1. Introduction
Biomembranes are primarily composed of amphiphilic phospholipid molecules organized in a bilayer structure. It is believed that the various components of biomembranes, such as phospholipids, cholesterol, and membrane proteins, are not uniformly mixed but rather form spatially heterogeneous raft structures. Because these lipid rafts may be involved in signal transduction and membrane trafficking, elucidation of raft formation is crucial for understanding cellular functions [
1,
2]. Bilayer membranes and vesicles (spherically closed bilayer membranes), consisting of various types of phospholipids, have been applied as model systems for biomembranes, and compositionally heterogeneous structures have been reported to emerge spontaneously under certain conditions. This phenomenon can be physicochemically interpreted as phase separation, and studies of the phase separation in lipid membranes in relation to raft formation are underway [
3].
Lipid membranes composed of saturated phospholipids, unsaturated phospholipids, and cholesterol are commonly used in phase separation studies. In these systems, a liquid-ordered (L
o) phase rich in saturated lipids and cholesterol and a liquid-disordered (L
d) phase rich in unsaturated lipids coexist, and the L
o phase is regarded as a model of lipid rafts [
4,
5,
6]. The interfacial tension at the phase-separated domain boundary, which is referred to as line tension, is an important parameter when considering the stability of the phase separation, and control over the line tension may allow control over the phase-separated structures. The line tension can be measured from the thermal fluctuation of the phase-separated domain boundary [
7,
8,
9] or by micropipette aspiration [
10]. Experimental measurements of the line tension near the critical point [
11,
12] and in the presence of hybrid lipids [
13,
14,
15,
16], vitamin E [
17,
18], and local anesthetics [
19] have been performed.
Unsaturated fatty acids have also been examined from the perspective of line tension control. These compounds have attracted attention, largely owing to their association with health. For example, polyunsaturated fatty acids (PUFAs) present in foods, such as docosahexaenoic acid and eicosapentaenoic acid, may reduce the risk of lifestyle diseases, such as type 2 diabetes mellitus (T2DM) [
20] and cardiovascular disease [
21]. Moreover, free fatty acids and cholesterol are also related to obesity, and abnormal lipid metabolism, as well as T2DM [
22]. In particular, it was suggested that cholesterol catalyzes the aggregation of islet amyloid polypeptides, which are linked to T2DM [
23,
24]. In addition to PUFAs, monounsaturated fatty acids (MUFAs), such as oleic acid, may exert a beneficial effect on health by increasing the level of high-density lipoprotein cholesterol [
25]. Although the effects of unsaturated fatty acids on membrane organization (e.g., phase separation) are not well understood, it is known that the biomembrane composition, which plays a crucial role in membrane organization, is influenced by the diet, including cholesterol and fatty acid intake. Furthermore, the phase behavior of fatty-acid-containing lipid membranes has been observed experimentally, and oleic acid was found to significantly decrease the line tension at the domain boundary [
26].
The line tension is also affected by the osmotic pressure. Because a lipid membrane is semipermeable, osmotic pressure is generated upon sandwiching the membrane between solutions of different solute concentrations. Consequently, vesicles shrink or swell depending on the solute concentration gradient across the membrane. In the body, osmotic pressure plays essential roles in metabolism, growth, development, and fluid homeostasis. Furthermore, the membrane lateral tension, resulting from the osmotic pressure, influences the activity of mechanosensitive channels in the membrane [
27,
28]. For example, in a hypotonic solution, where the external solute concentration is lower than the internal solute concentration, vesicles swell, owing to the water inflow and the resulting membrane lateral tension promotes the formation of phase-separated structures and increases the line tension [
29,
30,
31,
32]. The swelling of a vesicle also suppresses its thermal membrane fluctuation and decreases the entropy associated with membrane fluctuation [
33]. In general, a homogeneous fluid membrane strongly fluctuates as a result of its flexibility, whereas a phase-separated membrane fluctuates significantly less because of the presence of the rigid ordered phase. The homogeneous membrane is strongly affected by the fluctuation suppression due to swelling and becomes unstable. Therefore, under osmotic stress, the phase-separated structure is relatively stable, the phase separation is promoted, and the line tension increases.
The addition of molecules such as unsaturated fatty acids is a chemical approach to controlling the line tension, whereas the application of osmotic pressure is a physical approach. In this study, we combine these two approaches to achieve control over the line tension. As the unsaturated fatty acids, three MUFAs featuring different chain lengths, namely, palmitoleic acid, oleic acid, and eicosenoic acid, were used. First, the phase separation of the MUFA-containing lipid membranes was observed by fluorescence microscopy. From the domain boundary fluctuation, we quantitatively measured the line tension in the absence and presence of osmotic pressure. The effects of the MUFAs and osmotic pressure on the shape of the vesicles were examined by confocal laser scanning microscopy. In addition, we examined the interaction between lipids and the MUFAs by differential scanning calorimetry (DSC). Finally, we discuss the mechanisms of line tension control by the addition of MUFAs and the application of osmotic pressure.
2. Materials and Methods
2.1. Materials
The saturated phospholipid 1,2-dipalmitoyl-
sn-glycero-3-phosphocholine (DPPC), the unsaturated phospholipid 1,2-dioleoyl-
sn-glycero-3-phosphocholine (DOPC), and cholesterol (Chol) were purchased from Avanti Polar Lipids (Alabaster, AL, USA). Rhodamine B 1,2-dihexadecanoyl-
sn-glycero-3-phosphoethanolamine (triethylammonium salt; Rho-DHPE) and 1,2-dipalmitoyl-
sn-glycero-3-phosphoethanolamine-
N-(7-nitrobenz-2-1,3-benzoxadiazol-4-yl) (ammonium salt; NBD-PE) were obtained from Thermo Fisher Scientific (Waltham, MA, USA) and Avanti Polar Lipids, respectively. Oleic acid (C18:1,
cis-9; OA), palmitoleic acid (C16:1,
cis-9; PaA), and
d-(+)-glucose were acquired from Nacalai Tesque (Kyoto, Japan). Eicosenoic acid (C20:1,
cis-9; EiA) was obtained from Larodan Fine Chemicals (Malmö, Sweden). Ultrapure water (specific resistivity ≥ 18 MΩ·cm) was obtained from a Millipore Milli-Q purification system (Burlington, MA, USA). The chemical structures of the lipids and MUFAs used in this study are shown in
Figure S1.
2.2. Vesicle Preparation
Vesicles were prepared using the natural swelling method. The phospholipids, cholesterol, and MUFAs (0.2 mM) and the fluorescent probes (Rho-DHPE and NBD-PE, 0.1 μM) were dissolved in chloroform to obtain stock solutions, which were then mixed in glass test tubes to afford the desired compositions. Lipid films were prepared by evaporating the organic solvent under a flow of nitrogen gas, then dried under vacuum for 3 h. The films were then pre-heated using 5 µL of Milli-Q water for 10 min at 55 °C and hydrated using 200 mM glucose solution for 3 h at 37 °C.
In the osmotic pressure experiment, swollen vesicles were obtained by mixing the vesicle solution with Milli-Q water in a ratio of 1:9. This dilution step reduced the glucose concentration outside of the vesicles from 200 to 20 mM, such that the difference in glucose concentration between the inside and outside of the vesicles was ΔC = Ci − Co = 180 mM, where Ci and Co denote the glucose concentrations inside and outside of the vesicles, respectively. Consequently, water influx occurred, owing to the concentration gradient across the lipid membrane.
2.3. Microscopy
The vesicle solutions were placed on a glass slide and covered with a smaller glass coverslip, with a spacing of ca. 0.1 mm. The vesicle solutions were observed at room temperature under a fluorescence microscope (IX71, Olympus, Tokyo, Japan) and a confocal laser scanning microscope (FV1000, Olympus, Tokyo, Japan). Rho-DHPE and NBD-PE were used as fluorescent dyes. The fluorescence of Rho-DHPE was monitored using a standard U-MWIG filter set, with an excitation wavelength (λex) of 530–550 nm and an emission wavelength (λem) of 575 nm. The fluorescence of NBD-PE was monitored using a U-MNIBA filter, with a λex of 470–495 nm and a λem of 510–550 nm. The observation time for each slide was restricted to 90 s to avoid the occurrence of photo-oxidation.
2.4. Flicker Spectroscopy of Domain Boundary Fluctuation
The line tension (
) at a fluctuating domain boundary was determined by analyzing a movie of a phase-separated fluctuating domain (100×, objective lens) recorded over > 3 s (>90 frames, 30 frames/s). The pixel size of the obtained microscopic images was 100 nm × 100 nm. By binarizing the obtained movie with ImageJ, the radial fluctuation of the domain was identified. The domain radius (
r) was plotted as a function of the polar angle (
) [
7,
8,
9] and expressed in terms of a Fourier series expansion, as shown by the following equation:
where
is the average domain radius,
is the mode number, and
and
are the Fourier coefficients. The relationship between the fluctuation and the excess free energy can be written as
where
is the line tension. The free energy for each independent mode is
from the generalized equipartition theorem, where
is the Boltzmann constant and
is the absolute temperature. Therefore, we can obtain
where
is the average value for all images. The line tension
can be obtained by fitting the experimental data with Equation (3). Five to ten fluctuating domains were analyzed for each composition and condition. The details of the analysis are provided in
Figure S2.
2.5. Differential Scanning Calorimetry
DPPC, cholesterol, and the MUFAs were dissolved in chloroform at a concentration of 300 mM. Lipid films were then prepared by mixing these components in the desired ratio with a total volume of 30 μL. The chloroform was removed under nitrogen gas flow and the samples were dried in a vacuum desiccator for 3 h. For the hydration step, 60 μL of Milli-Q water was deposited onto the film to obtain a final concentration of 150 mM. The solutions were mixed and homogenized using a vortex mixer and sonicated at 60 °C for 1 h to produce the vesicles from the bottom of the tube. All of the thermographs were recorded on a DSC822 system (Mettler Toledo, Switzerland) from 20 to 60 °C, with a heating/cooling rate of 5 °C/min. Approximately 12 μL of the homogenized sample was placed into the aluminum sample pan. The same weight of Milli-Q water was used in the reference cell. Each sample was measured over three cycles of heating and cooling and each measurement was repeated at least three times to ensure the reproducibility of the results.
The obtained asymmetric thermographs were assumed to be described by a linear combination of two independent transitions. The thermographs were deconvolved by performing Gaussian two-peak fitting analysis in OriginPro 2018.
4. Discussion
We found that the addition of PaA or OA decreased the line tension. Because PaA and OA are MUFAs, they were mainly partitioned into the DOPC-rich Ld phase. However, a limited amount of PaA or OA, which are smaller than phospholipids, could also be incorporated into the DPPC-rich Lo phase. As demonstrated by the DSC results, PaA and OA interacted with DPPC unfavorably. The small amount of PaA or OA slightly disturbed the DPPC chain ordering, and the physical property differences between the Lo and Ld phases (e.g., chain ordering, membrane thickness, and spontaneous curvature) became smaller. As a result, the line tension decreased upon the addition of PaA or OA.
We observed S
o/L
d phase separation in the EiA-containing lipid membranes. Moreover, in the DSC measurements, the lower- and higher-temperature peaks did not change very much as the EiA content increased. The formation of the S
o phase means that the amount of Chol in the DPPC-rich phase has decreased. This suggests that EiA interacts more favorably with DPPC than Chol. Since EiA has a double bond, EiA may disturb the chain ordering of DPPC. On the other hand, EiA has a longer hydrophobic chain than PaA and OA, and it is expected to have a sufficiently strong hydrophobic interaction between DPPC and EiA. Therefore, EiA interacted more favorably with DPPC than Chol, and less Chol interacted with DPPC. As a result, the S
o phase formation was induced. To make this point clear, it will be important to investigate the localization of EiA and Chol in the phase-separated membranes by using specific fluorescent probes and spectroscopy. The experimental results were summarized in a schematic illustration in
Figure S4.
The application of osmotic pressure suppressed the membrane fluctuation. Therefore, the osmotic stress promoted the formation of a rigid phase with relatively small membrane fluctuation [
29]. PaA and OA disturb the chain ordering of the L
o phase, resulting in large membrane fluctuation. Such behavior was not favored in the presence of osmotic stress. Thus, the amount of PaA or OA in the L
o phase decreased to prevent the membrane from undergoing large fluctuations, and the line tension reduction due to PaA or OA addition became smaller upon the application of osmotic stress. In the case of EiA, the rigid S
o phase was formed. Because this is the same trend as the stabilization induced by the application of osmotic pressure, it is likely that no osmotic-pressure-induced changes were observed. In the present study, we considered the localization of MUFAs in phase-separated membranes based on the results of fluorescence microscopy and DSC measurements. In the future, it will be important to investigate the localization directly using MUFA-specific fluorescent probes and infrared spectroscopy.
In addition, we have discussed the effects of the osmotic pressure mainly based on the suppression of the membrane fluctuation. Phase separation is usually interpreted in terms of the lipid interactions and the mixing entropy. The suppression of the membrane fluctuation is the suppression of the vertical movement of the membrane, while the mixing entropy and interactions within membranes are lateral. Therefore, we believe that the changes in the interactions and mixing entropy due to the osmotic stress are limited. However, it will be important to consider what extent they actually change with the osmotic pressure, both experimentally and theoretically.
It is interesting that the phase behavior for PaA and OA was markedly different from that for EiA, despite only small structural differences between the three MUFAs. MUFAs disturb the chain ordering of DPPC, owing to the presence of the double bonds, although the greater chain length of EiA may also lead to strong hydrophobic interactions. PaA and OA have smaller chain lengths, and the disruptive influence of the unsaturated chains presumably outweighed the hydrophobic attraction. In other words, in cis-9 MUFAs, the phase behavior of lipid membranes (DOPC/DPPC/Chol) appears to change significantly between C18 (OA) and C20 (EiA). In the future, it would be valuable to systematically investigate the phase behavior for MUFAs with different double bond positions.
For PaA and OA, the vesicles with liquid domains became polyhedral when the line tension decreased. In the case of EiA, the vesicles with solid domains became polyhedral, even though the line tension did not decrease. These deformation mechanisms have been explained by a theoretical model [
26]. When the vesicle shape is described by the curvature energy, line energy at the phase-separated domain boundary, and membrane surface energy, the domains should be flat to decrease the curvature energy. However, the formation of one flat domain greatly deforms the vesicle shape and increases the membrane surface energy. Therefore, the energy is reduced by increasing the number of domains and making the vesicle more spherical. Although the line energy increases upon increasing the number of domains, it does not increase markedly for a small line tension. Consequently, a polyhedral vesicle with many flat domains is formed. Next, although EiA did not reduce the line tension, we also observed polyhedral vesicles for the EiA-containing membranes. This occurred because EiA transforms the L
o phase into the S
o phase, with a dramatic increase in the bending rigidity. To decrease the curvature energy, the domains become flat. This behavior, resulting in rigid and flattened domains, was also observed for a saturated fatty acid and a
trans fatty acid [
26], with which EiA displayed similar behavior. Upon the application of osmotic pressure, these polyhedral vesicles disappeared, owing to the large energetic penalty associated with a non-spherical shape in the case of high internal pressure.
5. Conclusions
We have examined the phase behavior of MUFA-containing lipid membranes and measured the line tension at the phase-separated domain boundary by flicker spectroscopy. PaA and OA significantly decreased the line tension, resulting in domain boundary fluctuation. Moreover, the line tension increased in the presence of osmotic stress for both MUFA systems. On the other hand, EiA did not reduce the line tension, but rather transformed the Lo phase into the So phase. Because of the line tension reduction for PaA and OA and the domain rigidification for EiA, the vesicle shape became polyhedral. The obtained results were supported by the DSC measurements.
The line tension is an important factor that governs the stability of phase separation. Under isothermal conditions, we found that the line tension can be decreased or increased by the addition of MUFAs or the application of osmotic stress. By combining these two approaches, it may be possible to precisely control the line tension without changing the temperature.