Mammalian strategies to regulate transcription: Transcription factor sumoylation and cis-regulatory region identification

The revelation that the 6 billion base pairs of the human genome harbor relatively few (30,000) genes was one of the biggest surprises accompanying the first genome annotation in 2001. When compared to the 20,000 genes of the nematode Caenorhabditis elegans, it was immediately clear that gene number is not the defining factor underlying the behavioral and phenotypic complexity of mammals. Interestingly, 5-10% of human genes appear to code for transcription factors. Moreover, these transcription factors frequently participate in a diverse array of interactions with other transcription regulatory proteins, known as combinatorial regulation, and can be subject to a wide variety of post-translational modifications, such as sumoylation, which can tune their function. The intricate tapestry of transcriptional regulation may ultimately be the key to unlocking the complexity of mammalian species. However, deciphering the transcriptional code has proven to be a challenging venture, both at the level of transcription factor regulation and discovering cis-regulatory regions in the vast non-coding space. This thesis examines how covalent attachment of the polypeptide SUMO can influence the transactivation capacity of the transcription factors Smad4 and MEF2A and to explore the features defining the genomic sites of combinatorial cis-regulation in mammals. Through these studies, I determined that both Smad4 and MEF2A can serve as substrates for sumoylation but that this modification activates Smad4 but attenuates MEF2A activity. I also showed that mammalian cis-regulatory regions can be defined as clusters of transcription factor binding sites that are conserved in order and in spacing across mammalian genomes; such sites can be employed in combinatorial regulation. We call these regions Pattern-defined R egulatory Islands (PRIs). Unexpectedly, I also found a PRI that encompasses an essential exon of the gene ADAMTS5 that relies on binding sites in the exonic region. This is the first example of coding sequence moonlighting as an enhancer. Altogether, this thesis work enriches our understanding of mammalian transcriptional regulation.