Relating phase separation and thickness mismatch in model lipid membranes.

Abstract

Membrane lipids can spontaneously undergo a process called phase separation wherein they organize into domains with different compositions. Phase separation has been proposed as a mechanism in the formation of heterogeneous lipid regions (e.g. rafts) in cell membranes. The activity and partitioning of membrane proteins vary depending on the local lipid environment due both to hydrophobic thickness matching between the protein and lipids and to lipid packing (also known as lipid order). A major question is how these two parameters, thickness and order, vary with lipid composition. Many previous studies have treated membrane order and thickness as inextricably linked; thicker regions always have higher order. Other studies have attributed membrane thickness to properties of phase separation, including the temperature below which phases appear, Tmix, and the size of phase domains. In this thesis we present two studies that test whether the previously observed relationships among phase thickness, order, and Tmix are as straightforward as previously thought. For the first study, we investigate whether in liquid-liquid phase separated membranes the more ordered phase need be thicker than the less ordered phase. We use non-canonical ternary lipid mixtures designed to produce thicker Ld than Lo phases and measure mixing temperatures in fluorescence. We find that few of these candidate systems produce coexisting liquid phases in vesicles. We also directly measure membrane thickness by atomic force microscopy and find that none of the systems produce thicker Ld than Lo phases under standard experimental conditions, although one system does after photo-oxidation. Interestingly, we find no simple monotonic relationship exists between the physical parameters of single-component lipid membranes and the highest demixing temperature of the corresponding ternary membranes. These results highlight the surprisingly robust propensity for the Lo phase to be thicker than the Ld phase, regardless of lipid composition. In our second study, we provide an experimental counterpoint to the previously found trend that mixing temperatures depend strongly on the thickness mismatch of the two phases. We use measure mixing temperatures and thickness differences of membranes composed of different mole ratios of DiPhyPC/DPPC/chol in fluorescence microscopy and atomic force microscopy. We show trends in which membrane thickness differences do not uniquely predict mixing temperatures. Our results highlight the importance of membrane composition on the physical properties of phase separation.