Computational design of soluble variants of transmembrane proteins: an information theoretic approach

Abstract

The study examines and implements a strategy to build soluble analogues of various hydrophobic transmembane proteins. The design is done using an information-theoretic approach which allows one to perturb certain properties of the protein while keeping the others constant. Knowledge-based force fields are used to obtain self and cross-interaction energies. A novel postprocessing technique is added to the theory to obtain a library of analogues with controlled but varying degrees of solubility. Not all mutations suggested by the information theoretic approach are kept. Only the significant mutations as determined by a threshold value are made. The technique also applies a sliding window protocol and implements point mutations at certain structurally and functionally 'non-conserved' locations. The library approach provides greater flexibility during actual experimental synthesis of the mutant. The overall structural and functional properties are preserved even in the altered context and this is validated by alignment and homology modelling.