Design of a metallo-bZIP protein that discriminates between CRE and AP1 target sites: selection against AP1.

The bZIP class of dimeric DNA binding proteins is characterized by a motif containing two discrete domains: a DNA contact domain defined by conserved basic and hydrophobic residues (basic domain), and a dimerization domain identified by a heptad repeat of leucine residues (zipper domain). Molecules are constructed in which the GCN4 dimerization domain is replaced by a series of stereochemically defined metal ion complexes that alter systematically the relative orientation of the basic domain peptides. Both the affinity and the specificity of DNA binding are modulated by seemingly small changes in metal complex stereochemistry. Although GCN4 binds CRE (ATGACGTCAT) and AP1 (ATGACTCAT) target sites with comparable affinity, one metallo-bZIP peptide ([G29Ts]2Fe) prefers the CRE by 4 kcal.mol-1 (1 cal = 4.184 J). Competition experiments performed with several DNAs demonstrate that discrimination between CRE and AP1 is dominated by selection against the AP1 site.