Localization of cholesterol in membranes of erythrocyte ghosts

Cold shock hemolysis in human erythrocytes studied by spin probe method and freeze-fracture electron microscopy

When human erythrocytes are osmotically stressed or chemically treated, they hemolyze on cooling below 10 degrees C (called cold shock). We have studied the effects of osmotic stress and cooling on the state of membrane by the spin-probe method and freeze-fracture electron microscopy. At room temperature, the membrane fluidity detected by 12-doxyl stearate spin probe showed a steady decrease with osmolality in hypertonic NaCl solutions up to 900 mOsm/kg, above which it remained unchanged. In hypertonic sucrose solutions, the electron paramagnetic resonance spectra showed an additional pair of absorptions, indicating development of regions, in the membrane, further immobilized than in NaCl solutions. Mobility of a cholesterol analogue probe, androstane, did not show change by hypertonicity, but the spectral intensity dropped at 1,200 mOsm/kg, probably due to formation of loose aggregates in the cholesterol phase. On cooling the osmotically stressed cells in NaCl solution, the isotropic rotational correlation time vs. inverse temperature plot of 12-doxyl stearate probe exhibited a step-wise discontinuity at approximately 10 degrees C, suggestive of a drastic transition in the state of the membrane. At about the same temperature, the freeze-fracture pattern of osmotically stressed cells revealed the development of large wrinkles and aggregation of membrane particles, in contrast to the case of the cells in isotonicity. Significance of these findings in understanding cold shock hemolysis is discussed.

Equilibrium energy analysis of freeze-fracture planes in membranes

We have used equilibrium energy calculations to determine the most probable freeze-fracture planes in a lipid bilayer. Using a pairwise-summation computer method, we have generated numerical values for the Van der Waals potentials (electron shell repulsion, dispersion forces and electrostatic interactions) between molecules. We have compared our theoretical predictions with the experimental conclusion that the fracture planes occur normally between lipid molecules. These calculations also provide information about the composition of intramembranous particles, the potential for local clustering of single lipid types in the fluid membrane, and the importance of lipid molecules to the function of membrane proteins such as voltage-sensitive ion channels.

Thermotropic lipid phase separations in human erythrocyte ghosts and cholesterol-enriched rat liver plasma membranes

Electron spin resonance (ESR) studies of human erythrocyte ghosts labeled with 5-nitroxide stearate, I(12,3), indicate that a temperature-dependent lipid phase separation occurs with a high onset at 38 degrees C. Cooling below 38 degrees C induces I(12,3) clustering. Similar phase separations were previously identified in human platelet and cholesterol-loaded [cholesterol/phospholipid molar ratio (C/P) = 0.85] rat liver plasma membranes [L.M. Gordon et al., 1983; J. Membrane Biol. 76; 139-149]; these were attributed to redistribution of endogenous lipid components such that I(12,3) is excluded from cholesterol-rich domains and tends to reside in cholesterol-poor domains. Further enrichment of rat liver plasma membranes to C/P ratios of 0.94-0.98 creates an "artificial" system equivalent to human erythrocyte ghosts (C/P = 0.90), using such criteria as probe flexibility, temperature dependent I(12,3) clustering; and polarity of the probe environment. Consequently, cholesterol-rich and -poor domains probably exist in both erythrocyte ghosts and high cholesterol liver membranes at physiologic temperatures. The temperature dependence of cold-induced hypertonic lysis of intact human erythrocytes was examined by incubating cells in 0.9 M sucrose for 10 min at 1 degree C intervals between 9 and 46 degrees C (Stage 1), and then subjecting them to 0 degrees C for 10 min (Stage 2). Plots of released hemoglobin are approx. sigmoidal, with no lysis below 18 degrees C and maximal lysis above 40 degrees C. The protective effect of low temperatures during Stage 1 may be due to the formation of cholesterol-rich domains that alter the bilayer distribution and/or conformation of critical membrane-associated proteins.

Thermotropic lipid phase separations in human platelet and rat liver plasma membranes

Electron spin resonance (ESR) studies were conducted on human platelet plasma membranes using 5-nitroxide stearate, I(12,3). The polarity-corrected order parameter S and polarity-uncorrected order parameters S(T parallel) and S(T perpendicular) were independent of probe concentration at low I(12.3)/membrane protein ratios. At higher ratios, S and S(T perpendicular) decreased with increasing probe concentration while S(T parallel) remained unchanged. This is the result of enhanced radical interactions due to probe clustering. A lipid phase separation occurs in platelet membranes that segregates I(12,3) for temperatures less than 37 degrees C. As Arrhenius plots of platelet acid phosphatase activity exhibit a break at 35 to 36 degrees C, this enzyme activity may be influenced by the above phase separation. Similar experiments were performed on native [cholesterol/phospholipid ratio (C/P) = 0.71] and cholesterol-enriched [C/P = 0.85] rat liver plasma membranes. At 36 degrees C, cholesterol loading reduces I(12,3) flexibility and decreases the probe ratio at which radical interactions are apparent. The latter effects are attributed to the formation of cholesterol-rich lipid domains, and to the inability of I(12,3) to partition into these domains because of steric hinderance. Cholesterol enrichment increases both the high temperature onset of the phase separation occurring in liver membranes from 28 degrees to 37 degrees C and the percentage of probe-excluding, cholesterol-rich lipid domains at elevated temperatures. A model is discussed attributing the lipid phase separation in native liver plasma membranes to cholesterol-rich and -poor domains. As I(12,3) behaves similarly in cholesterol-enriched liver and human platelet plasma membranes, cholesterol-rich and -poor domains probably exist in both systems at physiologic temperatures.

Ultrastructural localization of cholesterol by enzyme histochemistry

The histochemical localization of cholesterol using oxidized diaminobenzidine as the final reaction product was studied at the electron microscopical level and compared with the digitonin method of cholesterol localization based on cholesterol digitonide as the final reaction product. Tissue chopper sections of fixed rat adrenal glands were incubated at 37 degrees C in a medium consisting of 0.8 units/ml cholesterol oxidase, 1.4 units/ml cholesterol ester hydrolase, 50 units/ml horseradish peroxidase, 0.5 mg/ml diaminobenzidine, 0.1% v/v Triton X-100 (or Surfal) and an endogenous peroxidase inhibitor in 0.1 M phosphate buffer, pH 7.0. An electron-dense osmiophilic reaction product was observed in many lipid droplets, intracellular vesicles and focally around mitochondria. Appropriate control experiments indicated that deposition of reaction product depended on the presence of cholesterol and the necessary enzymes. Comparison studies using digitonin confirmed the presence of cholesterol in the lipid droplets, but ultrastructural distortion limited the resolution of the more discrete deposits of cholesterol than the digitonin method because it does not cause distortion of cellular ultrastructure attributed to the formation of cholesterol digitonide. The enzyme method or a combination of enzyme and digitonin enable localization of free, esterified or total cholesterol.

The effect of membrane cholesterol on the sodium pump in red blood cells

1. Human red blood cells (R.B.C.) were incubated with phosphatidylcholine vesicles to obtain partial depletion of their membrane cholesterol. The kinetic parameters of the Na pump (affinities and maximal fluxes) were determined in these cells and compared with controls. 2. In the presence of internal K, progressive cholesterol depletion induces gradually (i) an increase of the maximal flux (ii) a reduction of the apparent affinity for internal Na. 3. In the absence of internal K, cholesterol depletion has no apparent effect, indicating that the observed changes were mediated by the internal K. Cholesterol depletion acts by increasing these two K effects: (i) activation of the maximal fluxes, and (ii) reduction of the internal Na affinity. 4. The cholesterol effect (on the K-pump interaction) is specific; it occurs without any change (i) in the affinities for internal Na and inorganic phosphate and for external Na and K, and (ii) in the number of pumping sites. 5. Under physiological conditions, cholesterol reduces the cation translocation rate and increases the selectivity of the pump for internal Na.

Digitonin induced alterations of the erythrocyte membrane as visible by freeze-fracturing

Treatment of erythrocytes by low concentrations of digitonin results in the formation of elongated, bulged membrane areas free from particles and intramembraneous tubular structures emerged from these domains. At higher concentrations the tubular structures are present also outside of the membrane. A second, not bulged type of particle-free areas is more or less rounded, but mostly of rectangular shape. The rate of this type of domains increases with the concentration of digitonin. Breaks in these areas lead to a formation of sheets which results in a breakdown of the membrane structure finally into a brittle mass of many small sheets lying irregularly one upon the other. Already during the first steps of this membrane breakdown hemolysis takes place. Cross-linking of the proteins with glutaraldehyde does not entirely block up the membrane alterations. In fixed erythrocytes the formation fo elongated domains is restricted, no tubular structures are present and no dislocations of particles can be observed. Nevertheless smooth sheets localized also outside of the membrane are formed. However the investigated ghosts are not stabilized by glutaraldehyde against the effects of digitonin. Particle dislocations as well as all the other membrane alterations are present. The implications of the obtained results are: 1. The elongated domains and the tubular structures are probably not digitonin-cholesterol-complexes. 2. The formation of crystalline regions by digitonin-cholesterol-complexes destroys the membrane structure.

The molecular architecture of myelin: identification of the external surface membrane components

Basic information concerning the molecular organization of the myelin membrane is an intrinsic requirement for understanding the neurochemical events leading to myelination, as well as the potential mechanism of demyelination that might exist at the molecular level for a variety of neurological diseases. The application of chemical, enzymatic, fluorescent, and immunological membrane probes has contributed significantly to this end, although the diverse structural complexity of the myelin sheath has permitted only a rudimentary understanding of its molecular organization. Nevertheless, compelling evidence is accumulating which suggests that components of myelin are asymmetrically distributed in the membrane. Such membrane asymmetry should not only provide important clues to the mechanisms of membrane assembly in the process of myelination, but should also serve as a paradigm for potential functional asymmetry of the individual components at the molecular level. One particularly useful membrane probe is galactose oxidase which has the capacity for identifying surface galactose residues in both glycoproteins and glycolipids on the external surface of the myelin sheath. The identification of these surface components on the myelin sheath is of primary importance since such components might be more readily susceptible to immunological damage or act as a viral receptor which ultimately might lead to demyelination.

The reactivity of human erythrocyte membrane cholesterol with a cholesterol oxidase

Cholesterol oxidase (EC 1.1.3.6, Brevibacterium sp.), which catalyzes the reaction: cholesterol + O2 leads to delta4-cholestenone + H2O2, has no effect on the cholesterol of intact (human) erythrocytes and of "resealed" ghosts, when it is present only outside these ghosts. The cholesterol of "leaky" ghosts, of "resealed" ghosts with enzyme trapped within, and of "inside-out" vesicles, was completely oxidized. This pattern indicates that the inner ((cytoplasmic) membrane surface must be exposed to the enzyme for the reaction to occur, and that outer surface cholesterol only becomes reactive after the membrane has been degraded by the oxidation of inner surface cholesterol. The enzymatic oxidations followed monotonic first-order kinetics, and hence gave no evidence to support the two states of cholesterol in the membrane that had been postulated earlier from studies on the plasma lipoprotein extraction of cholesterol from the membrane.

Multiple thermotropic state transitions in erythrocyte membranes. A laser-Raman study of the CH-stretching and acoustical regions

1. We have examined the Raman scattering from erythrocyte ghosts at 2700 cm-1 -3000 cm-1 (CH-stretching region), that at low frequencies due to acoustical vibrations, as well as that due to v (-C=C-) and v (=C-C=) of beta-carotene, in the temperature range of -30 degrees C to +30 degrees C. For comparison we have also evaluated the Rama spectra of liposomes composed of lecithins+/-cholesterol. 2. Plots vs. temperature of the intensities (I) of the approx. 2880 cm-1 and 2932 cm-1 bands relative to the intensity of the thermally stable 2850 cm-1 band, i.e. the (I2880/I2850) and (I2932/I2850) ratios, revealed discontinuities in both erythrocyte membranes and liposomes. 3. In erythrocyte membranes, plots of (I approximately 2880/I2850) and (I approximately 2932/I 2850) reveal several discontinuities. Those reported by (I approximately 2880/I 2850) occur at +17 degrees C and approx. -4 degrees C. The ratio (I approximately 2932/I 2850) reveals a transition between -20 degrees C and +10 degrees C, 1/2 maximal at -8 degrees C. The ratio (I 1527/I 1158), representing the relative intensity of v (-C=C-) vs. v (=C-C=) of membrane-bound beta-carotene, changes discontinuously with the 17 degrees C transition, but is stable over the rest of the temperature change studied. 4. The data indicate that a major membrane-state transition takes place below 0 degrees C but that an additional transition occurs at approx. 17 degrees C.

Analysis of membrane halves: cholesterol

Membrane splitting by freeze-fracture has been used as a preparative tool for chemical analysis of outer and inner "half"-membranes. In a previous report I showed that monolayers of human erythrocytes, bound to cationized glass, fracture nonrandomly, producing membrane fractions substantially enriched in outer or inner "halves." For the present study cells were used in quantities compatible with microanalysis. For quantitation the total amount of membrane present and the fractional enrichment of outer and inner "half"-membranes were determined. Cholesterol was examined by quantitative thin-layer chromatography modified to assay nanogram amounts. Comparison of lipids extracted from intact membranes with lipids from fractured membranes indicated that cholesterol was asymmetrically distributed across the plane of the membrane, more being present on the exterior side than on the interior side.

Variations of lipid-protein interactions in erythrocyte ghosts as a function of temperature and pH in physiological and non-physiological ranges. A study using a paramagnetic quenching of protein fluorescence by nitroxide lipid analogues

1. Incorporation of stearic acid nitroxides into erythrocyte ghosts markedly depresses the fluorescence of membrane protein tryptophan residues. 5-Nitroxide stearate quenches fluorescence more efficiently than 16-nitroxide stearate. Both compounds exhibit dynamic (diffusion-limited) quenching above 0.28 mumol/mg protein and static quenching at lower nitroxide protein ratios. Static quenching can be attributed to high affinity binding of nitroxide stearates by membrane protein. The dynamic phase represents distribution of the stearate analogues into a fluid lipid system. 2. Protein fluorophores accessible to quenching by a cholesterol analogue, androstane nitroxide, are saturated at low nitroxide/protein ratios (less than 0.14 mumol/mg protein), without resolution of a static quenching phase. This suggests that sterols are segregated away from protein, probably in CLusters". 3. Paramagnetic quenching by stearate nitroxides increases abruptly between 35 and 50 degrees C. This discontinuous enhancement of quenching by temperature is reversible up to 41 degrees C but irreversible at higher temperatures. The discontinuity is also diminished by lowering pH from 7.3 through 6.5 to 6.0. Quenching by androstane nitroxide increases linearly with temperature up to approx. 41 degrees C and then rises exponentially. We attribute the reversible quenching thermotropism detected by stearate derivatives to reversible, thermotropic unfolding and/or depolymerisation of membrane proteins. The irreversible phase, detected also by the sterol derivative can be attributed to non-reversible protein denaturation. 4. Paramagnetic quenching of membrane tryptophan fluorescence by stearate derivatives is minimal at approx. pH 7.1 (35 degrees C) and increases sharply at lower and higher pH values, suggesting that two categories of protein residues, titrating between pH 6 and 8, profoundly influence the association of fatty acyl chains and penetrating protein segments. Quenching by androstane nitroxide exhibits no significant variation between pH 6 and 8, consistent with other data indicating that erythrocyte membrane sterols are segregated from membrane proteins, probably in clusters. 5. Our new approach confirms previous suggestions of a boundary layer of lipid in close association with some proteins in erythrocyte membranes, as well as experiments indicating that the lipid status in this boundary layer depends on that state of membrane proteins. However, sterols appear to be largely excluded from this boundary domain. Our data further show that lipid-protein interactions in erythrocyte membranes can vary significantly with fluctuations of temperature and pH in the physiological range.

Phospholipid unsaturation and plasma membrane organization

A comparison has been made between the unsaturation of plasma-membrane phospholipids,present in the human erythrocyte, rat liver, mouse liver and a rapidly growing rat hepatoma. Of the double bonds present in the hydrocarbon chains of the membrane phospholipids,onethird is contributed by sphingomyelin plus phosphatidyl choline and the remainder by phosphatidyl serine, ethanolamine and inositol. Assuming that the phospholipids are asymmetrically distributed in the two leaflets of the bilayer in general, the consequences of this asymmetry in combination with cholesterol content and fatty acid distribution on plasma membrane organization and function are discussed. It is suggested, that the organizational disposition of plasma membrane components other than phospholipids is at least related if not dependent upon the latter's asymmetric distribution in the bilayer.