Effect of Ionic Strength on Dynamics of Supported Phosphatidylcholine

Apr 15, 2013 - Lipid bilayers are generally thought of as a structural matrix for different .... Finding a constant value of D over a large q range co...
2 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/Langmuir

Effect of Ionic Strength on Dynamics of Supported Phosphatidylcholine Lipid Bilayer Revealed by FRAPP and Langmuir−Blodgett Transfer Ratios Frédéric F. Harb and Bernard Tinland* CNRS, UMR7325, 13288, Marseille, France, and Aix-Marseille Univ., CINaM, 13288, Marseille, France S Supporting Information *

ABSTRACT: To determine how lipid bilayer/support interactions are affected by ionic strength, we carried out lipid diffusion coefficient measurements by fluorescence recovery after patterned photobleaching (FRAPP) and transfer ratio measurements using a Langmuir balance on supported bilayers of phosphatidylcholine lipids. The main effect of increasing ionic strength is shown to be enhanced diffusion of the lipids due to a decrease in the electrostatic interaction between the bilayer and the support. We experimentally confirm that the two main parameters governing bilayer behavior are electrostatic interaction and bilayer/support distance. Both these parameters can therefore be used to vary the potential that acts on the bilayer. Additionally, our findings show that FRAPP is an extremely sensitive tool to study interaction effects: here, variations in diffusion coefficient as well as the presence or absence of leaflet decoupling. a negative ζ potential in Milli-Q water. Zimmermann et al. studied the fluidity modulation of phospholipid bilayers by electrolyte ions and pH.22 Here, we systematically examined the effect of ionic strength on the dynamics of supported phosphatidylcholine lipid bilayers. For a unilamellar supported lipid bilayer (SLB) prepared using Langmuir−Blodgett deposition, we measured the variation in diffusion coefficient and transfer ratio of lipids on mica and glass, using increasing salt concentrations. Diffusion coefficients as a function of ionic strength were also measured for double and compared to multibilayer behavior.

1. INTRODUCTION Lipid bilayers are generally thought of as a structural matrix for different molecules like proteins and cholesterol. Cell membranes control how small molecules, water, and ions diffuse through them. The structure and behavior of lipids are affected by pH, ionic strength, and ion type in solution.1,2 The ions most commonly encountered are Na+, K+, Cl−, Ca2+, and Mg2+. Divalent ions like Ca2+ play a role in mitochondrial membranes, while monovalent ions like Na+ and K+ modulate plasma membrane properties.3−5 Recent studies show that monovalent ions can have a significant impact on lipid bilayers, i.e., increasing lipid−lipid interaction and membrane compression.6−9 Garcia-Maynes et al.7 detected this interaction by measuring nanoindentation forces using an AFM tip. Other studies examine the membrane potential through lipid mobility as a function of ionic strength.10,11 Results show that cations are deeply inserted in the membrane, reaching the carbonylated region and forming complexes with lipids. This cation−lipid link impacts the structural properties and dynamics of phosphocholine membranes: a decrease in the diffusion coefficient is predicted and experimentally observed.9,12−14 The effect of salt and charge is still under debate; a monovalent cation has been found to bind 3 lipids,9 while other studies show that 2−9 ions bind 100 lipids.15,16 The bilayers in question are either free-standing bilayers (vesicles or multibilayer) or, more rarely, supported lipid bilayers. Here we used PC-type lipids, which bear a phosphatidylcholine headgroup. This zwitterionic group is globally uncharged according to some authors,17−19 but according to others,7,20,21 it gives rise to © 2013 American Chemical Society

2. MATERIALS AND METHODS 2.1. Sample Preparation. Phospholipids were purchased from Avanti Polar Lipids and used without further purification. SLBs were prepared using 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) or 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC). Supports were either glass plates (Marienfield, cut edges, France) for single supported bilayers or mica slides for n-bilayers (JBG-Metafix, France). Mica slides were cleaved on both sides just before use. Glass slides were cleaned by immersion for 10 min in freshly prepared alcoholic NaOH (followed by thorough Milli-Q water rinse) and then fully rinsed and sonicated 3 times for 5 min in ultrapure water. For the measurement of phospholipid diffusion, a small fraction of fluorescently labeled phospholipids were incorporated in the samples (0.1%) for fluorescence recovery experiments. We used fatty acid Received: December 15, 2012 Revised: April 13, 2013 Published: April 15, 2013 5540

dx.doi.org/10.1021/la304962n | Langmuir 2013, 29, 5540−5546

Langmuir

Article

different temperatures, consistently finding a linear dependence of time on q2 (data not shown). Finding a constant value of D over a large q range confirms that, as often reported, the Langmuir−Blodgett balance used under appropriate conditions leads to defect-free bilayers on a large scale. More details about the FRAPP setup and typical signals are shown in the Supporting Information.

labeled lipids with NBD (nitrobenzoxadiazole): 16:0−12:0 NBD PC for DPPC. 2.2. Langmuir−Blodgett Deposition. The Langmuir−Blodgett (LB) transfer technique was used to prepare bilayers supported on planar hydrophilic substrates.23 Lipid molecules solubilized in chloroform (1 mg/mL) were deposited on the water subphase (18 MΩ·cm, Milli-Q, pH ≈ 6, 15 °C) of a Langmuir trough (KSV Minitrough 361 mm × 74 mm, Finland), equilibrated for 10 min to allow complete solvent evaporation, and then compressed to 42 mN/ m. The first, second, or third (in the case of double bilayers) monolayer of lipids was transferred to the substrate by pulling the support (25 mm × 25 mm) up and down at a speed of 5 mm/min. The fourth monolayer (in the case of double bilayers), closing the bilayer, was transferred according to the out-of-equilibrium Langmuir− Schaefer (LS) technique: the “hydrophobic” substrate in air was rotated to the horizontal position and lowered through the interfacial monolayer (for details, see ref 24). 2.3. Transfer Ratio. The Langmuir balance allows the quantitative determination of the transfer ratio of the lipid from the surface of the trough to the support. It is defined as the decrease in water−air monolayer area during depositing divided by the area of the substrate pulled through the monolayer. It strongly influences the quality of samples, as shown by structural studies based on AFM25 or neutrons and X-ray reflectivity measurements.24 Its value represents an indirect evaluation of the bilayer/support interaction. A high transfer ratio also indicates successful layer-by-layer depositing on a solid substrate. Additionally, a global transfer ratio value equal to its instantaneous value continuously measured during the whole deposition process indicates defect-free deposition (see examples in the Supporting Information). SLBs were kept under water throughout and used directly after preparation. The transfer ratio as a function of ionic strength was assessed by replacing water with salt solution in the subphase. 2.4. Multibilayer Preparation. The multibilayer was formed via the lipid film hydration method. 150 μL of lipids in chloroform (2 mg/ mL) was first desiccated under house vacuum for 1 h on a mica support, the whole contained in a well-sealed cell. Then the films were rehydrated at room temperature in ultrapure water (18 MΩ·cm, MilliQ) for 1 h. For DPPC, assuming a lipid head molecular area of 0.63 nm2, we estimated from the initial quantity of lipid a number of bilayers close to 500 in the multibilayer. 2.5. FRAPP Measurements. We used fluorescence recovery after pattern photobleaching (FRAPP)26 to investigate the diffusion of lipids in the SLB. Briefly, the light beam of an etalon-stabilized monomode Ar laser (1 W at 488 nm) was split, and the two equivalent beams crossed on the sample, providing an interference fringe pattern. The fringe spacing i = 2π/q (set by the crossing angle θ, q = 4π/λ sin(θ/2) is the wave vector) ranged from 1 to 80 μm and defined the diffusion distance. A bleach pulse set to 3 s with 0.5 W laser intensity wrote a fringe pattern into the sample, which, due to diffusion, disappeared with time. After the bleach pulse, the beam intensity is reduced to a few milliwatts: under these circumstances, we observed no further bleaching of the NBD labeled lipid. Unless specified otherwise, data were fitted to a single-exponential e−t/τq, and D was calculated from τq = 1/Dq2. In a previous study23 on the diffusivity of SLB in water on glass and on mica, we detailed how FRAPP can be used to discriminate mono- or biexponential systems corresponding to leaflet decoupling. Error bars in figures refer to standard deviation obtained from diffusion measurements at different temperatures, each consisting of 5−7 measurements. The following two special features of the setup are worthy of note. (i) The use of the same periodic sinusoidal pattern for bleaching and reading allows a single mode of the diffusion equation to be probed. The analysis is more precise, and thus it is easier to detect small variations in the diffusion coefficient or to discriminate between different populations. (ii) The experimental setup also makes it possible to easily vary the interfringe value and to directly check the validity of the diffusion law in the reciprocal space. To test whether the diffusion was Brownian or not, we carried out measurements for differing fringe widths on different systems at

3. RESULTS AND DISCUSSION 3.1. Transfer Ratio. Figure 1 shows the transfer ratio of DPPC on mica for unilamellar supported bilayers. The

Figure 1. Transfer ratio of DPPC on mica as a function of the total ionic strength of the subphase. Tags point marks where ionic strength results from NaCl + CaCl2. Other points correspond to pure NaCl solutions. Filled red squares and open green squares represent the transfer of the first and the second monolayer, respectively. To shorten x-axis, ionic strength corresponding to transfer ratios in H2O is set at 10−5 mM. The dashed green line is a guide for the eye.

subphase is either ultrapure water (pH = 6) or a saline solution (NaCl or CaCl2) (pH = 7.1). The x-axis represents the equivalent ionic strength. The transfer ratio of the first monolayer is always excellent (≥1), indicating an affinity of the lipids for the surface of the support (red filled squares) which remains sufficiently high to ensure good lipid transfer whatever the ionic strength. The transfer ratio of the second monolayer is close to 1 up to an ionic strength of 10 mM and then decreases drastically to a negative value close to −1. The presence of divalent Ca2+ions, which might have been expected to form bridges between the proximal leaflet and the support, does not modify this behavior, suggesting that this technique is not sufficiently sensitive to detect the potential bridging effect of divalent ions. The curve expresses the change from a situation (low salt, strong interaction) where the lipid head/support interaction is greater than the acyl first layer/acyl second layer interaction, and thus the bilayer can be completed, to the inverse situation, namely: when salt concentration increases, the acyl first layer/ acyl second layer interaction becomes greater than the lipid head/support interaction, resulting in the peeling of the first layer. In other words, even though the strength of the lipid head/substrate interaction decreases with increasing salt concentration, it remains sufficiently high until 10−2 M to continue to ensure good transfer ratios during the first deposition. At ionic strengths greater that 10−2 M, the second monolayer, needed “to close” the bilayer, is not transferred to 5541

dx.doi.org/10.1021/la304962n | Langmuir 2013, 29, 5540−5546

Langmuir

Article

Figure 2. Schematic representation of a supported double bilayer according to the Debye length (as a function of ionic strength).

the surface; on the contrary, the first monolayer is completely cleaned off the surface. This means that the bilayer/support interaction energy decreases as ionic strength increases. Using theoretical models should enable the magnitude of these interactions to be assessed. 3.2. Diffusion Coefficients. To offset the above difficulty in closing the bilayer at high ionic strengths, we decided to build our bilayer from a Langmuir film on a water subphase. Then, the support bearing the bilayer was transferred into a crystallizer (100 mL) filled with the phosphate buffered saline solution (1 mM) at the desired ionic strength and at a pH adjusted to its physiological value (pH ≈ 7). According to Zimmerman et al., 22 the diffusion of the zwitterionic phospholipid does not change between pH 6 and 8. During the course of our experiments, after transfer from the water subphase, two samples were left in an electrolyte solution at the desired concentration, namely 10 mM and 1 M, for differing times (half an hour up to 24 h). We carried out FRAPP measurements at differing incubation times and observed no differences in diffusivity for a given ionic strength. Consequently, we consider that after 30 min for ionic equilibration all the water, including the water film between the bilayer and the support, has been replaced by the salt solution. The SLB was then transferred to the FRAPP cell, itself filled with the same saline solution, to carry out diffusion coefficient measurements. Our observations are complementary to the work of Bockmann et al.,9 who observed that the presence of NaCl divided roughly by two the diffusion coefficient both in the gel and in the fluid regime of BODIPY-DHPE. However, while they carried out their experiments on a multibilayer where the support is expected to play no role, in our experiments, we specifically tested the effect of the presence of the support. The result shows that the lipid/support interaction is far greater than might be expected, going well beyond merely binding some lipids with Na+ in the plane of the bilayer. Figure 2 is a schematic representation (roughly to scale) of the system with a support + water film (e ≅ 1 nm) + a first bilayer + a second bilayer separated from the first by a water layer 2 nm thick.24 Other studies report that interbilayer water thickness can be less than 2 nm, e.g., papers based on neutron reflection.27−30 However, these systems differ greatly from our system in their structure (first monolayer grafted to the substrate, odd number

of layers or curvature and stacking in multilamellar vesicles), which may explain the difference in thicknesses found for the interbilayer water film. Because our double bilayer system is very similar to the system of Charitat et al.,24 we chose 2 nm as an order of magnitude to draw our schematic representation and for comparison with the Debye length. Plotting the Debye length κ−1 on y axis, and assuming that the interaction has greatly decreased after κ−1, the value at which the interaction potential is 1/e, everything happens as if the main interaction holding the bilayer was the electrostatic interaction, in addition to the van der Waals, bilayer/solvent, leaflet/leaflet, and lipid/lipid interactions. All behaviors can consistently be described using the schematic representation in Figure 2. Diffusion Coefficients in Unilamellar Supported Bilayer of DPPC on Mica. Figure 3 shows the dependence of the diffusion coefficient on temperature at differing ionic strengths of a micasupported DPPC bilayer. In water and in 1 and 10 mM NaCl, we observed (i) a decoupling between distal and proximal leaflets23 and (ii) the successive steps of the transition between different phases on mica: gel phase/pretransition/ripple phase/

Figure 3. Diffusion coefficient of DPPC in SLB on mica as a function of temperature and NaCl concentration. The abbreviations PROX and DIST represent the proximal and the distal leaflet, respectively. 5542

dx.doi.org/10.1021/la304962n | Langmuir 2013, 29, 5540−5546

Langmuir

Article

main transition/fluid phase. To our knowledge, this is the first study to provide values for diffusion coefficients over the whole range of physical changes, particularly in the gel phase. The D = f(T) curves of the proximal (PROX) leaflets in water, 1 mM NaCl, and 10 mM NaCl are very similar. The D = f(T) curves of the distal (DIST) leaflets, however, while similar in water and in 1 mM NaCl, show higher diffusion coefficients in 10 mM NaCl. In 100 mM and 1 M NaCl, there is no more decoupling. Diffusion coefficients in the gel phase have increased by several orders of magnitude and are now close to 6 × 10−11 cm2/s. In the fluid phase, curves join together with those of distal leaflets at other ionic strengths. The common value (≈ 6.5 × 10−8 cm2/s) is close to that measured in free-standing bilayers (≈ 10−7 cm2/s) of PC lipids.31,32 Thus, when NaCl is added and ionic strength is increased, leaflet decoupling disappears and the diffusion coefficient increases. We interpret this behavior as follows: In water, κ−1 tends toward infinity, in 1 mM NaCl its value is 9 nm, a single supported bilayer is entirely situated in a strong potential, and the bilayer/support interaction is maximum. Since the proximal leaflet is more affected by the action of the support and its roughness than the distal leaflet, its diffusion coefficient is lower and leaflet decoupling is observed.23 As the NaCl concentration increases, the interaction decreases, and thus the diffusion coefficient increases. In NaCl 10 mM (κ−1 ≈ 3 nm) there is still decoupling, the PC heads of the distal leaflet, which are roughly 5−6 nm away from the support, experience a weaker potential, and the diffusion coefficient of the lipids increases strongly. In NaCl 100 mM and 1 M, κ−1 is respectively 0.9 and 0.3 nm, the potential is weak (the remaining interaction is not zero since the bilayer does not leave the surface as vesicles), but the interaction is no longer strong and differentiated enough to induce leaflet decoupling. The diffusion coefficient in the gel phase has again increased compared to diffusion in all the other ionic strengths. Diffusion Coefficients in Unilamellar Supported Bilayer of DPPC on Glass. Figure 4 shows similar curves, obtained with an SLB of DPPC on glass. The bilayer/glass interaction is weaker.23 Indeed, at the same ionic strength, we observe that diffusion coefficients of lipids on glass are always faster than those on mica, when ionic strengths are low. In 100 mM and 1 M NaCl, however, they tend to gather. Curves in 1 M and 100 mM NaCl on glass and on mica are identical. On glass, in water, we did not observe the decoupling found on mica. While diffusion coefficients increase by 2 orders of magnitude from glass to mica in the gel phase, the difference is only a factor 3 in the fluid phase. The new feature is that we now observe a distal leaflet decoupling in 10 mM NaCl: the diffusion curve of the proximal leaflet remains identical to that in water, while diffusion of the distal leaflet increases. Results from Figure 4 can be interpreted in the same way as on mica, reinforced by the behavior observed in 10 mM NaCl. Although in water, 100 mM, and 1 M NaCl we observed no decoupling for reasons already described (weak interaction of lipids with glass and, in saline solutions, screening of the electrostatic interaction), decoupling is observed in 10 mM NaCl with κ−1 ≈ 3 nm: the proximal leaflet remains in a strong potential, and the distal leaflet (particularly the PC heads which are 5−6 nm away from the support) is in a weaker potential. As a result, they do not have the same diffusion coefficients.

Figure 4. Diffusion coefficients of DPPC in SLB on glass as a function of temperature for differing NaCl concentrations. In the gel phase D varies over several orders of magnitude. In the fluid phase, curves tend to gather. The abbreviations PROX and DIST represent the proximal and the distal leaflet, respectively.

Consequently, by adjusting the ionic strength of the solution, we were able to induce leaflet decoupling. Diffusion Coefficients in Unilamellar Supported Bilayer of DPPC on Mica in Presence of a Divalent Cation Ca2+. Using a (1 M NaCl + 2 mM CaCl2) solution induced no variation in the diffusion coefficient, despite the presence of the divalent cation Ca2+ with its potential to bridge the proximal leaflet and the support. This may be explained as follows: since κ−1 is ≈0.3 nm, the electrostatic interaction is consistently screened to a great extent, preventing any specific effect from the divalent ion. In contrast, in (100 mM NaCl + 2 mM CaCl2): κ−1 is ≈0.9 nm, a value at which Ca2+ ions are able to bridge the support and the proximal leaflet, assuming a water film 1 nm thick, and decoupling is observed (Figure 5) where no decoupling was observed in pure 100 mM NaCl. Additionally, it shows that

Figure 5. Diffusion coefficient of DPPC in single SLB on mica as a function of the temperature and of ionic strength in presence of Ca2+. The abbreviations PROX and DIST represent the proximal and the distal leaflet, respectively. 5543

dx.doi.org/10.1021/la304962n | Langmuir 2013, 29, 5540−5546

Langmuir

Article

Diffusion Coefficients in the Second Bilayer of Double Bilayer and in Multibilayer of DPPC on Mica. The cases of the double and the multibilayer are more complicated. We observed that, in water, lipid diffusion in the second bilayer of a double bilayer was faster than in the first bilayer, which was closer to the support (at the same ionic strength, the diffusion coefficient is 10 times greater), and that there was no more decoupling of the leaflets. Figure 6 seems to show that in NaCl 10 mM and 1 M the second bilayer is decreasingly “held” by the electrostatic interaction and consequently its diffusion coefficient increases. The curve at the top of Figure 6a, enlarged in Figure 6b, represents the dependence of the diffusion coefficient of a multibilayer on temperature in water. The values of D indicate an even weaker interaction with the support due to a greater distance: the value in gel is higher than in the previous situations, the value in the fluid regime is similar to that for a free-standing bilayer. Thus, at first glance, the increase in diffusion coefficient and the disappearance of decoupling could be attributed to the increased distance between this second bilayer and the support. This second bilayer, for a given ionic strength, is less affected by the strength of the potential than the first bilayer, nor does it experience the roughness of the support, which is minimal with mica (atomically flat). However, as pointed out by a reviewer, we notice that Debye length at 10 mM is 3 nm, and the effect of support on the second bilayer (at least 7 nm away from the support) should be almost completely shielded at this concentration. Thus, further change at higher salt concentrations would not be expected, leaving unexplained the effect observed in Figure 6, i.e., that diffusivity was further increased by increasing NaCl concentration from 10 mM to 1 M. Similarly, if we assume that the second bilayer behaves as a fully free-standing bilayer (being out of range with respect to the electrostatic interaction), how can we explain the fact that the multibilayer (which should mimic a free bilayer even better) has a faster diffusion coefficient than the second bilayer of a double bilayer? Consequently, in these cases, increased diffusivity cannot be explained solely by effect of support, owing to its decreased

FRAPP is more sensitive (through D) than the Langmuir balance (through the transfer ratio) to the variation in the strength of interactions. Finally, from what we have observed in (a), (b), and (c) and, assuming that the potential is significantly reduced at κ−1, we deduce that the bilayer/support water film thickness is between 0.9 and 3 nm. Diffusion Coefficients in Unilamellar SLB of DMPC. To test the possible influence of the hydrocarbon chain length of the lipid, we studied the effect of differing NaCl concentrations on unilamellar SLB of DMPC on glass and on mica. Qualitatively, the behavior is the same as with DPPC: (i) In water, we observed that the diffusion coefficient increases from mica to glass (×10 in the gel phase). (ii) In 100 mM NaCl, in the gel phase, the diffusion coefficient is increased by a factor of 10 compared to DMPC on glass in water while in the fluid phase, the diffusion coefficient tends to become similar to that of a free-standing bilayer. (iii) On mica, leaflet decoupling is observed. Table 1 shows relevant values of D in the different phases. Table 1. Values of Diffusion Coefficient in the Gel, Ripple, and Fluid Phases for DMPC SLB Supported on Glass and on Mica D (cm2/s) DMPC

gel phase

100 mM NaCl/glass H2O/glass H2O/mica

2.5 × 10−9

distal proximal

2 × 10−10 3.5 × 10−11 2.8 × 10−12

ripple phase

fluid phase 1.2 × 10−7

−9

1.7 × 10 2 × 10−10

4.8 × 10−8 6.5 × 10−8 1.7 × 10−8

The bilayer−bilayer and bilayer−support interactions were quantitatively estimated based on DLVO and non-DLVO forces in the works of Nabika et al.33 and Tero et al.,34 respectively. A quantitative paper using our experimental data, numerical simulations, and theoretical models is in preparation.

Figure 6. (a) Diffusion coefficient of DPPC in the second bilayer of a double bilayer supported on mica as a function of temperature and NaCl concentration. (□), (○), and (■) correspond to water, 10 mM NaCl, and 1 M NaCl, respectively. The red rectangle contains the curve for a multibilayer in water. (b) is an enlargement of this rectangle. 5544

dx.doi.org/10.1021/la304962n | Langmuir 2013, 29, 5540−5546

Langmuir

Article

(4) Binder, H.; Zschornig, O. The effect of metal cations on the phase behavior and hydration characteristics of phospholipid membranes. Chem. Phys. Lipids 2002, 115 (1−2), 39−61. (5) Altenbach, C.; Seelig, J. Ca2+ binding to phosphatidylcholine bilayers as studied by deuterium magnetic resonance. Evidence for the formation of a Ca2+ complex with two phospholipid molecules. Biochemistry 1984, 23 (17), 3913−20. (6) Garcia-Manyes, S.; Oncins, G.; Sanz, F. Effect of pH and ionic strength on phospholipid nanomechanics and on deposition process onto hydrophilic surfaces measured by AFM. Electrochim. Acta 2006, 51 (24), 5029−5036. (7) Garcia-Manyes, S.; Oncins, G.; Sanz, F. Effect of ion-binding and chemical phospholipid structure on the nanomechanics of lipid bilayers studied by force spectroscopy. Biophys. J. 2005, 89 (3), 1812−26. (8) Fukuma, T.; Higgins, M. J.; Jarvis, S. P. Direct imaging of lipidion network formation under physiological conditions by frequency modulation atomic force microscopy. Phys. Rev. Lett. 2007, 98 (10), 106101. (9) Bockmann, R. A.; Hac, A.; Heimburg, T.; Grubmuller, H. Effect of sodium chloride on a lipid bilayer. Biophys. J. 2003, 85 (3), 1647− 1655. (10) Makino, K.; Yamada, T.; Kimura, M.; Oka, T.; Ohshima, H.; Kondo, T. Temperature-induced and ionic strength-induced conformational-changes in the lipid head group region of liposomes as suggested by zeta-potential Data. Biophys. Chem. 1991, 41 (2), 175− 183. (11) Eisenberg, M.; Gresalfi, T.; Riccio, T.; Mclaughlin, S. Adsorption of mono-valent cations to bilayer membranes containing negative phospholipids. Biochemistry 1979, 18 (23), 5213−5223. (12) Venable, R. M.; Zhang, Y. H.; Hardy, B. J.; Pastor, R. W. Molecular-dynamics simulations of a lipid bilayer and of hexadecane an investigation of membrane fluidity. Science 1993, 262 (5131), 223− 226. (13) Vanderploeg, P.; Berendsen, H. J. C. Molecular-dynamics simulation of a bilayer-membrane. J. Chem. Phys. 1982, 76 (6), 3271− 3276. (14) Cascales, J. J. L.; Berendsen, H. J. C.; delaTorre, J. G. Molecular dynamics simulation of water between two charged layers of dipalmitoylphosphatidylserine. J. Phys. Chem. 1996, 100 (21), 8621− 8627. (15) Tatulian, S. A. Binding of alkaline-earth metal-cations and some anions to phosphatidylcholine liposomes. Eur. J. Biochem. 1987, 170 (1−2), 413−420. (16) Pandit, S. A.; Bostick, D.; Berkowitz, M. L. Molecular dynamics simulation of a dipalmitoylphosphatidylcholine bilayer with NaCl. Biophys. J. 2003, 84 (6), 3743−3750. (17) Vacha, R.; Jurkiewicz, P.; Petrov, M.; Berkowitz, M. L.; Bockmann, R. A.; Barucha-Kraszewska, J.; Hof, M.; Jungwirth, P. Mechanism of interaction of monovalent ions with phosphatidylcholine lipid membranes. J. Phys. Chem. B 2010, 114 (29), 9504−9. (18) Gurtovenko, A. A.; Vattulainen, I. Membrane potential and electrostatics of phospholipid bilayers with asymmetric transmembrane distribution of anionic lipids. J. Phys. Chem. B 2008, 112 (15), 4629− 34. (19) Garcia-Celma, J. J.; Hatahet, L.; Kunz, W.; Fendler, K. Specific anion and cation binding to lipid membranes investigated on a solid supported membrane. Langmuir 2007, 23 (20), 10074−80. (20) Zimmermann, R.; Kuttner, D.; Renner, L.; Kaufmann, M.; Zitzmann, J.; Muller, M.; Werner, C. Charging and structure of zwitterionic supported bilayer lipid membranes studied by streaming current measurements, fluorescence microscopy, and attenuated total reflection Fourier transform infrared spectroscopy. Biointerphases 2009, 4 (1), 1−6. (21) Zhou, Y.; Raphael, R. M. Solution pH alters mechanical and electrical properties of phosphatidylcholine membranes: relation between interfacial electrostatics, intramembrane potential, and bending elasticity. Biophys. J. 2007, 92 (7), 2451−62.

potential and increased distance. Something else must be at work. We hypothesize a contribution indirectly related to the increased mobility of the first bilayer: The first bilayer experiences friction with and feels the potential of an immobile and solid support, mainly in the range of the corresponding Debye length. Interaction is thus high. As ionic strength increases, this interaction decreases and diffusivity increases. (i) When the second bilayer interacts, it is with a soft “support”, the first bilayer, thus involving lower friction and increasing its diffusion. (ii) The “support” of the second bilayer is mobile: therefore, increasing the diffusivity of the first bilayer when increasing the salt concentration may induce higher diffusivity in the second bilayer, somewhat like a conveyor belt effect. (iii) A possible transmission of interactions trough somewhat like a stacking effect may occur. These possible mechanisms could be hypothesized for the multibilayer case and will be tested in forthcoming work.

4. CONCLUSION To better understand lipid bilayer/support interactions, we carried out diffusion coefficient measurements by FRAPP and transfer ratio measurements using a Langmuir balance on supported bilayers of phosphatidylcholine lipids. This showed that the main effect of increased ionic strength is to enhance diffusion of the lipids due to a decrease in the electrostatic interaction between the bilayer and the support. For unimamellar SLBs, the two main parameters that govern bilayer behavior are confirmed to be electrostatic interaction and bilayer/support distance, both enabling the potential that acts on the bilayer to be varied. FRAPP measurements of the diffusion coefficients are shown here to be sufficiently “fine” to provide valuable input for tuning bilayer dynamics, even at the leaflet level.

■

ASSOCIATED CONTENT

S Supporting Information *

Schematic of the FRAPP setup, typical FRAPP signals, and global/instantaneous transfer ratios. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (B.T.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Aix-Marseille University for financial support to Frédéric Harb and Marjorie Sweetko for English language editing.

■

REFERENCES

(1) Verkleij, A. J.; de Kruyff, B.; Ververgaert, P. H.; Tocanne, J. F.; van Deenen, L. L. The influence of pH, Ca2+ and protein on the thermotropic behaviour of the negatively charged phospholipid, phosphatidylglycerol. Biochim. Biophys. Acta 1974, 339 (3), 432−7. (2) Hauser, H.; Shipley, G. G. Interactions of monovalent cations with phosphatidylserine bilayer membranes. Biochemistry 1983, 22 (9), 2171−8. (3) Herbette, L.; Napolitano, C. A.; McDaniel, R. V. Direct determination of the calcium profile structure for dipalmitoyllecithin multilayers using neutron diffraction. Biophys. J. 1984, 46 (6), 677−85. 5545

dx.doi.org/10.1021/la304962n | Langmuir 2013, 29, 5540−5546

Langmuir

Article

(22) Zimmermann, R.; Kuttner, D.; Renner, L.; Kaufmann, M.; Werner, C. Fluidity modulation of phospholipid bilayers by electrolyte ions: insights from fluorescence microscopy and microslit electrokinetic experiments. J. Phys. Chem. A 2012, 116 (25), 6519−25. (23) Scomparin, C.; Lecuyer, S.; Ferreira, M.; Charitat, T.; Tinland, B. Diffusion in supported lipid bilayers: Influence of substrate and preparation technique on the internal dynamics. Eur. Phys. J. E 2009, 28 (2), 211−220. (24) Charitat, T.; Bellet-Amalric, E.; Fragneto, G.; Graner, F. Adsorbed and free lipid bilayers at the solid-liquid interface. Eur. Phys. J. B 1999, 8 (4), 583−593. (25) Bassereau, P.; Pincet, F. Quantitative analysis of holes in supported bilayers providing the adsorption energy of surfactants on solid substrate. Langmuir 1997, 13 (26), 7003−7007. (26) Davoust, J.; Devaux, P. F.; Leger, L. Fringe pattern photobleaching, a new method for the measurement of transportcoefficients of biological macromolecules. EMBO J. 1982, 1 (10), 1233−1238. (27) Rand, R. P.; Fuller, N. L.; Gruner, S. M.; Parsegian, V. A. Membrane curvature, lipid segregation, and structural transitions for phospholipids under dual-solvent stress. Biochemistry 1990, 29 (1), 76−87. (28) Nagle, J. F.; Tristram-Nagle, S. Structure of lipid bilayers. Biochim. Biophys. Acta 2000, 1469 (3), 159−95. (29) Kinoshita, K.; Furuike, S.; Yamazaki, M. Intermembrane distance in multilamellar vesicles of phosphatidylcholine depends on the interaction free energy between solvents and the hydrophilic segments of the membrane surface. Biophys. Chem. 1998, 74 (3), 237− 49. (30) Hughes, A. V.; Howse, J. R.; Dabkowska, A.; Jones, R. A.; Lawrence, M. J.; Roser, S. J. Floating lipid bilayers deposited on chemically grafted phosphatidylcholine surfaces. Langmuir 2008, 24 (5), 1989−99. (31) Ladha, S.; Mackie, A. R.; Harvey, L. J.; Clark, D. C.; Lea, E. J. A.; Brullemans, M.; Duclohier, H. Lateral diffusion in planar lipid bilayers: A fluorescence recovery after photobleaching investigation of its modulation by lipid composition, cholesterol, or alamethicin content and divalent cations. Biophys. J. 1996, 71 (3), 1364−1373. (32) Sonnleitner, A.; Schutz, G. J.; Schmidt, T. Free Brownian motion of individual lipid molecules in biomembranes. Biophys. J. 1999, 77 (5), 2638−2642. (33) Nabika, H.; Fukasawa, A.; Murakoshi, K. Control of the structure of self-spreading lipid membrane by changing electrolyte concentration. Langmuir 2006, 22 (26), 10927−31. (34) Tero, R.; Ujihara, T.; Urisu, T. Lipid bilayer membrane with atomic step structure: supported bilayer on a step-and-terrace TiO2(100) surface. Langmuir 2008, 24 (20), 11567−76.

5546

dx.doi.org/10.1021/la304962n | Langmuir 2013, 29, 5540−5546