Cell morphology changes are used to examine the interaction of exogenous phosphatidylserine and phosphatidylethanolamine with human erythrocytes. Short-chain saturated lipids transfer from liposomes to cells, inducing shape changes that are indicative of their incorporation into, and in some cases translocation across, the cell membrane bilayer. Dioleoylphosphatidylserine and low concentrations of dilauroyl- and dimyristoylphosphatidylserine induce stomatocytosis. At higher concentrations, dilauroylphosphatidylserine and dimyristoylphosphatidylserine induce a biphasic shape change: the cells crenate initially but rapidly revert to a discocytic and eventually stomatocytic shape. The extent of these shape changes is dose dependent and increases with increasing hydrophilicity of the phospholipid. Cells treated with dilauroylphosphatidylethanolamine and bovine brain lysophosphatidylserine exhibit a similar biphasic shape change but revert to discocytes rather than stomatocytes. These shape changes are not a result of vesicle--cell fusion nor can they be accounted for by cholesterol depletion. The reversion from crenated to stomatocytic forms is dependent on intracellular ATP and Mg2+ concentrations and the state of protein sulfhydryl groups. The present results are consistent with the existence of a Mg2+- and ATP-dependent protein in erythrocytes that selectively translocates aminophospholipids to the membrane inner monolayer engendering aminophospholipid asymmetry.

Monolayers at the air-water interface were prepared from lipids extracted from human red blood cells. Epifluorescence microscopy was used to show that monolayers simulating the inner and outer leaflets of the red cell membrane form immiscible liquid phases with critical points at surface pressures of 21 and 29 dyn͞cm. At these pressures the monolayer lipid density is comparable to that in the red cell membrane. This suggests that lipid bilayers of a red blood cell are near a miscibility critical point, which should significantly affect the biophysical properties of the red cell membrane.[S0031-9007(98)07922-8] 64.75. + g, 87.22.Bt, Epifluorescence microscope studies have shown that mixtures of cholesterol and phospholipids form immiscible liquid phases in monolayers at the air-water interface [1][2][3][4]. There is indirect evidence that immiscibility also occurs in bilayers of these lipids [5][6][7]. These results raise the question of whether immiscible liquid domains exist in biological membranes. To this end, we extracted the lipids from membranes of human red blood cells (erythrocytes). Red blood cell membranes contain over 250 lipid species [8], most of which are arranged asymmetrically across the membrane [9]. The major headgroup classes of the extracted erythrocyte lipids were separated by thin layer chromatography, then reconstituted with 50 mol % cholesterol so as to mimic compositions of the inner and outer erythrocyte membrane leaflets. Monolayers of these lipids at the air-water interface are found to form immiscible liquid phases. Figures 1 and 2 show epifluorescence micrographs of monolayers of reconstituted lipids simulating the inner and outer leaflets. Contrast between phases is achieved by 0.2 mol % of a fluorescent probe that is preferentially excluded from the cholesterol-rich phase [4]. Two immiscible liquid phases coexist to surface pressures greater than 20 dyn͞cm. Domain shapes characteristic of proximity to a critical point are observed at 21 dyn͞cm in the simulated inner leaflet [ Fig. 1(b)] and at 29 dyn͞cm in the outer leaflet [Figs. 2(c) and 2(d)]. The fingering of the domains in Fig. 2(f) indicates the monolayer is close to the critical composition. "Fingered" domains are circular domains from which stripes emanate. At higher pressures the monolayers are homogeneous as in Figs. 1(c) and 2(e). The domain shape changes are reversible.It is helpful to consider these results in terms of a simple thermodynamic model. Mixtures of phospholipids alone have not generally been found to exhibit liquid-liquid immiscibility in monolayers. However, immiscibility is seen in binary mixtures of cholesterol with the phospholipids phosphatidylcholine [1-4], phosphatidylethanolamine [10], or phosphatidylserine [11]. Cholesterol and egg sphingomyelin also exhibit immiscibility (data not shown).Thus, all of the red cell's major lipid components can potentially contribute to the observed immiscibility. This leads to the question of how binary properties manifest themselves in a multicomponent mixture. C...

ATP-depleted human erythrocytes lose their smooth discoid shape and adopt a spiny, crenated form. This shape change coincides with the conversion of phosphatidylinositol-4,5-bisphosphate to phosphatidylinositol and phosphatidic acid to diacylglycerol. Both crenation and lipid dephosphorylation are accelerated by iodoacetamide, and both are reversed by nutrient supplementation. The observed changes in lipid populations should shrink the membrane inner monolayer by 0.6%, consistent with estimates of bilayer imbalance in crenated cells. These observations suggest that metabolic crenation arises from a loss of inner monolayer area secondary to the degradation of phosphatidylinositol-4,5-bisphosphate and phosphatidic acid. A related process, crenation after Ca2+ loading, appears to arise from a loss inositides by a different pathway.

When human erythrocytes are incubated with certain phospholipids, the cells become spiculate echinocytes, resembling red cells subjected to metabolic starvation or Ca2+ loading. The present study examines (1) the mode of binding of saturated phosphatidylcholines and egg lysophosphatidylcholine to erythrocytes and (2) the quantitative relationship between phospholipid incorporation and red cell shape. We find that the phospholipids studied become intercalated into erythrocyte membranes, not simply adsorbed to the cell surface. Spin-labeling and radiolabeling data show that the incorporation of (4 +/- 1) X 10(6) molecules of exogenous phosphatidylcholine per cell converts discocytes to stage 3 echinocytes with about 35 conical spicules. This amount of lipid incorporation is estimated to expand the red cell membrane outer monolayer by 1.7% +/- 0.6%. Calculations of the inner and outer monolayer surface areas of model discocytes and stage 3 echinocytes yield an estimated difference of 0.7% +/- 0.2%.

The rate of phospholipid transfer from sonicated phospholipid vesicles to human erythrocytes has been studied as a function of membrane concentration and lipid acyl chain composition. Phospholipid transfer exhibits saturable first-order kinetics with respect to both cell and vesicle membrane concentrations. This kinetic behavior is consistent either with transfer during transient contact between cell and vesicle surfaces (but only if the fraction of the cell surface susceptible to such interaction is small) or with transfer of monomers through the aqueous phase. The acyl chain composition of the transferred phospholipid affects the transfer kinetics profoundly; for homologous saturated phosphatidylcholines, the rate of transfer decreases exponentially with increasing acyl chain length. This behavior is consistent with passage of phospholipid monomers through a polar phase, which might be the bulk aqueous phase( as in the monomer transfer model) or the hydrated head-group regions of a cell-vesicle complex (transient collision model). Collisional transfer also predicts that intercell transfer of phospholipids should be slow compared to cell-vesicle transfer, as surface charge and steric effects should prevent close apposition of donor and acceptor membranes. This is not found; dilauroylphosphatidylcholine transfers rapidly between red cells. Thus, the observed relationship between acyl chain length and intermembrane phospholipid transfer rates likely reflects the energetics of monomer transfer through the aqueous phase.