Membrane Fusion of Giant Unilamellar Vesicles of Neutral

Faculty of Science, Shizuoka University, 836 Oya, Shizuoka 422-8529, Japan. Received February 6, 2004. In Final Form: May 1, 2004. Membrane fusions of...
0 downloads 0 Views 465KB Size
5160

Langmuir 2004, 20, 5160-5164

Membrane Fusion of Giant Unilamellar Vesicles of Neutral Phospholipid Membranes Induced by La3+ Tomoki Tanaka† and Masahito Yamazaki*,†,‡ Materials Science, Graduate School of Science and Engineering, and Department of Physics, Faculty of Science, Shizuoka University, 836 Oya, Shizuoka 422-8529, Japan Received February 6, 2004. In Final Form: May 1, 2004 Membrane fusions of vesicles of biomembranes play various important roles in cells, but their mechanisms are unclear and controversial. In the present study, we found that 30 µM to 1 mM La3+ induced membrane fusion of two giant unilamellar vesicles (GUVs) composed of a mixture of dioleoylphosphatidylcholine (DOPC) and dipalmitoleoylphosphatidylethanolamine (DPOPE). We succeeded in observing a process of this membrane fusion in detail. First, two GUVs became strongly associated, with a partition membrane between them composed of two bilayers, one from each GUV. Then, the partition membrane was suddenly broken at one site on its edge. The area of this breakage site gradually spread, until it was completely separated from the GUV to complete the membrane fusion. Here, we propose a new model (i.e., the partition breakage model) for the mechanism of La3+-induced membrane fusion of GUVs.

1. Introduction Membrane fusions of vesicles of biomembranes, including cell fusion, play various important roles in cells. Although there have been many studies of membrane fusion, the mechanisms involved are still unclear and controversial.1,2 For example, in membrane fusion induced by SNARE proteins, the mechanism of the association between vesicles and cell membranes has been revealed, but the mechanism of their membrane fusion is still not clear.3,4 At present, two models of membrane fusion are popular. One is the so-called stalk model.5,6 In this model, after the formation of the stalk between the associated vesicles, a hemifusion occurs followed by formation of a bilayer between the associated vesicles. The other model is the defect model.2,7 In this model, a local disorder is created in the membrane in the two associated vesicles, followed by membrane fusion. To elucidate the mechanism and process of membrane fusion, many studies of membrane fusion between small vesicles (diameter less than 200 nm) have been conducted using various techniques. However, in these studies, average physical properties of many instances of membrane fusion have been measured; thus, no direct information about single instances of membrane fusion has been obtained. Recently, giant unilamellar vesicles (GUVs) of phospholipid membranes with diameters greater than 10 µm have been used in studies of elastic properties of phospholipid membranes and shape changes of vesicles.8-11 In these experiments, physical properties of single GUVs such as elastic constant and shape changes have been * To whom correspondence may be addressed at Department of Physics, Faculty of Science, Shizuoka University, 836 Oya, Shizuoka 422-8529, Japan. Tel and fax: 81-54-238-4741. E-mail: spmyama@ ipc.shizuoka.ac.jp. † Materials Science, Graduate School of Science and Engineering. ‡ Department Physics, Faculty of Science. (1) Blumenthal, R.; Clague, M. J.; Durell, S. R.; Epand, R. Chem. Rev. 2003, 103, 53. (2) Cevc G.; Richardsen, H. Adv. Drug Delivery Rev. 1999, 38, 207. (3) Pariati, F.; McNew, J. A.; Fukuda, R.; Miller, R.; So¨llner, T. H.; Rothman, J. E. Nature 2000, 407, 194. (4) Hu, C.; Ahmed, M.; Melia, T. J.; So¨llner, T. H.; Mayer, T.; Rothman, J. E. Science 2003, 300, 1745. (5) Chernomordik, L.; Chanturiya, A.; Green, J.; Zimmerberg, J. Biophys. J. 1995, 69, 922. (6) Siegel, D. P.; Epand, R. M. Biophys. J. 1997, 73, 3089. (7) Yamazaki, M.; Ito, T. Biochemistry 1990, 29, 1309.

observed. Such findings suggest that investigation of membrane fusion of single GUVs could help to elucidate its mechanism. We have hypothesized that the lateral tension of lipid membranes plays an important role in their structure and stability in processes such as phase transition and membrane fusion. Previously, we have investigated effects of La3+ and Gd3+ on membranes composed of phospholipids such as phosphatidylcholine (PC) and phosphatidylethanolamine (PE), which have no net charge at neutral pH.12 The binding of La3+ and Gd3+ on PC membranes was indicated by several experimental techniques; for example, using the titration calorimetry,13 the binding constant K of La3+ on PC membrane was determined to be 4 × 103, and the measurement of the ζ-potential14 shows that, for example, ζ-potentials of PC membranes in the presence of 100 µM and 1 mM were 18 and 32 mV, respectively, and its analysis shows the binding constant K of Gd3+ on PC membrane was 1 × 103. We found that chain-melting transition temperatures of PC and PE membranes increased with an increase in La3+ concentration and even at low concentrations (100 µM to 1 mM) their increases were evident, indicating that the lateral compression pressure of the membrane increases with an increase in La3+ concentration.11,15 Addition of 10 to 100 µM La3+ (or Gd3+) through a micropipet near a GUV of dioleoyl-PC (DOPC; C18:1) membrane induced several kinds of shape changes, which can be explained by the decrease in the area of the outer monolayer membrane of the GUV.11 The effect of La3+ on the shape of PC-GUV was the same as that of Gd3+, probably because these ions have the same charge and their sizes were similar.13 The decrease in membrane area of the outer monolayer is (8) Sackmann, E. In Structure and dynamics of membranes; Lipowsky, R.; Sackmann, E., Eds.; Elsevier/North-Holland: Amsterdam, 1995, pp 213-304. (9) Ka¨s, J.; Sackmann, E. Biophys. J. 1991, 60, 825. (10) Mathivet, L.; Cribier, S.; Devaux, P. F. Biophys. J. 1996, 70, 1112. (11) Tanaka, T.; Tamba, Y.; Masum, S. M.: Yamashita, Y.; Yamazaki, Y. Biochim. Biophys. Acta 2002, 1564, 173. (12) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: New York, 1992. (13) Lehrmann, R.; Seelig, J. Biochim. Biophys. Acta 1994, 1189, 89. (14) Ermakov, Y. A.; Averbakh, A. Z.; Arbuzova, A. B.; Sukharev, S. I. Membr. Cell Biol. 1998, 12, 411. (15) Tanaka, T.; Li, S. J.; Kinoshita, K.; Yamazaki, M. Biochim. Biophys. Acta 2001, 1515, 189.

10.1021/la049681s CCC: $27.50 © 2004 American Chemical Society Published on Web 05/26/2004

Letters

Langmuir, Vol. 20, No. 13, 2004 5161

induced by the lateral compression pressure of the membrane due to the interaction of La3+ with the PC membrane interface. We also found that La3+ stabilizes the HII phase rather than the LR phase in the PE membrane, which can be explained by the increase in the lateral compression pressure of the membrane at the local sites.15 In the present study, we investigated the effect of La3+ on GUVs composed of a mixture of DOPC and dipalmitoleoyl-PE (DPOPE, C16:1) (DPOPE/DOPCGUVs), and found that addition of low concentrations of La3+ induced membrane fusion between two DPOPE/ DOPC-GUVs. We also obtained detailed images of the process of membrane fusion. Part of this research was presented at the 40th Annual Meeting of the Biophysical Society of Japan.16 2. Materials and Methods 2.1. Materials. DOPC and DPOPE were purchased from Avanti Polar Lipids Inc. LaCl3 (g99.9%) and poly(ethylene glycol) (average molecular weight, 7500) [PEG 6K] were purchased from Wako Chemical Co. BODIPY-cholesterol was purchased from Molecular Probes Inc. FITC-dextran (average molecular weight, 10 000) was purchased from Sigma-Aldrich Co. 2.2. Formation of GUVs and Observation of GUVs Using a Microscope. Phospholipid GUVs were prepared by natural swelling of dry lipid film and were observed with an inverted fluorescence, phase-contrast microscope (IX-70, Olympus, Tokyo, Japan), using the standard method described in detail in our previous paper.11,17 Fluorescence images of GUVs were recorded on a video recorder using an EB-CCD camera (C7190-23, Hamamatsu Photonics, Hamamatsu, Japan), which is a highly sensitive fluorescence camera. A 10-µL sample of GUV solution (0.1 M sucrose solution; internal solution) was diluted into 290 µL of 0.1 M glucose aqueous solution containing 2% (w/v) PEG 6K (external solution). The resulting mixture was then transferred into a handmade microchamber. 2.3. Association and Membrane Fusion of GUVs. The external solution containing various concentrations of La3+ was added into the vicinity of a GUV through a 10-µm-diameter glass micropipet, the position of which was controlled by a micromanipulator (MMW-23, Narishige, Tokyo, Japan).11,17 To induce membrane fusion between two GUVs, we used two methods. In method 1 (the one-pipet method), after we positioned the micropipet containing La3+ solution in the chamber near a GUV, we added a small amount of La3+ solution, and made contact with the GUV on the upper surface of the micropipet to adsorb the GUV on its surface. Then, we moved this GUV toward another GUV using the micromanipulator and made contact between the two GUVs. Finally, we added La3+ solution slowly through the micropipet, and observed shape changes in the two GUVs using the phase-contrast microscope or the fluorescence microscope. In method 2 (the two-pipet method), we held a GUV by aspiration at the tip of the micropipet, which contains 0.1% (w/v) BSA in 0.1 M glucose solution. The aspiration pressure was small, and the tension of the membrane due to the aspiration was less than 0.1 mN/m. Then, we moved this GUV toward another GUV using the micromanipulator, and made contact between the two GUVs. Finally, we added La3+ solution near these GUVs through another micropipet located at an opposite side of the holding micropipet, the position of which was controlled by another micromanipulator, and observed shape changes in the two GUVs. 2.4. Internal Content Mixing during GUV Membrane Fusion. Two kinds of GUVs were prepared as follows. GUV containing 1 µM FITC-dextran as internal solution was prepared by natural swelling in 0.1 M sucrose solution containing 1 µM FITC-dextran. To prepare GUV with membrane labeled by BODIPY-chol, a lipid mixture (0.1 mol % BODIPY-chol and 99.9 mol % DOPC+DPOPE) in chloroform was dried. All other methods were the same as in section 2.2. (16) Tanaka, T.; Yamazaki, M. Biophys. Jpn. 2002, 42S, 134. (17) Yamashita, Y.; Masum, S. M.; Tanaka, T.; Yamazaki, M. Langmuir 2002, 18, 9638.

3. Results and Discussion We investigated effect of La3+ on DPOPE/DOPC-GUVs using method 1. At first, a GUV was adsorbed on the surface of a micropipet and was contacted with another GUV. When we added 100 µM to 1 mM La3+ solution from the micropipet into the vicinity of two DOPC-GUVs, association between the two GUVs occurred, but membrane fusion did not occur (Figure 1A). Interaction of La3+ with membranes of the GUVs induced suppression (or absence) of undulation motion of membranes, causing the association of GUVs. It is well-known that the suppression of undulation motion induces association of GUVs in other systems.10 On the other hand, Figure 1B shows the shape change of two spherical 30 mol %-DPOPE/70 mol %-DOPCGUV (i.e., 30%DPOPE/70%DOPC-GUV) in 2% (w/v) PEG 6K aqueous solution induced by addition of 100 µM La3+ solution through a micropipet near the GUV. During the addition of La3+, the two GUVs became associated with each other (Figure 1B(1)), followed by a gradual increase in the contact area between these GUVs (Figure 1B(2,3)). Further addition of La3+ induced membrane fusion between these GUVs to produce a larger spherical GUV (Figure 1B(4)). Addition of 1 mM La3+ also induced membrane fusion between these GUVs, but 10 µM La3+ did not induce membrane fusion or membrane association. At g30 µM La3+, membrane fusion occurred. PEG-6K was added to reduce breakage of GUV during its interaction with La3+. The efficiency of the membrane fusion depended on the concentration of DPOPE in the DPOPE/DOPC membrane: at e10 mol % DPOPE, there was no fusion; at 20 mol % DPOPE, 50% of associated GUVs fused; at g30 mol %, 100% of associated GUVs fused. During membrane fusion between two DPOPE/DOPC-GUVs, the total volume of the two GUVs remained constant; i.e., the sum of the volumes of the two original vesicles (V1 + V2) was equal to the volume of the fused vesicle (V3). However, total surface area was decreased by membrane fusion; i.e., the surface area of the fused vesicle (A3) was less than the sum of the surface areas of the two original vesicles (A1 + A2) (Figure 1D). In the above experiments, we continued to add a given concentration of La3+ solution near the GUV through the micropipet slowly, but at the same time the diffusion of La3+ from the vicinity of the GUV into the bulk phase occurred. Thereby, we observed these phenomena at the steady-state condition of La3+ concentration near the GUV, not at the equilibrium condition.11 Therefore, the La3+ concentration near the GUV at the steady-state condition in our experiments may be a little lower than that of the La3+ solution in the micropipet. When we used method 2 (the two-pipet method) to bring GUVs into contact, similar membrane fusion between DPOPE/DOPC-GUVs occurred (Figure 1E). Membrane fusion occurred under the same conditions (La3+ concentration and DPOPE concentration) as described above in the membrane fusion using method 1 (Figure 1B). A detailed process of membrane fusion was the same as those observed in method 1, which is described below (see Figure 2). To elucidate the mechanism of the La3+-induced membrane fusion, we investigated the process of membrane fusion. Figure 2 shows, in detail, the steps in 100 µM La3+-induced membrane fusion of 30% DPOPE/70% DOPC-GUV. In Figure 2A, at 1/30 s, two GUVs were strongly associated with each other, and a partition membrane composed of two bilayers of each GUV had formed between the GUVs. At 2/30 s, the partition membrane had suddenly broken at one end (i.e., dis-

5162

Langmuir, Vol. 20, No. 13, 2004

Figure 1. (A) Association of DOPC-GUV induced by 100 µM La3+. The time after starting injection of La3+ solution through the micropipet is (1) 0 s, (2) 2 s, and (3) 4 s. The scale is the same as (B). (B) Membrane fusion of 30% DPOPE/70% DOPC-GUV induced by 100 µM La3+. The time after starting injection of La3+ solution through the micropipet is (1) 0 s, (2) 2 s, (3) 4 s, and (4) 6 s. (C) A scheme of diameter, surface area, and volume of two GUVs and the resulting large fused GUV. (D) Relationship between the ratio of area of the fused GUV (A3) to the sum of the area of the two small GUVs (A1 + A2) and the ratio of radius of the two small GUVs (R1/R2). (O) 100 µM La3+ and (b) 200 µM La3+. A theoretical curve was calculated based on the fact that the total volume of the two small GUVs (V1 + V2) was equal to that of the fused vesicle (V3). (E) Membrane fusion of 30% DPOPE/70% DOPC-GUV induced by 100 µM La3+. One GUV was supported at the tip of a micropipet using light aspiration. The time after starting injection of La3+ solution through the other micropipet is (1) 0 s, (2) 15 s, (3) 20 s, and (4) 22 s. All bars in the pictures (B, E) correspond to 20 µm.

connected from the GUV membranes). Subsequently, the cross sectional length of the partition membrane gradually decreased, finally forming a small structure that was difficult to define. In some cases of membrane fusion, after the partition membrane suddenly broke, it curled to form a smaller spherical vesicle (Figure 2B). The area of the

Letters

partition was almost equal to the decrease in the total surface area (At) at the membrane fusion. With two GUVs arranged perpendicular to the focus plane of the microscope during the membrane fusion, we were able to directly observe changes in the partition membrane (Figure 2C). In Figure 2C, at first (Figure 2C(1)), two GUVs were arranged parallel to the focus plane. We then moved the GUVs into the perpendicular arrangement, using the micromanipulator (Figure 2C(2)). At 3/30 s, a lower part of the partition membrane was already broken and out of focus, but an upper part of the partition membrane was connected to the membrane of the GUV and remained in focus. The area of the upper part of the partition membrane then decreased gradually. At 37/30 s, only a small area of the partition membrane remained. Finally, at 2 s, the partition membrane was disconnected completely (Figure 2C(4)). The results shown in Figure 2 clearly indicate that, at the membrane fusion, a partition membrane was suddenly broken at one end, followed by a gradual increase in the area of this breakage from the site of the first breakage, and ending with complete disconnection of the partition membrane from the GUV to form the smaller vesicle. Results of the internal contents mixing experiment are shown in Figure 3. The left GUV contained the watersoluble fluorescent probe FITC-dextran, and the right GUV contained the lipid fluorescent probe BODIPY-chol. In Figure 3B(3), two GUVs were strongly associated with each other, and a partition membrane had formed between them. In Figure 3B(4), the upper side of the partition membrane was suddenly broken and the FITCdextran began to diffuse from the left GUV to the right GUV. In Figure 3B(6), the FITC-dextran had completely diffused throughout the fused GUV. In Figures 3B(4) and 3B(5), a curled partition membrane can be seen at the lower area. These images show that there was no leakage of internal contents during the membrane fusion. On the basis of these results, we propose a mechanism of La3+-induced membrane fusion of GUVs (Figure 4). After association of two GUVs, addition of La3+ near the GUVs increases the density of La3+ bound in the outer monolayer membrane facing the buffer, increasing the lateral compression pressure of the monolayer membrane (Figure 4(1)). This induces fusion between the outer monolayer membranes at the edge of the partition membrane (Figure 4(2)), if the membrane contains a high concentration of DPOPE, because such a membrane favors formation of a negative high curvature.18 As a result, chain packing at the edge of the partition membrane (i.e., at the interstitial hydrocarbon region indicated by a green triangle in Figure 4) is destabilized, causing breakage of the membrane at one site on the edge (Figure 4(3)). Then, the area of this breakage site gradually increases from the site of the first breakage, eventually causing the partition membrane to separate from the GUVs, and membrane fusion of the two GUVs is completed simultaneously. The separated partition membrane is a sheet in water, and therefore acyl chains face water at its side. This destabilizes the partition membrane, causing it to curl into a smaller liposome inside the large fused GUV (Figure 4(4)). Here, we compare our new model of the membrane fusion (we call it “the partition breakage model”) with the stalk model and the defect model to elucidate their difference. In the partition breakage model, the first stage of the membrane fusion is merging of the outer monolayer (18) Kinoshita, K.; Li, S. J., Yamazaki, M. Eur. Biophys. J. 2001, 30, 207.

Letters

Langmuir, Vol. 20, No. 13, 2004 5163

Figure 2. Phase contrast images of intermediate structures of membrane fusion of 30% DPOPE/70% DOPC-GUV induced by addition of 100 µM La3+. Numbers under the photographs indicate time (seconds). Figure (C)-2 shows detail of the process between panels 3 and 4 of (C)-1. Arrows indicate partition membrane. The bar is equal to 10 µm.

membrane at the fusion site, which is the same as in the stalk model, but in the defect model the merge of the monolayer is not always necessary. In the stalk model, at the second stage of the membrane fusion, the stalk is formed at the fusion site and then the hemifusion intermediate is formed, but in the partition breakage model no stalk and no hemifusion intermediate are formed. At the next stage of the stalk model, the so-called fusion

pore is formed and its size increases with time, which is a similar phenomenon to the breakage of the partition membrane and the pore increases with an increase in the area of the breakage site of the partition membrane in the partition breakage model. The large difference between the stalk model and the partition breakage model is the change of the area of membrane after the fusion; in the stalk model the membrane area does not change, how-

5164

Langmuir, Vol. 20, No. 13, 2004

Letters

Figure 3. Phase contrast image (A) and fluorescence microscope image (B) of membrane fusion of 30% DPOPE/70% DOPC-GUVs induced by addition of 100 µM La3+. The left GUV contains FITC-dextran, and the right GUV contains BODIPY-chol. Arrows indicate the partition membrane. Bar is 20 µm.

Figure 4. Scheme of the proposed mechanism of La3+-induced membrane fusion of DPOPE/DOPC-GUVs. A detailed description is provided in the text. Red arrows indicate the La3+-induced lateral compression pressure of the membranes. The green triangle indicates the interstitial hydrocarbon region where free energy of chain packing is very large.

ever, in the partition breakage model the total membrane area of vesicles decreases due to the loss of the partition membranes from the fused vesicle. Therefore, to verify the model of membrane fusion experimentally, we should investigate the change of the membrane area of vesicles after membrane fusion and also existence of a new smaller vesicle inside the fused larger vesicle after the fusion.

Confirming the proposed mechanism of the membrane fusion, i.e., the partition breakage model, requires more experimental and theoretical investigation. The present findings on the La3+-induced membrane fusion of GUVs provide new and useful information about the process of membrane fusion. LA049681S