Water-solubilization of integral membrane proteins

We developed computational methods to water-solubilize membrane proteins based on the observation that the cores of membrane- and water-soluble proteins are similar in their amino acid composition, with the differences being in the solvent exposed residues-water-soluble proteins have hydrophilic, while membrane proteins have hydrophobic exteriors. Thus, it should be possible to water-solubilize a membrane protein by mutating exposed residues to hydrophilic residues. Water-solubilized membrane proteins may retain the structures and binding functions of their membrane-soluble predecessors, being more amenable to solution studies, crystallization, and use in NMR. To test this, we chose two membrane proteins, phospholamban (PLB), and KcsA. Phospholamban contains 52 residues, exists in a monomer-pentamer equilibrium, and is a reversible inhibitor of the Ca2+-dependent ATPase SERCA2a. Phosphorylation of PLB promotes pentameriation, relieving monomer-induced inhibition of SERCA2a. We used a pairwise amino acid potential to computationally redesign PLB, mutating membrane-exposed positions to hydrophilic residues, maintaining the apolar LeuaIled core, generating water-soluble PLB (WSPLB). WSPLB was cloned, expressed, and purified, and found to be water-soluble, α-helical, pentameric, with the degree of pentamerization increasing upon phosphorylation. Truncated WSPLB constructs were synthesized, and it was observed that the truncated peptide 31–52, which contained only the LeuaIle d core, was tetrameric, while inclusion of interhelical hydrogen bonding residues Gln22(g), Gln23(a) Gln26(d), Asn27(e), GIn29(g), and Asn30(a) promoted pentamerization. The x-ray crystal structure of WSPLB 21–52 was solved to 1.8 Å resolution, showing an antiparallel tetrameric coiled-coil. A model was developed where WSPLB switches from an antiparallel tetramer, with inclusion of interhelical hydrogen bonding residues promoting parallel pentamers. Computational methods were extended to water-solubilize the tetrameric bacterial ion channel KcsA. Water-soluble KcsA (WSK-3) was developed, cloned, expressed, and purified and found to be water-soluble, α-helical, and largely tetrameric. Importantly, mutants of WSK-3 which included a binding site for the scorpion toxin agitoxin2, bound to the toxin with the same stoichiometry and affinity as KcsA. WSK-3 also preferentially bound tetraethylammonium over tetramethylammonium with the same affinity as KcsA. These studies demonstrate that membrane proteins can be water-soluibilized while retaining the proper structure and binding properties as their membrane-soluble predecessors