Measuring stress-induced DNA methylation in apomictic Dandelions

The success or continuous existence of species requires continuous adaptation to changes in the environment to survive and contribute offspring to the next generation. Selection acts on the phenotype, which is in turn determined by the complex interplay of genetic, epigenetic and environmental variation. (Natural) selection leads to ‘survival of the fittest’ or best-adapted individuals to their local environment, ultimately determining which individuals contribute offspring to the next generation. Understanding the mechanisms by which epigenetic and genetic variation can arise and get passed on through generations determines our understanding of inheritance and evolution. Hitherto, the mechanistic understanding of genetics has shaped the scientific view of inheritance and evolution, leading to the gene-centered paradigm of Neo-Darwinism. However, recent studies indicate that besides genetic (DNA sequence) variation, epigenetic variation can also be transmitted between generations. Further studies on the properties and transgenerational dynamics of epigenetic variation are needed to enhance our understanding of heritability and evolution.

Epigenetic variation has distinct properties and different transgenerational dynamics compared to genetic variation. Epigenetic variation helps to regulate gene expression and determines the different cell types and function in eukaryotes. The main function of DNA methylation, an important part of the epigenetic code, is to prevent the spread of selfish genetic elements in the genome and to establish the different cellular profiles observed in multicellular organisms. One differentiating feature of epigenetic variation compared to genetic variation is that (specific) epigenetic variation can arise under the influence of stress. This can enable a trans-generational stress-response of organisms which can have a positive influence on the phenotype and (natural) selection on either the (enhanced level of) transgenerational phenotypic plasticity or the epigenetic variation itself, potentially influencing natural selection and ultimately evolution. Where genetic variation can be characterized as hard-inheritance, the inheritance of epigenetic variation is often referred to as ‘soft-inheritance’ due to the lower transgenerational stability and resetting that occurs in the intergenerational transfer of epigenetic variation. Epigenetic variation is also often dependent on, or a downstream consequence, of genetic variation, suggesting that it is (in part) determined by genetic variation.

Mechanistic studies in model species have contributed greatly to the understanding of the molecular mechanisms that control the dynamics of different epigenetic marks present in multicellular organisms. In plants, studies in the model plant Arabidopsis thaliana have resulted in deciphering the most important molecular mechanisms and actors, giving an ever-increasing insight into the dynamics of epigenetic regulation of cells and organisms. A key feature of model systems is the ability to ‘switch’ off certain genes or molecular pathways, for instance via the experimental use of mutants, enabling the study of their role in the heritability of epigenetic marks. DNA methylation is a well-studied epigenetic mark, which has shown high stability even in transgenerational experiments.

From the perspective of studying epigenetic variation, plants are particularly interesting for several reasons, most importantly: 1) The separation between soma and germline, the Weismann barrier, is less strict in plants compared to other eukaryotes, as in higher plants

germline cells are formed during floral development from somatic cells (which can occur throughout the life of the plant), whereas in most eukaryotes germline cell development is restricted to a defined point (early) in the organismal development. 2) The sessile nature of plants makes an adaptive plastic response to changing environments an important feature, a plant cannot just walk away when the going gets tough. 3) The transgenerational stability of DNA methylation is higher in plants compared to other eukaryotes such as mammals, in which epigenetic information is erased during germline reprogramming. These factors combined suggest that the potential importance of epigenetic variation in plants might be high.

In this thesis, I focus on studying DNA methylation in apomictic Dandelions, applying Next Generation Sequencing (NGS) approaches to the study of this non-model plant species. Apomictic dandelions produce seeds that are genetically identical to the ‘mother’ plant, which makes it easier to study the influence of epigenetic variation without confounding effects of genetic variation. Working with Next Generation Sequencing data is still relatively new and therefore not always optimized for specific types of analysis. I discovered a distinct error pattern in RNAseq data that indicated an artificial source of variation that could be traced back to the way the RNAseq libraries were constructed. The first publication of this thesis contains a technical analysis of such artefacts present in RNAseq data, suggesting that these errors are related to random hexamer mispriming during library construction (Chapter 2).

The main goal of my work is to better understand the role of epigenetic variation in adaptation and plasticity of plants. This role remains poorly understood. This is in part due to the lack of high-resolution techniques that allow for the detailed study of epigenetic marks such as DNA methylation in non-model organisms. Existing techniques for measuring DNA methylation such as methylation sensitive AFLPs offer only information on DNA methylation variation in an anonymous and limited fashion. The plummeting costs of sequencing techniques have enabled large-scale genotyping efforts (focusing on genetic variation only) for a wide variety of non-model organisms. Here, I extended this popular genotyping by sequencing technique, to allow for sequencing-based epigenotyping or epiGBS (chapter 3), which allows for measuring DNA methylation and genetic variation in hundreds of samples simultaneously. I have extensively validated the approach, providing evidence that with the right design, the accuracy of the DNA methylation measurements with epiGBS are as high as those with the gold standard Whole Genome Bisulfite Sequencing.

An important aim of my PhD research was to investigate the stability of (stress induced) DNA methylation variation in apomictic dandelions and the potential of phenotypic variation underpinned by DNA methylation variation to be subjected to selection. I therefore studied the transgenerational stability of both stress induced and natural DNA methylation variation in different genotypes of apomictic dandelions in a six-generation experiment, comparing DNA methylation patterns between generations and tracking changes in them (chapter 4) using epiGBS. I found clear but limited evidence for environmental induction of heritable DNA methylation changes after application of Jasmonic Acid. Furthermore, I found a significant negative relation between the similarity of  DNA methylation patterns and intergenerational distance, indicating epigenetic divergence over generations. I conclude that DNA methylation in both CG and CHG (where H can be any nucleotide except for G) sequence context are heritable and that environmental perturbation can result in heritable DNA methylation changes which are however not widespread.

A prerequisite for epigenetic variation to contribute to adaptation is that epigenetic variants that affect the phenotype are heritable. To test whether an epigenetics-based selection response is possible, at least over the time course of a few generations, I selected early flowering for two subsequent generations in three genotypes of apomictic dandelions. This selection effort included lines that received a stress pre-treatment with either Jasmonic Acid or 5-azacytidine, to determine if stress-induced DNA methylation variation would increase the capacity to respond to selection. The selection experiment on flowering time (chapter 5) resulted in a shift in flowering time for all treatments in a young apomict, suggesting that natural and heritable epigenetic variation can underpin quantitative traits such as flowering time. I also found evidence for treatment induced (epi)genetic variation leading to a stronger selection response in one out of 3 genotypes. This suggests that stress- induced heritable epigenetic variation can lead to a selection response. Further study is however required to rule out genetic variants and to study the long-term stability of the variation selected upon.

Finally, in the General Discussion I summarize the findings, putting them in context with recently published studies. I reflect on the state of the field of ecological epigenetics and in what sense the epiGBS technique that I developed and other emerging techniques can contribute to a better understanding of the role of epigenetic variation in ecology and evolution. I reflect on the place of epiGBS compared to other techniques. I point out that with the growing evidence of the inadequacy and misinterpretation of MS-AFLP results a systematic review of such results by replicating the studies employing sequencing based techniques such as epiGBS instead of MS-AFLP is in order.