Experimental alteration of phospholipid-protein interactions within the human erythrocyte membrane. Dependence on glycolytic metabolism.

Phosphatidylethanolamine in freshly drawn human erythrocytes is trinitrophenylated by 2,4,6-trinitrobenzene sulfonic acid only slowly and to a maximum of 32%. After different preincubation procedures at 37 degrees C in saline media in the absence of glucose (24 h without additive, 1-5 h with 8 mM hexanol or 1-4 h with the SH reagent, 5 mM tetrathionate) the rate of subsequent trinitrophenylation of phosphatidylethanolamine, in the absence of the additives, is greatly enhanced and the amount of phospholipid reacting increased. Glucose or inosine prevent these effects, inhibitors of glycosis abolish this protection. The results indicate that in fresh as well as in glycolysing incubated erythrocytes phosphatidylethanolamine in the outer layer of the membrane lipid is shielded by a protein. Conformational changes of this protein induced by metabolic starvation and perturbing agents expose the phospholipid head group to 2, 4, 6-trinitrobenzene sulfonic acid. In addition, a "flip-flop" of phosphatidylethanolamine from the inner to the outer layer may also contribute to the effects observed.

Possible relationship between membrane proteins and phospholipid asymmetry in the human erythrocyte membrane.

After incubation of human erythrocytes at 37 degrees C in the absence of glucose (A) for 24 h, (B) for 4 h with 8 mM hexanol or (C) for 3 h with SH reagents, phosphatidylethanolamine becomes partly susceptible to hydrolysis by phospholipase A2 from Naja naja. The presence of glucose during the pretreatments suppresses this effect, except in the case of SH reagents that inhibit glycolysis. After incubation with tetrathionate, up to 45% of the phosphatidylethanolamine is degraded by the enzyme, an amount considerably in excess of the 20% attacked in fresh erythrocytes. Pancreatic phospholipase A2, an enzyme unable to hydrolyse the phospholipids of intact erythrocytes, partially degrades phosphatidylcholine and phosphatidylethanolamine of erythrocytes pretreated with hexanol or SH reagents. Reagents capable of oxidizing SH groups to disulfides (tetrathionate, o-iodosobenzoate and hydroquinone) even render susceptible to pancreatic phospholipase A2 phosphatidylserine, a phospholipid supposed to be entirely located in the inner lipid layer of the membrane. Alkylating or acylating SH reagents have no such effect. It is postulated that disulfide bond formation between membrane protein SH groups leads to an alteration in protein-phospholipid interactions and consequently induces a reorientation of phospholipids between the inner and the outer membrane lipid layer.

Evidence for a role of the multidrug resistance protein (MRP) in the outward translocation of NBD-phospholipids in the erythrocyte membrane.

Phosphatidylserine (PS) containing a 7-nitrobenz-2-oxa-1, 3-diazol-4-yl- (NBD-) hexanoyl residue, like native PS, preferentially distributes into the inner membrane leaflet of human erythrocytes. In the case of NBD-PS, this preference results from two opposite active processes, an inward translocation mediated by the aminophospholipid flippase and an outward translocation mediated by an ill-defined floppase. Selective inhibition of this floppase by alkylating reagents or cationic and anionic drugs increases the extent of accumulation of NBD-PS in the inner membrane leaflet from about 70% in control cells to about 90%. Different inhibitor sensitivities of the flippase and the floppase strongly suggest that both represent different entities. The floppase was characterized in further detail by comparing inhibitory effects of various compounds on this translocase with their effects on known primary active transport systems for amphiphilic compounds. The inhibitory effects of various drugs, glutathione conjugates and GSSG on the floppase activity closely correlate with those reported for the active transport by the multidrug resistance protein (MRP) while only poorly going parallel with those for the active transport by the low affinity pump for glutathione conjugates and the multidrug resistance MDR1 P-glycoprotein. The NBD-phospholipid floppase activity of the erythrocyte is thus probably a function of MRP.

Lateral segregation of membrane lipids and formation of stable rod-shaped membrane projections in erythrocytes treated with long-chain alcohols.

Human erythrocytes, incubated with sonicated dispersions of phosphatidylcholine, cholesterol and saturated straight-chain alcohols (C16-C18) develop stiff, rod-shaped, hemoglobin-containing membrane projections within 120 min. The number of these 'rods' varies (1-3 per cell), they reach a length of up to 14 microns (twice the cell diameter) and a thickness of 0.3-1.0 microns. 'Rods' may be separated from 'residual cells' by shear flow and centrifugation without severe hemolysis. Lipid analyses carried out on residual cells and rods indicate lateral segregation of the phospholipids of the outer leaf of the membrane lipid bilayer (phosphatidylcholine and sphingomyelin) and of the alcohol applied. Phosphatidylcholine accumulates in the residual cells, sphingomyelin and the alcohol in the rods. No differences in membrane protein patterns were observed between rods and residual cells. The rod-shape is dependent on the presence of the alcohol, extraction of the alcohol converts rods into hemoglobin-containing spheres without lysis. The formation of rods, which is indicative of a lateral phase separation, is discussed in terms of lipid-lipid interactions and with respect to parameters determining the shape of cells.

Alcohols produce reversible and irreversible acceleration of phospholipid flip-flop in the human erythrocyte membrane.

The slow, non-mediated transmembrane movement of the lipid probes lysophosphatidylcholine, NBD-phosphatidylcholine and NBD-phosphatidylserine in human erythrocytes becomes highly enhanced in the presence of 1-alkanols (C2-C8) and 1,2-alkane diols (C4-C8). Above a threshold concentration characteristic for each alcohol, flip rates increase exponentially with the alcohol concentration. The equieffective concentrations of the alcohols decrease about 3-fold per methylene added. All 1-alkanols studied are equieffective at comparable calculated membrane concentrations. This is also observed or the 1,2-alkane diols, albeit at a 5-fold lower membrane concentration. At low alcohol concentrations, flip enhancement is reversible to a major extent upon removal of the alcohol. In contrast, a residual irreversible flip acceleration is observed following removal of the alcohol after a treatment at higher concentrations. The threshold concentrations to produce irreversible flip acceleration by 1-alkanols and 1,2-alkane diols are 1.5- and 3-fold higher than those for flip acceleration in the presence of the corresponding alcohols. A causal role in reversible flip-acceleration of a global increase of membrane fluidity or membrane polarity seems to be unlikely. Alcohols may act by increasing the probability of formation of transient structural defects in the hydrophobic barrier that already occur in the native membrane. Membrane defects responsible for irreversible flip-acceleration may result from alterations of membrane skeletal proteins by alcohols.

Transbilayer reorientation of phospholipid probes in the human erythrocyte membrane. Lessons from studies on electroporated and resealed cells.

In order to characterize in more detail the previously observed (Dressler et al. (1983) Biochim. Biophys. Acta 732, 304-307) increases in transbilayer mobility of phospholipids in the erythrocyte membrane following electroporation at 0 degrees C and subsequent resealing at 37 degrees C of the cells, we have studied rates of flip and flop as well as steady state distributions of the fluorescent N-(NBD)-aminohexanoyl-analogues of the four major membrane phospholipids. Measurements comprised the passive non-mediated components as well as those mediated by specific translocases (flippase and floppase). The major new findings and insights can be summarized as follows. (1) The enhancement of passive transbilayer mobility which increases with the strength, duration, and number of field pulses at 0 degrees C, cannot be fully reversed by subsequent resealing at 37 degrees C. Flip-flop remains considerably elevated relative to the original values.(2) Enhanced mobilities induced by electroporation differ for the probes studied in the sequence SM <<< PS << PC < PE. Other membrane perturbations going along with enhanced flip-flop share only in part this pattern. (3) Mediated, ATP-dependent components of flip and flop of the probes are suppressed in electroporated/resealed cells, partly due to loss of cellular Mg2+, partly - in case of flippase - due to competition by externalized endogenous PS. (4) Electroporated/resealed cells provide an elegant means to demonstrate the contribution of various components of flip and flop to the steady state transbilayer distribution of phospholipids, in particular the role of passive mobility. The new, detailed information on the displacements of phospholipid between the two leaflets of the membrane bilayer in porated/resealed cells will help to understand erythrocyte shape changes following poration and during resealing (Henszen et al. (1993) Biol. Chem. Hoppe-Seyler 374, 114).

Pathways for flip-flop of mono- and di-anionic phospholipids in the erythrocyte membrane.

The inward translocations (flip), from the outer to the inner membrane leaflet of human erythrocytes, of di-anionic NBD-labeled phospholipids containing as a head group phosphate esters of glycolate, butyrate and hydroxyethanesulfonate are slow processes (k = 0.005-0.008 h-1, 37 degrees C) at pH 7.4. A decrease of pH highly stimulates the flip. A major role of the anion exchanger (AE1), band 3, in this flip is indicated by (a) the strong inhibition of the flip (55-85%) by stilbene disulfonates and other inhibitors of anion transport, (b) the stimulation and loss of pH dependence of the flip after modification of band 3 by Woodward's reagent K and NaBH4, and (c) the stimulation of the flip after proteolytic cleavage of band 3 by papain. The flip of mono-anionic NBD-phospholipids with phosphate esters of glycerol, glycol, methanol, butanol and benzyl alcohol is much faster than that of their dianionic analogs (k = 0.04 to > 3.0 h-1, 37 degrees C). It is inhibited by stilbene disulfonates to a decreasing extent (35 to 0%) and is not affected by several reversible inhibitors of anion exchange. This indicates a minor component of band-3-mediated flip and a major component of nonmediated flip. The outward translocations (flop), from the inner to outer membrane leaflet, of both mono- and di-anionic phospholipids are very fast (1.0-5.9 h-1), ATP-dependent and inhibitable by vanadate, fluoride, SH-reagents or Mg(2+)-depletion of cells and thereby likely to be largely mediated by a 'floppase'. The stationary distributions of the NBD-labeled anionic phospholipids are asymmetric to an extent (outer to inner leaflet ratio 2-9) correlating with the ratio of the rates of the outward and the inward translocation. Thus, asymmetry is largely abolished by blockage of the floppase-mediated translocation.

Phospholipid dependence of the anion transport system of the human erythrocyte membrane. Studies on reconstituted band 3/lipid vesicles.

Band 3 protein extracted from human erythrocyte membranes by Triton X-100 was recombined with the major classes of phospholipid occurring in the erythrocyte membrane. The resulting vesicle systems were characterized with respect to recoveries, phospholipid composition, protein content and vesicle size as well as capacity and activation energy of sulfate transport. Transport was classified into band-3-specific fluxes and unspecific permeability by inhibitors. Transport number (sulfate ions per band 3 per minute) served as a measure of functional therapy after reconstitution. The transport properties of band 3 proved to be insensitive to replacement of phosphatidylcholine by phosphatidylethanolamine, while sphingomyelin and phosphatidylserine gradually inactivated band-3-specific anion transport when present at mole fractions exceeding 30 mol%. The activation energy of transport remained unaltered in spite of the decrease in transport numbers. The results, which are discussed in terms of requirements of band 3 protein function with respect to the fluidity and surface charge of its lipid environment, provide a new piece of evidence that the transport function of band 3 protein depends on the properties of its lipid environment just as the catalytic properties of some other membrane enzymes. The well-established species differences in anion transport (Gruber, W. and Deuticke, B. (1973) J. Membrane Biol. 13, 19-36) may to some extent reflect this lipid dependence.

Mediated transport of anions in band 3-phospholipid vesicles.

Band 3 protein, extracted from human erythrocyte membranes by Triton X-100, was recombined with egg lecithin/cholesterol mixtures to form small unilamellar vesicles at a yield of 15-20%. These systems exhibited sulfate fluxes which were inhibitable by stilbene disulfonates and other inhibitors. Maximal inhibition could only be obtained when inhibitors were present at both membrane surfaces. Inhibitor constants I50 were higher than in the native membrane. Quantitatively, transport function was retained at least 60%, as related to the amount of protein involved. Sulfate transport in the recombinates resembled transport in the native membrane with respect to temperature dependence (Ea = 29-32 kcal/mol), pH dependence between pH 6.5 and 7.8, and the relationship between new and exchanges fluxes. In contrast to the native cell, concentration dependence was linear up to 80 mM sulfate, which may be indicative of a lowered affinity for the substrate. Lactate transport in these systems, although substantial, was insensitive to stilbene disulfonates as well as to mercurials, indicating that band 3 is not involved in the specific monocarboxylate transfer in the erythrocyte. Anion transport in band 3-lipid recombinates was insensitive to cholesterol between 0 and 27 mol%. Treatment with proteases, while not affecting transport per se, abolished sensitivity to stilbene disulfonate inhibitors. These observations indicate a number of disturbances of band 3 after recombination, in spite of a preservation of the major transport properties.

Topology of membrane sulfhydryl groups in the human erythrocyte. Demonstration of a non-reactive population in intrinsic proteins.

A major fraction of the protein sulfhydryl groups of human erythrocyte membranes can be oxidized to disulfide bonds by the lipid soluble reagent, diamide, and the hydrophilic reagent, tetrathionate. Furthermore, the same fraction also reacts with the monofunctional reagent, N-ethylmaleimide. About 20% of the SH groups, however, do not react with any of these agents even upon prolonged treatment and increased concentrations. These 'non-reacting' SH groups were now localized by a procedure involving blockage of the accessible SH groups by non-labeled N-ethylmaleimide or by diamide, subsequent isolation and solubilization of the membranes in SDS and labelling of the now accessible, residual SH groups with N-[ethyl-2-3H]ethylmaleimide. The distribution of the radioactivity over the peptide fractions shows that the non-reacting SH groups are mainly localized in the intrinsic proteins, while essentially all of the SH groups of the extrinsic protein, spectrin, are reactive. After solubilization of the membranes with Triton X-100 the non-reacting SH groups became reactive towards N-ethylmaleimide. It is proposed that lack of reaction of SH groups in the native membranes is due to their localization within the hydrophobic core of the membrane.

Penetration of 2,4,5-trinitrobenzenesulfonate into human erythrocytes. Consequences for studies on phospholipid asymmetry.

The glutathione content of human erythrocytes rapidly diminishes when cells are exposed to 2,4,6-trinitrobenzenesulfonate (20 mumol/l cells) at 37 degrees C. Even at 0 degrees C a slow decrease in glutathione content is observed. The uptake of trinitrobenzenesulfonate by the cells is retarded by inhibitors of the inorganic anion exchange system, indicating that trinitrobenzenesulfonate enters the cells by this pathway. The disappearance of glutathione most probably results from the reaction: 2 GSH + trinitrobenzenesulfonate leads to GSSG + aminodinitrobenzenesulfonate. The reaction of trinitrobenzenesulfonate with glutathione occurs prior to its covalent binding to amino groups of hemoglobin which makes this reaction a more sensitive method of detection of penetration of trinitrobenzenesulfonate into erythrocytes. Results of studies on the asymmetric distribution of phospholipids using trinitrobenzenesulfonate as the only probe should be reconsidered in the light of these new data.

Leak formation in human erythrocytes by the radical-forming oxidant t-butylhydroperoxide.

Oxidative damage of human erythrocytes by the lipoperoxide analogue, t-butylhydroperoxide, has been characterized with regard to ion-permeable leaks formed in the membrane. The formation of these leaks is not correlated with oxidative denaturation of hemoglobin and its precipitation at the membrane. It is also not related to the oxidation of membrane protein SH-groups. A close, although not simply proportional correlation could be demonstrated between leak formation and phospholipid peroxidation as monitored by occurrence of malondialdehyde. The two processes showed similar dependences on exposure time, concentration and temperature. Both were stimulated by the addition of azide as a ligand of ferric heme iron, and suppressed by the anti-oxidant, butylated hydroxytoluene. The leak pathway permits solute permeation with a temperature dependency of bulk diffusion in water and discriminates nonelectrolytes according to size. Discrimination among alkali chlorides corresponds to their free solution mobility; sodium halides are discriminated more effectively. Apparent radii of about 0.5-0.7 nm can be assigned to the defects, while apparent numbers of defects per cell as low as 0.1-0.2 suggest that the defects are dynamic in nature.

Progressive oxidative membrane damage in erythrocytes after pulse treatment with t-butylhydroperoxide.

Development of membrane damage in erythrocytes in the presence of the radical-forming oxidant t-butylhydroperoxide is a well established fact (see, for example, Deuticke et al. (1986) Biochim. Biophys. Acta 854, 169-183). We have now demonstrated that a mere pulse treatment of erythrocytes (5-15 min) with this agent leads to subsequent development of progressive oxidative membrane damage in spite of the absence of exogenous oxidant. Damage comprises the occurrence of ion leakiness and subsequent colloid-osmotic lysis, enhancement of the transbilayer mobility of phospholipid analogues, and lipid peroxidation. There is, however, only very little concomitant oxidation and precipitation of hemoglobin. Defect formation is not due to oxidation of SH-groups nor is it directly related to lipid peroxidation, since it can be suppressed by thiourea without concommitant inhibition of lipid peroxidation. This 'spontaneous' development of membrane damage can be antagonized by metabolic substrates and by desferrioxamine, indicating that lack of protective metabolic resources as well as the presence of catalytic metal (iron) complexes are required for the development of membrane damage. This progressive development of injury in cells only temporarily exposed to an exogenous oxidant may be regarded as a more appropriate model for oxidative membrane damage under pathophysiological conditions in vivo than cells exposed to continuous damage by exogenous oxidants.

Translocation of oleic acid across the erythrocyte membrane. Evidence for a fast process.

To clarify divergent views concerning the mechanism of fatty acid translocation across biomembranes this issue was now investigated in human erythrocytes. Translocation rates of exogenously inserted radioactive oleic acid across the membrane of native cells were derived from the time-dependent increase of the fraction of radioactivity becoming non-extractable by albumin. No accumulation of non-extractable unesterified oleic acid occurred. The rate of transfer was markedly suppressed by SH-reagents and by ATP-depletion. The suppression, however, resulted from a mere decrease of incorporation of oleic acid into phospholipids and was not accompanied by an increase of non-extractable unesterified oleic acid. These findings were reconcilable with the concept of a slow, possibly carrier-mediated fatty acid transfer as well as a very fast presumably, diffusional process not resolvable by the albumin extraction procedure. This ambiguity was resolved by using resealed ghosts, which are unable to incorporate oleic acid into phospholipids. In such ghosts all of the oleic acid inserted into the membrane remains extractable by albumin even after prolonged incubation. On the other hand, ghosts containing albumin accumulated non-extractable oleic acid. The rate of accumulation was beyond the time resolution of the albumin extraction procedure at 4 degrees C. Oleic acid uptake into albumin-containing ghosts became kinetically resolvable when the fatty acid was added as a complex with albumin. Correspondingly, time-resolvable release of oleic acid, originally complexed to internal albumin, into an albumin-containing medium was demonstrated at 4 degrees C. Rate and extent of these redistributions of oleic acid were dependent on the concentrations of internal and external albumin. This indicates limitation by the dissociation of oleic acid from albumin and not its translocation across the membrane. Translocation of oleic acid, which is probably a simple diffusive flip-flop process, must therefore occur with a half-time of less than 15 s. These findings raise doubts on the physiological role of presently discussed concepts of a carrier-mediated translocation of fatty acids across plasma membranes.

Extensive electroporation abolishes experimentally induced shape transformations of erythrocytes: a consequence of phospholipid symmetrization?

As shown in earlier work (M.M. Henszen et al., Mol. Membr. Biol. 14 (1997) 195-204), exposure of erythrocytes to single brief electric field pulses (5-7 kV cm(-1)) enhances the transbilayer mobility of phospholipids and produces echinocytes which can subsequently be transformed into stomatocytes in an ATP-dependent process. These shape transformations arise from partly reversible changes of the transbilayer disposition of phospholipids, in agreement with the bilayer couple concept. Extensive membrane modification by repetitive (

Acceleration of phospholipid flip-flop in the erythrocyte membrane by detergents differing in polar head group and alkyl chain length.

The detergents, alkyltrimethylammonium bromide, N-alkyl-N, N-dimethyl-3-ammonio-1-propanesulfonate (zwittergent), alkane sulfonate, alkylsulfate, alkyl-beta-D-glucopyranoside, alkyl-beta-D-maltoside, dodecanoyl-N-methylglucamide, polyethylene glycol monoalkyl ether and Triton X-100, all produce a concentration-dependent acceleration of the slow passive transbilayer movement of NBD-labeled phosphatidylcholine in the human erythrocyte membrane. Above a threshold concentration, which was well below the CMC and characteristic for each detergent, the flip rate increases exponentially upon an increase of the detergent concentration in the medium. The detergent-induced flip correlates with reported membrane-expanding effects of the detergents at antihemolytic concentrations. From the dependence of the detergent concentration required for a defined flip acceleration on the estimated membrane volume, membrane/water partition coefficients for the detergents could be determined and effective detergent concentrations in the membrane calculated. The effective membrane concentrations are similar for most types of detergents but are 10-fold lower for octaethylene glycol monoalkyl ether and Triton X-100. The effectiveness of a given type of detergent is rather independent of its alkyl chain length. Since detergents do not reduce the high temperature dependence of the flip process the detergent-induced flip is proposed to be due to an enhanced probability of formation of transient hydrophobic structural defects in the membrane barrier which may result from perturbation of the interfacial region of the bilayer by inserted detergent molecules.

Fast translocation of phosphatidylcholine to the outer membrane leaflet after its synthesis at the inner membrane surface in human erythrocytes.

The translocation rate of [14C]phosphatidylcholine to the outer membrane leaflet of human erythrocytes after its primary synthesis from lysophosphatidylcholine by acylation with 14C-labeled oleic or palmitic acid in the inner leaflet has been measured by following the time-dependent increase of cleavability of 14C-labeled phospholipids by external phospholipase A2 (5 min, 37 degrees C). Immediately after a short acylation time period of 10 min about 20% of the newly synthesized [14C]phosphatidylcholine are already detectable in the outer leaflet. After an incubation of 1 h at 37 degrees C following 10 min of acylation the fractions of labeled and native phosphatidylcholine accessible to the lipase are identical, which demonstrates that [14C]phosphatidylcholine has attained the same asymmetric distribution as its endogenous analogue. The calculated halftime of the outward translocation is about 20 min and its activation energy is low, 30 kJ/mol. Translocation is inhibited by a 5 min treatment with phenylglyoxal following acylation. A fast translocation is not observed for newly synthesized phosphatidylethanolamine. Results suggest a selective, protein-mediated outward translocation of newly synthesized phosphatidylcholine.

Influence of enzymatic phospholipid cleavage on the permeability of the erythrocyte membrane. I. Transport of non-electrolytes via the lipid domain.

In order to further elucidate the influence of membrane lipids on transport via the lipid domain of the erythrocyte membrane, simple non-electrolyte diffusion was investigated by tracer flux measurements in whole cells after cleavage of up to 65% of phosphatidylcholine or sphingomyelin by phospholipase A2 from Naja naja, or by sphingomyelinase. A new type of labelled model non-electrolyte was used in this study, readily available by reacting a non-labelled thiol with a labelled alkylating SH-reagent. In spite of the marked enzymatic alterations of the membrane, which lead to the occurrence of large quantities of lysophosphatidylcholine and long chain fatty acids, or of ceramide, the permeability of the lipid domain remained unaffected. This finding is very surprising, since the physical properties of the lipid phase (microviscosity, structure of the membrane interface) are likely to be perturbed in the enzyme-treated membranes. Sphingomyelinase-treated cells undergo stomatocytic shape changes followed by deep invaginations of the membrane and finally endocytosis, while phospholipase A2-treated cells essentially maintain their normal shape.

Influence of enzymatic phospholipid cleavage on the permeability of the erythrocyte membrane. II. Protein-mediated transfer of monosaccharides and anions.

In order to investigate the influence of membrane lipids on transport via the protein domain of the erythrocyte membrane, a number of facilitated diffusion processes was studied by tracer flux techniques in whole cells after cleavage of up to 65% of the phosphatidylcholine or the sphingomyelin by phospholipase A2 from Naja naja or bee venom, or by sphingomyelinase, respectively. The mediated fluxes of L-arabinose, which is transported by the glucose carrier, and of L-lactate, which uses a specific monocarboxylate carrier, were markedly inhibited by cleavage of either phosphatidylcholine or sphingomyelin. These phospholipid dependencies are in line with earlier data on cholesterol dependencies (Deuticke, B. (1977) Rev. Physiol. Biochem. Pharmacol. 78, 1-97). They can only in part be explained by changes of membrane fluidity. More specific interactions of the degradation products with the carrier proteins seem also to play a role. Sulfate and oxalate transfer, which proceed via the inorganic anion-exchange system, are essentially unaffected by cleavage of phosphatidylcholine and less sensitive to sphingomyelin cleavage than the two other processes. This also agrees with earlier data on cholesterol independency of sulfate transfer. The inorganic anion-exchange protein thus seems to be less dependent on the surrounding lipids in its conformation and its mode of action than the two other carriers.

Bacterial cytotoxins, amphotericin B and local anesthetics enhance transbilayer mobility of phospholipids in erythrocyte membranes. Consequences for phospholipid asymmetry.

Incorporation of the channel-forming polyene antibiotic amphotericin B and of cytotoxins from Staphylococcus aureus (alpha-toxin) or Pseudomonas aeruginosa into erythrocyte membranes results in a concentration-dependent enhancement of the flip rates of exogenous lysophosphatidylcholine. The flip rate is also enhanced by incorporation of tetracaine and dibucaine. Removal of tetracaine and amphotericin B from the cells normalizes the flip rates. In parallel to the enhancement of flip rates, alpha-toxin produces a loss of transmembrane asymmetry of both phosphatidylethanolamine and phosphatidylserine. Pretreatment of cells with amphotericin or high concentrations (over 2.5 mmol . l-1) of tetracaine, followed by removal of the perturbing agent by washing, produces a selective loss of the asymmetric orientation of phosphatidylethanolamine to the inner membrane layer, as evaluated by the accessibility of the lipid towards cleavage by phospholipase A2. The extent to which asymmetry is lost depends on the time of pretreatment with amphotericin or tetracaine, indicating a limitation by the rate of reorientation of phosphatidylethanolamine to the outer membrane surface. Evaluation of the accessibility of phosphatidylethanolamine towards cleavage by phospholipase A2 in the presence of local anesthetics indicates accessible fractions much higher than those obtained after removal of the perturbant. In the presence of tetracaine, endofacial phosphatidylethanolamine seems somehow to become accessible to phospholipase A2. Phosphatidylserine does not exhibit this peculiarity. The results indicate that various types of perturbation of the lipid domain of the erythrocyte membrane may enhance the transbilayer mobility of phospholipids as well as destabilize the asymmetric distribution of aminophospholipids. However, as in other instances reported previously (Haest, C.W.M., Erusalimsky, J., Dressler, V., Kunze, I. and Deuticke B. (1983) Biomed. Biochim. Acta 42, 17-21), there is no tight coupling between transbilayer mobility and destabilization of asymmetry of the transbilayer distribution of phospholipids.