pubs.acs.org/Langmuir © 2010 American Chemical Society
Quantification of the Layer of Hydration of a Supported Lipid Bilayer Theodore J. Zwang, Will R. Fletcher, Thomas J. Lane, and Malkiat. S. Johal* Chemistry Department, Pomona College, 645 North College Avenue, Claremont, California 91711 Received January 19, 2010. Revised Manuscript Received February 16, 2010 Dual polarization interferometry (DPI) and quartz-crystal microgravimetry (QCM-D) were used to investigate the adsorption of DOPC vesicles to a solid hydrophilic surface. The layer of hydration formed between a self-assembled DOPC bilayer and a silica solid support was probed in assemblies constructed using H2O and D2O buffers. We used QCM-D to measure the mass of the bilayer, including the mass contribution of the coupled solvent that resides between the membrane-solid interface. The mass of only the DOPC in the bilayer was resolved using DPI. By comparing these two measurements, and also accounting for the bulk phase effects on mass, we have been able to determine the mass of water below the bilayer. The thickness of this hydration layer, calculated by relating its mass to the density of the layer, was determined to be 10.46 A˚ ( 0.15 A˚ for trapped D2O and 10.21 A˚ ( 0.40 A˚ for trapped H2O, in agreement with measurements obtained by other methods. This work establishes the feasibility of concurrently using DPI and QCM-D to gauge the extent of hydration in thin films.
Introduction Lipid membranes are essential for cellular function, acting both as a selective barrier between the cell interior and its environment and as an interface for cell signaling responses and cell communication. Creating model membranes allow for the complexities of membranes to be characterized in hopes of furthering the understanding of lipid membrane function and physiology. One such complex feature, the cellular hydration state, is dynamic in vivo and changes within minutes under the influence of aniso-osmolarity, hormones, nutrients, and oxidative stress. Volume regulatory mechanisms act as dampeners in order to prevent excessive deviations in hydration that may be harmful to the cell. Small fluctuations of cell hydration can also act as signals for cellular metabolism and gene expression.1 Therefore, it is important to be able to characterize the layer of hydration at the lipid-substrate interface, which represents the cellular hydration in biomimetic systems, in order to more accurately mimic biological membranes. In this Letter, we investigate the layer of hydration formed between a self-assembled DOPC bilayer and the silica solid support it is adsorbed to by comparing data from the quartz crystal microbalance with dissipation monitoring (QCM-D) and the dual polarization interferometer (DPI). QCM-D measures the mass of an adsorbate, including coupled solvent, in real time. Using various sensor surfaces, QCM-D has been used extensively to characterize self-assembling lipid bilayers.2-5 In contrast, DPI measures the mass due to the adsorbate molecules only, assuming that adsorption does not significantly change the refractive index *Author to whom correspondence should be addressed. malkiat.johal@ pomona.edu; fax: (909) 607-7726; http://pages.pomona.edu/∼msj04747/.
(1) Haussinger, D. Biochem. J. 1996, 313, 697–710. (2) Viitala, T.; Hautala, J. T.; Vuorinen, J.; Wiedmer, S. K. Langmuir 2007, 23, 609–618. (3) Anderson, T. H.; Min, Y.; Weirich, K. L.; Zeng, H.; Fygenson, D.; Israelachvili, J. N. Langmuir, Article ASAP • DOI: 10.1021/la900181c • Publication Date (Web): 08 April 2009 (4) Richter, R. P.; Brat, R.; Brisson, A. R. Langmuir 2006, 22, 3497–3505. (5) Seantier, B.; Breffa, C.; Felix, O.; Decher, G. J. Phys. Chem. B 2005, 109, 21755–21765. (6) Cross, G. H.; Reeves, A. A.; Brand, S.; Popplewell, J. F.; Peel, L. L.; Swann, M.; Freeman, N. J. Biosens. Bioelectron. 2003, 19, 383–390. (7) Lu, J. R.; Swann, M. J.; Peel, L. L.; Freeman, N. J. Langmuir 2004, 20, 1827– 1832.
4598 DOI: 10.1021/la100275v
of the solvent at the solid-liquid interface. DPI has been used to characterize the formation of thin films at the solid-liquid interface,6-10 but has only recently been used to characterize lipid deposition and bilayer formation.11,12 We have capitalized upon the discrepancy between the solvent sensitivities of these techniques in order to calculate thin film solvent content. Whereas QCM-D is well-established in the field, DPI is a new noninvasive optical technique that uses a waveguide interferometer to determine the thickness and effective refractive index (RI) of a thin film, such as a bilayer. The instrument monitors both transverse magnetic and electric polarization modes, each of which generate evanescent fields that extend into the sensing region of the waveguide, but are of different intensities and decay at different rates with respect to the sensor normal. By using a reference waveguide to create a two-slit interference situation, each polarized mode will generate its own interference pattern, which, due to the evanescent field, will be sensitive to changes in the RI at the surface of the waveguide. Consequently, each will provide a separate calculation of the film RI as a function of film thickness, dependent on the surface excess and refractive index increment of the adsorbate. Only one unique pair of absolute RI and thickness values will generate the observed effective RI for both polarized modes, and this pair represents the value of the absolute RI and thickness of the film on the sensing waveguide surface. When calculating the unique solution pair, DPI determines the RI and thickness of lipid bilayers to within a fraction of an angstrom simultaneously and in real time. Furthermore, since RI is related to the film density, DPI can be used to calculate the mass of the lipid bilayer.6 Using QCM-D we have been able to measure the mass of the bilayer, including the mass of the solvent. Using DPI, we have been able to resolve the mass of the DOPC in the bilayer, (8) Swann, M. J.; Peel, L. L.; Carrington, S.; Freeman, N. J. Anal. Biochem. 2004, 329, 190–198. (9) Lane, T. J.; Fletcher, W. R.; Gormally, M. V.; Johal, M. S. Langmuir 2008, 24, 10633–10636. (10) Ro, T.; Murzyn, K.; Milhaud, J.; Karttunen, M.; Pasenkiewicz-Gierula, M. J. Phys. Chem. B 2009, 113, 2378–2387. (11) Mashaghi, A.; Swann, M.; Popplewell, J.; Textor, M.; Reimhult, E. Anal. Chem. 2008, 80, 3666–3676. (12) Terry, C. J.; Popplewell, J. F.; Swann, M. J.; Freeman, N. J.; Fernig, D. G. Biosens. Bioelectron. 2006, 22, 627–632.
Published on Web 02/26/2010
Langmuir 2010, 26(7), 4598–4601
Zwang et al.
Letter
excluding the mass of any solvent incorporated in the bilayer. By comparing these two measurements, we have been able to determine the mass of the coupled solvent in a bilayer and therefore the resulting thickness of the layer of hydration between a bilayer and its solid support. The thickness of the hydration layer was determined to be 10.46 A˚ ( 0.15 A˚ for trapped D2O and 10.21 A˚ ( 0.40 A˚ for trapped H2O, in agreement with measurements obtained by other methods.13-15
Experimental Section Materials. 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC) suspended in chloroform (Avanti Polar Lipids, Inc. CAS5664895-4), HEPES (Free Acid, Omnipur, CAS7365-45-9), deuterium oxide (99 atom % D Aldrich, CAS7789-20-0), and sodium chloride 99.5% (Aldrich, CAS7646-14-5) were used as received. Buffer was prepared with 10 mM HEPES and 150 mM NaCl in either ultrapure water (resistivity >18MΩ cm, Milli-Q) or deuterium oxide. The pH was adjusted to 7.4, with small additions of dilute NaOH or HCl as needed. The deuterium oxide buffer’s pH was made to have an effective pH 7.4 as calculated using data from Covington et al.16 Vesicles were prepared by drying the chloroform with inert nitrogen gas to create a thin lipid film, which was then dried overnight in a vacuum manifold. The lipids were then resuspended in either ultrapure water or deuterium oxide and sonicated in a bath sonicator for one hour on ice. The vesicles were then extruded 21 times through 100 nm membranes using Avanti Polar Lipid’s mini extruder. The vesicles formed in water were then added to the water buffer, and the vesicles formed in deuterium oxide were then added to the deuterium oxide buffer. As the main focus of this experiment is to test the viability of our method in characterizing the interfacial layer of hydration, choice of materials and procedure were based on an experiment conducted by Mashaghi et al.11 in which a simple lipid bilayer was successfully formed and characterized. DPI. An AnaLight Flex Bio200 dual-polarization interferometer (DPI, Farfield Scientific, Inc., Cheshire, U.K.) was used to obtain the optical properties of DOPC bilayers. The instrument utilizes a helium-neon laser (632.8 nm), a polarizer, a sensor consisting of two adjacent optical waveguides (AnaChip), and a CCD camera. Each sensor-chip’s active surface (0.15 cm2) was unmodified silicon oxynitride and had an average roughness of (5 nm. The fluidic system coupled to the instrument was comprising standard HPLC components: an autosampler (Jasco AS2055 Plus), a pump (Harvard Apparatus), and two three-way valves to direct flow to the two active channels of the chip independently. The two polarizations of light are sent through the chip’s surface, producing the interference patterns that constitute the raw data output. Density and thickness values of the lipid bilayer film were resolved simultaneously to less than 1 pg/ mm2 and 10 pm, respectively. Data was analyzed in AnaLight Explorer (Farfield Scientific) to calculate thickness, density, and mass values. Where appropriate, the lipid bilayer was modeled as an anisotropic layer, fixing the measured thickness while opening a new parameter, the layer anisotropy. This method is outlined in detail elsewhere,11 and has been found to give good results. Lipid bilayers were formed by flowing a 50 μg/L dispersion of vesicles across the DPI sensor at a flow rate of 10 μL per minute, and internal temperature was fixed at 25.0 °C ( 0.01 °C for all experiments. QCM-D. Real-time frequency and dissipation data were collected using a quartz crystal microbalance with dissipation (13) Roark, M.; Feller, S. E. Langmuir 2008, 24, 12469–12473. (14) Doshi, D. A.; Dattelbaum, A. M.; Watkins, E. B.; Brinker, C. J.; Swanson, B. I.; Shreve, A. P.; Parikh, A. N.; Majewski, J. Langmuir 2005, 21, 2865–2870. (15) Koenig, B. W.; Krueger, S.; Orts, W. J.; Majkrzak, C. F.; Berk, N. F.; Silverton, J. V.; Gawrisch, K. Langmuir 1996, 12, 1343–1350. (16) Covington, A. K.; Paabo, M.; Robinson, R. A.; Bates, R. G. Anal. Chem. 1968, 40, 700–706.
Langmuir 2010, 26(7), 4598–4601
Figure 1. DOPC bilayer formation monitored by QCM-D, showing both the frequency and the dissipation response of the 3rd harmonic. Data for both H2O (a) and D2O (b) situations are shown. Regions correspond to formed bilayer under pH 7.4 HEPES (A), under pure solvent (B), and again under HEPES (C). (QCM-D) monitoring (E4, Q-Sense, Gothenburg, Sweden). The QCM-D sensor, mounted in a liquid flow cell (40 μL), consisted of an AT-cut piezoelectric quartz crystal disk operated at 4.95 MHz. Crystals were exposed to solution on the active SiO2 surface (0.2 cm2 in area, 50 nm thick) and were coated on the opposite side with a Au electrode (100 nm thick). All QCM-D crystals were optically polished with a root-mean-square roughness less than 3 nm. Crystals were decontaminated by UV-ozonation for 10 min, treated with 2 vol % Hellmanex solution (Hellma GmbH & Co.) for 15-30 min, rinsed with ultrapure water, blown dry with N2, and finally treated again with UV-ozonation before use. A stable baseline was obtained by flushing the QCM-D flow cell with ultrapure water prior to vesicle adsorption, accomplished by flowing the HEPES buffered vesicle dispersion through the cell. Flow cell temperature was fixed at 25.00 ( 0.02 °C, and a peristaltic pump (Ismatec ISM935C) was used to flow solution through the cell at a constant rate of 100 μL/min. Details of QCM-D principles and operation can be found elsewhere.17 DOPC vesicles were prepared in HEPES buffered H2O or D2O. In all QCM-D experiments, a baseline was achieved by flowing pure H2O over a SiO2-coated QCM-D sensor. Bilayers were formed by flowing 0.05 mg/mL buffered DOPC dispersions across the SiO2 QCM crystal, until after a rupture event took place and the frequency stabilized. After this point, pure H2O (not D2O) was rinsed over the sensor surface. This sequence can be represented as H2O (pure) f DOPC (HEPES) f H2O (pure) f H2O (HEPES). In a separate experiment, the sequence H2O (pure) f DOPC (d-HEPES) f H2O (pure) f D2O (d-HEPES), was followed, where d-HEPES indicated that the buffer was prepared in D2O. These experiments then represent two situations: (1) H2O trapped between the bilayer and the solid support, and (2) D2O trapped between the bilayer and the solid support. In both cases, H2O is the bulk phase. Data were acquired using QSoft (Q-Sense) software and analyzed with QTools (Q-Sense).
Results and Discussion Since vesicle fusion and subsequent rupture at the silica interface has been observed previously using QCM-D,2-5 we discuss only briefly the changes in frequency (ΔF) and dissipation (ΔD). Figure 1a shows changes in ΔF and ΔD for the third harmonic after exposing the SiO2-coated QCM-D sensor to DOPC dispersion for the sequence H2O (pure) f DOPC (HEPES) f H2O (pure)f DOPC (HEPES). Once vesicles are exposed to the surface, vesicle adsorption occurs and results in a significant decrease in ΔF (17) Rodahl, M.; Hook, F.; Krozer, A.; Brzezinski, P.; Kasemo, B. Rev. Sci. Instrum. 1995, 66, 3924–3930.
DOI: 10.1021/la100275v
4599
Letter
(Figure 1). The corresponding ΔD values increase, reflecting the build-up of a loosely adsorbed vesicle layer that is separated from the silica sensor by a trapped layer of solvent. The vesicle rupture event is subsequently observed as an increase in ΔF and a decrease in ΔD. These observations in ΔF and ΔD are consistent with the loss of water within the vesicles and reorganization of the lipids in the layer leading to the formation of a stable DOPC bilayer on top of a trapped layer of solvent (Figure 1, A). In region A (Figure 1a), a layer of H2O (HEPES) is present between the solid support and the DOPC bilayer. In this assembly, bulk H2O (HEPES) is present above the membrane. This film is then exposed to a H2O rinse, leading to a bilayer under pure H2O (Figure 1a, B) and containing a thin layer of H2O (HEPES) between the membrane and the solid support. The film is exposed to a H2O (HEPES) rinse, resulting in a bilayer at C under similar conditions to that at A (i.e., H2O, HEPES) above and below the membrane. By switching again back to the buffer, we are able to demonstrate that the bilayer at region C retains the same mass as it had at region A, despite exposure to different conditions (unbuffered H2O). In agreement with Seantier et al.,5 we have not observed a change in adsorbed mass with a change in pH; because of this, we are not concerned with discrepancies in measuring the mass of preformed bilayers under buffered or unbuffered conditions. In Figure 1a, region B represents the ΔF (-24.39 Hz) value corresponding only to the bilayer and the coupled hydration (H2O) layer beneath it. In the thin, elastic limit, we can directly relate the change in frequency and the mass adsorbed, as proven by Sauerbrey.18 Using this method, the mass of the bilayer and the coupled hydration layer beneath it was found to be 431.8 ng/cm2. The ΔD value is near zero indicating a rigid and strongly coupled membrane at the SiO2 surface, which supports the accuracy of the Sauerbrey approximation. In a separate experiment, in which the sequence H2O (pure) f DOPC (d-HEPES) f H2O (pure) f D2O (d-HEPES) was followed, a bilayer with a deuterated hydration layer (d-HEPES) beneath it was created (Figure 1b). The ΔF value and the Sauerbrey mass for this layer were -26.64 Hz and 471.5 ng/cm2, respectively. The difference between this mass and that reported from Figure 1a (39.7 ng/cm2) reflects the fact that one assembly contains a layer of D2O and the other contains H2O. These experiments also suggest that the layer of hydration is not in exchange with bulk solvent. In the transition between regions A and C, no hysteresis is observed, indicating D2O molecules trapped below the membrane do not exchange with H2O molecules above the membrane. This gives us confidence that we have indeed quantified the mass of the lipid bilayer and the layer of hydration beneath it, without any exchange interference. Figure 2a shows the mass of the bilayer determined using DPI. In this experiment, a baseline was obtained using a HEPES buffer (H2O). This was followed by flowing DOPC (HEPES) over the DPI waveguide in order to form a stable lipid bilayer similar to the one formed in the QCM experiment. Previously, it has been found that the QCM and DPI sensor surfaces are negligibly different.19 Data indicate that the final bilayers formed on each system are completely analogous. Other conditions were constant, ensuring the bilayer formed on one system should be directly comparable to a bilayer formed on the other. The DPI measurement is insensitive to solvent, so the determined mass was only for the lipid bilayer adsorbed to the (18) Sauerbrey, G Z. Phys. 1959, 155, 206–222. (19) Aulin, C.; Verga, I.; Claesson, R. M.; Waberg, L.; Lindstrom, T. Langmuir 2008, 24, 2509–2518.
4600 DOI: 10.1021/la100275v
Zwang et al.
Figure 2. DPI response upon adsorption of a DOPC bilayer. Reported is the mass of the layer assuming a fixed film thickness of 4.5 nm. Reported are bilayers under both H2O (a) and D2O (b) solvents.
waveguide and did not include the bulk solvent or the trapped layer of hydration. In order to obtain a valid mass measurement, we have used the anisotropic modeling developed by Farfield,12 which requires a fixed value of either the refractive index or thickness. For this study, the thickness was set to 4.5 nm and the mass was calculated from the variation in the RI. Figure 2 shows the results, which are distinctly different than the QCM data. There is a complete absence of evidence of vesicle fusion and rupture, since DPI is insensitive to the changes in coupled solvent that make these phenomenon so prevalent in the QCM data. From the final, equilibrium mass, the “dry” DOPC bilayer was determined to be 330 ng/cm2. Figure 2b shows these measurements using D2O as a solvent instead of H2O, yielding a “dry” mass of 356 ng/cm2. This difference can be explained by the results from a molecular dynamics simulation conducted by Rog et al.20 which determined that a bilayer hydrated by deuterium oxide is of higher density than those hydrated in water. This is believed to be due to the relative strength of hydrogen bonds between the phosphatidylcholine headgroups and deuterium oxide, which are stronger than hydrogen bonds in the aqueous system. By subtracting the “dry” mass from the corresponding Sauerbrey masses, we obtain the water content at the membrane-support interface. Relating this mass to the density of the solvent trapped under the bilayer, which is assumed to have density similar to the bulk,11 the thickness of the layer of hydration is obtained. The thickness of the layer of hydration was 10.21 ( 0.40 A˚ and 10.46 ( 0.15 A˚ for H2O and D2O layers, respectively. We note that there seems to be no significant thickness difference between these two situations. These values are summarized in Table 1. The values we have obtained agree well with the literature; previous studies focused on scattering rather than gravimetric techniques to determine the thickness of the layer of hydration. Roark et al.,13 through use of a molecular dynamics, postulated that the layer of hydration is between 10 and 15 A˚ thick. Using neutron scattering techniques, Doshi et al.14 found for the silicaPOPC system that the thickness of the layer of hydration, including the lipid headgroup, was 14.6 A˚. The hydration layer was not discernible due to lack of contrast. Koenig et al.15 attempted to separate the layer of hydration and found the thickness of the water layer between the solid support and lipid headgroup to be (20) Rog, T.; Murzyn, K.; Milhaud, J.; Karttunen, M.; Pasenkiewicz-Gierula, M. J. Phys. Chem. B 2009, 113, 2378–2387.
Langmuir 2010, 26(7), 4598–4601
Zwang et al.
Letter Table 1. Layer of Hydration Thicknessa
trapped layer
combined LBþLoH mass (QCM-D)
LB mass (DPI)
LoH mass
LoH density
LoH thickness (mass/density)
471.1 ng/cm2 356 ng/cm2 115.5 ng/cm2 ( 1.7 ng/cm2) 1.104 g/cm3 10.46 A˚ ( 0.15 A˚ D2O 431.8 ng/cm2 330 ng/cm2 101.8 ng/cm2 ( 4.0 ng/cm2) 0.997 g/cm3 10.21 A˚ ( 0.40 A˚ H2O a Summary of results (LB, Lipid Bilayer; LoH, Layer of Hydration). Additive mass of lipid bilayer and layer of hydration measured with QCM-D. Mass of lipid bilayer measured with DPI. Layer of hydration mass determined by subtracting the lipid bilayer mass from the additive mass. Layer of hydration thickness determined by relating the trapped layer’s mass to its density.
between 11.5 and 16 A˚ using a specially designed reflection setup. The thickness values obtained from neutron scattering are inherently plagued by uncertainty. Due to hydration of the head groups, surface roughness, and the highly dynamic state of the bilayer, no definite scattering plane exists, and so data modeling is difficult, leading to errors and uncertainty. Our method circumvents these limitations, providing an accurate quantification of the layer of hydration.
Conclusion The layer of hydration formed between a self-assembled DOPC bilayer and a silica solid support was probed in assemblies constructed using H2O and D2O buffers. The mass of the bilayer including the mass of the DOPC and the solvent was measured using QCM-D. The mass of the “dry” bilayer was resolved using
Langmuir 2010, 26(7), 4598–4601
DPI. By comparing these two measurements, we have been able to determine the mass of the hydration of a bilayer. The thickness of the layer of hydration between a bilayer and its solid support was calculated by relating this mass to the density of the layer. The value obtained for the thickness of the hydration layer, 10.21 ( 0.40 A˚ for H2O, is in agreement with measurements obtained by other methods. The accuracy and ease of these measurements leads us to conclude that this experimental method would be appropriate for similar treatment of other systems, being especially well suited to biological applications. Acknowledgment. This research was funded in part by the Gordon and Betty Moore Foundation, and the Arnold and Mabel Beckman Foundation. We also thank the Pomona College SURP program and the Department of Chemistry for continued support.
DOI: 10.1021/la100275v
4601
