Comparison Actin- and Glass-Supported Phospholipid Bilayer Diffusion Coefficients

Abstract

The formation of biomimetic lipid membranes has the potential to provide insights into cellular lipid membrane dynamics. The construction of such membranes necessitates not only the utilization of appropriate lipids, but also physiologically relevant substrate/support materials. The substrate materials employed have been shown to have demonstrable effects on the behavior of the overlying lipid membrane, and thus must be studied before use as a model cushion support. To our knowledge, we report the formation and investigation of a novel actin protein-supported lipid membrane. Specifically, inner leaflet lateral mobility of globular actin-supported DMPC (1,2-dimyristoyl-sn-glycero-3-phosphocholine) bilayers, deposited via the Langmuir-Blodgett/Langmuir Schaefer methodology, was investigated by z-scan fluorescence correlation spectroscopy across a temperature range of 20–44°C. The actin substrate was found to decrease the diffusion coefficient when compared to an identical membrane supported on glass. The depression of the diffusion coefficient occurred across all measured temperatures. These results indicated that the actin substrate exerted a direct effect on the fluidity of the lipid membrane and highlighted the fact that the choice of substrate/support is critical in studies of model lipid membranes.

Introduction

Phospholipid membranes play a critical role in the establishment and maintenance of biological homeostasis (1). In the cell, lipid membranes surround organelles and form the plasma membrane. These supramolecular structures perform a variety of activities such as maintaining electrochemical gradients (2), aiding in the accumulation and utilization of essential metabolic molecules (3), and conducting intracellular signaling cascades (4), in addition to numerous cellular activities.

The underlying cellular cytoskeleton shapes and orders membrane constituents that, in part, define cellular membrane function (5). The cytoskeleton is largely composed of the protein found in all eukaryotic cells, actin. Actin is a 42-kDa cytosolic protein capable of dynamic and reversible polymerization from its globular, monomeric G-actin state, to its filamentous F-actin state through the binding and subsequent hydrolysis of ATP (6). This dual-conformational character allows actin to play many critical roles in a cell. Specifically, actin has been implicated in facilitating protein and vesicle trafficking (7,8), cell motility (9), and morphology (10), and has also been shown to interact indirectly with membranes via lipid-binding proteins (11). These interactions allow for the actin cytoskeleton to coordinate membrane domains (12), form cellular structures (13), and transport various factors to and from the membrane inner leaflet (14). Further understanding of membrane structure and function, as related to actin, promises to provide greater insight into specific physiological and pathophysiological cellular events (15,16).

Although studies of actin/lipid membrane dynamics may be performed in situ within live cells, such studies are potentially complicated by the presence of additional cellular constituents. Model cellular membranes provide an excellent alternate platform for investigating actin/lipid interactions in a highly controlled environment. Model membrane systems have been fabricated in a variety of geometries including planar lipid mono-, bi-, and multilayers, as well as various-sized vesicles (3,17). Planar lipid membranes are particularly amenable to microscopic and spectroscopic investigations (3,18) and lend structural verisimilitude to the eukaryotic plasma membrane. Planar phospholipid bilayer systems have been developed on solid (19), tethered (20), and cushioned supports (21). Clearly, to optimize the biological relevance of supported planar membrane systems, the choice of the supporting surface and the sample environment are critical. Further, detailed knowledge of substrate effects on lipid bilayer dynamics is essential for interpretation of lipid membrane behavior (2,22,23). As noted by Seeger et al. (24), the inner leaflet of the membrane comes into direct contact with the supporting substrate, therefore the perturbation of the planar membrane system by the supporting substrate must be fully characterized. To date, studies (reviewed elsewhere (18)) have suggested both strong and weak coupling of the planar membrane to the supporting surface.

As stated above, in vivo cellular membranes are supported by an actin cytoskeletal network (interacting with the inner leaflet), while the outer leaflet of the membrane is exposed to varying aqueous environments or the extracellular matrix. As such, a planar membrane system with the inner leaflet supported by an actin cushion and the outer leaflet exposed to an aqueous environment would provide a highly physiologically relevant system for investigating membrane biophysics. Previous work (reviewed by Vogel and Schwille (25)) has investigated actin interactions with lipid monolayers (26), biofunctionalized monolayers (27), and bilayers (28). Actin has been encapsulated inside giant unilamellar vesicles for the study of membrane organization (29) and the mechanical influence on the hybrid membrane system (30). Additionally, F-actin has been investigated as an inducer of single ionic channels (31) and separately, the binding affinity of actin to phospholipid monolayers has been studied (26,27).

Although a wealth of cushioned and solid-supported planar membrane systems have been developed (reviewed by Kiessling et al. (32)), actin has not, to the authors’ knowledge, been employed as the supporting surface for a phospholipid membrane. As such, to our knowledge, this work represents the first-known realization of an actin-supported planar phospholipid membrane. The actin support was fabricated via a novel self-assembly procedure, and characterized using surface immunochemistry and ellipsometry. Phospholipid bilayers were deposited by the Langmuir-Blodgett/Langmuir Schaefer (LB/LS) technique. Inner leaflet lateral mobility was determined as a function of temperature via z-scan fluorescence correlation spectroscopy (FCS). The lateral mobility and main-phase transition temperature of the actin-supported inner leaflet were compared to values determined from glass-supported bilayers. Importantly, to our knowledge, the assessment of the novel actin-supported phospholipid membrane described herein presents factors to be considered in the development of model membrane systems.

Materials and Methods

Actin purification

Actin was extracted from acetonic powder, prepared from rabbit skeletal muscle, following the protocol of Pardee and Spudich (33). Purified actin concentration was evaluated by a Bradford protein assay (34) on a BioTek Synergy 2 Microplate Reader (Winooski, VT) and subsequently diluted to 30 μg/mL in G-buffer (5 mM tris(hydroxymethyl)aminomethane, 0.2 mM calcium chloride, 0.2 mM adenosine triphosphate, and 0.5 mM dithiothreitol, pH 7.9) (33). Before use, actin solutions were dialyzed against G-buffer for 48 h. A final dialysis was performed against PBS (phosphate-buffered saline) (137 mM sodium chloride, 2.7 mM potassium chloride, 8.1 mM sodium phosphate, and 1.76 mM potassium phosphate, pH 7.8) for 24 h, to remove residual G-buffer components from the protein solution. Isolated actin was used within one week of initial purification.

Surface functionalization

A cleaned (35) No. 1.5 square cover glass (22 mm; Corning, Corning, NY) was used as the substrate for actin functionalization. Actin surface functionalization was accomplished using a modified procedure from Okada et al. (36) and An et al. (37). The cover glass was incubated for 3 h at 24°C in 5 mM APTMS ((3-aminopropyl)trimethoxysilane; Gelest, Morrisville, PA), then the APTMS was dissolved in a 5:1 v/v solution of ACS-certified acetone (Fisher Scientific, Suwanee, GA) and 18.2 MΩ⋅cm water. After incubation, APTMS-functionalized cover-glass substrates were rinsed in a 5:1 v/v acetone/18.2 MΩ⋅cm water solution and dried under nitrogen flow. Directly before use, APTMS-functionalized cover-glass substrates were incubated for 1 h in 10% v/v glutaraldehyde (Sigma-Aldrich, St. Louis, MO) in 18.2 MΩ⋅cm water. The functionalized cover glasses were rinsed with 18.2 MΩ⋅cm water, and immersed for 2 h in a 30 μg/mL G-actin solution. Actin-modified cover-glass substrates were subsequently rinsed with 18.2 MΩ⋅cm water before use.

Surface characterization

Actin-modified cover-glass substrates were assessed for actin coverage on the cover glass using immunochemical staining. Cover-glass substrates were incubated for 8 h at 24°C in 1:200 v/v anti-C4 actin antibody (Millipore, Billerica, MA) in G-buffer with 2% v/v bovine BSA (serum albumin; Sigma-Aldrich), followed by rinsing with G-buffer. Control cover-glass substrates were incubated solely in G-buffer. The substrates were subsequently incubated overnight at 24°C in 1:200 v/v FITC-conjugated goat-anti-mouse antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) in G-buffer with 2% BSA. After secondary incubation, cover-glass samples were rinsed with excess G-buffer and mounted on a glass microscope slide using a 1:1 v/v solution of ACS-certified glycerol (Fisher Scientific) and PBS containing 1% by volume 1,4-diazabicyclo[2.2.2]octane (Sigma-Aldrich). Cover-glass substrates were imaged on a FluoView laser scanning confocal microscope (Olympus America, Melville, NY) using a 40× objective and images were collected. As a separate step, ellipsometry was performed to assess the thickness of the actin-functionalized substrate. The substrates employed were polished, monocrystalline silicon wafers, prepared using the method previously described for the cover glass. The substrates were measured using a model No. M-2000V spectroscopic ellipsometer (J.A. Woollam, Lincoln, NE) at an incident angle of 75° in air. Ellipsometric data was analyzed via the software WVASE (J.A. Woollam), employing a refractive index for actin of 1.59 (38).

Bilayer preparation

DMPC (1,2-dimyristoyl-sn-glycero-3-phosphocholine) and RhoPE (1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine-n-(lissamine rhodamine B sulfonyl) (ammonium salt))were purchased from Avanti Polar Lipids (Alabaster, AL), stored at −20°C, and used without purification. Preparation of bilayers has been described elsewhere (35). It is noted that RhoPE was included only in the inner leaflet of the supported bilayers fabricated in this work.

Fluorescence recovery after photobleaching

Bilayers were prepared as described, employing 1 mol % RhoPE in the DMPC inner leaflet on actin-modified cover-glass substrates and separately, on cover-glass substrates. Fluorescence recovery after photobleaching (FRAP) data was collected employing a FluoView 1000 Laser Scanning Confocal Microscope with a UPlanApo 20× 0.70 NA objective (Olympus America), and an ORCA-ER digital camera (Hamamatsu, Hamamatsu City, Japan). Bleaching time was 5 s over a circular area with radius of 2.728 μm. FRAP data was collected from multiple sites such that the mean image intensity at each time point was computer-captured and stored. Data was normalized, accounting for bleaching during the measurement and total intensity, following from Phair et al. (39). Normalized curves were averaged yielding one aggregate FRAP curve with the associated standard deviation of each time point.

Z-scan fluorescence correlation spectroscopy

A variation of fluorescence correlation spectroscopy, z-scan FCS, was first suggested by Sorscher and Klein (40), later actualized by Benda et al. (41), and has been reviewed in detail by Machán and Hof (18). Details of the instrumentation employed in this work have been described in Allgeyer et al. (42). Measurements of actin-supported, and separately, glass-supported phospholipid bilayers were made and analyzed following the scheme presented by Sterling et al. (35). Briefly, for each substrate, correlation curves were collected at each of 12 axial positions and fit to the two-dimensional diffusion model (Eq. 1) employing a Levenberg-Marquardt fit routine with the standard deviation serving as the fitting weight (1/σ²),

$$G{\left(\tau \right)}_{2D}=\frac{1}{N}{\left(1+\frac{\tau }{{\tau }_D}\right)}^{-1},$$ (1)

where N is the particle number in the observation volume, τ_D is the diffusion time, and τ is time (43,44).

For each temperature, the diffusion time and the particle number, resulting from the two-dimensional diffusion model fit (Eq. 1), were individually plotted as a function of the axial position. The parabolic dependences of τ_D and N on axial sample position were fit with Eqs. 2 and 3, respectively, as

$${\tau }_D=\frac{w_0^2}{4D}\left(1+\frac{{\lambda }_0^2\Delta z^2}{{\pi }^2{n}^2{w}_0^4}\right)$$ (2)

and

$$N=\pi c{w}_0^2\left(1+\frac{{\lambda }_0^2\Delta z^2}{{\pi }^2{n}^2{w}_0^4}\right),$$ (3)

where w₀ is the radius of the beam in the focal plane, c is the average concentration of fluorophores in the focal plane, λ₀ is the excitation laser wavelength, Δz is the relative axial position of the sample, and n is the refractive index of the medium (water, for this work) (41,42). The mean calibration free beam waist recovered from measurements on the actin and glass-supported bilayers described herein was (332 ± 9) nm, which is within experimental uncertainty of the beam waist determined using the reference fluor Alexa 546 in solution, assuming a diffusion coefficient of 341 μm² s⁻¹ as published by Petrásek and Schwille (45) of (324 ± 8.7) nm.

Results and Discussion

Assessment of the uniformity, distribution, and thickness of the actin monomers on the functionalized substrates was performed before bilayer deposition. Fig. S1 A in the Supporting Material presents a confocal fluorescence image of an actin-functionalized cover-glass substrate exposed to immunochemical staining. As indicated by the uniformity of the staining, the surface of the cover glass was fully coated with actin monomers. The lack of fluorescence in the no-primary antibody control (Fig. S1 B) confirmed the specificity of the staining for actin. Fig. S1, A and B, indicated that the method developed for actin functionalization of the cover glass was highly effective in adsorbing actin onto the substrate. Additionally, the staining uniformity demonstrated that there are no local defects perceptible at the optical level. Ellipsometric measurements were performed for each step of the functionalization procedure for a series of samples and indicated progressive film thickness increases. The mean thickness of the actin layer on the functionalized substrates was ∼46 Å. Actin layer thickness measured at the center of given functionalized substrates was consistent with thickness measured at the periphery. It is well known that surface roughness of the cushion supporting a membrane is critical, because defects present in the supporting substrate can exert direct topological effects on the membrane (46), as well as affect membrane structural integrity and fluidity (47). As such, the uniformity and distribution of the actin on the cover glass indicated that the actin-functionalized substrates were likely good cushions for subsequent phospholipid depositions.

With confirmation of uniform actin coverage on cover-glass substrates, DMPC bilayers (with RhoPE in the inner leaflet) were deposited on actin-functionalized substrates with a mean LB transfer ratio of 1.0 ± 0.1. Fig. 1, A–F, presents a schematic representation of the actin-supported bilayer while Fig. 1 G is a wide-field fluorescence image of an actin-supported DMPC bilayer excited with a 543-nm laser beam focused at the back aperture (∼1 mW of power). Fluorescence was collected through the same objective and laser excitation was removed with appropriate filters. Fluorescence images were captured with a Luca EMCCD camera (Andor Technology, Belfast, UK).

Figure 1.

Schematic representation of an actin-supported DMPC bilayer (not to scale). A cover glass (A) is functionalized with APTMS (B), glutaraldehyde (C), and G-actin (D), before LB deposition of RhoPE:DMPC (E) and LS deposition of DMPC (F), resulting in an actin-supported phospholipid bilayer. (G) Wide-field fluorescence image of an actin-supported DMPC bilayer doped with RhoPE in the inner leaflet. Note that the imaged area in (G) is not uniformly illuminated due to the Gaussian profile of the excitation beam. To see this figure in color, go online.

To confirm bilayer continuity and fluidity, FRAP experiments were performed on actin-supported bilayers and separately, glass-supported bilayers each containing 1 mol % RhoPE in the DMPC inner leaflet. Fig. S2 presents the averaged fluorescence recovery curves for each substrate type, indicating that bilayers on each substrate do exhibit fluorescence recovery. Fitting each of the curves (39) yielded ∼90% mobile fraction for each substrate type, with recovery half-times of 6.8 and 3.4 s for the actin-supported bilayers and glass-supported bilayers, respectively.

After visual confirmation of the presence and fluidity of a phospholipid bilayer on the actin-functionalized substrate, the lateral diffusion of the phospholipids comprising the RhoPE-doped inner leaflet was studied in detail by z-scan FCS. Diffusion coefficients of the RhoPE-doped inner leaflet were collected over a temperature range of 20–44°C, from low to high temperature, in increments of 1–2°C for three separate actin-supported DMPC membranes. Correlation curves, collected at each of 12 axial positions, were fit with a Levenberg-Marquardt routine with the standard deviation serving as the fitting weight (1/σ²). Fig. 2 (top) presents a sample correlation curve from a single axial position at 38°C fit to the two-dimensional diffusion model (Eq. 1). For each temperature, the diffusion time and the particle number, each resulting from the two-dimensional diffusion model fit routine, were individually plotted as a function of that axial position (individual data points in Fig. 2, bottom). Employing Eqs. 2 and 3, the diffusion time and particle number z-scan FCS data were fit, respectively, to determine the diffusion coefficient of RhoPE within the inner leaflet of the DMPC bilayer. It is expected that the resultant diffusion coefficients reflect the DMPC mobility due to the low concentration of RhoPE in the preparation of the supported bilayer (35,44,48).

Figure 2.

(Top) Sample FCS autocorrelation curve and two-dimensional diffusion model fit for the inner leaflet of an actin-supported DMPC bilayer at 38°C. (Bottom) Diffusion time versus relative axial position and particle number versus relative axial position fit with Eqs. 2 and 3, respectively, to yield the diffusion coefficient and effective concentration. A weighted average was employed to compute the diffusion coefficient for each temperature while standard error propagation methods were used to determine the uncertainties. To see this figure in color, go online.

Fig. 3 presents the mean diffusion coefficient as a function of temperature for three separate actin-supported bilayers (circles) yielding a phase transition curve. The mean diffusion coefficients and the uncertainties were determined from a weighted average, with the error propagating from the correlation curves and z-scan FCS-plot fit routines. The diffusion coefficients increased with temperature over a wide range of temperatures (23°C – 36°C). To enable comparison, DMPC bilayers were deposited on three separate bare cover glasses by the same method and diffusion coefficients for the RhoPE-doped inner leaflet were determined by z-scan FCS over the same temperature range. The diffusion coefficients for the glass-supported membranes (Fig. 3, triangles) also increased with temperature, but over a more narrow range (22–30°C). The main-phase transition temperature for each model membrane was determined via a sigmoidal fit method (49,50). The Boltzmann sigmoidal line shape employed in this work is

$$D\left(T\right)=\frac{D_0-D_f}{1+e^{\left(T-T_m\right)/\Delta T}}+D_f,$$ (4)

where D₀ is the initial (lowest) diffusion coefficient, D_f is the final (highest) diffusion coefficient, T_m is the main-phase transition temperature, and ΔT is the change in temperature over the range with the greatest change in D. The sigmoidal fits for the actin- and glass-supported DMPC bilayer phase transition curves (dotted lines) yielded inner-leaflet main-phase transition temperatures, T_m, of 32.7 ± 0.5°C and 24.3 ± 0.1°C, respectively. Previously reported main-phase transition temperatures for glass-supported DMPC bilayers fabricated by the LB/LS method range from 18 ± 0.8°C (51) to 20.4–24°C (21), while the standard main-phase transition temperature of DMPC is reported as 23°C (52). The main-phase transition temperature for the glass-supported DMPC bilayer is within error of the range reported by Baumgart and Offenhäusser (21). The main-phase transition temperature reported by Scomparin et al. (51) was determined by fitting their diffusion coefficient versus temperature data with a power law (noted as an underestimation). It is noted that the main-phase transition temperature of the actin-supported DMPC bilayer was significantly higher than all values reported for glass-supported DMPC bilayers.

Figure 3.

Diffusion coefficients plotted as a function of temperature for the inner leaflet of DMPC membranes supported on actin-functionalized substrates (circles) and separately, on glass substrates (triangles). The phase transition curves were fit with a Boltzmann sigmoidal line shape yielding main-phase transition temperatures of 32.7°C for the actin-supported bilayers and 24.3°C for the glass-supported bilayers. Measured main-phase transition temperatures of the inner leaflet of DMPC membranes, supported on, respectively, actin-functionalized substrates (vertical dashed lines) and glass substrates (dotted lines). (Inset) Phase transition curve of the inner leaflet of the DMPC membranes supported on actin-functionalized substrates. To see this figure in color, go online.

The elevated main-phase transition temperature of DMPC inner leaflets on actin-versus-glass substrates may potentially be a result of the high anionic charge density of the actin layer. Actin is an anionic protein that has been shown to interact with charged lipids, in both the monomeric (G-actin) and polymeric (F-actin) forms (53). Additionally, liposomes have been shown to induce G-actin to form F-actin, a process that is charge-dependent (54). However, Le Bihan et al. (55) and separately, Bouchard et al. (56) have found that zwitterionic lipids, including DMPC, have very little interaction with actin in situ. In instances where interaction between zwitterionic phospholipids and actin has been shown to be present, the interaction has been mediated by the presence of Mg²⁺, resulting in binding of actin to the phospholipids (57). It is noted that in this work, no Mg²⁺ was used in the preparation of the supports or phospholipid bilayers; however, Mg²⁺ was employed to polymerize actin during its preparation. Mg²⁺ was removed from actin by G-buffer and subsequent PBS washes of the actin solution; therefore, any phospholipid/actin interactions stemming from Mg²⁺ mediation should have been significantly decreased.

An alternative source of the elevated main-phase transition temperature of the DMPC inner leaflet on the actin-versus-glass substrate may be physical surface features of the actin substrate. It is noted that while the ellipsometric data indicated a relatively uniform thickness, they do not provide detailed root mean-square roughness values. As stated above, surface roughness is strongly correlated with membrane mobility. Indeed, if there are chains of F-actin on the surface, the fluidity of the membrane may be affected by penetration of F-actin filaments into the bilayer (31). The actin-functionalized substrates in this work were fabricated with G-actin to reduce surface defects that may stem from F-actin filaments based on previous work (58). While the cell membrane is often associated with F-actin in vivo, G-actin has been shown to exist at equimolar concentrations to F-actin in the cortical and submembrane cytoplasm (59). Indeed, the regions proximal to the cell membrane that have been shown to have increased G-actin concentration over F-actin concentration are at the leading edge of motile cells (60), within neuronal lamellipodia, at the leading edge of the advancing growth cone (61), and in the cortical cytoplasm of unfertilized oocytes (62). G-actin has also been shown to play a critical role in the functional polarization of osteoclasts (63).

The data obtained in this work was analyzed to probe the nature of diffusion via application of the submicron confinement model, or FCS diffusion law, reported by Wawrezinieck et al. (64). Three types of lateral diffusion may be explored with the submicron confinement model: free diffusion, diffusion within microdomains (so-called lipid rafts), and diffusion hindered by a meshwork, such as the actin cytoskeleton (64). The submicron confinement model was developed by employing FCS measurements with a varying observation area, and the method was subsequently adapted to the z-scan FCS technique by Humpolícková et al. (65). Indeed, Humpolícková et al. (65) demonstrated that the change in axial position during the course of a z-scan FCS measurement allowed the resulting data to be analyzed via the submicron confinement model by utilizing the change in the effective fluorophore concentration with respect to the area probed. Therefore, the apparent diffusion time, τ_D^app, in the submicron diffusion model when employing z-scan FCS data, may be expressed as

$${\tau }_D^{\text{app}}=t_0+\frac{w_0^2}{4D_{\text{eff}}}\frac{N}{N_0},$$ (5)

where D_eff is the apparent diffusion coefficient, w₀ is defined by Eqs. 2 and 3, N₀ is the effective concentration at the minimum beam waist, N is defined by Eq. 3, and t₀ is a constant. The resultant plot may be fit with a linear function to extract the value of t₀ and permit the following assessment of the lateral diffusion behavior: t₀ is zero for free diffusion, positive for diffusion in isolated microdomains/rafts, and negative for diffusion in a meshwork (64). Fig. 4 (top) presents sample FCS diffusion law data with linear fits for actin (circles) and glass-supported bilayers (triangles) at 38°C.

Figure 4.

(Top) Sample data of the apparent diffusion time, τ_D^app, as a function of the ratio of the particle number, N, to the particle number at the minimum waist, N₀. The data has been fit with a linear function to determine the value of t₀ (detailed in Eq. 5), for the inner leaflet of an actin-supported DMPC bilayer and separately, a glass-supported DMPC bilayer, at 38°C. (Bottom) The value of t₀ was determined for each model membrane and plotted as a function of temperature. The inset shows the variance in the actin and glass supported phospholipids at higher temperatures. To see this figure in color, go online.

Applying the same analysis to the z-scan FCS data collected at each temperature, t₀ may be plotted as a function of temperature (Fig. 4, bottom) for the actin and the glass-supported bilayers. Examination of the data indicates that t₀ is greater than zero at temperatures <∼31°C for the actin-supported bilayers, suggesting diffusion hindered by microdomains. At temperatures >∼31°C, the inner leaflet of the actin-supported bilayers has t₀ values near zero, indicative of free diffusion. It is noted that the change of the diffusion mode occurred at approximately the main-phase transition temperature determined from the diffusion coefficients measured via z-scan FCS. Examination of the data for the glass-supported bilayers suggests similar results, with the change in the diffusion mode potentially occurring near the measured main-phase transition temperature of 24.3°C. However, the uncertainty of the data in the temperature range of 20–24°C prohibits any absolute statements about the diffusion behavior. Regardless, it is clear that there is a substrate-dependent difference in behavior in the 20–24°C temperature range when comparing the actin-supported and glass-supported bilayers.

The change in the diffusion mode, from diffusion hindered by microdomains to free diffusion, reported in this work agrees with previous trends noted for hydrogel-supported DMPC bilayers (35). The diffusion modes in this work are also in good agreement with results of solid-supported phospholipid bilayers prepared by vesicle fusion (66). Recent work by Heinemann et al. (67), employing actin filaments anchored on top of solid-supported phospholipid bilayers, also utilized the submicron diffusion model (64). Heinemann et al. (67) reported that diffusion of the lipids in the presence of actin filaments exhibited a degree of confinement due to a meshwork (negative value for t₀). Combined with the findings of this work, diffusion of phospholipids in the presence of actin appears to display a dependence on the monomeric/polymeric state of actin.

Conclusions

To our knowledge, this work described for the first time an actin-supported DMPC membrane. Employing a novel self-assembly procedure, cover-glass substrates were successfully cross-linked to G-actin monomers, resulting in an actin-functionalized substrate amenable to phospholipid bilayer deposition. LB/LS deposition of DMPC on the actin-functionalized substrate was confirmed via the LB transfer ratio and visual observation of incorporated fluorescently tagged phospholipids. Further, incorporation of fluorescently tagged phospholipids during the deposition of the inner leaflet of the DMPC membranes enabled determination of diffusion behavior employing z-scan FCS. The resulting diffusion coefficients and phase transition curve revealed a significant interaction between the phospholipids and the actin when compared to similar DMPC bilayers supported on glass. The strong perturbative effect was not expected because actin has been reported in the literature to have very little interaction with zwitterionic phospholipids such as DMPC. Further analysis of the z-scan FCS data via the submicron confinement model indicated that the diffusion mode of the phospholipids changed from diffusion hindered by microdomains to free diffusion as the temperature increased through the measured main-phase transition temperature. Development of the model membrane system described in this work supports the hypothesis that the choice of substrate may substantially influence lipid diffusion behavior. Subsequently, careful consideration of cushion material is required to achieve relevant model membrane systems.