Molecular Exchange through the Vesicle Membrane of Siloxane

May 15, 2007 - ... are gradually swollen, and large multilamellar vesicles disappear. ... Sabine Himmelein , Nora Sporenberg , Monika Schönhoff , and...
0 downloads 0 Views 317KB Size
Download PDF
J. Phys. Chem. B 2007, 111, 6161-6166

6161

Molecular Exchange through the Vesicle Membrane of Siloxane Surfactant in Water/ Glycerol Mixed Solvents Yun Yan,*,† Heinz Hoffmann,*,† Alina Leson,‡ and Christian Mayer‡ BZKG, UniVersita¨t Bayreuth, Gottlieb-Keim-Strasse 60. D-95448 Bayreuth, Germany, and UniVersita¨t Duisburg-Essen, Fachbereich Chemie, D-45141 Essen, Germany ReceiVed: February 12, 2007; In Final Form: April 2, 2007

The effect of glycerol on the permeability of vesicle membranes of a siloxane surfactant, the block copolymer polyethyleneoxide-b-polydimethylsiloxane-polyethyleneoxide, (EO)15-(DMS)15-(EO)15, was studied with freeze-fracture transmission electron microscopy (FF-TEM) and pulsed-field gradient nuclear magnetic resonance (PFG-NMR) spectroscopy. The FF-TEM results show that, in pure water, the surfactant can form small vesicles with diameters of less than 25 nm, as well as a few multilamellar vesicles with diameters larger than 250 nm. Gradual substitution of water with glycerol to a glycerol content of 40% leads to significant structural transformations: small vesicles are gradually swollen, and large multilamellar vesicles disappear. A glycerol content of 60% results in the complete disintegration of the vesicles into membrane fragments. PFG-NMR measurements indicate that the vesicle membrane does not represent an effective barrier for water molecules on the NMR time scale; hence, the average residence time of water in the encapsulated state is below τb ) 2 ms. In contrast, the average residence time of glycerol molecules in the encapsulated state can be as large as τb ) 910 ms. The permeability of the vesicle membrane increases with increasing glycerol concentration in the solvent: At a concentration of 40%, the residence time τb is lowered to approximately 290 ms. After vesicle destruction at higher glycerol concentrations, a small glycerol fraction is still bound by membrane fragments that are formed after the disintegration of the vesicles.

1. Introduction Vesicles are well-known to be models for biological cell membranes because they have great potential for the uptake, encapsulation, and controlled release of water-soluble components.1,2 Therefore, the permeability of the vesicle membranes is of special interest.3-8 It not only represents an important parameter for any given combination of vesicle and encapsulated component but also provides important information on the molecular structures and physical characteristics of the vesicle wall. It has been reported that the permeability of the vesicle membrane can be affected by small organic molecules, which change the thickness and the polarity profile of the bilayers.9-11 As an example, ethanol has a significant influence on the permeation characteristics of polymer bilayer membranes.11 In practical applications, many surfactants are used in combination with glycerol. For instance, the formulations of pharmaceutical agents, personal care products, detergents, and foods often contain glycerol to improve sensory qualities.12 Therefore, it is important to study the permeability of vesicle membranes in water/glycerol mixed systems. The vesicle systems used in this study were prepared from a siloxane surfactant, the block copolymer polyethyleneoxide-bpolydimethylsiloxane-polyethyleneoxide, (EO)15-(DMS)15(EO)15, which is used as an additive in cosmetics and is commercially available under the name IM-22. In aqueous solutions, it can self-assemble into vesicles13 with the DMS segments coiled into the wall of the vesicle membrane.14-16 In * To whom correspondence should be addressed. E-mail: yun.yan@ wur.nl (Y.Y.), [email protected] (H.H.). † Universita ¨ t Bayreuth. ‡ Universita ¨ t Duisburg-Essen.

a previous work, we found that the gradual substitution of water by glycerol swells the LR phase of IM-22, i.e., the interlamellar distance increases with increasing glycerol content.17 The swelling of the LR phase in water upon addition of glycerol is explained by the matching of the refractive index of the solvent to the refractive index of the surfactant. The matching of the refractive index lowers the attraction between the bilayers, which allows them to swell to a larger separation. In this research, we focused on the influence of the glycerol content on the permeability of vesicle membranes of IM-22. The tool we used to study molecular exchange through the vesicle membrane was the combination of nuclear magnetic resonance echo experiments with pulsed-field gradients (PFGNMR spectroscopy). In a given vesicle system, PFG-NMR studies can differentiate the molecules in the continuous external aqueous phase from those in the encapsulated state inside the vesicles. At the same time, this technique allows one to observe the exchange process between the internal volume of the vesicles and the continuous external phase.18-22 2. Experimental Section Vesicle Preparation. For vesicle preparation, aqueous solutions of 5% polyethyleneoxide-b-polydimethylsiloxane-polyethyleneoxide, (EO)15-(DMS)15-(EO)15 (sold under the commercial name IM-22 and referred to as IM-22 in the following text), were prepared in a 100 mL flask by adding 95% water to a carefully determined amount of IM-22. Vesicles in water/ glycerol mixed solvents (20%, 40%, and 60% glycerol contents) were prepared by replacing the water by glycerol/water mixtures with volume ratios of 2:8, 4:6, and 6:4 (glycerol/water). All

10.1021/jp0711848 CCC: $37.00 © 2007 American Chemical Society Published on Web 05/15/2007

6162 J. Phys. Chem. B, Vol. 111, No. 22, 2007

Yan et al.

mixtures were homogenized by vortex mixing and allowed to equilibrate for 24 h at room temperature. Viscosities. The viscosity of the water/glycerol mixtures was obtained from P&G Chemicals as the original weight fraction of the glycerol. Then, the value was transformed into volume fraction by employing the corresponding density value. The densities of the water/glycerol mixtures were obtained by averaging the densities of water and glycerol at different volume ratios. The slight variation in partial molar volume caused by mixing was assumed to be negligible. This was verified by weighing water/glycerol mixtures of known volumes. According to the above information, different mass ratios for water/glycerol mixtures were transformed into different volume ratios. A master curve was developed using these viscosity-volume fraction relations. Then, the viscosities for 20%, 40%, and 60% glycerol solutions were estimated from the master curve. Freeze-Fracture Transmission Electron Microscopy (FFTEM). For FF-TEM experiments, a small amount of sample was placed on a copper disk of 0.1 mm thickness and covered with a second copper disk. The sample was frozen by plunging this sandwich into liquid propane that had been cooled with liquid nitrogen. Fracturing and replication were carried out in a freeze-fracture apparatus (Balzers BAF 400, Wiesbaden, Germany) at a temperature of -140 °C. Pt/C was deposited at an angle of 45°. The replica was examined in a CEM 902 electron microscope (Zeiss, Oberkochen, Germany). PFG-NMR Measurements. To study molecular exchange through the vesicle membrane, both water and glycerol molecules were traced simultaneously. The samples were filled into regular 5 mm NMR sample tubes and used for NMR experiments without further preparation. PFG-NMR experiments were performed following a general procedure that has been described elsewhere.18,19 All measurements were performed at 293 K on a Bruker Avance 400 spectrometer (Bruker AG, Karlsruhe, Germany) equipped with a BAFPA 40 gradient amplifier and a Bruker DIFF30 probe. The instrument was tuned to 400 MHz proton frequency, and gradient pulses were adjusted to gradient strengths G between 0 and 4 T/ m with δ ) 1.2 ms duration. A Hahn echo pulse program (90°-τ1-180°-τ1-echo) and a stimulated echo pulse program (90°-τ1-90°-τ2-90°-τ1echo) were used, each in combination with gradient pulses during each τ1 waiting period. The spacing ∆ between gradient pulses was varied between 6 and 15 ms for the observation of water (Hahn echo) and between 50 and 250 ms for the observation of glycerol (stimulated echo). Sets of PFG-NMR experiments were started under variation of the gradient strength G while the corresponding echo amplitudes I were determined. In the presence of free diffusion with a diffusion coefficient D, the decay of the echo intensity I with respect to the original value I0 (for G ) 0) can be related to the gyromagnetic ratio γ of the observed nucleus, the strength G and duration δ of the pulse gradients, the separation ∆ of the gradient pulses, and the self-diffusion coefficient D according to

I/I0 ) exp[-γ δ G D(∆ - δ/3)] 2 2

2

(1)

In a plot of ln(I/I0) vs the parameter γ2δ2G2(∆ - δ/3), such a decay will be represented by a straight line with a slope of -D in the case of free diffusion. In the presence of an encapsulated fraction of the observed molecules, this dependence becomes more complicated, as it reflects (i) the self-diffusion constant Da of the free fraction, (ii) the diffusion constant Db for the Brownian motion of the encapsulated fraction, and (iii) the molecular exchange between the free and encapsulated fractions. With the approximation that (∆ - δ/3) ≈ ∆ (for the glycerol

experiments, ∆ and δ/3 differ by at least 2 orders of magnitude), the resulting shape of the echo decay plot is given by the following set of equations20

I/I0 ) P′a exp(-k′a∆) + P′b exp(-k′b∆)

(2)

k′a ) C1 - C2

(3)

k′b ) C1 + C2

(4)

with

[

]

(5)

[

]

(6)

1 1 1 (Pb - Pa)(ka - kb) + + τa τb 1 4 P′a ) + 2 C2 1 1 1 (Pb - Pa)(ka - kb) + + 4 τ τ a b 1 P′b ) 2 C2

( [(

) )

C1 )

1 1 1 k + kb + + 2 a τa τb

C2 )

1 1 1 2 4 k a - kb + + 2 τa τb τaτb

(7)

]

1/2

(8)

ka ) γ2δ2G2Da

(9)

kb ) γ2δ2G2Db

(10)

and

where Pa, Pb and τa, τb represent the populations and average residence times of observed molecules in the (a) free and (b) encapsulated states, respectively. According to this set of equations, the echo decay curve can be analyzed for all important system parameters: The diffusion constant Da primarily determines the initial steep slope of the decay plot, and the diffusion constant for the Brownian motion of the vesicles Db dominates the slope of the final shallow part of the decay curve. The absolute level of the shallow part in the extrapolation for G, ∆ f 0 is determined by the encapsulated fraction Pb, and the variation of this level with the pulse separation ∆ reflects the exchange rate as determined by the average residence time in the encapsulated state τb. The reliability of the detected parameters was finally checked by a comparison between the experimental data and a corresponding simulation of the echo decay according to eqs 2-10. The hydrodynamic radius Rh of the vesicles can be calculated from the diffusion constant for the Brownian motion Db according to

Rh ) kBT/(6πηDb)

(11)

where kB is the Boltzmann constant, T is the absolute temperature, and η is the viscosity of the solvent. 3. Results 3.1. Viscosity of Water/Glycerol Mixed Solutions. For a full interpretation of the diffusion experiment, it is necessary to determine the solution viscosities of the water/glycerol mixtures. Figure 1 shows the variation of the viscosity of the aqueous solutions with the volume fraction of glycerol. The viscosity values used in this study are listed in Table 1.

Permeability of Siloxane Surfactant Vesicle Membranes

Figure 1. Variation of the viscosity of water/glycerol mixed solvents with the volume fraction (φ) of glycerol at 293 K.

TABLE 1: Viscosities of Water/Glycerol Mixtures at Different Glycerol Volume Fractions (293 K) φ

η (103 Pa s)

0.00 0.20 0.40 0.60

1.005 2.27 5.90 20.51

3.2. FF-TEM Results on the Effect of Glycerol on Vesicle Structure. The change of the solvent’s properties brings about a transition of the aggregation behavior of the IM-22 molecules; hence, one expects structural changes of the vesicles with increasing glycerol concentration. Figure 2 shows FF-TEM micrographs of 5% IM-22 in water and in different water/ glycerol mixed solvents. For 0% glycerol, the images show large amounts of SUVs (small unilamellar vesicles) with radii less than 25 nm, as well as a few LMVs (large multilamellar vesicles) with radii as large as 250 nm. When the glycerol content is increased to 20%, the SUVs are more pronounced, and the LMVs are still preserved. When the glycerol content is increased to 40%, the SUVs with diameters of less than 50 nm are hardly detectable, and their size clearly increases. At the same time, the collapse of the LMVs is observed. Finally, at a glycerol content of 60%, most of the vesicles are destroyed, and open membrane fragments appear (Figure 2d). 3.3. PFG-NMR Results. 3.3.1. Self-Diffusion of Water. The echo decay curve of a pulsed-field gradient NMR experiment allows for the observation of the diffusion profile of any given component in a heterogeneous system. Figure 3 shows a set of echo decay curves observed for the water signal of an aqueous dispersion of 5% IM-22 vesicles for different spacings ∆ between the two gradient pulses. Plotting logarithmic echo intensities, ln Irel ) ln I/I0, versus the parameter γ2δ2G2(∆-δ/ 3) yields linear decay functions in the case of free diffusion. In fact, for all pulse spacings, linear dependencies for ln Irel vs γ2δ2G2(∆-δ/3) are observed for the full echo decay down to Irel ) exp(-8) ≈ 3.4 × 10-4. No signs for a separation between free and encapsulated fractions of water are detected. A similar observation was made in the presence of glycerol (data not shown). In the absence of glycerol, the common slope of all five curves reflects a self-diffusion constant of Da ) (2.03 × 10-9 ( 0.016 × 10-9) m2/s, which is slightly smaller than the coefficient for the free self-diffusion of pure water at 2.26 × 10-9 m2/s for 298 K (source, AIP). 3.3.2. Self-Diffusion of Glycerol. As a larger molecule, glycerol is expected to present a different diffusion behavior

J. Phys. Chem. B, Vol. 111, No. 22, 2007 6163 than water. Figure 4A-C shows the echo decay curves for glycerol signals in 20%, 40%, and 60% glycerol systems, respectively. The symbols in the plot refer to the measured echo intensities, and the solid lines represent the best-fit simulations according to eqs 2-10. In the case of the glycerol signal, two clearly different slopes are observed for each given pulse separation from 50 to 250 ms. The initial very steep slope down to ln (I/I0) ) -2 corresponds to the free diffusion of glycerol in the continuous phase, which contributes to more than 90% of the overall glycerol content. The common slope of all five curves reflects the self-diffusion constant of glycerol in the continuous phase of the given solution viscosity. The final slope, which again is shared by all five echo decay curves, is related to the Brownian motion of the vesicles. Finally, the dependence of the echo intensities on ∆ for larger gradients (hence, the “spread” of the curves) indicates the loss of the encapsulated fraction of glycerol over time as a result of the continuous molecular exchange through the vesicle membranes.20 The rate constant for the exchange of glycerol molecules between (a) the continuous phase and (b) the vesicle content can be obtained by fitting the echo decay curves as described in previous publications.19,20 It is characterized by an average residence time τb of a glycerol molecule in the encapsulated state. The parameters Da, Db, τb, and the entrapped populations of glycerol Pb obtained from the PFG-NMR simulation for 20%, 40%, and 60% glycerol system are plotted against the glycerol content in Figures 5 and 6. Using the solution viscosity and the diffusion coefficient Db, which refers to the Brownian motion of the vesicles, the average hydrodynamic radii of the vesicles (or other membrane binding structures) can be calculated according to eq 11. With increasing volume fraction of glycerol, the hydrodynamic radius of the vesicles presents a maximum value at 40% glycerol (Figure 7). 4. Discussion 4.1. Vesicle Morphology. Observation of an aqueous IM22 system in a transmission electron microscope yields clear indications of the presence of vesicles (Figure 2). In addition to a large number of small vesicles with radii smaller than 25 nm, several large multilamellar vesicles with radii of up to 250 nm are detected in the micrographs. The smaller vesicles show only minor deviations from the ideal spherical shape. Increasing glycerol content initially leads to a collapse of the large multilamellar vesicles (at 40% glycerol content) and finally to the destruction of the small vesicles into membrane fractures (at 60% glycerol content). The disintegration of the vesicle membranes can be attributed to a change in solvent properties. As the concentration of glycerol increases, the polarity of the solvent adapts to the overall polarity profile of the IM-22 molecules, leading to larger membrane undulations and promoting its disassembly. For intact vesicles, the diffusion experiments on glycerol molecules show a clear separation between two domains: a large, free fraction and a smaller, encapsulated fraction. The analysis of the echo decay plots in Figure 4 yields data on the relative amount of the encapsulated glycerol fraction. When the glycerol content is increased from 20% to 40%, the entrapped population of glycerol is reduced from 9.5% to 7.5%, indicating the partial destruction of the vesicles. Interestingly, the encapsulated fraction does not disappear completely after the full destruction of all vesicles at 60% glycerol. At this point, a residual content of 3.3% can be

6164 J. Phys. Chem. B, Vol. 111, No. 22, 2007

Yan et al.

Figure 2. FF-TEM images of 5% IM-22 in (a) water, (b) 20% glycerol, (c) 40% glycerol, and (d) 60% glycerol.

Figure 3. Overlapping echo decay curves observed on the water signal of 5% IM-22 in aqueous solution (without glycerol) under variation of the field gradient strength G for different spacings (6 ms e ∆ e 15 ms) between the two gradient pulses of a PFG-Hahn echo sequence.

attributed to the encapsulated state. Presumably, this fraction of glycerol is bound inside the hydrophobic layers of open membrane fragments. With increasing glycerol content, the solvophobicity of the DMS groups decreases, which enables the polar glycerol molecules to integrate into the molecular packing of the DMS layers, probably even binding to the DMS groups. Because the basic mechanism for the integration of glycerol molecules observed for the membrane fragments also represents a crucial step for the membrane permeation observed for vesicles, it is not surprising that the exchange rates are similar for the two processes (see τb for 40% and 60% glycerol in Figure 5). In both cases, the kinetics of the exchange are determined by the intrusion of the glycerol molecule into the hydrophobic layer of the membrane. 4.2. Vesicle Size. The average hydrodynamic radii of the vesicles are calculated from the solution viscosity and the diffusion coefficient Db (which refers to the Brownian motion of the vesicles) according to eq 11. With an increasing volume

fraction of glycerol, the average hydrodynamic radius of the vesicles increases from 9.1 nm (20% glycerol) to 12.9 nm (40% glycerol) (Figure 7) and then decreases sharply to 2.5 nm. This trend is in accordance with the visual observations made by transmission electron microscope (Figure 2), where the small vesicles appear to grow with increasing glycerol concentration up to 40% glycerol. The calculated hydrodynamic radii are averages over all existing colloidal particles in the dispersion. Actually, aside from SUVs and LMVs, one has to expect large amounts of micelles with radii lower than 2 nm, which are beyond the resolution limits of FF-TEM.13 Therefore, the calculated hydrodynamic radii are smaller than the present FFTEM observations. Obviously, the contribution of the LMVs to these values is relatively small. The broken structures observed for glycerol contents of around 60% obviously derive from the broken vesicle membranes, which seem to accommodate small amounts of glycerol (see above). Because of the breaking of vesicles, the resulting membrane fractures are much smaller than the original vesicles, which leads to a dramatic decrease of the average hydrodynamic radius to 2.5 nm. Of course, at this stage, the membrane fractures are obviously nonspherical. In fact, shear thinning behavior has already been observed at this composition. Consequently, eq 11 is no longer accurate for the calculation of the radius at 60% glycerol, and the result marks only a rough trend. 4.3. Vesicle Membrane Permeability for Water Molecules. The detection of the self-diffusion profile for mobile components in the vesicle dispersion generally yields valuable information on the specific permeability of the vesicle membrane. For the current system, water and glycerol are the natural candidates for an analysis of vesicle membrane properties. When water molecules are observed in a PFG-NMR experiment, the echo decay curves reflect only a linear dependence

Permeability of Siloxane Surfactant Vesicle Membranes

J. Phys. Chem. B, Vol. 111, No. 22, 2007 6165

Figure 6. Relative entrapment of glycerol with increasing glycerol content in the 5% IM-22 system.

Figure 4. Sets of echo decay curves for 5% IM-22 in different water/ glycerol mixed solvents. Glycerol contents (by volume): (A) 20%, (B) 40%, (C) 60%. Each individual plot refers to a given pulse spacing ∆ as shown in the insets. Simulation data sets are as follows: 20% glycerol (Da ) 4.48 × 10-10 m2/s, Db ) 1.04 × 10-11 m2/s, Pb ) 9.5%, τb ) 910 ms), 40% glycerol (Da ) 2.45 × 10-10 m2/s, Db ) 2.83 × 10-12 m2/s, Pb ) 7.5%, τb ) 287 ms), 60% glycerol (Da ) 1.21 × 10-10 m2/s, Db ) 4.62 × 10-12 m2/s, Pb ) 3.3%, τb ) 289 ms).

Figure 5. Fitting parameters plotted against the glycerol content for the 5% IM-22 system.

with a slope equal to the negative self-diffusion constant (-D). This means that average residence time for water molecules inside the vesicles is very short, presumably on the order of 2 ms or less. This rapid exchange indicates that the vesicle membranes of IM-22 do not represent a significant barrier for the self-diffusion of water molecules. Previous results have shown that the permeability of the vesicle membranes for water molecules strongly depends on the polarity of the hydrophobic part of the membrane.20 In the case of the IM-22 vesicle system, the observed high permeability of the vesicle membrane is

Figure 7. Average diameters of the particles calculated from the diffusion coefficient of Brownian motion as a function of the volume fraction φ of glycerol.

probably due to the very flexible and relatively polar nature of DMS groups and to the small thickness of the hydrophobic layer.14 Further, with glycerol present in the solvent, the hydrophobic part of the membrane is expected to contain a certain amount of glycerol, which will further reduce the ability of this layer to form a barrier against water molecules. 4.4. Vesicle Membrane Permeability for Glycerol Molecules. The observation of glycerol molecules in the diffusion experiments yields a completely different result. In this case, two distinctly different slopes are found for the echo decay curves in all three glycerol-containing samples, indicating a much slower exchange through the vesicle membranes than for water. This is primarily due to the larger size of glycerol molecules, which is related to a higher activation energy for the membrane transfer. The permeability of the vesicle membranes changes dramatically with the glycerol content. The increase of the glycerol concentration from 20% to 40% is followed by an increase of the exchange rate by more than a factor of 3. This is expressed by a decrease of the average residence time from 910 to 287 ms. It should be pointed out that the membrane thickness does not change with the glycerol content as revealed by SAXS experiments in our previous report.17 It is expected that the amount of glycerol that resides in the hydrophobic layer of the vesicle membrane is proportional to the glycerol concentration in the solution. Hence, the glycerol concentration in the membrane is assumed to be twice as large in the 40% system as it is in the 20% system. At the same time,

6166 J. Phys. Chem. B, Vol. 111, No. 22, 2007 the permeability for glycerol molecules increases. Obviously, the energy barrier for the permeation process is strongly reduced by the presence of glycerol molecules in the hydrophobic layer. Presumably, the glycerol molecules lead to an increase of the polarity as well as of the flexibility of the hydrophobic section of the block copolymer, a phenomenon that has previously been observed for ethanol in poly(isoprene-block-ethylene oxide) vesicles.11 This effect of small molecules of intermediate polarity might be of general importance for the intentional variation of the permeability of vesicle membranes. 5. Conclusion The permeability of vesicle membranes with increasing glycerol was studied by PFG-NMR experiments, and the results were related to FF-TEM observations. Generally, the exchange of water molecules characterized by an average residence time in the encapsulated state of τb e 2 ms is too fast to be observed with the given approach, which means that vesicle membranes of IM-22 are not a significant barrier for water diffusion. In contrast, glycerol molecules show an average residence time of up to τb ) 910 ms, which is primarily a consequence of their larger size. The permeability of the vesicle membrane for glycerol increases with increasing glycerol concentration in the solvent: At a concentration of 40%, the residence time τb is lowered to approximately 290 ms. After vesicle destruction at higher glycerol concentrations, a small glycerol fraction is still bound in the hydrophobic region of open membrane fragments that result from the disintegration of the vesicles. The exchange of glycerol molecules between the membrane fragments and the continuous phase is characterized by rate constants similar to those for the vesicles. This indicates that the intrusion of glycerol into the hydrophobic region of the vesicle membrane determines the time scale of the membrane permeation process. Acknowledgment. The authors gratefully acknowledge the generous donation of samples of IM-22 by Wacker Chemie AG,

Yan et al. Munich, Germany. We also thank Dr. Walter Richter (Klinikum der FSU Jena, Elektronenmikroskopisches Zentrum, Ziegelmu¨hlenweg 1, 07740 Jena, Germany) for help with the FFR-TEM experiments. References and Notes (1) Galme¨s, A.; Besalduch, J.; Bargay, J.; Novo, A.; Morey, M.; Guerra, J. M.; Duran, M. A. Transfusion 1999, 39, 70. (2) Schiffer, C. A.; Aisner, J.; Wernik, M. D. New Engl. J. Med. 1978, 299, 7. (3) Lawaczeck, R. J. Membr. Biol. 1979, 51, 229. (4) Finkelstein, A. Water MoVement Through Lipid Bilayers, Pores, and Plasma Membranes: Theory and Reality; Wiley-Interscience: New York, 1987. (5) Paula, S.; Volkov, A. G.; Van Hoek, A. N.; Haine, T. H.; Deamer, D. W. Biophys. J. 1996, 70, 339. (6) Huster, D.; Jin, A. J.; Arnold, K.; Gawrisch, K. Biophys. J. 1997, 73, 855. (7) Groth, C.; Bender, J.; Nyden, M. Colloids Surf. A 2003, 228, 64. (8) Battaglia, G.; Ryan, A. J.; Tomas, S. Langmuir 2006, 22, 4910. (9) Komatsu, H.; Okada, S. Biophys. J. 1996, 70, WP217. (10) Ly, H. V.; Longo, M. L. Biophys. J. 2004, 87, 1013. (11) Bauer, A., Kopschu¨tz, C.; Stolzenburg, M.; Fo¨rster, S.; Mayer, C. J. Membr. Sci. 2006, 284, 1. (12) Martin, M., Swaebrick, J., Ananthapadamanabhan, K. P., Eds. Physical Pharmacy; Lea & Febiger Press: Philadelphia, PA, 1983. (13) Yan, Y.; Hoffmann, H.; Drechsler, M.; Talmon, Y.; Makarsky, E. J. Phys. Chem. B 2006, 110, 5621. (14) Gradzielski, M.; Hoffmann, H.; Robisch, P.; Ulbricht, W. Tenside Surf. Deterg. 1990, 27, 366. (15) Schmaucks, G.; Sonnek, G.; Wustneck, R.; Herbst, M.; Ramm, M. Langmuir 1992, 8, 1724. (16) Hill, R. M. Langmuir 1993, 9, 2789. (17) Yan, Y.; Hoffmann, H.; Makarsky, A.; Richter, W.; Talmon, Y. J. Phys. Chem. B 2007, 111, in press. (18) Bauer, A.; Hauschild, S.; Stolzenburg, M.; Fo¨rster, S.; Mayer, C. Chem. Phys. Lett. 2006, 419, 430. (19) Rumplecker, A.; Fo¨rster, S.; Za¨hres, M.; Mayer, C. J. Chem. Phys. 2004, 120, 8740. (20) Mayer, C.; Bauer, A. Prog. Colloid Polym. Sci. 2006, 133, 22. (21) Waldeck, A. R.; Kuchel, P. W.; Lennon, A. J.; Chapman, B. E. Prog. Nucl. Magn. Reson. Spectrosc. 1997, 30, 39. (22) Kennedy, A.; Long. C. J.; Hmel, P. J.; Hicks, R.; Reid, T. J. Spectroscopy 2004, 18, 265.