Feigenson Lab

Overview

Cell membranes exhibit complex behavior, in part because of the large number of different components, but also because of the variety of different types of interactions. This complex behavior can result in lateral organization of components on a range of size scales. Lipid-only mixtures can model the range of size scales, from a few nanometers up to many microns. The size of such compositional heterogeneities for coexisting liquid-disordered + liquid-ordered (Ld + Lo) phases can be controlled entirely by lipid composition for mixtures such as SM/DOPC/POPC/chol. In one compositional region of special interest because of its connection to cell membrane rafts, nanometer-scale domains of Ld phase and Lo phase coexist over a wide range of compositions. These observed behaviors can be better understood in terms of the fundamental molecular interactions and phase properties by use of computational methods of molecular dynamics and Monte Carlo simulations.

Phase Diagrams reveal any limits of solubility when components are mixed. Identification of the phases is one key aspect of solving a phase diagram, and locating phase boundaries in composition and temperature space is another. Important and related questions of experimental design involve the number of components to study, and the compositional resolution of the phase boundaries. We find that 3 components are the minimum to exhibit membrane heterogeneities at the scale of tens of nanometers, and solving such 3-component phase diagrams is an important part of our research program. But in order to control the size of Ld + Lo domains, a minimum of 4 components are needed. Apparently, it is the line tension between Ld/Lo domains that is controlled by lipid composition. Experimental methods that are especially useful for solving phase diagrams include Förster resonance energy transfer (FRET), fluorescence microscopy imaging of giant unilamellar vesicles (GUVs), and differential scanning calorimetry (DSC).
Modulated phase behavior describes morphology of coexisting liquid Ld + Lo phases where the line tension is not great enough to cause minimal phase perimeter, thus shapes arise that can appear as "elongated dots", stripes, or versions of honeycombs. In these cases, bending energies together with dipole repulsion are competing with line tension. These behaviors can be observed with conventional widefield fluorescence microscopy, multi-photon microscopy, confocal fluorescence microscopy, or even DIC microscopy. Special care is required to minimize light-induced artifacts. We handle this problem by using a dye/lipid ratio of 1/5000.
A Competing Interactions Model enables simulation of the sizes and shapes of coexisting lipid domains. We identify two types of interactions that can compete with line tension to produce small and non-circular phase domains: (1) the difference in bending energies between Ld and Lo phases; and (2) the difference in dipole repulsive potential between Ld and Lo phases. Bending energies operate only on curved bilayers, which do occur on cell membranes, for example at microvilli, and on model membranes such as GUVs and smaller liposomes. Dipolar repulsion operates whenever Ld + Lo coexist on curved or flat bilayers.
Molecular Dynamics simulations enable study of many bilayer behaviors that are not clear from wet lab experiments. We use GROMACS, with our own computer cluster, and with the NSF resource XSEDE. We are particularly interested in (1) the size transition of coexisting Ld + Lo phase domains from nanoscopic to macroscopic; (2) the nature of the Ld/Lo interface; and (3) interactions of membrane proteins with the bilayer that change bilayer phase behavior or phase morphology.