Determination of specific protein-DNA interactions using a combination of experimental and computational analysis

In the post-genomic era, accurate and rapid characterization of specific DNA binding sites is needed to understand how regulators interact with the genome to coordinate transcription programs. Using yeast transcription factor Leu3p as a model molecule, I combined high throughput approaches with computational analysis to determine DNA binding motifs and to define the determinants of DNA target selection in vivo.; I measured the in vitro equilibrium dissociation constants of 43 Leu3p-binding sites and established that the free energy of binding can be approximated as a sum of free energy contributions from each base-pair. This allows the construction of a Leu3p binding motif to predict binding affinity for all potential binding sites in the genome. From the binding affinity of individual sites, the probability that at least one site is occupied within a defined segment upstream of a gene was calculated for all genes in yeast. This probability is substantially better at predicting regulation by Leu3p than is the number of binding sites.; I then developed a new method for rapid determination of PWMs for DNA-binding proteins. In DNA ImmunoPrecipitation with microarray detection (DIP-chip), protein•DNA complexes are isolated from an in vitro mixture of purified protein and naked genomic DNA. Whole-genome DNA microarrays are used to identify protein-bound DNA fragments. The sequence of the identified fragments was used as input to motif discovery programs to derive PWMs. I demonstrated that Leu3p PWMs defined by DIP-chip are as effective at predicting the location of bound Leu3p in vivo as PWMs defined by conventional in vitro methods.; The genome-wide binding specificity of Leu3p in vitro was compared with the specificity in vivo to determine how chromatin influences DNA target selection. I found that the DNA sequence motif recognized by Leu3 in vitro and in vivo is functionally indistinguishable, but Leu3p binds to different genomic locations in vitro than it does in vivo. Accounting for both sequence motifs and nucleosome occupancy improves prediction of protein-DNA interactions in vivo. These results strongly suggest a genome-wide role for nucleosome occupancy in binding site accessibility and selection.