pubs.acs.org/Langmuir © 2010 American Chemical Society
Lipid Bilayer Blanketing versus Penetrating Silica Colloidal Crystals Angela R. Soemo† and Mary J. Wirth*,‡ †
Department of Chemistry & Biochemistry, University of Arizona, Tucson, Arizona 85721 and ‡Department of Chemistry, Purdue University, West Lafayette, Indiana 47907 Received October 13, 2009. Revised Manuscript Received December 22, 2009
Fluorescence intensities and diffusion coefficients were measured for a labeled lipid in 1-palmitoyl-2-oleoyl-sn-glycero3-phosphocholine (POPC) bilayers deposited by vesicle fusion on two types of silica colloidal crystals, one made from 145 nm particles and the other from 490 nm particles. The bilayer is shown to blanket the top of the silica colloidal crystal for the 145 nm case and to coat the entire surface area for the 450 nm case. The diffusion coefficient of the labeled lipid is shown to be the same in both cases, showing that the diffusion coefficient is not an indicator of bilayer penetration.
Introduction Phospolipid bilayers have been widely used as analogues to mimic the cell membrane.1 These artificial systems maintain many of the fundamental attributes of cell membranes and lend themselves to study under controlled condition and composition.2 The value of supported lipid bilayers on solid surfaces is that these are potentially useful as platforms for biotechnology applications, for which miniaturization and enhanced membrane stability is required, such as drug binding to transmembrane proteins.3,4 Dye-labeled lipids have been electrophoretically separated in supported lipid bilayers,5 and membrane proteins differentially migrate in electric fields,6 pointing to the possibility that membrane proteins might someday be separated in an environment that conserves protein function. By contrast, the challenge with transmembrane protein separations is that a significant amount of the protein extends into the aqueous medium on both sides of the membrane. Transmembrane proteins have been studied in supported bilayers on a polyacrylamide film, for which a tiny electrophoretic mobility was observed,7 but these films are difficult to control. Transmembrane proteins have been supported in free-standing bilayers over microwells in silicon supports,8 but this was not designed for electrophoresis. A solid support with contiguous pores might someday enable electrophoresis of transmembrane proteins in lipid bilayers. Silica colloidal crystals are solids with contiguous pores.9 These were first demonstrated to support lipid bilayers by using crystals *To whom correspondence should be addressed. E-mail:
[email protected]. (1) Richter, R. P.; Berat, R.; Brisson, A. R. Formation of solid-supported lipid bilayers: An integrated view. Langmuir 2006, 22, 3497-3505. (2) Forstner, M. B.; Yee, C. K.; Parikh, A. N.; Groves, J. T. Lipid lateral mobility and membrane phase structure modulation by protein binding. J. Am. Chem. Soc. 2006, 128, 15221-15227. (3) Fang, Y.; Peng, J. L.; Ferrie, A. M.; Burkhalter, R. S. Air-stable G proteincoupled receptor microarrays and ligand binding characteristics. Anal. Chem. 2006, 78, 149-155. (4) Tokimoto, T.; Bethea, T. R. C.; Zhou, M.; Ghosh, I.; Wirth, M. J. Probing Orientations of Single Fluorescent Labels on a Peptide Reversibly Binding to the Human Delta-Opioid Receptor. Appl. Spectrosc. 2007, 61, 130-137. (5) Daniel, S.; Diaz, A. J.; Martinez, K. M.; Bench, B. J.; Albertorio, F.; Cremer, P. S. Separation of membrane-bound compounds by solid-supported bilayer electrophoresis. J. Am. Chem. Soc. 2007, 129, 8072-8073. (6) Groves, J. T.; W€ulfing, C.; Boxer, S. G. Biophys. J. 1996, 71, 2716–2723. (7) Smith, E. A.; Coym, J. W.; Cowell, S. M.; Tokimoto, T.; Hruby, V. J.; Yamamura, H. I.; Wirth, M. J. Lipid bilayers on polyacrylamide brushes for inclusion of membrane proteins. Langmuir 2005, 21, 9644-9650. (8) Ganesan, P. V.; Boxer, S. G. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 5627– 5632. (9) Bertone, J. F.; Jiang, P.; Hwang, K. S.; Mittleman, D. M.; Colvin, V. L. Phys. Rev. Lett. 1999, 83, 300–303.
2196 DOI: 10.1021/la9038914
made of 330 nm particles,10 The diffusion coefficient of a dyelabeled lipid in the supported bilayer was shown to be much faster for the colloidal crystal compared to micrometer sized particles as a support, leading to the conclusion that the bilayer on the colloidal crystal was blanketing the surface. The same type of colloidal crystal, phospholipid, and vesicle size were found later to give a bilayer that significantly penetrated the interior surfaces of the colloidal crystal, as indicated by the fluorescence being more than 10 higher than that for a planar substrate.11 Penetration is advantageous in its own right for protecting the integrity of the bilayer,11 and the contiguity of penetrated bilayers in beds of micrometer sized particles has since been shown to allow fast electrophoretic transport of dye-labeled lipids.12 Live biological cells grown on silica colloidal crystals13 remain as the one example where it is still reasonable to conclude that the membrane blankets the colloidal crystal. The purpose of this Letter is to investigate whether blanketing of a silica colloidal crystal by a bilayer is possible. The approach is to use smaller nanoparticles and make measurements of both fluorescence intensity and lateral diffusion of a membrane probe.
Experimental Section Silica particles were prepared according to the St€ ober method,14 calcined at 600 °C overnight15 and then resuspended in absolute ethanol. These were deposited onto cleaned glass coverslips by vertical deposition from the meniscus, as described by Jiang et al.16 Scanning electron microscopy (SEM) (Hitachi S4800) was used to characterize the structure and thickness of the colloidal crystals. A 10 mg/mL solution of 1-palmitoyl2-oleoyl-sn-glycero-3-phosphocholine (POPC) (Avanti Polar Lipids) with 2 mol % 1-palmitoyl-2-{12-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]lauroyl}-sn-glycero-3-phosphocholine (NBD-PC) (Avanti Polar Lipids) in chloroform was extruded through 100 nm pores using the procedure of Mui et al.17 Supported phospholipid bilayers were prepared by adding the vesicle solution to a cleaned (10) Brozell, A. M.; Muha, M. A.; Sanii, B.; Parikh, A. N. J. Am. Chem. Soc. 2006, 128, 62–63. (11) Ross, E. E.; Wirth, M. J. Silica colloidal crystals as three-dimensional scaffolds for supported lipid films. Langmuir 2008, 24, 1629-1634. (12) Suzuki, K.; Hosokawa, K.; Maeda, M. J. Am. Chem. Soc. 2008, 130, 1542–1543. (13) Velarde, T. R. C.; Wirth, M. J. Appl. Spectrosc. 2008, 62, 611–616. (14) Stober, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26, 62–69. (15) Chabanov, A. A.; Jun, Y.; Norris, D. J. Appl. Phys. Lett. 2004, 84, 3573-3575. (16) Jiang, P.; Bertone, J. F.; Hwang, K. S.; Colvin, V. L. Chem. Mater. 1999, 11, 2132–2140. (17) Mui, B.; Chow, L.; Hope, M. J. Liposomes, 2003, Part A 2003, 367, 3–14.
Published on Web 01/21/2010
Langmuir 2010, 26(4), 2196–2199
Soemo and Wirth
Letter
substrate and allowing it equilibrate in the dark for 30 min, followed by gentle rinsing 20 times with vesicle-free buffer. Diffusion coefficients were determined for NBD in lipid bilayers using the method of fluorescence recovery after photobleaching (FRAP).18 A Nikon Eclipse TE 2000 microscope with a FRAP mirror, a 10 air objective, and a TRITC blue filter cube was used with a Cascade 512b CCD camera and Winview software. The light source was switched between an air-cooled argon ion laser (10 mW all lines) and a mercury lamp, and fluorescence images were captured versus time using a Cascade 521b (Roper Scientific) CCD camera and Winview software. A typical FRAP data set contained 450 frames at 1 s/frame exposure with a 100 ms (100 mW) photobleach pulse. To correct for the slight photobleaching by the mercury lamp, the fluorescence intensities were expressed as a fraction relative to the intensity of the lipid bilayer that was not photobleached by the laser. The data were fit to a biexponential recovery curve. yðtÞ ¼ A1 e -t=τ1 þ A2 e -t=τ2 Since the laser-induced bleach profile was found to be Gaussian, the time constants, τi, were related to the diffusion coefficient according to the following equation Di ¼ ω2 =4τi where D is the diffusion coefficient and ω is the experimentally determined radius of the bleach profile. For reliable error analysis, six different slides for each of the three types of coatings were used for the measurements of diffusion coefficients. For each of the six slides, no fewer than 9 different spots were chosen and the results averaged together. All six slides for each of the three types of coatings were in agreement within the intraslide error.
Results and Discussion Figure 1 shows SEM images of the silica colloidal crystals made from (A) the 490 nm and (B) the 145 nm particles. The images show that virtually all of the surface area is filled by the [111] and [100] planes of the face center cubic crystals and that there are occasional vacancy defects. Rare, larger vacancies reveal the planar substrate underneath, and the lack of a sublayer indicates that the colloidal crystal is a monolayer. These are typical of the dozens of specimens from each of the two particle sizes for these studies. NBD-doped lipid bilayers were formed by fusion of the 100 nm unilamellar vesicles onto the colloidal crystals. Most fluorescence micrographs showed uniform intensity of the NBD fluorescence. Figure 2 shows a few unusual fluorescence micrographs, which were only seen for the case of the 490 nm silica particles. The green region is twice as bright as the blue in each case. These unusual micrographs show that there are occasional regions where the fluorescence is twice as bright. The brighter regions were suspected to be regions having two layers of nanoparticles. No such regions of brighter fluorescence were observed for the colloidal crystals made of 145 nm particles after exposure to these vesicles. This is preliminary evidence that the bilayer penetrates into the colloidal crystals made of 490 nm particles to make the fluorescence twice as bright but blankets the colloidal crystals made of 145 nm particles. To investigate this intensity behavior systematically, colloidal crystals of varying thicknesses were studied. Making the colloidal crystals by vertical deposition gives a gradient of thicknesses, which enables choosing thicker regions. Fluorescence intensities (18) Axelrod, D.; Koppel, D. E.; Schlessinger, J.; Elson, E.; Webb, W. W. Biophys. J. 1976, 16, 1055–1069.
Langmuir 2010, 26(4), 2196–2199
Figure 1. Scanning electron micrographs of silica colloidal crystals on the same size scale for (A) 490 ( 10 nm particles and (B) 145 ( 3 nm particles.
were measured in regions that were marked, and then these regions were examined by SEM to determine the cross-sectional thickness. Figure 3A presents a plot of NBD fluorescence intensity versus colloidal crystal thickness for the lipid bilayers deposited on the two types of colloidal crystals. The NBD fluorescence intensity is shown to increase with thickness for the colloidal crystals made of 490 nm particles, indicating penetration of the bilayer, whereas the intensity remains constant with varying thickness for the colloidal crystals made of the 145 nm particles, indicating that the bilayer blankets the colloidal crystal. This supports the main conclusion of the paper, which is that the bilayer blankets the colloidal crystal made of the 145 nm particles and it penetrates the colloidal crystal for the case of the 490 nm particles. Figure 3B provides a drawing to illustrate bilayer blanketing versus bilayer penetration, aiding in explaining why the fluorescence intensity increases with bilayer penetration. It is possible that the pore sizes of the colloidal crystals determine whether vesicles can reach the interior of a colloidal crystal. The relation between vesicle rupture and topography has been studied for 2D surfaces19 but not for 3D materials. For facecentered cubic crystals, the pore diameter in the [100] plane is 41% of the particle size. The 100 nm vesicles could fit through the pores for the 490 nm case but not for the 145 nm case, suggesting an explanation for why there is blanketing in the 145 nm case but penetration of the bilayer for the 490 nm case. The intensity data of Figure 3A are quantitatively consistent. First, for the bilayer deposited on a monolayer of colloidal crystals made of 145 nm particles, the intensity is twice that for the case of the planar surface. This is explained by basic geometry: (19) Lee, S. W.; Na, Y. J.; Lee, S. D. Langmuir 2009, 25, 5421–5425.
DOI: 10.1021/la9038914
2197
Letter
Soemo and Wirth
Figure 2. Fluorescence micrographs showing regions of stark nonuniformity in fluorescence intensity. The fluorescence intensity jumps digitally, indicating that the thickness jumps digitally, e.g., from one to two layers of particles. Scale bar is the same for all images (100 μm).
Figure 4. Representative FRAP data. (A) Images of fluorescence recovery. (B) Fluorescence recovery curves for a lipid bilayer on bare silica (black) and on colloidal crystals made of 490 nm (blue) and 145 nm (red) particles. Figure 3. (A) Plot of fluorescence intensity versus thickness of bilayer. (B) Cartoon of lipid bilayer blanketing (left) and penetrating (right) the colloidal crystal.
the ratio of the area of the top half of the spheres to the underlying planar region is 1.8. This agrees well with the experimentally observed factor of 2.1 ( 0.3 intensity ratio for the colloidal crystal relative to the planar surface. Second, for the bilayer deposited on a monolayer made of the 490 nm particles, the intensity is twice as high as that for a monolayer of the 145 nm particles, consistent with the bilayer covering the entire sphere for the 490 case while blanketing only a hemisphere for the 145 nm case. Third, addition of a second layer of particles doubles the intensity of the 490 nm case, indicating that this second layer is also thoroughly penetrated by the bilayer. These results point to the bilayer following the surface contours of the particles of the colloidal crystal in both cases, differing only in that it blankets in one case and covers the entire surface area on the other case. Since previous work pointed to diffusion as being related to blanketing versus penetration, the lateral diffusion coefficients of 2198 DOI: 10.1021/la9038914
the NBD probe in the penetrating versus blanketing bilayers were measured by fluorescence recovery after photobleaching. A typical set of recovery curves is shown in Figure 4, comparing the planar surface and the monolayer colloidal crystals made of 490 and 145 nm particles. The recovery curves fit to double exponentials, as expected.20 One can see from the raw data of Figure 4 that the diffusion is slower for the colloidal crystals compared to that of the flat surface. Once can also see from the raw data that behaviors for the two colloidal crystals are indistinguishable. Precise agreement was obtained among five different slides for each of the three types of supports for the bilayer, as summarized in Table 1. The diffusion coefficients for the two types of colloidal crystals are thus the same, despite one having a bilayer that is penetrating and the other having a bilayer that is blanketing. These results reveal that one cannot use the diffusion coefficient to make a conclusion about whether the bilayer is blanketing or penetrating. (20) Smith, E. A.; Coym, J. W.; Cowell, S. M.; Tokimoto, T.; Hruby, V. J.; Yamamura, H. I.; Wirth, M. J. Langmuir 2005, 21, 9644-9650.
Langmuir 2010, 26(4), 2196–2199
Soemo and Wirth
Letter
Table 1. Diffusion Parameters for Labeled Lipid in POPC Lipid Bilayers on Fused Silica Slides versus Silica Colloidal Crystals Comprising the Two Different Particle Diametersa D1 (109 cm2/s)
D2 (109 cm2/s)
fraction fast component
% recovery
bare silica 4.89 ( 0.22 1.086 ( 0.43 0.782 ( 0.019 0.98 ( 0.01 490 nm 2.49 ( 0.36 0.509 ( 0.52 0.624 ( 0.033 0.94 ( 0.02 particles 145 nm 2.26 ( 0.20 0.537 ( 0.42 0.573 ( 0.033 0.95 ( 0.01 particles a The reported numbers are the interslide averages for six different slides of each sample type. The error reported is the 95% confidence interval.
It may seem at first surprising that the diffusion coefficient would be the same for blanketing versus penetrating bilayers, but this can be explained by a simple geometric idea: the lateral contour lengths are the same whether the membrane probe diffuses over a sphere versus a hemisphere. This can be visualized with the help of Figure 3B. In the case of a penetrated colloidal
Langmuir 2010, 26(4), 2196–2199
crystal, when a lipid probe reaches the intersection of two particles, it makes no difference whether it chooses to proceed above or below the adjacent sphere; the distance that it must diffuse to make lateral progress is the same. The same argument can be applied to even thicker colloidal crystals that are penetrated. Quantitatively, the diffusion coefficient is inversely proportional to the square of the contour length, which is the surface area of the hemisphere in this case, and this contour length is 1.8 times higher than that of a planar surface. This agrees with the diffusion coefficients being 2-fold lower for both colloidal crystals compared to that for the flat surface. These results are the first experimental confirmation that a bilayer can blanket the surface of a silica colloidal crystal. The results further show, with a plausible explanation, that the lateral diffusion coefficient is the same whether the bilayer is blanketing or penetrating. Acknowledgment. This work was supported by NSF under Grant CHE-0649508.
DOI: 10.1021/la9038914
2199
