Bioconjugate Chem. 1997, 8, 941−945
941
TECHNICAL NOTES Bilayer Distribution of Phosphatidylserine and Phosphatidylethanolamine in Lipid Vesicles Maria Teresa Roy, Montserrat Gallardo, and Joan Estelrich* Unitat de Fisicoquı´mica, Facultat de Farma`cia, Universitat de Barcelona, Avinguda Joan XXIII s/n, 08028 Barcelona (Catalonia), Spain. Received May 28, 1997X
The distribution of phosphatidylethanolamine (PE) and phosphatydilserine (PS) in liposomes was studied as a function of aminophospholipid concentration using fluorescamine as labeling reagent. The method is suitable for such determination since, in the assay conditions, fluorescamine does not penetrate the vesicles nor does it disrupt them. The liposomes were obtained by sonication, extrusion, or mechanical dispersion (MLV). For any kind of vesicle, the percentage of PS in the external monolayer is higher than that obtained for PE in the corresponding vesicles. In extruded PS liposomes, this aminophospholipid is located preferentially in the outer layer, while for PE liposomes the localization depends on the size of vesicle. Sonicated liposomes present an asymmetrical distribution of both aminophospholipids, and the external location of PS or PE always predominates. In contrast, in MLV, aminophospholipids are mainly found in the inner layers of the vesicles, except for liposomes formed by the lowest PS proportion. A remarkable feature of PS liposomes is the reduction of vesicle size, especially in MLV liposomes, in comparison with neutral liposomes.
INTRODUCTION
The molecular architecture of biological membranes is highly asymmetric. Not only is the orientation of membrane proteins asymmetric, but the distribution of phospholipids and fatty acids is also asymmetric (1). The asymmetrical distribution of membrane phospholipids across the lipid bilayer seems to be responsible for cell functions such as coagulation, membrane fusion, and stability (2). In erythrocytes, the best documented system, phosphatidylserine (PS), phosphatidylethanolamine (PE), and probably phosphatidylinositol (PI) are located mainly in the inner monolayer, while phosphatidylcholine (PC) and sphingomyelin (SM) are mostly found in the outermost monolayer (3). This highly asymmetric orientation of lipids is essential to normal homeostasis. Increased exposure of PS on the outer surface of the red blood cell is a signal for sequestration by the reticuloendothelial system. Furthermore, partial loss of PS asymmetry occurs, for instance, in some pathological red blood cells, such as sickle cells. Similarly, tumorigenic cells are bound and lysed by activated monocytes or macrophages, as they express much more PS in their membrane outer leaflet than their nontumorogenic counterparts (4). In platelets, SM and PE are located predominantly in the surface membrane, while PS and PI are in the intracellular membrane (5). Moreover, in platelets, transbilayer asymmetry is rapidly lost upon stimulation by certain platelet antagonists. This phenomenon fulfills an important physiological function in the hemostatic process, since two sequential reactions of the coagulation cascade are dramatically accelerated in the presence of an anionic phospholipid surface (6, 7). PS and PI are * Author to whom correspondence should be addressed (telephone + 34-3-4024554; fax + 34-3-4021886; e-mail estelric@ farmacia.far.ub.es). X Abstract published in Advance ACS Abstracts, October 15, 1997.
S1043-1802(97)00105-5 CCC: $14.00
active in thrombin generation but have no role in factor Xa formation. PC and PE do not contribute to thrombin formation but are active in the formation of factor Xa. Although in principle any negatively charged phospholipid is capable of providing a catalytic surface, membranes containing PS exhibit the highest procoagulant activity (8). In view of the weak procoagulant activity of PI and its conversion in the PI cycle upon cell activation, PS is likely to be solely responsible for the formation of procoagulant lipid surfaces in stimulated platelets. It has been known for several decades that platelet activation may result in formation of plateletderived microparticles with clot-promoting activity. The importance of this material for platelet procoagulant activity has long been underrated. Recently, however, it has become evident that the particles are derived from the cell surface by shedding from the plasma membrane, and shedding of microvesicles appears to be closely associated with surface exposure of PS (7). Electromicroscopy reveals these particles as uni- or paucilamellars. Since this structure is quite similar to liposomes, lipid vesicles could mimic the platelet-derived microparticles or the platelets themselves. Thus, an ability to generate liposomal systems exhibiting asymmetric transbilayer distributions of PS (or PE) may have a number of applications. To this end, we have undertaken the study of the distribution of PE and PS in liposomes obtained by extrusion, sonication, or simple dispersion. These results may help to identify the kinds of liposomes that provide the desired aminophospholipid distribution. This may lead, in turn, to the development of models of biological membrane with a particular PE or PS distribution. EXPERIMENTAL PROCEDURES
Materials. Soybean PC (Lipoid S-100) was purchased from Lipoid (Ludwigshafen, Germany). PS from bovine spinal cord and PE from egg yolk were obtained from © 1997 American Chemical Society
942 Bioconjugate Chem., Vol. 8, No. 6, 1997
Roy et al.
Table 1. Average Size (Expressed as z-Average Diameter ( SD, n g 3) of the Different Kinds of Liposomes Used liposome composition type of liposome extrusion
sonication MLV
pore size (nm) 400 200 100 50
% PE 5
10
20
% PS 30
50
5
10
20
30
250 ( 11 247 ( 8 242 ( 10 238 ( 13 210 ( 10 195 ( 27 184 ( 10 174 ( 10 171 ( 10 184 ( 6 186 (6 170 ( 6 167 ( 3 164 ( 7 137 ( 16 139 ( 2 140 ( 11 140 ( 5 111 ( 7 110 ( 8 104 ( 9 109 ( 4 113 ( 1 98 ( 5 95 ( 2 97 ( 8 99 ( 8 76 ( 4 78 ( 4 78 ( 6 82 ( 6 84 ( 4 72 ( 3 73 ( 2 72 ( 5 74 ( 5 131 ( 20 127 ( 25 119 ( 19 136 ( 26 117 ( 19 86 ( 8 91 ( 8 76 ( 8 80 ( 7 1350 ( 150 1200 ( 190 1010 ( 210 1160 ( 210 1140 ( 110 560 ( 100 650 ( 160 640 ( 180 650 ( 130
Lipid Products (Nutfield, U.K.). Fluorescamine and Triton X-100 were purchased from Sigma. Organic solvents (methanol, ethanol, chloroform) were obtained from Merck (Darmstadt, Germany) and used without purification. Acetone was of spectroscopic grade (Merck). Saline solution was made with 150 mM NaCl. Borate buffer (pH 8.25) was made at 50 mM concentration and NaCl was added to give an isotonic solution, 310 ( 10 mOsm/kg. Preparation of Multilamellar Liposomes (MLV). Phospholipids (PC alone or PC with PS or PE) were dissolved in chloroform in a round-bottom flask and dried in a rotary evaporator under reduced pressure at 50 °C to form a thin film on the flask. The film was hydrated with saline solution to give a lipid concentration of 10 µmol/mL. Multilamellar liposomes were formed by constant vortexing for 4 min on a vortex mixer and sonication in a bath sonifier for 4 min. Preparation of Sonicated Liposomes. MLV were sonicated on an ice-water bath with a Branson sonifier (Labsonic U) equipped with a microtip probe and operating at 57 W output for 6 min with three cycles of 2 min and 30 s intervals stand-by in ice bath. The resulting suspension was passed through 0.45 µm membrane filters to remove the possible titanium particles from the sonifier probe. Preparation of Extruded Liposomes. MLV were downsized to form oligolamellar vesicles by extrusion at 50 °C in an extruder device (Lipex Biomembranes, Canada) through polycarbonate membrane filters of variable pore size under nitrogen pressures of up to 55 × 105 N‚m-2 (9). Liposomes were extruded sequentially through polycarbonate filters (0.8, 0.4, 0.2., 0.1, and 0.05 µm (Nucleopore, U.S.A.) to obtain liposomes of nominal size of 400, 200, 100, and 50 nm. Determination of Aminophospholipids. PE and PS were determined spectrofluorometrically using fluorescamine as labeling reagent (10). This reagent was used by Cheung and Forte (11) to label aminophospholipids, but the method is similar to that used by Barenholz et al. (12) to determine PE in liposomes by means of 2,4,6-trinitrobenzenesulfonic acid (TNBS). It was confirmed in advance that fluorescamine had no disruptive effect on liposome structure (data not shown). Determination of Aminophospholipids in the Outer Vesicle Surface. At room temperature, aliquots (100-200 µL) of liposomes were diluted with 2.0 mL of borate buffer (pH 8.25). Fluorescamine solution (0.1 mL; 3 mg in 100 mL of acetone) was added, and the vesicle sample was shaken vigorously for 30 s. After 1 min, 2 mL of 1.6% Triton X-100 in borate buffer was added to the sample, followed by mixing. The resulting fluorescence was read within 2 h of reaction by exciting the sample at 381 nm and measuring the emitted radiation at 471 nm. Determination of Total Aminophospholipids in the Vesicle. Aliquots (100-200 µL) of liposomes were disrupted with 2.0 mL of 1.6% Triton X-100 in borate buffer (pH 8.25). Fluorescamine solution (0.1 mL) was
50
0
161 ( 11 255 ( 16 134 ( 12 180 ( 15 106 ( 9 102 ( 5 81 ( 6 84 ( 5 83 ( 11 121 ( 20 590 ( 190 1350 ( 50
added, and the vesicle sample was shaken vigorously for 30 s. After 1 min, 2 mL of borate buffer was added to the sample, followed by mixing. The samples were read as described above. Photon Correlation Spectroscopy. Vesicle size distribution was determined by photon correlation spectroscopy with an Autosizer II spectrometer (Malvern Instruments, U.K.) at 37 °C. The vesicles obtained by extrusion or sonication, which afforded a unimodal vesicle distribution, were sized by the method of cumulant analysis (13), while the exponential sampling method (14), which does not assume any particular form of distribution, was used to size MLV. RESULTS
Fluorescamine Assay. The first stage of this investigation was aimed at establishing the validity of the fluorescamine labeling procedures for assaying transmembrane distributions of lipids containing primary amino groups in liposomal systems. We checked the linearity of fluorescence intensity between the range from 0.0125 to 0.300 µmol of aminophospholipid. On the other hand, we found there was no difference between the results of total and external aminophospholipid when the substance assayed did not present such differences, namely in solutions of bovine serum albumin. In this way, a solution of such protein gave a values of 153 ( 4 (arbitrary units of fluorescence) (n ) 6) when it was assayed as external amino grous and 157 ( 5 (n ) 6) for total amino groups. A comparison of means showed that at 95.0% confidence interval there were no significant differences between both means (t ) 1.610 < 2.228). Finally, we checked that at the reagent concentration used, the extent of the labeling was not influenced by the presence of a pH gradient (interior acidic). Liposomes with an acidic interior (achieved with citrate buffer) gave the same result as those obtained in saline solution. Thus, we deduced that, in the experimental conditions of the assay, liposomal membranes were impermeable to fluorescamine and, consequently, we prepared the liposomal samples in saline solution. Extrusion Liposomes. To assess the distribution of aminophospholipids, three kinds of liposomes were examined. Table 1 shows the z-average diameters obtained. On the whole, diameters of PE-containing extruded liposomes were similar to these obtained with neutral liposomes (for example, those formed solely by PC, the sizes of which are displayed in the last column of the table). However, the diameter of vesicles with >20% PE and extruded through 400 and 200 nm membranes was slightly lower that the obtained for PC liposomes extruded through membranes of the same pore size (255 and 180 nm, respectively). In PS-containing liposomes the reduction of vesicle size was even more evident in those vesicles downsized through membranes of 400 and 200 nm. At any PS percentage, the average values were 50% irrespective of composition or size. When both distributions were compared, it was observed that below 20% concentration of aminolipid in the membrane, the differences were relatively important, but above such concentration, broadly speaking, the percentages of external amine groups became the same. Sonicated Liposomes. Diameters of PE liposomes were greater than those of vesicles containing PS (Table 1). Diameters of PS sonicated vesicles were comparable to those of the vesicles extruded by membranes of 50 nm pore size. The aminophospholipid distribution followed the pattern observed in extruded liposomes (Figure 1). At low aminolipid concentrations, both PE and PS showed a tendency to concentrate in the outer surface. This location was reduced in extent when concentration increased. On the other hand, at almost all concentrations, PS was found in a higher extent in the external layer than PE. Mechanical Dispersion Liposomes. Percentages of
Figure 2. Average percentage of external aminophospholipids determined in multilamellar liposomes using the fluorescamine assay: (O) PS; (b) PE.
external aminophospholipids and z-average diameters of multilamellar liposomes are shown in Figure 2 and Table 1, respectively. PE multilamellar liposomes present the classical characteristics of this kind of vesicle: their average diameters were outside the range above the micrometer with relatively important associated polydispersities. PS liposomes, in contrast, showed a lower sizesin comparison with neutral liposomessand the extent of dispersity was lower, too. As in the other types of vesicle, the increment of aminophospholipid percentage favored their localization in the inner layer of the vesicle. DISCUSSION
Our results show that liposomes containing PS are smaller than PE vesicles or plain PC vesicles. This difference is especially marked in multilamellar liposomes and appears to be independent of PS concentration. This phenomenon has been also noted in other charged vesicles, namely liposomes containing phosphatidic acid, but not to the extent shown in PS liposomes. This fact could be explained by the different structure and morphology of PS dispersions in comparison to those of PC and other neutral or isoelectric phospholipids. In contrast to egg PC, aqueous dispersions of PS exist as a single lamellar liquid crystalline phase up to water content of ∼75%. In the absence of salt, pure PS incorporates all of the water between the lipid bilayers, increasing the lamellar repeat distance. The ability of PS bilayers to incorporate large amounts of water is a consequence of the net negative charge of the PS molecule at pH between 5 and 9, giving rise to large repulsive forces between adjacent bilayer sheets. Pure egg PC behaves differently and can incorporate water only up to ∼40% water; further addition of water produces a twophase system, a lamellar liquid crystalline phase in excess water (15). As such liposomes have been formed in saline solution (0.16 M NaCl), the salt must reduce the double-layer repulsion between adjacent bilayers. As
944 Bioconjugate Chem., Vol. 8, No. 6, 1997
a result of that, water is extruded from the interbilayer space consistent with a decrease of lamellar repeat distance. Hauser (16) proposed that at high water content the multilamellar structures break up, with each lamella sealing off to form closed unilamellar vesicles. The addition of salt would induce aggregation and fusion of the vesicles leading, in equilibrium, to multilamellar structures. The relatively moderate presence of sodium in our liposomal suspension should limit the transformation of multilamellar structures into unilamellar vesicles, and this could explain the intermediate size of PS multilamellar liposomes. Concerning the aminophospholipid distribution in extruded and sonicated liposomes, there is a predisposition of such lipids to locate in the external monolayer. This is tru in general for PS liposomes and characteristic of those PE liposomes extruded through membranes of pore below 200 nm. The preferential localization of PE and PS in the external monolayer is in agreement with the results of Massari et al. (17), who found 67% of external amines in egg PC containing 40% of PS. In contrast, Berden et al. (18) and Barsukov et al. (19) state that in sonicated liposomes, PE or PS had preference for the inside layer of the bilayer. However, the experimental conditions are not equivalent to ours. For instance, these authors did not measure the vesicle size and, although in theory these liposomes might have been unilamellar at the moment of their formation, they could have suffered aggregation and/or fusion during the relatively long time that a measurement by NMR takes. As it has been already seen in Table 1, the localization of PE shows a clear dependence on actual vesicle size. In PS liposomes the difference of size between the vesicles extruded by the biggest membrane pore and those passed by the lowest are smaller, and this is evidenced by minor and not so clear differences in the PS distribution. In liposomes obtained by mechanical dispersion the percentage of external aminophospholipid is always
