GPMV measurements

Rat basal leukemia (RBL)-2H3 cells (23) were maintained in minimum essential media with 20% fetal bovine serum and 0.1% gentamycin at 37°C in 5% CO₂. XTC-2 cells (24) were maintained in L-15 media diluted 1:1.5 with water for amphibian cells with 10% fetal bovine serum, sodium bicarbonate (2.47 g/L), pen strep (100 units/mL) at room temperature in 5% CO₂. Freshly seeded cells were incubated in complete media for at least 18 h at the growth temperature indicated before GPMV isolation. All culture reagents were purchased from Thermo Fisher Scientific (Waltham, MA). Other reagents were purchased from Sigma Aldrich (St. Louis, MO) at the highest available purity unless otherwise indicated.

GPMVs from RBL cells were prepared through incubation with low concentrations of dithiothreitol (DTT, 2 mM) and formaldehyde (25 mM) in the presence of calcium (2 mM) for 1 h, as described previously (21). GPMVs isolated from RBL cells contain 20–40 mol % cholesterol, sphingolipids, phospholipids, and gangliosides typically associated with the plasma membrane, along with many transmembrane and peripheral plasma membrane proteins (18, 25, 26). For XTC-2-derived GPMVs, the vesiculation buffer was diluted 1:5 while maintaining calcium, formaldehyde, and DTT concentrations, and cells were incubated for at least 2 h at room temperature. Before GPMV formation, cells were labeled with DiI-C12 (Life Technologies, Carlsbad, CA; 2 μg/mL in 1% methanol) for 10 min at room temperature. GPMVs probed at atmospheric pressure were imaged on an inverted microscope (IX81; Olympus, Center Valley, PA) with a 40× air objective (0.95 NA), and epi-illumination using an Hg lamp and Cy3 filter set (Chroma Technology, Bellows Falls, VT). Temperature was controlled using a home-built Peltier stage, described previously (21), coupled to a proportional-integral-derivative controller (Oven Industries, Mechanicsburg, PA), and images were recorded using an sCMOS camera (Neo; Andor, South Windsor, CT).

GPMV suspensions with hexadecanol were prepared using either supersaturated solutions or equilibrated solutions. To make supersaturated solutions, hexadecanol was suspended in either dimethylsulfoxide (DMSO) or ethanol using volumes corresponding to the final concentrations indicated in the figures, then mixed directly with the GPMV suspension in aqueous buffer while mixing. The maximum DMSO concentration used was 0.5% v/v, and previous work demonstrates that T_c is not affected by DMSO treatment alone (21). Equilibrated solutions were prepared by first adding a concentrated hexadecanol stock directly to the GPMV suspension such that it precipitated out of solution. Then, the desired volume of ethanol was added and the solution mixed to facilitate the resuspension of hexadecanol. In all cases where ethanol and hexadecanol were added in combination, they were maintained at a 120 mM ethanol to 30 μM hexadecanol ratio, as indicated in figures. Equilibrated solutions could be made for 60 mM:1.5 μM, 120 mM:3 μM, 180 mM:4.5 μM, and 240 mM:6 μM (ethanol/hexadecanol ratio), although for higher concentrations some small fraction of hexadecanol lacked solubility. Only supersaturated solutions could be made for 600 mM:15 μM ethanol/hexadecanol.

Measurements conducted at elevated hydrostatic pressures were made using a custom-built microscopy-compatible pressure cell mounted on a Nikon Eclipse TE2000-E inverted microscope as described previously (27, 28) with a 20× extra-long-working-distance air objective (0.4 NA) and G2-A filter set (Nikon Instruments, Richmond, UK) The pressure cell temperature was controlled via a circulating water bath. Images were acquired using an sCMOS based camera (Zyla; Andor, Belfast, United Kingdom) and recorded using custom-built software with temperature and pressure logging.

GPMV transition temperatures at constant pressure were measured as described previously (21) and as illustrated in Fig. 1 A. Briefly, images were acquired of fields of GPMVs over a range of temperatures such that at least 100 vesicles were detected at each temperature. After imaging, individual vesicles were identified as having a single liquid phase or two coexisting liquid phases. This information was compiled into a plot showing the percentage of vesicles with two liquid phases as a function of temperature, which was fit to a sigmoid function to extrapolate the temperature at which 50% of vesicles contained two coexisting liquid phases, % phase separated = 100 × (1 − 1/(1 + e^{−(T−T_c)/B})), where B is a parameter describing the width of the transition. The width of the transition within a population of vesicles ( ∼ 10°C) is much broader than the width of the transition for a single GPMV (<2°C) (19), most likely due to heterogeniety in composition between GPMVs (29).

Determination of the average critical temperature or pressure of DiIC₁₂-labeled GPMVs through fluorescence imaging. (A) Fields containing multiple GPMVs were imaged over a range of temperatures and at fixed pressure, with representative subsets of images shown on the right. At high temperatures, most GPMVs appear uniform, whereas an increasing fraction of vesicles appear phase separated as the temperature is lowered, with phase-separated vesicles indicated by yellow arrows. From these images, we manually tabulate the fraction of GPMVs that contain two coexisting liquid phases as a function of temperature, constructing the plot on the left. These points are fit to the sigmoid function described in Materials and Methods to determine the extrapolated temperature at which 50% of vesicles contain coexisting liquid phases. (B) Fields containing multiple GPMVs were imaged over a range of pressures at fixed temperature, and representative subsets of images are shown on the right. At low pressure, most GPMVs appear uniform, whereas an increasing fraction of vesicles appear phase separated as pressure is increased. As with the fixed-pressure data in (A), these points are fit to the sigmoid function described in Materials and Methods to determine the extrapolated pressure at which 50% of vesicles contain coexisting liquid phases. To see this figure in color, go online.

We have previously demonstrated that these GPMVs pass through a critical temperature at the transition, even in the presence of anesthetics. Therefore, we refer to this temperature as the average critical temperature (T_c) of the sample. Errors in single measurements of T_c (σ_Tc) are 68% confidence-interval estimates of this parameter determined directly from the fit. For the example shown in Fig. 1 A, the T_c is 19.5 ± 0.6°C. We generally observe that average critical temperatures measured in this way vary in the range 12°C < T_c < 27°C for untreated RBL-derived GPMVs prepared as described above, depending on the growth conditions of the cells from which the GPMVs were derived (29). XTC-2 cells produced GPMVs with average critical temperatures of 14.8 ± 0.4° and 20.7 ± 0.4°C. Error bounds for a transition temperature shift (ΔT_c) are given by σM2+σC2, where σ_C is the error in measuring the T_c of the untreated control and σ_M is the error in measuring the T_c of the treated sample. The average critical pressure in the sample (P_c) at constant temperature was determined by a similar procedure, as illustrated in Fig. 1 B, although fit to a slightly different equation, % phase separated = 100 × (1/(1 + e^{−(P−P_c)/B})). As with previous studies in which only temperature was varied (19), we find that the fraction of phase-separated vesicles present in a population of vesicles at a given temperature/pressure does not depend on the order in which temperatures or pressures are sampled, suggesting that the transition is fully reversible.