Influence of Perfluorinated Compounds on the Properties of Model

Aug 2, 2007 - Mitchell B. Rosen , Judith R. Schmid , J. Christopher Corton , Robert D. Zehr , Kaberi P. Das , Barbara D. Abbott , Christopher Lau...
0 downloads 0 Views 921KB Size
Download PDF
9908

J. Phys. Chem. B 2007, 111, 9908-9918

Influence of Perfluorinated Compounds on the Properties of Model Lipid Membranes Dorota Matyszewska,† Kirsi Tappura,‡ Greger Ora1 dd,§ and Renata Bilewicz*,† Department of Chemistry, UniVersity of Warsaw, ul. Pasteura 1, 02093 Warsaw, Poland, VTT Technical Research Centre of Finland, P.O. Box 1300, FI-33101 Tampere, Finland, and Department of Chemistry, Umeå UniVersity, S-901 87 Umeå, Sweden ReceiVed: December 22, 2006; In Final Form: June 4, 2007

The influence of selected perfluorinated compounds (PFCs), perfluorooctanoic acid (PFOA) or perfluorooctanesulfonic acid (PFOS), on the structure and organization of lipid membranes was investigated using model membranesslipid monolayers and bilayers. The simplest modelsa lipid monolayerswas studied at the airwater interface using the Langmuir-Blodgett technique with surface pressure and surface potential measurements. Lipid bilayers were characterized by NMR techniques and molecular dynamics simulations. Two phospholipids, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and 1,2-dimyristoyl-sn-glycero-3phosphocholine (DMPC), characterized by different surface properties have been chosen as components of the model membranes. For a DPPC monolayer, a phase transition from the liquid-expanded state to the liquidcondensed state can be observed upon compression at room temperature, while a DMPC monolayer under the same conditions remains in the liquid-expanded state. For each of the two lipids, the presence of both PFOA and PFOS leads to the formation of a more fluidic layer at the air-water interface. Pulsed field gradient NMR measurements of the lateral diffusion coefficient (DL) of DMPC and PFOA in oriented bilayers reveal that, upon addition of PFOA to DMPC bilayers, DL of DMPC decreases for small amounts of PFOA, while larger additions produce an increased DL. The DL values of PFOA were found to be slightly larger than those for DMPC, probably as a consequence of the water solubility of PFOA. Furthermore, 31P and 2H NMR showed that the gel-liquid crystalline phase transition temperature decreased by the addition of PFOA for concentrations of 5 mol % and above, indicating a destabilizing effect of PFOA on the membranes. Deuterium order parameters of deuterated DMPC were found to increase slightly upon increasing the PFOA concentration. The monolayer experiments reveal that PFOS also penetrates slowly into already preformed lipid layers, leading to a change of their properties with time. These experimental observations are in qualitative agreement with the computational results obtained from the molecular dynamics simulations showing a slow migration of PFCs from the surrounding water phase into DPPC and DMPC bilayers.

Introduction Perfluorinated compounds (PFCs) are fully fluorinated fatty acid analogues used in production of materials such as lubricants, paints, cosmetics, and fire-fighting foams. These compounds have recently attracted scientists’ attention due to their presence in the environment including humans.1-3 Owing to their stable nature, PFCs can be accumulated in higher order species, and it is thus of great importance to study the longterm effects of perfluorinated compounds on the biosphere. So far it has been shown that PFCs may cause alterations in cell membrane properties,4-6 and this rather alarming feature resulted in numerous further studies. Lipid monolayers although much less complex than biological membranes have been shown to be excellent model interfaces to study the interactions with solution species.7-9 They are especially convenient for the studies of interactions of the lipid layers with various pollutants or drugs since these compounds can be purposely added to the subphase in a controlled concentration.10-13 Monolayers of 1,2-dipalmitoyl-sn-glycero* To whom correspondence should be addressed. E-mail: bilewicz@ chem.uw.edu.pl. Phone: +48 22 8220211 ext 345. Fax: +48 22 8224889. † University of Warsaw. ‡ VTT Technical Research Centre of Finland. § Umeå University.

3-phosphocholine (DPPC) are well characterized14-19 and often used as models of the outer membrane cell leaflet. Another lipid which can be used as a component of such model cell membranes is 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), with a structure that slightly differs from the DPPC structure (Figure 1). The difference in hydrocarbon chain length leads to different surface properties of the two lipids such as their fluidity.20,21 The use of the pulsed field gradient (pfg) NMR technique of measuring lateral diffusion coefficients of molecules included in macroscopically oriented bilayers22 has proven to be a powerful approach for studies of membrane packing properties.23 In model membranes the lipid lateral diffusion coefficients (DL) are sensitive to changes in lipid packing induced by incorporating, e.g., sterols, and/or changes in temperature. The influence of membrane additives can thus be conveniently studied by this technique. Macroscopically oriented model membranes are also well suited for studies of the gel-liquid crystalline phase transition since the sharp lines in proton and phosphorus spectra originating from the oriented part of the samples disappear as the gel phase is entered.24 In the present study, we describe the effect of selected fluorinated compounds upon the organization of lipid Langmuir monolayers of DPPC and DMPC spread at the air-water

10.1021/jp068874g CCC: $37.00 © 2007 American Chemical Society Published on Web 08/02/2007

Influence of PFCs on Model Lipid Membranes

Figure 1. Structures of the lipids and perfluorinated compounds studied: (a) DPPC, (b) DMPC, (c) PFOS, (d) PFOA.

interfaces. The monolayers are characterized by surface pressure (π) and electric surface potential (∆V) measurements.25,26 Bilayer model membranes composed of DMPC with various amounts of perfluorooctanoic acid (PFOA) are studied by 1H, 19F, 2H, and 31P NMR. The lateral diffusion coefficients of DMPC and PFOA, obtained from pfg-NMR experiments in oriented lipid bilayers, are used to probe the changes in lipid packing induced by the addition of PFOA, and this is compared with the changes in deuterium order parameters imposed on DMPC by PFOA. 31P and 2H NMR data give information on the changes in the main transition temperature of the system. The effect of PFCs on the lipid DPPC and DMPC bilayers is also investigated by molecular dynamics (MD) simulations. Experimental Section Materials and Methods. DPPC, DMPC, and DMPC-d54 were purchased from Avanti Polar Lipids Inc., PFOA was purchased from Aldrich, and perfluorooctanesulfonic acid (PFOS) was purchased from Apollo Scientific Ltd. Chloroform used for preparing lipid solutions was purchased from POCh Gliwice, Poland. Distilled water was passed through a Milli-Q water purification system (resistivity 18.2 MΩ). Langmuir-Blodgett Technique. Surface pressure and surface potential vs area per molecule isotherms were recorded using a KSV LB trough 5000 (KSV Ltd., Finland) equipped with two hydrophobic barriers, a Wilhelmy balance with a paper plate used as a surface pressure sensor, and a 5000SP surface potential meter (vibrating capacitor method). Software version KSV-5000 was used to control the experiments. To protect the experimental setup from dust, it was placed in the laminar flow

J. Phys. Chem. B, Vol. 111, No. 33, 2007 9909 hood in which temperature was kept constant. The surface potential and surface pressure were recorded simultaneously as a function of molecular area. The accuracy of the measurements was (0.2 Å2 for the area per molecule, (0.1 mN m-1 for the surface pressure, and (5 mV for the surface potential. Water with or without perfluorinated compound was used as the subphase. The spreading solutions were prepared by dissolving DPPC or DMPC samples in chloroform with a typical concentration of 1 mg/mL. Samples of 40-50 µL of these solutions were spread on the air-water interface. After spreading, the solution was left for 10 min for solvent evaporation. Compression of the barriers was performed at a speed of 7.5 cm2/min. NMR Experiments. Macroscopically oriented bilayers were prepared by the solvent evaporation technique as described in detail elsewhere.22 Briefly, thin layers of DMPC mixed with PFOA were deposited on thin glass plates that were subsequently stacked and placed in a sample holder where they were exposed to a humid atmosphere. During the hydration of the sample, the bilayers become macroscopically oriented parallel to the surface of the glass plate. The samples were buffered to approximately pH 7.2 by addition of appropriate amounts of phosphate buffer to the DMPC/PFOA mixtures before the glass plate deposition. The NMR experiments were performed on a Varian/Chemagnetics CMX Infinity spectrometer (Fort Collins, CO) using a specially designed 1H/BB goniometer probe in which the sample can be rotated in the magnetic field around a horizontal axis. The temperature was controlled to (0.5 °C by a heated air stream passing the sample, and temperature scans were performed from 40 to 10 °C (DMPC) in steps of 3 °C with a waiting time of 10 min after each temperature change. 2H NMR spectra were obtained at 61.48 MHz with a quadrupole echo sequence,27 using composite pulses to obtain complete spectral coverage.28 A total of 128 scans were collected with a recycle time of 0.3 s, and the spectra were Fouriertransformed after a 20 Hz line broadening. For the 2H experiments the bilayer normal was oriented parallel to the main magnetic field, thereby maximizing the separation of the Pake doublets in the spectrum. The deuterium order parameter (SCD) can be calculated from the frequency separation of the doublets as |SCD| ) ∆νq/255 kHz.27 Lateral diffusion coefficient determinations were performed by the pfg-NMR technique22 in which the proton (400.15 MHz) and fluorine (376.81 MHz) frequencies were used for DMPC and PFOA, respectively. This makes it possible to separately study the two molecules in the bilayers. To remove unwanted static interactions in the 1H experiments, the samples were oriented with the bilayer normal at 54.7° with respect to the main magnetic field by maximizing the signal amplitude in a pfg-NMR spin-echo experiment.22 The stimulated spin echo29 was used with the gradient amplitude varied between 0.48 and 9.52 T/m and all other parameters kept constant during the experiment. DL was then extracted from the attenuation of the spectral amplitude as a function of the gradient strength. For the proton experiments the gradient pulse width was 3 ms and the diffusion time was either 15 or 100 ms, while the corresponding parameters for the fluorine experiments were 1.5 and 10 or 50 ms, respectively. The obtained DL values were not dependent on the diffusion time. 31P NMR spectra were recorded at 162.13 MHz using the Hahn echo sequence with WALTZ decoupling of the protons.30 A total of 32 scans were collected with a recycle time of 1 s, and the spectra were Fourier-transformed after a 50 Hz line

9910 J. Phys. Chem. B, Vol. 111, No. 33, 2007

Matyszewska et al.

broadening. The 31P experiments were performed with the bilayer normal at 54.7° with respect to the main magnetic field, giving a single line at the isotropic chemical shift for the oriented bilayers.31 Molecular Dynamics Simulations. MD simulations were performed on different systems comprised of a DPPC bilayer, water, and PFOS and of a DMPC bilayer, water, and PFOA with the GROMACS32 molecular dynamics simulation suite (versions 3.3-3.31). The lipid bilayers of 128 DPPC or DMPC molecules, with 64 phospholipid molecules in each leaflet, were fully hydrated by the surrounding 3655-4921 water molecules (the exact number depending on the system). The number of PFOS molecules, initially randomly embedded in water, was 14 or 22, corresponding to 7.7 wt % (0.15 M) and 11.6 wt % (0.23 M) PFOS, respectively, while the initial number of PFOA molecules in water was 6 or 14 corresponding to 2.7 wt % (0.07 M) and 5.8 wt % (0.14 M) PFOA, respectively. The simulations of the bilayers were performed both in pure water and in water with the two concentrations of PFOS or PFOA. In the simulation box the lipid bilayer was sandwiched between the layers of water or the water/PFC mixtures. The water molecules (SPC)33 were included explicitly, and periodic boundary conditions were used to minimize the artificial boundary effects. The temperature was controlled by weak coupling to a heat bath of 7 °C (280 K), 27 °C (300 K), or 52 °C (325 K) with a time constant of 0.1 ps.34 A similar procedure with a time constant of 10 ps was used to adjust the pressure to 1 bar, while the height of the simulation box (i.e., the z direction, which is perpendicular to the plane of the bilayer) was allowed to vary independently of the lateral directions. The time step was 2 fs with all bond lengths constrained with the LINCS algorithm35 in nonwater molecules and with SETTLE36 in water molecules. The cutoff distance for the nonbonded interactions was 0.9 nm. However, to avoid the possible artifacts created when the long-range electrostatic interactions are cut off, the particle-mesh Ewald37,38 method was applied to the electrostatic interactions. Initial atomic velocities were randomly generated from the Maxwell distribution at 7 °C (280 K), 27 °C (300 K), or 52 °C (325 K). The initial configurations of the lipid bilayers were obtained from http:// moose.bio.ucalgary.ca, with the DPPC configuration corresponding to the final structure of run E of ref 36. A previously validated united atom model was used for the lipids,39,40 while the basis of the topology and parameter information for PFOS and PFOA was generated with the PRODRG server,41 after which the values, especially the correctness of the partial charges, were manually checked to follow the conventions of the GROMACS force field.42,43 Both PFOS and PFOA were treated as anions with a charge of -1. After insertion of the PFOS or PFOA molecules and water in the simulation box on both sides of the bilayer, the system was energy-minimized and then the solvent equilibrated for at least 10 ps to remove solvent-based bad contacts, while the other molecules were restrained by applying harmonic potentials with

Figure 2. π-area per molecule isotherms of DPPC on subphases containing different solutions of (a) PFOA and (b) PFOS. Insets: compression modulus (Cs-1) vs surface pressure (π) plots.

a force constant of 1000 (kJ/mol)/nm2. The length of the production runs was 80 ns. Results and Discussion Influence of Perfluorinated Compounds on DPPC Langmuir Monolayers. Isotherms of DPPC on subphases containing two different perfluorinated compounds (PFOA and PFOS) have been recorded and compared with DPPC isotherms on water. For each fluorinated compound, two different concentrations (10-5 and 10-4 M) in the subphase were investigated. A similar effect can be observed for both PFCs. Molecules of perfluorinated compounds are incorporated into the DPPC monolayer, which can be easily noticed especially in the beginning of the monolayer compression. The area per molecule, A0 (obtained by extrapolating the tangent to the isotherm in the liquid-condensed (LC) region to zero surface pressure), increases when PFOA or PFOS is present in the subphase (Table 1). The difference can be clearly seen at higher concentrations, and the effect is more significant for PFOS (Figure 2). At higher surface

TABLE 1: Characteristic Parameters for DPPC Langmuir Monolayers on Different PFOA and PFOS Concentrations in the Subphase concn (M) of perfluorinated compd in the subphase water PFOA PFOS

0 10-5 10-4 10-5 10-4

A0 (Å2)

Acoll (Å2)

πcoll (mN/m)

max Cs-1 (mN/m)

53.0 ( 0.1 54.0 ( 0.3 60.5 ( 0.9 55.6 ( 0.5 63.0 ( 0.9

38.9 ( 0.8 38.0 ( 0.7 35.4 ( 0.6 32.5 ( 0.0 38.7 ( 0.3

43.6 ( 0.4 42.4 ( 0.7 41.9 ( 0.9 40.6 ( 1.1 42.3 ( 0.6

191.2 ( 3.0 171.8 ( 4.7 127.7 ( 5.4 149.7 ( 6.1 145.9 ( 7.4

Influence of PFCs on Model Lipid Membranes pressures PFOA is expelled from the monolayer, while PFOS still interacts with the condensed monolayer. This difference in the nature of the interactions of the two perfluorinated compounds with lipid monolayers might be attributed both to the difference in the length of the hydrophobic chain and to the difference in the polar groups of the two compounds. The influence of perfluorinated compounds on the area per molecule corresponding to the collapse of the monolayer (Acoll) and on the value of the collapse surface pressure (πcoll) is not that significant. The presence of PFCs in the subphase also influences other properties of the DPPC monolayer such as its fluidity. With the increasing concentration in the subphase, the phase transition from liquid-expanded (LE) to LC becomes less visible or finally disappears. As a result, the DPPC monolayer becomes less condensed. These changes are visualized by the isotherm shape and also by the compression modulus (Cs-1) (Figure 2). The compression modulus (reciprocal of compressibility) is defined as44-48

Cs-1 ) -A(dπ/dA) where A is the area per molecule and π is the surface pressure. The phase transition of DPPC occurring at approximately 70 Å2 may be clearly seen as a minimum in the Cs-1 versus surface pressure plot. However, this minimum becomes almost invisible in the presence of higher PFC concentrations in the subphase (10-4 M) (insets in Figure 2), which stays in good agreement with the results obtained by other authors, who suggested that perfluorinated compounds may be incorporated into the lipid membranes and thus may increase the membrane fluidity.4,5 The incorporation is more pronounced in the case of PFOS; for 10-4 M, the increase in A0 is approximately 10 Å2 compared to that of the DPPC monolayer on water. Moreover, the isotherm is significantly shifted toward greater areas per molecule, and the minimum in the Cs-1 versus surface pressure plot cannot be observed, which proves the fluidizing effect of PFOS. The comparison of maximum Cs-1 values for 10-4 M PFC solutions given in Table 1 leads to similar conclusions. Since values well above 150 mN/m indicate the formation of solidlike layers and the maximum Cs-1 values for 10-4 M PFC solutions are less than 150 mN/m, the tendency that a more liquidlike state of the monolayers is formed is confirmed. In the case of the higher concentration of PFOA the minimum in the Cs-1-π dependency can be observed, which might suggest that LE-LC phase transition still occurs and the influence of PFOA is less significant than that of PFOS.5 To find whether PFOS molecules not only may be incorporated into a lipid monolayer during its formation, but also are able to penetrate an already preorganized layer, the time dependencies of the surface pressure were measured. The DPPC monolayer was compressed to a selected surface pressure value, then barriers of the trough were stopped, and changes of the surface pressure in time were observed. The experiments for monolayers compressed to π ) 4 mN/m (liquid-expanded region) and π ) 35 mN/m (liquid-condensed region) were carried out. Shortly after compression to the chosen surface pressure and stopping of the barriers, PFOS solution was injected into the subphase to achieve the final concentration of 10-5 M. Comparison of the results for stability experiments of DPPC monolayers on water as the subphase and after injection of PFOS are shown in Figure 3. For the DPPC monolayer compressed to 4 mN/m a significant decrease in surface pressure in time can be observed, which means that the monolayer is not highly stable. Interestingly,

J. Phys. Chem. B, Vol. 111, No. 33, 2007 9911

Figure 3. Stability of DPPC monolayers on water and after injection of PFOS solution into the subphase (final concentration 10-5 M) compressed to two different surface pressure values.

TABLE 2: Values of the Area per Molecule and Collapse Pressure for the DMPC Monolayer and Different PFOA Concentrations in the Subphase PFOA concn (M) in the subphase

A0 (Å2)

Acoll (Å2)

πcoll (mN/m)

0 10-5 10-4

73.0 ( 0.3 79.6 ( 0.5 82.6 ( 0.8

43.9 ( 0.5 41.0 ( 0.6 40.8 ( 0.9

47.0 ( 0.6 49.5 ( 1.0 49.5 ( 0.9

after injection of PFOS into the subphase, the decrease in surface pressure is stopped and it remains at the level of approximately 1.5 mN/m. This may suggest that PFOS molecules can adsorb onto the air/water interface and penetrate the DPPC monolayer in the liquid state and prevent it from destabilizing with time. A similar conclusion may be drawn from the second stability experiment. After the injection of PFOS into the subphase, on which the DPPC monolayer is compressed to 35 mN/m, the surface pressure increases in time, reflecting slow incorporation of PFOS molecules into the layer. After a certain time the surface pressure attains a constant value, indicating stabilization of the monolayer on the subphase containing the given concentration of PFOS. Although in both cases the process of penetration can be observed, the changes are more pronounced in the case of monolayers compressed to 4 mN/m. Incorporation of PFOS molecules is thus much easier when the DPPC monolayer is in the liquid-expanded state. Influence of Perfluorinated Compounds on DMPC Langmuir Monolayers. As was mentioned in the Introduction, the DMPC monolayer at room temperature differs significantly from DPPC in that the phase transition from the LE phase to the LC phase does not occur and only the LE phase can be observed. Therefore, the influence of the two perfluorinated compounds on DMPC monolayers has also been investigated and compared to the influence on DPPC monolayers. Also for the DMPC monolayer a significant increase in A0 can be observed with increasing PFOA concentration in the subphase (Table 2). Thus, PFOA molecules are being incorporated into the liquid DMPC monolayer. The difference between the values of A0 measured in the absence and presence of 10-4 M PFOA in the subphase is approximately 10 Å2, which shows that there are strong interactions between PFOA and the DMPC monolayer. The values of πcoll and Acoll are not significantly influenced by increasing the PFOA concentration in the subphase. The fluidizing effect of PFOA on DMPC monolayers was also studied under decreased temperature conditions. At 13 °C the phase transition from the LE phase to the LC phase can be

9912 J. Phys. Chem. B, Vol. 111, No. 33, 2007

Matyszewska et al.

Figure 4. π-area per molecule isotherms of DMPC monolayers on subphases containing different PFOA concentrations (T ) 22 °C).

Figure 5. π-area per molecule isotherms of DMPC for different PFOA concentrations in the subphase (T ) 13 °C). Inset: compression modulus (Cs-1) vs surface pressure (π) plot.

observed for the DMPC monolayer, making it similar to the DPPC monolayer behavior at room temperature (Figure 5). The phase transition becomes gradually less developed when the PFOA concentration in the subphase increases and finally disappears for the highest PFOA concentration (10-3 M) (Figure 5). The conclusions drawn from the analysis of the influence of PFOA on DMPC isotherms are confirmed by measurements of the Cs-1 coefficient values (inset in Figure 5). When the PFOA concentration in the subphase increases, the maximum values of Cs-1 decrease from approximately 155 mN/m for the DMPC monolayer on water (without the presence of PFOA) to 150, 135, and 80 mN/m for 10-5, 10-4, and 10-3 M PFOA concentrations in the subphase, respectively. At the same time, the minimum, which indicates the phase transition from LE to LC becomes poorly developed and disappears when the PFOA concentration is 10-3 M. The comparison of the results obtained for both studied PFCs confirms that PFOS interactions with the DMPC monolayer are much stronger than those of PFOA (Table 3). The difference in A0 is significant when we compare the values obtained for

Figure 6. Comparison of the (a) surface potential and (b) dipole moment for DMPC monolayers on a subphase containing PFOA and PFOS (T ) 22 °C).

the same concentrations of the two perfluorinated compounds. πcoll and Acoll obtained for DMPC isotherms in the presence of PFOS are in good agreement with the values obtained for DMPC on water. Significant changes may also be observed in the surface potential (∆V) and dipole moment (µ) obtained for the two perfluorinated compounds. A semiquantitative analysis of the surface potential isotherm was made using the Helmholtz equation:44,48,49

µ ) 0(∆V)A where µ is the vertical component of the dipole moment, A is the area per molecule, and  and 0 are the permittivities of the monolayer and of a vacuum, respectively. The maximum apparent dipole moment is calculated as µA) µ/. The onset of the increase in surface potential in the presence of perfluorinated compounds is shifted toward larger areas per molecule, and in the case of PFOS ∆V becomes finally smaller than in the absence of the fluorinated compound (Figure 6). The same difference may be observed for the dipole moment

TABLE 3: Characteristic Parameters for DMPC Langmuir Monolayers on the Subphase Containing Different PFCs subphase

A0 (Å2)

Acoll (Å2)

πcoll (mN/m)

dipole momenta (D)

mean molecular areab (Å2)

water 10-4 M PFOA 10-4 M PFOS

73.0 ( 0.3 82.6 ( 0.8 92.7 ( 0.8

43.9 ( 0.5 40.8 ( 0.9 44.8 ( 0.8

47.0 ( 0.6 49.5 ( 0.9 47.6 ( 1.1

1.03 ( 0.05 0.94 ( 0.05 0.60 ( 0.10

170 ( 5 195 ( 10 290 ( 15

a Dipole moment defined as the difference between the maximum value of the dipole moment and the dipole moment at the beginning of monolayer compression. b Area per molecule corresponding to the maximum value of the dipole moment.

Influence of PFCs on Model Lipid Membranes

Figure 7. Stacked 31P NMR spectra obtained from oriented samples with the bilayer normal at 54.7° with respect to the main magnetic field as a function of temperature. The width of each spectrum is 20 kHz.

defined as the difference between the maximum value of the dipole moment and the dipole moment at the beginning of monolayer compression (Table 3). The maximum value of the dipole moment occurs for PFOS at larger areas per molecule than for PFOA and much larger areas per molecule than for the pure water subphase. It is also lowest in the case of PFOS. Such a shift of the onset of the surface potential and dipole moment suggests that when PFCs are present in the subphase, the reorientation of molecules while the monolayer is formed is promoted and appears quicker, that is, at greater areas per molecule. Moreover, a decrease in the value of ∆V and the dipole moment can be observed. Such phenomena may be attributed to the negative contribution from C-F dipoles of the PFC molecules incorporated into the monolayer.50,51 The observed changes are much more pronounced in the case of PFOS, which confirms stronger interactions of this compound with lipid membranes.5 Influence of PFOA on the Gel-Liquid Crystalline Phase Transition Temperature of DMPC Bilayers. Figure 7 displays stacked 31P NMR spectra of bilayer samples with the bilayer normal at 54.7° with respect to the main magnetic field. Each row consists of 11 spectra, corresponding to the temperatures 10-40 °C in 3 °C steps. The amount of PFOA increases from bottom to top as indicated in the figure. The spectra are the sum of two types of line shapes. The so-called powder pattern arising from the unoriented part of the lamellar phase gives rise to a broad line shape with a high-field peak and a low-field shoulder,31 and superposed on this is a sharp line from the oriented bilayers. The frequency of this signal is dependent on the angle between the bilayer normal and the main magnetic field, and for bilayers oriented at 54.7° the signal is positioned at the isotropic chemical shift. Due to the high degree of orientation of the bilayers, the powder line shape is almost buried in the noise; see, e.g., Figure 1 in ref 22 for an example of a less well oriented sample. The phase transition is clearly noticeable by the disappearance of the narrow line corresponding to the oriented part of the bilayers, caused by an excessive line broadening as the gel phase is formed. For the pure DMPC bilayer the phase transition occurs between 22 and 25 °C, in agreement with previous studies on oriented bilayers.24 An addition of 2 mol % PFOA gives no visible change in the phase transition, but for higher amounts of PFOA the phase transition gradually shifts toward lower temperatures. At 10% PFOA the narrow peak is still visible at 16 °C, and for 20% PFOA a narrow signal is present all the way down to 10 °C. This shows that PFOA disrupts the packing

J. Phys. Chem. B, Vol. 111, No. 33, 2007 9913

Figure 8. 2H NMR spectra of macroscopically oriented DMPC bilayers with varying amounts of PFOA. The bilayer normal is parallel to the main magnetic field. The temperature increases in 3 °C steps from 10 (bottom) to 40 (top) °C.

Figure 9. 2H NMR spectra of macroscopically oriented DMPC bilayers with varying amounts of PFOA at 28 °C. The bilayer normal is parallel to the main magnetic field.

properties of the bilayers such that the highly packed gel phase is gradually suppressed as the PFOA concentration is increased. The same information can also be obtained from the 2H spectra of DMPC-d54 shown in Figure 8. The spectra at high temperatures consist of a series of Pake doublets, each corresponding to an individual C-2H2 segment. As the gel phase is entered at lower temperatures, the spectra change into broad, featureless line shapes with a much larger splitting. The phase transition can thus easily be detected and is found to lie between 19 and 22 °C for pure DMPC. The difference in transition temperature, compared to that found with 31P NMR, is related to the labeling of the acyl chains, which lowers the transition by approximately 2-3 °C. When PFOA is added, the transition is broadened and lowered in the same way as previously seen by 31P NMR. Influence of PFOA on the Order Parameters of DMPC Bilayers. The order parameter (SCD) can be calculated from the quadrupole splittings in the deuterium spectra, and the obtained order parameters are in good agreement with earlier published data for DMPC bilayers.52-54 In Figure 9 the spectra at 28 °C are shown for DMPC bilayers with varying amounts of PFOA. As PFOA is incorporated into the bilayers the splittings increase slightly. It is in general not possible to assign the splittings of perdeuterated lipids to individual C-2H bond segments. Studies using lipids deuterated at specific positions in the lipid chain have revealed that, apart from small deviations near the headgroup, SCD decreases monotonically with increasing carbon number. Furthermore, the order parameter profile exhibits a plateau of nearly constant values, corresponding to the maximal splittings obtained in the spectra. This plateau is usually assigned to the C-2H segments adjacent to the carbonyl; e.g., segment

9914 J. Phys. Chem. B, Vol. 111, No. 33, 2007

Matyszewska et al.

Figure 11. DL obtained for PFOA in DMPC bilayers as a function of temperature for three different PFOA contents.

Figure 10. DL of DMPC as a function of temperature and PFOA concentration. Inset: DL of DMPC as a function of the PFOA concentration at 31 °C.

numbers 3-8 of the sn-2 chain of DMPC all have order parameters between 0.221 and 0.234 at 30 °C.54 As a quantitative measure of the changes in the overall order of the bilayers we take the maximal splitting in each spectrum, which gives an order parameter of 0.230 for DMPC, 0.236 for DMPC + 5% PFOA, and 0.239 for DMPC + 10% PFOA. The increase in order induced by PFOA is thus moderate. Influence of PFOA on the Lipid Lateral Diffusion Coefficient in DMPC Bilayers. Figure 10 shows the DL values of DMPC obtained for different temperatures and PFOA concentrations. The DL values obtained for DMPC in the absence of PFOA are in good agreement with earlier published data.24 DL increases monotonically with temperature but shows a more complex behavior with PFOA content. For small additions of PFOA, DL decreases somewhat, but for PFOA additions of more than 5 mol % an increase in DL is observed. This behavior is more or less independent of temperature and is highlighted for 31 °C in the inset in which the error bars denote 1 standard deviation based on two experiments with different diffusion times. The lateral diffusion of membrane-bound species is sensitive to the packing of the lipid chains such that the diffusion coefficient is increased if the packing gets disrupted. According to the free volume model put forward by Almeida et al.,55 the lateral diffusion decreases as the free volume in the bilayers decreases. This means that changes in the packing of the lipid bilayers are accompanied by changes in the lateral diffusion of the lipids. This approach has been used to explain lateral diffusion behavior in systems containing phospholipids and cholesterol.23 Using the same approach in the present system indicates that PFOA included in the membrane at low concentrations actually increases the lipid packing, but as the concentration is increased further a disordering effect is seen. This behavior is different from that reported by deuterium NMR, in which the order was found to increase slightly with increasing PFOA content. The reason for this discrepancy is not clear. One explanation might be that factors other than the free volume are important in the diffusion process in systems containing fluorinated chains, which differ in flexibility and hydrophobicity from the hydrocarbon analogues. Further studies are needed to make any conclusions on this.

The diffusion coefficients of PFOA are found to be about 50% larger than the corresponding values for DMPC (Figure 11). This is quite surprising since in general all molecules in a bilayer have the same lateral diffusion coefficient, reflecting the fact that the lateral diffusion is governed by the physicochemical properties of the bilayer as a whole, and not by the properties of the individual lipids.56 However, as mentioned in the above paragraph, perfluorinated molecules might have a behavior different from that of the corresponding hydrocarbons. Futhermore, it is not clear whether the larger value of DL for PFOA is really an effect of a larger lateral diffusion of PFOA since it is expected to partition into the water phase to a significant degree. The observed diffusion coefficient would then be a weighted average of the diffusion in the two phases, DLobsd ) P1DL1 + (1 - P1)DL2. Since one would expect the diffusion in the water phase to be much faster, the observed diffusion would be larger than that for a truly membrane bound molecule. Using the value of 600 µm2/s for diffusion of PFOA in a water solution at 25 °C (unpublished data), and the values for the lipid diffusion for diffusion of PFOA in the bilayer, the fraction of PFOA in the water phase is found to be less than 1%. This shows that even a small amount of PFOA in the water phase can have a rather large influence on the obtained diffusion coefficient. Penetration of PFOS into DPPC bilayers: Molecular Dynamics Simulations. MD simulations of a DPPC bilayer in pure water and in water with 0.15 M (7.7 wt %) and 0.23 M (11.6 wt %) PFOS were performed. Figure 12 shows the time evolution of the partial mass density distributions of DPPC and PFOS across the simulation box perpendicular to the bilayer (in the z direction) for 0.15 and 0.23 M PFOS at 27 °C (300 K). The average mass density of DPPC in pure water (no PFOS) is also shown for comparison. No clear correlation can be detected between the density of DPPC and the concentration (or existence) of PFOS for the concentrations and run times studied. However, in both cases, the PFOS molecules slowly penetrate the lipid bilayer from the water phase where they were initially embedded. This is also obvious from the snapshot of the trajectory of a DPPC simulation in water with 0.23 M PFOS at 27 °C shown in Figure 13a. Simulations with the higher concentration of PFOS were repeated at 52 °C (325 K), where the lipid is known to exhibit the biologically most relevant fluid phase.57 Figure 14 depicts the mass density distributions at different time intervals for those simulations. The results indicate that the more fluidic phase of the lipid makes the drift of PFOS into the bilayer easier. The difference can be also seen in Figure 13, which presents the frames taken at the time point of 50 ns from the trajectories of the simulations at 27 and 52 °C: PFOS molecules have entered deeper into the bilayer at 52 °C. The results also suggest that

Influence of PFCs on Model Lipid Membranes

J. Phys. Chem. B, Vol. 111, No. 33, 2007 9915

Figure 12. Time evolution of the partial mass densities across the simulation box perpendicular to the bilayer (in the z direction) for (a) DPPC in water with 0.15 M PFOS and (b) DPPC in water with 0.23 M PFOS at 27 °C. The M-shaped curves correspond to DPPC and the smaller hills close to the horizontal axis to PFOS. The average mass density of DPPC in pure water (no PFOS) is also shown for comparison. The distribution of water is not shown.

the main features of the PFOS distributions in the bilayers are typically reached in about 30-40 ns. A kind of “area per lipid” values were calculated simply by dividing the x-y plane of the simulation box by the number of lipid molecules in one leaflet independent of whether there are PFOS molecules also incorporated into the bilayer. Comparison of the areas per lipid for the system of 0.23 M PFOS at the two temperatures shows an increase from 66 ( 1 Å2 at 27 °C to 69 ( 1 Å2 at 52 °C, which are similar to the corresponding values of the bilayer in pure water. This suggests that when PFOS is incorporated into the bilayer, the system kept under constant pressure does not expand, but appears more compressed. This can also be seen in Figure 15, where the lipid tail deuterium order parameter (SCD)39 is depicted as a function of the carbon atom along the lipid tails for the initial 0.23 M PFOS concentration corresponding to 17 mol % PFOS in the lipid bilayer. The Langmuir monolayer experiments, where the conditions most closely resembled those of the MD simulations, exhibited consistent results showing an increase in the surface pressure after injection of the PFOS solution into the subphase on which the DPPC monolayer was first compressed (Figure 3). Both of the results suggest improved packing of the lipids when PFOS penetrates a preformed DPPC layer; i.e., PFOS tends to force the dipalmitoyl chains of DPPC to decrease the average tilt angle and align closer to the normal of the layer

Figure 13. Cross-sections of the simulation box perpendicular to the plane of the bilayer: snapshots of the simulations of the DPPC bilayer (vertically in the center) in water with 0.23 M PFOS (in green) at the time point of 50 ns (a) at 27 °C and (b) at 52 °C.

surface. Something similar was previously observed for perfluorinated carboxylic acids with DPPC.58 To conclude, the computational results presented above are in accordance with the experimental findings of the Langmuir monolayer studies. Both methods suggest that PFOS is able to penetrate DPPC layers and thus may be potentially harmful to wildlife1-3 as well as to humans. Penetration of PFOA into DMPC Bilayers: Molecular Dynamics Simulations. MD simulations of a DMPC bilayer in pure water and in water with 0.07 M (2.7 wt %) and 0.14 M (5.8 wt %) PFOA were also performed at 27 °C (300 K) and at 7 °C (280 K). The time evolutions of the partial mass density distributions of DMPC and PFOA across the simulation box were very similar to those obtained for DPPC and PFOS in Figures 12-14, showing a fast incorporation of PFOA into the DMPC bilayer (Figure 16). Different from the DPPC/PFOS system, however, a not so clear influence of temperature on the ease of the diffusion of PFOA into the bilayer was observed for the two temperatures studied in the simulations, but a

9916 J. Phys. Chem. B, Vol. 111, No. 33, 2007

Matyszewska et al.

Figure 14. Time evolution of the partial mass densities across the simulation box perpendicular to the bilayer (in the z direction) for DPPC (M-shaped curves) in water with 0.23 M PFOS (smaller hills close to the horizontal axis) at 52 °C. The average mass density of DPPC in pure water (no PFOS) is also shown for comparison. The distribution of water is not shown.

Figure 15. Deuterium order parameter (SDC) for the phospholipid tails for DPPC under different conditions at 52 °C (325 K). The 17 mol % PFOS in DPPC corresponds to the initial concentration of 0.23 M in water. The order parameter is calculated as the average over the two tails and over the time intervals displayed in the legend.

relatively fast incorporation was detected in both cases. Interestingly, the incorporation of PFOA into the DMPC bilayer appeared to increase the order of atoms 9-13 of the phospholipid tails for 5 mol % PFOA but decrease that of atoms 2-8 in all the conditions studied (Figure 17), while no change was observed in the order of atoms 10-13 for 10 mol % PFOA. This kind of position-dependent behavior is not observed in the NMR results shown above. Furthermore, the magnitudes of the order parameters are significantly smaller than those obtained by the NMR experiments. The reason for these differences between the experimental and computational deuterium order parameters with PFOA is not clear at the moment. However, the overall result of the simulations showing a tendency of PFOA to incorporate into DMPC bilayers and influence the lipid properties is qualitatively consistent with the experimental results. It should be pointed out that the force field used for modeling the lipids39,40 was optimized for the liquid crystalline phase, i.e., for the biologically most relevant fluid phase. This means that the phase transitions from the liquid crystalline phase to the gel phase cannot be expected to be properly reproduced,

Figure 16. Time evolution of the partial mass densities across the simulation box perpendicular to the bilayer (in the z direction) for DMPC in water with 0.14 M PFOA at (a) 7 °C and (b) 27 °C. The M-shaped curves correspond to DMPC and the smaller hills close to the horizontal axis to PFOA. The average mass density of DMPC in pure water (no PFOA) is also shown for comparison. The distribution of water is not shown.

Figure 17. Deuterium order parameter (SDC) for the phospholipid tails for DMPC under different conditions at 27 °C (300 K). The 5 and 10 mol % PFOA in DMPC correspond to the initial concentrations of 0.07 and 0.14 M in water, respectively. The order parameter is calculated as the average over the two tails.

but the lipids are forced to remain longer in a kind of fluidic phase as the temperature decreases. This was also seen in the simulations performed (see, e.g., Figure 13) and should be

Influence of PFCs on Model Lipid Membranes considered when the results of the simulations are compared with those of the experiments. This is also why the deuterium order parameters of DPPC and DMPC (in Figures 15 and 17) are only shown for the temperatures that are above the known phase transition values and can, thus, be considered more reliable. When the order parameters obtained for the pure DMPC and DPPC systems are compared with earlier studies, it can be seen that the values of Figures 15 and 17 are somewhat smaller than the corresponding values of the experimental works54,59,60 but in very good accordance with the results of the computational papers (see, e.g., ref 61 and references therein). The differences between the experimental and computational deuterium order parameters are most likely related to the detailed parameters of the nonbonded interactions of the force field including the Lennard-Jones cutoff distance used. The earlier computational papers typically consider such computational and experimental results to be in good agreement. Conclusions The PFCs in the subphase exert a strong influence on the properties of the DPPC and DMPC monolayers such as their fluidity and phase transitions. PFOS seems to affect lipids stronger than PFOA. The phase transition from liquid to solid becomes less developed with increasing concentration of the perfluorinated compound in the subphase and finally disappears for high concentrations. As a result, the monolayer becomes more fluid. These changes may be observed in the isotherm shape and also in the compression modulus (Cs-1 coefficient) values. After the injection of PFOS into the subphase, on which a DPPC monolayer was first compressed, the surface pressure increases and then becomes constant in time, reflecting a stabilizing effect due to slow incorporation of PFOS molecules into the layer. Thus, although the perfluorinated compound incorporated into the monolayer during its formation makes the layer more liquid, its incorporation into a preformed layer can also improve the packing and prevent the layer from destabilization at a longer time scale. The pfg-NMR studies of the DMPC oriented bilayers reveal that, upon an addition of PFOA to the DMPC bilayers, the lateral diffusion of DMPC decreases for small amounts of PFOA, indicating a condensing effect of PFOA on the membrane. At larger concentrations of PFOA (above 5 mol %) the lateral diffusion coefficient increases due to an increased fluidity of the bilayer. 31P and 2H NMR shows that the gel-liquid crystalline phase transition temperature decreases by the addition of PFOA for concentrations above 5 mol %. The phase transition temperature is also broadened. This shows that PFOA disrupts the packing properties of the bilayers in such a way that the highly packed gel phase is gradually suppressed with increasing PFOA concentration. The deuterium order parameters are found to increase only slightly upon incorporation of PFOA into the bilayers. The experimental observations are in qualitative agreement with the computational results obtained from the molecular dynamics simulations showing a slow migration of PFCs from the surrounding water phase into the already preformed DPPC and DMPC bilayers, leading to changes in their properties. This preliminary study on the effect of PFCs on the phase behavior of model membranes utilizes several methods, each reporting on both structural and dynamic properties of the systems. The investigations performed by different methods may as shown here lead sometimes to contradictory results. The discrepancies, underlined in our paper, emphasize the complexity of the problem and lack of full understanding of the behavior

J. Phys. Chem. B, Vol. 111, No. 33, 2007 9917 of perfluorinated compounds in biological and biomimetic systems. They point out the need for further in-depth studies to meet the environmental threats of these compounds. Acknowledgment. This work was financially supported by BBMO Project 78-UE-005199, Polish Ministry of Scientific Research, Information Technology Project No. PBZ 18-KBN098/T09/2003, and the Knut and Alice Wallenberg Foundation. References and Notes (1) Giesy, J. P.; Kannan, K. EnViron. Sci. Technol. 2001, 35, 13391342. (2) Hoff, P. T.; Van de Vijver, K.; Van Dongen, W.; Esmans, E. L.; Blust, R.; De Coen, W. M. EnViron. Toxicol. Chem. 2003, 22, 608-614. (3) Lau, C.; Butenhoff, J. L.; Rogers, J. M. Toxicol. Appl. Pharmacol. 2004, 198, 231-241. (4) Hu, W.; Jones, P. D.; De Coen, W.; King, L.; Fraker, P.; Newsted, J.; Giesy, J. P. Comp. Biochem. Physiol., C 2003, 135, 77-88. (5) Harada, K.; Xu, F.; Ono, K.; Iijima, T.; Kozumi, A. Biochem. Biophys. Res. Commun. 2005, 329, 487-494. (6) Lau, C.; Butenhoff, J. L.; Rogers, J. M. Toxicol. Appl. Pharmacol. 2004, 198, 231-241. (7) Maget-Dana, R. Biochim. Biophys. Acta 1999, 1462, 109-140. (8) Brezesinski, G.; Mohwald, H. AdV. Colloid Interface Sci. 2003, 100-102, 563-584. (9) Deleu, M.; Paquot, M.; Nylander, T. J. Colloid Interface Sci. 2005, 283, 358-365. (10) Hidalgo, A. A.; Caetano, W.; Tabak, M.; Oliveira, O. N., Jr. Biophys. Chem. 2004, 109, 85-104. (11) Souza, S. M. B.; Oliveira, O. N., Jr.; Scarpa, M. V.; Oliveira, A. G. Colloids Surf., B 2004, 36, 13-17. (12) Ha¸ c-Wydro, K.; Dynarowicz-Ła¸ tka, P. Colloids Surf., B 2006, 53, 64-71. (13) Jabłonowska, E.; Bilewicz, R. Thin Solid Films 2007, 515, 39623966. (14) Klopper, K. J.; Vanderlick, T. K. J. Colloid Interface Sci. 1996, 182, 220-229. (15) McConlogue, C. W.; Malamud, D.; Vanderlick, T. K. Biochim. Biophys. Acta 1998, 1372, 124-134. (16) Li, J. B.; Miller, R.; Volhardt, D.; Mohwald, H. Thin Solid Films 1998, 327-329, 84-86. (17) Lawrie, G. A.; Gentle, I. R.; Barnes, G. T. Colloids Surf., A 2000, 171, 217-224. (18) Minones, J.; Rodriguez, Patino, J. M.; Conde, O.; Carrera, C.; Seoane, R. Colloids Surf., A 2002, 203, 273-286. (19) Jyoti, A.; Prokop, R. M.; Li, J.; Volhardt, D.; Kwok, D. Y.; Miller, R.; Mohwald, H.; Neumann, A. W. Colloids Surf., A 1996, 116, 173-180. (20) Shao, L.; Feng, S. S. J. Colloid Interface Sci. 2004, 274, 55-68. (21) Gaboriaud, F.; Volinsky, R.; Bermana, A.; Jelinek, R. J. Colloid Interface Sci. 2005, 287, 191-197. (22) Ora¨dd, G.; Lindblom, G. Magn. Reson. Chem. 2004, 42, 123131. (23) Lindblom, G.; Ora¨dd, G.; Filippov, A. Chem. Phys. Lipids 2006, 141, 179-184. (24) Ora¨dd, G.; Lindblom, G. Biophys. J. 2004, 87, 980-987. (25) Langmuir, I. J. Am. Chem. Soc. 1917, 39, 1848-1906. (26) Nagle, J .F. J. Membr. Biol. 1976, 27, 233-250. (27) Davis, J. H. Biochim. Biophys. Acta 1983, 737, 117-171. (28) Raleigh, D. P.; Olejniczak, E. T.; Griffin, R. G. J. Magn. Reson. 1989, 81, 455-463. (29) Tanner, J. E. J. Chem. Phys. 1970, 52, 2523-2526. (30) Shaka, A. J.; Keeler, J.; Frenkiel, T.; Freeman, R. J. Magn. Reson. 1983, 52, 335-338. (31) Seelig, J. Biochim. Biophys. Acta 1978, 515, 105-140. (32) Lindahl, E.; Hell, B.; van der Spoel, D. J. Mol. Model. 2001, 7, 306-317. (33) Berendsen, H. J. C.; Potsma, J. P. M.; Gunsteren, W. F.; Hermans, J. Intermolecular Forces; D. Reidel Publishing Co.: Dordrecht, The Netherlands, 1981; pp 331-342. (34) Berendsen, H. J. C.; Potsma, J. P. M.; Gunsteren, W. F.; DiNola, A.; Haak, J. R. J. Chem. Phys. 1984, 18, 3684-3690. (35) Hess, B.; Beker, H.; Berendsen, H. J. C.; Fraaije, J. G. E. M. J. Comput. Chem. 1997, 18, 1463-1472. (36) Miyamoto, S.; Kollman, P. A. J. Comput. Chem. 1992, 13, 952962. (37) Darden, T.; York, D.; Pedersen, L. J. Chem. Phys. 1993, 98, 1008910092. (38) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. J. Chem. Phys. 1995, 103, 8577-8593.

9918 J. Phys. Chem. B, Vol. 111, No. 33, 2007 (39) Tieleman, D. P.; Berendsen, H. J. C. J. Chem. Phys. 1996, 105, 4871-4880. (40) Berger, O.; Edholm, O.; Jahning, A. Biophys. J. 1997, 72, 20022013. (41) van Aalten, D. M. F.; Bywater, R.; Findlay, J. B. C.; Hendlich, M.; Hooft, R. W. W.; Vriend, G. J. Comput.-Aided Mol. Des. 1996, 10, 255-262. (42) van Gunsteren, W. F.; Daura, X.; Mark, A. Encyclopedia of ComputationalChemistry; John Wiley & Sons: New York, 1998; p 1211. (43) van Gunsteren, W. F.; Berendsen, H. J. GROMOS87 Manual; Biomos BV: Groningen, The Netherlands, 1987. (44) Davies, J. T.; Rideal, E. K. Interfacial Phenomena, 2nd ed.; Academic Press: New York, 1963; p 265. (45) Harkins, W. D. The Physical Chemistry of Surface Films; Reinhold: New York, 1952; p 107. (46) Reis, T. Introduction a la Chimie Physique des Surface; Dunod: Paris, 1952; p 111. (47) Ocko, B. M.; Wu, X. Z.; Sirota, E. B.; Sinha, S. K.; Gang, O.; Deutsch, M. Phys. ReV. E 1997, 55, 3164-3182. (48) Broniatowski, M.; Sandez Macho, I.; Dynarowicz-Ła¸ tka, P. Thin Solid Films 2005, 493, 249-257. (49) Leite, V. B. P.; Cavalli, A.; Oliveira, O. N., Jr. Phys. ReV. E 1998, 57, 6835-6839. (50) Cavalli, A.; Dhanabalan, A.; Giacometti, J. A.; Oliveira, O. N., Jr. Synth. Met. 1999, 102, 1411.

Matyszewska et al. (51) Hendlinger, P.; Laschewsky, A.; Bertrand, P.; Delcorte, A.; Legras, R.; Nysten, B.; Mo¨bius, D. Langmuir 1997, 13, 310-319. (52) Lu, D.; Vavasour, I.; Morrow, M. R. Biophys. J. 1995, 68, 574583. (53) Trouard, T. P.; Nevzorov, A. A.; Alam, T. M.; Job, C.; Zajizek, J.; Brown, M. F. J. Chem. Phys. 1999, 110, 8802-8818. (54) Douliez, J.-P.; Le´onard, A.; Dufourc, E. J. Biophys. J. 1995, 68, 1727-1739. (55) Almeida, P. F. F.; Vaz, W. L. C.; Thompson, T. E. Biochemistry 1992, 31, 6739-6747. (56) Ora¨dd, G.; Westerman, P. W.; Lindblom, G. Biophys. J. 2005, 89, 315-320. (57) Pabst, G.; Amenitsch, H.; Kharakoz, D. P.; Laggner, P.; Rappolt, M. Phys. ReV. E 2004, 70, 021908. (58) Nakahara, H.; Nakamura, S.; Kawasaki, H.; Shibata, O. Colloids Surf., B 2005, 41, 285-298. (59) Brown, M. F.; Seelig, J.; Ha¨berlen, U. J. Chem. Phys. 1974, 70, 5045-5053. (60) Petrache, H. I; Dodd S. W.; Brown, M. F. Biophys. J. 2000, 79, 3172-3192. (61) Repa´lova´, J.; Holopainen, J. M.; Morrow, M. R.; McDonald, M. C.; Capkova´, P.; Vattulainen, I. Biophys. J. 2005, 88, 3398-3410.