Detergent Effects on Membranes at Subsolubilizing Concentrations

pubs.acs.org/Langmuir © 2010 American Chemical Society

Detergent Effects on Membranes at Subsolubilizing Concentrations: Transmembrane Lipid Motion, Bilayer Permeabilization, and Vesicle Lysis/ Reassembly Are Independent Phenomena Hasna Ahyayauch,† Mohammed Bennouna,‡ Alicia Alonso,† and F
elix M. Go~ni*,† †

Unidad de Biofı´sica (Centro Mixto CSIC-UPV/EHU) and Departamento de Bioquı´mica, Universidad del Paı´s Vasco, Aptdo. 644, 48080 Bilbao, Spain, and ‡Department of Biology, Faculty of Sciences Semlalia, Marrakech, Morocco Received November 4, 2009. Revised Manuscript Received February 3, 2010

Soluble amphiphiles, or detergents, are known to produce a number of structural and dynamic effects on membranes, even at concentrations below those causing membrane solubilization (i.e. in the so-called stage I of detergentmembrane interaction). The main subsolubilizing detergent effects on membranes are transmembrane lipid motion (flip-flop), breakdown of the membrane permeability barrier (leakage), and vesicle lysis/reassembly. For a proper understanding of membrane solubilization by detergents, it is important to assess whether the various effects seen at subsolubilizing surfactant concentrations occur independently from each other or are interconnected by cause-effect relationships so that they can be interpreted as necessary steps in the overall process of solubilization. To answer this question, we have explored the three above-mentioned effects (i.e., flip-flop, leakage, and lysis/reassembly) apart from solubilization in model (large unilamellar vesicles) and cell (erythrocyte) membranes. Five structurally different surfactants, namely, chlorpromazine, imipramine, Triton X-100, sodium dodecylsulfate, and sodium deoxycholate have been used. Each of them behaves in a unique way. Our results reveal that lipid flip-flop, vesicle leakage, and vesicle lysis/reassembly occur independently between them and with respect to bilayer solubilization so that they cannot be considered to be necessary parts of a higher-order unified process of membrane solubilization by detergents.

Introduction Detergents (or surfactants) are molecules that are easily dispersed in water and contain topologically separate hydrophobic and hydrophilic parts. Such an architecture ensures their ability to interact with the cell membrane and modify its properties. Surfactant-induced vesicle solubilization is often interpreted in terms of a three-stage model.1-3 During the first stage, surfactant molecules distribute between the aqueous and lipidic compartments and the latter, often in the form of a vesicle bilayer, start to lose their original shape/sphericity and physical stability. In stage 2, the surfactant-saturated vesicle bilayer coexists with lipiddetergent mixed fragments/micelles. Increasing the surfactantto-lipid ratio finally forces all surfactant-enriched bilayer fragments to convert to small, mixed amphipath micelles (step 3) (i.e., complete bilayer solubilization). As mixed micelles diminish in size they also accumulate more surfactant molecules. The most effective detergents finally create minute complexes containing just a few lipid and surfactant molecules held together mostly by hydrophobic forces.4 Lipid solubilization can proceed quickly or else take hours to complete5-10 depending on the molecular properties and concentration of the employed amphipaths. Stage I is the *Corresponding author. Fax: þ34 94 601 33 60. E-mail: [email protected].

(1) Helenius, A.; Simons, K. Biochim. Biophys. Acta 1975, 415, 29–79. (2) Lichtenberg, D. Biochim. Biophys. Acta 1985, 821, 470–478. (3) Heerklotz, H. Q. Rev. Biophys. 2008, 41, 205–264. (4) Almgren, M. Biochim. Biophys. Acta 2000, 1508, 146–163. (5) Partearroyo, M. A.; Alonso, A.; Goni, F. M.; Tribout, M.; Paredes, S. J. Colloid Interface Sci. 1996, 178, 156–159. (6) Madden, T. D.; Cullis, P. R. J. Biol. Chem. 1984, 259, 7655–7658. (7) Schubert, R. W.; Schmid, K. H.; Roth, H. J. Chem. Phys. Lipids 1991, 58, 121–129. (8) Viguera, A. R.; Gonz
alez-Ma~nas, J. M.; Taneva, S.; Go~ni, F. M. Biochim. Biophys. Acta 1994, 1196, 76–80. (9) Lasch, J. Biochim. Biophys. Acta 1995, 1241, 269–292. (10) Heerklotz, H.; Seelig, J. Biochim. Biophys. Acta 2000, 1508, 69–85.

Langmuir 2010, 26(10), 7307–7313

least well understood, yet for the correct use of detergents, it is important to have detailed knowledge of how and in which amounts they interact with integral membrane proteins and membrane lipid under both solubilizing and nonsolubilizing conditions. Furthermore, at nonsolubilizing concentrations, surfactants serve useful purposes in biological experimentation as agents that permeabilize or perturb membrane structure in various ways.11-14 Among the surfactant-induced phenomena under subsolubilizing conditions (stage I), detergent monomer binding to bilayers has been studied by isothermal titration calorimetry.10,15,16 Detergent effects at this stage include membrane permeabilization17,18 and changes in lipid chain order.16,19 Alonso et al.20,21 observed that Triton X-100 and other detergents caused, at concentrations below those producing bilayer solubilization, an increase in the turbidity of the vesicle suspension. This was interpreted in terms of the phenomenon of lysis and the reassembly of lipid vesicles. This (11) Andersen, J. P.; le Maire, M.; Kragh-Hansen, U.; Champeil, P.; Moeller, J. V. Eur. J. Biochem. 1983, 134, 205–214. (12) McIntosh, D. B.; Davidson, G. A. Biochemistry 1984, 23, 1959–1965. (13) Kragh-Hansen, U.; le Maire, M.; Noel, J. P.; Gulik-Krzywicki, T.; Moeller, J. V. Biochemistry 1993, 32, 1648–1656. (14) de Foresta, B.; Henao, F.; Champeil, P. Eur. J. Biochem. 1994, 223, 359– 369. (15) Tsamaloukas, A. D.; Keller, S.; Heeklotz, H. Nat. Protoc. 2007, 2, 695–704. (16) Arnulphi, C.; Sot, J.; Garcı´ a-Pacios, M.; Arrondo, J. L.; Alonso, A.; Go~ni, F. M. Biophys. J. 2007, 93, 3504–3514. (17) Ruiz, J.; Go~ni, F. M.; Alonso, A. Biochim. Biophys. Acta 1988, 937, 127– 134. (18) Ahyayauch, H.; Requero, M. A.; Alonso, A.; Bennouna, M.; Go~ni, F. M. J. Colloid Interface Sci. 2002, 256, 284–289. (19) Go~ni, F. M.; Urbaneja, M. A.; Arrondo, J. L.; Alonso, A.; Durrani, A. A.; Chapman, D. Eur. J. Biochem. 1986, 160, 659–665. (20) Alonso, A.; S
aez, R.; Villena, A.; Go~ni, F. M. J. Membr. Biol. 1982, 67, 55– 62. (21) Alonso, A.; Villena, A.; Go~ni, F. M. FEBS Lett. 1981, 123, 200–204.

Published on Web 02/19/2010

DOI: 10.1021/la904194a

7307

Article

Ahyayauch et al.

detergent effect was later characterized using cryo-transmission electron microscopy by Edwards and Almgren.22 Changes in the optical properties of merocyanine were used by Kaschny and Go~ni23 to detect a number of these detergent effects at stage I. Later, a novel effect of surfactants at subsolubilizing concentrations was described, namely, the induction of transmembrane, or flip-flop, lipid motion in model24 and erythrocyte25 membranes. In this article, we investigate stage I effects, namely, surfactantinduced flip-flop, vesicle lysis and reassembly, and surfactantinduced permeabilization, and the correlation between surfactant concentrations inducing subsolubilization effects and those causing solubilization. In particular, it was intended to determine whether these various phenomena occurred in a sequential way so that each could be understood as a result of the previous one. For this purpose, membrane perturbations caused by different surfactants, namely, chlorpromazine (CPZ), imipramine (IP), Triton X-100 (TX), sodium dodecylsulfate (SDS), and deoxycholate (DOX), were comparatively examined. Chlorpromazine (CPZ) is a phenothiazine that is used as an antipsychotic drug, and the structurally related imipramine (IP) is an antidepressant. Both are amphipathic molecules, a property that is believed to facilitate their targeting of cell membranes.26,27 CPZ and IP have also been found to induce membrane leakage and solubilization.18,28 TX was included in this study because it is probably the most commonly used nonionic surfactant in biochemical research. SDS and DOX are also well-known representatives of the anionic and steroid surfactants, respectively. We have observed that each of these five detergents has unique behavior at stage I and that flip-flop lipid motion, vesicle lysis and reassembly, membrane permeabilization, and solubilization do not appear to be interrelated phenomena.

Preparation of Small Unilamellar Vesicles (SUVs). Small unilamellar vesicles (SUVs) were prepared by the sonication of MLV in a bath sonicator for 20 min so that a clear suspension of SUVs was obtained. The SUV solution was then centrifuged to pellet larger lipid aggregates. The clear supernatant was retrieved for use in experiments. Turbidity. Liposome suspensions were mixed with the same volumes of the appropriate Triton solutions in the same buffer. Final lipid concentration, measured as lipid phosphorus, was 1 mM. Preliminary measurements had shown that a few minutes was enough for the various surfactants to equilibrate with the LUV suspensions. Still, the mixtures were left to equilibrate for 1 h at the desired temperature, and solubilization was assessed from the changes in turbidity.29 Turbidity was measured as the absorbance at 500 nm in a UVIKON spectrophotometer (Kontron Instruments, Milan, Italy) with continuous sample stirring. Turbidity values were normalized by setting 100% as the turbidity of the LUV suspension, 1 mM in lipid, in the absence of detergent, and 0% turbidity corresponded to pure buffer. Release of Vesicle Aqueous Contents. The leakage of vesicular aqueous contents was assayed with ANTS and DPX entrapped in the liposomes according to Ellens et al.30 Nonentrapped probes were removed by passing the LUV through a Sephadex G-75 column, which was eluted with 150 mM NaCl, 10 mM Hepes at pH 7.4. Assays were performed at 100 μM lipid in a total volume of 1 mL, with continuous stirring at 37 °C. Changes in fluorescence intensity were recorded with an Aminco-Bowman (Urbana, IL) AB-2 spectrofluorometer using 1 mL quartz cuvettes with continuous stirring. Excitation and emission wavelengths were 355 and 520 nm, respectively. An interference filter with a nominal cutoff value of 475 nm was placed in the emission light path to minimize the scattered-light contribution of the vesicles to the fluorescence signal. When the leakage reached equilibrium, excess Triton X-100 was added to induce 100% release. The percent release was computed as shown

Materials and Methods Surfactants, perylene, and bovine serum albumin (BSA) were purchased from Sigma (St. Louis, MO). Egg-yolk phosphatidylcholine was grade I from Lipid Products (South Nutfield, U.K.), and C6-NBD-PS and 1-oleoyl-2-[6(7-nitrobenz-2-oxa-1,3-diazol4-yl)amino]caproyl-sn-glycero-3-phosphoserine were obtained from Avanti Polar Lipids (Alabaster, AL). All other reagents were of analytical grade. The critical micellar concentration of IP was measured as an increase in 5 μM ANS fluorescence emission (excitation at 360 nm) in 20 mM HEPES buffer (pH 7.0) at room temperature (21 ( 1 °C). The IP cmc was estimated to be 5  10-2 M. Preparation of Large Unilamellar Vesicles (LUVs). The lipids were dissolved in chloroform/methanol (2:1 v/v) and mixed as required, and the solvent was evaporated exhaustively. Multilamellar vesicles (MLV) were prepared by hydrating the dry lipids in buffer with vortex shaking. Lipids were hydrated in 10 mM HEPES, 150 mM NaCl at pH 7.4 (Hepes buffer). Large unilamellar vesicles were prepared by the extrusion method (10 passages) with filters having a 0.1 μm pore diameter. Vesicle size was measured by quasi-elastic light scattering in a Malvern ZetaSizer 4 spectrometer. The average diameter of the vesicles was in all cases ca. 100 nm. (22) Edwards, K.; Almgren, M. Langmuir 1992, 8, 824–832. (23) Kaschny, P.; Go~ni, F. M. Eur. J. Biochem. 1992, 207, 1085–1091. (24) Pantaler, E.; Kamp, D.; Haest, W. M. Biochim. Biophys. Acta 2000, 1509, 397–408. (25) Akel, A.; Hermle, T.; Niemoeller, O. M.; Kempe, D. S.; Lang, P. A.; Attanasio, P.; Podolski, M.; Wieder, T.; Lang, F. Eur. J. Pharmacol. 2006, 532, 11– 17. (26) Seeman, P. Pharmacol. Rev. 1972, 24, 583–655. (27) Schreier, S.; Malheiros, S. V. P.; de Paula, E. Biochim. Biophys. Acta 2000, 1508, 210–234. (28) Ahyayauch, H.; Go~ni, F. M.; Bennouna, M. Int. J. Pharm. 2004, 279, 51– 58.

7308 DOI: 10.1021/la904194a

%release ¼

ðFf -Fo Þ  100 ðF100 -Fo Þ

where Ff, F100, and Fo are the fluorescence intensity values observed after the addition of surfactant, after the addition of excess Triton X-100, and before any addition, respectively. LUV Labeled in the Inner Monolayer with NBD-PE. PC LUV was prepared by using the method described above, including 0.6 mol % NBD-PE. To eliminate the outer leaflet label, we used 0.6 mM sodium dithionite, which reduced outer NBD and thus abolished its fluorescence. Reduction was monitored in an SLM Aminco spectrofluorometer at room temperature with a continuously stirred cuvette. Excitation and emission wavelengths were 465 and 530 nm, respectively. A cutoff filter (515 nm) was used to prevent contamination from scattered light. To separate LUV from excess sodium dithionite, a Sephadex G-75 chromatography column was used. After 30 min of incubation with surfactant, 50 μL of a 1 M dithionite solution was injected into the cuvette, resulting in a rapid decrease in fluorescence intensity. This fluorescence decline corresponds to the dithionite-mediated reduction of fluorophores localized in the outer membrane leaflet. From the fluorescence intensities before (F0) and after (FR) the addition of dithionite, the flip-flop was estimated according to this equation: F(%) = (1 - (FR/F0))  100. Attenuation of Perylene Fluorescence. The fluorescence attenuation of perylene (a hydrophobic fluorescent probe) by CPZ and IP was measured for small unilamelar vesicles (SUV) formed in the presence of perylene. The final concentrations of fluorescent probe and phospholipid were 0.25 and 150 μM, respectively. The surfactant was added to the SUV suspension (29) Go~ni, F. M.; Alonso, A. Biochim. Biophys. Acta 2000, 1508, 51–68. (30) Ellens, H.; Bentz, J.; Szoka, F. C. Biochemistry 1985, 24, 3099–3106.

Langmuir 2010, 26(10), 7307–7313

Ahyayauch et al.

Article

at different concentrations. The fluorescence intensity was measured using a Jobin Yvon spectrofluorimeter. The excitation and emission wavelengths were 413 and 473 nm, respectively. Isolation of RBC. Whole blood was obtained from healthy volunteers by venipuncture, using acid citrate/dextrose as an anticoagulant. In a typical experiment, RBC were centrifuged, the plasma was removed, and the RBC were washed five times in triple volumes of Hepes buffer. Cells were finally resuspended in the same buffer. Labeling of RBC with C6-NBD PS. Washed human RBC were resuspended in Hepes buffer to a cell concentration of 5  108 cells/mL and loaded with 2 μM C6-NBD-PS for 1 h at 37 °C. To measure outward movement, residual C6-NBD lipid remaining in the cells’ outer monolayer was removed by washing for 5 min with ice-cold 0.5% BSA prior to the experiment.

Translocation of RBC Membrane Lipids from the Inner to the Outer Leaflet. The outward movement of C6-NBD-PS was measured using the BSA back-exchange procedure as described by Connor et al.31 The labeled RBC were incubated with different surfactant concentrations. Briefly, 200 μL aliquots from the cell suspension were removed at the indicated time intervals and placed on ice for 5 min in the presence or absence of 1% BSA. Pellets obtained after 3 min of centrifugation at 12 000g were solubilized in 2 mL of 1% (w/v) Triton X-100, and the amount of externalized probe was determined by comparing the fluorescence intensity associated with the cells before and after back-exchange. The amount of probe extracted into BSA was related to the control sample incubated without surfactants, the value of which was established to be 100%. The fluorescence assay was measured at 37 °C using an SLM 8100 spectrofluorometer equipped with a circulating water bath. The excitation and emission wavelengths were 480 and 525 nm, respectively. Solubilization of RBC. The washed erythrocytes were incubated with different detergent concentrations. After incubation, the mixture was centrifuged and the supernatant was removed. Then, chloroform/methanol/HCl (66/33/1) was added to the supernatant to extract the solubilized phospholipids. Finally, the organic phase was evaporated under nitrogen, and the phospholipids were quantified with a phosphate assay. Hemolysis Assays. The hemolysis assay was performed in 1 mL test tubes by mixing the erythrocyte suspension (gave an A412 value of 0.6) with the required surfactant concentrations. The mixtures were incubated at 37 °C for 1 h with gentle shaking. After centrifugation at 1700g for 5 min, hemolytic activity was measured as an increase in the A412 value (i.e., increase in hemoglobin content) of the supernatant.

Results and Discussion When low concentrations of detergent are dispersed in an aqueous medium, a fraction of the molecules will orient themselves at the air-water interface and the rest will remain dispersed at random in the bulk water phase. With increasing detergent concentration, a point is reached, called the critical micelle concentration, beyond which the surfactant molecules in solution become organized in spherical structures, the micelles, so that multiple equilibrium is established among detergent molecules at the airwater interface, detergent monomers dispersed in the bulk water phase, and detergent molecules in part of the micelles.3,32 When, in addition to water and detergent, the system contains membrane vesicles in suspension, the multiple equilibrium just mentioned is made more complex, with the detergent molecules incorporated into the membrane bilayers or at surfactant concentrations causing solubilization, as also found in lipid-detergent mixed micelles.1 (31) Connor, J.; Bucana, C.; Fidler, I. J.; Schroit, A. J. Proc. Natl. Acad. Sci. U. S.A. 1991, 86, 3184–3188. (32) Tanford, C. The Hydrophobic Effect: Formation of Micelles and Biological Membranes; Wiley: New York, 1980.

Langmuir 2010, 26(10), 7307–7313

Figure 1. Destabilization of liposomal and erythrocyte membranes by IP. (A) Large unilamellar vesicles. (B) Erythrocytes. (b) NBD-phospholipid translocation from the inner to the outer membrane leaflet. (9) Membrane permeabilization, vesicle leakage (from liposomes), or hemolysis (from red blood cells). (O) Membrane solubilization, assayed as the change in turbidity (liposomes) or phospholipid solubilization (erythrocytes).

The detergent-induced phenomena described in this paper, consisting of transbilayer lipid motion or flip-flop, the breakdown of the permeability barrier, and vesicle lysis and reassembly, all occur at subsolubilizing detergent concentrations in so-called stage I of detergent-membrane interactions.1,33 This work deals with the properties of surfactant molecules from the point of view of their ability to (i) induce transmembrane motions in membranes, (ii) impair the membrane permeability barrier (i.e., induce leakage), (iii) cause the lysis and reassembly of vesicle bilayers, and (iv) solubilize membranes. First, we investigated changes in the flip-flop rate caused by the detergent molecules under study. The flop (transfer from the inner to the outer monolayer) of phospholipid probes NBD-PS and NBD-PE from the inner to the outer membrane leaflet of human erythrocytes and liposomes, respectively, is accelerated in the presence of all of the detergents investigated, and the phenomenon is detected at detergent concentrations below the cmc (Figures 1-5, filled circles, and Table 1). This is in agreement with the observations by Pantaler et al.24 The data as a function of IP concentration are shown in Figure 1. Both in erythrocytes and liposomes, flip-flop increased significantly with surfactant concentration. In both cases, the effect started to be seen at ca. 10-5 M IP, and a halfmaximal effect was seen at ca. 10-4 M IP. Note that in RBC the extent of flip-flop achieved at subsolubilizing IP concentrations was much lower than in liposomes. This could be attributed to a bilayer-stabilizing effect due to the presence of proteins, both membrane and skeletal. Solubilization of the surfactant-treated vesicle suspensions and erythrocytes is also shown in Figure 1A,B, respectively (open symbols). The results show that flip-flop (33) Lichtenberg, D.; Robson, M R. J.; Dennis, E. A. Biochim. Biophys. Acta 1983, 737, 285–304.

DOI: 10.1021/la904194a

7309

Article

Figure 2. Destabilization of liposomal and erythrocyte membranes by CPZ. (A) Large unilamellar vesicles. (B) Erythrocytes. (b) NBD-phospholipid translocation from the inner to the outer membrane leaflet. (9) Membrane permeabilization, vesicle leakage (from liposomes), or hemolysis (from red blood cells). (O) Membrane solubilization, assayed as a change in turbidity (liposomes) or phospholipid solubilization (erythrocytes).

Figure 3. Destabilization of liposomal and erythrocyte membranes by TX. (A) Large unilamellar vesicles. (B) Erythrocytes. (b) NBD-phospholipid translocation from the inner to the outer membrane leaflet. (9) Membrane permeabilization, vesicle leakage (from liposomes), or hemolysis (from red blood cells). (O) Membrane solubilization, assayed as a change in turbidity (liposomes) or phospholipid solubilization (erythrocytes). 7310 DOI: 10.1021/la904194a

Ahyayauch et al.

Figure 4. Destabilization of liposomal and erythrocyte membranes by SDS. (A) Large unilamellar vesicles. (B) Erythrocytes. (b) NBD-phospholipid translocation from the inner to the outer membrane leaflet. (9) Membrane permeabilization, vesicle leakage (from liposomes), or hemolysis (from red blood cells). (O) Membrane solubilization, assayed as a change in turbidity (liposomes) or phospholipid solubilization (erythrocytes).

occurred before any bilayer solubilization took place. CPZ produces, in the same way as IP, a concentration-dependent acceleration of the transbilayer movement of NBD-phospholipid (Figure 2 and Table 1). As with IP, transbilayer lipid motion becomes apparent at ca. 10-5M CPZ, in agreement with the results reported by Akel et al.25 For erythrocyte membranes, both IP and CPZ reach a half-maximal increase in transmembrane motion at 10-4 M (Table 1). Triton X-100 causes an increased flip-flop rate only at or above ∼10-4 M both in erythrocyte membranes and in liposomes (Figure 3). However, at variance with IP and CPZ, TX does not cause a large extent of transbilayer lipid motion (