Protein-Rich Rafts in Hybrid Polymer/Lipid Giant Unilamellar Vesicles

Considerable attention has been dedicated to lipid rafts due to their importance in numerous cell functions such as membrane trafficking, polarization, and signaling. Next to studies in living cells, artificial micrometer-sized vesicles with a minimal set of components are established as a major tool to understand the phase separation dynamics and their intimate interplay with membrane proteins. In parallel, mixtures of phospholipids and certain amphiphilic polymers simultaneously offer an interface for proteins and mimic this segregation behavior, presenting a tangible synthetic alternative for fundamental studies and bottom-up design of cellular mimics. However, the simultaneous insertion of complex and sensitive membrane proteins is experimentally challenging and thus far has been largely limited to natural lipids. Here, we present the co-reconstitution of the proton pump bo3 oxidase and the proton consumer ATP synthase in hybrid polymer/lipid giant unilamellar vesicles (GUVs) via fusion/electroformation. Variations of the current method allow for tailored reconstitution protocols and control of the vesicle size. In particular, mixing of protein-free and protein-functionalized nanosized vesicles in the electroformation film results in larger GUVs, while separate reconstitution of the respiratory enzymes enables higher ATP synthesis rates. Furthermore, protein labeling provides a synthetic mechanism for phase separation and protein sequestration, mimicking lipid- and protein-mediated domain formation in nature. The latter means opens further possibilities for re-enacting phenomena like supercomplex assembly or symmetry breaking and enriches the toolbox of bottom-up synthetic biology.


Text S2: Preparation of LUVs
Hybrids were prepared from PDMS-g-PEO:soy PC:PE-Rho mixture (70:29.97:0.01,mol%).5 mg of lipid/polymer mixture in chloroform:MeOH (2:1, v/v) was deposited in a glass vial and the solvent was removed by evaporation under a gentle stream of nitrogen for 30 min.The thin lipid/polymer film was rehydrated with 200 mM sucrose, 1 mM Tris-HCl (pH 7.5) and re-suspended to a final lipid concentration of 5 mg ml −1 by vortexing.The suspension of multilamellar vesicles (MLVs) was subjected to 5 freezethaw cycles (1 min lN2, then water bath at 35 °C until thawed completely, followed by 30 s vortexing).Finally, the size and lamellarity of vesicles was unified by extrusion (21 times) through a 100 nm pore (polycarbonate membrane, Mini Extruder).

PDMS-g-PEO facilitates LUVs preparation and membrane solubilization
In the first step polymer/lipid mixture in chloroform:MeOH (2:1, v/v) was deposited in glass vials, next solvent was evaporated under gentle N2 stream (Figure 1) and the thin lipid/polymer film was re-suspended in aqueous media.PDMS-g-PEO is liquid at room temperature, leading to facile re-suspension of polymer/lipid film in comparison to lipid (phosphatidylcholine, PC) film (only 2-3 min of vortexing) and much more unlabored than block copolymer film (which requires heating of the media to facilitate the process or mixing for 1-2 days [4]).Second step of LUVs preparation were freeze-thaw cycles, which formed unilamellar vesicles from multilamellar ones.For liposomes and hybrids freeze-thaw cycles were needed to obtain unilamellar vesicles with a single peak observed in dynamic light scattering (DLS) after extrusion, but for polymersomes freeze-thaw cycles could be skipped -polymer prefers to organize in single bilayer than forming multilayers (observed by cryo-TEM [5]).In third step, unilamellar vesicles with various sizes were unified by extrusion.Because of the liquid state of PDMS-g-PEO room temperature (22 °C) was sufficient when extruding hybrids or polymersomes (for comparisons, 60 °C is needed for PBD-b-PEO [4]).Furthermore, for extruding PDMS-g-PEO:PC hybrids lower mechanical force was needed than for liposomes with the same mass concentration.The latter might be due to ~8× lower vesicle concentration (determined by TRPS, tunable resistive pulse sensing).After hybrid film rehydration, hybrids had a maximum diameter of 300 nm (verified by DLS) and were extruded through a membrane with 100 nm sized-pores without difficulty.
While there are three main approaches for insertion of MPs into LUVs, i.e. organic solvent-mediated reconstitution, direct incorporation into preformed liposomes and detergent-mediated reconstitution, the last one is commonly used for energy-transducing MP because of its high efficiency and activity retention.The reconstitution is performed in presence of detergent at concentrations of the vesicle saturation point or even higher, i.e. above the vesicle to micelle transition [6].The fluid membranes of liposomes are comparatively easy to solubilize (low detergent concentrations are needed), while the degree of solubilization for reconstitution is determined for the individual MP.On the other hand, in the case of the more rigid membranes of polymersomes, the solubilization requires higher detergent concentrations and stronger detergents (e.g.Triton X-100 for PBD-b-PEO).For efficient micellization of PBD-b-PEO polymer/lipid mixture with sodium cholate 8.6× higher concentrations were used than for reconstitution of bo3 oxidase in PC and PDMS-g-PEO:PC vesicles [4,7].In case of octyl glucoside, 16× higher concentration was needed for reconstitution in block copolymer LUVs than for reconstitution of bo3 oxidase in graft copolymer LUVs [4,7].In cases when micellization of block polymer/lipid mixture is possible, the transition to vesicles after detergent removal is difficult to achieve [4].For hybrids make of PDMS-g-PEO graft copolymer various detergents (including mild ones as sodium cholate, which do not damage sensitive proteins) can be used and only very low concentrations are needed, even much lower than for solubilization of liposomes [7].This is most likely possible due to the intrinsic surfactant proprieties of PDMS-g-PEO [8].

Monitoring respiratory-driven ATP synthesis in proteoLUVs
Measurements of respiration-driven ATP production were performed via monitoring the luminescence of luciferin/luciferase assay.First, 112.8 µl of reaction buffer (20 mM Tris, pH 7.5, 20 mM KH2PO4, 2.5 mM MgSO4, 135 mM sucrose; ~200 mOsmol kg −1 ), 2.26 µl of luciferin/luciferase reagent CLSII, 3.77 µl of 9.96 mM ADP (ultra-pure) and 2.26 µl of proteoLUVs was mixed in 1.5 ml Eppendorf tube by three short burst of vortexing, and baseline was recorded for ~2 min.As standard, 2.26 µl of 2 µM ATP (final concentration 36.16nM) was added and recorded for another ~2 min.To start the reaction, 1.5 µl freshly mixed DTT/Q1 (6 µl 1 M DTT mixed with 0.25 µl 80 mM Q1) was added.When adding ATP and DTT/Q1, the sample was vortexed in three short bursts before continuing the measurement.ATP synthesis was recorded for around 15 min.The ATP production rates (Figure S3) were reported as the average of 3 replicates, with standard deviation.

Preparation of proteo-GUVs
The 0.5 mg of PDMS-g-PEO:soy PS:PE-Rho mixture (70:29.95:0.05mol%) were dissolved in 500 or 495.5 µl of diethyl ether, for the control and for preparation of protein-functionalized GUVs, respectively.Next, bo3 oxidase and F1FO-ATPase were added at the final concentration of 96.3 and 48.2 nM, respectively (to obtain polymer/lipid-to-bo3 oxidase-to-F1FO-ATPase molar ratio of 8,900:2:1), and the mixture was vortexed for 30 sec.Next, 15 µl of those mixtures were spread on each of both glass slides of a homemade electroformation device composed of two electrodes (glass slides coated with indium tin oxide with resistivity of 55 Ω) and a silicone spacer (1.81 mm thick).Solvent was evaporated under a gentle stream of nitrogen and the electroformation chamber was filled with 1 mM Tris (pH 7.5), 200 mM sucrose.A sine wave at 2 V, 10 Hz, was applied for 60 min, followed by detachment step with square wave at 1 V, 2 Hz, for 15 min.

Text S6: Preparation of proteoGUVs Preparation of bo3-F1FO-GUVs
The initial fusion/electroformation experiments were performed with protein-free LUVs.Once the protocol to obtain high yield of GUVs with appropriate size (> 10 µm) was established, proteoGUVs were prepared from proteoLUVs.Varying the dehydration and electroformation steps allowed to obtain quality hybrid GUVs (with diameter of 10-30 µm, high yield, without intravesicular structures, and high protein loading).
Dehydration conditions.For successful fusion of LUV membranes and formation of >20 μm hybrid GUVs, it was crucial to deposit on ITO-coated glass slides a layer of LUVs with optimal thickness.Various vesicle concentrations (0.1-10 mg ml −1 ) and different deposition procedures were tested (spreading 2-20 µl of vesicle suspension on plasma cleaned ITO slides or depositing 0.2-2 µl droplets of vesicle suspension).Deposition of the LUV suspension in droplets was more efficient for vesicle fusion (the DLS peak intensity at 100 nm decreased substantially or even disappeared).Large droplets of relatively highly concentrated vesicles had to be deposited; this can be explained by the lower amount (~8×) of hybrid particles (i.e.LUVs) shown by TRPS, compared to the lipid ones for the same mass concentration.Spreading the droplets or depositing smaller (<1 μl) droplets decreased the fusion efficiency (large amount of LUVs was left in the samples after electroformation).Various dehydration procedures were tested: 2 h and overnight in desiccator (at room temperature and 4 °C) in presence and absence of saturated NaCl [9], and 30-60 min at room temperature.The water removal from hybrid LUVs films was facilitated in comparison to lipid LUV film.This is most likely associated to the different concentration of deposited vesicles.The hybrid LUV film behaved similarly to the polymer film because of the high molar percent of polymer in hybrids (70 mol%).
Electroformation protocol.Various electroformation protocols were tested aiming to obtain a high yield of >10 μm GUVs.With most commonly used protocols for preparation of lipid GUVs in low (10 Hz, 1.1 V, 1-3 h) and high salt buffers (500 Hz, 1.1 V, 1-3 h), hybrid bo3-GUVs were formed with maximum diameter of around 1 μm.With the protocol that we previously used for preparation of hybrid GUVs (conventional electroformation from lipid/polymer film, i.e. sine wave for 40 min (2 V, 10 Hz), followed by square wave for 15 min (1 V, 2 Hz) [7]), larger (>5 µm) protein-free GUVs were formed, but the size of bo3-GUVs was still around 1 µm.Assuming the need of slower initial swelling to prevent the early LUVs film detachment, we applied a protocol consisting of three subsequent steps: first, the voltage was slowly increased for 42 min (starting with 50 mV and increasing in 6 min-steps to 1.1 V), in which the fused membranes film started to swell.Second, swelling and growing continued at constant voltage and third, the GUVs detached at an elevated voltage and decreased frequency.Comparison of hybrid GUVs prepared with the two different electroformation protocols is shown in Figure S5.Extending the second part from 2 h to overnight (ca.12 h) led to increased yield of 20-30 μm GUVs.For the most experiments, the yield was high enough with shorter second swelling part, and to retain protein activity (which was decreasing faster at room temperature), the shorter, 2 h swelling step was preferred.For the overnight protocol, during the first electroformation step chamber was at room temperature, while for next two steps chamber was transferred to an ice box (to avoid moisture contact, chamber was protected with plastic bag).

Figure S3 .
Figure S3.Influence of freeze-thaw cycles on the activity of LUVs with co-reconstituted bo3 oxidase and F1FO-ATPase, measured via respiratory-driven ATP synthesis.ns, not significant for P > 0.05; *P ≤ 0.05.

Figure S4 :
Figure S4: Analysis of protein insertion

Figure S6 .
Figure S6.Respiratory-driven ATP synthesis in hybrid GUVs formed by rehydration under an AC electrical field of a dried film prepared from a solution of polymers, lipids and proteins in diethyl ether.A) An example of ATP measurement in empty and protein-functionalized GUVs: ATP standard added for internal calibration; proton pumping activated by DTT and Q1; arrows indicate additions and vortexing.B) Comparison of ATP synthesis rates in empty and protein-functionalized GUVs.

Figure S7 .
Figure S7.Hybrid GUVs with co-reconstituted bo3 oxidase and F1FO-ATPase used for measurement of ATP synthesis.GUVs were formed by rehydration under an AC electrical field of a dried film prepared from a solution of polymers, lipids and proteins in diethyl ether.Membrane was tagged with 0.05 mol% PE-Rho (red).

Figure S9 .
Figure S9.PDMS-g-PEO:PC (70:30, molar ratio) hybrids size distribution by intensity before and after reconstitution of bo3 oxidase and after dehydration and electroformation, determined by DLS.Time of dehydration was only 30 min, which was insufficient and large portion of LUVs remained unfused.

Figure S17 .
Figure S17.3D of phase separated hybrid GUVs with reconstituted bo3 oxidase-ATTO 425 and F1FO-ATPase-ATTO 620 (magenta) on day 4. Membrane was labeled with PE-Rho (red).ATTO 425 channel is not shown due to high bleaching effect while taking Z-stacks.

Figure S34 :
Figure S34: Fluorescence intensity of proteins in heterogeneous and homogenous hybrid GUVs