Article pubs.acs.org/Langmuir
Thermal Annealing Triggers Collapse of Biphasic Supported Lipid Bilayers into Multilayer Islands Sean F. Gilmore,† Darryl Y. Sasaki,∥ and Atul N. Parikh*,‡,§ †
Applied Science Graduate Group, ‡Department of Biomedical Engineering, and §Department of Chemical Engineering and Materials Science, University of California, Davis, Davis, California 95616, United States ∥ Bioscience Group, Sandia National Laboratory, Livermore, California, 94550, United States S Supporting Information *
ABSTRACT: The collapse of phase-separating single, supported lipid bilayers, consisting of mixtures of a zwitterionic phospholipid (POPC) and an anionic lipid (DPPA) upon thermal annealing in the presence of ions is examined using a combination of scanning probe, epifluorescence, and ellipsometric microscopies. We find that thermal annealing in the presence of ions in the bathing medium induces an irreversible transition from domain-textured, single supported bilayers to one comprising islands of multibilayer stacks, whose lateral area decays with lamellarity, producing pyramidal staircase “mesa” topography. The higher order lamellae are almost invariably localized above the anionic-lipid rich, gel-phase domains in the parent bilayer and depends on the ions in the bathing medium. The collapse mechanism appears to involve synergistic influences of two independent mechanisms: (1) stabilization of the incipient headgroup−headgroup interface in the emergent multibilayer configuration facilitated by ions in the bath and (2) domain-boundary templated folding. This collapse mechanism is consistent with previous theoretical predictions of topography-induced rippling instability in collapsing lipid monolayers and suggests the role of the mismatch in height and/or spontaneous curvature at domain boundaries in the collapse of phase-separated single supported bilayers.
1. INTRODUCTION
SLBs derive their stability and structural integrity through a combination of electrostatic, hydration, and van der Waals interactions.4 These forces are present in interactions between membrane amphiphiles, as well as between the membrane molecules and the surrounding ambient phase (i.e., hydrophobic effect and membrane-substrate adhesion energy).4 The balance of these interactions is thus strongly modulated not only by the type of the lipids used, which include size and charge of the headgroups and degree of chain unsaturation, but also the properties of the ambient phase (i.e., ionic strength and osmotic effects). When these factors combine, such as for mixed SLBs consisting of gel and fluid phase lipids, which also carry net headgroup charges, the criteria for bilayer stability become complicated, giving rise to conditions for destabilization of lamellar single SLBs. However, mechanisms that lead to instabilities and collapse in the structures of SLBs submerged in the aqueous phase are largely unknown. In this vein, instabilities and collapse conditions for Langmuir and Langmuir−Blodgett monolayers have been well documented. In Langmuir monolayers, a dominant class of instability, which characterizes monolayer collapse,16,17 results in the transition of two-dimensional (2D) amphiphilic monolayers18,19 at the air−water interface into three-dimen-
Single supported lipid bilayers (SLBs) represent a class of biomolecular interfacial films composed of two apposing monolayers of phospholipids stabilized at the aqueous interface of hydrophilic solids.1−3 They form spontaneously when lipid vesicles rupture and spread spontaneously onto hydrophilic surfaces submerged in aqueous environments.4 The interface between the bilayer and the substrate is lubricated by a thin layer of water (6−15 Å thickness on silica surfaces),5,6 allowing membrane molecules to recapitulate lateral contiguity and fluidity reminiscent of lipid membranes of vesicles and living cells.7,8 Continued interest in these interfacial films can be attributed to two main factors.3 First, they are proving to be useful, molecularly tailored models for characterizing essential biophysical properties of cellular membranes including lateral diffusivities and phase behavior of membrane molecules using surface sensitive probes including atomic force microscopy,9,10 X-ray11 and neutron reflectivity,6 quartz crystal microgravimetry,12 and surface plasmon resonance spectroscopy13 as well as fluorescence microscopy based methods. Second, because supported membranes integrate the biophysical membrane structure with a solid surface, they have potential for applications in the design of biocompatible surfaces, biosensors for diagnosis and detection of analytes that bind to membrane targets, and emerging classes of novel membrane-based biomimetic devices (e.g., photovoltaics).2,3,14,15 © 2014 American Chemical Society
Received: February 9, 2014 Revised: April 4, 2014 Published: April 7, 2014 4962
dx.doi.org/10.1021/la5005424 | Langmuir 2014, 30, 4962−4969
Langmuir
Article
phase domains in the parent bilayer. This unusual case of membrane binding in which fragments of contiguous bilayer become unbound from the substrate producing stacked mesas nucleated at the gel-phase domains lend support to the theoretical argument that nucleation of buckled folds during collapse is promoted by the mismatch in height and/or spontaneous curvature at domain boundaries.35
sional (3D) structures. Typically, this occurs when lateral pressure exceeds equilibrium spreading pressure of the amphiphile. The phenomenon is extensively studied in Langmuir monolayers primarily because of its relevance to folding of monolayers consisting of pulmonary lung surfactant,20−23 a feature of normal lung function.24,25 A significant body of literature describing mechanisms of monolayer collapse invokes one of the two major pathways depending on the nature of the 3D structures formed. In what is generally referred to as slow monolayer collapse,26,27 the instability arises due to lowering of the energy barrier for the nucleation of the more stable, bulk 3D phase above a critical surface pressure.28 Subsequent growth of the soluble bulk phase fragments then replaces the initial 2D interfacial monolayer. In contrast, when the 3D structures consist of discrete numbers of stacked monolayers or multilayers, the collapse does not follow the homogeneous nucleation and growth model above. Rather, it proceeds by mechanical processes of fracture or buckling depending on the elastic and cohesive properties of the interfacial monolayer.29−31 Generally, liquid-condensed (LC) or solid phase monolayers collapse by fracture whereas monolayers in the liquid-expanded (LE) phase solubilize.32 In the phase coexistence regime, on the other hand, monolayers consisting of coexisting LE and LC phases collapse by buckling; here, localized, large amplitude folds perpendicular to the compression direction become contiguous with the flat monolayer.32 Subsequent rupture of the folds then forms disjointed multilamellar islands.33 Theory suggests that the homogeneous monolayers become unstable to buckling only at zero surface tension, thus requiring impractically high surface pressures (π > 72 mN m−1).34 Experimentally, however, collapse occurs at much lower surface pressures, which is thought to result from the presence of defects by lowering the energetic barrier for buckling. In nonuniform monolayers, the differences in the spontaneous curvatures and/or heights of the coexisting phases or “mesas” at the domain boundaries, can serve as defects. Using elastic models for biphasic monolayers consisting of domains of differing elastic properties, Diamant and co-workers have theoretically shown that topographic features of “mesas” become pronounced producing overhangs, as the monolayer is compressed, ultimately becoming unstable resulting in collapse.35 In addition to destabilization of the Langmuir monolayers at the air−water interface, collapse in Langmuir− Blodgett films supported on solid substrates is also reported. Spatial variations in substrate properties, transfer conditions, and changes in the aqueous phase have all been shown to destabilize mono-, bi-, and multilayer Langmuir−Blodgett films.36−38 Here, we describe a structural instability leading to collapse of a single, biphasic phospholipid bilayer at the solid−liquid interface triggered by thermal annealing.2 The coexisting phases, differing in heights, elasticity, and headgroup charge, feature domain interfaces, which provide a pathway for structural destabilization. We find that single lipid bilayers, prepared by substrate-mediated vesicle fusion, consisting of binary mixtures of phase-separating zwitterionic and anionic lipids15 respond to thermal annealing by undergoing an irreversible transition from domain-textured 2D bilayers to one comprising islands of multibilayer stacks. The resultant stack shows discrete jumps in bilayer heights corresponding to integer numbers of bilayers. We find that the multilayer islands are almost invariably localized above the anionic-lipid rich, gel-
2. MATERIALS AND METHODS 1. Preparation of Silicon Substrates. Prime-grade silicon wafers were cut to approximately 20 mm squares. Silicon substrates were cleaned using a piranha etch solution consisting of 4:1 sulfuric acid and hydrogen peroxide (hydrogen peroxide was added to the sulfuric acid at 55 °C) at 100 °C (Caution! This mixture reacts violently with organic materials and must be handled with extreme care.) by immersing the substrates for 5−10 min. Cleaned substrates were rinsed with copious amounts of water and stored in deionized water prior to use. Typically, cleaned substrates were used within 1−2 days. 2. Preparation of Lipid Vesicle Suspensions. Phospholipids including 1-palmitoyl-2-oleoyl-sn-glycerophosphatidylcholine (POPC), 1,2-dihexadecanoyl-sn-glycero-3-phosphate (sodium salt) (DPPA), and 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) were purchased from Avanti Polar Lipids (Alabaster, AL). The fluorescent probe β-BODIPY 530/550 C5-HPC was purchased from Invitrogen (Grand Island, NY). MOPS buffer (3-(N-morpholino)propanesulfonic acid (20 mM), NaCl (100 mM)) was adjusted to pH 7.4 using 10% aqueous sodium hydroxide solution. Lipid films were prepared from stock solutions of POPC and DPPC in chloroform and DPPA in 50% methanol/chloroform, doped with 0.3% BODIPY 530/550 HPC for imaging by fluorescence microscopy. Vesicles were prepared by the standard tip sonication technique.39 3. Formation of Planar, Supported Lipid Bilayers. Substrates cleaned using Piranha-etch treatment described above were rinsed with fresh deionized water and then dried using a stream of nitrogen gas. A silicone spacer was applied to the substrate (Grace Biolabs, Bend, Oregon,) and the substrate wetted with 200 μL of MOPS buffer (20 mM MOPS, 100 mM NaCl, pH 7.4) for 2 min. Subsequently, 170 μL of the solution was replaced by the same volume of vesicle suspension. The vesicle solution was allowed to incubate with the substrate for 10 min at 21 °C., during which time vesicle fusion produces single supported bilayers. The samples are then washed in 170 μL increments of buffer (20 mM MOPS, 100 mM NaCl, pH 7.4). 4. Thermal Annealing of Planar, Supported Lipid Bilayers. Thermal annealing of supported bilayers was carried out using a dry bath (Torrey Pines Scientific, Carlsbad, CA). The samples were heated to 45 °C and maintained at that temperature for 2 min. Selected samples were annealed for 1 h. The samples were then removed and allowed to cool at 21 °C for 10 min. Samples were then fixed to a 60 mm sample dish using silicone epoxy. 5. Atomic Force Microscopy (AFM). AFM measurements were taken using a Veeco Dimension 3100 (Veeco Metrology, Santa Barbara, California). A silicon nitride tip, mounted to an in-liquid tipholder, with a spring constant of 0.03 N/m (Veeco Metrology, Santa Barbara, California) was used for scanning in contact mode. After the tip was mounted, the tip and wet cell were lowered into the MOPS buffer and allowed to sit in solution for 15 min prior to the first scan. Supported bilayer samples were scanned in the solution that they were annealed in. Data was analyzed using Gwyddion (A public-domain data SPM data analysis software). For the analyses of AFM images for domain size distribution, a Python script was used. Raw data (512 × 512 height data per scan) were sorted based on measured height values and grouped into bins of width 0.1 nm. Resulting histograms were plotted using the MatPlotLib library. 6. Imaging Ellipsometry. Ellipsometric measurements were carried out on a commercial imaging ellipsometer with a wet-cell (iElli 2000, Nanofilm, Germany) using a 532 nm laser at 3% power and at 55° inclination. Imaging of the unannealed bilayer on silicon substrates containing 20/79.3/0.7 DPPA/POPC/BODIPY was carried out in the previously mentioned MOPS buffer. 4963
dx.doi.org/10.1021/la5005424 | Langmuir 2014, 30, 4962−4969
Langmuir
Article
Figure 1. (a) Three-dimensional representation of ellipsometrically derived thickness map for a bilayer on silicon containing 20% DPPA, 80% POPC, and 0.3% BODIPY-DHPE. (b) Epifluorescence data for a supported lipid bilayer consisting of a biphasic POPC/DPPA mixture doped with BODIPY 530/550 (59.7/40/0.3 molar ratio), 20×, scale bar 50 μm. (c) Line scan corresponding to the epifluorescence data showing two depressions in intensity between 150 and 200 μm.
configuration (∼3−5 um2/pixel), the bilayer appears uniform devoid of large, microscopic defects or domains. Note, however, that these measurements do not rule out the presence of any submicroscopic defects (
