Kinetics of the Adhesion of DMPC Liposomes on a Mercury Electrode

Sep 14, 2006 - ... show the pathway of contact, hemifusion, flickering fusion pore formation, ..... Freeze-fracture electron microscopy revealed that ...
0 downloads 0 Views 203KB Size
Download PDF
Langmuir 2006, 22, 10723-10731

10723

Kinetics of the Adhesion of DMPC Liposomes on a Mercury Electrode. Effect of Lamellarity, Phase Composition, Size and Curvature of Liposomes, and Presence of the Pore Forming Peptide Mastoparan X† Victor Agmo Herna´ndez and Fritz Scholz* Institut fu¨r Biochemie, UniVersita¨t Greifswald, 17489 Greifswald, Soldmannstr. 23, Germany ReceiVed April 5, 2006. In Final Form: July 31, 2006 Liposomes suspended in aqueous electrolyte solutions can adhere at mercury electrodes. The adhesion is a complex process that starts with the docking and opening and leads to a spreading, finally resulting in the formation of islands of adsorbed lecithin molecules. The adhesion process can be followed by chronoamperometry, and a detailed analysis of the macroscopic and microscopic kinetics can be performed yielding rate constants and activation parameters. By using giant unilamellar liposomes and multilamellar liposomes, the effect of lamellarity and liposome size could be elucidated for liposomes in the liquid crystalline, gel, and superlattice phase states. Below the phase transition temperature, the time constant of opening of the liposomes (i.e., the irreversible binding of the lecithin molecules on the preliminary contact interface liposome|mercury and the therewith associated disintegration of the liposome membrane on that spot) is shown to be strongly size dependent. The activation energy, however, of that process is size independent with the exception of very small liposomes. That size dependence of time constants is a result of the size dependence of the initial contact area. The time constant and the activation energies of the spreading step exhibit a strong size dependence, which could be shown to be due to the size dependence of rate and activation energy of pore formation. Pore formation is necessary to release the solution included in the liposomes. This understanding was corroborated by addition of the pore inducing peptide Mastoparan X to the liposome suspension. The obtained results show that electrochemical studies of liposome adhesion on mercury electrodes can be used as a biomimetic tool to understand the effect of membrane properties on vesicle fusion.

Introduction Liposomes are frequently used to study the behavior of biological membranes and as biomimetic systems to understand the processes of vesicular transport in living systems.1-4 Zimmerberg and co-workers,5 for example, have reported that the energy barriers of lipid bilayer fusion in exocytosis, viral fusion, and trafficking are mainly the same as those for the fusion of phospholipid membranes. Related to such observations, Lee and Lentz have simulated the mechanism of secretory and viral fusion through the fusion of small unilamellar liposomes.6 Giant unilamellar vesicles (GUVs) are often used in membrane-related studies because of their resemblance to actual cell membranes and because, if appropriately dyed, they may be studied directly by microscopy and micropipet aspiration techniques.7-11 Studies concerning shape and phase transitions,7,8 determination of mechanical properties,9-11 and raft formation in vesicles with †

Part of the Electrochemistry special issue. * Corresponding author. Tel.: +49-(0)3834-86-4450. Fax: +49-(0)383486-4451. E-mail: [email protected]. (1) Arnold, K. Molekulare Mechanismen der Membranfusion; Akademie Verlag: Berlin, 1994. (2) Hellberg, D.; Scholz, F.; Schubert, F.; Lovric, M.; Omanovic, D.; Agmo Hernandez, V.; Thede, R. J. Phys. Chem. B 2005, 109, 14715. (3) Evans, D. F.; Wennerstro¨m, H. The colloidal domain: where physics, chemistry and biology meet, 2nd ed.; Wiley-VCH: New York, 1999. (4) Lipowsky, R. Encyclopedia of applied physics; Wiley-VCH: New York, 1998; Vol. 23, p 199. (5) Zimmerberg, J.; Chernomordik, L. V. AdV. Drug DeliVery ReV. 1999, 38, 197. (6) Lee, J.; Lentz, B. R. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 9274. (7) Bagatolli, L. A.; Gratton, E. Biophys. J. 1999, 77, 2090. (8) Needham, D.; Evans, E. Biochemistry 1988, 27, 8261. (9) Evans, E.; Rawicz, W. Phys. ReV. Lett. 1990, 64, 2094. (10) Olbrich, K.; Rawicz, W.; Needham, D.; Evans, E. Biophys. J. 2000, 79, 321. (11) Rawicz, W.; Olbrich, K. C.; McInstosh, T.; Needham, D.; Evans, E. Biophys. J. 2000, 79, 328.

coexisting fluid domains12-13 have been carried out with GUVs. Experiments with spherical GUVs of different radii are promising to study size dependent properties. Several methods have been developed in order to obtain GUVs. One of these methods consists of exposing dried crystalline lipid films to aqueous solutions for long periods of time (up to 48 h).14 A variation of this method uses low ac fields and high temperatures to produce giant liposomes in pure water.15 At certain conditions, monodisperse giant vesicles can be produced this way.16 Repetitive freezethaw cycles have been reported to produce also GUVs.17 Another method, developed by Moscho and co-workers,18 allows the preparation of GUVs in high yield in just 2 min, although the resulting size distribution is quite scattered. A disadvantage of working with GUVs is that they occupy slightly higher energy levels than multilamellar vesicles (MLVs) and, therefore, aggregate and fuse more rapidly to MLVs than the small unilamellar vesicles (SUVs).18 The interaction of liposomes with solid surfaces has been addressed in several papers;19-20 however, here we are focusing on the interaction with a charged mercury surface, as this allows time-resolved studies of the interaction process on a microsecond to millisecond scale. In this manuscript, we give further proof that electrochemical tools are excellent to (12) Baumgart, T.; Hess, S. T.; Webb, W. W. Nature 2003, 425, 821. (13) Baumgart, T.; Das, S.; Webb, W. W.; Jenkins, J. T. Biophys. J. 2005, 89, 1067. (14) Mueller, P.; Chien, T. F.; Rudy, B. Biophys. J. 1983, 44, 375. (15) Angelova, M. I.; Soleau, S.; Meleard, P.; Faucon, J. F.; Bothorel, P. Prog. Colloid. Polym. Sci. 1992, 89, 127. (16) Taylor, P.; Xu, C.; Fletcher, P. D. I.; Paunov, V. N. Phys. Chem. Chem. Phys. 2003, 5, 4918. (17) Traikia, M.; Warscawski, D. E.; Recouvreur, M.; Cartaud, J.; Devaux, P. F. Eur. Biophys. J. 2000, 29, 184. (18) Moscho, A.; Orwar, O.; Chiu, D. T.; Modi, B. P.; Zare, R. N. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 11443. (19) Teschke, O.; de Souza, E. F. Langmuir 2002, 18, 6513. (20) Martı´, M.; Barsukov, L. I.; Fonollosa, J.; Parra, J. L.; Sukhanov, S. V.; Coderch, L. Langmuir 2004, 20, 3068.

10.1021/la060908o CCC: $33.50 © 2006 American Chemical Society Published on Web 09/14/2006

10724 Langmuir, Vol. 22, No. 25, 2006

Herna´ ndez and Scholz

study the interaction of particles with electrodes, as it has been demonstrated also in several publications of Zˇ utic´ et al.21,22 and Compton et al.23-27.

Detection of Liposomes on a Mercury Electrode and Determination of the Kinetic Parameters of Adhesion (a) Macroscopic Kinetics. Previously,2,28 we have shown that liposomes adhere on mercury electrodes, and this process can be followed by chronoamperometric measurements. The basis of these measurements is the specific structure and the properties of the electrochemical double layer at electrodes.29 The adhesion of a liposome leads to the formation of a lecithin island on the electrode with a local charge density that differs considerably from that of the original electrode-aqueous solution interface. This causes capacitive spikes. Similar studies have been performed before to study different adhesion or contact events.30 In our case, the overall charge of one peak (Qlip) is related to the surface area of the lecithin island (Aisland) by the relation

Aisland ) Qlip(∆q)-1

(1)

where ∆q is the change of charge density between the bare electrode (i.e., the mercury|aqueous electrolyte interface) and the mercury covered with a lecithin monolayer. Qlip is the total charge displaced by one liposome (i.e., the overall charge underneath one peak). If the area occupied by a single lecithin molecule (Alecithin) is known, then the number of molecules that form one liposome can be calculated

nlecithin ) Aisland(Alecithin)-1

(2)

In these experiments, only peaks with a total charge exceeding 1 pC were easily detectable. Hence, the number of detectable events depends on ∆q and thus also on the electrode potential. For low values of ∆q [e.g., at potentials near to the point of zero charge (pzc)], only the largest liposomes will prompt a detectable charge effect. In the case of a large ∆q, smaller liposomes will also be detectable, and therefore, the peak frequency will increase. Optimal values of ∆q are found at potentials quite distant from the potential of the pzc but not too near to the desorption potentials. At potentials negative to the pzc, positive peaks result, and at potentials positive to the pzc, negative peaks result. The peak frequency has a maximum at -1 V and at -0.2 V, at least for the electrolyte solutions used in the experiments. At more negative or positive potentials respectively, the peak frequency decreases because the interaction of mercury with the electrolyte ions and with water becomes so much stronger than the hydrophobic interaction with the tails of the DMPC molecules that no adsorption can take place. Measuring the temperature dependence of the peak frequency gives access to the activation energy of (21) Ivosˇevic´, N.; Tomaic´, J.; Zˇ utic´, V. Langmuir 1994, 10, 2415. (22) Ivosˇevic, N.; Zˇ utic´, V.; Tomaic´, J. Langmuir 1999, 15, 7063. (23) Maisonhaute, E.; White, P. C.; Compton, R. G. J. Phys. Chem. B 2001, 105, 12087. (24) Maisonhaute, E.; Brookes, B. A.; Compton, R. G. J. Phys. Chem. B 2002, 106, 3166. (25) Banks, C. E.; Rees, N. V.; Compton, R. G. J. Phys. Chem. B 2002, 106, 5810. (26) Rees, N. V.; Banks, C. E.; Compton, R. G. J. Phys. Chem. B 2004, 108, 18391. (27) Davies, T. J.; Lowe, E. R.; Wilkins, S. J.; Compton, R. G. Chem. Phys. Chem. 2005, 6, 1340. (28) Hellberg, D.; Scholz, F.; Schauer, F.; Weitschies, W. Electrochem. Commun. 2002, 4, 305. (29) Stojek, Z. In Electroanalytical Methods; Scholz, F., Ed.; Springer: Berlin, 2005; p 3. (30) Scholz, F.; Hellberg, D.; Harnisch, F.; Hummel, A.; Hasse. U. Electrochem. Commun. 2004, 6, 929.

the overall adhesion process. The term “overall adhesion process” refers to the macroscopic kinetics of the process (i.e., to the number of adhesion events per time and surface area unit). The relation between the peak frequency (f) and the rate of adhesion (J) in mol s-1 cm-2 is given by

J ) f(ASMDENA)-1

(3)

where ASMDE is the electrode area and NA is the Avogadro constant. The unit mol refers to the number of liposomes that adhere. Using this equation, we may study the macroscopic kinetics of adhesion (i.e., the kinetics of the entire ensemble of liposomes in the suspension). Similar measurements have been done before by Wegener and co-workers31 using the quartz crystal microbalance (QCM) technique, but avidin had to be preadsorbed on the electrode surface and the liposomes had to be functionalized. Our technique has the advantage that it allows the kinetics to be measured on a bare mercury electrode surface and without the need to modify the liposome with additional compounds. It is also an important advantage that the surface of a mercury electrode can be renewed in an extremely reproducible way by dispatching the mercury droplet and formation of a new stationary mercury drop. Thus, repetitive measurements are possible under completely identical surface conditions. It goes without saying that mercury provides an ideally isotropic surface, what adds tremendously to the reproducibility of data. (b) Microscopic Kinetics. Beside the possibility to count the peaks for studying the macroscopic kinetics, the single peaks can be analyzed to access most interesting information on the microscopic kinetics (i.e., the kinetics of single adhesion events). It has been shown before that integration of the current-time transients allows fitting the resulting charge-time transient by the following simple equation:2

Q(t) ) Q0 + Q1(1 - exp-t/τ1) + Q2(1 - exp-t/τ2)

(4)

The first term on the right side has been interpreted as being due to the docking of the liposome on the mercury surface. Until now, this process is too fast for the employed equipment to be resolved in time. The docking is understood as a step in which probably only a few lecithin molecules that are in a turnedaround position (i.e., with the lipophilic chain pointing outward) are anchoring the liposome on the mercury surface. The second term is understood as being caused by the opening of the liposome and the third term shows the spreading of the lecithin island on the mercury surface. Opening of the liposome means here the irreversible restructuring of the liposome-mercury interface resulting in a liposome that is bound via a layer of lecithin molecules at that interface. It does not mean the formation of an open pore to the outer solution, as this is a step of the spreading process (see further down). The parameters τ1 and τ2 are the time constants of the opening and the spreading process, respectively. By analyzing peaks that have been measured at different temperatures, the activation energies of the two time dependent steps of the liposome adhesion can be determined. In our previous work we have proposed the following theoretical scheme:2 K0

k0

K1

k1

K2

k2

K3

L {\} L′ 98 LD {\} LD* 98 LO {\} LO• 98 LE {\} LA (5) where the symbols have the following meaning: L, free liposome; L′, liposome in contact with mercury surface; LD, docked liposome (“one lecithin molecule adsorbed on mercury”); LD*, docked (31) Reiss, B. A.; Janshoff, A.; Steinem, C.; Seebach, J.; Wegener, J. Langmuir 2003, 19, 1816.

Kinetics of the Adhesion of DMPC Liposomes

Langmuir, Vol. 22, No. 25, 2006 10725

Figure 1. Proposed mechanism of the adhesion of liposomes on a mercury electrode (Copyright 2005, American Chemical Society, reprinted with permission from ref 2).

liposome in deformed state; LO, opened liposome; LO*, adsorbed open liposome; LE, “deconvoluted” liposome (i.e., lecithin island that is not yet adsorbed); LA, island of adsorbed lecithin molecules. A model of that mechanism is shown in Figure 1. It must be noted here that not all steps formulated in eq 5 have a real meaning as some of them will proceed simultaneously; however, they are necessary for deciphering the mechanism. Solving the appropriate differential equations, the following expression has been derived

A ) A0 + A1(1 - e-t/([1+K1]/k1K1)) + A2(1 - e-t/([1+K2]/k2K2)) (6) A comparison of eqs 6 and 4 reveals the meaning of the time constants τ1and τ2

τ1 )

1 + K1 k1K1

(7)

τ2 )

1 + K2 k2K2

(8)

The thermodynamic equilibrium constants (K1 and K2) of adsorption decrease with increasing temperature. When K1 and K2 are smaller than 1 at higher temperatures, apparent negative activation energies are possible, whereas for lower temperatures, they may be larger than 1, which may lead to the result that they

cancel in eqs 7 and 8 and thus allow the observation of positive activation energies. In the present work, we have studied in detail how the kinetics of adhesion depends on the lamellarity, phase composition, size, and curvature of the DMPC liposomes and on the presence of the pore-inducing peptide Mastoparan X. Experimental Details 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) was a donation of the Lipoid GmbH, Ludwigshafen, Germany, and it was used without further purification. The electrolyte potassium chloride Suprapur was obtained from Merck, Darmstadt, Germany. Mastoparan X was purchased from Bachem (Bubendorf, Switzerland). The water used was prepared from deionized water by distilling it once with alkaline permanganate solution to oxidize traces of organic compounds followed by another distillation. Before measuring, the suspensions were deaerated for 20 min with high-purity nitrogen. Electrochemical measurements were performed with an Autolab PSTAT 10 (Eco Chemie, Utrecht, Netherlands) for the pure lecithin GUVs and with an Autolab PGSTAT 12 (Eco Chemie, Utrecht, Netherlands) to study the effect of lytic peptides. Both of them were interfaced to a P4 PC in conjunction with an electrode stand VA 663 (Methrom, Herisau, Switzerland). A multimode mercury electrode was used as working electrode, a platinum rod served as auxiliary electrode and an Ag|AgCl (3 M KCl, E ) 0.208 V vs SHE) electrode was used as reference electrode. The surface area of the mercury drop was 0.48 mm2, as determined by weighing 50 drops. The interaction of the liposomes with the electrode surface was recorded

10726 Langmuir, Vol. 22, No. 25, 2006 in chronoamperometric measurements performed within 1 s with sampling each 50 µs or within 0.040 s with sampling each 1.333 µs in the case of high-resolution measurements. For counting and analysis of the adhesion peaks, the software “Signal Counter” was used.32 The solutions were thermostated with an accuracy of (0.2 K. The distribution of liposome diameters was detected by lightscattering measurements with a Zetasizer 3000 HS (Malvern Instruments, Herrenberg, Germany). In all electrochemical experiments, five measurements were performed at each temperature. The peak frequency reported is the average of these five measurements. For calculation of the activation energy of the macroscopic process, the peak frequency was determined at 30 different temperatures. For the microscopic kinetics, 10-30 peaks for each temperature were integrated and fitted. At least 10 different temperatures have been chosen for these studies. For liposome suspensions containing Mastoparan X, experiments were repeated two times with new suspensions in a similar way as for the Mastoparan X-free suspensions. Giant Unilamellar Vesicles. The technique reported by Moscho et al.18 for a rapid preparation of GUVs was followed with some modifications. This technique has been selected because of the rather hostile environment (high ionic force) in which we had to suspend the vesicles. A total of 3 mg of lecithin were dissolved in a mixture containing between 2.1 and 4.2 mL of chloroform and 210-440 µL of methanol. A total of 30 mL of a 0.1 M solution of KCl were then added carefully by pouring along the flask walls. Then, the organic solvent was removed with the help of a rotary evaporator (Laborota 4000, Heidolph, Nu¨rnberg, Germany) using a Rotavac control pump (Heidolph, Nu¨rnberg, Germany) at 40 °C and a final pressure of 10 mbar. The rotation speed was 40 rpm. This way, a clear suspension containing a high yield of GUVs could be obtained. Temperature variation experiments (with 1 K increments) were performed by slowly cooling the freshly prepared liposomes. After reaching the phase transition temperature, the suspension was cooled to 2 °C, and the temperature was then slowly raised in 1 K increments and the adhesion events were measured at each temperature. This temperature program had the objective to minimize the shape changes,33-34 especially when the lecithin was in the liquid crystalline phase, and to be able to observe the different transitions in the gel phase. Calculations of the average diameter of the produced GUVs based on our electrochemical measurements gave a value of 2.14 µm, and the light scattering measurements gave an average diameter of 1.999 µm. This is a convincing indication that we have indeed produced GUVs.

Results and Discussion Giant Unilamellar Vesicles. Figure 2 shows one adhesion signal (current versus time) and the integrated signal as charge versus time trace together with the curve fitted to the experimental data by using eq 4. The use of GUVs remarkably reduced the data scattering that we had in previous experiments with MLVs2 (Figure 3). Table 1 shows a comparison of the kinetic parameters obtained with MLVs and GUVs at temperatures above and below the phase transition temperature (PTT). As it can be seen, the values of the activation energies of opening and spreading differ for both cases, suggesting that the lamellarity of the liposomes plays an important role in the mechanism of adhesion. This may be related to the formation of defects in the lipid membranes as a function of lamellarity, which has been already reported by Ku¨nneke and co-workers.35 They observed very clear breakthrough events for tips approaching liposomes adsorbed on mica when the liposomes had unilamellar bilayer membranes. For liposomes with multilamellar membranes no breakthrough events (32) Designed by Dario Omanovic´. Rudjer Boskovic Institute, Zagreb, Croatia. (E-mail: [email protected]). (33) Berndl, K.; Ka¨s, J.; Lipowsky, R.; Sackmann, E. Europhys. Lett. 1990, 13, 659. (34) Sackmann, E. FEBS Lett. 1994, 346, 3. (35) Ku¨nneke, S.; Kru¨ger, D.; Jahnshoff, A. Biophys. J. 2004, 86, 1545.

Herna´ ndez and Scholz

Figure 2. Adhesion signal (current versus time; top), and the integrated signal (bottom) as charge versus time trace together with the curve fitted to the experimental data by using eq 4. Signal obtained from a 0.1 g/L DMPC suspension in 0.1 M KCl at a potential of -0.9 V and a temperature of 30 °C.

have been observed and the adhesion of the liposomes on the mica was weak. The results shown in Table 1 are in agreement with these findings. The time constants for GUVs are significantly smaller than for MLVs, and the activation energies, at least in the gel phase, are also smaller for the unilamellar vesicles. For liposomes with membranes in the liquid crystalline state, negative apparent activation energies were observed for the macroscopic kinetics and the spreading step of the microscopic kinetics. This may be explained by small equilibrium constants of adsorption (see eqs 7 and 8). Figure 4 clearly shows that there is a temperature range, i.e., from 13 to 23 °C, in which the peak frequency clearly decreases with temperature. Needham and Evans8 have reported the existence of a phase pretransition in this temperature range where a superlattice phase more fluidic than the normal gel phase is formed. In this region, we observe a negative activation energy for the overall process. It is a typical feature of the results shown in Figure 4 that the data scattering in the pretransition range is remarkably small, compared to the data for temperatures above the phase transition temperature. This indicates that at the PTT really something important happens, and therefore, we do not believe that it would be reasonable to treat the data for the pretransition range and for T > PTT together. Hence, the different slopes for these two temperatures are obviously real and not an artifact. Also in the case of incorporation of Mastoparan X, we

Kinetics of the Adhesion of DMPC Liposomes

Langmuir, Vol. 22, No. 25, 2006 10727

Figure 4. Arrhenius plot for the overall adhesion kinetics.

Figure 3. Dependence of τ1 (top) and τ2 (bottom) on the total charge of adhesion (Qlip) at 40 °C for multillamelar liposomes (b) and unilamellar vesicles (0). Table 1. Comparison of Kinetic Data Obtained with MLVs and GUVsa parameter T < PTT Ea overall Ea openingb Ea spreadingb τ1 at 10°Cb τ2 at 10°Cb T > PTT Ea overall Ea opening Ea spreading τ1 at 40°C τ2 at 40°C

MLVs

GUVs

89 kJ/mol 73 kJ/mol 45 kJ/mol 3.57 × 10-4s 1.2 × 10-3 s

59.3 kJ/mol 49.6 kJ/mol 33.8 kJ/mol 2.53 × 10-4 s 1.05 × 10-3 s

-10 kJ/mol 9 kJ/mol -15 kJ/mol 7.42 × 10-5 s 7.61 × 10-4 s

-9.3 kJ/mol 34.7 kJ/mol -9 kJ/mol 5.3 × 10-5 s 5.83 × 10-4 s

a Data for MLVs taken from Hellberg et al.2 b At temperatures below the PTT, average values are reported because at these temperatures a size dependence has been observed. Above the PTT, no size dependence could be seen.

have seen two distinctly different slopes for these temperature ranges (not shown here). Size Dependence of the Time Constants of the Microscopic Kinetics of Adhesion of Gel-Phase Liposomes. For temperatures below the PTT, there is a slight, however distinct, dependence (cf. Figure 5) of τ1 and τ2 on the total charge of adhesion (and thus on the size of the liposome). For MLVs only, a scattering of the data has been observed.2 Although both time constants show a similar dependence on the total charge of adhesion, it would not be reasonable to assume the same cause in both cases.

Figure 5. Dependence of τ1 (top) and τ2 (bottom) on Qlip at 10 °C for multillamelar liposomes (b) and unilamellar vesicles (0).

The kinetics of each of the processes, according to the model (Figure 1), follows different mechanisms. For the kinetics of opening, the most important is the dynamics of molecules in the small part of the liposome surface that touches the mercury surface, whereas for the spreading, the dynamics of unfolding the liposome membrane is important. Later on, we shall discuss the latter in more detail. The dependence of τ1 on the total charge of adhesion can be fitted to the following equation:

τ1 ) a1Qlip1/2 + a2

(9)

The linearized graph is shown in Figure 6 (this figure shows just one plot, and it has to be stressed that also for other temperatures

10728 Langmuir, Vol. 22, No. 25, 2006

Herna´ ndez and Scholz

Figure 6. Dependence of τ1 on the square root of Qlip at 10 °C.

Figure 8. Dependence of τ1 on r at 23 °C.

Figure 9. τ1 as a function of Ac for different arbitrary Dc values at 23 °C.

Figure 7. Activation energy of opening as a function of the total charge of adhesion for liposomes in the gel phase.

below the PTT a linear regression gave the best fit). As the time constants are size dependent, the activation energy of opening should be calculated separately for different vesicles sizes. Therefore, the data for all temperatures below PTT were fitted according to eq 9, and different values of the activation energy were calculated taking the fitted values of τ1 for different liposome sizes. Figure 7 shows the resulting dependence of the activation energy of opening on Qlip. On a first glance, this looks like a contradiction since Figure 6 shows that the opening is slower for larger liposomes, whereas in Figure 7, it is shown that the energy barrier of the opening process is independent of size, except for the very small liposomes by which, counterintuitively, the activation energy reaches a very high value. The dependency of τ1 on the total charge of adhesion resembles that of the radius of the liposome, assuming a spherical configuration

r)

( ) Qlip 8π∆q

1/2

+ LL

(10)

where LL is the length of a lecithin molecule. Therefore, according to the experimental data, a linear relationship exists between τ1 and the radius of the liposome (Figure 8). Extrapolating that line to r ) 0, the time constant of opening of an infinitely small liposome can be estimated. The activation energy of such a process may be related to the energy of activation required for a single lecithin molecule to turn its polar head into the interior of a

membrane in order to be adsorbed on the electrode. Finding the intercept at different temperatures, the activation energy of that process could be estimated to be 93 kJ/mol. At the beginning of the opening step, there is a very small area of the liposome surface that is in contact with the electrode. That initial contact area Ac may be assumed to be the area of a cap taken from the spherical liposome at a distance Dc from the edge, and it would be, in terms of r

Ac ) 2πDcr

(11)

The relationship between τ1 and Ac is therefore a line with a slope dependent on Dc (Figure 9). The value of the later parameter must be very small, since just a small deformation of the liposome takes place in the beginning of the opening process. Some of the lecithin molecules forming the contact area have been already turned and adsorbed in the mercury surface during the previous “docking step” (see Figure 1). However, the large majority of the molecules is still showing their polar heads to the electrode and, therefore, have not yet been adsorbed and the surface charge has not been displaced (Figure 10a). This “polar-head-to-mercury” orientation of the molecules is strongly disfavored since the mercury surface is hydrophobic. However, the already adsorbed molecules will then act as nuclei for the growing of islands of adsorbed lecithin (Figure 10b). At the end of the opening process all molecules in the contact area have been adsorbed and spreading of the rest of the liposome has begun (Figure 10c). It has been reported that perturbations in the bilayer of the liquid crystalline phase propagate in the membrane.36 In the case of liposomes above the PTT, the turned-around molecules will diffuse laterally in the membrane, sliding on the mercury surface thus homog-

Kinetics of the Adhesion of DMPC Liposomes

Langmuir, Vol. 22, No. 25, 2006 10729

Figure 11. τ2 fitted as a function of Qlip at 17 °C.

liposome, only one nucleus may be present, and the adsorbed island may grow without limitation. In a large liposome, with a large contact area, many islands may start to grow at the same time, and that must lead to a serious problem: Since the growing of the islands needs the lecithin molecules to turn around, a process for which they need space, the simultaneous growing of islands will be impeded and thus slowed. This can explain the decrease of the rate constant of opening with increasing size of liposomes. It still needs an explanation why the opening of the smallest liposomes exhibits such large activation energies. An answer can be suggested on the basis of the previous arguments: For a very small liposome, it is possible that the area in contact with the electrode will not possess a sufficient number of turnedaround molecules to start the island growing process. The liposome will therefore need energy to flip over some more lecithin molecules in order to form a higher surface concentration of nuclei. For larger liposomes, there may be always a sufficiently high number of nuclei in contact with the mercury surface and, very probably, the surface concentration of such nuclei will be the same, independent of the size of the contact area. Therefore, the activation energy for liposomes larger than a critical size will be independent of the total charge of adhesion as shown in Figure 7. In the case of τ2, the size dependence can be fitted using an equation of the form (see Figure 11) Figure 10. Detailed scheme of the opening process showing the differences between the process in the gel and in the liquid crystalline phase.

enizing the contact area (Figure 10b). Also, as the molecules can rather freely glide laterally, the lecithin molecules forming the inner part of the membrane will easily displace the molecules between them and the electrode. The size of the contact area, and therefore of the liposome, will not affect the rate of opening, as Figure 3 suggests. On the other hand, the perturbations in the gel phase will not diffuse, and located islands of adsorbed molecules will grow (Figure 10b). The space available for the creation of new islands and for growing of the already formed will decrease with time. For the molecules in the inner layer, it will be rather difficult to move toward the mercury and to adsorb there. Assuming that the initial surface concentration of adsorbed molecules (i.e., the number of adsorbed molecules per unit of contact area) is independent of size of the liposomes, we will find more nuclei when the contact area is large. For a very small (36) May, S. Curr. Opin. Colloid Interface Sci. 2000, 5, 244. (37) Sens, P.; Safran, S. A. Europhys. Lett. 1998, 43, 95.

τ2 ) b1lnQlip + b2

(12)

(Figure 11 just shows one plot for 17 °C. It has to be stressed that, also for the temperatures 10, 13, 20, and 23 °C (plots not shown here), eq 12 allowed the best fitting.) The activation energy as a function of the liposome size is shown in Figure 12. Contrary to the activation energy of opening, that of spreading increases with size. It is interesting to see that for small liposomes (Qlip < 14 pC) the opening process has a larger activation energy, whereas for large liposomes, the spreading process needs a higher activation energy. The dependence shown in Figure 12 can be explained as follows: When a liposome has docked on the mercury and “opened” in the sense that the lecithin molecules on the contact have turned around and adsorbed on mercury, a spreading of the liposome needs the release of the inner solution. This, in principle, could be achieved via two mechanisms, either by diffusion of the inner water through the membrane, or by release through pores. If pore formation happens, and this is the most likely scenario for energetic reasons,10 the rate of spreading will be limited by that

10730 Langmuir, Vol. 22, No. 25, 2006

Herna´ ndez and Scholz

Figure 12. Size dependence of the activation energy of spreading.

process. There are many models to explain the formation of pores in membranes.37-39 Many of them agree in that the pore formation is favored by a large free energy per unit surface area, i.e., by a large surface tension (σ) that characterizes the membrane and disfavored by large values of the edge energy or line tension (Γ) arising from the lipid molecules that will come in contact with water upon pore formation.36,37According to that model, the energy barrier for the formation of a persisting pore is

δE ) πΓ2σ-1

Figure 13. Activation energy of spreading as a function of size for a suspension of liposomes containing 0.1 µM of Mastoparan X.

(13)

The surface free energy of the membrane arises largely from the bending energy and may be related to the curvature using the Helfrich equation40

1 σ ) σ0 + kc(J - J0)2 + khK 2

(14)

where kc is the bending modulus, kh is the Gaussian bending modulus, J is the total curvature (J ) 2r-1 for a spherical vesicle), K is the Gaussian curvature (K ) r-2 for a sphere), σ0 is the surface tension of a plane membrane, and J0 is the spontaneous curvature corresponding to the minimum of the interfacial free energy, and it is zero in the case of symmetrical bilayers. Therefore, the value of σ will be smaller for larger liposomes. On the other hand, the value of Γ depends only on the nature of the lipids and their interaction with water. It is independent of the liposome size. Then the energy barrier will increase with the size of the vesicles, in agreement with our results. A more recent model developed by Santangelo and co-workers41 considers the fluctuations in the membrane and concludes that for large liposomes it is even possible that the effective surface tension is negative, and therefore, it would act to prevent the formation of pores. For small liposomes, their results approach the model cited before. Although our observations are in good agreement with both models, we further tried to corroborate the involvement of pore formation in the mechanism of spreading by adding a specific pore forming agent to the liposome suspension. For that purpose, we have chosen a concentration of 0.1 µM of Mastoparan X, a tetradecapeptide found as a component of the wasp venom (38) Litser, J. D. Phys. Lett. 1975, 53A, 193. (39) Shillcock, J. C.; Seifert, U. Biophys. J. 1998, 74, 1754. (40) Helfrich, W. Z. Naturforsch. 1973, 28, 693. (41) Farago, O.; Santangelo, C. D. J. Chem. Phys. 2005, 122, 0449011.

Figure 14. Activation energy of opening as a function of size for a suspension of liposomes containing 0.1 µM of Mastoparan X.

produced by Vespa xanthoptera. This peptide has a lytic activity, even at very low concentrations, creating pores in the membrane without destroying it.42-44 Figure 13 shows the values of the activation energies for liposomes modified in this way. Clearly there is a very distinct decrease of the energy barrier needed to start the spreading process. These findings are a very strong support of the idea that pore formation is the rate determining step of liposome spreading. Interestingly, the opening activation energy of small liposomes was also diminished for liposomes suspended in solutions of the pore forming peptide. For larger liposomes, no effect was observed. A distinct size dependence of the activation energy of opening is observed in such solutions (Figure 14). As stated above, the opening process takes place only at the contacting area at the beginning of the adhesion. If there are pores in the membrane, they may occur also in such a contact area. In a small liposome, the area of a pore will represent a larger fraction of the total contact area than in a larger one. (42) Schwarz, G.; Arbuzova, A. Biochim. Biophys. Acta 1995, 1239, 51. (43) Whiles, J. A.; Brasseur, R.; Glover, K. J.; Melacini, G.; Komives, E. A.; Vold, R. R. Biophys. J. 2001, 80, 280. (44) Yu, K.; Kang, S.; Kim, S. D.; Ryu, P. D.; Kim, Y. J. Biomol. Struct. Dyn. 2001, 18, 595.

Kinetics of the Adhesion of DMPC Liposomes

Besides, according to our model, the fraction of the total area that comes in initial contact with the electrode is larger for small vesicles, and therefore, the relative number of pores in the contact area is also larger. Pores may work also as nuclei for the turning around of molecules, since the lecithin molecules at the edge are already partially turned around. Nay it is even perceivable that the complete mechanism of docking and opening may change when pores are present. This is a feasible explanation of the decrease of the activation energy of opening for smaller liposomes containing the pore-forming peptide.

Conclusions The use of GUVs instead of MLVs remarkably increased the accuracy of the experimentally determined kinetic data, as the scattering caused by different morphologies is reduced. The observations made so far allow us to correlate the kinetic data with physical properties of the membrane such as compressibility and bending energies. The electrochemical tool developed for the analysis of liposomes has proved to be excellent for acquisition of data that are related to the stability of lipid membranes.

Langmuir, Vol. 22, No. 25, 2006 10731

With the obtained data, a very detailed model of the mechanism of adhesion of liposomes on a mercury electrode can be proposed and experimentally supported. The different steps involved (docking, opening, and spreading) can be described both qualitatively and quantitatively. The rate determining steps of the two main microscopic processes of adhesion (i.e., the opening and the spreading of a liposome) were found to be the turning around of lecithin molecules, in the case of opening, and the formation of pores in the membrane, for the case of spreading. The adhesion of liposomes on the surface of a mercury electrode is kin to vesicle fusion, and thus, the obtained results may be helpful to gain a better understanding of these processes that are of fundamental importance for life. Acknowledgment. V.A.H. acknowledges provision of a DAAD-Conacyt scholarship. F.S. acknowledges financial support by Fonds der Chemischen Industrie. The authors gladly acknowledge provision of lecithin samples by Lipoid GmbH, Ludwigshafen, Germany. LA060908O