Two related and biologically important families of eukaryotic transcription factors, defined by the promoter elements they interact with, contain DNA- binding domains having he bZIP structural motif. These are the AP-1 family, which includes the Jun and Fos oncoproteins, and the ATF/CREB proteins recognize related, but not identical DNA sequences, and individual members of these families have distinct dimerization properties. The interactions within and between the AP-1 and ATR/CREB protein families results in a large number of distinct DNA-binding transcription actors, and it is extremely likely that this diversity is important in gene regulation and oncogenesis. The goals of this proposal are to determine the detailed molecular basis for both dimerization and DNA-binding specificity of bZIP proteins, and to precisely define the DNA-binding specificities of the individual AP-1 and ATF/CREB proteins and the functional heterodimers. The major experimental approaches will involve 1) degenerate cassette mutagenesis to generate extensive and diverse protein libraries of Fos, Jun, and chimeric derivatives, 2) powerful genetic selections and screens in yeast cells to identify functional or non-functional proteins including those with novel sequences or interesting properties such as altered specificity, 3) biochemical characterization of the dimerization and DNA- binding properties of the resultant proteins, and 4) precisely determining DNA-binding specificity of the individual AP-1, ATF/CREB and functional heterodimeric proteins by the random selection method in which high- affinity binding sites are selected from a collection of random-sequence oligonucleotides. These experiments should lead to detailed knowledge of the critical protein-protein and protein-DNA interactions that determine how dimerization, DNA-binding, and half-site spacing specificity is achieved in bZIP proteins. As such, this knowledge should provide crucial complementary information to biophysical and X-ray crystallographic studies of bZIP proteins carried out in other laboratories. In addition, knowledge of the precise DNA sequence recognition properties of the various homodimer and heterodimer combinations should yield important clues as to the potential target genes that are directly affected by AP-1 and ATF/CREB proteins.
