Role of hydrophobic and electrostatic interactions in coiled coil stability and specificity.

We have screened two coiled coil-forming libraries in which core a and electrostatic e/g positions have been partially randomized. We observed the relative ability of these residues to confer coiled coil stability using a protein-fragment complementation assay. Our studies continue with the Jun/Fos activator protein-1 (AP-1) leucine zipper complex, as it provides a valid therapeutic target, while representing one of the more simplistic examples of quaternary structure. In mammalian cells, 28 possible dimeric interactions result from combinations of cJun, JunB, JunD, cFos, FosB, Fra1, and Fra2. Consequently, peptides designed to target particular oncogenic members must bind with high affinity and also be specific if they are to function as desired. We have therefore tested the ability of core and electrostatic interactions to confer stable and specific peptides. A previously selected peptide (FosW) formed the template for the core and electrostatic libraries. The winner from the core randomization (FosW(Core)) bound specifically to cJun relative to cFos, FosB, Fra1, Fra2, and the FosW(Core) homodimer, as verified by thermal melting analyses and growth competitions in the presence of either a negative control "mock" peptide or a competitor fusion peptide (cFos-FosB-Fra1-Fra2). In contrast, the winner from the electrostatic e/g randomization (FosW(e/g)) bound to all respective complexes with high stability, suggesting that the more significant energetic contributions made by core residues may be enough to generate specificity as a consequence of positive design.