Flip-Flop of Phospholipids in Vesicles - American Chemical Society

Apr 22, 2009 - We applied a time-resolved small-angle neutron scattering technique to vesicle systems to determine interparticle transfer and flip-flo...
0 downloads 0 Views 452KB Size
J. Phys. Chem. B 2009, 113, 6745–6748

6745

Flip-Flop of Phospholipids in Vesicles: Kinetic Analysis with Time-Resolved Small-Angle Neutron Scattering Minoru Nakano,*,† Masakazu Fukuda,† Takayuki Kudo,† Naoya Matsuzaki,† Takuto Azuma,† Kazuhisa Sekine,† Hitoshi Endo,‡ and Tetsurou Handa† Graduate School of Pharmaceutical Sciences, Kyoto UniVersity, Sakyo-ku, Kyoto 606-8501, Japan, and Institute for Solid State Physics, The UniVersity of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba, 277-8581, Japan ReceiVed: January 31, 2009; ReVised Manuscript ReceiVed: April 2, 2009

We applied a time-resolved small-angle neutron scattering technique to vesicle systems to determine interparticle transfer and flip-flop of phospholipids. Measurements were performed for large unilamellar vesicles, consisting of dimyristoylphosphatidylcholine (DMPC), 1-palmitoyl-2-oleoylphosphatidylcholine (POPC), or 1-palmitoyl2-oleoylphosphatidic acid (POPA), which differ either in their acyl chains or headgroup. POPC, which is analogous to naturally occurring phosphatidylcholines, exhibited no transbilayer transfer and very slow interbilayer migration. POPC on the inner leaflet of vesicles did not flop even when phospholipase D converted all POPC molecules on the outer leaflet into POPA, which was shown to exhibit fast flip-flop. From these results, together with the observation that the flip-flop of DMPC was entirely inhibited in the presence of cholesterol, it is deduced that the flip-flop of phosphatidylcholines does not take place spontaneously in cellular plasma membranes rich in cholesterol and that it requires enzymatic activities of energy-dependent and/or -independent flippases/floppases. Introduction Interbilayer transport and transbilayer movement of phospholipids are crucial for cell growth, development, and survival and are controlled by lipid transfer proteins and translocase enzymes.1-4 The endoplasmic reticulum (ER) in which phospholipids are newly synthesized on the cytosolic leaflet maintains membrane symmetry, presumably by flippase activity,5 which is bidirectional and energy-independent.6,7 On the other hand, the plasma membrane retains asymmetric lipid distribution via aminophospholipid translocase that mediates the unidirectional transport of phosphatidylserine and phosphatidylethanolamine from the ectoplasmic to cytoplasmic leaflet of the bilayer.8 Disruption of the asymmetry in cells by the action of phospholipid scramblase is involved in apoptosis and is associated with increased binding and phagocytosis of these cells by macrophages.9 Thus, understanding and control of these lipid dynamics quantitatively is a key challenge in biophysics and cell biology. Previously, we succeeded in determining lipid dynamics using small-angle neutron scattering (SANS).10 This technique takes advantage of the large difference in the scattering length density between hydrogenated and deuterated lipids, and the exchange of these lipids between large unilamellar vesicles (LUVs) results in a decrease in the scattering intensity, which can be detected in situ by time-resolved SANS (TR-SANS) measurements. Different from fluorescence spectroscopy11-14 and electron spin resonance,15,16 which are widely used for investigating lipid flipflops, this technique does not require fluorescence- or spin-labels into lipids that change physical and chemical properties.17 Here, we demonstrate with TR-SANS measurements that interbilayer and transbilayer transfers of phospholipids are * To whom corresponding should be addressed. Tel.: +81-75-753-4565. Fax: +81-75-753-4601. E-mail: [email protected]. † Kyoto University. ‡ The University of Tokyo.

strongly influenced by their chemical structure (acyl chain length and headgroup size) and the presence of cholesterol. These dynamic properties are discussed with relevance to symmetric/ asymmetric lipid distribution in biomembranes. Experimental Methods LUV Preparation. Dimyristoylphosphatidylcholine (DMPC), 1-palmitoyl-2-oleoylphosphatidylcholine (POPC), 1-palmitoyl2-oleoylphosphatidic acid (POPA), and their deuterated compounds, d54-DMPC, d31-POPC, d31-POPA (Avanti, Alabaster, AL), and cholesterol (Chol) (Sigma-Aldrich, St. Louis, MO) were used as received. The solvent used in this study was Trisbuffered saline (10 mM Tris, 150 mM NaCl, 1 mM EDTA, and 0.01% NaN3, pH 7.4) containing an appropriate volume of D2O (Sigma-Aldrich) and H2O. The volume fraction of D2O in the buffer was 50, 45, and 40% for DMPC/Chol LUVs with 0, 20, and 40 mol % Chol, respectively, and 33 and 30% for POPA and POPC LUVs, respectively. The lipids were hydrated in the buffer, repeatedly freeze-thawed, and extruded through a polycarbonate membrane with a pore size of 100 nm (Avestin, Ottawa, Canada) using LiposoFast (Avestin). Peptide-containing LUVs were prepared by mixing KALP23 (Ac-GKKL(AL)8KKANH2) (Hayashi Kasei, Osaka, Japan) with lipids before hydration.12 In addition to LUVs consisting of either deuterated lipids (D-LUV) or hydrogenated lipids (H-LUV), LUVs consisting of a 1:1 mixture of both lipids (D/H-LUV) were prepared by mixing these lipids before hydration. Concentration of phosphatidylcholine (DMPC and POPC) was determined using an enzymatic assay kit for choline (Wako, Osaka, Japan). Concentration of POPA was determined by the method of Bartlett.18 It was certified by dynamic light scattering measurements (Photal FPAR-1000; Otsuka Electronic Co., Osaka) that the LUV size was unchanged during the SANS experiment. SANS. The phospholipid concentration of each LUV preparation was set to 20 mM for DMPC and 30 mM for POPC and POPA. SANS measurements were performed by SANS-U of

10.1021/jp900913w CCC: $40.75  2009 American Chemical Society Published on Web 04/22/2009

6746

J. Phys. Chem. B, Vol. 113, No. 19, 2009

Nakano et al.

Institute for Solid State Physics, the University of Tokyo, at the research reactor JRR-3, Tokai, Japan.19 The wavelength (λ) of the neutron source was 7 Å (∆λ/λ ) 10%), and the sampleto-detector distance was set to 4 m. Samples were measured in quartz cells (Nippon Silica Glass, Tokyo) with a pass length of 2 mm. Each measurement was started immediately after mixing an equivalent volume of D-LUV and H-LUV. Scattering data for 3-8 min (depending on the intensity) were accumulated in time-resolved measurements to calculate the count rate (total counts/second). The count rate for the solvent was then subtracted. The normalized contrast, ∆F(t)/∆F(0), was calculated from the following equation

√I(t) - √I(∞) ∆F(t) ) ∆F(0) √I(0) - √I(∞)

(1)

Figure 1. Contrast decays after mixing D- and H-LUV of DMPC/ Chol mixtures (20 mM for DMPC) with different Chol concentrations at 37.0 °C. Solid lines are fitting curves according to eq 2.

where I(t) is the count rate at time t after LUV mixing. I(∞) is the count rate from D/H-LUV, and I(0) is the average count rate for D- and H-LUV. The decay curves were fitted to determine the rate constants of the exchange (kex) and flip-flop (kf) with the following equation, which is a modified form of the equation reported10

[( ) ( ( ) (

)

kex + 2kf + X ∆F(t) 1 kf ) (1 - C) - exp t + ∆F(0) 2 X 2 kex + 2kf - X 1 kf + exp t +C 2 X 2

)]

(2)

2 1/2 where, X ) (4kf2 + kex ) . The constant C, which denotes a nonexchangeable fraction, was involved in this equation to take into account the presence of multilamellar vesicles (see Supporting Information for SANS data analysis). Phospholipase D Assay. Hydrolysis of POPC that produces POPA and choline was monitored by detecting the choline production by an enzymatic assay. Phospholipase D (PLD) from Streptomyces sp. (Sigma-Aldrich) was mixed with LUVs at final concentrations of 1.25 units/mL PLD and 200 µM POPC in Tris-buffered saline (10 mM Tris, 150 mM NaCl, pH 7.4) at 37 °C in the absence and presence of 1.0 vol % heptaethylene glycol monododecyl ether (HED). At several intervals, 10 µL was withdrawn from the reaction mixture and added to 190 µL of fluorogenic solution containing 0.4 units choline oxidase from Alcaligenes sp. (Wako), 0.4 units horseradish perxoidase (Sigma-Aldrich), and 30 nmol Amplex red (Invitrogen, Eugene, OR) in Tris-buffered saline. After the mixture was kept for 20 min at 37 °C, its fluorescence intensity was measured with a HITACHI F-4500 Fluorescence Spectrophotometer with excitation and emission wavelengths of 566 and 590 nm, respectively. The choline concentration was determined from a calibration curve with choline chloride solution.

Results Previously, we succeeded in detecting the intervesicular exchange and flip-flop of DMPC by TR-SANS.10 In the present study, we first investigated the effect of Chol on the dynamics of DMPC, since Chol, which has a rigid steroid backbone, is known to increase membrane packing and reduce fluidity. As shown in Figure 1, contrast decay curves obtained at 37 °C were affected by Chol. The intervesicular exchange of DMPC was almost independent of the Chol content (the half-lives (t1/2 ) (ln 2)/k) of 130, 170, and 170 min for 0, 20, and 40 mol % Chol, respectively), suggesting that Chol barely affects the transition state of the lipid exchange, where DMPC molecules start falling out of the bilayer and expose their acyl chains to water. On the other hand, flip-flop was markedly hampered (t1/2

Figure 2. Contrast decays after mixing D- and H-LUV of POPC (30 mM) containing 0 (circles) and 0.5 mol % KALP23 (dots) at 37.0 °C in the absence and presence of MβCD. Solid lines are fitting curves of the data in the absence of MβCD according to eq 2.

of 350, 1300, and >7000 min in the presence of 0, 20, and 40 mol % Chol, respectively). The flip-flop almost disappeared for bilayers with 40 mol % Chol, which are known to be in a liquidordered phase.20 The data suggest that phospholipids can barely rotate longitudinally in such rigid membranes. We next investigated the dynamics of POPC whose acyl chains are longer than those of DMPC and are the most abundant constituents (C16 and C18:1) in naturally occurring phospholipids. The contrast decay curve of POPC represented very slow lipid exchange (t1/2 ∼ 90 h) in the absence of methyl-βcyclodextrin (MβCD), as shown in Figure 2. MβCD, which is known to catalyze the interparticle transfer of phospholipids and Chol by forming clathrate compounds,21 facilitated the POPC exchange in a concentration-dependent manner. Surprisingly, the decay curves reached a constant value of ca. 0.55 and the value was kept unchanged for at least 2600 min (not shown), which suggests the absence of flip-flop (t1/2 > 1000 h). The disagreement of the constant value with the ideal value (0.50) was due to the presence of multilamellar species in the LUV preparations. POPA has identical acyl chains but a smaller headgroup than POPC. From these structural differences, POPA is expected to show slower interparticle transfer and faster flip-flop. The dynamics of this lipid were investigated using LUVs prepared at pH 7.4, as shown in Figure 3. The contrast did not decrease after mixing D- and H-LUVs, suggesting no exchange of POPA. Even in the presence of MβCD, the exchange rate of POPA was about 4-fold slower than that of POPC at the same MβCD concentration. On the contrary, POPA was found to flip-flop from the observation that the decay curves were well below 0.5 when the intervesicular transfer of POPA was induced by

Flip-Flop Detection by TR-SANS

Figure 3. Contrast decays after mixing D- and H-LUV of POPA (30 mM) containing 0 (circles) and 0.5 mol % KALP23 (dots) at 37.0 °C in the absence and presence of MβCD. Solid lines are fitting curves of the data in the absence of MβCD according to eq 2.

J. Phys. Chem. B, Vol. 113, No. 19, 2009 6747

Figure 5. Schematic representation of the constraints during phospholipid flip-flop. (a) DMPC can traverse across the bilayer, which brings about bilayer disturbances within a permissible range. (b) POPC, with longer acyl chains compared with DMPC, does not show flipflop since in this process POPC must bring its polar headgroup through a longer hydrophobic passage in a thicker membrane, and/or invert its own longer acyl chains to the other side of the bilayer, which disturbs bilayer integrity.

though POPA was shown to exhibit flip-flop in the time scale of hours (Figure 3), it did not promote the flip-flop of POPC at all. Discussion

Figure 4. Time course of hydrolysis of POPC LUV (200 µM) by PLD at 37 °C in the absence and presence of 1.0 vol% HED.

MβCD. MβCD did not affect the flip-flop rate at lower concentrations (t1/2 of 450 and 430 min at 0.5 and 1.0 mM MβCD, respectively), but 1.3-fold accelerated it at 5.0 mM (t1/2 of 330 min) at which level the intervesicular transfer was fast enough to observe an inflection point in the decay curve. It has been reported that the transmembrane helix of hydrophobic peptides incorporated into LUVs facilitates the flipflop of fluorescence-labeled phospholipids, including phosphatidic acid.12 We therefore examined if a transmembrane peptide, KALP23 (Ac-GKKL(AL)8KKA-NH2),12 could also facilitate the flip-flop of POPC and POPA. The results are shown in Figure 2 and 3. Overlap of the decay curves of POPC in the absence and presence of 0.5 mol % KALP23 (Figure 2) clearly indicates that KALP23 affected neither the intervesicular transfer nor the flip-flop of POPC. In fact, the upward shift of the decay curves of POPA in the presence of 0.5 mol % KALP23 (Figure 3) suggests that the flip-flop of POPA was inhibited by the peptide, presumably due to the electrostatic interaction between positive residues of the peptide and phosphate group. Taken together, it is concluded that the mere presence of transmembrane helices has no effect on the flip-flop of naturally occurring lipids. Phospholipase D (PLD) catalyzes hydrolysis of phosphatidylcholine to generate phosphatidic acid and choline. PLD is considered to hydrolyze phosphatidylcholine molecules on the outer leaflet when it acts against LUVs. We monitored the PLD reaction against 200 µM POPC LUV in the absence and presence of a surfactant (HED), as shown in Figure 4. In the absence of the surfactant, the reaction proceeded until the half of the total lipids were converted; thereafter, no further hydrolysis was observed. This result suggests that PLD transforms POPC LUVs to asymmetric POPC/POPA LUVs. Al-

With recent progress in research dealing with asymmetric bilayers,22,23 it has become more desirable to know how frequently lipids flip in membranes. Previously, we demonstrated that the kinetics of intervesicular and transbilayer lipid migrations could be detected in situ by TR-SANS,10 a technique first applied to detect the unimer exchange in polymer micelles.24-26 This method involves the simple process of mixing deuterated and hydrogenated compounds, and enables us to determine both exchange and flip-flop rates unambiguously.10 Especially, the benefit of TR-SANS is that flip-flop of lipids before sample mixing does not influence the determination of its rate, which is strictly distinct from methods using asymmetrically labeled bilayers, where measurements must be performed immediately before the asymmetry disappears.12,17 Similarly to the method using radioisotopes27,28 and fluorophores,29 the determination of the flip-flop rate by TR-SANS is limited to systems that involve faster intervesicular exchange than the flip-flop. However, we demonstrated that the exchange rate could be controlled by using MβCD with little disturbance of the flip-flop. This could be also valid for Chol-containing vesicles, since MβCD have been reported to accelerate the phospholipid transfer by considerably larger factors than they accelerate the transfer of cholesterol.21 Therefore, TR-SANS is well suited to determine the flip-flop of phospholipids. If a certain protein that possesses lipid translocase activity can be incorporated into LUVs, the proteinmediated flip-flop can be detected by this method. In this case, since lipid-to-protein ratio is controllable, this method potentially quantifies how frequently one enzyme moves one lipid. Flip-flop is a process in which the polar headgroup of a lipid passes through a hydrophobic bilayer. In contrast with DMPC, which showed a relatively fast flip-flop, POPC failed to flipflop despite the fact that the headgroup of both lipids is the same. This presumably comes from constraints that POPC, as compared with DMPC, must bring its polar headgroup through a longer hydrophobic passage in a thicker membrane, and/or invert its own longer acyl chains to the other side of the bilayer, which disturbs bilayer integrity (Figure 5). Cellular phospholipids that are synthesized in the cytosolic leaflet of the ER must flip into the luminal side to keep the balance of mass in each leaflet. Thus, the presence of flippase in the biogenic membrane has been postulated.5 Kol et al.

6748

J. Phys. Chem. B, Vol. 113, No. 19, 2009

showed that transmembrane peptides induce the flop of 2-[6[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]caproyl] (C6NBD) phospholipids (phosphatidylglycerol, phosphatidylethanolamine, and phosphatidic acid) in dioleoylphosphatidylcholine (DOPC) vesicles,12 which suggests that the mere presence of membranespanning helices of proteins may account for the fast translocation of phospholipids in the ER. However, in the present study, KALP23 neither induced POPC’s flip-flop, nor accelerated POPA’s. This contradiction could be ascribed to the use of DOPC as a platform in the study of Kol et al., since an introduction of unsaturated acyl chains increases the membrane’s fluidity, compared with POPC and POPA. The use of labeled lipids could also affect the data, that is, KALP23 might be able to induce the flip-flop of NBD-labeled lipids but not natural lipids. Our result with more naturally relevant systems supports an idea that a specific activity of the putative flippases or a specific transmembrane peptide sequences is indeed required for the flip-flop of phospholipids in the biogenic membrane.30 POPA exhibited flip-flop with t1/2 of ca. 7 h at pH 7.4 and 37 °C. Since POPC LUVs were hydrolyzed in the same conditions in the PLD assay, POPA molecules produced at the outer leaflet were expected to flip into the inner leaflet. We anticipated that the flip of POPA would force POPC to turn out of the interior, thereby hydrolyzing all the POPC molecules. However, the reaction was arrested after half the total lipids were hydrolyzed, suggesting the high persistency of POPC against flop, which resulted in the formation of POPC/POPA asymmetric vesicles. Plasma membranes are known to contain higher amounts of Chol (30-40%) than the ER (ca. 5%). Flip-flop of DMPC, which was much faster than that of the naturally occurring phosphatidylcholines, was entirely inhibited by 40% Chol. Therefore, it is deduced that phospholipids in the cellular plasma membranes are secured against transbilayer migration unless assisted by enzymatic activities of energy-dependent and/or -independent flippases/floppases. In conclusion, we succeeded in monitoring the flip-flop of phospholipids in model membranes using TR-SANS with the aid of MβCD. With this assay, flip-flop with half-lives of 0.5-50 h was measurable. TR-SANS detected the flip-flop of DMPC and POPA but not POPC. In addition, this method revealed that Chol inhibits the flip-flop, while the transmembrane peptide, KALP23, has no stimulation effect. Provided that proteoliposomes with an enzyme that mediates the flip-flop are available,31 this technique will give quantitative data for its activity. Acknowledgment. This work was carried out by joint research of the Institute for Solid State Physics, the University of Tokyo (Proposal No. 7615 and 7616). This study was supported by Grants-in-aid for Scientific Research from the Japanese Ministry of Education, Culture, Sports, Science, and Technology (Nos. 17390011, 20050017, and 20790032) and the

Nakano et al. program for the Promotion of Fundamental Studies in Health Science of the National Institute of Biomedical Innovation. Supporting Information Available: The detail of the SANS data analysis, including determination of the vesicle lamellarity from SANS decay data collected in the presence of MβCD or 31 P NMR. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Graham, T. R. Trends Cell Biol. 2004, 14, 670. (2) Holthuis, J. C. M.; Levine, T. P. Nat. ReV. Mol. Cell Biol. 2005, 6, 209. (3) Kol, M. A.; de Kroon, A. I. P. M.; Killian, J. A.; de Kruijff, B. Biochemistry 2004, 43, 2673. (4) Pomorski, T.; Holthuis, J. C. M.; Herrmann, A.; van Meer, G. J. Cell Sci. 2004, 117, 805. (5) Bretscher, M. S. Science 1973, 181, 622. (6) Bishop, W. R.; Bell, R. M. Cell 1985, 42, 51. (7) Buton, X.; Morrot, G.; Fellmann, P.; Seigneuret, M. J. Biol. Chem. 1996, 271, 6651. (8) Seigneuret, M.; Devaux, P. F. Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751. (9) Mcevoy, L.; Williamson, P.; Schlegel, R. A. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 3311. (10) Nakano, M.; Fukuda, M.; Kudo, T.; Endo, H.; Handa, T. Phys. ReV. Lett. 2007, 98, 238101. (11) Hrafnsdottir, S.; Nichols, J. W.; Menon, A. K. Biochemistry 1997, 36, 4969. (12) Kol, M. A.; van Laak, A. N. C.; Rijkers, D. T. S.; Killian, J. A.; de Kroon, A. I. P. M.; de Kruijff, B. Biochemistry 2003, 42, 231. (13) Kubelt, J.; Menon, A. K.; Muller, P.; Herrmann, A. Biochemistry 2002, 41, 5605. (14) Marx, U.; Lassmann, G.; Holzhutter, H. G.; Wustner, D.; Muller, P.; Hohlig, A.; Kubelt, J.; Herrmann, A. Biophys. J. 2000, 78, 2628. (15) Kornberg, R. D.; Mcconnel, H. M. Biochemistry 1971, 10, 1111. (16) Mcnamee, M. G.; Mcconnel, H. M. Biochemistry 1973, 12, 2951. (17) Liu, J.; Conboy, J. C. Biophys. J. 2005, 89, 2522. (18) Bartlett, G. R. J. Biol. Chem. 1959, 234, 466. (19) Okabe, S.; Nagao, M.; Karino, T.; Watanabe, S.; Adachi, T.; Shimizu, H.; Shibayama, M. J. Appl. Crystallogr. 2005, 38, 1035. (20) Almeida, P. F. F.; Vaz, W. L. C.; Thompson, T. E. Biochemistry 1992, 31, 6739. (21) Leventis, R.; Silvius, J. R. Biophys. J. 2001, 81, 2257. (22) Collins, M. D.; Keller, S. L. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 124. (23) Kiessling, V.; Crane, J. M.; Tamm, L. K. Biophys. J. 2006, 91, 3313. (24) Lund, R.; Willner, L.; Richter, D.; Dormidontova, E. E. Macromolecules 2006, 39, 4566. (25) Lund, R.; Willner, L.; Stellbrink, J.; Lindner, P.; Richter, D. Phys. ReV. Lett. 2006, 96, 068302. (26) Willner, L.; Poppe, A.; Allgaier, J.; Monkenbusch, M.; Richter, D. Europhys. Lett. 2001, 55, 667. (27) Wimley, W. C.; Thompson, T. E. Biochemistry 1990, 29, 1296. (28) Wimley, W. C.; Thompson, T. E. Biochemistry 1991, 30, 1702. (29) Bai, J. N.; Pagano, R. E. Biochemistry 1997, 36, 8840. (30) Pomorski, T.; Menon, A. K. Cell. Mol. Life Sci. 2006, 63, 2908. (31) Vehring, S.; Pakkiri, L.; Schroer, A.; Alder-Baerens, N.; Herrmann, A.; Menon, A. K.; Pomorski, T. Eukaryotic Cell 2007, 6, 1625.

JP900913W