A] Preparation of the GUVs of high-melting lipids and cholesterol at physiological conditions displaying lo-ld domains
We have mixed SUVs of DPPC/chol (0.538/0.462) with DOPC-proteoliposomes such that the molar fraction of the lipids in the final mixture (sample I) is DPPC/chol/DOPC (0.35/0.30/0.35). For control experiments we have mixed SUVs of DPPC/chol (0.538/0.462) with DOPC such that the molar fraction of the lipids in the final mixture (sample II) is DPPC/chol/DOPC (0.26/0.22/0.52). The SUVs suspension is mixed inside an eppendorf and frozen in liquid N2, followed by thawing on ice. The freeze-thaw process is repeated a maximum of the three times in the same eppendorf. Many well separated drops of 1 µl of the SUVs suspension is placed on the Pt-wires (inbuilt part of an electroformation chamber) and the chamber is placed inside a desiccator containing saturated NaCl solution with small vacuum sufficient to tight hold the chamber overnight (for max. 12 hrs.) in dark at 4oC. Later the individual partially dried samples are hydrated with the appropriate solvent at ~ 37oC and an a.c. electric field is applied to the Pt-wires. The temperature of the chamber is maintained at 37oC throughout the electroformation process. After the electroformation is completed, the temperature of the chamber is changed slowly to 23oC and fluorescence observations are made at the same temperature. The GUVs are shown in the Fig. 1.
We have prepared GUVs containing proteins of a ternary mixture in the sucrose-buffer (and in 200 mM sucrose for GUVs without NKA) solution and have transferred 50 μl of the GUVs suspension into a fluid observation chamber filled with 950 μl of equi-osmolar glucose-buffer (or glucose) solution for fluorescence observations. It is important to have a well-balanced osmotic pressure of the two sugar solutions in order to minimize the membrane’s thermal conformations. The density difference between sucrose and glucose leads to sinking of the vesicles onto the bottom of the observation chamber. The vesicles were allowed for an hour to settle down and the chamber was covered from top to prevent evaporation and fluid flow during the experiment.
B] Quantification of the compositional variation among GUVs displaying lo-ld domains
After about an hour, we select a quasi-spherical vesicle for taking 3D confocal stacks of the GUVs, at fixed confocal slice thickness and scan speed. Having all the 2D confocal stacks of a GUV from bottom to top, we align these with respect to each other to construct a 3D vesicle body, as described earlier in . In some cases, we require to correct for the vesicle’s motion in the horizontal plane in order to align all the stacks on top of each other, for extracting a radius (r) and centroid (c) of the quasi-spherical vesicle. Individual vesicles are modelled as a sphere with a surface modelled by a triangulated mesh and each mesh-point resembling a volume pixel (voxel). We count the intensity values of the two fluorescence probes (NaP and RhdPE) in each voxel and project it onto the surface of the vesicle to map the areal-intensity counts of NaP and RhdPE. The area fraction of the fluorescence intensity of the two probes represents directly the area fraction of the two membrane phases and is well tested in  for a number of ternary lipid compositions.
Fig. 1b shows the fluorescence intensity counts of the two dyes (corresponding to the two membrane phases) on the surface of a GUV. Using the intensity counts, we estimate the area fraction of the lo phase: A(lo)/A plotted in the Fig. 1c for a batch of vesicles prepared. As a control experiment, we have made an estimate of the lipid compositional state in the GUVs without NKA (Fig. 1d-f) in the lo-ld coexistence region. The A(lo) /A is 0.57 ± 3% (SEM, N=16) for sample I and is 0.387 ± 1% (SEM, N=14) for sample II. The error bars include contributions from experimental and software analysis and constitute an upper bound on compositional variations. In order to compare the values of the area fraction of the lo phase in GUVs prepared by SUVs mixing and by conventional methods (dissolving lipids in chloroform), we have included the color diagram in the Fig. 2a, displaying A(lo)/A for a batch of vesicles prepared by the conventional method, which is taken from . The two samples are indicated on top of the color diagram by black dots in order to highlight the measured A(lo)/A for these ternary compositions and the values closely match with shown in the Fig. 1 for the respective samples. We find the same area fraction as predicted by equilibrium thermodynamics in the GUVs with and without NKA and that it is not affected by the presence of protein in the non-active state with the SUVs fusion method. We have found that the macroscopic lo and the ld domains are visible in GUVs both with (Fig. 1a-c) and without (Fig. 1d-f) NKA. Both of the samples I and II are prepared by mixing SUVs of DPPC/chol (0.538/0.462) and DOPC with different mole fractions of DOPC SUVs and therefore, lie on the same line A1B1, as shown in the Fig. 2a. Moreover the sensitivity of the protocol is clearly indicated by the fact that by changing the moles of DOPC with respect to DPPC/chol, (as for the samples I and II) all the ternary lipid compositions in the coexistence region lying on the line A1B1 can be experimentally accessed without having to prepare new SUVs. The protocol could be generalized to work with three populations of SUVs (or more) to reach an overall composition (x1, x2, x3) in the coexistence region if the molar ratios of the populations are chosen appropriately, as described in .
C] Applications of the protocol to prepare sphingomyelin-chol-DOPC GUVs.
We have applied the above-mentioned protocol to sphingomyelin-containing lipid samples. We have mixed PSM/chol (1/1) with DOPC-proteoliposomes such that the molar fraction of the lipids in the final mixture (sample III) is (1/1/1), and for the control experiments we have used DOPC SUVs instead of proteoliposomes. A GUV of sample III is shown in the Fig. 3a with lateral domain that both the dyes are excluded from the lo phase, as shown in Fig. 3b. A z-stack of the GUV of Fig. 3a is shown in Fig. 3c that confirms the presence of the domain structure. Thus, here we show qualitatively that our protocol works with the sphingomyelin lipid mixture at 37oC. The thawed mixed-SUVs suspensions are left on ice (in 4 oC room) for an extra hour prior to the re-freezing process in liquid N2, for the best results.
D] Protein density and activity measurements in GUVs
The protein content in GUVs of samples I and III is measured and found to be almost 30% with respect to the initial proteoliposomes. The specific hydrolytic activity of reconstituted NKA in GUVs is estimated from the measured hydrolytic activity of NKA with an n-o orientation assuming that the fraction of enzyme with this orientation is preserved from the proteoliposomes, where it is measured to be ~33%. The specific activity in GUVs is found to be below 10±2 µmol⋅mg–1⋅h¬–1 (mean, n = 6) for DPPC at 23℃ and around 5±0.01 µmol⋅mg–1⋅h¬–1 (mean, n = 2) at 42℃ for the PSM mixture, where n denotes the number of independent measurements for different batches of vesicles. A lower activity of NKA has previously been reported in the case where NKA is reconstituted into lo/ld phase-separated proteoliposomes with poly-unsaturated lipids (PUFA) . Based on functional analysis of NKA , structural  and generalized polarisation measurements , it was suggested that a lower hydrolytic activity would result if the NKA localize at the two-phase domain boundaries, a hypothesis that we have verified and confirmed in a recent paper (Bhatia et. al., submitted).
- Formation of GUVs with and without proteins (Timing ~ Overnight + 3-5 hrs)
1a) Dissolve DOPC, DPPC and cholesterol in CHCl3 in 3 separate glass vials (4 ml) at 8.65 mM.
1b) Prepare RhdPE and NaP dyes stock in CHCl3 at 1 mM.
1c) Prepare 2.844788 mM of PSM (1 ml) in CHCl3.
1d) Prepare 1 ml of the following lipid mixtures: DPPC:cholesterol (53.8:46.2) at 8.65 mM containing RhdPE and NaP at 0.4 mole%.
1e) PSM:cholesterol (1:1) at 2.844788 mM containing RhdPE and NaP at 0.4 mole%.
1f) DOPC 8.65 mM (no dye).
- SUVs preparation (Timing ~ 4 hrs): Around a) 500 µL of the 8.65 mM DPPC:cholesterol (53.8:46.2), b) 1 ml of the 2.844788 mM PSM:cholesterol (1:1) and c) 500 µL of the 8.65 mM DOPC (no dye) lipid mixtures solution are individually placed in a flask and chloroform is removed from the sample by using a rotatory evaporator at 50ºC for about an hour. The sample flask is kept in vacuum for about an hour to remove any residual chloroform at room temperature (23ºC). 500 µL of milli-Q water is added to the flask to hydrate the lipids and is mixed using the rotatory evaporator without vacuum-tight conduit at 45℃ (23ºC for DOPC) for about an hour. The hydrated lipid sample is transferred into an eppendorf, and 500 µL of milli-Q water is re-added to the sample flask and mixed, resulting in an overall 1mL volume (measured with a micropipette) of the hydrated lipid sample. A tip-ultrasonicator is used to prepare SUVs of 1 mL of the lipid solutions in water inside a glass vial kept in an ice-bath (to prevent heat-induced chemical degradation of lipids) in the following sequence: one step of sonication for 10 s and break for 5 s at 2 W power, for total sonication and break time of 20 mins and 10 mins respectively. In this way, SUVs of concentration a) 4.325 mM of DPPC:chol (53.8:46.2), b) 2.844788 mM of PSM:cholesterol (1:1) both in water containing RhPE and NaP and c) 4.325 mM of DOPC (no dye) in water are prepared. In case the SUVs have been stored at -20℃, then SUVs are required to be ultrasonicated with the same protocol as mentioned above and after sonication SUVs are extruded in the liquid phase, using membrane filters.
3a) For sample I: DOPC-proteoliposomes are mixed with 4.325 mM DPPC/cholesterol (53.8/46.2) SUVs in the total molar ratio DPPC/cholesterol/DOPC (0.35/0.3/0.35), and for sample II: DOPC SUVs are mixed with 4.325 mM DPPC/cholesterol (53.8/46.2) SUVs in the total molar ratio DPPC/cholesterol/DOPC (0.26/0.22/0.52) and for sample III: DOPC SUVs are mixed with 2.844788 mM PSM/cholesterol (1/1) SUVs in the total molar ratio PSM/cholesterol/DOPC (1/1/1).
3b) The mixed SUVs suspension is freeze-thawed max. 3 times and many 1 μl drops of the SUVs suspension are coated on the Pt-wire being well separated.
3c) The sample is dehydrated in a desiccator with minimal pressure drop inside enough to hold the container tight during incubation, overnight ( 12 hrs. max.). For protein samples, the desiccator is kept in dark at 4ºC for partial dehydration of SUV suspension deposits in the presence of saturated NaCl-salt solution. (Timing 12 hrs. max.).
- 200 mM sucrose or sucrose buffer preheated at 37oC is introduced in the electroformation chamber and an a.c. electric field is applied. 4b) For GUVs with no proteins, the a.c. field is applied in the following sequence: 10 Hz (0.2 VPP for 5’, 0.5 VPP for 10’, 1 VPP for 20’, 1.5 VPP for 20’, 2 VPP for 30’), 4 Hz, 2 VPP for 30’. (Timing ~ 2 hrs.).
4c) For GUVs containing proteins, the a.c. field is applied in the following sequence: 50 Hz (0.15 VPP (peak-to-peak voltage) for 2 mins, 0.3 VPP for 5 mins, 0.6 VPP for 5 mins, 1 VPP for 10 mins, 1.5 VPP for 10 mins, 2 VPP for 15 mins, 3 VPP for 120 mins), 10 Hz (2 VPP for 45 mins). (Timing ~ 4 hrs.).
4d) After electroformation, the temperature is changed at a rate of 0.02o C/min to 23oC.
4e) The GUVs can be stored at R.T. (and at 4oC for containing proteins) for later use.
- CRITICAL STEPS
a) SUVs are unstable and form large unilamellar vesicles (LUVs) by fusion. The mixing protocol is reproducible only for SUVs and not for LUVs.
b) If the mixed SUVs suspension is thawed faster by increasing temperature then the results are not reproducible.
c) If the mixed SUVs suspension is completely dried on Pt-wires (due to an reduction in the saturated vapour pressure) and not partially dried then the GUVs do not form.
d) If the thawed mixed SUVs suspensions containing PSM are left for an extra hour on ice (in the 4 o C room) before the re-freezing process in the liquid N2, the results are the best. The suspension is also mixed by using a pipette before freezing.
- Acquiring 3D confocal stacks of GUVs (Timing ~ max. 2 hrs)
a) The fluid chamber is filled with 350 μl osmotically matched glucose solution (buffer in the case of GUVs containing proteins) and then we add 50μl of the GUVs suspension at the centre of the chamber giving a total volume of 400μl of GUVs suspension and solvent, as shown in Fig. 1.
b) CRITICAL STEP
It is important that the osmotic pressure of glucose solution and GUVs suspensions match. Usually, the osmotic pressure of the GUVs suspension is slightly different from the solvent in which these are prepared after electroformation. If the GUVs are floppy then membrane undulations are enhanced.
c) GUVs are allowed to settle down for about an hour.
d) Quasi-spherical GUV are selected and its bottom and top edge is found by moving the focus manually, in small steps.
e) CRITICAL STEP It is important to choose a bidirectional scan speed fast enough which allows for confocal stack acquisition of all the 2D cross-section of the GUV from bottom to top. Having all the 2D cross-sections we can combine those to construct a 3D vesicle body.
- Measuring protein density and activity in active GUVs (Timing ~ 1 day)
a) NKA activity and density in the active GUVs are measured according to the Baginsky  and to the Peterson’s modification of the Lowry method [16, 17].
b) The specific hydrolytic activity of reconstituted NKA in GUVs is estimated from the measured hydrolytic activity of NKA with a n-o orientation assuming that the fraction of enzyme with this orientation is preserved from the proteoliposomes, where it is measured to be ~33 %.