Anal. Chem. 2005, 77, 6096-6099
Infrared Spectroscopy of Fluid Lipid Bilayers Marshall C. Hull, Lee R. Cambrea, and Jennifer S. Hovis*
Department of Chemistry, Purdue University, West Lafayette, Indiana 47907-2018
Infrared spectroscopy is a powerful technique for examining lipid bilayers; however, it says little about the fluidity of the bilayersa key physical aspect. It is shown here that it is possible to both acquire spectroscopic data of supported lipid bilayer samples and make measurements of the membrane fluidity. Attenuated total reflectionFourier transform infrared spectroscopy (ATR-FT-IR) is used to obtain the spectroscopic information and fluorescence recovery after photobleaching (FRAP) is used to determine the fluidity of the samples. In the infrared spectra of lipid bilayers composed of 1-palmitoyl-2-oleoylsn-glycero-3-phosphocholine, the following major peaks were observed; νas(CH3) 2954 cm-1, νs(CH3) 2870 cm-1, νas(CH2) 2924 cm-1, νs(CH2) 2852 cm-1, ν(CdO) 1734 cm-1, δ(CH2) 1463-1473 cm-1, νas(PO2-) 1226 cm-1, νs(PO2-) 1084 cm-1, and νas(N+(CH3)3) 973 cm-1. The diffusion coefficient of the same lipid bilayer was measured to be 3.5 ( 0.5 µm2/s with visual recovery also noted through use of epifluorescence microscopy. FRAP and visual data confirm the formation of a uniform, mobile supported lipid bilayer. The combination of ATR-FT-IR and FRAP provides complementary data giving a more complete picture of fully hydrated model membrane systems. Supported lipid bilayers have proven to be highly useful systems for examining lipid bilayers.1-3 One of the key aspects of the solid supports is that they allow the lipids to retain their natural fluidity.1,4,5 Lipid fluidity is known to affect a number of membrane processes; it can facilitate the interaction of membrane proteins with their lipophilic ligands, regulate the function of various receptors, and affect the response of microorganisms to environmental stresses.6-9 To develop a more complete understanding of these processes, we would like to be able to do both spectroscopy and microscopy on the samples. In this paper, we will show that by using silicon for the solid support it is possible to do the following: image supported lipid bilayers using epifluo* Corresponding author. Phone: (765) 494-4115. Fax: (765) 494-0239. E-mail:
[email protected]. (1) Brian, A.; McConnell, H. M. Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 61596163. (2) Sackmann, E. Science 1996, 271, 43-48. (3) Richter, R. P.; Him, J. L. K.; Brisson, A. Mater. Today 2003, 6, 32-37. (4) Groves, J. T.; Boxer, S. G. Biophys. J. 1995, 69, 1972-1975. (5) Cremer, P. S.; Boxer, S. G. J. Phys. Chem. B 1999, 103, 2554-2559. (6) Beney, L.; Gervais, P. Appl. Microbiol. Biot. 2001, 57, 34-42. (7) Jain, M. K.; White, H. B. Adv. Lipid Res. 1977, 15, 1-60. (8) Heldin, C. H. Cell 1995, 80, 213-223. (9) Mammen, M.; Chio, S. K.; Whitesides, G. M. Angew Chem., Int. Ed. 1998, 37, 2755-2794.
6096 Analytical Chemistry, Vol. 77, No. 18, September 15, 2005
rescence microscopy; determine the diffusion coefficient of the lipids using fluorescence recovery after photobleaching (FRAP); and collect infrared spectra using attenuated total reflection Fourier transform infrared spectroscopy (ATR-FT-IR)sall on the same sample. ATR-FT-IR is a powerful technique for obtaining information about the conformation of lipid headgroups (e.g., νCN+C, νPO2-), the hydration/hydrogen bonding between and around lipids (e.g., νCO, νPO2-), and the orientation/conformation of the hydrocarbon chains (e.g., νCH2, νCH3).10,11 ATR-FT-IR also allows one to obtain important information about the orientation of lipids in bilayers and the intermolecular interactions involved in assembly. For example, the peaks arising from the phosphate and carbonyl groups (hydrogen bond acceptors) are known to shift when participating in intermolecular hydrogen bonding.12 There are a number of papers in the literature using FT-IR or ATR-FT-IR to study lipids and supported lipid bilayers. To our knowledge, however, it has not been shown that the lipids in these experiments undergo long-range motion. As lipids in cellular systems undergo long-range motion, and this in fact is one of the hallmarks of the cell membrane, it is important to be able to determine that the systems studied spectroscopically retain this key feature. FRAP allows one to ensure that long-range diffusion is occurring and gives information about the fluidity of the membranes. Infrared spectra, on the other hand, give no information about long-range motion and limited information about membrane fluidity; a small increase in ν(CH2) from ∼2849 and ∼2917 cm-1 to ∼2853 and ∼2923 cm-1 has been observed when lipids go from the ordered gel phase to the disordered liquid crystalline phase.13 This same phase transition causes changes in the diffusion coefficient of 2 orders of magnitude, clearly indicating that FRAP is a vastly more sensitive technique for measuring diffusion.14 Typically, ATR-FT-IR of lipids has been done using germanium as the ATR element. Germanium has good transmission properties through the two most important spectral regions 900-1800 and 2800-3100 cm-1; however, germanium becomes problematic due to its lack of a stable oxide layer.15 It has been well documented that fluid-supported lipid bilayers form on very few surfaces (mica;16,17 glass, quartz, and silicon with native oxide layers;1,5,18 (10) (11) (12) (13) (14) (15)
Tamm, L. K.; Tatulian, S. A. Q. Rev. Biophys. J. 1997, 30, 365-429. Tatulian, S. A. Biochemistry 2003, 42, 11898-11907. Lewis, R. N. A. H.; McElhaney, R. N. Biophys. J. 1993, 64, 1081-1096. Mantsch, H. H.; Lewis, R. N. A. H. Chem. Phys. Lipids 1991, 57, 213-226. Mouritsen, O. G. Life-As a Matter of Fat; Springer-Verlag: Berlin, 2005. Hovis, J. S.; Hamers, R. J.; Greenlief, C. M. Surf. Sci. 1999, 440, L815L819. (16) McKiernan, A. E.; Ratto, T. V.; Longo, M. L. Biophys. J. 2000, 79, 26052615. (17) Ratto, T. V.; Longo, M. L. Langmuir 2003, 19, 1788-1793. 10.1021/ac050990c CCC: $30.25
© 2005 American Chemical Society Published on Web 08/09/2005
oxidized PDMS;19,20 and titanium dioxide21), and to our knowledge, it has never been established that they form on germanium with or without an oxide layer. Vesicle fusion is the most common method for creating supported lipid bilayers. However, if the surface conditions are not correct, the vesicles will simply sit on the surface. Consequently, we have elected to use silicon, which has a stable oxide layer that fluid lipid bilayers have been documented to form on. The formation of supported lipid bilayers is highly dependent on having exactly the right surface chemistry;5,19 thus, silicon has another advantagesdouble-sided polished silicon wafers are inexpensive, easy to obtain, and from them ATR elements can be created in-house. As a result, it is possible to use a new ATR element for every experiment, allowing one to retain the necessary control of the surface chemistry. Silicon, like germanium, is equally sensitive to dipoles laying parallel to the surface as to those lying perpendicular to the surface; by using polarized light, orientational information can be obtained in addition to chemical information.22 It is commonly assumed that light will not pass through silicon in the region between 900 and 1800 cm-1, which is why germanium has been favored; this is not quite true as the transmission drops off but never to zero.23 By using a shorter path length than has traditionally been used and by taking advantage of improvements to spectrometers and detectors this region is accessible in silicon.24 EXPERIMENTAL SECTION Preparation of Large Unilamellar Vesicles (LUVs). LUVs composed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC, Avanti Polar Lipids) with the addition of 5 mol % 1-oleoyl2-[6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl]-sn-glycero3-phosphocholine (NBD-PC, Avanti Polar Lipids) were prepared by vesicle extrusion.1,5 Lipids dissolved in chloroform were combined in appropriate amounts, dried under N2, and placed under vacuum for at least 45 min. The lipids were then reconstituted with a buffer solution (300 mM KCl, 50 mM KH2PO4, and 0.1 mM EDTA, adjusted to pH 7.4 with 1 M KOH) to a concentration of 2 mg/mL and passed 21 times through an Avanti extruder containing a polycarbonate membrane with a 50-nm pore size. Following extrusion, the LUV solution was centrifuged for 5 min at 14 000 rpm (Eppendorf Minispin Plus). The extruded vesicles were held at 20 °C, shielded from light, and used within 1 day. Preparation of Silicon Attenuated Total Reflection (ATR) Elements. Double-side polished silicon (001) wafers (>10 Ω‚cm resistivity, ∼525 µm thick, Silicon Inc.) were cut to 15 mm × 9 mm. A polymer, Crystal Bond 509 (Electron Microscopy Sciences), was used to adhere the wafer to an aluminum block with a 45° angle; the polymer protects the optically flat front and back surfaces of the wafer. The short edges of the wafer were beveled to 45° angles through use of a series of 120-1500-grit SiO2 (18) Keller, C. A.; Kasemo, B. Biophys. J. 1998, 75, 1397-1402. (19) Lenz, P.; Ajo-Franklin; M., C.; Boxer, S. G. Langmuir 2004, 20, 1109211099. (20) Hovis, J. S. B.; Steven, G. Langmuir 2001, 17, 3400-3405. (21) Starr, T. E.; Thompson, N. L. Langmuir 2000, 16, 10301-10308. (22) Chabal, Y. J. Surf Sci. Rep. 1988, 8, 211-357. (23) Harrick, N. J. Internal Reflection Spectroscopy; Interscience Publishers: New York, 1967. (24) Queeney, K. T.; Fukidome, H.; Chaban, E. E.; Chabal, Y. J. J. Phys. Chem. B 2001, 105, 3903-3907.
Figure 1. Schematic illustration of flow cell with ATR element. A supported lipid bilayer is shown on both sides of the ATR element.
sandpapers and then mirror polished using 5- and 0.3-µm alumina lapping film (Fiberlap Technologies); excess water was present during all polishing stages. The polymer was softened with heat to release the wafer from the block. Excess polymer was removed from the wafer through sonication with solvents: twice in acetone, once with methanol, and once with electronic grade methanol (99% residue free). Each sonication step lasted 15 min. Although, the Si(001) surface is protected by a 3-4-layer-thick oxide layer, which is hydrophilic, we found that vesicle fusion proceeded more rapidly if the wafers were exposed to a 4:1 mixture of sulfuric acid and hydrogen peroxide (piranha etch) for 60 min. This additional step renders the surface more hydrophilic (presumably through the creation of more surface hydroxyl groups); additionally, this treatment does not change the surface topography.25,26 After piranha treatment, the silicon wafer was rinsed exhaustively in ultrapure water (NANOpure Ultrapure Water System, Barnstead). ATR-Fourier Transform Infrared Spectroscopy (FT-IR). ATR-FT-IR spectra were obtained using a Nicolet 470 FT-IR equipped with a MCTA* detector and a custom-designed ATR setup. The light is sent out of the spectrometer into a plexiglass box continuously purged by a N2 gas generator (Parker Balston model 74-5041). Infrared light is focused onto the edge of the ATR element (15 mm × 9 mm × 525 µm silicon wafer) using silvercoated off-axis parabolic mirrors, collected, and focused onto the MCTA* detectorswhich has been moved out of the spectrometer and into the purged plexiglass box. Rather than sending the IR light in normal to the sample beveled edge, as is typical in an ATR experiment, the light is sent in at a 45° angle. The 45° incident angle results in a 57° angle at the silicon/bilayer interface (to the surface normal). The light bounces ∼19 times before exiting; the shorter than usual path length coupled with the shallower than usual reflection angle helps reduce the amount of absorption by the silicon crystal. The custom-made sample flow cell (54.8 mm2 area of exposure) is machined from Delrin (Figure 1 adapted from ref 10). The flow cell and ATR element are held together using a Delrin vise attached to a magnetic mount on a linear translational stage. The magnetic mount facilitates the removal of the flow cell assembly from the purged plexiglass box. The translational stage allows for the interferogram signal to be maximized before an experiment is started. For both the background and sample spectra, 300 scans were signal averaged at a resolution of 4 cm-1 using Happ-Genzel apodization and zero filling. A background of the silicon ATR element and the buffer system was collected and before introducing the lipids. To form supported lipid bilayers, LUVs were injected (25) Yee, C. K.; Amweg, M. L.; Parikh, A. N. J. Am. Chem. Soc. 2004, 126, 13962-13972. (26) Kiessling, V.; Tamm, L., K. Biophys. J. 2003, 84, 408-418.
Analytical Chemistry, Vol. 77, No. 18, September 15, 2005
6097
into the flow cell and allowed to incubate for 30 min. Once the bilayer had formed, buffer (same as used for vesicle extrusion) was flushed through the flow cell to rinse away excess vesicles and the sample spectrum was obtained. The flow cell was then removed from the ATR-FT-IR setup and disassembled under ultrapure water, taking care not to scrape or expose the silicon wafer to air (bilayers tightly adhere to the supports and, consequently, moving the samples will not alter them unless they are exposed to air or roughly scraped). The wafer was then placed into a custom glass sample holder, sandwiched with a coverslip, and placed onto a Delrin holder with wells of ultrapure water to keep the silicon wafer hydrated throughout imaging and bleaching of the bilayer. Imaging and FRAP. A Nikon TE2000 fluorescence microscope equipped with a Cascade 512B CCD camera (Roper Scientific) was used to image the bilayers. NBD fluorophores were imaged using a NBD filter set (Chroma Technology Corp.). For the FRAP experiments, a silicon avalanche photodiode (PerkinElmer Optoelectronics) was used for photon counting and a 25mW argon ion laser (488 nm Melles Griot) was used to bleach a 14-µm-radius spot in the bilayer (bleach time of 1 s, less than 1% of total recovery time). Fluorescence recovery was monitored by inserting a 5× neutral density filter (Oriel Instruments) in the beam path, reducing the laser intensity to 250 nW. Counts were monitored using a custom LabVIEW program that controls the shutter between the laser and the sample. In the FRAP technique, fluorophores covalently attached to a small percentage of the lipids are irreversibly bleached in a large area; as the lipids are undergoing Brownian motion, the fluorescence intensity in the bleach region returns to the original value as unbleached lipids diffuse in and the bleached lipids diffuse out. The diffusion coefficient was determined by fitting the recovery curve with a solution to the differential equation for lateral transport of a molecule by diffusion,27 using the method of Soumpasis.28 RESULTS AND DISCUSSION The transmission through one of our silicon ATR elements convolved with the detector response is shown in Figure 2. The MCTA* detector cuts out below 850 cm-1, and a drop in the detector sensitivity contributes to the steep drop in transmission below 1500 cm-1; the bulk of the drop in transmission below this point is due to absorption of IR light by silicon phonon modes.23 From 4000 to 1500 cm-1, the transmission through silicon is flat;23 in this region, the changes in transmission are due to the detector response and absorption by the Delrin flow cell. We have examined whether the Delrin peaks appearing in the lipid C-H stretching region (2990-2812 cm-1) change over time, due perhaps to spreading of the material; we find the peaks do not change over at least 24 h. In Figure 3A, we show an IR spectrum acquired after a supported lipid bilayer of POPC with 5 mol % NBD-PC was formed. The ratio of the spectrum was taken against one with buffer in the flow cell; consequently, a decrease in the water bending (1645 cm-1) and stretching (3490 and 3280 cm-1) modes are observed as the lipids replace buffer at the silicon wafer surface. There are (27) Axelrod, D.; Kopp, D. E.; Schlessinger, J.; Elson, E.; Webb, W. W. Biophys. J. 1976, 16, 1055-1069. (28) Soumpasis, D. M. Biophys. J. 1983, 41, 95-97.
6098 Analytical Chemistry, Vol. 77, No. 18, September 15, 2005
Figure 2. Infrared transmission through a silicon ATR element convolved with the detector response.
Figure 3. (A) Background corrected infrared spectrum acquired after formation of a bilayer of 95 mol % POPC/5 mol % NBD-PC on a silicon ATR element. The spectrum was acquired at 4-cm-1 resolution. (B) Same spectrum as Figure 3a showing specific lipid peaks in more detail.
two primary contributions to the water stretching mode. The coupled OH symmetric stretching mode, located around 3280 cm-1, arises from tetrahedrally coordinated water molecules at the interface29sthe “icelike” molecular structure.30-33 A more (29) Kim, J.; Kim, G.; Cremer, P. S. Langmuir 2001, 17, 7255-7260. (30) Scherer, J. R.; Go, M. K.; Kint, S. J. Phys. Chem. B 1974, 78, 1304-1313. (31) Eisenberg, D.; Kauzmann, W. The Structure and Properties of Water; Oxford University Press: New York, 1969.
Figure 4. FRAP recovery data of 95% POPC and 5% NBD-PC bilayer from Figure 3. Inset pictures show visual recovery of bleached spot.
disordered hydrogen bonding or a “waterlike” molecular structure,29 located around 3490 cm-1, arises from either an OH symmetric stretch from asymmetrically hydrogen-bonded water31 or water molecules with bifurcated hydrogen bonds.34 In Figure 3B, the peaks arising from the lipids are shown in further detail; the major peaks are assigned as follows:35 CH alkene stretch (3010 cm-1), CH3 antisymmetric stretch (2954 cm-1), CH3 symmetric stretch (2870 cm-1) CH2 antisymmetric stretch (2924 cm-1), CH2 symmetric stretch (2852 cm-1), CdO stretch (1734 cm-1), CH2 scissoring (1463-1473 cm-1), PO2- antisymmetric stretch (1226 cm-1), PO2-symmetric stretch (1084 cm-1), and N+(CH3)3 antisymmetric stretch (973 cm-1). The dip in the PO2- symmetric stretch at ∼1100 cm-1 is due to noise; this is where the transmission through silicon drops to nearly zero. Due to the decrease in water, it is difficult to see the alkene antisymmetric and symmetric stretching regions, arising from the one double bond in the lipid tail region; when bilayers are formed in D2O, the peaks are clearly seen. With 300 scans (6 min), the peaks in the region from 4000 to 1200 cm-1 can be clearly resolved. To resolve the peaks below 1200 cm-1, it is necessary to acquire 1600 scans (45 min). The stability of the system has been examined, and it is found that the ratio of the spectra against background spectra can be acquired up to the previous 24 h. To confirm that fluid-supported lipid bilayers were indeed formed on the surface, the sample was transferred to the epifluorescence microscopesbeing careful to ensure that during transfer the sample remained fully hydrated. In Figure 4, fluo(32) Walrafen, G. E. Raman and infrared spectral investigations of water structure; Plenum: New York, 1972. (33) Yalamanchili, M. R.; Alia, A. A.; Miller, J. D. Langmuir 1996, 12, 41764184. (34) Giguere, P. A. J. Raman Spectrosc. 1984, 15, 354-359. (35) Krimm, S.; Bandekar, J. Adv. Protein Chem. 1986, 38, 181-364.
rescence recovery after photobleaching data is shown, confirming that the bilayer is fluid and undergoing long-range motion. The data were fit using the method of Soumpasis28 giving a diffusion coefficient of 3.8 µm2/s. The residuals are shown in the plot; they clearly indicate that the data fits to theory very well. The average diffusion coefficient for three trials was determined to be 3.5 ( 0.5 µm2/s. The measured diffusion coefficient is very similar to what we measure for POPC bilayers on glass substrates. Visual fluorescence recovery (over 95%) can be observed from the time lapse images inset in Figure 4. It is logistically impossible to measure the diffusion across the entire area that the IR beam samples; however, the entire sample was examined with the microscope and the fluorescence found to be uniform across the entire area (indicative of the complete formation of a supported lipid bilayer). Supported lipid bilayers examined with ATR-FT-IR/FRAP where a bilayer had been formed on only one side of the wafer show identical spectra as Figure 3; the two-sided spectra were shown due to a better signal-to-noise ratio. As FRAP is a technique vastly more sensitive to lipid diffusion than IR spectroscopy, no correlations between the IR measurements and the FRAP data are made other than to note that the location of the CH2 peaks in Figure 3, 2852 and 2924 cm-1, correspond well with previous reports that the CH2 peaks of fluid bilayers are found at ∼2853 and ∼2923 cm-1.13 In this technical note, FRAP is used to ensure that the infrared spectrum collected is that of a lipid bilayer undergoing long-range motion and not intact vesicles. The ability to use two separate techniques on the same sample will provide a better understanding of how lipid orientation and intermolecular interactions affect diffusion, a critically important cellular process. CONCLUSIONS In this technical note, we have shown that by using silicon as the support it is possible to both acquire infrared spectra of lipid bilayers and determine the fluidity of the lipids. Using both of the techniques, ATR-FT-IR and FRAP, a more complete picture of a fully hydrated model membrane system can be obtained. Depending on the spectral region of interest, the solvent can be changed to D2O (a bending mode at 1210 cm-1 and stretching modes at 2540 and 2450 cm-1). Peaks obscured by the H2O bending and stretching modes, e.g., the HsCdCsH and CdO peaks, appear more clearly in a D2O system. The ability to obtain chemical, fluidity, and visual data on the same sample will greatly enable future studies of lipid-lipid and lipid-protein interactions. ACKNOWLEDGMENT J.S.H. is a recipient of a Career Award in the Biomedical Sciences from the Burroughs Welcome Fund. Received for review June 3, 2005. Accepted July 11, 2005. AC050990C
Analytical Chemistry, Vol. 77, No. 18, September 15, 2005
6099
