Domain formation in cholesterol–phospholipid membranes exposed to adhesive surfaces or environments

Abstract

Domain formation in binary mixtures of cholesterol and a single phospholipid has been studied for a long time but the nature of these domains is still a matter of some debate. One interpretation of these domains is that they arise from the coexistence of liquid-ordered and liquid-disordered phases within the membranes. Here, we study the effect of adhesive surfaces and environments on the proposed phase behavior theoretically. The adhesion or partial support of the membranes leads to two or several membrane segments that exhibit different molecular interactions with their environments and, thus, have different molecular affinities to these environments. We show that the affinity contrasts between different environments strongly affect the phase behavior of the membranes. First, the affinity contrasts confine the domain formation spatially to single membrane segments. Second, these contrasts lead to a partitioning of the coexistence region in the composition–temperature plane into distinct coexistence regions for the different membrane segments. Third, the range of membrane compositions, for which one can observe domain formation in one of the membrane segments, is always reduced compared to the composition range of the free membrane. Furthermore, small affinity contrasts represent singular perturbations in the sense that the phase diagrams change in a discontinuous manner as one varies the affinity contrast from small positive to small negative values. All of these results are quite general because they follow from thermodynamic stability alone and apply to any two-component membrane with fluid–fluid coexistence. In addition to these generic features, we also provide a computational scheme, by which one can explicitly calculate the partitioning of the coexistence regions for a specific binary mixture, and illustrate this scheme for binary cholesterol–DPPC membranes. Both the generic and the specific features of domain formation as predicted by our theory are accessible to experimental studies.