Missing heritability and soft inheritance of morphology and metabolism in Arabidopsis

The plant phenotype is shaped by complex interactions between its genotype and the environment. Although the genotype is stable and determined by the genomic sequence, plants are able to respond flexibly to changes in environmental conditions by orchestrated signal transduction pathways. The genomic sequence may change with each generation through chromosome rearrangements, meiotic recombination and spontaneous mutations. Through natural selection on these randomly induced changes, genotypes become adapted to their local environment. Because different genotypes adapt to different environments, natural variation within species expands in time and gives rise to a wide variety of genotypes and phenotypes. The genetic architecture that specifies the phenotype can be investigated by analyzing different genotypes in the same environment and associate the phenotypic variation with molecular markers that discriminate the genotypes. Recent advances in next-generation sequencing technology enabled the fast sequencing of entire genomes, and in <em>Arabidopsis</em><em>thaliana</em>alone, more than 1000 different genotypes have been fully resequenced. The sequencing allows the association of phenotypic variation with large numbers of single nucleotide polymorphisms (SNPs) that greatly enhance resolution in genome-wide association studies (GWAS). GWAS on human diseases suffer from missing heritability that is most likely caused by the genetic architecture of the disease traits. Many variants of small effect or rare variants most likely determine a large part of the genetic variation and these variants are difficult to identify in GWAS due to lack of statistical power. In plants, several GWAS have been performed and they have identified previously validated genes and genes involved in monogenic disease resistance, but elucidating quantitative traits such as many agronomic important traits might be problematic in plants as well. Chapter 2 describes a GWA study in which quantitative morphological traits, such as leaf area, flowering time and branching were examined in 350 accessions of Arabidopsis for association with about 200,000 SNPs. The morphological traits showed extensive variation and were highly heritable, but GWA mapping could not identify the genetic variants that explain the heritability. Therefore, missing heritability was addressed using genomic selection models and these models confirmed the quantitative complex architecture of the morphological traits. Based upon these results, the heritability was assumed to be hidden below the significance threshold, and indeed lowering the significance threshold enabled the identification of many candidate genes that have been implicated to play a role in the phenotype directly or indirectly, in previous studies. One candidate gene was studied in more detail; natural variants of <em>ACS11</em>, an ethylene biosynthesis gene, associated significantly with the petiole to leaf length ratio. <em>ACS11</em>is indeed expressed in petioles and ectopically supplied ethylene abolished the difference in the phenotype of natural variants at this locus, strongly suggesting that <em>ACS11</em>is involved in the regulation of petiole growth. However, lowering the significance threshold also increases the number of false-positive associations, non-causal alleles that co-segregate with the trait values. Because regulation of the morphological traits occurs at multiple intermediate levels, increased certainty on the associations can be obtained by performing GWA mapping on the intermediate levels from genotype to phenotype such as gene expression, and protein and metabolite content. Chapter 3 describes a literature survey into the multi-dimensional regulation of metabolic networks that are regulated by inputs from the clock, the communication between cells and between source and sink tissues, and the environment. The metabolic status of the plant can be seen as the final product of the interaction with the environment, and as such, it can serve as a blueprint for growth and development. Chapter 4 describes the abundant variation in enzyme activities and metabolites involved in primary carbon and nitrogen metabolism. The metabolite and enzyme activity data were analyzed together with plant biomass data, and many pleiotropic regulators were identified with opposite effects on primary metabolism and biomass formation. Natural variants in two stress-responsive genes were oppositely associated with biomass and many enzymes and metabolites involved in primary metabolism, suggesting that higher enzyme activities and higher levels of sugars and proteins might be needed to support plant resistance to stress at the expense of growth. Some studies indicated that epigenetic variation, independent of the genetic SNPs, may contribute to missing heritability. Epigenetic inheritance is defined as the inheritance of phenotypic variation to future generations without changes in DNA sequence. Epigenetic variation is caused by variation in chromatin marks such as DNA methylation, histone modifications and small RNAs. Recently, a recombinant inbred line (RIL) population was developed in Arabidopsis where the chromosomes are differentially methylated in lines with an otherwise isogenic background by crossing wild-type Col-0 with a hypomethylated <em>ddm1-2</em>mutant. Chapter 5 describes the epigenetic regulation of morphology and phenotypic plasticity by studying morphological variation in 99 epiRILs under control and saline conditions. The morphology and plasticity trait values were associated with differentially methylated regions (DMRs) that were used as molecular markers in QTL mapping. Many QTLs for various morphological traits and phenotypic plasticity parameters co-located, suggesting pleiotropic epigenetic regulation of growth, morphology and plasticity. Furthermore, methylation variation in the promoter of a salt-tolerance gene, <em>HIGH-AFFINITY K⁺TRANSPORTER1 (HKT1)</em>associated significantly with leaf area, especially under saline conditions. To gain more insight into the epigenetic regulation of plant growth and morphology, chapter 6 describes the epigenetic regulation of secondary metabolite levels in leaves and flowers and studies the relationship with the morphological traits determined in chapter 5. Many of the QTLs that were found for growth and morphology overlapped with the QTLs for metabolic traits, and suggest pleiotropic regulation. Furthermore, subsets of the metabolites correlated well with the morphological traits and might thus be regulated by the same loci. The majority of metabolite QTLs was detected for glucosinolates and flavonoids in the flowers, and methylation variation was observed for some of the biosynthetic pathway genes of these compounds when comparing Col-0 and <em>ddm1-2</em>, which indicates a role for epigenetic regulation of these biosynthesis pathways. Although stable, natural epialleles have been found in plant species and the environment can induce hypo- and hypermethylation of DNA, it remains elusive whether environmentally-induced epigenetic changes can be inherited to subsequent generations, independent of genetic variation. Chapter 7 describes the transgenerational inheritance of phenotypic variation in progeny derived from a common Arabidopsis founder line. The progeny of stressed parents and grandparents showed variation in morphological traits, metabolite accumulation and gene expression. For example, many salt-responsive genes were up-regulated in progeny of salt-stressed grandparents. The responses to biotic (methyljasmonate) and abiotic (salt) stress differed strongly and this suggests that different environments can cause different transgenerational responses. Because all lines are derived from a single ancestor, epigenetic variation and not DNA variation is most likely causal for the phenotypic variation. Further studies are, however, needed to provide conclusive evidence for transgenerational inheritance. Chapter 8 provides a synthesis of the work and discusses the GWA studies in the light of missing heritability, genetic architecture and the verification of candidate genes. The work on epigenetic regulation of phenotypic plasticity, morphology and metabolism is discussed in relation to Lamarckian soft inheritance that gained new enthusiasm after some recent discoveries in the field of epigenetics. And finally, the metabolomics work is discussed in the light of the growth-defense hypothesis that states that investments in defense occur at the expense of growth.