Gene-specific regulation of transcription is achieved through the binding of transcription factors to DNA sequences. Many Eukaryotic transcription factors maintain affinity differences between target and non-target sequences that appear too small to explain the specificity observed for the genes they regulate. How is specificity achieved in Eukaryotic gene expression? In eukaryotes, DNA is spooled around histone protein octamers to form nucleosomes. The nucleosome represses transcription by acting as a barrier to the binding of transcription factors. Thus, gene activation requires the recruitment of ATP-dependent chromatin remodelers which remove nucleosomes covering important regulatory sequences. However, promoter nucleosome structure is heterogeneous even under activating conditions. Why does the cell expend energy to maintain heterogenous promoter chromatin in the promoters of actively transcribing genes?
In Chapter 1, I present a model of gene transcription which represents a unified solution to these questions, among others. I show that activator mediated ATP dependent stochastic removal and reformation of nucleosomes on promoter DNA may be used for the kinetic proofreading of activator-DNA interactions. The specificity enhancement due to kinetic proofreading is an archetype that, in part, can be used to explain the observed specificity in Eukaryotic gene expression. I show that contrary to expectation, heterogeneity in promoter chromatin structure reduces the variation observed in gene expression. Additionally, I provide insight into the necessity of transcriptional bursting for regulated, highly expressed genes.
In Chapter 2, I present a number of experimental tests of the proofreading model. We observe transcriptional bursting, chromatin remodeling and activator binding at a classic model gene, PHO5, in Saccharomyces cerevisiae. I show that transcriptional bursting of PHO5 occurs in at least two distinct timescales, an expectation of the proofreading model. In addition, I show that mutation of a single chromatin remodeler, Isw2, is sufficient to disrupt correlation at the longer timescale. I present a model of kinetic proofreading of activator specificity by Isw2 and test conjectures such a model purports.
In chapter 3, I present a technique for studying eukaryotic gene expression by generating and testing the expression of >400,000 permuted synthetic cassettes generated from 26 genes from Saccharomyces cerevisiae.