range interaction we developed a spectroscopic method to chlorophyll-containing vesicles, we were able to characterize the orientation of the porphyrin ring within the membrane. Moreover, the chlorophyll fluorescence is also quenched by a water-soluble spin label. Therefore the porphyrin ring appears to be orientated in the polar head group region of the lipid layer, but not to be protruding out into the water phase. This conclusion is confirmed by the use of pyrene derivatives. Fluorescence quenching by a water-soluble spin label within the lipid matrix is observed even in the rigid state of the membrane. Fluorescence lifetime measurements suggest the existence of two different quenching mechanisms: (1) a static quenching occurring below the lipid phase transition temperature, and (2) an additional dynamic quenching taking place in the fluid state of the lipid bilayer.

Polymyxin binding to charged lipid membranes. An example of cooperative lipid-protein interaction.

The binding of polymyxin-B to lipid bilayer vesicles of synthetic phosphatidic acid was studied using fluorescence, ESR spectroscopy and electron microscopy. 1,6-Diphenylhexatriene (which exhibits polarized fluorescence) and pyrene decanoic acid (which forms excimers) were used as fluorescence probes to study the lipid phase transition. The polymyxin binds strongly to negatively charged lipid layers. As a result of lipid/polymyxin chain-chain interactions, the transition temperature of the lipid. This can be explained in terms of a slight expansion of the crystalline lipid lattice (Lindeman's rule). Upon addition of polymyxin to phosphatidic acid vesicles two rather sharp phase transitions (width deltaT = 5 degrees C) are observed. The upper transition (at Tu) is that of the pure lipid and the lower transition (at T1) concerns the lipid bound to the peptide. The sharpness of these transitions strongly indicates that the bilayer is characterized by a heterogeneous lateral distribution of free and bound lipid regions, one in the crystalline and the other in the fluid state. Such a domain structure was directly observed by electron microscopy (freeze etching technique). In (1 : 1) mixtures of dipalmitoyl phosphatidic acid and egg lecithin, polymyxin induces the formation of domains of charged lipid within the fluid regions of egg lecithin. With both fluorescence methods the fraction of lipid bound to polymyxin-B as a function of the peptide concentration was determined. S-shaped binding curves were obtained. The same type of binding curve is obtained for the interaction of Ca2+ with phosphatidic acid lamellae, while the binding of polylysine to such membranes is characterized by a linear or Langmuir type binding curve. The S-shaped binding curve can be explained in terms of a cooperative lipid-ligand (Ca2+, polymyxin) interaction. A model is proposed which explains the association of polymyxin within the membrane plane in terms of elastic forces caused by the elastic distortion of the (liquid crystalline) lipid layer by this highly asymmetric peptide.

Virus capping on mycoplasma cells and its effect on membrane structure.

The capping of mycoplasmavirus L3 on the surface of Acholeplasma laidlawii was investigated. In electron microscope studies we observed a reduced capping after treatment of the host cell with energy-blocking agents. Other drugs inhibiting ligand capping on eucariotic cells had no effect. Changes in membrane structure after virus adsorption were observed spectroscopically using the excimer fluorescence technique. The results are interpreted in terms of a lipid-protein phase separation in connection with virus capping.

Incorporation of highly purified melittin into phosphatidylcholine bilayer vesicles.

Melittin free of phospholipase A2 was prepared. In the absence of salt this highly pure protein starts to aggregate in solution at a protein concentration of Cp greater than 10(-3) M. In high salt solution (2 M) aggregation starts at Cp greater than 10(-6) M. This was determined from the blue shift of the intrinsic fluorescence of the protein. Reinvestigation of the quenching behaviour clearly shows that self-aggregation cannot be deduced from quenching experiments using nitrate or 2,2,6,6-tetramethylpiperidine-1-oxyl as quencher. The incorporation of melittin into phosphatidylcholine bilayer vesicles was studied by fluorescence quenching and by energy-transfer experiments using 2- and 6-anthroyloxypalmitic acid as acceptor and peptide tryptophan as donor. Incorporation of melittin into small unilamellar vesicles was found to be reduced below the lipid phase transition temperature, Tt, whereas it incorporates and distributes more randomly above Tt. Cooling the temperature below Tt after incubation at T greater than Tt leads to a deeper incorporation of the peptide into the lipid bilayer due to electrostatic interaction between the lipid phosphate groups and the positively charged amino acids. This stabilizing effect is lost above Tt and melittin is extruded to the polar phase. Quenching experiments support this finding. EPR measurements clearly demonstrate that even in the presence of high amounts of melittin up to 10 mol% with respect to the lipid broadening of the phase transition curves was only observed with fatty acid spin labels, where the doxyl group is localized near the bilayer surface. The order degree of the inner part of the bilayer remains almost unchanged even in the presence of high melittin content.