Vesicles Balance Osmotic Stress with Bending Energy That Can Be Released to Form Daughter Vesicles

The bending energy of the lipid membrane is central to biological processes involving vesicles, such as endocytosis and exocytosis. To illustrate the role of bending energy in these processes, we study the response of single-component giant unilamellar vesicles (GUVs) subjected to external osmotic stress by glucose addition. For osmotic pressures exceeding 0.15 atm, an abrupt shape change from spherical to prolate occurs, showing that the osmotic pressure is balanced by the free energy of membrane bending. After equilibration, the external glucose solution was exchanged for pure water, yielding rapid formation of monodisperse daughter vesicles inside the GUVs through an endocytosis-like process. Our theoretical analysis shows that this process requires significant free energies stored in the deformed membrane to be kinetically allowed. The results indicate that bending energies stored in GUVs are much higher than previously implicated, with potential consequences for vesicle fusion/fission and the osmotic regulation in living cells.


GUV preparation.
GUVs were prepared on indium tin oxide (ITO)-coated coverslips (30-60 Ohms/Sq, Sigma-Aldrich) by the electroformation method 1, 2 in a fluidic flow channel. 2 By using this fluidic flow channel we can easily exchange solutions and also observe free vesicles that do not attach or sediment on surfaces, which excludes phenomena induced by surface adhesion. Stock solutions of either DOPC or DMPC with 5 mol% cholesterol, or DMPC with 30 mol% cholesterol were prepared in chloroform/methanol (9:1 volume ratio) at a concentration of 0.2 mg/ml. To all lipid samples, 0.5 mol% the fluorescent lipid analogue Liss Rhod PE (red) was added. To prepare GUVs, the ITO-coated coverslips were first cleaned with ethanol and dried by nitrogen gas. 10 µL lipid solution was then deposited onto the conductive side of the ITO-coated coverslip and dried in a vacuum chamber overnight. The lipid coated coverslip was then mounted to the adhesive underside of a microchannel (Ibidi sticky-Slide VI 0.4). Another ITO-coated coverslip was attached to the top side of the microchannel with the conductive side towards sample solution. Next, conductive wires were used to connect the conductive sides of the two ITO-coated coverslips to the electrodes from the frequency generator. The AC electric field (10 Hz, 3V) was applied for 3 hours to generate the GUVs. The GUVs were prepared in the fluid state, DOPC at 20 °C, DMPC with cholesterol at 28 °C. DOPC GUVs were prepared in the liquid disordered phase, and DMPC/chol GUVs in the liquid ordered phase. 3 Experimental procedure. GUVs composed of either DOPC or DMPC/chol were prepared using the electroformation method. After completing the vesicle preparation, the GUVs were observed by CLSM. The preparation of GUVs and CLSM observation were carried out at 20 °C (DOPC) or 28 °C (DMPC/chol). Hypertonic osmotic gradients were generated by adding glucose or PEG2000 stock solutions to the fluidic flow channel with the prepared GUVs. Typically, 6.5-29 µL of 100 mM glucose were added to the sample (150 µL) to generate an osmotic gradient of 4-16 mM (0.1-0.4 atm). For proper mixing after injection, the fluidic flow channel cell was gently rotated several times. Due to the mixing, it was typically not possible to image the same vesicle at all steps of the experiments, and all figures therefore show representative images for the different conditions. After this, the two wells of the fluidic channel were sealed by parafilm to prevent water evaporation. The sample was then observed by using CLSM. For each condition investigated (step-wise increase in osmotic gradient and then addition of water at a fixed osmotic gradient) the experiment was repeated at least 3 times.

Confocal Laser Scanning Microscope.
The fluorescent GUVs were observed by confocal laser scanning microscope (CLSM, Leica SP5) operated in the inverted mode (D6000I). The temperature of the samples was controlled with an accuracy of 0.2 °C by mounting the CLSM to a thermostated enclosure. Samples were equilibrated for 2 hours at 20 or 28 °C before observation. The red fluorescence of the membrane lipid analogue (Liss Rhod PE) and the green fluorescence probe, Alexa Fluor™ 488 NHS Ester (Alexa488), were excited by using a HeNe laser at 543 nm and an argon-ion laser at 488 nm, respectively.
A time-lapse sequence of confocal fluorescence images (Movie S1) and the image stack from top-tobottom z-axis scans (Movie S2) clearly shows that the daughter vesicles formed inside DOPC GUVs are separated from the mother vesicle.

Movies
Movie S1. A time-lapse sequence of confocal fluorescence images of Liss Rhod PE labeled DOPC vesicle (red) after reversing osmotic gradient by rinsing with water and the fluorophore Alexa488 (green).

Movie S2.
Top to bottom z-axis scans of an unlabeled DOPC vesicle after reversing the osmotic gradient by rinsing with water and the fluorophore Alexa488 (green).      . Aspect ratio between the long (LL) and short axis (LS) of prolate DOPC vesicles present at 0.4 atm for vesicles of different sizes. The membrane area of the prolate vesicles were calculated using Eq.S1. Assuming that the membrane area is unchanged upon deformation, the corresponding radius of the original spherical vesicles was then calculated. 70 vesicles were analyzed. (S1)