UC Berkeley

One remarkable property of the eukaryotic cell is its ability to orchestrate the activities of thousands of genes in a complex temporal symphony of transcriptional expression. Development in multicellular species often requires that many genes lie dormant in early undifferentiated cellular lineages, awakening only in the tissues that they help define. Even single cellular species, such as the yeast Saccharomyces cerevisiae, need to keep some genes in a temporary, transcriptionally-repressed state until the onset of particular environmental conditions. This is no easy feat to accomplish, and cells use many different molecular mechanisms to do so; this includes the intricate interplay of many epigenetic regulatory systems, such as the post-translational modification of histones, the incorporation of histone variants, and a covalent but reversible modification of the DNA itself, DNA methylation.

In this dissertation, I describe a series of experiments designed to help understand the interaction between two of these epigenetic factors, DNA methylation and the histone variant H2A.Z, within the context of gene regulation. This work was conceived after initial mapping experiments in the model plant Arabidopsis thaliana revealed that the genome-wide distributions of H2A.Z and DNA methylation are strikingly anticorrelated. Additionally experiments have revealed that the basis for this relationship is the exclusion of H2A.Z from chromatin by the presence of DNA methylation, an epigenetic principle that appears likely to be an ancient invention conserved among both plants and animals.

To better understand what purpose this relationship might hold in eukaryotes, I developed an Arabidopsis partial loss-of-function h2a.z mutant, and surveyed its genome-wide RNA expression profile. These experiments revealed strong correlations between transcriptional misregulation in the h2a.z mutant, the presence of H2A.Z within gene bodies, and levels of gene responsiveness, a measure of the degree to which a gene's expression varies across tissue types or environmental conditions. As we have shown that the presence of DNA methylation antagonizes H2A.Z incorporation across the genome, we propose that one basal function of gene-body methylation, an ancient and yet mysterious chromatin feature found in many eukaryotes, may be the prevention of H2A.Z incorporation within the bodies of genes that need to be constitutively expressed.

How gene body methylation is targeted in the first place remains unclear. The fact that genic methylation in all species is almost exclusively limited to CG sites, even in plants which have two other contexts of methylation that they use for the silencing of transposons, suggests that the various methylation targeting machineries are somehow able to distinguish between gene sequences and their other heterochromatic targets. Recently, several mutants in Arabidopsis have been shown to accumulate non-CG methylation within gene bodies. In order to understand the mechanisms responsible for this hypermethylation of genes, we examined the methylation profiles of these mutants. We discovered that the hypermethylation phenotypes of these mutants are quite different from one another in several respects, including their correlation with normal genic methylation, their distribution patterns across the gene body, and their dependence on the endogenous RNAi machinery. This suggests that multiple mechanisms may be responsible for controlling genic methylation patterns.