Development and application of microarray technologies for the highly parallel analysis of the sequence specificity of DNA binding proteins

The interactions between transcription factors and their DNA binding sites are a primary step in the regulatory networks within cells. Many of these are key to pharmacogenomics and other quantitative traits. The development of DNA microarrays beyond hybridization-based assays and into the realm of characterizing the sequence specificities of DNA binding proteins such as transcription factors provides a critical link connecting the field of mRNA expression analysis with the burgeoning fields of proteomics and structural genomics. This technology presents an exciting technological advance since up until now technologies that characterized these critical DNA-protein interactions have been laborious and thus have not permitted more than a handful of DNA sequences to be examined. Furthermore, the types of experiments and analyses presented in this thesis will permit a more complete delineation of the interplay of the regulatory pathways present within cells.; In this thesis, I present the development of a high-throughput technology for probing the specificity of DNA-protein interactions, as well as the technology's various applications and analysis of the data that it provides. In Chapter 2 I show that single-stranded oligonucleotide arrays can be double-stranded by DNA polymerase and further biochemically modified by dam methylase , and that the resulting double-stranded DNA is accessible for interaction with DNA binding proteins. In Chapter 3 I present the quantitative measurement of binding affinities of transcription factors using DNA microarrays. I also show that this microarray technology can distinguish transcription factors with very similar binding site preferences. In Chapter 4 I present the results of searches I performed on the human genome using the recognition sites found in Chapter 3 for the human/mouse early growth factor EGR1. I also describe the nonindependence of the nucleotide positions of transcription factor recognition sites, and show that this nonindependence is responsible for superiority of complete reference tables based on protein binding microarray data over the use of binding site weight matrices and consensus sequences. In Chapter 5, I show that microarrays spotted with DNA fragments at least 1 kb in length permit detection of binding to a single specific site. Therefore, microarrays spotted with entire intergenic regions from smaller genomes such as S. cerevisiae or segments of intergenic regions from higher eukaryotes can be used for genome-wide identifications of transcription factor binding sites and their putatively regulated genes. I also present specific binding of the S. cerevisiae transcription factors Rpn4 and Zap1 to microarrays via phage display constructs of their DNA binding domains. In Chapter 6, I present the construction of genomic binding site knockouts and their subsequent mRNA expression analyses in order to validate the regulatory roles of these sites on the putatively regulated genes. In Chapter 7, I discuss issues regarding the types of microarrays, microarray slides, and DNA sequences to use for such experiments.