The interplay between DNA methylation and gene expression in Arabidopsis thaliana

Abstract

Plants often need to adapt to changing environments despite their immobility, and employ a wide range of regulatory controls to fine-tune gene expression under adverse conditions. Epigenetic changes such as covalent modifications to DNA or their associated histone proteins, and optimal packaging of chromatin within the cell can strongly influence gene expression without changing the underlying genetic sequence. DNA cytosine methylation, which refers to the addition of a methyl residue to cytosine nucleotide, is a well-studied epigenetic mark that is prevalent in several flowering plants and is known to induce gene expression changes at many genomic loci. In plants, DNA methylation occurs in three different nucleotide contexts - CG, CHG and CHH (where H = A,T,C) and is predominantly found at heterochromatic regions that carry repeat sequences and silenced transposable elements. In this dissertation, I studied the model plant species Arabidopsis thaliana to investigate how DNA methylation can directly and indirectly regulate gene expression when it occurs in cis to genes, and how it can affect genome and epigenome stability.

In my first project, I artificially introduced differential methylation in various promoter regions of a gene, to investigate whether downstream gene expression would be affected. I used the well-known fwa epiallele as an example - a locus where methylation marks are absent in the promoter, and thereby activate the expression of the FWA gene. I generated transgenic lines carrying the fwa epiallele with methylation targeted at three different regions of the FWA promoter and measured the induced methylation, altered gene expression and the flowering phenotype associated with the gene. I found that methylation at tandem-repeats in proximity to the gene exerted the greatest impact on downstream gene expression, as opposed to minimal impact when methylation was targeted further upstream.

Next, I investigated the consequences of genome-wide hypomethylation. I generated CRISPR-Cas9 knockouts of the MET1 gene (which encodes a DNA methyltransferase catalyzing CG methylation) in 18 natural accessions of A. thaliana. These mutant lines showed severe phenotypic defects and were compared with respective wild-type lines to examine differential methylation, chromatin accessibility and gene expression using bisulfite sequencing, ATAC sequencing and RNA sequencing. Firstly, I found that all met1 mutant lines exhibited very low levels of genome-wide CG methylation resulting from the inactivation of MET1. Furthermore, these mutants exhibited large changes in their chromatin accessibility and gene expression profiles, which quantitatively varied across different accessions. While many genes associated with transposable elements were activated in met1 mutants, the largest variation in gene expression between accessions occurred at protein-coding genes, many of which included known epialleles. Finally, I uncovered multiple epigenetic states of the same genes in different accessions, indicating complex and unique regulatory mechanisms associated with DNA methylation. Together, these results show that epigenetic patterns are often tightly linked to each other, and that perturbation of methylation is sufficient to catalyse multiple direct and indirect downstream effects which determine plant phenotype. Ultimately, understanding epigenetic mechanisms unique to different accessions can explain how natural populations have independently evolved to maintain optimal genetic and epigenetic stability in their diverse natural habitats.