J. Phys. Chem. B 2001, 105, 9879-9886
9879
Octyl Glucoside-Mediated Solubilization and Reconstitution of Liposomes: Structural and Kinetic Aspects‡ Olga Lo´ pez,† Mercedes Co´ cera,† Luisa Coderch,† Jose Luis Parra,† Leonid Barsukov,# and Alfonso de la Maza*,† Departamento de TensioactiVos, Instituto de InVestigaciones Quı´micas y Ambientales de Barcelona (IIQAB), Consejo Superior de InVestigaciones Cientı´ficas (CSIC), C/ Jordi Girona, 18-26, 08034 Barcelona, Spain, and Shemyakin-OVchinnikoV Institute of Biorganic Chemistry, Russian Academy of Sciences, UI. Miklukho-Maklaya, 16/10, 117871 Moscow V-437, Russia ReceiVed: January 23, 2001; In Final Form: July 6, 2001
Structural and kinetic aspects of the solubilization of phosphatidylcholine (PC) liposomes induced by the nonionic surfactant octyl glucoside (OG) and the reconstitution of PC vesicles by dilution of OG/PC mixed micellar systems were studied using dynamic light scattering and freeze-fracture electron microscopy. To this end, three regions were delimited in the equilibrium phase diagram of this interaction: a “vesicular” region formed only by mixed vesicles, a “coexistence” region, in which fragmented vesicles and mixed micelles coexisted, and a “micellar” region formed by only mixed micelles. A simple mechanism of liposomes solubilization is proposed based on the following points: (a) Up to saturation of vesicles by OG, a direct formation of mixed micelles within the bilayer occurred. (b) The progressive separation of the formed micelles from the liposome surface led to the complete solubilization of vesicles. (c) This separation would take place without formation of complex intermediate aggregates in equilibrated systems with the only presence of mixed micelles and fragmented vesicles. The systems placed in the micellar and vesicular regions were more stable with time than those placed in the coexistence region in which a growth of the fragmented vesicles and a reduction in their proportion were detected. The dilution (one step fast dilution) of OG/PC micellar solutions placed on the micellar phase boundary (corresponding to the effective surfactant to PC molar ratio for liposome solubilization, ReSOL) led to the formation of vesicles, in which the higher the total concentration of OG and PC in the initial system, the higher the size of the reconstituted vesicles. The kinetics of solubilization and reconstitution processes showed that the “induction time” (time needed for the processes to start) for the formation mixed micelles was higher that for the reconstitution of mixed vesicles, indicating that the reconstitution process was faster than that of solubilization.
Introduction Surfactants are indispensable reagents in the solubilization and reconstitution of membrane proteins.1-4 The need to find effective and predictable means to solubilize and reconstitute these membranes, as well as to devise appropriate protocols for biological research or pharmacological applications is one reason for interest in the study of membrane-surfactant interactions. In this field, one of the most commonly used amphiphilic compounds is the nonionic surfactant octyl glucoside (OG), whose structural properties have been recently studied.5 This surfactant has been demonstrated to be a “mild” surfactant with respect to its denaturing effect on proteins and shows a relatively high critical micelle concentration (CMC), this characteristic being suitable in processes of vesicle reconstitution by dialysis and dilution.6-10 A number of studies have been devoted to the understanding of the principles governing the interaction of surfactants with simplified membrane models such as phospholipid vesicles.7,11-13 This interaction leads to the breakdown of lamellar structures ‡ Abbreviations: OG, octyl glucoside; PC, phosphatidylcholine; DLS, dynamic light scattering; FFEM, freeze fracture electron microscopy; CMC critical micelle concentration. † IIQAB, CSIC. # Russian Academy of Sciences.
and their solubilization via formation of lipid-surfactant mixed micelles. According to Lichtenberg,14 this process has been described by a three-stage model (vesicle saturation, formation of mixed micelles, and complete vesicle solubilization). The vesicle formation upon surfactant depletion from mixed surfactant-phospholipid micelles (by means of dialysis or dilution with surfactant-free buffer) has been reported as the opposite process of bilayer solubilization, being also described as a threestage process: micellar equilibration, bilayer closure, and vesicle growth.15-17 Although solubilization studies offer detailed structural descriptions, some contradictory aspects are still found, particularly those related to the size, shape, and stability of the aggregates formed. Thus, several authors described an important growth of mixed vesicles at sublytic surfactant concentrations,18 and a number of intermediate aggregates of different shapes and stability.19-20 Other authors, however, claim a more simplistic transition ruled only by a slight increase in the size of vesicles and subsequent formation of mixed micelles. In this case, the complex intermediate aggregates were not detected.21 Concerning to the reconstitution process, several authors described that the presence of strong electrostatic interactions in the medium22 as well as the rate of surfactant removal and the concentration of lipid and surfactant critically affected the size of the reconstituted vesicles. Thus, upon slow surfactant
10.1021/jp010273w CCC: $20.00 © 2001 American Chemical Society Published on Web 09/14/2001
9880 J. Phys. Chem. B, Vol. 105, No. 40, 2001 depletion the structures formed were higher than those formed upon rapid surfactant removal.15,23-24 It has been also reported that the lower the proportion of surfactant in the reconstituted vesicles, the lower their final size.14,25 Although these studies reported useful information about the solubilization and reconstitution of vesicles, additional information is needed to clarify kinetic and structural aspects of both processes. In earlier papers, we reported the solubilization of PC liposomes by various surfactants.26-28 In the case of the nonionic Triton X-10029 and the anionic sodium dodecyl sulfate,30 we demonstrated that the solubilization was ruled by direct formation of mixed micelles within the bilayer without formation of complex intermediate aggregates. Kinetic studies showed that the particle size variations and the times associated with this transformation were attributed to the electrostatic and structural characteristics of each surfactant.31-32 Here, we seek to extend these investigations by studying structural and kinetic aspects involved in the solubilization and reconstitution of PC liposomes by addition or depletion of OG. To this end, freeze-fracture electron microscopy and dynamic light scattering (DLS) techniques based on the use of an Ar laser source were used. The use of these two complementary techniques opens up new avenues in the study of the solubilization and reconstitution of biological membranes by surfactants. Experimental Section Materials. Phosphatidylcholine (PC) was purified from egg lecithin (Merck, Darmstadt, Germany) by the method of Singleton33 and was shown to be pure by thin-layer chromatography (TLC). The nonionic surfactant n-octyl β-D-glucopyranoside (octyl glucoside, OG) was purchased from Sigma Chemicals Co. Tris(hydroximethyl)-aminomethane (TRIS) was obtained from Merck. TRIS buffer was prepared as 5.0 mM TRIS adjusted to pH 7.4 with HCl, containing 100 mM NaCl, and was filtered through Millipore membranes type GS 0.22 µm (Bedford, USA). Methods. Liposomes of a defined size (of about 200 nm) and PC concentrations ranging from 0.5 to 5.0 mM were prepared by extrusion of large unilamellar vesicles (through 800-200 nm polycarbonate membranes) previously obtained by reverse phase evaporation.27 A film was formed by removing the organic solvent by rotatory evaporation from a chloroform solution of PC. The lipid was then redissolved in diethyl ether, and TRIS buffer was added to the solution. Gentle sonication led to the formation of a W/O type emulsion. After evaporation of the diethyl ether under reduced pressure, a viscous gel was formed. The elimination of the final traces of the organic solvent at high vacuum transformed the gel into a liposome suspension, in which no traces of ether were detectable by NMR.34 Solubilizing Parameters. Different systems containing OG (concentrations ranging from 5.0 to 40.0 mM, critical micelle concentration, CMC, 18.0 mM27) and PC liposomes (PC concentrations ranging from 0.5 to 5.0 mM) were studied at the equilibrium. We consider that a system is in equilibrium when it reaches a stable state and the system remains so until it is altered by any other factor. The overall solubilization was characterized by two parameters termed ReSAT and ReSOL, according to the nomenclature adopted by Lichtenberg.14 These parameters corresponded respectively to the OG/PC molar ratios at which the surfactant saturated liposomes (ReSAT) and led to a complete solubilization of these structures (ReSOL). The determination of these parameters can be carried out on the basis of the linear dependence existing between the surfactant concentrations required to achieve these parameters and the PC
Lo´pez et al. concentration in liposomes, which can be described by the equation:
ST ) SW + Re[PC]
(1)
where ST is the total surfactant concentration, Re is the effective surfactant to PC molar ratio in bilayers (Re ) SB/PC), SB is the surfactant concentration in bilayers, SW is the surfactant concentration in the aqueous phase, and PC is the lipid concentration. The SW and the Re in each curve are the slope and the ordinate at the origin (zero PC concentration), respectively. Monitoring the Solubilization and Reconstitution Processes. Measurements of the hydrodynamic diameter (HD) of pure PC vesicles, pure OG micelles, and particles formed after mixing different concentrations of OG with PC liposomes were determined by means of a DLS technique using a photon correlator spectrometer (Malvern Autosizer 4700c PS/MV) equipped with an Ar laser source (wavelength 488 nm). Quartz cuvettes were filled with the samples, and all the experiments were thermostatically controlled at 25 °C. All the DLS measurements were made with a scattering angle of 90°. To acquire the full range of decay times necessary to determine the signal from both the large and the small particles, a low sample time value (2 µs) and a dilatation of 3 with parallel subcorrelators was used. This allowed for the correlograms to extend up to 1.7 ms. The DLS measurements were performed in equilibrium and nonequilibrium systems to determine the phase diagram of the transition and the kinetics of the process, respectively. Thus, the evolution with time was followed for 24 h after mixing OG with liposomes. We were able to obtain data every 30 s (20 s for DLS measurement followed by 10 s for the analysis of data). This period of 30 s was the minimum needed to obtain statistically reliable data. The analysis of the data was performed using CONTIN software provided by Malvern Instruments (England). The validity of the CONTIN results was tested by fitting a single or a biexponential to the correlation function. If a biexponential had to be fitted, first a single exponential was fitted to a long time range and the second exponential was then fitted to the residual. Both methods agreed fairly well. The results are given as diameters, and the percentages correspond to intensity values. To characterize the aggregates formed during the solubilization at equilibrium (10 h after mixing) and as a function of time, a freeze fracture electron microscopy (FFEM) study was carried out. The experiments were done according to the procedure described by Egelhaaf.35 About 1 µL of suspension was sandwiched between two copper platelets using a 400-mesh gold grid as spacer. Then, the samples were cryofixed by dipping into nitrogen-cooled liquid propane at -180 °C and fractured at -110 °C and 5 × 10-7 mbar in a BAF-060 freeze-etching apparatus (BAL-TEC, Liechtenstein). The replicas were obtained by unidirectional shadowing with 2 nm of Pt/C and 20 nm of C, and they were floated on distilled water and examined in a Hitachi H-600AB TEM at 75 kv. As for the reconstitution process, three mixed micellar systems containing the PC and OG concentrations at which the complete solubilization is reached (systems corresponding to the OG/PC molar ratio ReSOL) were prepared. These micellar systems were submitted to a fast-step dilution with surfactantfree buffer that reduced the OG and the PC concentrations to the half. Both the size of the particles formed and the evolution with time of this process were monitored by DLS following the experimental procedure described above.
Octyl Glucoside-Mediated Solubilization and Reconstitution
Figure 1. Phase diagram of the aggregates formed by mixtures of OG and PC in water, showing the phase boundary of the three domains: a “micellar” region formed by only mixed micelles, a “coexistence” region formed by mixed micelles and fragmented vesicles, and a “vesicular” region formed by only mixed vesicles. The data points (stars) correspond to the following systems: point 1 to pure liposomes; point 2 to the vesicular phase boundary; point 3 to the coexistence region; and point 4 to the micellar phase boundary. Filled squares indicate the straight line for the saturation, line SAT, and open squares indicate the straight line for the solubilization, line SOL.
Results Solubilization of Liposomes and Stability of the Systems. The size distribution curves for pure micelle OG solutions (OG concentrations ranging from 20.0 to 40.0 mM) and for pure PC liposomes were first determined by DLS. The curves exhibited in both cases a monomodal distribution with a hydrodynamic diameter (HD) of 5 and 186 nm, respectively. This micellar size is in line with that reported by Lorber et al. using various techniques.36 To study the solubilization of PC liposomes, increasing OG amounts were added to liposomes at different PC concentrations (from 0.5 to 5.0 mM) and the size of the resulting aggregates was measured 10 h after mixing. This time was chosen as the optimum period needed to reach the equilibrium in all the systems for the lipid concentration range used. At low OG concentrations, only mixed vesicles were detected (monomodal distribution curves with a HD of about 186 nm). Increasing OG amounts led to the formation of mixed micelles, which were detected as small aggregates (bimodal distribution curves with a new peak at 11 nm). Higher OG concentrations resulted in the formation of systems, in which fragmented vesicles (about 145 nm diameter) and mixed micelles coexisted. The higher the surfactant concentration, the higher the proportion of mixed micelles. Finally, only mixed micelles were detected (monomodal distribution curves with a peak of 11 nm), indicating the complete solubilization of liposomes. In addition, no peaks corresponding to intermediate complexes aggregates were found in equilibrated systems throughout the process. Some of the particles’ sizes studied (those about 150-180 nm) did not generate Rayleigh light scattering. However, this fact does not affect our results because the HD of the particles, which determines the different phases, was obtained using the Zave directly. This value was calculated by the cumulant method applied to the autocorrelation function described by Koppel.37 When plotting the OG versus the PC concentrations for systems at which the formation of mixed micelles starts (saturation of liposomes) and those with the only presence of mixed micelles (liposome solubilization), a linear dependence was obtained (Figure 1) with regression coefficients (r2) 0.992 and 0.993, respectively. The slope of these two straight lines corresponded to the surfactant to PC molar ratio at which the surfactant saturated liposomes (ReSAT, line SAT in Figure 1) and led to the complete liposome solubilization (ReSOL, line SOL in Figure 1), respectively.14 These values were 1.4, and 3.2,
J. Phys. Chem. B, Vol. 105, No. 40, 2001 9881 respectively, in agreement with previous solubilization studies.7-8,38 Hence, line SAT and line SOL delimit three regions: a “vesicular” region formed only by mixed vesicles, a “coexistence” region, in which fragmented vesicles and mixed micelles coexisted, and a “micellar” region formed by only mixed micelles. From a structural viewpoint, some representative FFEM micrographs reflecting the solubilization of PC liposomes by rising OG amounts at the equilibrium (10 h after mixing) are shown in Figure 2. The four systems studied are indicated in Figure 1. Micrograph 1 for the system 1 (2.5 mM PC) shows some pure PC vesicles with an HD of about 186 nm. Micrograph 2 shows a system placed at the vesicular phase boundary (2.5 mM PC/20 mM OG, system 2) formed by mixed vesicles with a similar size to that found for pure PC vesicles (in accordance with the DLS data). However, some alterations in the topology of the vesicles were detected. Micrograph 3 shows a system for the coexistence region (2.5 mM PC/25 mM OG, system 3). Vesicles with strong morphological alterations or even fragmented vesicles of about 145 nm diameter (arrows) together with small particles (arrowheads) corresponding to OG/PC mixed micelles were observed. It is noteworthy that no other intermediate complex aggregates were observed, in accordance with the DLS data. A system placed on the micellar phase boundary (2.5 mM PC/26 mM OG, system 4) is shown in the micrograph 4, in which the only presence of mixed micelles (arrowheads) was detected, in accordance with the DLS data. Micrograph 5 shows the system 3 (2.5 mM PC/25 mM OG) 24 h after mixing. It can be noted that no changes in the appearance or size of the micelles (arrowheads) were detected; however, an enlarging with changes in the topology of the fragmented vesicles were observed reflecting the unstability of the system with time. To study the solubilization kinetics, three periods of time were considered: (1) the “micelle induction” as the time needed for micellization to start to occur (previously described by Lo´pez et al.31); (2) the “equilibration time” as the subsequent period needed for the formation of stable particles both in percentage and in size, that is, the time needed to reach a temporary equilibrium state. This period will be associated with the end of the process for a given system (no more changes associated with solubilization were detected); (3) the “stability time” as the following period, in which the equilibrium system remained without changes. Figure 3 plots the particle size variations 30 s (A), 10 h (B), and 24 h (C) after mixing increasing OG concentrations (from 0 to 40 mM) with PC liposomes (2.5 mM) in independent samples. After 30 s of mixing and up to 26 mM OG, only one peak for mixed vesicles (HD of about 186 nm) was detected. Higher OG concentrations led to the presence of only mixed micelles (new peak for particles with an HD of 11 nm). After 10 and 24 h of mixing and up to 20 mM OG, the size of the mixed vesicles remained almost constant. However, in a range of OG concentrations from 20 to 26 mM, fragmented vesicles and mixed micelles coexisted with noticeable changes in the size of fragmented vesicles. OG concentrations higher than 26 mM gave rise to only mixed micelles. These structures were detected 30 s after mixing, and they remained stable for weeks. Table 1 shows the variation in size and in percentage of systems 3 (coexistence region) and 4 (micellar phase boundary) (Figure 1) with time. Both systems 3 and 4 showed similar micelle induction and equilibration times; however, only in system 4 was the complete solubilization achieved. Thus, the systems at equilibrium (10 h after mixing) consisted in the first case (system 3) of a mixture of 27-28% fragmented vesicles and 72-73% mixed micelles, and in the second (system 4) of
9882 J. Phys. Chem. B, Vol. 105, No. 40, 2001
Lo´pez et al.
Figure 2. Freeze-fracture electron microscopy images reflecting the solubilization of 2.5 mM PC liposomes by rising OG amounts at equilibrium (10 h after mixing). Micrograph 1 shows some pure PC vesicles (point 1, Figure 1). Micrograph 2 shows a system placed on the “vesicular phase boundary” (point 2, Figure 1). Micrograph 3 shows a system for the “coexistence” region (point 3, Figure 1). A system placed on the “micellar phase boundary” is shown in the micrograph 4 (point 4, Figure 1). Micrograph 5 shows the “coexistence” system of micrograph 3 but 24 h after. Structures are marked as follows: fragmented vesicles with arrows and mixed micelles with arrowheads.
100% mixed micelles. The stability time was in the first case of about 20 h (from this time the size of fragmented vesicles rose approximately two times and their proportion decreased), whereas in the second case the system was stable for weeks. From all these data, it can be noted that the solubilization of the system containing 26 mM OG was completed in 4.5 min
(Table 1), whereas more concentrated systems achieved total solubilization before 30 s (Figure 3). Thus, systems with a higher OG concentration led to micellar systems that were greatly stable and showed a faster solubilization kinetics. Reconstitution of Vesicles by Dilution. From the diagram of Figure 1, it is expected that mixed micellar solutions resulted
Octyl Glucoside-Mediated Solubilization and Reconstitution
J. Phys. Chem. B, Vol. 105, No. 40, 2001 9883 TABLE 1: Particle Size Distributions for the Systems 2.5 mM PC/25 mM OG (system 3) and 2.5 mM PC/26 mM OG (system 4) as a Function of Time during the Solubilization Processa 2.5 mM PC/25 mM OG (system 3) time (min) 0 0.5 1 1.5 3 3.5 4.5 10 60 300 600 800 1200 1440
first peak nm %
second peak nm %
11 11 11 12 12 11 11 12 12
186 187 186 186 150 145 144 144 145 144 151 150 152 295
57 73 70 72 73 72 73 74 80
100 100 100 100 100 43 27 30 28 27 28 27 26 20
2.5 mM PC/26 mM OG (system 4) first peak nm %
12 12 11 11 11 12 11 11 11 11
60 85 100 100 100 100 100 100 100 100
second peak nm % 190 189 190 188 190 190
100 100 100 100 40 15
a Results are given as HD and the percentage corresponds to intensity values.
Figure 3. Particle size variation of the aggregates formed by 2.5 mM PC liposomes and increasing concentrations of OG 30 s (A), 10 h (B), and 24 h (C) h after mixing.
in the spontaneous formation of vesicles when the vesicular phase boundary was crossed by dilution. To confirm this finding, various micellar systems were submitted to a fast one-step dilution. The experiments were designed in such way that the three mixed micellar systems were diluted starting from the same state. This state contained the OG and PC concentrations at which the complete solubilization was reached (data points on the line SOL of Figure 1). The slope of this line corresponded to the OG/PC molar ratio in the micelles (ReSOL) common in all the mixed micellar systems submitted to dilution. This dilution with buffer reduced the concentration of both components to half being the phase boundary for the vesicular region crossed in all cases. Taking into account that no surfactant removal was performed during the dilution process, we assume that the reconstituted vesicles included surfactant molecules and, hence, were mixed vesicles. The compositions of the three systems studied before dilution were 5.0 mM PC/35 mM OG, 2.5 mM PC/26 mM OG mM (system 4), and 1.25 mM PC/22 mM OG. The variation in the proportion of both components with respect to the ReSOL (effective surfactant to PC molar ratio in bilayers for vesicle solubilization) is because the total OG concentration (ST) added is the sum of the surfactant in bilayers (SB) and in the aqueous phase (SW), in accordance with the eq 1. In the three systems studied, the HD of the particles before
Figure 4. The continuous lines of this diagram plot the composition of systems 5.0 mM PC/35 mM OG (square), 2.5 mM PC/26 mM OG mM, system 4 (circle), and 1.25 mM PC/22 mM OG (triangle) before (open symbols) and after (filled symbols) a fast one-step dilution. The strong line plots the linear dependence between the reconstituted vesicle size and their PC concentration. The discontinuous lines correspond with line SOL and line SAT of Figure 1.
dilution was 11 nm (mixed micelles). After dilution, the size distribution curves for reconstituted vesicles also showed a monomodal distribution with HD values of 110, 60, and 45 nm, respectively, these sizes remaining stable for weeks. The composition of systems before and after dilution and the experimental size of the reconstituted vesicles are shown in Figure 4. A linear dependence was established between the size of the reconstituted vesicles and the PC concentrations. To study the kinetics of vesicle reconstitution by DLS, three periods of time were considered: (1) the “vesicle induction” time as the period needed for the vesicle reconstitution to start, (2) the “equilibration” time as the subsequent period needed for the complete formation of reconstituted vesicles (period associated with the end of the reconstitution process), and (3) the “stability” time as the period in which the system remained unchanged. Table 2 shows the kinetics of vesicle reconstitution for the systems 2.5 mM PC/26 mM OG (system 4) and 1.25 mM PC/ 22 mM OG. This reconstitution process resulted in vesicles with 60 and 45 nm diameter, respectively, which indicates a certain dependence between the size of the vesicles formed and the initial concentration of PC/OG. Although the vesicle induction time was similar in both systems (about 30 s), the equilibrium time increased as the concentration of PC and OG decreased (2 and 2.5 min, respectively). The experiments were performed
9884 J. Phys. Chem. B, Vol. 105, No. 40, 2001
Lo´pez et al.
TABLE 2: Particle Size Distributions for the Systems 2.5 mM PC/26 mM OG (system 4) and 1.25 mM PC/22 mM OG as a Function of Time during the Reconstitution Process by a Fast One-Step Dilutiona 2.5 mM PC/26 mM OG (system 4) time (min)
first peak nm %
0 0.5 1 1.5 2.0 2.5 5.0 10 60 600 1440
11 11 11 11
100 98 51 4
second peak nm % 56 58 59 60 60 60 60 60 60 60
2 49 96 100 100 100 100 100 100 100
1.25 mM PC/22 mM OG first peak nm % 11 11 11 11 11
100 99 79 48 5
second peak nm % 43 44 45 45 45 45 45 45 45 45
1 21 52 95 100 100 100 100 100 100
a Results are given as HD and the percentage corresponds to intensity values.
Figure 5. Variation in the percentage of mixed micelles as a function of time during the solubilization (A) and reconstitution by dilution (B) of PC vesicles for the systems formed by 2.5 mM PC/26 mM OG (9) (system 4) and by 1.25 mM PC/22 mM OG (O).
by dilution of micelles in which the PC and OG concentrations varied simultaneously. Considering this fact, to assess a dependence between the equilibrium time and the PC or OG concentration in the reconstitution process is difficult. Such dependence should be studied in future work, and experiments in which the PC or OG concentrations remain constant should be considered. The stability time, however, was in both cases higher than 24 h, indicating that in the period studied the stability of the vesicles formed was independent of the concentration of the components. In the system formed by 5.0 mM PC/35 mM OG, the reconstitution process, which resulted in vesicles with 110 nm diameter, was too rapid to be followed by DLS. To compare the kinetics of vesicle solubilization and reconstitution for the systems given in Table 2, the variation in the proportion of mixed micelles for these two processes are plotted versus time in Figure 5, panels A and B, respectively. The second process was faster basically due to the short induction time associated with the formation of mixed micelles.
Discussion From the data obtained using DLS and FFEM techniques, a simple mechanism of PC liposomes solubilization by OG is proposed. Although this mechanism is based on the three-stage model proposed by various authors,14,19,39-40 it differs in the following points: (a) When the surfactant molecules are added to a liposome suspension, they are distributed into the lamellar phase of liposomes up to a critical concentration. At this point, mixed micelles start to form within the saturated PC bilayers; this concept is named “in situ micellization”. This term was developed in previous papers to explain the first steps of the PC liposomes solubilization by other surfactants.29,30 Although these micelles within the bilayers were unstable structures, at this step of the interaction they remain in a relative steady state and their study by FFEM was as reliable as that for transmembrane proteins. (b) The progressive separation of the formed micelles from the liposome surface led to the complete solubilization of vesicles. (c) This separation would take place without formation of stable complex intermediate aggregates, with only the presence of mixed micelles and fragmented vesicles. The dimension of these two structures was different, about 11 nm diameter for the mixed micelles and about 150 nm for the fragmented vesicles. The data obtained by FFEM (micrograph 3) and by DLS (Table 1, system 3, 600 min) confirm these differences. Such results contrast partly with those reported in our previous work.27 This paper reported that at the beginning of the solubilization, DLS and negative staining-TEM showed liposomes of about 200 nm diameter, and at the end particles with 57 nm diameter. In addition, intermediate aggregates were detected. The improvement of the techniques in the present paper would explain the differences observed between the works. The use of an Ar laser (wavelength of 488 nm) instead of the He-Ne laser source (wavelength of 632 nm), and the use of FFEM instead of negative staining-TEM have allowed one to obtain more reliable results in the solubilization process. The technique of Cryo-TEM (CTEM) has proved to be very useful in the study of vesicle shape and size.18,19,41 However, this technique was unsuitable to visualize the “in situ” micellization because CTEM does not report topological information. In any case, one should taken into account that the size of the mixed vesicles and mixed micelles is dependent on the surfactant-lipid composition, on the rate of surfactant addition, and on the stability conditions of the systems.7,41 The solubilization mechanism was similar to that described for the nonionic surfactant Triton X-100 (TX-100) and the anionic sodium dodecyl sulfate (SDS).29-30 However, some specific structural differences were observed depending on the physicochemical characteristics of each surfactant. Thus, using OG the size of saturated vesicles was unaffected (about 185 nm diameter); in the case of TX-100 this size increased (from 160 to 200 nm diameter), and using SDS the vesicle saturation occurred with a slight contraction (from 190 to 170 nm diameter). The preservation of the vesicular size after incorporation of the OG molecules contrasts with the results previously reported by Edwards and Almgren.18 These authors claimed an important increase in the size of the vesicles saturated by surfactant. However, considering our results and the work reported by Kragh-Hansen et al.,21 the enlarging reported by Edwards and Almgren could be explained as a result of a fusion between vesicles rather than being due to real growth. Furthermore, the preservation of the OG saturated vesicles size is in conflict with some theoretical considerations, which justify that this incorporation led to a vesicular growth due to the increase in the total surface area of mixed vesicles. Taking into account
Octyl Glucoside-Mediated Solubilization and Reconstitution the ReSAT ) SB/PC ) 1.4, the molecular surface area of PC in the bilayer ) 65 Å2, the molecular surface area of OG ) 38 Å2, and the surface area of the sphere A ) 4πR2,41 a PC vesicle with initial size of 200 nm should increase its diameter by at least 40 nm due to the increase in the total surface area of mixed vesicle bilayer. The only way to maintain the same vesicle size would be to change the bilayer topology. In this sense, micrograph 2 in Figure 2 presents evidence of such alteration, so it is clearly observed that OG-containing vesicles have a cramped wavelike appearance (wavy shape)42,43 in contrast to the spherical smooth surface of pure PC vesicles. DLS studies show that the level of liposome solubilization (PC concentration 2.5 mM) and the kinetics of liposome solubilization were dependent on the OG concentration (Figure 3). Thus, the addition of OG concentrations lower than 20 mM (OG CMC 18 mM) led to the formation of stable mixed vesicles (even when 1 mM of OG is added). These mixed vesicles were detected 30 s after mixing liposomes and OG and they remained stable for weeks. However, the addition of OG concentrations starting from 27 mM led to the formation of mixed micelles also 30 s or less after mixing. Hence, for high OG concentrations, the sum of micelle induction and equilibration times were probably too rapid to be studied by DLS. Intermediate OG concentrations led to the formation of measurable systems by DLS, in some of which fragmented vesicles and mixed micelles coexisted (Table 1). The micelle induction time was inversely dependent on the surfactant concentration (3.5 and 3 min for OG concentrations of 25 and 26 mM, respectively). However, the stability time was dependent on the particles present at the equilibrium. The systems with coexistence of mixed micelles and fragmented vesicles were less stable. The loss of stability of system 3 was reflected in the growth of fragmented vesicles up to doubling the initial size and reduction in its proportion (Table 1). These variations could be explained by fusion or aggregation of the fragmented vesicles. Micrograph 3 and 5 of Figure 2 shows system 3 (2.5 mM PC/25 mM OG) 10 and 24 h after of mixing. Comparison of these two micrographs confirms the increase in size of fragmented vesicles with time. This increase in size suggests a fusion mechanism (no signs of aggregation were detected) that lead the fragmented vesicles to big vesicles. These fused vesicles exhibited similar morphology as the initial fragmented vesicles. Thus, the vesicle size in the coexistence systems decreased from 180 nm (initial size) to 150 nm (after 10 h) because they were converted to fragmented vesicles by the effect of the OG. The subsequent increase in size (detected 24 h after mixing) were possibly induced by the fusion of these particles due to their poor stability. It is a common observation that surfactants may induce vesicle fusion when interacted with liposomes at sublytic concentrations.8,19,44 However, in our experiments, the vesicle fusion seems to take place in the systems of coexistence micelle vesicle; in particular, this phenomenon occurred in the step of the separation of mixed micelles from the liposome surface. This fusion may be explained because under these conditions the membranes are in a fluctuating dynamic state characterized by rapid exchange of phospholipid that favors fusion phenomena as reported by Kragh-Hansen et al.21 Another possible cause could be the formation of holes on the vesicle surface, which could act as critical points of fusion. Comparison of solubilization kinetics of PC liposomes by two nonionic surfactants, OG and Triton X-100 (TX-100), shows that in systems in which vesicles and micelles coexisted the TX-100 had lower times of micelle induction and equilibration.32 This indicates that the solubilization induced by TX-100 was
J. Phys. Chem. B, Vol. 105, No. 40, 2001 9885 faster than that induced by OG. Taking into account that the CMC of OG was approximately 120 times higher than that of TX-100,26 a higher number of OG monomers may have to be incorporated in the bilayers before the micellization “in situ” starts. This fact could be the cause of the higher time of micelle induction for the OG. In this sense, Lasch et al.45 compared the interaction of these two surfactants with phospholipid membranes demonstrating that TX-100 (lower CMC than the OG) induced an optimization of packing of the lipid molecules. This appropriate packing could favor the mechanism of in situ micellization proposed here and, hence, could be related to the faster solubilization kinetics induced by the TX-100. In addition to the surfactant CMC, another aspect that should be considered in the kinetic study of solubilization is the hydrophobicity of the surfactants. We are aware that the hydrophilic/lipophilic balance of surfactants is not a relevant parameter to measure their penetration on membranes and that other parameters could be involved. However, Kragh-Hansen et al.21 associated the surfactant hydrophobicity with a higher rate of transbilayer movement (flip-flop) and, hence, with a faster solubilization kinetics. Consequently, given the higher hydrophobicity of OG (hydrophilic-lipophilic balance of OG and TX-100 12.6 and 13.6, respectively46,47), it would be expected that the OG exhibited a faster solubilization kinetics than the TX-100. However, in the case of these two surfactants with so different CMC values, the argument of hydrophobicity appears not to be appropriate. The surfactant CMC seems to be an important factor to explain the different solubilization kinetics induced by these surfactants. The work reported by Israelachvilli48 in which the hydrophilic and lipophilic properties of poly(ethylene oxide) are discussed supports our results. The TX-100 (containing ethylene oxide units) may be expected more easily than the OG to be inserted in the membranes, and this fact could be associated with the faster kinetic of solubilization induced by the TX-100. As for the reconstitution process, although the micellar size was initially the same (11 nm) a linear dependence was established between the size of the reconstituted vesicles and their composition (Figure 4). From a practical viewpoint, this finding opens the possibility to predict the size of reconstituted vesicles after a controlled one-step fast dilution of specific OGPC micellar systems (Figure 4). This dependence between the size of the vesicles and their OG-PC composition may be related to the work reported by Paternostre et al.8 concerning the study of the transition micelle-vesicle of egg-PC and OG. These authors described the existence of a unique vesicular state at the onset of the micellization that is identical regarding the shape, molecular organization, and initial size of the vesicles. This strongly supports the existence of an equilibrium size related to the curvature radius and the surfactant and lipid composition of the vesicles. Work based on the study of the reconstitution of liposomes have led to interesting information about the size of the particles in this process. Ollivon et al.49 reported that the rate of surfactant removal affects the size of the final reconstructed vesicles, and Das et al.50 estimated empirically that the vesicular diameter should be about four times larger than that of the micelles. Given our results in the reconstitution experiments, not only the considerations of these authors but also the total OG-PC concentration of the systems should be considered in the size prediction of the particles involved in the vesicle reconstitution process. Results reported by various authors have led to the claim that the molecular rearrangements during the vesicle reconstitution follow the same pathway as the solubilization event.15-16
9886 J. Phys. Chem. B, Vol. 105, No. 40, 2001 However, in the present work, it is demonstrated that the kinetics of reconstitution was shorter than that of solubilization (vesicle induction time 30 s and micelle induction time 3-3.5 min). However, solubilization and reconstitution are two different processes, and our results suggest that the different induction times could involve different requirements for the formation of mixed micelles and mixed vesicles. Thus, the formation of mixed micelles requires first the initial release of the surfactant monomers from pure micelles governed by a micellar relaxation time τ2.51 Second, the incorporation of these monomers into bilayers possibly through the hydrophilic holes created by these monomers on the PC polar heads or via formation of shortlived complexes surfactants-PC polar heads.52 Finally, the formation of mixed micelles would involve not only the adsorption of surfactant monomers into bilayers but also a desorption of mixed micelles from the bilayer surface.29 However, the reconstitution of vesicles from micellar solutions only would involve the release of surfactants monomers from the mixed micelles and the reorganization of the remaining lipid and surfactant molecules in vesicles in a simpler way than that described until now. Furthermore, other additional factors such as the different thermodynamics of these two processes13 could also be involved. Acknowledgment. We thank Dr. Carmen Lo´pez-Iglesias and Dr. David Bellido from the Serveis Cientı´fico-Te´cnicos, Universidad de Barcelona, for their skillful work at FFEM experiments. This work was supported by funds from DGICYT (Direccio´n General de Investigacio´n Cientı´fica y Te´cnica), Spain. References and Notes (1) Das, T. K. J. Phys. Chem. 1996, 100, 20143. (2) Parmar, M. M.; Edwards, K.; Madden, T. D. Biochim. Biophys. Acta 1999, 1421(1), 77. (3) Le Marie, M.; Champeil, P.; Moller, J. V. Biochim. Biophys. Acta 2000, 1508, 86. (4) Lacapere, J. J.; Robert, J. C.; Thomas-Soumarmon, A. Biochem. J. 2000, 345(2), 239. (5) Bogusz, S.; Venable, R. M.; Pastor, R. W. J. Phys. Chem. B 2000, 14, 5462. (6) Ollivon, M.; Lesieur, S.; Grabielle-Madelmont, C.; Paternostre, M. Biochim. Biophys. Acta 2000, 1508, 34. (7) Almog, S.; Litman, B. J.; Wimley, W.; Cohen, J.; Wachtel, E. J.; Barenholz, Y.; Ben-Shaul A.; Lichtenberg, D. Biochemistry 1990, 29, 4582. (8) Paternostre, M.; Meyer, O.; Grabielle-Madelmont, C.; Lesieur, S.; Ghanam, M.; Ollivon, M. Biophys. J. 1995, 69, 2476. (9) Angelov, B.; Ollivon, M.; Angelova, A. Langmuir 1999, 15, 8225. (10) Heerklotz, H.; Seelig, J. Biophys. J. 2000, 78, 2435. (11) Zheng, Y.; Lin, Z.; Zakin, J. L.; Talmon, Y.; Davis, H. T.; Scriven, C. E. J. Phys. Chem. B 2000, 104, 5263. (12) Partearroyo, A.; Alonso, A.; Gon˜i, F. M.; Tribout, M.; Paredes, S. J. Colloid Interface Sci. 1996, 178, 156. (13) Wenk, M. R.; Seelig, J. J. Phys. Chem. 1997, 101, 5224. (14) Lichtenberg, D.; Robson, R. J.; Dennis, E. A. Biochim. Biophys. Acta 1983, 737, 285. (15) Levy, D.; Gulik, A.; Seigneuret, M.; Rigaud, J. L. Biochemistry 1990, 29, 9480. (16) Knol, J.; Sjollena K.; Poolman. B. Biochemistry 1998, 37(46), 16410.
Lo´pez et al. (17) Rigaud, J. L.; Chami, M.; Lambert, O.; Levy, D.; Rank, J. L. Biochim. Biophys. Acta 2000, 1508, 112. (18) Edwards, K.; Almgren, M. Prog. Colloid. Polym. Sci. 1990, 82, 190. (19) Edwards, K.; Almgren, M. Langmuir 1992, 8, 824. (20) Polozova, A. I.; Dubachev, G. E.; Simonova, T. N.; Barsukov, L. I. FEBS Lett. 1995, 358, 17. (21) Kragh-Hansen, U.; Le Marie, M.; Moller, J. V. Biophys. J. 1998, 75, 2932. (22) Oberdisse, J.; Regev, O.; Porte, G. J. Phys. Chem. B 1998, 102, 1102. (23) Cladera, J.; Rigaud, J. L.; Villaverde, J.; Dun˜ach, M. Eur. J. Biochem. 1997, 243, 798. (24) Lambert, O.; Levy, D.; Ranck, J. L.; Leblanc, G.; Rigaud, J. L. Biophys. J. 1998, 74, 918. (25) Schurtenberger, P.; Sva¨rd, M.; Wehrli, E.; Lindman, B. Biochim. Biophys. Acta 1986, 882, 465. (26) De la Maza, A.; Parra, J. L. Biochem. J. 1994, 303, 907. (27) De la Maza, A.; Parra, J. L. Eur. J. Biochem. 1994, 226, 1029. (28) De la Maza, A.; Parra, J. L. Biophys. J. 1997, 72, 1668. (29) Lo´pez, O.; De la Maza, A.; Coderch, L.; Lo´pez-Iglesias, C.; Wehrli, E.; Parra, J. L. FEBS Lett. 1998, 426, 314. (30) Lo´pez, O.; Co´cera, M.; Wehrli, E.; Parra, J. L.; De la Maza, A. Arch. Biochem. Biophys. 1999, 367, 153. (31) Lo´pez, O.; Co´cera, M.; Pons, R.; Azemar, N.; De la Maza, A. Langmuir 1998, 14, 4671. (32) Lo´pez, O.; Co´cera, M.; Pons, R.; Azemar, N.; Lo´pez-Iglesias, C.; Wehrli, E.; Parra, J. L.; De la Maza, A. Langmuir 1999, 13, 4678. (33) Singleton, W. S.; Gray, M. S.; Brown, M. L.; White, J. L. J. Am. Oil Chem. 1965, 42, 53. (34) Allen, T. M. Removal of detergent and solvents and solvents traces from liposomes. In Liposome Technology, G. Gregoriadis, Ed.; CRC Press: Boca Raton, FL, 1986; Vol. 3. (35) Egelhaaf, S. U.; Wehrli, E.; Mu¨ller, M.; Adrian, M.; Schurtenberger, P. J. Microsc. 1996, 184, 214. (36) Lorber, B.; Bishop, J.; DeLucas, L. Biochim. Biophys. Acta 1990, 1023, 254. (37) Koppel, D. E. J. Chem. Phys. 1972, 57, 4814. (38) Keller, M.; Kerth, A.; Blume, A. Biochim. Biophys. Acta 1997, 1326(2), 178. (39) Jackson, M. L.; Schmidt, C. F.; Lichtenberg, D.; Litman B. J.; Albert, A. D. Biochemistry, 1982, 21, 4576. (40) Gon˜i, F. M.; Urbaneja, M. A.; Arrondo, J. L.; Alonso, A.; Durrani, A. A.; Chapman, D. Eur. J. Biochem. 1986, 160, 659. (41) Vinson, P. K.; Talmon, Y.; Walter, A. Biophys. J. 1989, 56, 669. (42) Entenmann, A.; Wedel, T.; Kolbert, G.; Kuhnel, W.; Bruch, J. P. Anat. Anz. 1999, 181, 237. (43) Southan, G.; Firtel, M.; Blackford, B. L.; Jericho, M. H.; Xu, W.; Mulhern, P. J.; Beveridge, T. J. J. Bacteriol. 1993, 175, 1946. (44) Silvander, M.; Karlsson, G.; Edwards, K. J. Colloids Interface Sci. 1996, 179, 104. (45) Lasch, J.; Hoffmann, J.; Omelyanenko, W. G.; Klibadov, A. A.; Torchilin, W. D.; Binder, H.; Gawrisch, K. Biochim. Biophys. Acta 1990, 1022, 171. (46) Shinoda, K.; Kunieda H. Phase properties of emulsions: PIT and HLB. In Encyclopedia of Emulsion Technology, Becher, P., Ed.; Marcel Dekker: New York, 1985; Vol. 1, p 337. (47) Van Ede, J.; Nigmeijer, J. R. J.; Welling-Wester, S. J. Chromatogr. 1989, 476, 319. (48) Israelachvilli, J.; Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 8378. (49) Ollivon, M.; Eidelman, O.; Blumenthal, R.; Walter, A. Biochemistry 1988, 27, 1695. (50) Das, T. K. J. Chem. Soc., Faraday Trans. 1996, 92, 4279. (51) Patist, A.; Chhabra, V.; Pagidipati, R.; Shah, R.; Shah, D. O. Langmuir 1997, 13, 432. (52) Lasic, D. D. Liposomes: From Physics to Applications; Elsevier Science Publishers B. V.: Amsterdam, 1993; Chapter 2.
