Abstract:

The loss of conformational freedom experienced by lipid chains in the vicinity of one, or two, impenetrable walls, representing the surfaces of hydrophobic transmembrane proteins, is calculated using a mean-field molecular-level chain packing theory. The hydrophobic thickness of the protein is set equal to that of the unperturbed lipid membrane (i.e., no ``hydrophobic mismatch''). The probability distributions of chain conformations, at all distances from the walls, are calculated by generating all conformations according to the rotational-isomeric-state model, and subjecting the system free energy to the requirement that the hydrophobic core of the membrane is liquid-like, and hence uniformly packed by chain segments. As long as the two protein surfaces are far apart, their interaction zones do not overlap, each extending over several molecular diameters. When the interaction zones begin to overlap, inter-protein repulsion sets in. At some intermediate distance the interaction turns strongly attractive, resulting from the depletion of (highly constrained) lipid tails from the volume separating the two surfaces. The chains confined between the hydrophobic surfaces are tilted away from the walls. Their tilt angle decreases monotonically with the distance from the walls, and with the distance between the walls. A nonmonotonic variation of the lipid-mediated interaction free energy between hydrophobic surfaces in membranes is also obtained using a simple, analytical, model in which chain conformations are grouped according to their director (end-to-end vector) orientations.