Mesoscopic simulations of phase transitions in lipid bilayers

Abstract

The formation of the rippled phase in biological membranes and its relation with anomalous swelling are still lacking a molecular explanation. Starting from all-atom simulations we use a mapping to create a mesoscopic model of the lipid dimyristoylphosphatidylcholine (DMPC) in water. We use this model to study the phase behaviour of lipid bilayers. Depending on the lipid structure and head group, our simulations reproduce the experimental phase diagrams. The anomalous swelling is caused by conformational changes of the lipid tails but is not directly related to the rippled phase. A key factor for the rippled phase is a frustration between the surface area of the heads and the lateral density of the tails.Phospholipids can self-assemble in water to form bilayer structures, which are seen as model systems of biological membranes.1 At high temperatures lipids form a flat fluid membrane (L a phase) and at low temperatures a flat gel phase in which the lipid tails are ordered and tilted (L b 0 phase). For certain phospholipids between the L a and L b 0 phase phases a stable supramolecular periodic structure was observed [2][3][4] which is characterized by a long-wavelength rippling of the bilayer (P b 0 phase) and a (anomalous) swelling of the membrane. This rippled phase has attracted the attention of many groups, 2-7 but still lacks a molecular understanding. 1,8 Using state of the art molecular dynamics simulations, it is possible to obtain detailed structural information of a single phase, for example, the gel phase, 9 but these are too time consuming to determine a complete phase diagram. In this work, we use an alternative approach in which realistic all-atom simulations are used to determine the effective intramolecular interaction parameters of a mesoscopic model. At this mesoscopic level simulations are four to five orders of magnitude more efficient, 10,11 allowing us to compute complete phase diagrams. In a dissipative particle dynamics (DPD) simulation a particle represents the centre of mass of a cluster of atoms. The total force on a particle consists of dissipative, random, and conservative forces. 12,13 For the conservative force we use a soft-repulsive interactionwhere r ij is the distance between particles i and j, a ij is the parameter characterising the interaction between two particles, and r c is the cut-off radius. In our mesoscopic model, we distinguish three types of particles, w, h, and t to mimic the water and the head-and tail-atoms of a lipid, respectively. A coarse graining procedure to map the interactions of realistic molecules on DPD interaction parameters has been developed by Groot and Warren. 14 In this procedure the value of the repulsion parameter is taken such that the DPD waterlike particles reproduce the compressibility of water. The interactions of the hydrophilic and hydrophobic particles (a ww ¼ a tt ¼ 25, a ht ¼ a wt ¼ 80, and a hw ¼ 15) are based on the Flory-Huggins solubility parameters. We vary the headhead (a hh ) interaction parameter to study the effect of changin...

Understanding the Phase Behavior of Coarse-Grained Model Lipid Bilayers through Computational Calorimetry

“…11,42 Various harmonic bond and bond angle interactions are set so as to match typical molecular flexibility in all-atom models. 43 MARTINI Model. The coarse-graining procedure in the MAR-TINI model is based on a four-to-one mapping of non-hydrogen atoms into beads.…”

Phase Behavior of Model Lipid Bilayers

“…A short communication on these results has been published previouly. 29 The resulting phase diagrams are presented in section 3. We will show that we can reproduce the various phases of the phospholipid bilayer, dependent on temperature and head-head interaction.…”

Variable tilt on lipid membranes

“…Accordingly, the relevant contribution to ∂W/∂δ is G d n. In the same way, 27) where 29) and the relevant contribution to ∂W/∂δ is ξ −1 G ξ φ. Altogether,…”