Ganglioside Partitioning and Aggregation in Phase-Separated

4 May 2007 - ... Andrew C. Benniston , Graeme Copley , Anthony Harriman , and David .... Nicholas E. Dickenson , Kevin P. Armendariz , Heath A. Huckab...
0 downloads 0 Views 958KB Size
6704

Langmuir 2007, 23, 6704-6711

Ganglioside Partitioning and Aggregation in Phase-Separated Monolayers Characterized by Bodipy GM1 Monomer/Dimer Emission Oana Coban, Melanie Burger, Mike Laliberte, Anatoli Ianoul, and Linda J. Johnston* Steacie Institute for Molecular Sciences, National Research Council Canada, Ottawa, Ontario K1A 0R6, Canada ReceiVed December 6, 2006. In Final Form: February 26, 2007 The distribution of Bodipy GM1 in monolayers of binary and ternary lipid mixtures with coexisting fluid and ordered phases has been examined using a combination of atomic force microscopy and near-field scanning optical microscopy. Monolayers deposited at high (30 mN/m) and low (5 or 10 mN/m) surface pressures were examined and compared to those containing the same concentration of unlabeled ganglioside. Measurements of monomer and dimer Bodipy emission were used to distinguish aggregated from dilute ganglioside levels. For binary DPPC/DOPC monolayers, Bodipy GM1 is distributed throughout both the fluid and ordered phases at low surface pressures, and both labeled and unlabeled gangliosides result in a reduction in the size of ordered DPPC domains at 0.4% and the appearance of small aligned ganglioside-rich domains at 4%. In agreement with earlier studies, GM1 is heterogeneously distributed in small islands in the condensed DPPC domains at high surface pressure. By contrast, Bodipy GM1 causes the disappearance of large DPPC domains at 0.4% and the formation of a new GM1-rich phase at 4%. The addition of both gangliosides leads to a comparable loss of large ordered domains at low surface pressure and the appearance of a new GM1-rich phase at 30 mN/m for ternary lipid mixtures containing cholesterol. The results demonstrate the complexity of GM1 partitioning and illustrate the utility of complementary AFM and high spatial resolution two-color fluorescence experiments for understanding Bodipy GM1 aggregation and distribution.

Introduction Lipid rafts are membrane microdomains composed of sphingolipids, cholesterol, and some signaling molecules that exist in a liquid-ordered phase that is distinct from the disordered fluid phase of the bulk cellular membrane.1-4 The detection and characterization of lipid rafts has attracted much attention during the past decade, largely because they are postulated to play a significant role in aggregating/organizing proteins and in regulating signaling pathways. The difficulty in detecting small dynamic membrane domains has led to considerable controversy, and there are many unanswered questions about the physical characterization and biological significance of rafts.5,6 Ganglioside GM1 has been widely used as a marker for lipid raft microdomains in cellular membranes and in models such as vesicles and supported phospholipid monolayers and bilayers. In many cases, GM1 is visualized by fluorescence following the selective binding of cholera toxin B subunit dye conjugates or antibody labeling.7-14 In other cases, Bodipy-labeled gangliosides have * Corresponding author. E-mail: [email protected]. (1) Simons, K.; Vaz, W. L. Annu. ReV. Biophys. Biomol. Struct. 2004, 33, 269-295. (2) Simons, K.; Ikonen, E. Science 2000, 290, 1721-1726. (3) Brown, D. A.; London, E. J. Biol. Chem. 2000, 275, 17221-17224. (4) Jacobson, K.; Dietrich, C. Cell Biol. 1999, 9, 87-91. (5) Lichtenberg, D.; Goni, F. M.; Heerklotz, H. Trends Biochem. Sci. 2005, 30, 430-436. (6) Munro, S. Cell 2003, 115, 377-388. (7) Bacia, K.; Schwille, P.; Kurzchalla, T. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 3272-3277. (8) Burns, A. R. Langmuir 2003, 19, 8358-8363. (9) Burns, A. R.; Frankel, D. J.; Buranda, T. Biophys. J. 2005, 89, 1081-1093. (10) Dietrich, C.; Bagatolli, L. A.; Volovyk, Z. N.; Thompson, N. L.; Levi, M.; Jacobson, K.; Gratton, E. Biophys. J. 2001, 80, 1417-1428. (11) Dietrich, C.; Volovyk, Z. N.; Levi, M.; Thompson, N. L.; Jacobson, K. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 10642-10647. (12) Hammond, A. T.; Heberle, F. A.; Baumgart, T.; Holowka, D.; Baird, B.; Feigenson, G. W. Proc. Natl. Acad. Sci. U.S.A. 2005, 101, 6320-6325. (13) Kahya, N.; Scherfeld, D.; Bacia, K.; Poolman, B.; Schwille, P. J. Biol. Chem. 2003, 278, 28109-28115. (14) Shaw, J. E.; Epand, R. F.; Epand, R. M.; Li, Z.; Bittman, R.; Yip, C. M. Biophys. J. 2006, 90, 2170-2178.

been used to visualize the GM1 distribution.7-9,14,15 Although the latter approach has the potential advantage of preventing protein-induced aggregation of the ganglioside, several studies have shown that the addition of a Bodipy fluorophore to the ceramide acyl chain significantly modifies the partitioning of GM1 between fluid and ordered membrane phases. For example, our earlier work demonstrated that GM1 was localized in sphingomyelin-cholesterol domains in monolayers of ternary lipid mixtures whereas Bodipy GM1 was located predominantly in the fluid DOPC phase.16 Studies of vesicles and supported bilayers showed similar changes in the partitioning of native and Bodipy-labeled GM1.7-9,14,15 This work raises concerns on the interpretation of results when Bodipy GM1 is used as a marker for lipid rafts in cellular studies.17 The use of Bodipy-labeled lipids for studies of lipid organization and trafficking has a number of advantages, including the high fluorescence quantum yield and stability of the Bodipy fluorophore and the ease of incorporation of Bodipy-labeled lipids into bilayer membranes.18 In addition, the fluorescence emission spectrum of Bodipy shows a concentration-dependent shift from green (515 nm) to red (630 nm) wavelengths, making it particularly attractive for probing the lipid density in cellular membranes.18,19 Although this concentration-dependent shift was originally assigned to excimer (excited-state dimer) emission,19 detailed photophysical studies demonstrate that the emission is due to a ground-state dimer.20 This study reported the detection (15) Samsonov, A. V.; Mihalyov, I.; Cohen, F. S. Biophys. J. 2001, 81, 14861500. (16) Burgos, P.; Yuan, C.; Viriot, M.-L.; Johnston, L. J. Langmuir 2003, 19, 8002-8009. (17) Veyrat-Durebex, C.; Pommerleau, L.; Langlois, D.; Gaudreau, P. J. Cell. Phys. 2005, 203, 335-344. (18) Marks, D. L.; Singh, R. D.; Choudhury, A.; Wheatley, C. L.; Pagano, R. E. Methods 2005, 36, 186-195. (19) Pagano, R. E.; Martin, O. C.; Kang, H. C.; Haugland, R. P. J. Cell Biol. 1991, 113, 1267-1279. (20) Bergstrom, F.; Mikhalyov, I.; Hagglof, P.; Wortmann, R.; Ny, T.; Johansson, L. B.-A. J. Am. Chem. Soc. 2002, 124, 196-204.

10.1021/la0635348 CCC: $37.00 Published 2007 by the American Chemical Society Published on Web 05/04/2007

Ganglioside Partitioning and Aggregation

of two Bodipy dimers with absorption maxima at 480 and 570 nm that are assigned to configurations with parallel and collinear transition dipoles, respectively, when Bodipy-labeled lipids are incorporated into either micelles or lipid vesicles. The dimer with the higher transition energy is nonfluorescent whereas the lower-energy species absorbing at 570 nm gives rise to the concentration-dependent red emission at 630 nm. The ratio of Bodipy monomer/dimer emission has been used extensively to examine the relative lipid concentrations in cellular compartments.18,19,21 Studies of a Bodipy-labeled phosphatidylcholine in lipid monolayers indicated that Bodipy does not promote aggregation, which is an important prerequisite for using monomer/dimer emission to study clustering.22 Other work in which Bodipy-tagged lipids were used to monitor caveolar-related endocytosis for several glycosphingolipids showed that varying either the carbohydrate headgroup or the sphingosine backbone chain length did not alter the pathway for lipid internalization.21 The dual-wavelength emission properties of Bodipy GM1 have also been used to provide evidence for ganglioside clustering in vesicles23,24 but have not yet been exploited to examine changes in aggregation as a function of partitioning between ordered and disordered domains in phase-separated membranes. Herein we report a detailed study of the distribution of GM1 and Bodipy GM1 in fluid versus condensed phases of supported monolayers using a combination of atomic force microscopy (AFM) and near-field scanning optical microscopy (NSOM). This work takes advantage of the high-resolution capabilities of AFM for probing small nanodomains such as GM1 clusters.25-27 In addition, the use of two-color NSOM16,28,29 to monitor Bodipy monomer and dimer emission provides a more complete understanding of the partitioning and aggregation properties of Bodipy GM1. The ability to obtain fluorescence images with ∼50 nm resolution is a significant advantage over other hybrid optical/scanning probe techniques such as AFM-confocal or AFM-total internal reflection fluorescence microscopy, which have diffraction-limited spatial resolution.8,9,30,31 Our results demonstrate that small amounts of both native and labeled gangliosides cause dramatic changes in the domain morphology and size for some mixtures and indicate that the distribution and aggregation behavior of Bodipy GM1 in phase-separated mixtures is more complex than previously reported. Materials and Methods Materials. 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), cholesterol (Chol), Bodipy HPC, and ganglioside GM1 (bovine brain) were from Avanti Polar Lipids (Alabaster, AL). Bodipy Fl C5-ganglioside GM1 was purchased from Invitrogen Canada (Burlington, ON, (21) Singh, R. D.; Puri, V.; Valiyaveettil, J. T.; Marks, D. L.; Bittman, R.; Pagano, R. E. Mol. Biol. Cell 2003, 14, 3254-3265. (22) Dahim, M.; Mizuno, N. K.; Li, X.-M.; Momsen, W. E.; Momsen, M. M.; Brockman, H. L. Biophys. J. 2002, 83, 1511-1524. (23) Kakio, A.; Nishimoto, A.; Yanagisawa, K.; Kozutsumi, Y.; Matsuzaki, K. J. Biol. Chem. 2001, 276, 24985-24990. (24) Kakio, A.; Nishimoto, S.; Yanagisawa, K.; Matsuzaki, K. Biochemistry 2002, 41, 7385-7390. (25) Yuan, C.; Johnston, L. J. Biophys. J. 2001, 81, 1059-1069. (26) Yuan, C.; Furlong, J.; Burgos, P.; Johnston, L. J. Biophys. J. 2002, 82, 2526-2535. (27) Weerachatyanukul, W.; Ira Kongmanas, K.; Tanphaichitr, N.; Johnston, L. J. Biochim. Biophys. Acta 2007, 1768, 299-310. (28) Murray, J.; Cuccia, L.; Ianoul, A.; Cheetham, J. C.; Johnston, L. J. ChemBioChem 2004, 5, 1-6. (29) Ianoul, A.; Burgos, P.; Lu, Z.; Taylor, R. S.; Johnston, L. J. Langmuir 2003, 19, 9246-9254. (30) Shaw, J. E.; Slade, A.; Yip, C. M. J. Am. Chem. Soc. 2003, 125, 1183811839. (31) Shaw, J. E.; Oreopoulos, J.; Wong, D.; Hsu, J. C. Y.; Yip, C. M. Surf. Interface Anal. 2006, 38, 1459-1471.

Langmuir, Vol. 23, No. 12, 2007 6705

Figure 1. AFM images of 1:1 DOPC/DPPC monolayers deposited at 10 (A) and 30 mN/m (B). The z scale for both images is 10 nm. The arrow represents the direction of monolayer transfer. Canada). Analytical-grade chloroform and methanol were purchased from ACP Chemicals (QC, Canada). Monolayer Preparation. Monolayers were prepared under nitrogen using a Langmuir-Blodgett trough (611 NIMA, Coventry, U.K.) placed in a glove box (830-Compact Glove Box, Plas Labs, Lansing, MI) to avoid the oxidation of unsaturated lipids at the air-water interface. Milli-Q water was used as the subphase. Sample solutions of lipids (ratios in mol %) freshly dissolved in chloroform or chloroform/methanol were spread on the water surface. After solvent evaporation (2-5 min), monolayers were annealed by at least two compression/expansion cycles at 150 cm2/min and once again at 75 cm2/min. After compression to the required pressure, monolayers were transferred to freshly cleaved mica by vertical deposition at a dipping speed of 2 mm/min. The surface pressure was measured to a precision of 0.1 mN/m using a Wilhelmy balance. Arrows indicating the direction of film transfer are shown in Figures 1 and 2 and are the same for all other images. For comparison, monolayers were also transferred to silicon (which has a thin layer of silicon dioxide on the surface) and to mica from a subphase containing NaCl. Scanning Probe Microscopy. AFM images were acquired in contact mode in air using a Multimode Nanoscope III (Digital Instruments, Santa Barbara, CA) equipped with a J piezoscanner (maximum scan size 120 × 120 µm2). Typically, regions ranging from 100 × 100 to 5 × 5 µm2 were imaged at a resolution of 512 pixels × 512 pixels using a scan rate of 1 Hz. For smaller scan sizes, we used an E scanning head (maximum scan size 16 × 16 µm2). V-shaped Si3N4 cantilevers (200 µm length, normal spring constant 0.12 N/m) with pyramidal tips having nominal curvature radii of ∼10 nm were used. The lateral size of monolayer domains or ganglioside clusters was determined using the Digital Instruments Nanoscope software (version 5.12b49) on flattened topography images. Differences in height were measured using the cross-section analysis routine. Fluorescence imaging was performed on an AFM/NSOM microscope consisting of a Digital Instruments Bioscope mounted on an inverted microscope (Axiovert 100 TV, Zeiss). The sample is scanned using a piezoelectric, 2D scanning stage with feedback control (E-50100, Physick Instrumente). All images were acquired using an air objective (Ldplan-Neofluar, 63×, Korr, NA ) 0.75, Zeiss). All experiments used bent double-etched optical fiber probes with aperture sizes of 70-100 nm and an estimated spring constant of 100 N/m. More details on the probe fabrication and characterization can be found elsewhere.32 Bodipy GM1 fluorescence was excited at 488 nm by an Ar+-Kr+ laser (Coherent, Innova 70 Spectrum). The monomer and dimer emissions at the same excitation power were sequentially detected by an avalanche photodiode (EG&G, SPCM-AQR-14, Perkin-Elmer, Vaudreuil, Canada) with a 488 nm Notch filter to remove residual excitation (Kaiser Optical Systems) and a band-pass filter (535AF26 for monomer and 590DF35 for dimer emission, Omega Optical). NSOM images were collected in tapping mode in air at a resolution of 512 pixels × 512 pixels and a typical line scan rate of 0.25 Hz. The signal in the dimer channel (32) Burgos, P.; Lu, Z.; Ianoul, A.; Hnatovsky, C.; Viriot, M.-L.; Johnston, L. J.; Taylor, R. S. J. Microsc. 2003, 211, 37-47.

6706 Langmuir, Vol. 23, No. 12, 2007

Coban et al.

Figure 2. AFM images of 1:1 DOPC/DPPC monolayers containing 0.4% (A) and 4% (D, G) GM1 and 0.4% (B) and 4% (E, H) Bodipy GM1 deposited at 10 mN/m. The z scale is 10 nm (A, B, and E), 20 nm (D and G), and 5 nm (H). NSOM images for monolayers with 0.4% (C) and 4% (F) Bodipy GM1 are displayed with monomer emission in green and dimer emission in red. The cross sections for the lines shown in images A, B, G, and H are plotted in I-L, respectively. All cross sections are plotted on a 10 nm height scale. The arrow represents the direction of monolayer transfer. due to residual monomer emission was ∼5% of the intensity measured for the monomer channel, which is based on NSOM imaging of a control monolayer containing 4% Bodipy HPC in 1:1 DOPC/DPPC. Pseudocoloring, image overlay, and final contrast adjustments were made using public domain Image J software (available at http:// rsb.info.nih.gov/ij).

Results DOPC/DPPC Monolayers. The distributions of GM1 and Bodipy GM1 have been studied in 1:1 DOPC/DPPC monolayers transferred to mica at two different pressures, 10 and 30 mN/m. The higher pressure is a reasonable estimate for the surface pressure in biological membranes,33,34 and the lower pressure has been selected for comparison with our earlier studies as well as literature data. Figure 1 shows representative AFM images of monolayers in the absence of ganglioside at the two pressures. At low surface pressure (Figure 1A), the monolayers have micrometer size domains (2 to 3 µm) that are ∼1 nm higher than the surrounding fluid phase and cover 12% of the surface. Raising (33) Feng, S.-S. Langmuir 1999, 15, 998-1010. (34) Nagle, J. F. J. Membr. Biol. 1976, 27, 233-250.

the pressure to 30 mN/m gives a mixture of small islands (∼200 nm) that are aligned in a striped pattern oriented perpendicular to the dipping direction and a few larger raised domains that have approximate diameters of 3 to 6 µm (Figure 1B). The higher phase accounts for ∼55% of the surface area. Pure DPPC monolayers have coexisting fluid and condensed phases at 10 mN/m,35 and the islands observed for the DOPC/DPPC mixture are assigned to condensed DPPC domains surrounded by a fluid DOPC/DPPC phase.36 At higher surface pressures, the raised islands and domains correspond to a condensed DPPC phase surrounded by a fluid phase that is predominantly (if not completely) DOPC.37,38 Monolayers of 1:1 DOPC/DPPC containing 0.4% GM1 deposited at 10 mN/m have very similar morphologies to those in the absence of ganglioside, as shown by the AFM image in Figure 2A. The domains are comparable in size (1 to 2 µm) and (35) Yuan, C.; Johnston, L. J. Biophys. J. 2000, 79, 2768-2781. (36) Coban, O.; Popov, J.; Burger, M.; Vobornik, D.; Johnston, L. J. Biophys. J. 2007, 92, 2842-2853. (37) Vie, V.; Mau, N. V.; Lesniewska, E.; Goudonnet, J. P.; Heitz F.; Le Grimellec, C. Langmuir 1998, 14, 4574-4583. (38) Yokoyama, S.; Ohta, Y.; Sakai, H.; Abe, M. Colloids Surf., B 2004, 34, 65-68.

Ganglioside Partitioning and Aggregation

Langmuir, Vol. 23, No. 12, 2007 6707

Figure 3. AFM images of 1:1 DOPC/DPPC monolayers containing 0.4% and 4% native GM1 (A, D, G) and Bodipy-labeled GM1 (B, E, H), respectively, deposited at 30 mN/m. The z scale is 10 nm for all images. NSOM images of the monolayers containing 0.4 and 4% fluorescently labeled GM1 are shown in Figure 3C and F. The cross-sections for the lines shown in images A, D, G, and H are plotted in I-L, respectively. All cross sections are plotted on a 10 nm height scale.

height (0.9 nm) to those in Figure 1A but account for a lower fraction of the surface area (9%), consistent with a higher fraction of DPPC in the fluid phase. There are also small raised islands in the center of most domains (∼300 nm in diameter, 1 to 1.2 nm above the domain) that we attribute to GM1-enriched areas, by analogy to literature results.35,37,38 Monolayers prepared with the same percentage of Bodipy GM1 are similar, except that the small, higher islands appear only in some domains and are much taller (4 nm, Figure 2B). In this case, NSOM provides a clearer picture of the ganglioside distribution. Figure 2C shows a fluorescence image that is an overlay of monomer (green) and dimer (red) channels, scanned sequentially over the same sample area. The monomer emission is visible as a slightly heterogeneous fluorescent matrix surrounding a number of dark domains that are approximately the same size as those observed by AFM (1-3 µm). The dimer emission shows predominantly bright fluorescent areas in the center of some of the small domains, which correlate with the raised dots in the AFM images. These results indicate that Bodipy GM1 is localized predominantly in the fluid phase, with more tightly packed small aggregates that give rise to dimer emission in the center of many of the condensed domains. AFM detects only aggregated GM1 and provides no information on monomers. A

control NSOM experiment for a DOPC/DPPC monolayer containing 4% Bodipy HPC showed uniform monomer emission in the fluid phase. The absence of dimeric Bodipy emission indicates that aggregation is promoted by the ganglioside rather than the Bodipy probe. The addition of a larger amount of GM1 or Bodipy GM1 to monolayers at 10 mN/m leads to more dramatic changes in morphology, with large-scale images for samples with 4% ganglioside showing branched stripes of a higher phase (Figure 2D,E) in an otherwise homogeneous monolayer. Note that in this case the stripes are aligned parallel to the transfer direction. Small-scale images for GM1 monolayers (Figure 2G) indicate that the stripes are formed by the alignment of many small domains (