BPS2026 – Defining binding specificity in transcription factors through knob-socket analysis

DNA binding proteins play important roles in regulating gene expression, yet a definitive model of recognition remains elusive. A knob-socket (KS) analysis was performed on a set of basic leucine zipper (bZIP) proteins binding to DNA double helix structures. The KS model simplifies tertiary and quaternary interactions into surfaces defined by 3 residue sockets that pack a 4th knob residue that can be mapped in two-dimensions to identify regular patterns of packing interactions. A model for DNA recognition of the 4 base half-site by a single bZIP α-helix identifies a quadripartite region made up of 9 noncontiguous α-helical residues oriented by a conserved set of 3 residues on an i+4 ridge. This region consists of 4 pockets each made up of a specific set of 4 α-helical residues that map to and recognize 1 of the 4 DNA base positions. The KS analysis reveals that the α-helix splits recognition with the first two on the positive (P) strand P1 and P2 and the last two on the negative (N) strand N3 and N4. From the KS analysis, the majority of recognition is associated with the C5 methyl group of dT. The bZIPs CREB and c/EBPα were chosen for this study due to their differing recognition of dT, where CREB packs dT’s methyl at the P2 and N4 positions and c/EBPα at the N3 and N4. Mutations were performed based on the hydrogen bonding analysis to swap the specificity and analyzed alongside the hydrogen bonding network and KS analysis to further understand the recognition of non-dT bases. Furthermore, electrophoretic mobility shift assay (EMSA) was used to further investigate the contributions of each base to specificity as well as negative design in the recognition sequence, particularly focusing on mutated residues.