Search

To investigate the relative importance of catalase and glutathione in erythrocyte oxidant defense, human and mouse (normal and acatalasemic) erythrocytes were reversibly lysed and resealed in the presence of exogenous catalase or glutathione. This resulted in an increase in intracellular catalase activity or glutathione concentration in the resealed erythrocytes while normal cellular structure, hemoglobin concentration, cell volume, cellular deformability, and adenosine triphosphate concentration were maintained. Resealing alone had no effect on oxidant sensitivity. In human cells, a threefold increase in catalase activity resulted in the maintenance of glutathione levels in response to hydrogen peroxide (H2O2) challenge. Reconstitution of congenitally acatalasemic mouse erythrocytes, which were extremely sensitive to even micromolar concentrations of H2O2 with purified catalase resulted in complete protection against H2O2. Indeed, the catalase-reconstituted acatalasemic cells were less sensitive to H2O2-mediated damage than were normal, catalase-replete mouse cells. In contrast, alteration of the glutathione status of human and mouse (normal and acatalasemic) cells had no significant effect on oxidant sensitivity. Even a five-fold increase in intracellular glutathione concentration (greater than 30 micromoles glutathione per gram of hemoglobin) in normal or catalase-deficient (azide-treated or acatalasemic) red blood cells had no protective effect against H2O2-mediated lipid peroxidation or methemoglobin generation. Similarly, depletion of glutathione by 1-chloro-2,4-dinitrobenzene also had no effect on erythrocyte H2O2 sensitivity. These results suggest an important role for catalase in protection against H2O2-mediated damage at physiologic levels and that catalase is as at least as important as glutathione in cellular defense against H2O2.

We report the application of high-performance liquid chromatographic (HPLC) separation with ultraviolet detection and direct, on-line, structural analyses by mass spectrometry of glycerobenzoate derivatives from complex mixtures of phospholipid molecular species. Individual phospholipids were resolved from total lipid extracts by thin-layer chromatography (TLC). Diradylglycerols were released from phospholipids by phospholipase-C treatment, converted to diradyl glycerobenzoates and subsequently separated by TLC into subclasses (alk-1-enylacyl, alkylacyl and diacyl types). The molecular species within each subclass were resolved by HPLC with an octadecyl reversed-phase column in acetonitrile-isopropanol (80:20, v/v). Individual peaks were quantitated at the picomole level by measuring absorbance at 230 nm. After post-column addition of methanol-0.2 M ammonium acetate (50:50, v/v), peaks were introduced through the thermospray interface into a VG Masslab 30-250 quadrupole mass spectrometer. Molecular species showed as base peaks the salt adducts of the molecular ion which permitted easy deduction of the overall fatty acyl composition. In addition, the diglyceride fragment of each species was found at [MH - 122]+ and two fragments formed by the loss of the fatty acyl groups (R) in the sn-1 or sn-2 position were found at [M - R1]+ and [M - R2]+, respectively. Since preferential release of either fatty acyl group was observed in positional isomers, the ratio of the intensity of these fragments gave information on the position of the fatty acyl groups in the individual HPLC peaks. We show that the use of on-line mass spectrometry, however, provides easy identification of all molecular species present in a complex phospholipid mixture, even when more than one molecular species is contained in an HPLC peak.

Having minimized spectroscopic interference by hemoglobin (Hb), peroxidation processes in intact erythrocytes could be monitored in a continuous assay using the fluorescent polyunsaturated fatty acid, parinaric acid (PnA), as a peroxidation probe. Control experiments to establish the character of the method are described in detail. As a practical application, comparative studies were performed to monitor the response of normal and sickle Hb-containing human erythrocytes to oxidative stress in the PnA assay. After 10 min of incubation with 200 microM cumene hydroperoxide (cumOOH), peroxidation of PnA was found to be enhanced in erythrocytes from sickle cell disease patients (SS: 48 +/- 9% (n = 6) of initial amount had been peroxidized) compared to healthy controls (AA: 30 +/- 4% (n = 9)). PnA peroxidation in erythrocytes from sickle cell trait individuals (AS: 30 +/- 3% (n = 4)) was equal to that in control cells. The increased oxidation of PnA in sickle erythrocytes was accompanied by enhanced oxidation of Hb (metHb and hemichrome formation), indicating that sickle Hb mediates enhanced cumOOH-derived radical generation. It is concluded that PnA can be a useful tool in studying membrane peroxidation processes in intact normal and pathological erythrocytes.

In view of a possible relationship between fish oil, lipid peroxidation, and atherosclerosis, the in vitro lipid peroxidation susceptibility of red blood cells (RBCs) from rabbits on conventional (-FO) and fish oil-enriched diets (+FO) was investigated. The diet caused substantial increases in the RBC concentrations of n-3 polyunsaturated fatty acids (PUFAs), in combination with decreases in the concentration of oleic acid (18:1) and linoleic acid (18:2). Cumene hydroperoxide-induced oxidative stress led to increased overall fatty acid peroxidation in +FO RBCs compared with with -FO RBCs, as quantitated by GLC fatty acid analysis. However, the increased overall susceptibility to lipid peroxidation of +FO RBCs was not reflected in increased peroxidation of every individual fatty acid. This was observed for endogenous arachidonic acid (20:4) as well as, in separate experiments, for exogenously added parinaric acid (PnA). The increased cumene hydroperoxide-induced PUFA oxidation in +FO RBCs was accompanied by a lesser extent of hemolysis. To account for these observations, it is proposed that the increased n-3 PUFA content of +FO RBCs serves as an oxidizable buffer. The present data suggest that oxidation of fatty acids can occur until a critically low level of intact phospholipid in the RBC membrane is reached, after which the membrane destabilizes and hemolysis occurs. At the same time, the PUFA buffer in +FO RBCs could also prevent oxidative damage to specific membrane proteins, which could also help prevent cell lysis.

We analyzed the molecular species composition of the glycosylphosphatidylinositol (GPI) anchor of Torpedo marmorata acetylcholinesterase (AChE) and compared it to that of the membrane phosphatidylinositol (PI) as well as the other major phospholipid classes of T. marmorata electrocytes. Purified amphiphilic AChE was treated with PI-specific phospholipase C in order to release the diradylglycerol moiety from the membrane anchoring domain. Subsequently, the diradylglycerols were derivatized into their diradylglycer-obenzoates and separated into subclasses (diacyl, alkylacyl, and alk-1-enylacyl types). The molecular species within each subclass were separated and quantitated by high performance liquid chromatography and UV detection and directly introduced through the thermospray interface into a mass spectrometer for identification. The PI moiety of the GPI anchor of AChE consisted exclusively of diacyl molecular species. Over 85% of the molecular species were composed of palmitoyl (16:0), stearoyl (18:0), and oleoyl (18:1) fatty acyl chains in the sn-1 and sn-2 positions. Less than 5% of the molecular species of the GPI anchor contained polyunsaturated fatty acyl chains, as compared to more than 70% of the diacyl molecular species of the PI from electrocyte membranes. Since the GPI anchors of AChE from both human and bovine erythrocytes contain alkylacyl molecular species of PI (Roberts, W. L., Myher, J. J., Kuksis, A., Low, M. G., and Rosenberry, T. L. (1988) J. Biol. Chem. 263, 18766-18775), our results on AChE from Torpedo demonstrate that the composition of the PI moiety of the GPI anchor of a protein is not characteristic for that protein but may vary between species.

To define a more sensitive and reliable method to determine changes in the overall cellular characteristics of erythrocytes after oxidative damage, we used a viscodiffractometric method (ektacytometry) to measure the effect of oxidative stress. Erythrocytes were incubated in the presence of hydrogen peroxide, t-butyl hydroperoxide, or cumene hydroperoxide in phosphate buffer. This treatment resulted in decreased cellular deformability of the intact erythrocytes. In addition, deformability and fragility measurements of the erythrocyte ghost membranes indicated an increased membrane dynamic rigidity and altered-mechanical stability as a consequence of oxidant stress. These changes were observed before the onset of hemolysis. The observed decrease in deformability was accompanied by oxidation of hemoglobin, alterations of membrane proteins, and lipid peroxidation. To continuously measure the time course of the decrease in deformability in intact erythrocytes under oxidative stress, a new ektacytometric method was developed. Erythrocytes were oxidatively challenged within the viscometer at a constant osmolality and shear stress. The change in deformability was monitored and a typical range was defined for erythrocytes from normal individuals. Comparison of erythrocytes from patients with sickle cell disease with those from normal individuals demonstrated a higher susceptibility of sickle red cells toward oxidative stress.

The present studies describe the distribution of phosphatidylinositol (PI) within the membrane bilayer of the human red blood cell (RBC) as well as its transbilayer mobility. The membrane bilayer distribution was determined by measuring the hydrolysis of PI in the exterior leaflet of the RBC membrane using a PI-specific phospholipase C and by extraction of PI from the exterior leaflet using bovine serum albumin. The transbilayer mobility of PI was measured by following the fate of radiolabeled PI which was first incorporated into the outer leaflet of the RBC membrane. Our results indicate that PI is asymmetrically distributed in the membrane, with approximately 80% located in the inner and 20% in the outer leaflet of the bilayer. The rate of transbilayer mobility of PI is similar to that for certain molecular species of phosphatidylcholine and much slower than that reported for the aminophospholipids in the RBC membrane.

Desferroxamine has been shown to exhibit potent antimalarial activity. However, it is unclear as to whether desferroxamine functions by the chelation of extracellular, intra-erythrocytic, or parasite-associated iron. In order to determine desferroxamine's site of action, we have employed a large molecular weight dextran derivative of desferroxamine (70 kDa) and a reversible osmotic lysis technique by which erythrocytes were intracellularly loaded with this chelator. The desferroxamine-dextran derivative has virtually identical iron-binding characteristics to desferroxamine but, unlike desferroxamine, it is unable to cross the erythrocyte membrane. As previously shown, desferroxamine added to culture media exhibited potent antimalarial activity (mean effective inhibitory dose (ED50) approximately 6 microM). However, extracellular desferroxamine-dextran showed antimalarial activity only at very high doses (ED50 greater than or equal to 180 microM), indicating that extracellular iron chelation is not involved in the antimalarial activity of desferroxamine. The intra-erythrocytic entrapment of the desferroxamine-dextran derivative also had no significant effect, except at very high concentrations, demonstrating that desferroxamine does not remove a non-haem iron source necessary for malarial replication. The results of this study clearly suggests that the antimalarial activity of desferroxamine is directly related to its ability to enter the parasitic compartment and not due to the chelation of extra- or intra-erythrocytic iron pools necessary for malarial growth.

Previous descriptions of hereditary high-phosphatidylcholine hemolytic anemia (HPCHA) have highlighted the association of increased erythrocyte membrane phosphatidylcholine with abnormal membrane cation permeability. We studied the function and composition of erythrocyte membranes from three individuals with HPCHA to characterize further the membrane abnormalities in this disorder. Despite significant macrocytosis, HPCHA red cells were dehydrated and showed an increased surface area to volume ratio compared to normal red cells. The passive efflux of K+ from HPCHA erythrocytes was increased fourfold at 37 degrees C. Total membrane phospholipid was increased 7-42%, largely due to excess phosphatidylcholine, which made up 35.8-37.2% of total phospholipid. Membrane cholesterol:phospholipid ratios were in the normal range. It appears that the excess phosphatidylcholine was not acquired during circulation, since plasma lipids were normal and all subpopulations of density-separated HPCHA erythrocytes were similarly abnormal. The ratio of total protein to phospholipid in white ghosts was increased, indicating that membrane protein was increased to an even greater extent than membrane lipids. No abnormal membrane proteins were identified by Coomassie or periodic acid Schiff (PAS) staining. Quantitation of the major membrane proteins indicated that the total protein excess in HPCHA membranes was due to a proportional increase in all major proteins. We conclude that HPCHA erythrocytes have excess membrane proteins and hypothesize that the changes in lipid composition and cation permeability are secondary to underlying protein abnormalities, which remain to be defined.

We have recently modified the dialysis tubing osmotic lysis and resealing method to examine the role of intracellular red blood cell (RBC) antioxidants. However, the potential effect of resealing on the RBC was not fully investigated. This study examined a number of cellular characteristics to determine the effects of physical lysis and resealing on the RBC. Following resealing, RBC exhibited normal morphology and at most only slight alterations in mean cell volume and mean cell hemoglobin concentration. RBC density distribution was significantly affected by resealing with increased populations of both light and dense cells, though the mean cell density was similar to that of control cells. Endogenous enzyme activities and adenosine triphosphate (ATP) concentration were unaffected by the resealing procedure. While reduced glutathione (GSH) concentration was decreased by 15%, RBC oxidant sensitivity was found to be unaltered. Cellular deformability of the resealed RBC was 80% to 90% that of the control cells. Membrane phospholipid and fatty acyl composition of the resealed RBC were unaffected when compared with matched control samples. Membrane transport, permeability, and Ca2(+)-mediated cellular vesiculation were minimally altered by resealing. Finally, entrapment of fluorescent compounds demonstrated that greater than 95% of the resealed RBC had incorporated exogenous agents. In summary, the osmotic lysis and resealing method described resulted in only minor changes in cellular characteristics while allowing for the efficient loading of compounds to which the RBC membrane is normally impermeable. Consequently, this method provides great potential for the selective modification of erythrocyte constituents in order to further define their roles within the RBC.

The influence of vitamins E and C on the initial stages of lipid peroxidation in human erythrocyte membranes was assessed with the fluorescent polyunsaturated fatty acid, parinaric acid, as probe molecule. Cumene hydroperoxide was used as initiator with either haemin-Fe3+ or Cu2+ as metal ion cofactor. The effect of vitamin C (pro- or antioxidant) appeared to be determined by the localisation of the metal ions, either in the water phase or in the membrane. Vitamin C is only able to reduce metal ions in the water phase, which results in acceleration of radical generation and subsequent enhancement of parinaric acid peroxidation. Thus, interaction of vitamin C with Cu2+ in the water phase led to drastically enhanced peroxidation of parinaric acid. In contrast, when only membrane-associated haemin-Fe3+ was present, vitamin C functioned as an antioxidant at all concentrations tested (0-10 microM). In a system with haemin-Fe3+ equilibrated between the water phase and the membranes, less than 5 microM vitamin C produced an overall prooxidant, and greater than 15 microM vitamin C an overall antioxidant effect. At vitamin C concentrations of 5-15 microM, continuous measurement of parinaric acid fluorescence revealed a shift in the vitamin C effect from antioxidant to prooxidant within the time-course of an assay. Vitamin E exhibited a protective effect on peroxidation initiated by cumene (per)oxyl radicals with haemin-Fe3+ as cofactor, by inducing a concentration-dependent extension of the lag-phase in parinaric acid peroxidation. Vitamin E appeared to be much more effective compared with vitamin C in scavenging radicals in this system. This indicates that vitamin C has only a limited ability to react with cumene (per)oxyl radicals in the membrane. The combination of vitamins E and C produced a protective effect on parinaric acid peroxidation exceeding the sum of their individual contributions. Moreover, the rate of vitamin E consumption was drastically lowered in the presence of vitamin C, whereas the rate of vitamin C consumption hardly decreased in the presence of vitamin E. The results are discussed in terms of a reaction scheme where the relative contributions of a number of reactions are considered to determine the total effect of added vitamin C or E. Vitamin E radicals constitute an additional substrate for vitamin C, resulting in a more than additive shift in the overall effect to the antioxidant side.

Qinghaosu, a potent antimalarial agent, has recently been shown to act via oxidative mechanisms. Hence, we have investigated what effect qinghaosu has on cellular oxidation in normal and oxidant-sensitive red blood cells (RBCs). At 500 mumol/L, qinghaosu was found to directly alter red cell deformability (DI) in both normal (hemoglobin AA) and abnormal (hemoglobins SS, AE, and EE) RBCs, with the maximum DI being 70% to 80% of that of untreated RBCs. Although concentrations of less than or equal to 200 mumol/L qinghaosu had a minimal effect on the maximum DI, qinghaosu was found to act as an efficient prooxidant at these concentrations. Cellular deformability was lost more rapidly in response to exogenous oxidants in the qinghaosu-treated RBCs than in control cells. Hemoglobin SS and EE RBCs pretreated with qinghaosu demonstrated a much more rapid decrease in cellular deformability than did the control RBCs in response to exogenous oxidants. Additionally, qinghaosu resulted in a dose-dependent increase in red cell lysis and methemoglobin generation while decreasing reduced glutathione concentration. As a consequence of qinghaosu challenge, a decrease in unsaturated fatty acids was noted. Deoxyqinghaosu, which lacks the endoperoxide bridge and is pharmacologically inactive, did not affect cellular deformability and did not function as a prooxidant. Additionally, deoxyqinghaosu had no effect on fatty acid composition, red cell lysis, methemoglobin generation, or reduced glutathione concentration. In conclusion, although the oxidative effects of qinghaosu on uninfected erythrocytes were only seen at concentrations much greater than that necessary for antimalarial activity, these results confirm the proposition that qinghaosu may act via oxidative mechanisms. Furthermore, qinghaosu-mediated oxidation was significantly increased in erythrocytes characterized by enhanced oxidant sensitivity caused by unstable hemoglobins.

To delineate further the role of superoxide dismutase (SOD) in red blood cell (RBC) oxidant defense, normal human erythrocytes were osmotically lysed and resealed in the presence of varying concentrations of exogenous SOD. This resulted in a dose-dependent increase in SOD activity in the resealed erythrocytes while maintaining nearly normal RBC hemoglobin concentration (less than 10% decrease from the control value), cell volume, and cellular deformability. Surprisingly, a five- or ninefold increase in SOD activity yielded no additional protection against superoxide-generating drugs (phenazine methosulfate or menadione sodium bisulfite). No significant differences were observed between the control and SOD-loaded RBCs in O2-driven methemoglobin formation or generation of thiobarbituric acid-reactive substances. In contrast, RBCs with elevated SOD activity pretreated with sodium azide (to block catalase activity) or 1-chloro-2,4-dinitrobenzene (to deplete reduced glutathione, GSH) showed significantly enhanced methemoglobin generation in response to superoxide generating drugs. No differential response was noted between the control, control-resealed, and SOD-loaded RBCs to oxidants other than superoxide. Based on our results and other data, we conclude that elevated SOD activity may imbalance cellular oxidant defense, resulting in enhanced oxidation due to the accelerated generation of H2O2, the product of O2- dismutation. This effect is significantly exacerbated under conditions in which H2O2 catabolism is altered.

The decrease in fluorescence of conjugated polyenic acyl chains is used as a sensitive assay for lipid peroxidation. The fatty acid cis-trans-trans-cis-9,11,13, 15-octadecatetraenoic acid (cis-parinaric acid) is introduced into liposomal membranes as free fatty acid or, by using the PC specific transfer protein from bovine liver, as 1-palmitoyl-2-cis-parinaroyl-sn-glycero-3-phosphocholine. The peroxidation process as monitored by the decrease in fluorescence intensity is compared with other peroxidation assay systems. Applications of the new assay system are discussed.

Several proteins are attached to the cell membrane by a glycosyl-phosphatidylinositol (GPI) anchor. In this report, we show that during vesiculation of human RBCs in vitro, two of these proteins, acetylcholinesterase and decay accelerating factor, redistribute on the cell surface and become enriched in the released vesicles. As a result, the remnant cells are depleted of these proteins. We suggest that alterations in the architecture of the RBC membrane that precede vesiculation lead to selective polarization of GPI-anchored proteins within the domain of the membrane destined to become a vesicle. Since vesiculation occurs in many cell types, and if the loss of GPI-anchored proteins accompanies this process, it may have important biologic significance.

The phospholipid composition, the distribution of phospholipids over the two membrane layers as well as the phosphatidylcholine-specific transfer protein-mediated exchangeability of phosphatidylcholine from the membrane, has been investigated in two types of abnormal erythrocytes--the McLeod phenotype and the Leach phenotype. The acanthocytic McLeod cells appeared to have a normal phospholipid composition and distribution, but the exchangeability of phosphatidylcholine was found to be markedly enhanced. Unlike control erythrocytes, in which 75% of all of the phosphatidylcholine can be exchanged during an 8 h incubation, the McLeod cell showed a complete exchange of this phospholipid within the same time period. This obviously indicates an enhanced transbilayer mobility of phosphatidylcholine in the membrane of McLeod cells. Erythrocytes of the Leach phenotype showed an elliptocytic shape and increased osmotic fragility, but no abnormalities were observed as to the composition and organisation of the phospholipid complement of their membranes.

The recently developed parinaric acid assay is shown to offer possibilities for studying peroxidation processes in biological membrane systems. Taking the human erythrocyte membrane as a model, several initiating systems were investigated, as well as the effect of residual hemoglobin in ghost membrane preparations. The effectivity of a radical generating system appeared to be strongly dependent upon whether radicals are generated at the membrane level or in the water phase. Thus, cumene hydroperoxide at concentrations of 1.0-1.5 mM was found to be a very efficient initiator of peroxidation in combination with submicromolar levels of hemin-Fe3+ as membrane-bound cofactor. In combination with cumene hydroperoxide, membrane-bound hemoglobin appeared to be about 6-times more effective in promoting peroxidation than hemoglobin in the water phase. Results comparing the behaviour of normal and sickle erythrocyte ghost suspensions in the peroxidation assay suggest that the increased oxidative stress on sickle erythrocyte membranes could be due to enhanced membrane binding of sickle hemoglobin, but also partly to a characteristically higher capability of sickle hemoglobin to promote peroxidation. The order of peroxidation-promoting capabilities that could be derived from the experiments was hemin greater than sickle hemoglobin greater than normal hemoglobin.

The phosphatidylcholine specific transfer protein (PCTP) from bovine liver was used to retailor the molecular species composition of phosphatidylcholine (PC) in the membrane of normal (AA) and sickleable (SS) human erythrocytes. Changes in molecular species composition of PC altered morphology as well as cellular deformability and stability as measured with ektacytometry. In normal cells, replacement of native PC with 1-palmitoyl,2-arachidonoyl PC (PAPC) resulted in a decrease in osmotic fragility with no change in hydration, whereas replacement with 1,2-dipalmitoyl PC (DPPC) led to an increased osmotic fragility and cellular hydration. Replacement of native PC by 1-palmitoyl,2-oleoyl PC (POPC) in normal cells had no apparent effect on these parameters. In contrast, replacement of native PC in sickle cells with either PAPC, DPPC or POPC led to cellular hydration. Facilitation of PC exchange between subpopulations of SS cells separated on buoyant density also led to cellular hydration. These observations suggest that the state of hydration of sickle cells can be modified by the fatty acyl composition of PC and illustrate a a role for the lipid core in the observed permeability changes in sickle erythrocytes. They also raise the interesting possibility that the state of cellular hydration of sickle cells may be modulated by altering the molecular species composition of the membrane phospholipids.

The steady state fluorescence anisotropy (rs) of 1-acyl-2-cis parinaroyl phosphatidylcholine (PnPC) was compared with that of diphenylhexatriene (DPH) in a variety of model- and biological membrane systems. The fluorescence anisotropy of both probes responded similarly to temperature changes and variations in the acyl chain composition in phosphatidylcholine (PC) liposomes. The presence of proteins and cholesterol increased rs for both DPH and PnPC in the biological membranes as compared to the isolated polar membrane lipids. Comparison of DPH and PnPC in dipalmitoyl-PC-liposomes with and without 50 mol% cholesterol, showed at temperatures above the phase transition of pure dipalmitoyl-PC the presence of cholesterol increased the rs-value for DPH strongly, whereas the rs-value for PnPC was much less affected. In the cholesterol-rich erythrocyte membrane as well as in microsomes from Morris hepatoma 7787, which have an increased cholesterol content as compared to normal rat liver microsomes, the rs of DPH was higher than that of PnPC. No large differences between the rs-values of both probes were evident in the normal cholesterol-poor rat liver microsomes. These effects are discussed in terms of structural differences between the probes and variation of cholesterol content. Alterations in the fatty acid composition of PC present in human erythrocyte membranes were introduced with the aid of a PC-specific transfer protein. Fluorescence anisotropy values of both probes hardly changed upon enrichment of the red cell membrane with either dipalmitoyl PC or 1-palmitoyl-2-arachidonyl PC.

This report describes the molecular species composition of phosphatidylcholines (PC) transferred from human erythrocytes to acceptor vesicles composed of cholesterol and single PC species in the presence of PC-specific transfer protein from bovine liver. The compositions of the PC isolated from the vesicles were determined by capillary GLC as the diacylglycerol trimethylsilyl ethers. The cellular PC species appearing in the acceptor vesicles were enriched in unsaturated species and showed a low content of dipalmitoyl PC compared to untreated erythrocytes. This trend was independent of the composition of the PC used to construct the acceptor vesicles and it was possible to determine that the relative rates of efflux of the palmitoyl-containing phosphatidylcholines decreased in the order: palmitoyl-linoleoyl greater than palmitoyl-oleoyl greater than dipalmitoyl and in the stearoyl series, stearoyl-linoleoyl greater than stearoyl-oleoyl. No clear trend was distinguished for the influence of chain-length on the efflux, thus preventing an unambiguous assignment of the order of removal of all species from the cell membrane. Results derived for arachidonoyl-containing species were compromised by evidence for oxidation occurring during incubations at 37 degrees C. To confirm that acyl selectivity was also possible during transfer in the absence of the transfer protein, the efflux of 14C-labeled soya PC and [14C]dipalmitoyl PC from prelabeled erythrocytes was measured using plasma as the acceptor. As predicted by the chromatographic analyses, 14C-labeled soya PC effused up to 10-times faster than [14C]dipalmitoyl PC from the red cell membrane. Thus, the more rapid transfer of unsaturated PC cannot be explained entirely as a specificity of the transfer protein and is consistent with the hypothesis that intermolecular interactions involving PC molecules within the erythrocyte membrane, become weaker with increasing unsaturation. The results suggest a potential role of PC-specific transfer protein as a probe of the nature of PC interactions within biological membranes.

Whole cell deformability and lipid determinations were performed on red cells from two patients who had acanthocytes in the peripheral blood (10% and 20% to 30%) and normal serum lipoprotein levels. One patient had typical chorea-acanthocytosis and the other had no clinical abnormalities associated with acanthocytosis. Red cells from the patient with chorea-acanthocytosis showed reduced deformability, as measured by a visco-diffractometric method (ektacytometry), which could be explained by the presence of increased numbers of dehydrated cells containing high concentrations of hemoglobin. The total cell population showed a modest increase in potassium efflux, which may be responsible for reduced cation content and dehydration in a subpopulation of cells. When the patients' red cells were separated into different density populations by centrifugation on density gradients, the cells of classic acanthocyte morphology were concentrated in the high-density layers. This was true for both patients, although the subject with acanthocytes and no clinical disorder had a normal red cell density distribution. Lipid analysis of both types of acanthocytes showed an increase in the relative proportion of sphingomyelin with respect to the glycerophospholipids. Total cholesterol and phospholipid levels were reduced in the chorea-acanthocytosis red cells, but the other acanthocytes did not differ significantly in total lipid content from normal control samples. Thus, the one common abnormal feature in these two forms of acanthocytosis is the increase in the proportion of red cell sphingomyelin. Because this is a very stable, immobile component of the membrane, we suggest that its relative enrichment could result from a defect in the transport and maintenance of glycerophospholipids. Further study of the dynamics of glycerophospholipid organization in acanthocytes may be useful in increasing our understanding of the genesis of abnormal, acanthocytic morphology.

The effect of dietary fat on the distribution of fatty acids within the bilayer of red cell membrane phosphatidylcholine was determined after feeding young female rats semipurified diets containing 10% of either hydrogenated cottonseed oil, menhaden oil, or corn oil for 4 days. Changes in fatty acyl composition of both plasma and erythrocyte phosphatidylcholine were noted. The composition of the fatty acids in phosphatidylcholine in the outer leaflet of the bilayer was determined after hydrolysis of outer leaflet phospholipids, first with phospholipase A2 and subsequently with sphingomyelinase. The distributions of acyl constituents of phosphatidylcholine in the inner and the outer monolayers of the membrane were found to be distinctively different, regardless of the dietary fat intake. For example, palmitic acid and oleic acid were predominantly present in the phosphatidylcholine of the inner monolayer, and arachidonic acid and docosahexaenoic acid were present in that of the outer monolayer. Despite such diet-induced changes in phosphatidylcholine fatty acid composition, no alteration in erythrocyte integrity was observed when monitored by either morphology or ektacytometry.

Using the phosphatidylcholine specific transfer protein from bovine liver, native phosphatidylcholine from intact human erythrocytes was replaced by a variety of different phosphatidylcholine species without altering the original phospholipid and cholesterol content. The replacement of native phosphatidylcholine by the disaturated species, 1,2-dipalmitoyl- and 1,2-distearoylphosphatidylcholine, proceeded at a low rate and extensive replacement could only be achieved by repeatedly adding fresh donor vesicles. The replacement by disaturated molecules was accompanied by a gradual increase in osmotic fragility of the cells, finally resulting in hemolysis when 40% of the native PC had been replaced. Up to this lytic concentration, the replacement did not affect the permeability of the membrane for potassium ions. Essentially, all of the PC in the outer monolayer of the membrane could be replaced by 1-palmitoyl-2-oleoyl- and 1-palmitoyl-2-linoleoylphosphatidylcholine. These replacements did not alter the osmotic fragility of the cells, nor the K+ permeability of the membrane. Increasing the total degree of unsaturation of the phosphatidylcholine species modified the properties of the membrane considerably. Replacement by 1,2-dilinoleoylphosphatidylcholine resulted in a progressive increase in osmotic fragility and hemolysis started to occur after 30% of the native PC had been replaced by this species. K+ permeability was found to be slightly increased in this case. Cells became leaky for K+ upon the introduction of 1-palmitoyl-2-arachidonoylphosphatidylcholine in the membrane. The increased permeability was also reflected by an apparent increase in the resistance of the cells against osmotic shock. The conclusions to be drawn are that (i) 1-palmitoyl-2-oleoyl- and 1-palmitoyl-2-linoleoylphosphatidylcholine are species which fit most optimally into the erythrocyte membrane; (ii) loss of membrane stability results from an increase in the degree of saturation of phosphatidylcholine (unsaturation index greater than 0.5) and (iii) the permeability is enhanced by increasing the content of highly unsaturated species (unsaturation index greater than 1.0).

The phospholipid composition and the distribution of phospholipids over the two leaflets of the membrane have been investigated for rabbit and horse erythrocyte membranes. Phosphatidylcholine (PC) comprises 39.4% and 41.3% of the total phospholipid complement of the rabbit and horse erythrocytes, respectively. In both membranes the distribution of this phospholipid is asymmetric: 70% of the PC is present in the outer layer of the rabbit membrane and 60% in that of the horse. The major species of this phospholipid class are the (1-palmitoyl-2-oleoyl)- and the (1-palmitoyl-2-linoleoyl)PC. The disaturated species, (1,2-dipalmitoyl)PC, is present in limited amounts only. Partial replacement of the native PC from intact erythrocytes was accomplished with a purified PC specific transfer protein from bovine liver. Replacement of the native PC species with (1-palmitoyl-2-oleoyl)PC up to 40% of the total PC complement had no effect on the osmotic fragility, the shape and the in vivo survival time of both erythrocyte species. Replacement of the native PC in both rabbit and horse erythrocytes with (1,2-dipalmitoyl)PC up to 20% gave rise to an increased osmotic fragility, a shape change from discocytic to echinocytic and a significant reduction in survival time measured after reinjection of the modified cells. At 30% replacement with (1,2-dipalmitoyl)PC the resulting spheroechinocytes appeared to be cleared from the circulation within 24 h after reinjection. The conclusion can be drawn that the repair mechanisms which may exist in vivo are insufficient to cope with the drastic changes in properties of the erythrocyte membrane which are induced by replacing more than 15% of the native PC by the dipalmitoyl species.

In a previous report it was shown that the replacement of native erythrocyte phosphatidylcholine (PC) with different PC species which have defined acyl chain compositions can lead to morphological changes (Kuypers, F.A., W. Berendsen, B. Roelofsen, J. A. F. Op den Kamp, and L.L.M. van Deenen, 1984, J. Cell Biol., 99:2260-2267). It was proposed that differences in molecular shape between the introduced PC species and normal erythrocyte PC caused the membrane to bend outwards or inwards, depending on the shape of the PC exchanged. To support this proposal, two requirements would have to be fulfilled: the exchange reaction would take place only with the outer lipid monolayer of the erythrocyte, and the extent of lipid transbilayer movement would be restricted. If this theory is correct, any treatment causing unilateral changes in lipid molecular shape should lead to predictable morphological changes. Since this hypothesis is a refinement of the coupled bilayer hypothesis, but so far lacks experimental support, we have sought other means to change lipid molecular shape unilaterally. Shape changes of human erythrocytes were induced by the replacement of native PC by various PC species using a phosphatidylcholine-specific transfer protein: by hydrolysis of phospholipids in intact cells using sphingomyelinase C or phospholipase A2, and by the combination of both procedures. The morphological changes were predictable; additive when both treatments were applied, and explicable on the basis of the geometry of the lipid molecules involved. The results strongly support the notion that lipid molecular shape affects erythrocyte morphology.

We have studied the erythrocyte membrane phospholipid organization in hereditary high red cell membrane phosphatidylcholine hemolytic anemia (HPCHA) and the response of these red cells during incubation with sonicated suspensions of dimyristoylphosphatidylcholine (DMPC). Although both the absolute and relative amounts of phosphatidylcholine were elevated in these red cells, the relative distribution of phosphatidylcholine on both sides of the membrane bilayer and the transbilayer mobility of phosphatidylcholine were normal. HPCHA erythrocytes showed elevated absolute amounts per cell of both protein and lipid and an increased cellular ratio of protein to phospholipid. Incubation of normal red cells with DMPC led to the formation of echinocytes, followed by the release of acetylcholinesterase-containing vesicles. Both echinocyte formation and vesiculation were markedly reduced in red cells from patients with HPCHA. Studies with red cells from patients with liver disease, cells that also have elevated relative amounts of membrane phosphatidylcholine and increased amounts of lipid per cell, revealed normal echinocyte formation and normal DMPC-induced vesiculation. We conclude that the altered lipid composition of HPCHA erythrocytes per se is not responsible for the observed reduction in DMPC-induced vesiculation, but that it is more likely the result of a modification in the protein moiety of these cells. This putative protein abnormality could enhance binding of phosphatidylcholine to red cell membranes and could explain the elevated phosphatidylcholine content of HPCHA erythrocytes and their inability to vesiculate.

To delineate further the underlying mechanism by which amphiphilic drugs can modulate vesicle release from human RBCs, we studied the effect of chlorpromazine on erythrocyte vesiculation induced by ATP depletion. This was correlated with turnover of the phosphoinositides as well as RBC deformability during the process since phosphoinositide metabolism may be involved in shape regulation of RBCs. Echinocytic shape transformation and subsequent vesiculation of RBCs, which commonly occur during ATP depletion, were inhibited by chlorpromazine. Furthermore, with a newly developed two-dimensional thin-layer chromatography separation of RBC membrane phospholipids, we showed that chlorpromazine significantly decreased the dephosphorylation of phosphatidylinositol-4,5-bisphosphate (PIP2) in both ATP-depleted RBCs as well as in cells with partly maintained ATP levels. Concomitantly, there was a smaller increase in the relative amount of phosphatidylinositol. In addition, chlorpromazine also inhibited the decreased in RBC deformability as well as the shift of osmotic fragility that occurs during ATP depletion of erythrocytes.

During incubation of intact human erythrocytes with sonicated dimyristoylphosphatidylcholine (DMPC) vesicles, the cells change their discoid morphology to form echinocytes and finally give rise to the release of membrane vesicles. In this process, the red cell membrane accumulates DMPC and loses up to 15% of its cholesterol. On the other hand, replacement of 25% of the endogenous phosphatidylcholine species by DMPC without affecting the cholesterol level of the erythrocytes can be achieved by incubation with DMPC/cholesterol (1:1, mol/mol) sonicated vesicles in the presence of the phosphatidylcholine-specific phospholipid-transfer protein from bovine liver. This replacement also gives rise to an echinocytic cell morphology, but no membrane vesiculation can be observed. However, the vesiculation process can as yet be initiated upon a subsequent decrease of the cholesterol level, by incubation of those modified cells in the presence of sonicated vesicles of pure egg phosphatidylcholine. Incubation of native erythrocytes with pure egg phosphatidylcholine vesicles, on the other hand, results in cholesterol depletion, but does neither induce the formation of echinocytes nor the release of membrane vesicles. Cellular ATP levels are not affected during these incubations. From these results, it can be concluded that a decrease in cholesterol content of the erythrocyte membrane is essential for the DMPC-induced vesiculation of those cells.

The rate of uptake of radioactive phosphatidylcholine molecules of different fatty acid composition in intact erythrocytes as facilitated by a phosphatidylcholine-specific transfer protein has been studied. When trace amounts of radiolabeled phosphatidylcholine molecules are present in donor vesicles consisting of egg phosphatidylcholine and cholesterol, the transfer of the radiolabeled species depends strongly on their fatty acyl composition: dipalmitoylphosphatidylcholine is transferred at the lowest rate, 1-saturated-2-unsaturated species are transferred faster and the highest rate is observed for dioleoyl phosphatidylcholine. Transfer of the various phosphatidylcholine molecules was measured furthermore using donor systems in which the bulk phosphatidylcholine was varied in its fatty acyl composition. Also in this type of experiment, the transfer protein preferentially stimulated transfer of unsaturated phosphatidylcholine molecules, especially from an environment containing more saturated molecules. Finally, the efflux of labeled phosphatidylcholine from intact erythrocytes to plasma in the absence of the phosphatidylcholine-specific transfer protein was studied and it became clear that in this case the nature of the effused molecules itself, rather than the composition of the bulk lipids, determined the effuse rates. An important conclusion to be drawn from these experiments is that radiolabeled phosphatidylcholine molecules, when used as markers for phospholipid exchange or transfer, should resemble in their fatty acid composition the composition of the bulk lipid in order to provide reliable data on rates and extents of the process studied.

Phosphatidylcholine-specific transfer protein from beef liver has been used to replace native phosphatidylcholine (PC) molecules from intact human erythrocytes by a variety of PC species differing in fatty acid composition. These replacements changed neither the total phospholipid content of the membrane, nor the composition of this fraction in terms of the various phospholipid classes. The morphology of the erythrocyte was not modified when native PC was replaced by 1-palmitoyl,2-oleoyl PC, 1-palmitoyl,2-linoleoyl PC, egg PC, or PC isolated from rat liver microsomes. Replacement with the disaturated species 1,2-dimyristoyl PC, 1,2-dipalmitoyl PC, and 1,2-distearoyl PC resulted in the formation of echinocytes and, at higher levels of replacement, in spheroechinocytes. Echinocyte-like erythrocytes were also observed after replacement with 1-palmitoyl,2-arachidonoyl PC, whereas stomatocytes were formed upon replacement with PC species containing two unsaturated fatty acids, e.g., 1,2-dioleoyl PC and 1,2-dilinoleoyl PC. The observations show that the erythrocyte membrane structure and the overall discoid cell shape of the human erythrocyte are optimally stabilized by PC species that contain one saturated and one mono- or diunsaturated fatty acid, and that the cell tolerates only limited variations in the species composition of its PC.