This protocol describes the formation of supported lipid bilayers and the addition of cytoskeletal filaments and motor proteins to study the dynamics of reconstituted, membrane-tethered cytoskeletal networks using fluorescence microscopy.
The surface of a living cell provides a versatile active platform for numerous cellular processes, which arise from the interplay of the plasma membrane with the underlying actin cortex. In the past decades, reconstituted, minimal systems based on supported lipid bilayers in combination with actin filament networks have proven to be very instrumental in unraveling basic mechanisms and consequences of membrane-tethered actin networks, as well as in studying the functions of individual membrane-associated proteins. Here, we describe how to reconstitute such active composite systems in vitro that consist of fluid supported lipid bilayers coupled via membrane-associated actin-binding proteins to dynamic actin filaments and myosin motors that can be readily observed via total internal reflection fluorescence microscopy. An open-chamber design allows one to assemble the system in a step-by-step manner and to systematically control many parameters such as linker protein concentration, actin concentration, actin filament length, actin/myosin ratio, as well as ATP levels. Finally, we discuss how to control the quality of the system, how to detect and troubleshoot commonly occurring problems, and some limitations of this system in comparison with the living cell surface.
The plasma membrane of a living animal cell constantly interacts with the adjacent actin cytoskeleton, and together they form an active composite material that fulfills a multitude of cellular functions1,2. To study processes at this lipid membrane-actin interface, the reconstitution of cytoskeletal networks on top of supported lipid bilayers (SLBs) has proven to be very helpful. This minimal system approach allows the precise control of cytoskeleton network components and lipid composition. Compared to the free-standing lipid membranes of giant unilamellar vesicles, the planar geometry of SLBs allows efficient use of state-of-the-art microscopy techniques such as super-resolution3,4, total internal reflection fluorescence (TIRF)5,6,7, or interferometric scattering8 to study the spatial organization and dynamics of cytoskeletal networks. TIRF provides the highest contrast for fluorescently labeled components, since the signal of unbound labeled molecules in the solution contributing to the background signal is minimal.
Here, we describe a basic protocol for the formation of actomyosin networks tethered to supported lipid bilayers, which are widely used in the field to study the physics of active, quasi-2D networks9,10,11 and their effect on membrane organisation3,5,12,13,14,15,16 (Figure 1). This approach is not limited to actin-based networks but can also be adapted easily to explore microtubules, intermediate filaments, or composite networks of mixed nature and to study a variety of interactions between lipid membrane proteins and cytoskeletal components using surface-sensitive microscopy methods.
To keep this protocol focused, we have excluded a detailed description of the purification and labeling of actin and myosin proteins or details about how to tune and control the contractility and organization of actomyosin networks. One should refer to other protocols that are published alongside this one in the JoVE Methods Collection, In Vitro Reconstitution of Cytoskeleton Networks for Biomaterials, Biophysics and Active Matter Research17.
Figure 1: Schematic of the in vitro actin-membrane active composite system. Created with Biorender. Please click here to view a larger version of this figure.
1. Reagents and equipment
Figure 2: Schematic showing the workflow from preparing multilamellar vesicles and small unilamellar vesicles to the formation of supported lipid bilayers. Created with Biorender. Please click here to view a larger version of this figure.
2. Reconstitution of membrane-tethered actin networks
Figure 3: Quality assessment of the bilayers with quick FRAP assay. Supported lipid bilayers (SLBs) prepared from DOPC and Ni-NTA lipids (98:2 mol%) are coated with HYE (10xHis-YFP-tagged membrane-actin linker). After the unbound protein is washed out, the fluorescent bilayer is imaged under a TIRF microscope. A small region on the bilayer is photobleached with high laser power, and the recovery of fluorescence is recorded. (A) A good bilayer always recovers fast, with an expected diffusion coefficient of 1-1.5 µm2/s for the lipid composition used in this case. (B) Bad bilayers recover very slowly or do not recover at all. (C) Representative images of bad bilayers: (C-i) a bilayer with holes, (C-ii) a bilayer with big, immobile lipid patches, and (C-iii) a bilayer with small, immobile dots. Please click here to view a larger version of this figure.
Figure 4: Schematic showing how to polymerize actin using the target buffer method. Please click here to view a larger version of this figure.
Figure 5: Spatial organization of HYE upon binding to F-actin. TIRF snapshots showing the spatial organization of HYE before and after the addition of actin filaments (labeled with Atto-635 maleimide). The HYE organization is homogenous before the addition of F-actin and becomes colocalized and coaligned along actin filaments. Please click here to view a larger version of this figure.
3. Data analysis
For representation, here a typical postbleach profile from the 1st image after photobleaching (image at t = 0 s in Figure 3A) and its fit to the following function28 (see Figure 6A) is shown:
The value of re (23.94 µm) calculated by the fit to this curve is very similar to the re calculated in step 2.8.4. (23.24 µm). Here, K is a bleach depth parameter that can be directly estimated from F0 (described in step 2.8.4.). Similarly, Figure 6B shows the recovery profile and its fit to the following function28:
We find the fitted value of the diffusion coefficient to be 1.34 µm2/s, a value that closely agrees with the value of 1.39 µm2/s that is calculated by the formula in step 2.8.4. Here, MF stands for the mobile fraction of the lipid bilayer that represents the fraction of the bleached population that recovers back. The mobility of lipid-anchored molecules depends, of course, on the lipid composition and its physical state (liquid or gel phase). For our experiments using DOPC-based lipid membranes, the mobility should be >1 µm2/s, and the mobile fraction should not be less than 0.9 to indicate a good lipid bilayer. We recommend the use of the manual fitting-free method for a quick test of the quality and mobility of the bilayer. The fitting method can be useful while automating the analysis for many FRAP curves. Further, if one wants to perform a more sophisticated FRAP experiment to systematically characterize diffusion in the system, we recommend the reader to this review from Lorén et al.30 for more detail on fitting models and potential pitfalls in experimental design.
Figure 6: Quantifying the diffusion coefficient of lipid bilayers. (A) Line profile of the first image after photobleaching (t = 0 s in Figure 3A) and its fit to equation 4 to calculate the effective bleach radius. (B) The recovery profile of the bleached region and its fit to equation 5 to calculate the diffusion coefficient and mobile fraction. Please click here to view a larger version of this figure.
A typical result of the experiments described above showing the dynamic assembly and organization of an acto-myosin network linked on a supported lipid bilayer imaged by TIRF microscopy is depicted in Figure 7 and Supplementary Video S1.
Figure 7 shows an image montage of the linker protein, F-actin, and myosin-II.
Figure 7: Contractile actomyosin flows drive local clustering of the membrane-actin linker protein HYE. TIRF snapshots of HYE (YFP-tagged), actin filaments (labeled with Atto-635 maleimide), and myosin II filaments (labeled with Atto-565 maleimide) upon addition of myosin II to an SLB containing HYE and F-actin. Time is indicated on the top: 0 min is immediately before fluorescent myofilaments started appearing in the TIRF field. HYE and F-actin are homogeneously distributed over the lipid bilayer prior to myosin addition (0 min). Myosin activity induces contractile actomyosin flows, which emerge into aster-like structures at the steady state (15 min), driving local clustering of the coupled membrane component (HYE). The lowest row is a merge of actin (yellow) and myosin II (magenta) images showing the organization of actin and myosin at different time points. The images used in making these montages were corrected in Fiji for background signal, non-uniform intensity patterns, and translational movement. Scale bar = 10 µm. For details, see Supplementary Video S1. Please click here to view a larger version of this figure.
Buffer Name | Composition | |
Lipid Rehydration buffer | 50 mM HEPES, 150 mM NaCl, 5% sucrose, pH 7.5 | |
SLB Formation buffer | 50 mM HEPES, 150 mM NaCl, pH 5-6 | |
SLB Storage buffer | 50 mM HEPES, 150 mM NaCl, pH 7.2 | |
Protein Dilution buffer | 20 mM HEPES, 100 mM KCl, 1mM TCEP or DTT, pH 7.2 | |
1X ME or Actin ion-exchange buffer | 50 mM MgCl2, 0.2 mM EGTA, 10 mM HEPES, pH 7.2 (store at 4°C) | |
1X KMEH or Actin polymerization buffer | 50 mM KCl, 1 mM MgCl2, 1 mM EGTA, 50 mM HEPES, pH 7.2 | |
100 mM ATP stock | 100 mM ATP disodium salt, 50 mM Tris, 50 mM NaCl, 5 mM MgCl2, 2 mM EGTA, pH 7.5 (store at -20°C) | |
2x Target buffer | 2x KMEH, 2 mg/ml BSA, 2mM ATP, 5mM TCEP (stored at 4°C) | |
G-buffer | 2 mM Tris, 0.1 mM CaCl2, 0.2 mM ATP, 0.5 mM TCEP, 0.04 % NaN3, pH 8 (store at 4°C) | |
Myosin II buffer | 500 mM KCl, 1 mM EDTA, 10-20 mM Hepes, pH 7.0 | |
Gel-filtration chromatography buffer | 50 mM Tris-HCl, 150-300 mM NaCl, 5 mM TCEP, 0.1% Tween-20, pH 7.5 | |
Capping protein Storage buffer | 10 mM Tris·Cl, 50 mM NaCl, 1 mM TCEP, pH 7.5, 20% glycerol |
Table 1: List of buffer compositions used in this protocol.
Common problems and their trouble shooting | Problem | Cause | Possible solutions | ||||||
1 | Lipid bilayer shows no diffusion | The most probable cause for this problem is dirty coverglass that can happen when the cleaning solution is aged or the heating did not take place during bath sonication. These bilayers have a ‘vesicular’ appearance because the bursted vesicles stick to the coverglass but do not fuse with each other. Using MLVs older than 6 weeks or SUVs older than 6 days, or adding low amounts of SUVs can also lead to vesicular bilayer formation. | Use fresh cleaning solution. Make sure the heater is turned on and the temperature is between 45-65°C. Use fresh lipid mixes.(Using a fluorescent lipid probe versus a fluorescent protein probe can sometimes manifest differently. E.g., if the bilayer has sub-diffraction defects and surface passivation step is skipped (or does not work), the lipid probe will show a uniform intensity distribution but the fluorescent protein probe may display bright fluorescent spots. ) | ||||||
2 | Lipid bilayer has bright patches | Long incubation of SUVs for bilayer formation can create a lipid bilayer that is overall diffusing but with occasional bright patches. These patches can be multi-layered bilayers that can attract large amounts of fluorescent probe. | 15-20 min incubation with SUVs is enough. Make sure the probe is not aggregating: a quick hard spin of the linker protein (300 x g for 15 min at 4 °C) can remove the aggregates | ||||||
3 | Lipid bilayer has dark holes | This happens when the bilayer is made from old SUVs and imaged for prolonged hours (> 4 hours post formation), or the pH of the solution changes drastically due to prolonged imaging (e.g., in the high ATP state and in the presence of certain oxygen scavengers), or when the surface is over-passivated with beta-Casein (adding too much beta-Casein for more than 10-15 min and or not washing it out). | Use fresh lipids. Reduce the imaging frame rate or the effective laser illumination time. Use buffers with higher buffering capacity. | ||||||
4 | Lipid bilayer shows slow diffusion | Lipid bilayers with high percentage of cholesterol, long saturated lipids or charged lipids diffuse slower. | In such cases, prepare your sample at a high temperature. One can also use a simple, tested lipid composition as a control along with complex and untested lipid compositions. Make sure the glass is clean. | ||||||
5 | Actin does not polymerize | Target buffer is old, G-actin stock is too old, old and new G-actin were co-polymerized. | Make sure the Ca2+ is replaced by Mg2+ before polymerization (using ME buffer). Use fresh ATP-Mg2+ stock. Use freshly recycled G-actin. Make sure the concentration of F-actin (in terms of G-actin) added to the bilayer is higher than 0.2 µM. For lower concentrations, use phalloidin stabilized F-actin. | ||||||
6 | Actin does not bind to the bilayer | Membrane-actin linker is not added or added at very low concentration—this can be inferred form the fluorescence of the linker protein. If the fluorescence is decent, the membrane-actin linker has lost actin-binding capacity. Also, if the linker protein is non-specifically bound to the glass surface (when the bilayer is bad), it might not recruit actin filaments. | Make sure the bilayer is diffusing. Use fresh linker protein | ||||||
7 | Fluorescent F-actin signal is weak | The ratio of labelled to dark actin is too low. Either the labelled actin or unlabelled actin is too old and they are not copolymerizing with each other. | Recycle actin again, and retry the ploymerization with freshly recycled actin. Photodamage can destroy or depolymerize F-actin; if possible, use red or far-red dyes for actin (and myosin). | ||||||
8 | Myosin does not show contractility | it may be observed that after adding ATP to the myosin-infused system, there is no contractility of the acto-myosin. | Check if myosin concentration or purity level is good. Use freshly recycled myosin (use within 6 weeks after recycling). Adding fresh ATP to the myosin mix can help. De-gassing buffers and using oxygen scavengers etc. can reduce photodamaging of the motors. Further information might be found in the protocols by Plastino et al. or Stam et al. of the same methods collection | ||||||
9 | Coverglass is not hydrophilic | Coverglass is not cleaned properly. | Clean hydrophilllic coverglass is crucial for the lipid bilayer formation. A useful, visual readout of the hydrophilicity of the coverglass after the cleaning protocol is to observe the wetting of the glass by water. Add a small volume of water to a flat-lying coverslip. The water will remain in the shape of a round droplet if the coverslip is not cleaned properly. However, the same volume of water will spread out and form a thin layer, on a treated hydrophilic coverglass. This wetting behaviour of the water on the cover glass surface can be used to ascertain if the cleaning steps with the cleaning solution/NaOH have worked . |
Table 2: Troubleshooting guide summarizing common problems and corresponding solutions.
Supplementary Video S1: Contractile actomyosin flows drive local clustering of the membrane-actin linker protein HYE. TIRF timelapse of HYE (YFP-tagged), actin filaments (labeled with Atto-635 maleimide), and myosin II filaments (labeled with Atto-565 maleimide) upon addition of myosin II to an SLB containing HYE and F-actin. Time is indicated on the top: 0 min is immediately before fluorescent myofilaments started appearing in the TIRF field. Scale bar = 10 µm. Please click here to download this File.
This protocol presents a versatile platform and a starting point to design experiments to study the membrane-cortex interface of cells. Critical steps are the preparation of clean glass slides, using fresh lipids for efficient SUV formation (both affecting the quality of SLBs), and the use of freshly recycled myosin II proteins for dynamic actin filament reorganization. When imaging dynamics over a long time, it is very important to incorporate an oxygen scavenger system (e.g., protocatechuic acid and protocatechuate 3 4-dioxygenase5,31).
The open-chamber design allows the sequential addition of components to an existing system without inducing lipid flows. This can be an important advantage over commonly used, closed-chamber approaches or work using encapsulated proteins within liposomes36. Contrary effects such as protein-induced membrane deformation cannot be studied with glass-adsorbed lipid bilayers.
The lipid bilayers can be formed with a wide range of lipid compositions. It begins with adsorption of the lipid vesicles to the hydrophilic glass surface, followed by either spontaneous vesicle rupture due to surface-vesicle and direct vesicle-vesicle interactions or the adsorbed vesicles reaching a critical coverage after which a small fraction of vesicles rupture, forming active edges, which eventually leads to the bilayer formation32. Besides glass, various substrates can be used to form supported lipid bilayers, such as Mica (e.g., for atomic force microscopy), soft substrates (e.g., poly-di-methyl-siloxane), polymer cushions33,34,35, spanning between holes of electron microscopy grids14. Droplet interface bilayers are another interesting method to create stable, free-standing lipid bilayers36. The inclusion of acto-myosin networks into vesicles or emulsions is a very powerful method to study this minimal system in a cell-like geometry37,38, and which is described in detail elsewhere39.
The authors have nothing to disclose.
This work was supported by the AXA research fund and the Warwick-Wellcome Quantitative Biomedicine Programme (Wellcome ISSF, RMRCB0058) for DVK, NCBS-TIFR for AB and ST, and the Wellcome-DBT Margdarshi fellowship (IA/M/15/1/502018) for SM. DVK would also like to thank the Biophysical Society for enabling the virtual networking event "Challenges in understanding multi-component cytoskeletal networks from the molecular to the meso-scale", which contributed to the creation of this protocol collection.
1,2 dipalmitoyl-sn-glycero-3-phosphoethanolamine-N- (lissamine rhodamine B sulfonyl) | Avanti Polar Lipids | 810158 | 16:0 RhoPE |
1,2-dioleoyl-sn-glycero-3- [(N-(5-amino-1 carboxypentyl) iminodiacetic acid) succinyl] (nickel salt) | Avanti Polar Lipids | 790404 | DGS-NTA-Ni2+ |
1,2-dioleoyl-sn-glycero-3-phosphocholine | Avanti Polar Lipids | 850375 | DOPC |
1,2-dipalmitoyl-sn-glycero-3-phosphocholine | Avanti Polar Lipids | 850355 | DPPC |
Amber glass vials | ThermoFisher | B7990-2A | |
ATP disodium salt | Sigma Aldrich | A26209 | |
Attofluor cell chamber | ThermoFisher | A7816 | |
Bath sonicator | GT Sonic | 1860QTS | |
beta-casein | Sigma Aldrich | C6905 | |
CaCl2 | ThermoFisher | 12135 | |
chloroform | Sigma Aldrich | 650471 | alternatively from Electron Microscopy Sciences, 50980296 |
Cover slips, #1, 25 mm diameter, Gold Seal | Harvard Apparatus | 64-0705B | |
Cover slips, #1, 40×22 mm, Gold Seal | ThermoFisher | 48404-031 | |
EDTA | ThermoFisher | G12635 | |
EGTA | Himedia | MB130 | |
Gas tight glass syringes, with removebla needle, blunt, volumes 10 µL, 100 µL, 500 µL | Hammilton | 1700 series | |
Hellmanex III | Hellma Analytics | Z805939 | cleaning solution |
HEPES | Himedia | RM380 | |
KaH2PO4 | ThermoFisher | G13405 | |
KCl | ThermoFisher | G13305 | |
KOH | ThermoFisher | G26708 | |
Lipid extruder | Avanti Polar Lipids | 61000-1EA | |
MgCl2 | ThermoFisher | G15535 | |
Microtip sonicator | Sonics | VC750 | 3 mm Tip diameter |
Na2CO3 | ThermoFisher | G15955 | |
NaCl | Himedia | GRM853 | |
NaH2PO4 | ThermoFisher | G15825 | |
NaOH | ThermoFisher | G27815 | |
Nikon Ti Eclipse TIRF microscope | Nikon | With a TIRF unit connected through a polarization-conserving optical fibre to an Agilent monolithic laser combiner MLC400 with multiple laser lines with a 100X, 1.45 NA Nikon Oil Objective with two 512 x 512-pixel EMCCD cameras (Photometrics Evolve 512) with a 100X, 1.45 NA Nikon Oil Objective with two 512 x 512-pixel EMCCD cameras (Photometrics Evolve 512) | |
NOA88 | Norland Products | 8801 | |
PTFE Coated Tweezer Style #2A | Structure Probe | 0S2AT-XD | |
Refrigerated microcentrifuge | Eppendorf | 5424R | |
Sucrose | ThermoFisher | G15925 | |
UV-illuminator | Novascan | PSD PRO-UV | needs vacuum and O2 supply |