ARTICLE pubs.acs.org/Langmuir
Atomic Force Microscopy Studies of a Floating-Bilayer Lipid Membrane on a Au(111) Surface Modified with a Hydrophilic Monolayer Annia H. Kycia,‡ Jingpeng Wang,‡ A. Rod Merrill,† and Jacek Lipkowski*,‡ †
Department of Molecular and Cellular Biology and ‡Department of Chemistry, University of Guelph, Guelph, Ontario, Canada N1G 2W1
bS Supporting Information ABSTRACT: The surface of a gold electrode was functionalized with a hydrophilic monolayer of 1-thio-β-D-glucose formed by spontaneous self-assembly. The LangmuirBlodgett/LangmuirSchaefer (LB/LS) method was then used to assemble a bilayer onto the modified Au(111) surface. The bilayer lipid membrane (BLM) was separated from the Au(111) electrode surface by incorporating the monosialoganglioside GM1 into the inner leaflet of a bilayer composed of 1,2dimyristoyl-sn-glycero-3-phosphocholine (DMPC) and cholesterol. To make the inner leaflet, monolayers of GM1/DMPC/cholesterol with mole ratios of 1:6:3, 2:5:3, and 3:4:3 were used. The outer leaflet was composed of a 7:3 mole ratio of DMPC/cholesterol. Because of the amphiphilic properties of GM1, the hydrophobic acyl chains were incorporated into the BLM, whereas the large hydrophilic carbohydrate headgroups were physically adsorbed to the Au(111) electrode surface, creating a “floating” BLM (fBLM). This model contained a water-rich reservoir between the BLM and the gold surface. In addition, because of the bilayer being physically adsorbed onto the support, the fluidity of the BLM was maintained. The compression isotherms were measured at the air/ water interface to determine the phase behavior and optimal transfer conditions. The images acquired using atomic force microscopy (AFM) and the forcedistance measurements showed that the structure of the fBLM evolved with increasing GM1 content from 10 to 30 mol %, undergoing a transition from a corrugated to a homogeneous phase. This change was associated with a significant increase in bilayer thickness (from ∼5.3 to 7.3 nm). The highest-quality fBLM was produced with 30 mol % GM1.
’ INTRODUCTION The objective of this work was to design a model biomimetic membrane that could be employed to study voltage-gated channel formation by membrane-bound colicin E1 channel peptide, whose structure and properties have been investigated extensively by one of the authors.1 To apply the voltage gate, the membrane must be supported at a conductive surface. In addition, infrared reflection absorption spectroscopy (IRRAS) will be used to investigate conformational changes corresponding to the open and closed states of the channel, and hence, the support must be reflective for IR radiation. Therefore, a gold electrode was chosen as the support in these studies. We have already had significant experience in working with gold-supported bilayer lipid membranes (BLMs).2,3 In the open state, the colicin E1 channel consists of a quasicircular arrangement of eight helices embedded in the membrane interfacial layer and anchored by the hydrophobic helical hairpin.4 To allow protein insertion and protect against denaturation, the membrane has to be separated from the gold surface by a 12-nm-thick water-rich layer. Numerous strategies developed to separate the lipid bilayer membrane from the solid substrate have been described in several review articles.3,5,6 These include polymer-cushioned bilayers,7 tethered bilayers r 2011 American Chemical Society
(tBLMs),812 and floating bilayers (fBLMs).13 Tethered bilayers812 and floating bilayers13 are the most convenient systems to assemble at a gold electrode surface. tBLMs are constructed by covalent attachment of functionalized lipids with a hydrophilic spacer to a solid support. Polyethylene glycol (PEG) is frequently employed as the hydrophilic spacer.8,12b,12c However, recent neutron reflectivity experiments9a,14 and IRRAS15 demonstrated that the PEG spacer might adopt a conformation in which the water content in the spacer region is low. The water content and the fluidity of the tBLM can be increased by dilution of the inner leaflet with small surface-active thiol molecules such as β-mercaptoethanol.9 Such a model membrane is termed a sparsely tethered BLM. Although a sparsely tBLM has the desired properties,12a the molecules used to assemble these architectures are functionalized lipids that are synthesized by a complex procedure.9,16 Floating bilayers have recently been reported as an improved model for lipid membranes.13 They have been formed by a combination of LangmuirBlodgett and LangmuirSchaefer Received: May 2, 2011 Revised: July 14, 2011 Published: July 18, 2011 10867
dx.doi.org/10.1021/la2016269 | Langmuir 2011, 27, 10867–10877
Langmuir
ARTICLE
Figure 1. (A) Schematic of the floating bilayer on a thioglucose monolayer on Au(111) constructed using LB/LS deposition techniques. The expected thickness of the entire film is approximately 7 nm. (BF) Chemical structures of the components used to build the floating lipid bilayer along with dimensions: (B) GM1, (C) cholesterol, (D) water, (E) 1-thio-β-D-glucose (∼0.66 nm) and (F) DMPC. (Molecular lengths of DMPC and GM1 obtained from refs 48 and 49, respectively.)
deposition on top of a supported bilayer13a,e or on top of a monolayer of S-layer protein.13b In this architecture, the bilayer floats about ∼2.4 nm over the supporting layer, and the lateral mobility of the lipids is significantly improved.13b The fBLM was selected as the most promising model for our future studies with the colicin channel since colicin E1 can be reconstituted into a supported bilayer using the LangmuirBlodgett technique.17 In this article, we describe the construction of a novel fBLM supported on a single-crystal gold electrode, Au(111). A schematic of the model is shown in Figure 1. The lipid bilayer membrane is separated from the gold electrode surface by incorporating the monosialoganglioside GM1 in the inner leaflet of the bilayer composed of the zwitterionic phospholipid 1,2-dimyristoyl-sn-glycerol-3-phosphocholine (DMPC) and cholesterol. The inner leaflet containing GM1, DMPC, and cholesterol is physically adsorbed at the gold electrode modified with a hydrophilic thiol, 1-thio-β-D-glucose. The self-assembled monolayer (SAM) of the sugar thiol on the gold substrate is used to improve the transfer and stability of the inner leaflet. The inner leaflet is organized at the air/water interface of the Langmuir trough and then deposited onto the modified gold support using the LangmuirBlodgett (LB) technique. The outer leaflet of the bilayer (7:3 DMPC/cholesterol) was deposited using the LangmuirSchaefer (LS) method. The completed bilayer (including the thioglucose SAM) has a theoretical thickness of ∼7 nm, assuming that the lipid acyl chains are in a fully extended state with the molecules perpendicular to the gold substrate surface (see Figure 1). GM1 acts as a “pillar”, supporting the lipid bilayer membrane from the modified gold surface. The headgroup of GM1 is composed
of four neutral sugar groups and a sialic acid residue. The presence of the large hydrophilic headgroup of GM1 sandwiched between the thioglucose-modified gold substrate and the lipid bilayer should encourage water entrapment under the lipid bilayer, thereby creating a water reservoir, which is essential for accommodating proteins. Numerous studies have shown phase segregation in GM1containing lipid mixtures.1828 To meet the objectives of this study, phase segregation in the fBLM must be avoided. To determine the amount of GM1 in the inner leaflet required to produce the most uniform, stable, and microdomain-free fBLM, several lipid mixtures of GM1, DMPC, and cholesterol were studied with the following GM1/DMPC/cholesterol mole ratios: 0:7:3, 1:6:3, 2:5:3, 3:4:3. Compression isotherms were measured to determine the stability and miscibility of monolayers formed by these mixtures at the air/water interface, and the phase behavior of the fBLMs formed at the modified gold surface was characterized using atomic force microscopy (AFM). There are several elements of novelty in this work. We describe how to engineer an fBLM with ∼2-nm water-rich region separating the membrane from the metal surface using commercially available compounds. The spacer region contains mainly water and biocompatible sugar groups. We show that the phase behavior of the mixed lipid/cholesterol/polysaccharide bilayer can be determined with the aid of AFM and that this information can be used to optimize the fBLM composition. In this study, 7:3 DMPC/cholesterol lipid composition was chosen because, at 24.5 °C, the temperature at which the AFM experiments were conducted, the lipids are in a liquid-ordered state and have sufficient lateral mobility in an fBLM.13b In 10868
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
Langmuir addition, we have previous experience with assembling goldsupported DMPC/cholesterol bilayers.29 The preliminary results of this fBLM studies were first presented at the Biophysical Society 53rd Annual Meeting in Boston, MA, in 2009.
’ EXPERIMENTAL SECTION Chemicals, Solutions, Glassware. 1,2-Dimyristoyl-sn-glycerol3-phosphocholine (DMPC), cholesterol (chol), and 1-thio-β-D-glucose were obtained from Sigma-Aldrich (St. Louis, MO, 99%). Ganglioside (GM1) was purchased from Avanti Polar Lipids (Alabaster, AL, 99%) in powder form. Monolayer spreading solutions of cholesterol and DMPC were prepared by dissolving these compounds in chloroform (Fisher, Pittsburgh, PA, Spectroanalyze grade). Monolayer spreading solutions of pure GM1 and mixtures of GM1, DMPC, and cholesterol were prepared by dissolving these compounds in chloroform containing 25% methanol (Fisher, Spectroanalyze grade). To make an accurate spreading solution for Langmuir trough studies, it was required that the mass of GM1 (10 mg) be reported to 0.01 mg. The spreading solution was then made directly in the vial containing GM1. The electrolyte for electrochemical and AFM measurements was prepared from NaF (VWR Scientific, Radnor, PA, Suprapur). The NaF powder was placed in a UV ozone chamber (Jelight, Irvine, CA) for 20 min to oxidize any organic impurities. A 1 mM NaF electrolyte was employed because the fBLM described in this work was intended for use in infrared reflection absorption spectroscopy (IRRAS) studies of the voltage-gated channel formation by the membrane-bound colicin E1 channel peptide. The IRRAS measurements were performed using a BaF2 IR transparent window. The NaF electrolyte was thus needed to suppress the solubility of the window. The water used was Milli-Q UV-plus ultrapure water (18.2 MΩ cm). All glassware was cleaned in hot acid bath (volume ratio 1:3 HNO3/ H2SO4) for 4050 min. The glassware was removed and rinsed thoroughly with Milli-Q water. If necessary, the glassware was dried in an oven at ∼40 °C overnight. Teflon pieces were cleaned in a piranha solution (volume ratio 1:3 H2O2/H2SO4) and then rinsed with copious amounts of Milli-Q ultra pure water. (Caution: Piranha solution is an extremely strong oxidant and should be handled very carefully!) Langmuir Trough: Compression Isotherms. Compression isotherms were recorded for the pure components (DMPC, cholesterol, GM1) and the four mixtures of these lipids. Stock solutions of DMPC and cholesterol were prepared by dissolving the compounds in chloroform with concentrations of 12 mg/mL. A 2 mg/mL solution of GM1 was prepared by dissolving the compound in a 1:4 volume ratio of methanol/chloroform. Mixed solutions with concentrations of 0.60.8 mg/mL were prepared from the respective stock solutions. Surface pressurearea isotherms were recorded using a KSV MiniMicro trough (KSV Instruments Ltd.., Helsinki, Finland) equipped with two hydrophilic barriers and a Wilhelmy balance with a filter paper (width = 10 mm) as the plate (surface area = 243 cm2). The trough was controlled by a computer using KSV LB5000 software. To control and maintain a constant temperature during experiments, the trough was equipped with a water jacket connected to a water circulating bath (Cole-Parmer, Vernon Hills, IL, Polystat temperature controller). The water subphase was maintained at a temperature of 30 ( 1 °C and adjusted to a pH of 1 by adding HCl. GM1 contains a sialic acid residue; hence, the acidic subphase prevents loss of GM1 into the subphase.20 Lipid solutions were spread carefully onto the subphase using a microsyringe, the monolayer was left for 15 min to ensure solvent evaporation, and then compression was initiated with a barrier speed of 10 mm/min for cholesterol, DMPC, and DMPC/ cholesterol mixtures or a barrier speed of 30 mm/min for pure GM1. Three compression isotherms were obtained for each lipid solution to
ARTICLE
ensure reproducibility. For each lipid solution, the mean molecular area was reproducible to within 2%. Gold Substrate Preparation. For AFM studies, the gold substrate consisted of a vapor-deposited gold film. An approximately 1-μmthick film of gold was sputtered onto a standard microscope glass slide coated with a 2-nm-thick layer of chromium to ensure good adhesion of the gold. As a result, a rough film was obtained consisting of gold nanoparticles with an average diameter of 20 nm. To produce large gold crystallites, the gold-coated glass slide was annealed in a muffle furnace at 675 °C for 1 min. Figure SI 1 of the Supporting Information shows AFM images of the flame-annealed sample. The image shows atomically flat Au(111) terraces, approximately 400 nm in diameter, that were used in AFM studies reported herein. Flame annealing is a well-known procedure to produce clean and well-ordered surfaces of noble-metal electrodes.31 To prevent chromium diffusion to the gold surface, all experiments described in this work were performed on gold slides that had been flame annealed once. Hydrophilic Modification of Gold. The Au(111) substrate was modified by immersion of the flame-annealed gold into a 2 mmol L1 solution of 1-thio-β-D-glucose in methanol, as the gold surface covered by the self-assembled monolayer of thioglucose is hydrophilic. The selfassembly of the thiol was allowed to progress for 20 h. Afterward, the modified Au(111) surface was rinsed using copious amounts of methanol and water. The film-coated electrode was then dried under a vacuum for 1 h. Bilayer Deposition. Inner Leaflet: LangmuirBlodgett Deposition. To create the “floating” bilayer on the thioglucose-modified Au(111) surface, a combination of LangmuirBlodgett (LB) and Langmuir Schaefer (LS) techniques was used. The subphase of the Langmuir trough was heated to 30 ( 1 °C to improve the homogeneity and fluidity of the compressed film.29 The first monolayer deposited onto the hydrophilically modified Au(111) surface contained DMPC, cholesterol, and GM1. The lipid mixture was spread onto the air water interface. The chloroform was allowed to evaporate for 15 min. The monolayer was compressed to and held at a film pressure of 40 mN/m. After the film had stabilized for ∼10 min, the electrode was withdrawn vertically at a speed of 26 mm/min. A transfer ratio of 1.0 ( 0.1 was obtained. Recent studies by Kaviratna et al.30 demonstrated that, during the 30 min needed to spread and transfer the monolayer, the acidic ester hydrolysis of phosphocholins spread at the surface of an acidic subphase of pH 1 is negligible. The monolayer-covered thiolated gold electrode was then allowed to dry for 2 h. Outer Leaflet: LangmuirSchaefer Touch. The second monolayer deposited onto the Au(111) surface contained 7:3 DMPC/cholesterol. The second leaflet was deposited using the LS method on a pure water subphase at 30 ( 1 °C. The monolayer was spread at the airwater interface of the Langmuir trough and compressed and held at a surface pressure of 40 mN/m. The thioglucose-modified gold substrate with the inner leaflet was horizontally brought into contact with the DMPC/cholesterol monolayer on the trough and then lifted upward. Garcia et al.3b provided direct evidence that the second leaflet was not dewetted when the electrode was lifted following the LS touch. The thickness of the films measured in this work is also consistent with the thickness of a bilayer, indicating that the LB/LS transfer was successful. AFM Instrumentation and Procedures. For AFM experiments a custom-made liquid cell was used. All parts were cleaned in piranha solution and then rinsed copiously with Milli-Q water. (Caution: Piranha solution is an extremely strong oxidant and should be handled very carefully!) A glass rod with a gold-coated clip was used to hold the chip containing the cantilever/tip. The glass rod was cleaned before each use by placing it in a beaker containing ethanol and sonicating (VWR Scientific, model 50D) for 30 min. Afterward, the glass rod was rinsed with Milli-Q water and dried under a stream of argon. Prior to 10869
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
Langmuir use, the chip was placed under the glass rod clip and exposed to ozone in a UV laminar flow chamber (Jelight, Irvine, CA) for 30 min to remove any organic impurities. The experiments involved two types of measurements: (i) forcedistance curves to determine the film thickness and its distribution and (ii) images of the film topography. Each type of measurement was performed using different instrumentation along with different cantilevers as described in the following sections. ForceDistance Curves. The AFM instrument was a PicoSPM microscope (Molecular Imaging, Tempe, AZ, USA) equipped with an AFMS1182 scanner head (Molecular Imaging, Tempe, AZ, USA) and Picoscan 5.2 software. The method used for measuring the force distance curves in this study is described in detail in a previous publication.29 The cantilevers were V-shaped, made of silicon nitride, with a silicon pyramidal tip (tip radius of curvature ≈ 20 nm), purchased from Veeco (Plainview, NY). A nominal frequency of 0.18 kHz and a nominal spring constant of 0.06 N/m were reported. The manufacturer’s estimated spring constant for the AFM cantilever has a wide range of variation.32 Therefore, the thermal tune method was used to calibrate the cantilever spring constant.33 The calibration demonstrated that the nominal spring constant for the Veeco cantilever used on the fBLM containing 10 mol % GM1 in the inner leaflet was 0.21 ( 0.04 N/m. For the fBLMs containing 20 and 30 mol % GM1 (Veeco, Plainview, NY, model SNL-10) the calibration of the cantilever demonstrated that the nominal spring constant was 0.13 ( 0.03 N/m. Figure SI 2 of the Supporting Information shows that these cantilevers displayed a linear response in the section of the curve corresponding to the elastic bending of the tip against the gold surface. To increase the repulsive electrostatic interactions between the AFM tip and the film-covered Au(111) surface, the force distance curves were measured in a dilute 1 mM NaF solution.29 Force curves were measured at atomically flat Au(111) terraces. Force distance curves were recorded by measuring the deflection of the cantilever versus the position of the sample mounted on the piezoelectric translator. The cantilever deflection versus piezo position curves were converted into force curves (force versus tipsubstrate distance) using software written in-house. For statistical data analysis, at least 100 force curve measurements were recorded at three to four different terraces on the sample.
Imaging Using MAC (Magnetic Alternating Current) Mode. Dynamic MAC-mode AFM imaging was carried out in a 1 mM NaF solution using an Agilent Technologies 5500 scanning probe microscope (Agilent, Palo Alto, CA, N9621-13601 MAC III Mode controller). This model has an improved version of MAC mode and produces higher-quality images than the PicoSPM instrument. The calibration of the cantilevers by the thermal tune method (Thermal K) was also performed on the 5500 microscope from Agilent Technologies because this instrument was provided with the Thermal K software. For the MAC-mode imaging, the cantilevers were rectangular-shaped and made of silicon and had a typical resonant frequency of 155 kHz in air (tip/cantilever B of type I MAC levers, supplied by Agilent, batch no. T102100045-25). The resonance frequency in solution was dampened to 103130 kHz. The tapping amplitude set point was set at 85% of the free amplitude. The nominal force constant of this cantilever was 1.75 N/m. The scanner (model N9520A-US07480132.xml) was calibrated and used for imaging. Topography, amplitude, and phase images were acquired simultaneously. All images were recorded at 24.5 ( 0.5 °C in a liquid cell, with a scan speed of 2.03.5 lines per second. The system was allowed to stabilize for at least 30 min before images were acquired. Data acquisition and analysis were carried out using PicoView 1.49 and Gwyddion v2.19 softwares, respectively. Note the difference in the values of the force constant for cantilevers used for forcedistance measurements and for MAC-mode imaging. Because the measured force is equal to the product of the force
ARTICLE
constant and the deflection, cantilevers with a force constant even smaller than 0.1 N/m are recommended to measure forces between the tip and substrate, to improve the precision of such measurements.34 In contrast, stiffer cantilevers are recommended for AC-mode (MACmode) imaging to overcome sticking of the AFM probe and to ensure stability of the oscillations. This is why the force constants of MAC levers are about 10 times larger than those for forcedistance measurements.35 To determine whether the AFM images were distorted by the tip, we performed tip deconvolution on our topography images using the module in the commercial software SPIP (Image Metrology A/S, Hørsholm, Denmark) which is based on the blind tip estimation algorithm.36 The tip geometry was determined from scanning electron microscopy. The tip had a conical shape, and its radius of curvature was ∼15 nm. However, after tip deconvolution, differences in the deconvoluted image compared to the original were negligible. As a result, images reported in this work are the original images. Scaling analysis was performed to quantitatively characterize the surface features of the films imaged using MAC mode (topography images). The topography of the images was characterized by the critical scaling length, Lc, which is a measure of the lateral correlation or periodicity of surface features in the direction parallel to the surface, and ξl, which is a measure of the roughness in the vertical direction, where 4ξl is equal to the amplitude, A.37 For example, in the case of a surface showing regular periodic stripes, Lc would be equal to the periodicity of the stripes and 4ξl = A would be equal to the peak-tovalley amplitude of the stripes.37 Scaling analysis is a standard method of digital image analysis. It is to some extent equivalent to autocorrelation analysis, and in fact, the limiting surface roughness ξl is equal to the square root of the initial value of the autocorrelation function, whereas the critical scaling length is related to the correlation length (σ) by the formula σ = LcR1/2/π, where R is the slope of the initial fragment of the scaling analysis plot.38 The advantage of scaling analysis is that it provides a means to assess whether the image size is sufficiently large to acquire an average of the dimensions of topographical features observed in the image. The plots of scaling analysis, shown in the Supporting Information (Figure SI 3), indicate that the limiting surface roughness and the critical scaling length can be determined correctly if the image size is about 4 times larger than Lc. The principle of the scaling analysis is explained in detail in Figure SI 3 of the Supporting Information. It should be noted that the maximum accuracy of the AFM height values is (0.1 nm. In addition, considering the lateral resolution of AFM, the Lc values have an error of (1 nm.
’ RESULTS Compression Isotherm Studies of DMPC, Cholesterol, and GM1. To determine the behavior of mixed monolayers of
cholesterol, DMPC, and GM1 at the air/water interface, compression isotherms were obtained by measuring the change in surface pressure as a function of the mean molecular area at the air/water interface at constant temperature. Figure 2a shows the compression isotherms for the pure components (cholesterol, DMPC, and GM1), and Figure 2b shows those for the binary 7:3 DMPC/cholesterol and ternary GM1/DMPC/ cholesterol monolayers at the air/water interface. The subphase was 0.1 M HCl at 30 ( 1 °C. The compression isotherms in Figure 2 allow for the determination of the limiting areas (Alim) for cholesterol, DMPC, and 7:3 DMPC/cholesterol. Their values are 36.4, 46.0, and 38.5 Å2/molecule, respectively. Taking into account differences in experimental conditions, these values of Alim are in reasonable agreement with the literature. For example, 36.5 Å2/molecule was reported for 10870
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
Langmuir
ARTICLE
Figure 2. Compression isotherms of the (a) pure components cholesterol (—), DMPC ( ), and GM1 ( 3 3 3 ) and (b) mixed monolayers of DMPC, cholesterol, and GM1 with the following mole ratios: 7:3 DMPC/cholesterol (thick solid line), 1:6:3 ( ), 2:5:3 ( 3 3 3 ), 3:4:3 (—) GM1/ DMPC/chol. The subphase was 0.1 M HCl at a temperature of 30 ( 1 °C.
cholesterol at 25 °C,39 45 Å2/molecule for DMPC at 28 °C,40 and ∼38 Å2/molecule for 7:3 DMPC/cholesterol at 23 °C.41 In the case of GM1, the experimental value determined in Figure 2a for Alim was 49 Å2/molecule. The Alim values found in the literature for compression isotherms of GM1 cover the range from 39 to 70 Å2/molecule.19 The experimental value of 49 Å2/molecule determined in this work falls well within this range. The shape of the isotherms of the pure components is also in good agreement with measurements published earlier under similar conditions.19,20,39,40 Detailed comparison of the isotherms presented in Figure 2a with the literature data is presented in the Supporting Information (Figure SI 4). For mixtures of cholesterol, DMPC, and GM1, the mole ratio of lipid acyl chains to cholesterol was maintained at 7:3. The addition of GM1 to the mixture replaces the zwitterionic polar headgroup of DMPC by the bulky carbohydrate polar headgroup of GM1. Therefore, upon addition of GM1 to the mixture, the monolayer is expected to be more expanded than the monolayer of 7:3 DMPC/cholesterol. However, Figure 2b shows that the compression isotherms of monolayers containing 10, 20, and 30 mol % GM1 are very similar to the compression isotherm of 7:3 DMPC/cholesterol. The lift-off surface pressure is shifted to 6568 Å2/molecule relative to those of pure DMPC (96 Å2/ molecule) and GM1 (145 Å2/molecule), indicating that a condensing effect takes place. The “lift-off” is defined as the point on the compression isotherm at which the surface pressure starts to increase.13a The value for Alim is 39.1 Å2/molecule for 10 mol % GM1 and 38.2 for 20 and 30 mol % GM1, suggesting that the acyl chains limit the packing area rather than the bulky headgroup of GM1. In addition, with increasing GM1 content from 0 to 30 mol %, a slight fluctuation in the collapse pressure was observed (51.7, 51.4, 55.6, and 54.7 mN/m, respectively). The compression isotherm studies at the air/water interface allowed for the determination of the excess mean molecular area, Aex, and the excess Gibbs energy of mixing ΔGex, as shown in Figure SI 5 of the Supporting Information. The negative sign of these values suggests that monolayers containing 10, 20, and 30 mol % GM1 are at least partially miscible with attractive interactions between the molecules. At a surface pressure of 40 mN/m (the value at which the monolayer is transferred
Figure 3. AFM topography image of a bilayer of DMPC/cholesterol (7:3) deposited on a monolayer of thioglucose on an annealed gold substrate. Images were recorded at room temperature (24.5 ( 0.5 °C) in 1 mM NaF. Image dimensions are 180 180 nm. The critical scaling length (Lc) and the amplitude surface roughness (A) were obtained from scaling analysis of the films.
onto the hydrophilic-modified single-crystal gold), ΔGex attains a minimum value, indicating that the attractive interactions between the components of the mixture are enhanced. Further characterization of the phase behavior of the mixed monolayers after deposition onto the thioglucose-modified Au(111) surface can be determined from the AFM images of the assembled fBLM. Atomic Force Microscopy. i. Imaging. MAC-mode AFM was employed to study the structure of the fBLM supported on a Au(111) substrate modified with a SAM of thioglucose. The imaging allows the visualization of the lateral organization of the fBLM. The majority of images presented here were recorded at a single Au(111) terrace to minimize contributions from the roughness of the gold substrate. MAC-mode imaging simultaneously produces topography, amplitude, and phase images. In this work, the topography images are presented because these were the images used for quantitative scaling analysis. Phase images were of importance to confirm the presence of phase separation in the fBLM and are shown when necessary. AFM imaging was initially performed on a bilayer composed of 7:3 DMPC/cholesterol supported on a gold substrate modified with a thioglucose monolayer (see Figure 3). This 10871
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
Langmuir
ARTICLE
Figure 4. (a) AFM topography image of an fBLM on a monolayer of thioglucose on an annealed gold substrate, containing 1:6:3 GM1/DMPC/ cholesterol in the inner leaflet and 7:3 DMPC/cholesterol in the outer leaflet. (b) Enlarged topography image showing the corrugated phase. Images were recorded at room temperature (24.5 ( 0.5 °C) in 1 mM NaF. For images a and b, the values in the upper right corner correspond to the critical scaling length (Lc) and amplitude surface roughness (A) obtained from scaling analysis of the films.
bilayer was used as a “control” for the fBLMs being investigated to ensure that the thioglucose monolayer was not influencing the phase behavior of the lipids in the bilayer. The topography image of a bilayer supported on a single Au(111) terrace in Figure 3 shows that the bilayer appears to be relatively featureless, except for some randomly distributed brighter and darker “spots” in the film. The scaling analysis of the topography image gives Lc = 22 ( 1 nm, which is a measure of the average distance between spots, and an amplitude roughness of A = 0.2 ( 0.1 nm. There is a ∼0.5-nm difference between the thickness of liquid-ordered and liquid-crystalline phases of the 7:3 DMPC/ cholesterol bilayer.29 The height difference between the features seen in Figure 3 is much less than this number. Therefore, the image in Figure 3 shows the absence of any phase segregation, suggesting that the fBLM of DMPC/cholesterol (7:3) is homogeneous, with the bilayer most likely in the liquid-ordered phase.42 The observance of a homogeneous phase agrees well with previous studies of DMPC/cholesterol (7:3) monolayers at the airwater interface, which demonstrated that such monolayers exist as a single phase at surface pressures greater than 12 mN/m.41 In addition, a previous AFM study of a DMPC/cholesterol (7:3) bilayer directly supported on a gold substrate revealed a homogeneous film.29 Based on these results, it can be concluded that the gold surface modified with the hydrophilic monolayer does not change the behavior of the lipid bilayer. Below, similar experimental techniques and conditions were applied to analyze the changes in the bilayer structure when GM1 was incorporated into the inner leaflet of the fBLM. Figure 4a shows the topography images for an fBLM containing 10 mol % GM1 in the inner leaflet, on a monolayer of thioglucose self-assembled at a Au(111) substrate. The image clearly shows that the film consists of brighter and darker corrugated domains. Figure 4b shows the image of one corrugated domain taken with higher resolution. Scaling analysis gave a critical scaling length of Lc = 7 ( 1 nm and an amplitude roughness of A = 0.3 ( 0.1 nm. These numbers are a measure of the average peak-to-peak distance between the corrugations and their average peak-to-valley height. The presence of the corrugated phase has never been observed for a bilayer composed of GM1, DMPC, and cholesterol. However, a corrugated phase characterized by an amplitude roughness of 0.6 ( 0.1 nm and a critical scaling length of 5.5 ( 0.5 nm at 24 °C was observed for an sBLM (i.e., supported BLM) of DMPC on gold.2j The periodicity of the corrugations observed in the fBLM is equivalent to those observed in the sBLM within the
experimental uncertainty. For the sBLM, these undulations were explained by the elastic stress in the film induced by spontaneous curvature, caused by the packing of the polar headgroups of DMPC molecules in direct contact with the metal surface.2j Phospholipid bilayers are known to form a ripple phase that is a transition from the ordered gel phase to the disordered liquid-crystalline phase. The ripple phase is characterized by regular, large-scale spatial undulations (longwavelength rippling) with a characteristic period between 10 and 30 nm and an anomalous swelling of the bilayer.43 Although Figure 4b resembles AFM images of the ripple phase published in the literature,44 the periodicity of corrugations in the bilayer supported at the gold surface is shorter than the periodicity of the ripple phase. With the addition of 20 mol % GM1 to the inner leaflet of the fBLM, a very interesting change in the bilayer structure was observed. The central image in Figure 5 shows the phase image of the fBLM deposited on a Au(111) substrate modified with a thioglucose SAM, recorded over several Au(111) terraces (dimensions of 1 μm 1 μm). At the gold surface, this film appears inhomogeneous and phase-segregated. One can distinguish four phases: (1) a uniform phase, (2) a corrugated phase, (3) a wormlike phase with corrugations, and (4) a wormlike phase. Surrounding the central image are the topography images of each phase. The image for region 1 reveals that the surface of the bilayer is quite uniform, characterized by an amplitude surface roughness of 0.12 ( 0.10 nm, similar to that observed for the fBLM containing 0 mol % GM1 (Figure 3). For regions 2 and 3 of Figure 5, the surface of the bilayer contains corrugated structures characterized by an average periodicity of 9 ( 1 nm similar to the periodicity of corrugations observed in the fBLM containing 10 mol % GM1 (Figure 4a). However, these corrugations span over uneven buckled domains with height differences on the order of 1 nm. The fourth region corresponds to a smooth elevated wormlike phase surrounded by a uniform lower phase. The height difference between the elevated and lower phases is 1.2 ( 0.1 nm, similar to the height difference observed in region 3. Thus, the wormlike phase can either contain corrugations as observed in region 3 or be formed without corrugations as shown in region 4. Figure 6 shows several large gold terraces covered by the wormlike phase. This type of submicrodomain with elevated wormlike structures or filament network has been observed previously for monolayers or bilayers of dipalmitoylphosphatidylcholine 10872
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
Langmuir
ARTICLE
Figure 5. Central image with dimensions of 1 1 μm is an AFM phase image of an fBLM containing 20 mol % GM1 in the inner leaflet, deposited on a monolayer of thioglucose on an annealed gold substrate. Four phases are identified within the film: (1) uniform phase, (2) corrugated phase, (3) wormlike phase with corrugations, and (4) wormlike phase, as shown by the corresponding AFM topography images. Images were recorded at room temperature (24.5 ( 0.5 °C) in 1 mM NaF.
Figure 6. (a) AFM topography image of an fBLM on a monolayer of thioglucose on an annealed gold substrate, containing 20 mol % GM1 in the inner leaflet. Image clearly shows the wormlike phase of the bilayer. Image was recorded at room temperature (24.5 ( 0.5 °C) in 1 mM NaF. Image dimensions are 1000 1000 nm. (b) Model of fBLM for the wormlike phase showing two regions: an ordered condensed domain and a disordered liquid-expanded (LE) region. The height difference between the condensed domain and the disordered LE DMPC domain is ∼1.2 nm. Adapted from ref 25.
(DPPC)/GM1 (9:1) and GM1/DPPC/cholesterol (1:6:3) deposited from the air/water interface onto a hydrophilic mica substrate.25,27,28 It was suggested that this structure results from the coexistence of an ordered condensed and a disordered liquidexpanded (LE) phase. The ordered phase was ∼0.7 nm taller than the surrounding disordered LE phase.25 GM1 preferentially resides in the more ordered condensed phase and distributes itself in clusters arranged into a filament network.27,28 This argument can be applied to explain the origin of the wormlike phase observed in the present study of the fBLM with 20 mol % GM1, where the “worms” were found to be elevated by 1.2 ( 0.1 nm. In the model shown in Figure 6b, the worms are described as condensed domains rich in GM1, in which the void volume between the GM1 molecules is filled by DMPC and cholesterol molecules, encouraging alignment of hydrocarbon chains, with the acyl chains in the most extended state. This phase is elevated by approximately 1 nm with respect to the surrounding liquid-expanded phase composed of DMPC lipids, in which DMPC acyl chains are in a
disordered state. In conclusion, the AFM images show that the fBLM containing 20 mol % GM1 is phase-separated. This ternary mixture is a complex system consisting of phaseseparated GM1-rich domains surrounded by domains with lower GM1 content. Although this phase behavior is unique and interesting, for our future work with proteins, a model membrane with such phase separation is undesirable. With the addition of 30 mol % GM1 to the inner leaflet of the fBLM, the film appeared to be free of phase boundaries and homogeneous (Figure 7). The image in Figure 7a reveals that this uniform film is continuous over the boundary of the Au(111) terraces. A higher-resolution image over a single Au(111) terrace shown in Figure 7b is characterized by an amplitude roughness of 0.2 ( 0.1 nm and an average separation of 22 ( 4 nm between the small bumps. These values characterizing the bilayer surface are in excellent agreement with the values obtained for the homogeneous bilayer of 7:3 DMPC/cholesterol shown in Figure 3. It seems that, with addition of 30 mol % GM1, 10873
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
Langmuir
ARTICLE
Figure 7. AFM topography images of an fBLM containing 30 mol % GM1 in the inner leaflet, deposited on a monolayer of thioglucose on an annealed gold substrate. Images were recorded at room temperature (24.5 ( 0.5 °C) in 1 mM NaF. (a) Image dimensions of 500 500 nm. (b) Enlargement of the image in panel a with dimensions of 200 200 nm showing a uniform film. The values in the upper right corner of image b correspond to the critical scaling length Lc and the amplitude of surface roughness (A) obtained from scaling analysis of the films. These values are similar to those obtained for the uniform bilayer of DMPC/cholesterol (7:3) on a monolayer of thioglucose shown in Figure 3 (Lc = 22 ( 1 nm, A = 0.2 ( 0.1 nm).
Figure 8. (A) Typical force curve measured on an fBLM containing 10 mol % GM1 in the inner leaflet prepared on a thioglucose SAM on an annealed gold slide in 1 mM NaF at room temperature. (B) Penetration force versus thickness for 115 measurements on the fBLM. (C) Histogram plot of the width of the jump-in of the fBLM with Gaussian fitting. The mean value determined from the Gaussian fitting is treated as the estimated thickness of the bilayer.
the corrugations and wormlike phases disappear to give a uniform and smooth film. It is interesting to note that Frey et al.25 reported a similar trend in their studies of monolayers of a similar system, DPPC/GM1 mixtures. They observed phase segregation at lower GM1 concentrations and formation of a more homogeneous film with 30 mol % GM1 content. They suggested that phospholipid molecules formed complexes with GM1 in which gangliosides are optimally surrounded by phospholipid molecules. In this state, the tilt angle of the acyl chains is decreased, and the headgroup of GM1 is extended.25 Therefore, the fBLM containing 30 mol %
GM1 might exist predominantly in a condensed state, as modeled in Figure 6b. Images shown in this section provide information about the film homogeneity/heterogeneity at dimensions less than 1 μm. Figure SI 6 of the Supporting Information shows 3 3 μm images of the gold slides covered by films with 20 and 30 mol % of GM1. The image of the film with 20 mol % of GM1 shows submicrodomains on terraces, whereas no inhomogeneity can be observed in the image of the film with 30 mol % GM1. ii. ForceDistance Curve Measurements. The thicknesses of the fBLMs containing 10, 20, and 30 mol % GM1 in the inner 10874
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
Langmuir
Figure 9. Breakthrough force and fBLM thickness plotted as a function of GM1 content (mol %) in the inner leaflet.
leaflet were determined from the measured forcedistance curves. Figure 8A shows a typical forcedistance curve measured for the fBLM composed of 10 mol % GM1 in the inner leaflet. The curve exhibits a weak repulsive interaction between the tip and the fBLM-covered gold surface at separations greater than 15 nm, governed by electrostatic forces that decay exponentially with distance. At separations of less than 15 nm, the tipsample interaction changes to a highly repulsive interaction because of an additional repulsive hydration force contributing to the measured forcedistance curve. When the distance between the tip and film-coated gold substrate is approximately 5 nm, a characteristic discontinuity is observed, followed by a vertical section that corresponds to elastic bending of the tip at the gold surface. The zero for the separation distance is determined by extrapolation of this section of the curve to the zero value of the force.45 The discontinuity corresponds to the penetration of the tip through the bilayer. Consequently, this “jump-in” distance is taken to be equal to the thickness of the film, and the force at which this event occurs is defined as the penetration force.45 In addition, the distribution of the measured film thickness is plotted (Figure 8C) along with the corresponding penetration forces (Figure 8B) determined from 115 forcedistance curves taken at several Au(111) terraces. The mean value of the Gaussian fitting of the histogram plot (4.57 ( 0.02 nm) is taken as the film thickness for the fBLM with 10 mol % GM1. The uncertainty of the measured thickness is always reported as √ the standard error (σ/ N), where σ is the standard deviation and N is the number of measurements. The average penetration force was calculated to be 2.7 ( 0.8 nN. It is important to note that the penetration force is dependent not only on the intrinsic properties of the bilayer being measured but also on the chemical properties of the tip, the magnitude of the cantilever spring constant, the radius of the AFM tip, and the approach velocity of the AFM tip in the force curve measurement. For all of these experiments, the same type of tip was used, and the approach velocity was maintained at 0.25 μm/s. The tip radius was taken as that specified from the manufacturer (20 nm), whereas the spring constant was determined from the thermal tune calibration technique33 as described in the Experimental Section. In Figure 9, the estimated thickness and penetration force are plotted as function of mole percentage of GM1 in the inner leaflet of the fBLM. For the fBLM containing 20 mol % GM1, the
ARTICLE
average penetration force was calculated to be 2.6 ( 0.8 nN, similar to the average force obtained for the fBLM containing 10 mol % GM1. The estimated film thickness for the fBLM with 20 mol % GM1 was determined to be 4.90 ( 0.06 nm. For the fBLM containing 30 mol % GM1, the average penetration force was calculated to be 9.2 ( 3.1 nN, which is about 3.5 times greater than the average force obtained for the fBLMs containing 10 and 20 mol % GM1. In addition, the estimated film thickness for the fBLM with 30 mol % GM is 5.61 ( 0.04 nm. In comparison to the fBLM containing 20 mol % GM1, the fBLM with 30 mol % GM1 has an estimated thickness that is ∼15% greater. These results are most likely because a homogeneous smooth film is produced with the incorporation of 30 mol % GM1, as shown in the AFM topography images. The estimated thicknesses of the fBLMs containing 10, 20, and 30 mol % GM1 in the inner leaflet are less than the theoretical thickness (∼7 nm) shown in Figure 1. However, the bilayer is elastically compressed before the tip ruptures the film. The jumpin distance measured from the force curve corresponds to the thickness of a compressed layer. The elastic deformation (δ) produced by the AFM tip under the load force can be determined using the formula derived from the Hertzian model46 !1=3 9F 2 δ¼ ð1Þ 16RE2 The radius of the tip curvature (R) was 20 nm, and the load force (F) was taken as the average penetration force. Because the effective compression modulus (E*) for GM1/DMPC/cholesterol is unknown, the E* value of a bilayer composed of DMPC and cholesterol was used (∼7 108 Nm2).47 The deformations (δ) for the fBLMs containing 10, 20, and 30 mol % GM1 were calculated to be 0.75, 0.72, and 1.69 nm, respectively. The deformation is 16%, 15%, and 30% of the estimated thickness from the force curve measurements of the respective fBLMs. The large elastic deformation for the fBLM containing 30 mol % GM1 is because of the large penetration force. However, the deformation is an estimate rather than an exact value. The average corrected thicknesses (penetration distance plus elastic deformation) for the fBLMs containing 10, 20, and 30 mol % GM1 are 5.32 ( 0.02, 5.62 ( 0.06, and 7.30 ( 0.04 nm, respectively. (The reported error for these values is associated with the uncertainty in the estimated thickness of the films.) The fBLM containing 30 mol % GM1 produces the thickest bilayer, which is closest to the theoretical value of ∼7 nm. To achieve such a thick bilayer, the polar group of GM1 should be oriented perpendicularly to the surface of the substrate, with the lipid acyl chains in their most extended state, as shown schematically in Figure 6b.
’ SUMMARY AND CONCLUSIONS We have designed a novel model of a BLM, floating on top of a hydrophilic monolayer of thioglucose and separated from the solid support by a ∼2-nm water-rich region. Bulky sugar headgroups of the ganglioside GM1 incorporated into the proximal leaflet of the bilayer act as hydrophilic pillars supporting the water-rich region separating the membrane from the solid surface. AFM was employed to characterize the structure of this membrane and to determine the optimal content of GM1. The results from the imaging and the thickness measurements for the fBLMs containing 10, 20, and 30 mol % GM1 are 10875
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
Langmuir
ARTICLE
information is available free of charge via the Internet at http:// pubs.acs.org/.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT This work was supported by grants from Natural Sciences and Engineering Council of Canada. J.L. acknowledges Canada Foundation for Innovation for the Canada Research Chair Award. ’ REFERENCES
Figure 10. AFM topography images combined with Gaussian thickness distributions from force curve measurements for the fBLMs composed of 10, 20, and 30 mol % GM1. From the thickness distribution (corrected for elastic deformation), it is observed that, with the addition of GM1 to the inner leaflet of the bilayer, there is an increase in the thickness of the bilayer. The images correlate well with the thickness distribution in that the 20 mol % GM1 fBLM, which is composed of four phases has the widest distribution of film thicknesses.
summarized in Figure 10, which relates the distributions of measured thickness values to the film topography images. For 10 and 30 mol % GM1 films, the film thickness distributions are narrow, consistent with the images that show a single phase for these bilayer compositions. In contrast, for the 20 mol % GM1 film, the distribution is much broader. This behavior is a consequence of the film inhomogeneity and the presence of different phases that exist within the bilayer, as shown in the AFM images. In conclusion, the fBLM containing 30 mol % GM1 yields the most uniform, smoothest, and thickest lipid bilayer. This fBLM should be a good model for further studies with transmembrane proteins. The results of this study can be considered as a proof of concept, with the method described in this work being employed to build similar fBLMs with different mixtures of lipids and polysaccharides containing larger sugar headgroups. The originality of this fBLM design is the use of biocompatible and hydrophilic sugar groups to create a waterrich space between the solid support and the membrane in which the lipid composition can vary.
’ ASSOCIATED CONTENT
bS
Supporting Information. Detailed description of the scaling analysis of AFM images, comparison of compression isotherms measured in this work to the literature data, excess mean molecular areas and excess Gibbs energies for mixed DMPC/cholesterol/GM1 monolayers, larger-scale AFM images of gold slides covered by fBLMs with 20% and 30% GM1. This
(1) (a) Tory, M. C.; Merrill, A. R. Biochim. Biophys. Acta 2002, 1564, 435. (b) Ho, D.; Lugo, M. R.; Lomize, A. L.; Pogozheva, I. D.; Singh, S. P.; Schwan, A. L.; Merrill, A. R. Biochemistry 2011, 50, 4830. (c) Ho, D.; Merrill, A. R. Biochemistry 2009, 48, 1369. (2) (a) Horswell, S. L.; Zamlynny, V.; Li, H.-Q.; Merrill, R. A.; Lipkowski, J. Faraday Discuss. 2002, 121, 405. (b) Zawisza, I.; Lachenwitzer, A.; Zamlynny, V.; Horswell, S. L.; Goddard, J. D.; Lipkowski, J. Biophys. J. 2003, 85, 4055. (c) Bin, X.; Zawisza, I.; Goddard, J. D.; Lipkowski, J. Langmuir 2005, 21, 330. (d) Bin, X.; Horswell, S. L.; Lipkowski, J. Biophys. J. 2005, 89, 592. (e) Bin, X.; Lipkowski, J. J. Phys. Chem. B 2006, 110, 26430. (f) Xu, S.; Szymanski, G.; Lipkowski, J. J. Am. Chem. Soc. 2004, 126, 12276. (g) Burgess, I.; Szymanski, G.; Li, M.; Horswell, S. L.; Lipkowski, J.; Majewski, J.; Satija, S. Colloids Surf. B: Biointerfaces 2005, 40, 117. (h) Burgess, I.; Szymanski, G.; Li, M.; Horswell, S. L.; Lipkowski, J.; Majewski, J.; Satija, S. Biophys. J. 2004, 86, 1763. (i) Sek, S.; Xu, S.; Chen, M.; Szymanski, G.; Lipkowski, J. J. Am. Chem. Soc. 2008, 130, 5736. (j) Li, M.; Chen, M.; Sheepwash, E.; Brosseau, C. L.; Li, H.-Q.; Pettinger, B.; Gruler, H.; Lipkowski, J. Langmuir 2008, 24, 10313. (3) (a) Zawisza, I.; Bin, X. M.; Lipkowski, J. Langmuir 2007, 23, 5180. (b) Garcia-Araez, N.; Brosseau, C. L.; Rodriguez, P.; Lipkowski, J. Langmuir 2006, 22, 10365. (c) Brosseau, C. L.; Bin, X.; Roscoe, S. G.; Lipkowski, J. J. Electroanal. Chem. 2008, 621, 222. (d) Brosseau, C. L.; Leitch, J.; Bin, X.; Roscoe, S. G.; Lipkowski, J. Langmuir 2008, 24, 13058. (4) (a) Zakharov, S. D.; Cramer, W. A. Front. Biosci. 2004, 9, 1311. (b) Lindeberg, M.; Zakharov, S. D.; Cramer, W. A. J. Mol. Biol. 2000, 295, 679. (5) Guidelli, R.; Aloisi, G.; Becucci, L.; Dolfi, A.; Moncelli, M. R.; Buoninsegni, F. T. J. Electroanal. Chem. 2001, 504, 1. (6) Rossi, C.; Chopineau, J. Eur. Biophys. J. 2007, 36, 955. (7) (a) Knoll, W.; Frank, C. W.; Heibel, C.; Naumann, R.; Offenhausser, A.; Ruhe, J.; Schmidt, E. K.; Shen, W. W.; Sinner, A. J. Biotechnol. 2000, 74, 137. (b) Wagner, M. L.; Tamm, L. K. Biophys. J. 2000, 79, 1400. (c) Castellana, E. T.; Cremer, P. S. Surf. Sci. Rep. 2006, 61, 429. (d) Lin, J.; Szymanski, J.; Searson, P. C.; Hristova, K. Langmuir 2010, 26, 12054. (8) (a) K€oper, I. Mol. Biosyst. 2007, 3, 651. (b) Knoll, W.; K€oper, I.; Naumann, R.; Sinner, E. K. Electrochim. Acta 2008, 53, 6680. (c) Becucci, L.; Innocenti, M.; Salvietti, E.; Rindi, A.; Pasquini, I.; Vassalli, M.; Foresti, L.; Guidelli, R. Electrochim. Acta 2008, 53, 6372. (d) Erbe, A.; Bushby, R. J.; Evans, S. D.; Jeuken, L. J. C. J. Phys. Chem. B 2007, 111, 3515. (e) Bunjes, N.; Schmidt, E. K.; Jonczyk, A.; Rippmann, F.; Beyer, D.; Ringsdorf, H.; Gr€aber, P.; Knoll, W.; Naumann, R Langmuir 1997, 13, 6188. (9) (a) McGillivray, D. J.; Valincius, G.; Vanderah, D. J.; Febo-Ayala, W.; Woodward, J. T.; Heinrich, F.; Kasianowicz, J. J.; L€osche, M. Biointerphases 2007, 2 (1), 21. (b) Atanasov, V.; Knorr, N.; Duran, R. S.; Ingebrandt, S.; Offenh€ausser, A.; Knoll, W.; K€oper, I. Biophys. J. 2005, 89, 1780. (10) Jeuken, L. J. C.; Bushby, R. J.; Evans, S. D. Electrochem. Commun. 2007, 9, 610. (11) Baumgart, T.; Kreiter, M.; Lauer, H.; Naumann, R.; Jung, G.; Jonczyk, A.; Offenh€ausser, A.; Knoll, W. J. Colloid Interface Sci. 2003, 258, 298. 10876
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
Langmuir (12) (a) Shenoy, S.; Moldovan, R.; Fitzpatrik, J.; Vanderah, D. J.; Deserno, M.; Loschoe, M. Soft Matter 2010, 6, 1263. (b) Wang, X.; Shindel, M. M.; Wang, S. W.; Ragan, R. Langmuir 2010, 26, 18239. (c) Munro, J. C.; Frank, C. W. Langmuir 2004, 20, 10567. (13) (a) Girard-Egrot, A. P.; Blum, L. J. Langmuir-Blodgett Technique for Synthesis of Biomimetic Lipid Membranes. In Nanobiotechnology of Biomimetic Membranes; Ferrari, M., Martin, D. K., Eds.; Fundamental Biomedical Technologies; Springer: Boston, MA, 2007; Vol. 1, Chapter 2, pp 2374. (b) Gy€orvary, E.; Wetzer, B.; Sleytr, U. B.; Sinner, A.; Offenh€ausser, A.; Knoll, W. Langmuir 1999, 15, 1337. (c) Abdulreda, M. H.; Moy, V. T. Biophys. J. 2007, 92, 4369. (d) Daillant, J.; Bellet-Amalric, E.; Braslau, A.; Charitat, T.; Fragneto, G.; Graner, F.; Mora, S.; Rieutord, F.; Stidder, B. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 11639. (e) Stidder, B.; Fragneto, G.; Roser, S. J. Langmuir 2005, 21, 9187. (14) (a) Junghaus, A.; Koper, I. Langmuir 2010, 26, 11035. (b) Vockenroth, I. K.; Rossi, C.; Shah, M. R.; Koper, I. Biointerphases 2009, 4, 19. (15) Leitch, J.; Kunze, J.; Goddard, J. D.; Schwan, A. J.; Faragher, R. J.; Naumann, R.; Knoll, W.; Dutcher, J. R.; Lipkowski, J. Langmuir 2009, 25 (17), 10354. (16) Heinrich, F.; Ng, T.; Vanderah, D. J.; Shekhar, P.; Mihailescu, M.; Nanda, H.; Loesche, M. Langmuir 2009, 25, 4219. (17) Tian, C.; Tetreault, E.; Huang, C. K.; Dahms, T. E. S. . Biochim. Biophys. Acta 2006, 1758, 693. (18) Simons, K.; Ikonen, E. Nature 1997, 387, 569. (19) Roefzaad, M.; Kluner, T.; Brand, I. Phys. Chem. Chem. Phys. 2009, 11, 1. (20) Ohta, Y.; Yokoyama, S.; Sakai, H.; Abe, M. Colloids Surf. B 2004, 33 (4), 191. (21) (a) Luckham, P.; Wood, J.; Froggatt, S.; Swart, R. J. Colloid Interface Sci. 1993, 156, 164. (b) Luckham, P.; Wood, J.; Swart, R. J. Colloid Interface Sci. 1993, 156, 173. (22) Reed, R. A.; Shipley, G. G. Biophys. J. 1996, 70, 1363. (23) Ali, S.; Minchey, S.; Janoff, A.; Mayhew, E. Biophys. J. 2000, 78, 246. (24) Beitinger, H.; Vogel, V.; M€obius, D.; Rahmann, H. Biochim. Biophys. Acta 1989, 984, 293. (25) Frey, S. L.; Chi, E. Y.; Arratia, C.; Majewski, J.; Kjaer, K.; Lee, K. Y. C. Biophys. J. 2008, 94 (8), 3047. (26) Hitzemann, R. J. Chem. Phys. Lipids 1987, 43, 25. (27) (a) Yuan, C.; Johnston, L. J. Biophys. J. 2000, 79, 2768. (b) Johnston, L. Langmuir 2007, 23, 5886. (c) Yuan, C.; Johnston, L. J. Biophys. J. 2001, 81, 1059. (d) Vie, V.; Van Mau, N.; Lesniewska, E.; Goudonnet, J. P.; Heitz, F.; Le Grimellec, C. Langmuir 1998, 14, 4574. (e) Ohta, Y.; Yokoyama, S.; Sakai, H.; Abe, M. Colloids Surf. B 2004, 33, 191. (f) Mou, J.; Yang, J.; Shao, Z. J. Mol. Biol. 1995, 248, 507. (g) Giocondi, M.-C.; Yamamoto, D.; Lesniewska, E.; Milhiet, P.-E.; Ando, T.; Le Grimellec, C. Biochim. Biophys. Acta 2010, 1798, 703. (28) Burgos, P.; Yuan, C.; Viriot, M.; Johnston, L. J. Langmuir 2003, 19, 8002. (29) Chen, M.; Li, M.; Brosseau, C. L.; Lipkowski, J. Langmuir 2009, 25, 1028. (30) Kaviratna, A. S.; Banerjee, R. Colloids Surf., A 2009, 345, 155. (31) Clavilier, J. In Interfacial Electrochemistry: Theory, Experiment, and Applications; Wieckowski, A., Ed.; Marcel Dekker: New York, 1999; p 231. (32) Han, W.; Serry, F. M. Force Spectroscopy with the Atomic Force Microscope; Application Note; Agilent Technologies: Palo Alto, CA, 2008; http://cp.literature.agilent.com/litweb/pdf/5989-8215EN.pdf. (Accessed 07/2011). (33) Hutter, J. L.; Bechhoefer, J. Rev. Sci. Instrum. 1993, 64, 1868. (34) Hinterdorfer, P.; Dufrene, Y. F. Nat. Methods 2006, 3, 347. (35) Johnson, W. T. How to Choose your MAC Lever; Application Note; Agilent Technologies: Palo Alto, CA, 2011; http://cp.literature. agilent.com/litweb/pdf/5989-9947EN.pdf. (Accessed 07/2011). (36) (a) Villarrubia, J. S. Surf. Sci. 1994, 321, 287. (b) Villarrubia, J. S. J. Res. Natl. Inst. Stand. Technol. 1997, 102 (4), 425. (37) Zhao, T.; Zagidulin, D.; Szymanski, G.; Lipkowski, J. Electrochim. Acta 2006, 51, 2255.
ARTICLE
(38) Tong, W. M.; Williams, R. S. Annu. Rev. Phys. Chem. 1994, 45, 401. (39) Gupta, R. K.; Suresh, K. A. Eur. Phys. J. E 2004, 14, 35. (40) Li, M.; Retter, U.; Lipkowski, J. Langmuir 2005, 21 (10), 4356. (41) Seul, M.; Sammon, M. J. Phys. Rev. Lett. 1990, 64 (16), 1903. (42) McMullen, T. P. W.; Lewis, R. N. A. H.; McElhaney, R. N. Biochemistry 1993, 32, 516. (43) (a) Tardieu, A.; Luzzatti, V.; Reman, F. C. J. Mol. Biol. 1973, 75, 711. (b) Janiak, M. J.; Small, D. M.; Shipley, G. G. Biochemistry 1976, 15, 4575. (c) Sengupta, K.; Raghunathan, V. A.; Katsaras, J. Phys. Rev. E 1999, 59, 2455. (d) Sun, W. J.; Tristram-Nagle, S.; Suter, R. M.; Nagle, J. F. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 7008. (e) Needham, D.; Evans, E. Biochemistry 1988, 27, 8261. (44) (a) Fang, Y.; Yang, J. J. Phys. Chem. 1996, 100, 15614. (b) Leidy, C.; Kaasgaard, T.; Crowe, J. H.; Mouritsen, O. G.; Jørgensen, K. Biophys. J. 2002, 83, 2625. (c) Kaasgaard, T.; Leidy, C.; Crowe, J. H.; Mouritsen, O. G.; Jørgensen, K. Biophys. J. 2003, 85, 350. (d) Woodward, J. T., IV; Zasadzinski, J. A. Phys. Rev. E 1996, 53, 3044. (e) Kaasgaard, T.; Leidy, C.; Ipsen, J. H.; Mouritsen, O. G.; Jørgensen, K. Single Mol. 2001, 2, 105. (45) Butt, H. J.; Cappella, B.; Kappl, M. Surf. Sci. Rep. 2005, 59, 1. (46) (a) Brown, M. F.; Thurmond, R. L.; Dodd, S. W.; Otten, D.; Beyer, K. J. Am. Chem. Soc. 2002, 124, 8471. (b) McWhirter, J. L.; Voth, G. A. Biophys. J. 2004, 87, 3242. (c) Duwe, H. P.; Sackmann, E. Physica A 1990, 163, 410. (47) Dufrene, Y. F.; Boland, T.; Schneider, J. W.; Barger, W. R.; Lee, G. U. Faraday Discuss. 1998, 111, 79. (48) Hauser, H.; Pascher, I.; Pearson, R. H.; Sundell, S. Biochim. Biophys. Acta 1981, 650, 21. (49) Majewski, J.; Kuhl, T. L.; Kjaer, K.; Smith, G. S. Biophys. J. 2001, 81, 2707.
10877
dx.doi.org/10.1021/la2016269 |Langmuir 2011, 27, 10867–10877
