pubs.acs.org/Langmuir © 2009 American Chemical Society
Direct Correlation of Structures and Nanomechanical Properties of Multicomponent Lipid Bilayers Ruby May A. Sullan,†,‡ James K. Li,‡ and Shan Zou†,* †
Steacie Institute for Molecular Sciences, National Research Council Canada, 100 Sussex Drive, Ottawa, ON K1A 0R6, Canada and ‡Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, ON M5S 3H6, Canada Received January 31, 2009. Revised Manuscript Received February 17, 2009 Exploring the fine structures and physicochemical properties of physiologically relevant membranes is crucial to understanding biological membrane functions including membrane mechanical stability. We report a direct correlation of the self-organized structures exhibited in phase-segregated supported lipid bilayers consisting of dioleoylphosphatidylcholine/egg sphingomyelin/cholesterol (DEC) in the absence and presence of ceramide (DEC-Ceramide) with their nanomechanical properties using AFM imaging and high-resolution force mapping. Direct incorporation of ceramide into phase-segregated supported lipid bilayers formed ceramideenriched domains, where the height topography was found to be imaging setpoint dependent. In contrast, liquid ordered domains in both DEC and DEC-Ceramide presented similar heights regardless of AFM imaging settings. Owing to its capability for simultaneous determination of the topology and interaction forces, AFMbased force mapping was used in our study to directly correlate the structures and mechanical responses of different coexisting phases. The intrinsic breakthrough forces, regarded as fingerprints of bilayer stability, along with elastic moduli, adhesion forces, and indentation of the different phases in the bilayers were systematically determined on the nanometer scale, and the results were presented as two-dimensional visual maps using a selfdeveloped code for force curves batch analysis. The mechanical stability and compactness were increased in both liquid ordered domains and fluid disordered phases of DEC-Ceramide, attributed to the influence of ceramide in the organization of the bilayer, as well as to the displacement of cholesterol as a result of the generation of ceramide-enriched domains. The use of AFM force mapping in studying phase segregation of multicomponent lipid membrane systems is a valuable complement to other biophysical techniques such as imaging and spectroscopy, as it provides unprecedented insight into lipid membrane mechanical properties and functions.
Introduction Membrane microdomains enriched in cholesterol and glycosphingolipids (sphingomyelin in particular) have been proposed as lateral structural components of the plasma membrane referred to as floating lipid rafts, onto which specific proteins attach within the bilayer.1 These functional platforms have been implicated in a number of important cell functions such as protein sorting, signal transduction, transcytosis, and potocytosis, as well as HIV-1 assembly and release.2-4 In exploring this hypothesis, lipid mixtures consisting of phosphatidylcholine (PC), sphingomyelin (SM), and cholesterol (Chol), have often been used to mimic the rafts in cells, and it is wellestablished that these mixtures give rise to coexisting liquid ordered phase or domains (SM/Chol-enriched) and fluid disordered phase (PC-enriched) in the bilayer.5-8 *Corresponding author addresses: Biomolecular Sensing and Imaging Group, Steacie Institute for Molecular Sciences, National Research Council Canada, 100 Sussex Drive, Rm 1095, Ottawa, ON K1A 0R6, Canada. Phone: 613 949-9675, Fax: 613 991-4278, E-mail shan.
[email protected]. (1) Simons, K.; Ikonen, E. Nature (London) 1997, 387, 569–572. (2) Brown, D. A.; London, E. Annu. Rev. Cell Dev. Biol. 1998, 14, 111–136. (3) Siskind, L. J.; Colombini, M. J. Biol. Chem. 2000, 275, 38640–38644. (4) Yeagle, P. The Structure of Biological Membranes, 2nd ed.; CRC Press LLC: Washington DC, 2005. (5) Dietrich, C.; Bagatolli, L. A.; Volovyk, Z. N.; Thompson, N. L.; Levi, M.; Jacobson, K.; Gratton, E. Biophys. J. 2001, 80, 1417–1428. (6) Milhiet, P. E.; Giocondi, M. C.; Baghdadi, O.; Ronzon, F.; Roux, B.; Le Grimellec, C. EMBO Rep. 2002, 3, 485–490. (7) Rinia, H. A.; de Kruijff, B. FEBS Lett. 2001, 504, 194–199. (8) Saslowsky, D. E.; Lawrence, J.; Ren, X. Y.; Brown, D. A.; Henderson, R. M.; Edwardson, J. M. J. Biol. Chem. 2002, 277, 26966–26970.
Langmuir 2009, 25(13), 7471–7477
Ceramide, a sphingolipid that is highly hydrophobic in nature and capable of forming extensive H-bond networks, is regarded as an important signaling molecule involved in a wide range of cellular processes.9,10 Several recent investigations have determined the effects of ceramide incorporation on coexisting fluid phase and ordered domains in phaseseparated binary and ternary lipid mixtures.11-17 Ceramide has been shown to increase membrane heterogeneity, and to displace cholesterol from rafts, in both model and cellular membranes.11,18-22 The generation of ceramide-enriched (9) Bollinger, C. R. Biochim. Biophys. Acta, Mol. Cell. Res. 2005, 1746, 284–294. (10) Hannun, Y. A. Science 1996, 274, 1855–1859. (11) Chiantia, S.; Kahya, N.; Ries, J.; Schwille, P. Biophys. J. 2006, 90, 4500–4508. (12) Chiantia, S.; Ries, J.; Chwastek, G.; Carrer, D.; Li, Z.; Bittman, R.; Schwille, P. Biochim. Biophys. Acta, Biomembr. 2008, 1778, 1356–1364. (13) Ira; Johnston, L. J. Langmuir 2006, 22, 11284–11289. (14) Ira; Zou, S.; Carter, D.; Vanderlip, S.; Johnston, L. J. Struct. Biol. 2008, submitted. (15) Lopez-Montero, I.; Velez, M.; Devaux, P. F. Biochim. Biophys. Acta, Biomembr. 2007, 1768, 553–561. (16) Silva, L. C.; de Almeida, R. F. M.; Castro, B. M.; Fedorov, A.; Prieto, M. Biophys. J. 2007, 92, 502–516. (17) Trajkovic, K.; Hsu, C.; Chiantia, S.; Rajendran, L.; Wenzel, D.; Wieland, F.; Schwille, P.; Brugger, B.; Simons, M. Science 2008, 319, 1244–1247. (18) Goni, F. M.; Alonso, A. Biochim. Biophys. Acta, Biomembr. 2006, 1758, 1902–1921. (19) Ali, M. R.; Cheng, K. H.; Huang, J. Biochemistry 2006, 45, 12629–12638. (20) Megha; Bakht, O.; London, E. J. Biol. Chem. 2006, 281, 21903–21913. (21) Megha; London, E. J. Biol. Chem. 2004, 279, 9997–10004. (22) Yu, C.; Alterman, M.; Dobrowsky, R. T. J. Lipid Res. 2005, 46, 1678–1691.
Published on Web 03/18/2009
DOI: 10.1021/la900395w
7471
Article
domains from the enzymatic hydrolysis of the SM phosphocholine headgroups has also been reported.11,14 Atomic force microscope (AFM) imaging, with spatial resolution attainable in the nanometer range, has been successfully applied to study the self-organized structures of raft models, e.g., PC/SM/Chol bilayers.5-8 Its capability in distinguishing domains collected in an image primarily relies on different thicknesses of coexisting phases in the bilayer. Hence, while high spatial resolution images of the different phases can be obtained, caution must be exercised in their interpretation, as the topography generated can be a strong function of imaging parameters. This is especially true for multicomponent lipid mixtures like PC/SM/Chol with ceramide. In such cases, AFM force spectroscopy becomes a valuable tool since it is capable of circumventing the ambiguities caused by variation in imaging setpoints. It should be emphasized that force spectroscopy is the only technique that can investigate the mechanics of membranes over an area with nanometer dimensions.23 AFM-based force spectroscopy has been used to probe different interaction forces in single component and binary mixtures of synthetic phospholipids, including hydration, steric, and Derjaguin-Landau-VerweyOverbeek (DLVO) interactions, as well as the elastic/plastic behavior of lipid bilayers.24-29 The effects of ionic strength on the adhesion between chemically modified AFM tips and lipid bilayers have also been reported.30,31 Force mapping (also known as force volume), a specialized imaging technique wherein the morphology and interaction forces are measured simultaneously, has been chosen in the current study. It has been available for some time with applications in early experiments to synaptic vesicles and polymer surfaces.32-35 Previous efforts involved reconstruction of a sequence of force curves to create two-dimensional (2D) adhesion maps for bilayer domains.24,25,36 Although these studies have provided valuable information on the physical properties of binary lipid mixtures, nanomechanical properties of multicomponent lipid mixtures with biological relevance such as PC/SM/Chol with ceramide are scarce. This scarcity of data probably arises from the lengthy analysis of an enormous number of force-separation curves collected in high-resolution force mapping.23 Because of this, relevant issues such as the mechanical stability of different phases in phase-segregated multicomponent bilayers and correlation of (23) Seantier, B.; Giocondi, M. C.; Le Grimellec, C.; Milhiet, P. E. Curr. Opin. Colloid Interface Sci. 2008, 13, 326–337. (24) Dufrene, Y. F.; Barger, W. R.; Green, J. B. D.; Lee, G. U. Langmuir 1997, 13, 4779–4784. (25) Dufrene, Y. F.; Boland, T.; Schneider, J. W.; Barger, W. R.; Lee, G. U. Faraday Discuss. 1998, 111, 79–94. (26) Franz, V.; Loi, S.; Muller, H.; Bamberg, E.; Butt, H. H. Colloids Surf., B 2002, 23, 191–200. (27) Schneider, J.; Barger, W.; Lee, G. U. Langmuir 2003, 19, 1899–1907. (28) Schneider, J.; Dufrene, Y. F.; Barger, W. R.; Lee, G. U. Biophys. J. 2000, 79, 1107–1118. (29) Weisenhorn, A. L.; Maivald, P.; Butt, H. J.; Hansma, P. K. Phys. Rev. B: Condens. Matter 1992, 45, 11226–11232. (30) Garcia-Manyes, S. Biophys. J. 2005, 89, 4261–4274. (31) Garcia-Manyes, S.; Oncins, G.; Sanz, F. Electrochim. Acta 2006, 51, 5029–5036. (32) Eaton, P.; Smith, J. R.; Graham, P.; Smart, J. D.; Nevell, T. G.; Tsibouklis, J. Langmuir 2002, 18, 3387–3389. (33) Laney, D. E.; Garcia, R. A.; Parsons, S. M.; Hansma, H. G. Biophys. J. 1997, 72. (34) Mizes, H. A.; Loh, K. G.; Miller, R. J. D.; Ahuja, S. K.; Grabowski, E. F. Appl. Phys. Lett. 1991, 59, 2901–2903. (35) Song, J.; Duval, J. F. L.; Stuart, M. A. C.; Hillborg, H.; Gunst, U.; Arlinghaus, H. F.; Vancso, G. J. Langmuir 2007, 23, 5430–5438. (36) Kruger, S.; Kruger, D.; Janshoff, A. ChemPhysChem 2004, 5, 989–997.
7472 DOI: 10.1021/la900395w
Sullan et al.
composition and structure with membrane mechanical properties have not been adequately addressed yet. In the present study, we developed a batch analysis algorithm to rapidly process the large number of force curves captured in high-resolution force mapping measurements, i.e., 128 64 pixels versus a previously achieved maximum of 32 32 pixels. In addition to conventional adhesion data, mechanical properties such as breakthrough force and elastic modulus of different phases in the supported lipid bilayers consisting of dioleoylphosphatidylcholine/egg sphingomyelin/ cholesterol (DEC) in the absence and presence of ceramide (DEC-Ceramide) bilayers were extracted automatically. Twodimensional visual maps reconstructed using this analysis code enable us to correlate the structures of different phases to their nanomechanical properties. The mechanical responses of different phases in DEC and DEC-Ceramide bilayers were systematically determined using AFM topography imaging and force mapping. The subtle contribution of ceramide and cholesterol to the mechanical stability of the bilayer provides valuable insight into how these lipid molecules are further involved in membrane functions. Materials and Methods Materials. All lipids: 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), egg sphingomyelin (ESM), ovine wool cholesterol (Chol), N-palmitoyl-D-erythro-sphingosine (ceramide), and 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) were purchased from Avanti Polar Lipids (Alabaster, AL) and used as received. HPLC-grade chloroform from ACP Chemicals Inc. (Montreal, QC), ACS grade methanol from Fisher Scientific (Ottawa, ON), and Milli-Q water deionized to a resistivity of 18 MΩ 3 cm-1 were used in all of the experiments. Preparation of Small Unilamellar Vesicles. Lipid mixtures were obtained by combining the appropriate molar ratios of the different lipid components using chloroform and methanol as solvents: DOPC/ESM/Chol in a 2:2:1 molar ratio, referred to as DEC, and DOPC/ESM/Ceramide/Chol in a 4:3:1:2 molar ratio (25% of the ESM in DEC was replaced with ceramide resulting in 10 mol % ceramide of the total lipid mixture) as DEC-Ceramide. The resulting solution was then exposed to a gentle stream of nitrogen and placed under vacuum overnight to further remove the solvents. The lipid film was hydrated to a final lipid concentration of 0.5 mg/mL for the DEC and 1 mg/mL for the DEC-Ceramide prior to use. Small unilamellar vesicles were obtained by sonicating the lipid solution to clarity (∼20-30 min) using a bath sonicator (Cole Parmer, Montreal, QC). Preparation of the Bilayer. Vesicle fusion protocols for both DEC and DEC-Ceramide lipid bilayers preparation were followed.11,13 Vesicle solutions of DEC containing 25 μg lipids and a final concentration of 10 mM CaCl2 were deposited on freshly cleaved mica substrates (20-30 μm thick) glued on glass coverslips affixed to a liquid cell. The sample was incubated at 45 °C for an hour, and slowly cooled to room temperature. Our imaging and force measurement results were based on more than 15 samples. Extensive washing using ∼150 mL with 18 MΩ 3 cm-1 Milli-Q H2O followed the incubation. For the DECCeramide, vesicle solutions of 50 μg lipids and a final concentration of 10 mM CaCl2 were incubated for 30 min at room temperature. Similarly, washing with excess 18 MΩ 3 cm-1 MilliQ H2O followed the incubation. About 10 DEC-Ceramide samples were prepared and examined in our experiments. AFM Imaging and Force Mapping. All AFM images were obtained using the Nanowizard II BioAFM (JPK Instruments, Berlin, Germany) mounted on an Olympus 1X81 inverted confocal microscope. Both contact and intermittent contact modes were used. Silicon nitride cantilevers (DNP-S, Veeco, CA) were used in contact mode imaging and force mapping Langmuir 2009, 25(13), 7471–7477
Sullan et al.
Article
Figure 1. AFM images and height profiles of a DEC-Ceramide bilayer on mica in contact mode using a low setpoint of F ∼ 0.5 nN (A,E) with
corresponding lateral deflection image (B), using a high setpoint of F ∼ 5.0 nN (C,F), and in the intermittent contact mode (D,G). Height profiles show that the highest features (subdomains) are 0.3 ( 0.1, 0.6 ( 0.1, and 0.2 ( 0.1 nm higher than the intermediate liquid ordered domains, which are 0.7 ( 0.1, 0.8 ( 0.1, and 1.2 ( 0.1 nm above the lowest fluid disordered phase, respectively. Regions outlined in circles in the height profiles indicate the subdomains.
measurements unless stated otherwise. The spring constant, typically in the range 0.06-0.28 N/m, was determined by the thermal noise method37 after the determination of the cantilever deflection sensitivity by indenting the AFM tip against a hard reference substrate (glass). Silicon nitride cantilevers (TR08-35, Veeco, CA) were used in the intermittent contact mode experiments with nominal spring constant of 0.57 N/m. All AFM imaging measurements were carried out on mica-onglass substrates fixed to a liquid cell, and the samples were kept hydrated at all times. All AFM images were plane-fit (1st order) using the JPK SPM Image Processing Software (JPK Instruments, Berlin, Germany). In force mapping, a set of force distance curves were collected over an area of a bilayer sample. This area is first divided into a grid pattern (e.g., 128 64 pixels), and the scanner performs a single force spectroscopy measurement at the center of every pixel, each time collecting extension and retraction forcedistance curves. The 2D visual maps were reconstructed from 128 64 or 64 64 grids, over an area up to a maximum of 3 μm 3 μm in dimension. An applied load in the range 4-12 nN was used, unless stated otherwise. With the unmodified Si3N4 tips, we can reliably distinguish between the DOPC-rich phase and the (ESM/Chol)-rich phase in topography, adhesion, and other force-based 2D maps and plots. Batch Analysis of the Force Curves. The collected force curves (approximately 4000 to 8000 curves per set) comprising the force map were batch analyzed using the self-developed algorithm implemented in IGOR Pro 6 (Wavemetrics, Portland, OR). For each curve breakthrough force, indentation, elastic moduli, and adhesion were calculated or extracted. In addition, using the (x, y) positions, force maps, contour plots, and pairwise scatter plots of these quantities were generated to ascertain any correlation of clusters of data with the spatial distribution of the phases in the lipid bilayer. The details of the analysis code are provided in the Supporting Information.
Results and Discussion SetPoint Dependent AFM Imaging of Multicomponent Lipid Mixture. Figure 1A,B shows the phase-segregated DEC-Ceramide bilayer images using contact mode at a low setpoint of F ∼ 0.5 nN. Three distinct phases were observed: features with the shortest height are ascribed to the fluid disordered (DOPC-rich) phase, the tallest height to the (37) Hutter, J. L.; Bechhoefer, J. Rev. Sci. Instrum. 1993, 64, 1868–1873.
Langmuir 2009, 25(13), 7471–7477
ceramide-enriched domains or subdomains, and the intermediate height to the liquid ordered (ESM/Chol-rich) domains or ESM rafts, consistent with previous studies.11,13 The subdomains that are 0.3 ( 0.1 nm higher than the ESM rafts were obtained when a low setpoint was applied to the DEC-Ceramide lipid bilayer (Figure 1E). However, with a larger imaging set point of F ∼ 5.0 nN in contact mode, the difference in height between the subdomains and the ESM rafts increased to 0.6 ( 0.1 nm (Figure 1F), and the ESM rafts are 0.8 ( 0.1 nm taller than the matrix of the fluid disordered phase, similar to that of the 0.7 ( 0.1 nm observed in the lower setpoint image (Figure 1A,E). Applying an even smaller force in the AFM topography imaging, the same area was reimaged using intermittent contact mode (Figure 1D). The height profiles show that the liquid ordered domains are 1.2 ( 0.1 nm above the fluid disordered phase and the subdomains are 0.2 ( 0.1 nm higher than the ESM rafts (Figure 1G). Comparison of the height profiles obtained in different AFM imaging modes shows that different heights of ceramide-enriched domains can be obtained as a consequence of varying setpoints, suggesting a sensitive mechanical response of the subdomains in DEC-Ceramide bilayers. In contrast, the liquid ordered domains in DEC-Ceramide exhibited similar heights regardless of the imaging settings. These results indicate that, in DEC-Ceramide bilayers, one cannot only rely on the AFM height image in understanding the behavior of the different phases in the lipid mixture.25,28 In addition, the topography itself cannot sufficiently provide accurate composition and structural information of the bilayer. We thus used force mapping coupled with AFM imaging to systematically characterize the different coexisting phases in multicomponent lipid bilayers, on the basis of their nanomechanical properties. Force Mapping on DEC. The major driving force that holds the amphiphilic molecules together in bilayers is not due to strong covalent or ionic bonds, but rather arises from weaker van der Waals, hydrophobic, hydrogen-bonding, and screened electrostatic interactions, and is greatly influenced by external conditions.31,38 It is therefore necessary to (38) Oncins, G.; Picas, L.; Hernandez-Borrell, J.; Garcia-Manyes, S.; Sanz, F. Biophys. J. 2007, 93, 2713–2725.
DOI: 10.1021/la900395w
7473
Article
Sullan et al.
Figure 2. AFM height image (A), corresponding height profile (B), and typical force curves for liquid ordered domains and fluid disordered phase, showing the expected height difference of ∼ 0.8 nm (C) for a DEC bilayer.
Figure 3. Breakthrough force map (A), contour map representation of the breakthrough forces (B), and the Young’s modulus map (C). first consider how an applied force and the presence of certain molecules, i.e., cholesterol and ceramide, will influence the organization of the lipid bilayer, and consequently its properties and functions. In contrast to the significant height differences of ceramide-enriched domains upon imaging settings, the ESM rafts in DEC-Ceramide were consistently ∼0.8 nm taller than the matrix of the fluid disordered phases (Figure 1A, C), which is in good agreement with reported values in pure DEC ternary mixtures.7,11,39 As an initial step in understanding how the addition of one more component (i.e., ceramide) to the ternary lipid mixture alters the mechanical response of the bilayer, force mapping was first performed on DEC bilayer without ceramide. Supporting Information Figure S1 shows a typical force-extension curve illustrating characteristic features of the force profile and corresponding measured quantities. The AFM height image of a DEC bilayer showing the coexistence of liquid ordered domains (brighter regions) and fluid disordered phase (darker matrix) is shown in Figure 2A. The ESM rafts are 0.8 nm above the DOPC-rich phase consistent with literature data (Figure 2B).7,11,39 Superimposing more than 200 curves recorded in both phases shows two distinct groups of breakthrough events. The separation distances of each group at a given force (e.g., at 1 nN) within the indentation region yielded 5.5 nm (liquid ordered domains) and 4.7 nm (fluid phases), respectively. This separation difference (height difference) of ∼0.8 nm is consistent with what has been observed in the AFM height image (Figure 2A,B). The adhesion map and the corresponding histogram of the same area are shown in Supporting Information Figure S2. The lighter regions represent areas with high adhesion and are ascribed to the liquid ordered domains, while the darker regions are areas with low adhesion and correspond to fluid disordered phase. The slight difference in the domain sizes is due to the different pixel densities between an AFM topographical image (512 512) and the map constructed from force-separation curves (128 64). In general, adhesion force is related to the contact area of the tip-lipid interaction and depends on external forces and (39) Ira; Johnston, L. Biochim. Biophys. Acta, Biomembr. 2008, 1778, 185–197.
7474 DOI: 10.1021/la900395w
other environmental conditions such as temperature, humidity, surface roughness, contaminant, and total interfacial contact time.40 Hence, it is not an intrinsic property of a bilayer and should not be used to discriminate mechanical properties of the different phases. Breakthrough forces and Young’s moduli, in contrast to adhesion, are two mechanical quantities that can be used to reliably characterize the properties of each phase in a natural multicomponent lipid mixture. A breakthrough force is the maximum force that the bilayer is able to withstand before rupture.27,41 It has been shown to be an intrinsic property of a bilayer and can be regarded as a fingerprint of bilayer stability. The Young’s modulus, on the other hand, is a measure of the stiffness of a material42,43 and provides information on the strength of cohesive forces among the lipid components within a bilayer. Values of the Young’s modulus in this work were obtained by fitting the indentation region of the force curve (see Supporting Information Figure S1) to the Sneddon model for a semi-infinite sample in mechanical contact with a paraboloidal-shaped tip44 (Supporting Information eq S2). In order to quantitatively compare the mechanical properties of different phases in the DEC bilayer, analysis of 8192 force-separation curves (128 64 array of force curves) using the self-developed code yielded the high-resolution 2D visual maps of breakthrough forces and elastic moduli, as shown in Figure 3. As shown in these visual maps, liquid ordered domains exhibit higher breakthrough forces, higher Young’s moduli, and higher adhesion than fluid disordered phases. In both phases, heterogeneity can be resolved at the nanometer scale and is more apparent in the contour map representation (Figure 3B). Here, the boundary between the coexisting phases is clearly defined, and the interface has breakthrough forces intermediate to that of the two phases. Comparison of the AFM height images of a DEC bilayer before (Supporting Information Figure S3A) and (40) Israelachvili, J. Electrostatic Forces Between Surfaces in Liquid. Intermolecular and Surface Forces; Academic Press: London, 2002. (41) Kunneke, S.; Kruger, D.; Janshoff, A. Biophys. J. 2004, 86, 1545–1553. (42) Guo, S.; Akhremitchev, B. B. Langmuir 2008, 24, 880–887. (43) Akhremitchev, B. B.; Walker, G. C. Langmuir 1999, 15, 5630–5634. (44) Sneddon, I. Int. J. Eng. Sci. 1965, 3, 47–57.
Langmuir 2009, 25(13), 7471–7477
Sullan et al.
Article
Figure 4. Adhesion force map (A), AFM height image after the force mapping (B), the corresponding lateral deflection image (C), breakthrough force map (D), contour representation of the breakthrough forces (E), and Young’s modulus map (F) of a DEC-Ceramide bilayer.
Figure 5. Histograms of the breakthrough forces of a DEC (A) and of a DEC-Ceramide (B) bilayer. Solid bars correspond to the liquid ordered domains, while hollow bars correspond to the fluid disordered phase. after (Supporting Information Figure S3D) force mapping shows that the liquid ordered domains retained their general shapes and relative positions from each other, indicating that no significant restructuring of the bilayer is being induced by force mapping. The breakthrough force map (Supporting Information Figure S3B) and adhesion map (Supporting Information Figure S3C) are included to show the state of the bilayer while force mapping is being performed. Force Mapping on DEC-Ceramide. Similar to DEC, highresolution force mapping measurements were carried out on the DEC-Ceramide bilayers. Figure 4A shows the adhesion force map of a DEC bilayer with ceramide. The low adhesion regions (black areas) in this map correspond to locations of ceramide-enriched domains, as can be inferred from both AFM height (Figure 4B) and lateral deflection (Figure 4C) images collected right after force mapping. A typical breakthrough force map, its contour plot, and the Young’s modulus map, constructed from a total of 4096 force-separation curves (64 64 array of force curves) were presented in Figure 4D,E,F, respectively. High breakthrough forces at the boundary and within the vicinity of the subdomains were observed. Similar to DEC, heterogeneities in the phases are visible in the force maps and contours. Curves with both breakthrough and adhesion events absent (termed hereafter as no profile curve) were consistently observed within the subdomains. As will be discussed in greater detail in a subsequent section, the lack of breakthrough events in the subdomains indicates packing behavior that is different from typical ESM rafts. Comparison of Breakthrough Forces: DEC vs DECCeramide. The histograms of breakthrough forces for DEC and DEC-Ceramide bilayers were constructed, respectively, as shown in Figure 5. For the DEC bilayer, a bimodal distribution with peaks at F ∼ 1.4 nN, ascribed to the fluid disordered phase, and at F ∼ 3.2 nN, corresponding to the Langmuir 2009, 25(13), 7471–7477
liquid ordered domains, was obtained. Shifts to much higher breakthrough forces were observed in the DEC-Ceramide bilayer. The histogram of the breakthrough forces in the rafts (solid bars in Figure 5B) in DEC-Ceramide shows a peak at F ∼ 5 nN, with the fluid disordered phase (hollow bars in Figure 5B) at F ∼ 4.1 nN. Breakthrough forces greater than 5 nN are from the regions within the vicinity of the ceramideenriched domains (see Figure 4E). It should be noted that the histogram of breakthrough forces for DEC-Ceramide only accounts for the liquid ordered domains and fluid disordered phases, as the subdomains exhibited no breakthrough events. The high-resolution visual maps together with the histogram distribution of the intrinsic breakthrough forces generated from our batch analysis code could directly correlate the fine structures of multicomponent lipid mixtures with their mechanical stability. Comparison of Young’s Modulus: DEC vs DEC-Ceramide. Fitting the indentation region in each of the extension curve (see Figure 2C and Supporting Information Figure S1) with the Sneddon model yielded the Young’s modulus values, and the distribution of these values of the individual phases were grouped and plotted in Figure 6. The peak at ∼80 MPa corresponds to the fluid disordered phase (hollow bars) and the peak at ∼140 MPa to the liquid ordered domains (solid bars), indicating less elastic deformation or a higher degree of compactness in the latter. These values are in close agreement with what has been obtained using electrocompression experiments.45,46 The two phases have overlapping elastic moduli in the 50-130 MPa range, and the distribution is broader in liquid ordered domains. Both observations (45) Cevc, G.; Marsh, D. Phospholipid Bilayers: Physical Principles and Models; John Wiley and Sons: New York, 1987. (46) Marsh, D. Handbook of Lipid Bilayers; CRC Press: Boca Raton, FL, 1987.
DOI: 10.1021/la900395w
7475
Article
Sullan et al.
Figure 6. Histograms of the elastic modulus of the individual phases in a DEC bilayer (A) and a DEC-Ceramide bilayer (B). The hollow bars correspond to the fluid disordered phase, while the solid bars correspond to the liquid ordered domains. point to the heterogeneity within the liquid ordered domains, which is not evident in the height topography images (Figures 2A and 4B). The histogram of Young’s modulus values measured on a DEC-Ceramide shows a peak at ∼185 MPa for the liquid ordered domains and one at ∼175 MPa for the fluid disordered phase, with greater overlap between the two phases. This indicates that heterogeneity and compactness are increased in both phases. This result provides direct evidence that it is the high nanomechanical stability and compactness of the rafts that accommodates proteins and other signaling molecules. Ceramide and Cholesterol Increase the Mechanical Stability in the DEC-Ceramide Bilayer. Breakthrough forces increased from 1.4 nN (in DOPC-rich phase, DEC) to 4.1 nN (in DOPC-rich phase, DEC-Ceramide) and from 3.2 nN (in ESM rafts, DEC) to 5.0 nN (in ESM rafts, DECCeramide). Two possible scenarios could have caused the significant increase in breakthrough forces in the DECCeramide bilayer: the effect of ceramide itself and the effect of displaced cholesterol as a result of ceramide-enriched domains generation. To assess the latter, force mapping was carried out on DEC111 (dioleoylphosphatidylcholine/ egg sphingomyelin/cholesterol in a 1:1:1 molar ratio), a lipid bilayer with 13% more cholesterol in composition than a DEC or DEC-Ceramide. In DEC-Ceramide, the maximum amount of cholesterol that could be expelled by the formation of ceramide subdomains is 10 mol %. The effect of 13% more cholesterol in DEC 111 on the changes of breakthrough forces (if there are any) is comparable to the effect of the 10 mol % cholesterol at the maximum (10 mol % is the total ceramide content) that could be expelled in DEC-Ceramide. The histogram of the breakthrough forces in DEC111 (Supporting Information Figure S4) has a bimodal distribution with peaks at F ∼ 1.4 nN and at F ∼ 3.4 nN. It is not surprising that the breakthrough force peak of the fluid disordered phase remains unchanged even at higher cholesterol concentration, because cholesterol preferentially packs in the ESM domains rather than in the fluid DOPC.47-49 The increase of about 200 pN in the breakthrough force of the liquid ordered domains, on the other hand, was likely due to the presence of more cholesterol in the rafts, since cholesterol favors ESM over DOPC.47-49 Furthermore, forces higher than 3.5 nN, not present in the DEC bilayer, were seen in DEC111, again implying the influence of increased cholesterol content on the mechanical stability of the bilayer. The 13% more cholesterol, which led to the slightly higher breakthrough forces of the liquid ordered domains in
DEC111, however, could not account for the significant breakthrough force increase in the liquid ordered domains (i.e., from 3.2 nN to 5.0 nN in DEC and DEC-Ceramide, respectively). This significant increase of breakthrough forces suggests that it cannot be attributed to the displacement of cholesterol by the generation of ceramide-enriched domains, but primarily to the effects of ceramide itself on the bilayer. It should be noted that the high forces (8-12 nN) were observed at the boundary and within the vicinity of the subdomains (Figure 4D), which is a strong indication that there is a high localization of ceramide in those regions. The increase in breakthrough force values by 2.7 nN in the fluid disordered phase and by 1.8 nN in the liquid ordered domains suggested that the enhanced stability of the bilayer is due to the presence of ceramide in both phases. It is surprising to us that, in the ceramide-enriched regions, no profile (no breakthrough-no adhesion) curves were observed. To ensure that this observation is not a consequence of having “no molecular contact between the probe and the lipid layer surfaces”,25 or due to some strong repulsive forces between the tip and the lipid surfaces, stiffer silicon cantilevers (k ∼ 7 N/m) were used to probe those ceramideenriched regions. Consistent with the results using the softer cantilevers (k < 0.25 N/m), no profile curves were obtained even at loading forces as high as 70 nN. This may provide additional support to an earlier study, which suggested that the ceramide-enriched region is a highly ordered and tightly packed gel phase, composed of both ceramide and ESM.11 To compare the physical state of the ceramide-enriched domains with a well-known gel phase, force mapping was applied to another phase-separated supported lipid bilayer consisting of DOPC and 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) in a 1:1 molar ratio. DPPC is known to form an ordered gel phase in the lipid bilayers.50,51 The histogram of the breakthrough forces of the DPPC gel phase (Supporting Information Figure S5C) indeed shows higher values of forces, F ∼ 6-12 nN, consistent with its more ordered packing. In contrast to DEC-Ceramide, breakthrough and adhesion events were observed in the DPPC gel phases. This indicates that the lipids in the ceramideenriched domains may have an even more ordered organization than a typical gel phase such as DPPC. AFM height images (Supporting Information Figure S5A,S5B) of the DOPC/DPPC bilayer after force mapping showed holes (4∼5 nm deep) in the DPPC domains, highlighting the less fluid nature of the DPPC gel phase, suggesting that the
(47) Sankaram, M. B.; Thompson, T. E. Biochemistry 1990, 29, 10670–10675. (48) Simons, K.; Ikonen, E. Science 2000, 290, 1721–1726. (49) Slotte, J. P. Chem. Phys. Lipids 1999, 102, 13–27.
(50) Choucair, A.; Chakrapani, M.; Chakravarthy, B.; Katsaras, J.; Johnston, L. J. Biochim. Biophys. Acta, Biomembr. 2007, 1768, 146–154. (51) Giocondi, M. C.; Vie, V.; Lesniewska, E.; Milhiet, P. E.; ZinkeAllmang, M.; Le Grimellec, C. Langmuir 2001, 17, 1653–1659.
7476 DOI: 10.1021/la900395w
Langmuir 2009, 25(13), 7471–7477
Sullan et al.
Article
disordered phase and the lower indentation-higher breakthrough force region with the liquid ordered domains. The high-resolution force mapping and pairwise scatter plots of the measured nanomechanical quantities allow the direct determination of the breakthrough forces, adhesion, and indentation of a lipid bilayer, as well as a straightforward correlation of those properties to the bilayer structures.
Conclusion Figure 7. Indentation plotted against the breakthrough force of a DEC bilayer.
presence of cholesterol leads to more efficient packing of the bilayer. This comparison further indicates that the ceramideenriched regions exhibit an organization different from that of gel phases. Heterogeneity in Distinct Phases Observed at the Nanoscale Level through Force Mapping. In addition to how cholesterol and ceramide affect the mechanical properties of a bilayer, it is worthwhile to examine the heterogeneity at the nanoscale level observed in the different coexisting phases. The histograms of breakthrough forces and elasticity reflect this wide distribution. This heterogeneity is more apparent in the visual maps and contour maps presented. In particular, the contour map representation of the breakthrough forces clearly defines the boundary between the liquid ordered domains and fluid disordered phases in the DEC system (Figure 3B) and for the subdomains in the case of DEC-Ceramide (Figure 4E). It is known that the major force that governs the self-assembly of amphiphiles into welldefined structures, such as micelles and bilayers, is derived from the hydrophobic attraction at the hydrocarbon-water interface, which induces the molecule to associate. The hydrophilic, electrostatic, or steric repulsion of the headgroups thus imposes the opposite requirement, and they remain in contact with water. Measuring distinct phase boundaries is an important practical consideration, because reliable data interpretation of multicomponent lipids with different compositions is needed in understanding the properties of a bilayer.52 AFM force mapping employed in this work strongly provides such lateral heterogeneity and distinct phase boundary information. In addition to the 2D visual maps, pairwise scatter plots of the measured mechanical quantities that were extracted using our self-developed batch analysis code, correlate the lipid organizations to those physicochemical properties. The scatter plot in Figure 7 shows two distinct clusters: higher indentation-lower breakthrough force and lower indentation-higher breakthrough force regions. This scatter plot further confirms that the higher indentation-lower breakthrough force region can be correlated with the fluid (52) Feigenson, G. W. Annu. Rev. Biophys. Biomol. Struct. 2007, 36, 63–77.
Langmuir 2009, 25(13), 7471–7477
We have demonstrated that high-resolution force mapping coupled with AFM imaging provides direct correlation of the organization of multicomponent lipid mixtures to their nanomechanical properties that were not previously achieved. The 2D visual maps of the intrinsic breakthrough forces, elastic moduli, and adhesion generated from the self-developed batch analysis code can directly correlate the fine structures of multicomponent lipid mixtures with their mechanical stability. In the ceramide-incorporated bilayer system being studied here, we observed an increase in membrane mechanical stability, which was attributed to the influence of ceramide in the lipid organization, as well as the displacement of cholesterol as a result of the generation of ceramide-enriched domains. We showed that ceramide-enriched domains exhibit a different packing behavior from the well-known gel phases, indicating a high mechanical rigidity. Our results provide strong evidence that mechanical stability and compactness is the basis of the ceramide-induced formation of signaling platforms in cell membranes, and hence, AFM force mapping is a valuable complement to other biophysical techniques currently used in studying multicomponent lipid bilayer mixtures. Our results on the effects of direct incorporation of ceramide into a ternary lipid mixture on the properties of the bilayer are closely related to several recent studies that examined the influence of ceramide on membranes treated with enzyme Sphingomyelinase.11,14-16,39 Force mapping on the ceramide-enriched regions by this enzymatic generation will provide a comparison to the direct incorporation of ceramide being studied here, and these experiments are presently under investigation. Acknowledgment. We thank Dr. Linda Johnston and Dr. Dusan Vobornik for numerous stimulating discussions. S.Z. would like to thank the NRC Nanometrology program, and R.S. acknowledges NSERC (RGPIN 31249) for the financial support. Supporting Information Available: Detailed description of the self-developed batch analysis code, adhesion force map and histogram, as well as force mapping results of DEC111 and DOPC/DPPC bilayers.This material is available free of charge via the Internet at http://pubs.acs.org.
DOI: 10.1021/la900395w
7477
