Nanodomain formation in lipid bilayers II: The influence of mixed-chain saturated lipids

Abstract

An important class of lipids found in biological membranes are composed of two structurally different hydrocarbon chains. Among these, low-melting lipids possessing both a saturated and unsaturated chain have been intensely studied because of their biological abundance and influence on lipid rafts. In contrast, much less is known about the biophysical effects of mixed chains in high-melting lipids. Here, we investigated two such lipids-MSPC (14:0-18:0 PC) and SMPC (18:0-14:0 PC)-to determine how chain length mismatch and acyl chain position on the glycerol backbone influence lateral organization. We studied the temperature- and composition-dependent phase behavior of liposomes composed of either mixed-chain or symmetric-chain high-melting lipids plus DOPC and cholesterol, using techniques sensitive to domain formation at both microscopic and nanoscopic length scales. All studied mixtures exhibited liquid-ordered (Lo) + liquid-disordered (Ld) phase coexistence with domains that were visible in confocal microscopy experiments. FRET measurements showed that all mixtures also exhibited nanoscopic heterogeneity at temperatures above the microscopic miscibility transition temperature, and cryo-EM imaging further revealed bilayer thickness variation consistent with coexisting Ld and Lo phases. Both the microscopic miscibility transition temperature, $\text{μm-}{T}_{mix}$ , and its nanoscopic counterpart, $\text{nm-}{T}_{mix}$ , were strongly correlated with the melting transition temperature of the saturated lipid; the sole exception was SMPC/DOPC/Chol, whose $\text{μm-}{T}_{mix}$ showed a significant negative deviation from the expected value, implying an enhanced propensity for nanoscopic phase separation in mixtures containing this high-melting species. These results point to strong effects of acyl chain position within mixed-chain high- $T_M$ lipids on the microscopic phase behavior of ternary mixtures.

Keywords: chain length mismatch; cryo-EM; lipid raft; liquid-disordered; liquid-ordered.

Figure 1.

High-$T_M$ lipids in this study. From left to right: DMPC (14:0–14:0 PC); 15:0-PC (15:0–15:0 PC); DPPC (16:0–16:0 PC); DSPC (18:0–18:0 PC); SMPC (18:0–14:0 PC); and MSPC (14:0–18:0 PC). $T_M$ values are taken from Koynova and Caffrey [42].

Figure 2.

Confocal fluorescence images of GUVs at 22 °C. Shown are slices near the GUV cap for three-component mixtures containing 1:1 high-$T_M/\text{DOPC}$ with increasing Chol mol% and different high-$T_M$ lipid as indicated: MSPC (top row); 15:0-PC (middle row); and SMPC (bottom row). GUVs were labeled with fluorescent dyes LRPE (red) and NBD-DSPE (green). Each image shows the composite of the red and green channels. Scale bars are 5 μm.

Figure 3.

Phase behavior of GUVs at 22 °C. Shown are ternary phase diagrams for: MSPC/DOPC/Chol (a); 15:0-PC/DOPC/Chol (b); and SMPC/DOPC/Chol (c). Each point in the diagram shows the predominant phase behavior at a specific composition: Ld+Lβ (blue triangles), Ld+Lo (green circles), uniform mixing (red squares).

Figure 4.

Microscopic miscibility transition temperature ($\text{μm-}{T}_{mix}$) determined from confocal microscopy of GUVs. (a) The fraction of phase separated vesicles vs. temperature (open symbols) for ternary mixtures composed of high-$T_M$ lipid/DOPC/Chol 40/40/20 mol% for the indicated high-$T_M$ lipids. Data for one of three experimental replicates is shown overlaid with a fitted sigmoidal function (solid lines). The vertical dashed lines indicate the midpoint of the sigmoidal function, $\text{μm-}{T}_{mix}$. (b) $\text{μm-}{T}_{mix}$ for ternary mixtures vs. the chain melting temperature, $T_M$, of the saturated lipid. Average values and uncertainties obtained from sample replicates (N = 3) are listed in Table 2. The dashed line is a linear fit (R² = 0.998) to the data points corresponding to symmetric chain high-$T_M$ lipids (i.e., DMPC, 15:0-PC, DPPC, and DSPC). Error bars are SEM determined from sample replicates. $T_M$ values are taken from Koynova and Caffrey [42].

Figure 5.

Cryo-EM images of ternary LUVs. All compositions are 40/40/20 mol%: DPPC/DOPC/Chol (a); MSPC/DOPC/Chol (b); 15:0-PC/DOPC/Chol (c); and SMPC/DOPC/Chol (d). Shown is a representative vesicle for each composition, without (top row) and with (middle row) an overlay of the local bilayer thickness, $D_{TT}$, as indicated by the color scale bar. Black image scale bars are 50 nm. Plotted in the bottom row are histograms of local bilayer thickness calculated as described in Methods, with the total number of vesicles contributing to each histogram denoted by $n$.

Figure 6.

Nanoscopic miscibility transition temperature ($\text{nm-}{T}_{mix}$) determined from FRET data. (a) Shown are FRET ratio vs. temperature data for ternary mixtures composed of high-$T_M$ lipid/DOPC/Chol 40/40/20 mol% for the indicated high-$T_M$ lipids. The solid lines are the best fit to a phenomenological model as described in Methods. Vertical lines indicate the best-fit value of $\text{nm-}{T}_{mix}$ determined from the model fits. Error bars are standard deviations determined from 10 sample replicates for SMPC/DOPC/Chol, 13 replicates for DPPC/DOPC/Chol, and 3 replicates for all other mixtures. (b) $\text{nm-}{T}_{mix}$ for ternary mixtures vs. the chain melting temperature, $T_M$, of the saturated lipid. The dashed line is a linear fit (R² = 0.999) to the data points corresponding to symmetric chain high-$T_M$ lipids (i.e., DMPC, 15:0-PC, DPPC, and DSPC). Error bars are SEM determined from sample replicates.

Figure 7.

Comparing the microscopic and nanoscopic miscibility transition temperatures of ternary mixtures. Shown is the correlation between $\text{μm-}{T}_{mix}$ determined from fluorescence microscopy (horizontal axis) and $\text{nm-}{T}_{mix}$ determined from FRET (vertical axis) for ternary mixtures composed of high-$T_M$ lipid/DOPC/Chol 40/40/20 mol% for the indicated high-$T_M$ lipids. The dashed line is a linear fit (R² = 0.998) to the data points for mixtures with symmetric chain high-$T_M$ lipids (i.e., DMPC, 15:0-PC, DPPC, and DSPC). The solid gray line corresponds to equality of $\text{nm-}{T}_{mix}$ and $\text{μm-}{T}_{mix}$ values and is shown for reference.