Ecogenomics of plant resistance to biotic and abiotic stresses

<strong>Summary</strong> In natural and agricultural ecosystems, plants are exposed to a wide diversity of abiotic and biotic stresses such as drought, salinity, pathogens and insect herbivores. Under natural conditions, these stresses do not occur in isolation but commonly occur simultaneously. However, plants have developed sophisticated mechanisms to survive and reproduce under suboptimal conditions. Genetic screenings and molecular genetic assays have shed light on the molecular players that provide resistance to single biotic and abiotic stresses. Induced defenses are attacker specific and phytohormones play an essential role in tailoring these defense responses. Because phytohormones display antagonistic and synergistic interactions, the question emerges how plants elicit an effective defense response when exposed to conflicting signals under multiple attack. Recent studies have shed light on this issue by studying the effects of combinations of stresses at the phenotypic, transcriptomic and genetic level. These studies have concluded that the responses to combined stresses can often not be predicted based on information about responses to the single stress situations or the phytohormones involved. Thus, combined stresses are starting to be regarded as a different state of stress in the plant. Studying the effects of combinations of stresses is relevant since they are more representative of the type of stresses experienced by plants in natural conditions. In a coordinated effort, responses of <em>Arabidopsis thaliana</em> to a range of abiotic and biotic stresses and stress combinations have been explored at the genetic, phenotypic, and transcriptional level. For this purpose we used an ecogenomic approach in which we integrated the assessment of phenotypic variation and Genome-Wide Association (GWA) analysis for a large number of <em>A. thaliana </em>accessions with an in-depth transcriptional analysis. The focus of this thesis is especially on (but not limited to) three stresses, i.e. drought, herbivory by <em>Pieris rapae</em> caterpillars, and infection by the necrotrophic fungal pathogen <em>Botrytis cinerea</em>. These stresses were chosen because the responses of <em>A. thaliana </em>to these three stresses are highly divergent but at the same time regulated by the plant hormones JA and/or ABA. Consequently, analysis of responses to combinatorial stresses is likely to yield information on signaling nodes that are involved in tailoring the plant’s adaptive response to combinations of these stresses. Responses of <em>A. thaliana </em>to other biotic and abiotic stresses are included in an integrative study (Chapter 6). We first investigated (Chapter 2) the extent of natural variation in the response to one abiotic stress (drought), four biotic stresses (<em>Pieris rapae </em>caterpillars<em>, Plutella xylostella </em>caterpillars<em>, Frankliniella occidentalis </em>thrips<em>, Myzus persicae </em>aphids) and two combined stresses (drought plus <em>P. rapae,</em> and <em>B. cinerea</em> plus <em>P. rapae</em>). Using 308 <em>A. thaliana </em>accessions originating from Europe, the native range of the species, we focused on the eco-evolutionary context of stress responses. We analyzed how the response to stress is influenced by geographical origin, genetic relatedness and life-cycle strategy, <em>i.e.</em> summer versus winter annual. We identified heritable genetic variation for responses to the different stresses. We found that winter annuals are more resistant to drought, aphids and thrips and summer annuals are more resistant to <em>P. rapae </em>and <em>P. xylostella </em>caterpillars and to the combined stresses of drought followed by <em>P. rapae </em>and infection by the fungus <em>B. cinerea </em>followed by herbivory by <em>P. rapae</em>. Furthermore, we found differential responses to drought along a longitudinal gradient. We further investigated, using <em>A. thaliana</em> accession Col-0, how phenotypic and whole-genome transcriptional responses to one stress are altered by a preceding or co-occurring stress (Chapters 3 and 4). The whole-transcriptomic profile of <em>A. thaliana </em>triggered by single and combined abiotic (drought) and biotic (herbivory by caterpillars of <em>P. rapae</em>, infection by <em>B. cinerea</em>) stresses was analyzed by RNA sequencing (RNA-seq). Comparative analysis of plant gene expression triggered by single and double stresses revealed a complex transcriptional reprogramming. Mathematical modelling of transcriptomic data, in combination with Gene Ontology analysis highlighted biological processes specifically affected by single and double stresses (Chapters 3). For example, ethylene (ET) biosynthetic genes were induced at 12 h by <em>B. cinerea</em> alone or drought followed by <em>B. cinerea</em> inoculation. This induction was delayed when plants were pretreated with <em>P. rapae</em> by inducing ET biosynthetic genes only 18 hours post inoculation. Other processes affected by combined stresses include wound response, systemic acquired resistance (SAR), water deprivation and ABA response, and camalexin biosynthesis. In Chapter 4, we focused on the stress imposed by <em>P. rapae</em> herbivory alone or in combination with prior exposure to drought or infection with <em>B. cinerea</em>. We found that pre-exposure to drought stress or <em>B. cinerea</em> infection resulted in a significantly different timing of the caterpillar-induced transcriptional changes. Additionally, the combination of drought and <em>P. rapae </em>induced an extensive downregulation of <em>A. thaliana</em> genes involved in defence against pathogens. Despite the larger reduction in plant biomass observed for plants exposed to drought plus <em>P. rapae</em> feeding compared to <em>P. rapae</em> feeding alone, this did not affect weight gain of this specialist caterpillar. In Chapter 5, we used univariate GWA to (1) understand the genetic architecture of resistance to the different stresses and (2) identify regions of the genome and possible candidate genes associated with variation in resistance to those stresses. In Chapter 5 a subset of the stresses addressed in Chapter 1 (<em>i.e.</em> drought, herbivory by <em>P. rapae</em> and <em>P. xylostella</em>, and the combined stresses drought plus <em>P. rapae</em> and <em>B. cinerea</em> plus <em>P. rapae</em>) were investigated. Results from GWA were integrated with expression data generated in Chapters 3 and 4 or available from the literature. We identified differences in genetic architecture and QTLs underlying variation in resistance to (1) <em>P. rapae</em> and<em>P. xylostella</em> and (2) resistance to <em>P. rapae</em> and combined stresses drought plus <em>P. rapae</em> and <em>B. cinerea</em> plus <em>P. rapae</em>. Furthermore, several of the QTLs identified contained genes that were differentially expressed in response to the relevant stress. For example, for <em>P. xylostella </em>one of the QTLs contained only two genes encoding cysteine proteases (<em>CP1</em> and <em>CP2</em>). The expression data indicated that these genes were induced by <em>P. rapae</em> and <em>P. xylostella </em>herbivory. In Chapter 6, the genetic architecture underlying plant resistance to 11 single stresses and some of their combinations was investigated. First, the genetic commonality underlying responses to different stresses was investigated by means of genetic correlations,, revealing that stresses that share phytohormonal signaling pathways also share part of their genetic architecture. For instance, a strong negative genetic correlation was observed between SA and JA inducers. Furthermore, multi-trait GWA identified candidate genes influencing the response to more than one stress. For example, a functional <em>RMG1</em> gene seems to be associated with susceptibility to herbivory by <em>P. rapae</em> and osmotic stress since loss of function mutants in <em>RMG1</em> displayed higher resistance to both stresses. Finally, multi-trait GWA was used to identify QTLs with contrasting and with similar effects on the response to (a) biotic or abiotic stresses and (b) belowground or aboveground stresses. Finally, In Chapter 7, I discuss the feasibility of obtaining plants that are resistant to multiple stresses from the point of view of genetic trade-offs and experimental limitations. The ecogenomic approach for gene discovery taken in this thesis is discussed, and recommendations are especially given on the use of herbivorous insects in quantitative genetic studies of stress resistance. Furthermore, alternatives to the use of insects in quantitative genetic studies of stress resistance are discussed and proposed. Finally, I discuss the feasibility of using an ecogenomic approach to study stress responses in other plant species than the model plant of molecular genetics, <em>A. thaliana</em>. A wealth of candidate genes was generated by taking an ecogenomic approach, in particular transcriptome analysis and GWA analysis. Functional characterization of these genes is a next challenge, especially in the context of multiple stress situations. These genes constitute a rich source of potential factors important for resistance to abiotic, biotic and combined stresses that in the future may be applied for crop improvement.