Shape Changes and Vesicle Fission of Giant ... - ACS Publications

In this report, we investigated the interaction of lysophosphatidylcholine (lyso-PC) ..... and SM/chol GUVs in the lo Phase Induced by Lysophosphatidy...
0 downloads 0 Views 401KB Size
Download PDF
9526

Langmuir 2004, 20, 9526-9534

Shape Changes and Vesicle Fission of Giant Unilamellar Vesicles of Liquid-Ordered Phase Membrane Induced by Lysophosphatidylcholine Tomoki Tanaka,†,§ Ryoko Sano,‡,§ Yuko Yamashita,‡ and Masahito Yamazaki*,†,‡ Materials Science, Graduate School of Science and Engineering, Shizuoka University, Shizuoka, 422-8529, Japan, and Department of Physics, Faculty of Science, Shizuoka University, Shizuoka, 422-8529, Japan Received February 28, 2004. In Final Form: August 12, 2004 Liquid-ordered phase (lo phase) of lipid membranes has properties that are intermediate between those of liquid-crystalline phase and those of gel phase and has attracted much attention in both biological and biophysical aspects. Rafts in the lo phase in biomembranes play important roles in cell function of mammalian cells such as signal transduction. In this report, we have prepared giant unilamellar vesicles (GUVs) of lipid membranes in the lo phase and investigated their physical properties using phase-contrast microscopy and fluorescence microscopy. GUVs of dipalmitoyl-phosphatidylcholine (DPPC)/cholesterol membranes and also GUVs of sphingomyelin (SM)/cholesterol membranes in the lo phase in water were formed at 20-37 °C successfully, when these membranes contained g30 mol % cholesterol. The diameters of GUVs of DPPC/cholesterol and SM/cholesterol membranes did not change from 50 to 28 °C, supporting that the membranes of these GUVs were in the lo phase. To elucidate the interaction of a substance with a long hydrocarbon chain with the lo phase membrane, we investigated the interaction of low concentrations (less than critical micelle concentration) of lysophosphatidylcholine (lyso-PC) with DPPC/cholesterol GUVs and SM/cholesterol GUVs in the lo phase. We found that lyso-PC induced several shape changes and vesicle fission of these GUVs above their threshold concentrations in water. The analysis of these shape changes indicates that lyso-PC can be partitioned into the external monolayer in the lo phase of the GUV from the aqueous solution. Threshold concentrations of lyso-PC in water to induce the shape changes and vesicle fission increased greatly with a decrease in chain length of lyso-PC. Thermodynamic analysis of this result indicates that shape changes and vesicle fission occur at threshold concentrations of lyso-PC in the membrane. These new findings on GUVs of the lo phase membranes indicate that substances with a long hydrocarbon chain such as lyso-PC can enter into the lo phase membrane and also the raft in the cell membrane. We have also proposed a mechanism for the lyso-PC-induced vesicle fission of GUVs.

1. Introduction Recently, the liquid-ordered phase (lo phase) of lipid membranes has attracted much attention in both biological and biophysical aspects. Binary mixture membranes of cholesterol and saturated phosphatidylcholines (PCs) such as dipalmitoylphosphatidylcholine (DPPC) are in the lo phase, which has properties that are intermediate between those of the liquid-crystalline (LR) phase and those of the gel phase.1-5 In the lo phase, the acyl chains of PC have high orientational order, but the lateral diffusion coefficient of lipids in the membrane is relatively high.3 Membranes of sphingomyelin (SM) and cholesterol mixture are also considered to form the lo phase.6 In cells, rafts or microdomains in plasma membranes are consid* To whom correspondence should be addressed. Mail: Department of Physics, Faculty of Science, Shizuoka University, 836 Oya, Shizuoka 422-8529, Japan. Tel and Fax: 81-54-238-4741. E-mail: [email protected]. † Materials Science. ‡ Department of Physics. § These authors contributed equally. (1) Ipsen, J. H.; Karlstro¨m, G.; Mouritsen, O. G.; Wennersto¨m, H.; Zuckermann, M. J. Biochim. Biophys. Acta. 1987, 905, 162. (2) Vist, M. R.; Davis, J. H. Biochemistry 1990, 29, 451. (3) Bloom, M.; Evans, E.; Mouritsen, O. G. Q. Rev. Biophys. 1991, 24, 293. Bloom, M.; Mouritsen, O. G. In Structure and dynamics of membranes; Lipowsky, R., Sackmann, E., Eds.; Elsevier/North-Holland: Amsterdam, 1995; p 65. (4) Almeida, P. F. F.; Vaz, W. L. C.; Thompson, T. E. Biochemistry 1992, 31, 6739. (5) Mateo, C. R.; Acun˜a, A. U.; Brochon, J.-C. Biophys. J. 1995, 68, 978. (6) Sankaram, M. B.; Thompson, T. E. Biochemistry 1990, 29, 10670.

ered to play important roles in signal transduction and cell migration.7-10 The rafts are composed of sphingolipids, cholesterol, and several proteins such as GPI-anchored proteins, which are isolated from cell membranes based on the resistance to solubilization by the nonionic detergent Triton X-100 at 4 °C.11-13 The rafts are considered to be in the lo phase.12,14 To understand the function and physical properties of the lo phase membrane and the rafts in more detail, we need more experimental analysis using various kinds of experiments. On the other hand, giant unilamellar vesicles (GUVs) of phospholipids with more than 10 µm diameter have been used for physical and biological investigations such as elastic properties of the phospholipid membranes,15-17 shape change of vesicles,18-23 interaction of cytoskeleton (7) Simon, K.; Ikonen, E. Nature 1997, 387, 569. (8) Simons, K.; Ikonen, E. Science 2000, 290, 1721. (9) Go´mez-Mo´uton, C.; Abad, J. L.; Mira, E.; Lacalle, R. A.; Gallardo, E.; Jime´nez-Baranda, S.; Illa, I.; Bernad, A.; Man˜es, S.; Martinez-A, C. Proc. Natl. Acad. Sci. U.S.A. 2001. 98, 9642. (10) Bodin, S.; Giuriato, S.; Ragab, J.; Humbel, B. M.; Viala, C.; Vieu, C.; Chap, H.; Payrastre, B. Biochemistry 2001, 40, 15290. (11) Brown, D. A.; Rose, J. K. Cell 1992, 68, 533. (12) Schroeder, R.; London, E.; Brown, D. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 12130. (13) Brown, D. A. J. Cell Sci. 1998, 111, 1. (14) Ahmed, S. N.; Brown, D. A.; London, E. Biochemistry 1997, 36, 10944. (15) Kwok, R.; Evans, E. Biophys. J. 1981, 35, 637. (16) Sackmann, E. In Structure and dynamics of membranes; Lipowsky, R., Sackmann, E., Eds.; Elsevier Science B. V.: Amsterdam, The Netherlands, 1995; p 213. (17) Sandre, O.; Moreaux, L.; Brochard-Wyart, F. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 10591.

10.1021/la049481g CCC: $27.50 © 2004 American Chemical Society Published on Web 09/23/2004

Shape Changes and Fission of GUVs

proteins with the membranes,24 reconstitution of artificial cells,25-27 domain formation,28 and membrane fusion.29 These studies have been considered helpful to understand the dynamics of biological membranes such as various structures of membranes in the cell, endocytosis, and membrane fusions. So far, it has been difficult to prepare GUVs of PC in aqueous solution containing high salt concentration, which hindered their more extensive use. However, recently, we have developed a new method for the preparation of GUVs in aqueous solution containing high salt concentration (up to 2 M), which will expand the possibilities for GUVs.26 Several studies of shape changes of GUVs show that “the GUV method”, that is, the observation of shape changes of GUVs induced by the interaction of a substance with the membrane, can detect change of the area of the membrane with high sensitivity, and thereby, we can get information on the interaction.20-23,27 Interaction of substances with the lo phase membrane is not well understood, compared with their interaction with the LR phase membrane. In this report, we investigated the interaction of lysophosphatidylcholine (lysoPC) with DPPC/cholesterol membranes and SM/cholesterol membranes in the lo phase using the GUV method. One of its purposes is to get information on whether substances with a long hydrocarbon chain can enter into membranes in the lo phase and also into the rafts in the cell membrane from aqueous solution. In the interaction of lyso-PC with the LR phase membrane, Devaux and his colleague reported that the addition of a high concentration (5 mM; much higher than the critical micelle concentration (cmc)) of lyso-PC from the outside of GUVs of egg-PC membrane induced their shape changes and concluded that the lyso-PC can be partitioned into the external monolayer in the LR phase of the GUV from the aqueous solution.20,21 However, in that case, we also have to consider the interaction between micelles of lyso-PC and GUVs. In this report, to eliminate effects of the micelles on the GUVs, we investigated the effect of low concentrations (much less than the cmc) of several kinds of lyso-PC. We found that lyso-PC induced various shape changes of GUVs of the lo phase membranes and also vesicle fission of these GUVs above threshold concentrations. 2. Materials and Methods 2.1. Materials and Sample Preparation. 1,2-Dipalmitoylsn-glycero-3-phosphocholine (DPPC), 1,2-dioleoyl-sn-glycero-3phosphocholine (DOPC), SM from brain, 1,2-dipalmitoyl-snglycero-3-phosphoethanolamine-N-[poly(ethylene glycol)2000] (PEG2K-DPPE), 1-palmitoyl-2-hydroxy-sn-glycero-3-phosphocholine (lyso-PC(16:0)), 1-myristoyl-2-hydroxy-sn-glycero-3-phos(18) Lipowski, R. Nature 1991, 349, 475. (19) Ka¨s, J.; Sackmann, E. Biophys. J. 1991, 60, 825. (20) Farge, E.; Devaux, P. F. Biophys. J. 1992, 61, 347. (21) Mathivet, L.; Cribier, S.; Devaux, P. F. Biophys. J. 1996, 70, 1112. (22) Tanaka, T.; Tamba, Y.; Masum, S. M.; Yamashita, Y.; Yamazaki, M. Biochim. Biophys. Acta 2002, 1564, 173. (23) Yamashita, Y.; Masum, S. M.; Tanaka, T.; Yamazaki, M. Langmuir 2002, 18, 9638. (24) Saitoh, A.; Takiguchi, K.; Tanaka, Y.; Hotani, H. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 1026. (25) Tsumoto, K.; Nomura, S. M.; Nakatani, Y.; Yoshikawa, K. Langmuir 2001, 17, 7225. (26) Yamashita, Y.; Oka, M.; Tanaka, T.; Yamazaki, M. Biochim. Biophys. Acta 2002, 1561, 134. (27) Yamazaki, M.; Tanaka, T.; Yamashita, Y. In Science, Technology and Education of Microscopy: An Overview; Mendez-Vilas, A., Ed.; FORMATEX: Madrid, Spain, 2003; p 606. (28) Baumgart, T.; Hess, S. T.; Webb, W. W. Nature 2003, 425, 821. (29) Tanaka, T.; Yamazaki, M. Langmuir 2004, 20, 5160.

Langmuir, Vol. 20, No. 22, 2004 9527 phocholine (lyso-PC(14:0)), 1-lauroyl-2-hydroxy-sn-glycero-3phosphocholine (lyso-PC(12:0)), and 1-decanoyl-2-hydroxy-snglycero-3-phosphocholine (lyso-PC(10:0)) were purchased from Avanti Polar Lipids Inc. (Alabaster, AL). 2-(12-(7-Nitrobenz-2oxa-1,3-diazol-4-yl) amino) dodecanoyl-1-hexadecanoyl-sn-glycero-3-phosphocholine (NBD-PC), β-BODIPY-530/550-C12-HPE (BODIPY-PE), and cholesteryl-BODIPY-FL-C12 (BODIPY-chol) were purchased from Molecular Probes Inc. (Eugene, OR). Cholesterol was purchased from Wako Chemical Co. (Tokyo, Japan). 2.2. Formation of GUVs and Observation of GUVs Using a Microscope. Phospholipid GUVs were prepared in water by natural swelling of dry lipid film at 20-37 °C as follows. Phospholipid mixtures (100 µL, 1 mM) in chloroform in a small glass bottle (5 mL) were dried by N2 gas to produce a thin, homogeneous lipid film, and then the solvent was completely removed by placing the bottle containing the dry lipid film in a vacuum desiccator connected to a rotary vacuum pump for more than 12 h. Water (10 µL) was added into this glass bottle, and it was incubated at 45 °C for 10 min (prehydration). Then, 1 mL of 0.1 M sucrose in water was added into the glass bottle, and it was placed in an incubator which the temperature was controlled at various temperatures (from 20 to 37 °C) for 2 h. GUV solution (20 µL; 0.1 M sucrose solution; internal solution) was diluted into 300 µL of 0.1 M glucose aqueous solution (external solution) and then transferred into a handmade microchamber. This chamber (1 cm × 1 cm wide and 3 mm high, internal volume of ∼0.3 mL) was formed on a slide glass by inserting a U-shaped silicone-rubber spacer between a cover glass and the slide glass. We observed GUVs using an inverted phase-contrast microscope (IX-70, Olympus, Tokyo, Japan) at 20 ( 2 °C. Phase-contrast images of GUVs were recorded through a charge-coupled-device (CCD) camera (DXC-108, SONY, Tokyo, Japan) on a video recorder. As the GUVs, we selected the vesicles of the membrane for which the contrast was very low and also for which the undulation motion was large. When we observed these vesicles containing a small percentage (0.5 mol %) of long-chain fluorescent phospholipid, NBD-PC, using the fluorescence microscope (IX-70, Olympus, Tokyo, Japan), the intensities of fluorescence from these vesicles were much lower than those from multilamellar vesicles (MLVs). A disadvantage of NBD-PC is rapid photobleaching during observation by fluorescence microscopy. To overcome it, we used BODIPY-PE, which is much stronger against photobleaching than NBD-PC. However, when we observed GUVs containing BODIPY-PE, a rapid shape change occurred, which may be due to a photochemical reaction of BODIPY-PE during the exposure of light. Thereby, instead of BODIPY-PE, we used BODIPY-chol which did not induce a shape change during the observation. Fluorescence images of GUVs containing BODIPY-chol were recorded on a video recorder using an EB-CCD camera (C7190-23, Hamamatsu Photonics, Hamamatsu, Japan), which is a very highly sensitive fluorescence camera. Fluorescence intensities of GUVs containing 0.02-0.05 mol % BODIPY-chol (or BODIPY-PE) and 0.05 mol % NBD-PC were sufficiently strong to observe them using the EB-CCD camera. In this case also, the fluorescence intensities of the GUVs were much lower than those from oligo-lamellar vesicles. It was impossible to observe MLVs and oligo-lamellar vesicles containing several membranes using the EB-CD camera because the fluorescence intensity of the vesicles was too large. 2.3. Shape Changes of GUVs by Addition of Lyso-PC. Various concentrations of lyso-PC in 0.1 M glucose solution were added slowly into the vicinity of a GUV through a 10-20 µm diameter glass micropipet, the position of which was controlled by a micromanipulator (MMW-23, Narishige, Tokyo, Japan).22,23 A glass micropipet was prepared as follows: First we pulled a glass tube with 1.0 mm diameter (G-1, Narishige, Japan) to a needle point using a puller (PP-83, Narishige, Japan) and then broke it by quick fracture at the desired tip diameter. The micropipet was filled with the external solution of GUVs containing various concentrations of lyso-PC using aspiration by a vacuum pump (DA-5D, ULVAC KIKO, Japan), and then it was held by the micromanipulator, enabling control of the position of the tip of the micropipet. The injection pressure was controlled by changing the height of a vertical column of water in the

9528

Langmuir, Vol. 20, No. 22, 2004

Tanaka et al.

U-shaped glass tube to which the micropipet was hydraulically connected.22,23 2.4. Observation of GUVs by the Phase-Contrast Microscope at Various Temperatures. To investigate the effect of temperature on GUVs, GUVs of DPPC, PEG2K-DPPE/DPPC, DPPC/cholesterol, PEG2K-DPPE/DPPC/cholesterol, SM, and SM/ cholesterol membranes in water were prepared at 50 °C as follows. The appropriate amounts of various lipids in chloroform were mixed to prepare various stock lipid (or lipid mixture) solutions (1 mM) in chloroform. The stock lipid solution (100 µL, 1 mM) in chloroform in a small glass tube (1 mL) was dried by N2 gas, and then the solvent was completely removed by placing the sample in a vacuum desiccator connected to a rotary vacuum pump for more than 12 h. Water (10 µL) was added into this glass tube, and it was incubated at 50 °C for 5 min (prehydration). Then, 200 µL of 0.1 M sucrose in water was added and incubated at 50 °C for 2 h. GUV solution (20 µL; 0.1 M sucrose solution; internal solution) was diluted into 300 µL of 0.1 M glucose aqueous solution (external solution) at 50 °C and then transferred into the handmade microchamber kept at 50 °C. We controlled the temperature of the solution containing GUVs in the handmade chamber installed in the inverted phase-contrast microscope using the stage-warmer (MP100DM, Kitazato, Tokyo, Japan). We decreased the temperature from 50 to 28 °C at a rate of 1 °C/min and measured the diameters of a few directions of a GUV and obtained their average value.

3. Results and Discussion 3.1. Temperature Dependence of the Diameter of DPPC/chol GUVs and That of DPPC GUVs. We tried to produce GUVs of DPPC/cholesterol membranes containing 0-45 mol % cholesterol in water at 20-37 °C, which is lower than the chain melting phase transition temperature (Tm) of the DPPC membrane (Tm ) 41.4 °C). We could not produce any GUVs of 100% DPPC membrane and 90 mol % DPPC/10 mol % cholesterol membrane. On the other hand, we could produce many GUVs of 60 mol % DPPC/40 mol % cholesterol (DPPC/40%chol) membrane and of 70 mol % DPPC/30 mol % cholesterol (DPPC/ 30%chol) membrane in water. For DPPC/cholesterol membranes containing medium concentrations of cholesterol (around 20 mol % cholesterol), we could produce several GUVs, but the number of GUVs was much lower than that of DPPC/30%chol GUVs or DPPC/40%chol GUVs. DPPC/30%chol GUVs were spherical in most cases, but DPPC/40%chol GUVs had several kinds of shapes other than sphere, such as a prolate shape (i.e., a prolate) and a tubelike shape (i.e., a tube). Most GUVs have diameters of 10-30 µm. As the GUVs, we selected the vesicles of the membrane for which the contrast was very low and also for which the undulation motion was large. Fluorescence intensities of these GUVs containing 0.020.05 mol % BODIPY-chol using the fluorescence microscope with the highly sensitive EB-CCD camera were much lower than those from oligo-lamellar vesicles (see details in Materials and Methods). To confirm that these GUVs are unilamellar vesicles, we also measured an isothermal area compressibility modulus (i.e., elastic modulus) KA by the micropipet aspiration technique.15 To select unilamellar vesicles for the measurement of the elastic modulus KA, we included 0.05 mol % BODIPY-chol into the DPPC/chol membrane and observed giant liposomes using the EB-CCD camera. For the DPPC/40%chol GUV in water at 20 °C, the value of KA of the membrane was 1200 ( 100 mN/m (n ) 21). This value is similar to that of other lo phase membranes (e.g., 870 mN/m for 1-stearoyl-2-oleoyl-sn-glycero-3-phosphocholine (SOPC)/ 50%chol),16 which is much higher than those of the LR phase membranes (100-200 mN/m).16 Both the results of fluorescence microscope and the elastic modulus indicate

Figure 1. (A) Temperature dependence of the diameter of DPPC/30%chol GUVs and DPPC GUVs from 50 to 28 °C: (3) DPPC/30%chol GUVs, (O) 0.2%PEG-lipid/DPPC/30%chol GUVs, and (b) 0.2%PEG-lipid/DPPC GUVs. (B) Temperature dependence of the diameter of SM/40%chol GUVs and SM GUVs from 50 to 28 °C: (O) SM/40%chol GUVs and (b) SM GUVs.

that most of the DPPC/chol GUVs were unilamellar vesicles. DPPC/chol membranes containing more than 25 mol % cholesterol at 20-37 °C are in the liquid-ordered phase (lo phase).1-3 Thereby, the above experimental results indicate that GUVs can be formed from the membranes in the lo phase, although they cannot be formed from the membrane in the gel phase. To confirm that membranes of the DPPC/chol GUVs are in the lo phase, we investigated the temperature dependence of the diameter of DPPC GUVs and that of DPPC/chol GUVs from 50 to 28 °C. At first, we tried to produce DPPC GUVs in water at 50 °C but failed. Thereby, to prepare DPPC GUVs, we incorporated a small amount (0.2 mol % of the total lipid) of a poly(ethylene glycol) [PEG]-grafted phospholipid, PEG2K-DPPE (PEG-lipid), in DPPC membrane and succeeded in the large production of 0.2%PEG-lipid/DPPC GUVs at 50 °C. As discussed in our previous paper, hydrophilic polymers attached on the surface of electrically neutral lipid membranes by incorporating a small amount of PEG-lipid in PC membrane increase the formability of GUVs of PC membranes in water and also in buffers containing high concentrations of salts.26,27 Figure 1A (b) shows that the diameter of 0.2%PEG-lipid/DPPC GUVs began to decrease at 42 ( 2 °C (n ) 4) when temperature decreased from 50 to 28 °C. It is reasonable to consider that this result was due to the chain-melting phase transition of the DPPC membrane, since its Tm is 41.4 °C. We estimated the difference of the average surface area per lipid molecule for the DPPC bilayer membrane between that in the liquid-crystalline phase at 50 °C (AF) and that in the gel phase at 28 °C (AG) from the diameters of this GUV; ∆A/AF ) (AF - AG)/AF ) 0.32 ( 0.08 (n ) 4). This value is almost the same as ∆A/AF ) 0.25, which is estimated by the sophisticated

Shape Changes and Fission of GUVs

analysis of X-ray diffraction (AF ) 64 Å2 at 50 °C and AG ) 47.9 Å2 at 20 °C).30 In contrast, as shown in Figure 1A (3), the diameter of DPPC/30%chol GUVs was almost constant or decreased slightly with a decrease in temperature from 50 to 28 °C. We obtained the same results for 0.2%PEG-lipid/DPPC/30%chol GUVs (Figure 1A (O)) and also for DPPC/40%chol GUVs. Comparison of the results of these DPPC/chol GUVs with those of DPPC GUVs supports that these DPPC/chol GUVs were in the lo phase, and it also shows that the temperature dependence of the surface area of the lo phase membrane was very small. Bagatolli and Gratton31 reported a similar result on the temperature dependence of the diameters of DPPC GUVs and those of DPPC/30%chol GUVs, which were prepared on platinum wire using electroformation.32 However, in their results, ∆A/AF of DPPC GUVs was 0.13, which was lower than our experimental result. At present, we do not know the reason for this discrepancy, but we suspect that the difference in the preparation method of GUVs is one of the main reasons. 3.2. Effect of Temperature on the Diameter of SM/ chol GUVs and That of SM GUVs. Cell membranes contain a relatively high concentration of SM, and SM is considered to be a key component of the raft.7,8,13 Membranes of SM and cholesterol mixture (SM/cholesterol membranes) also form the lo phase.6 Thereby, it is interesting to compare the formation of GUVs of SM/ cholesterol membrane and its physical properties with those of DPPC/chol GUVs. We tried to produce GUVs of SM/cholesterol membranes containing 0-50 mol % cholesterol in water at 20-37 °C, which is lower than Tm of SM. We could not produce any GUVs of 100% SM membrane and 90 mol % SM/10 mol % cholesterol membrane when the solution was incubated at 20-37 °C. On the other hand, we could produce many GUVs of SM/cholesterol membranes containing g30 mol % cholesterol in water. In most cases, these SM/chol GUVs were spherical, but there were other shapes such as prolate shapes, tubelike shapes, and pearls on a string. When we observed SM/30%chol GUVs (or SM/40%chol GUVs) containing 0.05 mol % BODIPY-chol, their fluorescence intensities were much lower than those from oligo-lamellar vesicles. SM/cholesterol membranes containing 30 mol % cholesterol at 20-37 °C are in the lo phase.6 Thereby, the above experimental results indicate that GUVs can be formed from the membranes in the lo phase. We also investigated the temperature dependence of the diameter of SM GUVs and SM/40%chol GUVs from 50 to 28 °C. We could prepare SM GUVs in water at 50 °C. While the temperature decreased from 50 to 28 °C, the diameter of SM GUVs began to decrease at 43 ( 2 °C (n ) 6) and at lower than 38 °C the diameters were almost constant (Figure 1B (b)). It is reasonable to consider that this result was due to the chain-melting phase transition of the SM membrane, because the brain SM has various kinds of acyl chains, and thereby, the temperature range of the chain-melting phase transition was very broad around 41 °C.33 We estimated the difference of the average surface area per lipid molecule for the SM bilayer membrane between that in the liquid-crystalline phase at 50 °C (AF) and that in the gel phase at 28 °C (AG) from (30) Nagle, J. F.; Tristram-Nagle, S. Biochim. Biophys. Acta. 2000, 1469, 159. (31) Bagatolli, L. A.; Gratton, E. Biophys. J. 1999, 77, 2101. (32) Angelova, M. I.; Dimitrov, D. S. Faraday Discuss. Chem. Soc. 1986, 81, 303. (33) Do¨bereiner, H.-G.; Ka¨s, J.; Noppl, D.; Sprenger, I.; Sackmann, E. Biophys. J. 1993, 65, 1396.

Langmuir, Vol. 20, No. 22, 2004 9529

the diameters of this GUV; ∆A/AF ) (AF - AG)/AF ) 0.24 ( 0.02 (n ) 6). On the other hand, as shown in Figure 1B (O), the diameter of the SM/40%chol GUVs was almost constant or decreased slightly with a decrease in temperature from 50 to 28 °C. This result also supports that SM/40%chol GUVs were in the lo phase. 3.3. Shape Changes and Vesicle Fission of DPPC/ chol GUVs and SM/chol GUVs in the lo Phase Induced by Lysophosphatidylcholine. At first, we investigated the effect of low concentrations of lyso-PC(16:0) with a long acyl chain (cmc, 7 µM)34,35 on GUVs of the lo phase membrane. Figure 2 shows several kinds of shape changes of a DPPC/40%chol GUV in the lo phase induced by addition of 0.2 µM lyso-PC(16:0) in 0.1 M glucose aqueous solution through a 10 µm diameter micropipet near the GUV at 20 ( 2 °C. In Figure 2A, at first (in the absence of lyso-PC), the GUV was a prolate (Figure 2A-(1)). After the addition of lyso-PC(16:0), the shape changed into a pear (Figure 2A-(2)) and then into an asymmetrical two-spheres connected by a narrow neck (Figure 2A-(3)) (here, the meaning of “asymmetric” is that the sizes of the two spheres are different). Finally the diameter of the neck became very small (we define this shape as vesiculation; see also a following analysis of the shape) (Figure 2A-(4)) (n ) 10). The shape change from the two-spheres connected by a narrow neck (Figure 2A(3)) to that connected to a very narrow neck (i.e., vesiculation) (Figure 2A-(4)) completed in less than 1 s, and thereby, this shape change can be considered as a transition (i.e., vesiculation transition). In some cases, a prolate changed into a symmetrical dumbbell and then into a symmetrical two-spheres connected by a narrow neck. To determine the reversibility of the shape change, the addition of lyso-PC was stopped after the shape change of the GUV completed, and then we observed the shape change of the GUV. Figure 2A-(5)-(8) shows the time course of the shape change of the GUV during the stop of the addition of lyso-PC. At first, the two-spheres connected by a narrow neck changed into the pear (Figure 2A-(7)) and then into the prolate (Figure 2A-(8)). During the stop of the addition, lyso-PC diffused away from the vicinity of the GUV into the bulk solution, inducing the decrease in lyso-PC concentration near the GUV, and then the partition of lyso-PC into the membrane decreases (i.e., lyso-PC molecules in the outer monolayer of the GUV move into the aqueous solution). This result indicates that the 0.2 µM lyso-PC induced shape change of DPPC/40%chol GUVs was reversible, indicating that no vesicle fission occurred. The reversible shape changes of GUVs were reported in other cases, such as the interaction of La3+ (Gd3+) with GUVs or the de novo designed peptide with GUVs.22,23 When we added 0.2 µM lyso-PC(16:0) near a cylindrical GUV (Figure 2B-(1)), the shape changed into a GUV made of a series of many spherical vesicles connected by a narrow neck (so-called “pearls on a string”) (Figure 2B-(4)) (n ) 6). To determine the reversibility of the shape change, the addition of lyso-PC was stopped after the shape change of the GUV completed, and then we observed the shape of the GUV (Figure 2B-(5)-(8)). This shape change was also reversible, indicating no vesicle fission occurred. We did not observe these shape changes when we added 0.01 µM lyso-PC(16:0). These data on the lyso-PC-induced shape changes of GUVs of the lo phase membrane indicate that the flip-flop of lyso-PC (i.e., exchange of lyso-PC between the external monolayer and the internal monolayer in one bilayer) in the lo phase (34) Haberland, M. E.; Reynolds, J. A. J. Biol. Chem. 1975, 250, 6636. (35) Kumar, V. V.; Baumann, W. J. Biophys. J. 1991, 59, 103.

9530

Langmuir, Vol. 20, No. 22, 2004

Tanaka et al.

Figure 2. Shape change of a DPPC/40%chol GUV induced by the addition of 0.2 µM lyso-PC(16:0) at 20 °C. (A) The prolate changed into a pear and then into an asymmetrical two-spheres connected by a narrow neck. The time after starting injection of 0.2 µM lyso-PC solution through the micropipet is (1) 0 s, (2) 26 s, (3) 30 s, and (4) 31 s for the pictures. After the addition of lyso-PC was stopped, the shape change was reversed. The time after stopping injection of lyso-PC is (5) 0 s, (6) 180 s, (7) 190 s, and (8) 195 s. The bar in the picture corresponds to 10 µm. (B) The cylinder changed into pearls on a string. The time after starting injection of 0.2 µM lyso-PC solution through the micropipet is (1) 0 s, (2) 50 s, (3) 74 s, and (4) 104 s. After the addition of lyso-PC was stopped, the shape change was reversed. The time after stopping injection of lyso-PC is (5) 0 s, (6) 22 s, (7) 44 s, and (8) 102 s. The bar in the picture corresponds to 10 µm.

membrane is very slow, that is, there is almost no flip-flop within 1 min (which is the time scale of the shape changes). As far as we know, there is no other experimental data on the flip-flop of lyso-PC in the lo phase. In lipid membranes in the LR phase, rates of flip-flop of lyso-PC were reported to be very slow,36,37 although in cells such as erythrocyte its flip-flop is significant.38 Next, we investigated the effect of higher concentrations (1 µM; less than its cmc) of lyso-PC(16:0) on DPPC/40%chol GUVs in the lo phase. In Figure 3A, at first (in the absence of lyso-PC), the GUV was a prolate (Figure 3A-(1)). After the addition of 1 µM lyso-PC(16:0), the shape changed into a pear (Figure 3A-(2)) and then into an asymmetrical two-spheres connected by a neck (Figure 3A-(3)), and finally two spherical vesicles were separated (Figure 3A(4)) (n ) 11). In some cases, a prolate changed into a symmetrical dumbbell and then into a symmetrical twosphere connected by a narrow neck, and finally two spherical vesicles were separated. To determine the reversibility of the shape change, the addition of lyso-PC was stopped after the shape change of the GUV completed, and then we observed the shape of the GUV. The distance between two spherical vesicles increased with an increase in time (Figure 3A-(5)). This result indicates that the 1 µM lyso-PC induced shape change of DPPC/40%chol GUVs was not reversible, that is, vesicle fission occurred. When (36) Needham, D.; Zhelev, D. V. Ann. Biomed. Eng. 1995, 23, 287. (37) Bhamidipati, S. P.; Hamilton, J. A. Biochemistry 1995, 34, 5666. (38) Schwichtenho¨vel, C.; Deuticke, B.; Haest, C. W. M. Biochim. Biophys. Acta 1992, 1111, 35.

we added 1 µM lyso-PC(16:0) near a cylindrical GUV (Figure 3B-(1); C-(1)), the shape changed into pearls on a string (Figure 3B-(3); C-(2)), and finally several spherical vesicles were separated (Figure 3B-(4); C-(3)) (n ) 6 for both cases (B and C)). To determine the reversibility of the shape change, the addition of lyso-PC was stopped after the shape change of the GUV completed, and then we observed the shape of the GUV. The distance between these spherical vesicles increased with an increase in time (Figure 3B-(5); C-(4)), indicating that the 1 µM lyso-PC induced shape change of DPPC/40%chol GUVs was not reversible, that is, vesicle fission occurred. We obtained the same result in the interaction of 3 µM lyso-PC(16:0) on DPPC/40%chol GUVs. In the interactions of lyso-PC(16:0) (e10 µM) with SM/ 40%chol GUVs, similar results were obtained; for example, 0.1 µM lyso-PC(16:0) induced reversible shape changes of SM/40%chol GUVs (n ) 8), which are the same as those in DPPC/40%chol GUVs, and 1 µM lyso-PC(16:0) induced vesicle fission of GUVs (n ) 12). In contrast, in the interactions of lyso-PC(16:0) (e10 µM) with DOPC/ 40%chol GUVs or DOPC GUVs, the same shape changes were observed, but vesicle fission did not occur (n ) 10 for both cases). In all the above experiments, we added lyso-PC solution near GUVs using the same method as that we successfully applied to the investigation on the interaction of La3+ (or Gd3+) and de novo designed peptides on DOPC GUVs.22,23 In these experiments, we added a given concentration of substance (lyso-PC, La3+, and peptides) solution near the

Shape Changes and Fission of GUVs

Langmuir, Vol. 20, No. 22, 2004 9531

Figure 3. Shape change of a DPPC/40%chol GUV induced by the addition of 1 µM lyso-PC(16:0) at 20 °C. (A) The prolate changed into a pear and then into an asymmetrical two-spheres connected by a neck. The time after starting injection of 1 µM lyso-PC solution through the micropipet is (1) 0 s, (2) 17 s, (3) 18 s, and (4) 40 s. After the addition of lyso-PC was stopped, the distance between two spherical vesicles increased with an increase in time. The time after stopping injection of lyso-PC is (5) 240 s. (B) The cylinder changed into pearls on a string. The time after starting injection of 1 µM lyso-PC solution through the micropipet is (1) 0 s, (2) 21 s, (3) 25 s, and (4) 45 s. After the addition of lyso-PC was stopped, the distance between spherical vesicles increased with an increase in time. The time after stopping injection of lyso-PC is (5) 270 s. (C) The cylinder changed into pearls on a string. The time after starting injection of 1 µM lyso-PC solution through the micropipet is (1) 0 s, (2) 22 s, and (3) 37 s. After the addition of lyso-PC was stopped, the distance between spherical vesicles increased with an increase in time. The time after stopping injection of lyso-PC is (4) 180 s. The bar in the picture corresponds to 10 µm.

GUV (the distance between the GUV and the tip of micropipet is about 100 µm) through the micropipet so slowly that no shape changes of the GUV occurred due to the pressure of water. We also made control experiments and confirmed its validity; that is, we added the same buffer using the same method, but we could not observe any shape changes of GUVs. When we added lower concentrations of substances below the threshold concentration, no shape changes were observed. Moreover, the de novo designed peptide induced the same shape changes of DOPC GUVs as those of DPPC/chol GUVs induced by lyso-PC, whereas La3+ (or Gd3+) induced the completely opposite type shape changes of DOPC GUVs such as the discocyte via stomatocyte to inside budded shape transformation and the two-spheres connected by a narrow neck changed into the prolate transformation.22,23 The former shape changes can be explained by the increase in the area of the external monolayer membrane of GUVs, but the latter shape change can be explained by the decrease in the area of the external monolayer membrane on the basis of the bilayer-couple model or the areadifference-elasticity model (see the discussion below). We also investigated the effects of La3+ on the shapes of DPPC/ chol GUVs in the lo phase and observed the discocyte via stomatocyte to inside budded shape transformation and the two-spheres connected by a narrow neck changed into the prolate transformation, which are completely opposite shape changes observed in the interaction of lyso-PC with the same DPPC/chol GUVs using the same method (Sano et al., manuscript in preparation). These data clearly show that there is no artifact due to the pressure of water during the addition of the substance solution through the micropipet, and thereby, this method, that is, the GUV method, is useful to investigate the interaction of substances with GUVs. In these experiments, we continued to add a given concentration of substance (lyso-PC, La3+, and peptides) solution near the GUV, but at the same

time the diffusion of substance from the vicinity of the GUV into the bulk phase occurred. Thereby, we observed these phenomena at the steady-state condition of substance concentration near the GUV, not at the equilibrium condition.22 Therefore, the substance concentration near the GUV at the steady-state condition may be a little lower than that of the solution in the micropipet. Experiments to check the reversibility of shape changes clearly show the difference between vesicle fission and reversible shape change. As described above, the experiments of the reversibility of the shape change indicate that 1 µM lyso-PC induced vesicle fission of various kinds of DPPC/40%chol GUVs, but 0.2 µM lyso-PC did not induce it. However, when we observed only the appearance of GUVs (such as shape and the distance between small spherical vesicles) at the final stage of shape changes, it was a little difficult to determine whether vesicle fission occurred or not. We developed a new method to discriminate the two different states. In this method, after shape changes of GUVs complete, we aspirate gently a smaller spherical vesicle of the two-spheres connected by a neck by the suction of water using a micropipet located at the vicinity of the vesicle but not in contact with the vesicle. We can reasonably expect that if there is no vesicle fission, two vesicles would be connected by a long, narrow tube after the aspiration and that if vesicle fission occurs, there would be no tube connected with vesicles. Figure 4 shows two examples of these experiments. In Figure 4Aa, 1 µM lyso-PC(16:0) was added near a prolate of DOPC/40%chol membrane in the LR phase containing 0.05% BODIPYchol (Figure 4Aa-(1)), and the shape changed into an asymmetrical two-spheres connected by a neck (Figure 4Aa-(2)). Then, a smaller vesicle of the deformed GUV (Figure 4Aa-(2)) was slowly aspirated by the suction of water using the micropipet. A narrow tube between spherical vesicles appeared, and its length increased with an increase in time of aspiration (Figure 4Ab-(3)-(5)). A

9532

Langmuir, Vol. 20, No. 22, 2004

Figure 4. Discrimination of the fission and no fission from GUVs. (Aa) 1 µM lyso-PC(16:0) was added near a prolate (1) of DOPC/40%chol membrane in the LR phase containing 0.05% BODIPY-chol; the shape changed into an asymmetrical twospheres connected by a neck (2). (Ab) Then, a smaller vesicle was aspirated using the micropipet. A narrow tube between spherical vesicles appeared, and its length increased with an increase in time of aspiration ((3)-(5)). The time after the aspiration of the smaller vesicle is (3) 4 s, (4) 7 s, and (5) 40 s. Panel (6) is a fluorescence microscopy image of panel (5), which clearly shows the existence of the long, narrow tube between spherical vesicles. (Ba) 1 µM lyso-PC(16:0) was added near a prolate (1) of DPPC/40%chol membrane in the lo phase containing 0.05% BODIPY-chol; the shape changed into an asymmetrical two-spheres connected by a neck ((2),(3)). (Bb) Then, a smaller vesicle was aspirated using the micropipet. The time after the aspiration of the smaller vesicle is (3) 3 s, (4) 13 s, and (5) 28 s. Panel (6) is a fluorescence microscopy image of panel (5), which clearly shows no existence of the long, narrow tube between spherical vesicles. The bar in the picture corresponds to 10 µm.

fluorescence microscopy image (Figure 4Ab-(6)) also clearly shows the existence of the long, narrow tube between spherical vesicles. On the other hand, in Figure 4Ba, 1 µM lyso-PC(16:0) was added near a prolate of DPPC/ 40%chol membrane containing 0.05% BODIPY-chol in the lo phase (Figure 4Ba-(1)), the shape changed into an asymmetrical two-spheres connected by a neck (Figure 4Ba-(2)), and then it seems that two spherical vesicles were separated (Figure 4Ba-(3)). When a smaller vesicle of the separated GUVs (Figure 4Ba-(3)) was slowly aspirated using the micropipet, the distance between two vesicles increased and a narrow tube between spherical vesicles did not appear (Figure 4Bb-(4)-(6)). A fluorescence microscopy image (Figure 4Bb-(7)) also clearly shows no existence of the long, narrow tube between spherical vesicles. What is the mechanism for the shape changes of DPPC/ 40%chol GUVs (or SM/40%chol GUVs) in the lo phase and DOPC/40%chol (or DOPC) GUVs in the LR phase?

Tanaka et al.

Shapes of GUVs of lipid membranes are determined by the minimum of the elastic energy of the closed membrane of the GUV. In the bilayer-couple model, the elastic energy of the GUV (Wel) is due to only the bending energy (Wb) of its membrane since in this model it is assumed that the monolayer membrane cannot stretch elastically, and thereby, its minimum is determined for a given area A, a given volume V, and also a given difference between the area of the external monolayer (Aex) and that of the internal monolayer (Ain) in the bilayer membrane of the GUV, ∆A ()Aex - Ain).22,23,39-41 Recently, the area-difference-elasticity model (ADE model) (i.e., the generalized bilayercouple model) has been proposed.42,43 In this model, the area of each monolayer is not fixed to the equilibrium area but can increase or decrease (i.e., the monolayer membrane can stretch elastically) to increase the nonlocal elastic energy of the membranes. Thereby, in this ADE model, the elastic energy of the GUV (Wel) can be expressed as a sum of the membrane bending energy (Wb) and the energy of the relative monolayer stretching (Wr) which is proportional to (∆A - ∆A0)2 (i.e., Wr ∝ (∆A - ∆A0)2), where ∆A0 ()A0ex - A0in) is the area difference between the two monolayers in the bilayer membrane at equilibrium (i.e., nonstretched monolayers). In the ADE model, the shape of the GUV is determined by the minimization of the membrane elastic energy (Wel) for a given area A, a given volume V, and also a given equilibrium (i.e., relaxed) area difference ∆A0. The analysis based on the ADE model shows that under the condition of the constant volume of the GUV, the shape changes as follows: with an increase in ∆A0, (1) prolate f pear (i.e., asymmetric prolate) f two-spheres connected by a narrow neck, (2) dumbbell (i.e., symmetric prolate) f two-spheres connected by a narrow neck, and (3) cylinder f pearls on a string.43 Further increase in ∆A0 induces vesiculation where the neck diameter goes to zero at a critical value of ∆A0.43 However, above the critical value of ∆A0, the ADE model does not give us a good prediction, and also it cannot explain vesicle fission since the vesicle fission involves a topological change of the membrane. These shape changes predicted by the above analysis based on the ADE model are completely the same as the lyso-PC-induced shape changes of DPPC/40%chol GUVs (Figure 2). Thereby, this analysis shows that lyso-PC entered into the external monolayer of the GUVs from the outside aqueous solution, increasing ∆A0 of the GUVs. This is the main reason for the lyso-PC-induced shape changes of the GUVs. Moreover, this analysis indicates that monomer lyso-PC in aqueous solution can enter into the lo phase membrane, suggesting that substances with a long hydrocarbon chain such as lyso-PC can enter into the lo phase membrane and also into the raft in the cell membrane. Finally, we investigated the effect of low concentrations of the other three kinds of lyso-PC with different length of acyl chains (lyso-PC(14:0), lyso-PC(12:0), and lyso-PC(10:0)) on DPPC/40%chol GUVs of the lo phase membrane, to elucidate the effect of acyl chain length of lyso-PC on the lyso-PC-induced shape changes and vesicle fission. All three lyso-PCs induced the same shape changes and vesicle fission as observed in the interaction of lyso-PC(16:0). In Table 1, we summarized threshold concentrations of lyso-PC which we added to induce shape changes of the GUV (i.e., prolate f two-spheres connected by a (39) Svetina, S.; Zeks, B. Eur. Biophys. J. 1989, 17, 101. (40) Seifert, U.; Berndl, K.; Lipowsky, R. Phys. Rev. A 1991, 44, 1182. (41) Iglic, A.; Kralj-Iglic, V.; Majhenc, J. J. Biomech. 1999, 32, 1343. (42) Heinrich, V.; Svetina, S.; Zeks, B. Phys. Rev. E 1993, 48, 3112. (43) Miao, L.; Seifert, U.; Wortis, M.; Do¨bereiner, H.-G. Phys. Rev. E 1994, 49, 5389.

Shape Changes and Fission of GUVs

Langmuir, Vol. 20, No. 22, 2004 9533

Table 1. Threshold Concentrations of Lyso-PC-Induced Shape Changes and Membrane Fission of DPPC/40%chol GUVs threshold threshold concentration concentration of lyso-PC in of lyso-PC in water to induce water to induce shape changes vesicle fission of GUVs of GUVs cmc of Cw,/(shape change) Cw,/(vesicle fission) lyso-PC (µM) (µM) (µM)a

w,* Cmem,* ) KACw,* A ) KBCB , ∴

Cw,* B Cw,* A

)

KA ) 10 KB

(3)

Here, we consider that the difference of chain length of lyso-PC affects only the standard chemical potential of lyso-PC in water due to the difference of hydrophobic energy, but we assume that the standard chemical potential of lyso-PC in lipid membrane does not change, ) µmem,0 . If we consider that the same that is, µmem,0 A B threshold concentrations of lyso-PC in the membrane, that ) Cmem,* ) Cmem,*, induced shape changes and is, Cmem* A B vesicle fission of the GUVs, we can calculate the ratio of the threshold concentration of lyso-PC(B) in water, Cw,* B , to that of lyso-PC(A) in water, Cw,* A , as follows.

As shown in Table 1, the experimental results agree with eq 3 almost well, although the data of lyso-PC(10:0) deviate from the expected values from eq 3 a little. Probably, the standard chemical potential of lyso-PC(10:0) in lipid membrane is larger than that of lyso-PC(16:0) due to the mismatch of the hydrocarbon chains in DPPC/40%chol membranes. However, we can conclude that the above thermodynamic analysis can reasonably explain the dependence of both threshold concentrations on the chain length of lyso-PC, indicating that the shape changes and vesicle fission of the GUVs of the lo phase membrane occur at threshold concentrations of lyso-PC in the membrane. Determination of the threshold concentrations of lysoPC in the membrane to induce shape changes and vesicle fission of GUVs is important. However, at present, we do not have a highly sensitive method to determine the concentration of lyso-PC in the membrane, and thereby, it is difficult to determine these concentrations because they are very low. The GUV method to detect the area change in the interaction of substances with the external monolayer of the GUV, which we used in the experiments of shape changes of GUVs, can detect a much smaller area change of the GUV than the micropipet method does.20-23 However, the GUV method can give us only qualitative information on the area change, not quantitative data (i.e., the concentration of the substance in the membrane). Further investigation to determine the concentrations of lyso-PC in the membrane to induce shape changes and vesicle fission of GUVs is necessary. Our experimental results show that threshold concentrations of lyso-PC in the membrane are required for the vesicle fission of the GUV of the lo phase membrane. At present, we do not know the mechanism of the lyso-PCinduced vesicle fission of the DPPC/40%chol GUVs (or SM/40%chol GUVs) in the lo phase. However, on the basis of the data and the analysis described above, we can consider the following process for the lyso-PC-induced vesicle fission. Lyso-PC molecules enter into the external monolayer membrane of the GUV, to decrease the Gibbs free energy of the total system. Then, shape change of the GUV occurs until the elastic energy of the GUV (Wel) becomes minimum, and the shape becomes the twospheres connected to a very narrow neck (e.g., Figure 2A(4)) (Figure 5B-(2)). This induces stretching of the internal monolayer membrane; especially at the region of the narrow neck its stretching is very large due to the large curvature of the membrane (Figure 5B-(2)), since we can reasonably consider that no flip-flop occurs between the external and the internal monolayer on the basis of the reversibility of the lyso-PC-induced shape changes below the threshold concentration. When the lyso-PC concentration in the external monolayer increases and reaches the threshold value, the breakdown of the internal monolayers occurs at the region of the narrow neck and then they reseal to reduce the high curvature of the internal monolayers (Figure 5B-(3)). To decrease the elastic energy of the high-curvature region of the internal monolayer at the neck, the GUV becomes a shape with a low-curvature internal monolayer (Figure 5B-(4)). However, in the GUV, there is still a high-energy interstitial hydrocarbon region at the narrow neck (indicated by blue triangles in Figure 5B-(4)).45 To decrease the free energy

(44) Israelachvili, J. Intermolecular and surface forces, 2nd ed.; Academic Press: San Diego, 1992.

(45) Kinoshita, K.; Li, S. J.; Yamazaki, M. Eur. Biophys. J. 2001, 30, 207.

5 × 102 2 × 101 9 × 10-1 7 × 10-2

lyso-PC(10:0) lyso-PC(12:0) lyso-PC(14:0) lyso-PC(16:0) a

1 × 103 4 × 101 8 5 × 10-1

7 × 103 7 × 102 7 × 101 7

From ref 41.

narrow neck) and those to induce vesicle fission of the GUV. These threshold concentrations are defined as lysoPC concentrations at which shape changes (or vesicle fission) occurred at 50% of the examined GUVs. Both threshold concentrations increased greatly with a decrease in chain length (i.e., a decrease in the number of carbons of the hydrocarbon chain), and they were much less than the cmc of each lyso-PC. The interaction free energy between hydrocarbon and water (i.e., the so-called hydrophobic energy) increases with an increase in the number of CH2 groups, and its increment per CH2 group is 2.8 kJ mol-1 for lyso-PC.44 This hydrophobic energy unstabilizes monomer lyso-PC in water, and thereby, the partition constant, K, of lyso-PC from water to lipid membrane increases with an increase in the number of CH2 groups. We can consider this phenomenon quantitatively. The partition constant, KA, of a kind of lyso-PC (here, we call it lyso-PC(A)) can be expressed as follows:

KA )

Cmem A Cw A

) exp[-(µmem,0 - µw,0 A A )/RT]

(1)

and Cw where Cmem A A are the concentration of lyso-PC(A) in lipid membrane and that in water, respectively, and µmem,0 and µw,0 are the standard chemical potential of A A lyso-PC(A) in lipid membrane and that in water, respectively. When lyso-PC(A) has a longer acyl chain than lysoPC(B), and the difference in the number of CH2 groups is 2, we can calculate a ratio of the partition constant of lyso-PC(A), KA, to that of lyso-PC(B), KB, using eq 1 as follows.

KA w,0 ) exp[(µw,0 A - µB )/RT] ) exp(5.6 kJ/2.4 kJ) ) 10 KB (2)

9534

Langmuir, Vol. 20, No. 22, 2004

Tanaka et al.

Figure 5. Scheme of the proposed mechanism of lyso-PC-induced vesicle fission of DPPC/chol GUVs in the lo phase. A detailed description is provided in the text. Red arrows indicate the lateral pressure of the internal membranes induced by the shape change of the GUV. Blue triangles indicate the interstitial hydrocarbon region where the free energy of chain packing is very large.

of this shape, the merge of outer monolayer membranes occurs, and vesicle fission completes (Figure 5B-(5), Figure 5A-(2)). This process for the lyso-PC-induced vesicle fission is a mirror image of the “stalk model” of vesicle fusion, which is one of the most popular models for vesicle fusion between the influenza virus and endosome inside cells.46-49 In this model, Figure 5B-(1), -(3), and -(4) correspond to the fusion pore, the hemi-fusion intermediate, and the (46) Blumenthal, R.; Clague, M. J.; Durell, S. R.; Epand, R. Chem. Rev. 2003, 103, 53. (47) Cevc, G.; Richardsen, H. Adv. Drug Delivery Rev. 1999, 38, 207. (48) Chernomordik, L.; Chanturiya, A.; Green, J.; Zimmerberg, J. Biophys. J. 1995, 69, 922. (49) Siegel, D. P.; Epand, R. M. Biophys. J. 1997, 73, 3089.

stalk, respectively. For the vesicle fusion based on the stalk model, cone-shape type lipids are required to form a negative curvature,46-49 although for the lyso-PC-induced vesicle fission such kinds of lipids are not necessary. To elucidate the detailed mechanism of the vesicle fission, we need further experimental and theoretical studies. Acknowledgment. This work was supported in part by a Grant-in-Aid for General Scientific Research C (No. 15510099) from the Ministry of Education, Science, and Culture (Japan) to M.Y. LA049481G