A Model of bZIP Protein/DNA Recognition using Knob-Socket Analysis

Abstract

The basic leucine zipper protein (bZIP) family binds specific DNA sequences through interactions between the individual α-helical coils and respective half-sites. Despite previous studies on target sequences and affinities of both homo- and heterodimers, the recognition and binding mechanism of bZIP dimers are not fully understood. Despite the relatively high homology between the subfamilies, dimers of bZIP coils have shown high variability in target sequences, ranging from conjoined and variably-spaced target sequences to novel and emergent sequences. To help elucidate the determinants of protein-DNA recognition, a novel Knob-Socket analysis of the packing at the protein and nucleic acid interface is able to produce an insightful model of half-site recognition specificity between a protein α-helix and the DNA double helix. The model not only characterizes the protein sequence motifs that recognize DNA bases, but uniquely is able to describe the DNA motifs that specifically pack protein residues. The bZIP coils studied include Jun, Fos, CREB, and C/EBP, through the analyses of their respective PDB files. For each half-site, the protein ɑ-helix splits recognition on both the positive and negative strands of the DNA double helix. Half of the α-helix packs into the first two nucleotides of the positive DNA strand, and the other packs into the last two nucleotides on the negative DNA strand. Along the i+4 ridge, a set of three conserved residues provides a dividing line on the protein helix between positive and negative DNA strand recognition. These three residues are also involved in a quadripartite region that specifically packs and recognizes the methyl group on a thymidine. Recognition of other bases are more complex and involve characterization of the DNA sequence motifs that pack and recognize residues on the protein α-helix.