INTRODUCTION

Pathogens and herbivores play an important role in maintaining biodiversity in natural populations (Bever, Mangan, & Alexander, 2015; Kursar et al., 2009). The threats imposed by pathogens on humans and on managed, food production systems have motivated research that aims to predict where pathogens will occur and how risks of infection evolve (Gilligan, 2002; Koff, 1992; Woolhouse, Taylor, & Haydon, 2001). Pathogens can only occur where they have susceptible hosts, and hence, resistance diversity is the key determinant of disease dynamics. Thus, our ability to understand how diversity in resistance is generated and maintained underlies our ability to predict and prevent disease emergence and epidemics. In agriculture increasing the diversity of crops - even from a monoculture to a mixture of two cultivars - has been shown to reduce disease levels significantly (Mundt, 2002b; Zhu et al., 2000). Natural host populations typically support diversity in resistance phenotypes (Laine, Burdon, Dodds, & Thrall, 2011; Salvaudon, Giraud, & Shykoff, 2008), and limited data available to date show that increasing resistance diversity decreases disease risk also in the wild (Jousimo et al., 2014a).
Hosts and pathogens are assumed to coevolve through Red Queen dynamics, where the pathogen overcomes host’s defenses and the host in turn responds with new counter-defenses (Hamilton, 1980; Jaenike, 1978). Theory predicts such reciprocal coevolutionary selection to be a powerful mechanism for maintaining diversity in both host and parasite populations, as the selection rate for resistance depends on the frequency of parasite alleles, and vice versa, in a negative indirect frequency‐dependent manner (Bergelson, Kreitman, Stahl, & Tian, 2001; Leonard, 1977). There are numerous examples of pathogens overcoming host resistance mechanisms, both from agriculture and from the wild (Mundt, 2002a, 2014). While evidence of resistance evolving under pathogen attack in the wild is scarce (Laine, 2006), there is ample support for coevolution from local adaptation studies where parasite/host fitness is measured in sympatry vs. allopatry (Greischar & Koskella, 2007; Hoeksema & Forde, 2008). To date, a handful of ground-breaking studies have demonstrated that fluctuations in resistance and infectivity in natural systems match the predictions of coevolutionary selection (Decaestecker et al., 2007; Gómez & Buckling, 2011; Thrall et al., 2012).
The interaction between plants and their pathogens is mediated by complex defense mechanisms, incorporated in several layers of defense. Thick and waxy cell walls are an example of a mechanical defense barrier against pathogen invasion (Miedes, Vanholme, Boerjan, & Molina, 2014). Upon the arrival of a pathogen, the pathogen-associated molecular patterns (PAMPs) trigger the so-called PAMP-triggered immunity (PTI) response, which can stop the pathogen infection even before it begins. If the pathogen succeeds in overcoming these first two defense layers, the next layer is effector triggered immunity (ETI), involving the recognition of pathogen virulent factors (effector proteins), either directly or indirectly by recognizing the modifications that they make to the plant proteins (Jones & Dangl, 2006). Many of the proteins involved in intracellular pathogen recognition belong to nucleotide-binding–leucine-rich repeat (NLR) receptors (Monteiro & Nishimura, 2018). After pathogen recognition a multitude of different signaling pathways, including production of reactive oxygen species, elevated Ca2+ and MAP kinases lead to activation of plant defenses. These defenses include the induction of stress hormones salicylic acid, jasmonic acid and ethylene, as well as extensive transcriptional re-programming ultimately resulting in the production of defensive compounds, such as toxic secondary metabolites and antimicrobial chemicals and enzymes. Against some pathogens, plants also activate the hyper-sensitive response, programmed cell death, in which the plant rapidly kills the cells surrounding the infection to prevent the spread to nearby tissues (Coll, Epple, & Dangl, 2011; Egorov & Odintsova, 2012).
Within these defense responses, NLR proteins play a key role. They are involved in recognition of the pathogen’s effector proteins both directly and indirectly, as well as in triggering the plant immune response (Meunier & Broz, 2017). NLRs have also been shown to be involved in signaling and transcript regulation (Chisholm, Coaker, Day, & Staskawicz, 2006; Jones & Dangl, 2006). Moreover, NLR genes play an important role in local adaptation and habitat expansion of plants (Stam et al., 2017; Thrall et al., 2012). The antagonistic interaction between plant NLR and pathogen effector proteins is considered to have a profound effect on the evolution of both organisms, shaping their genomes and gene repertoire (Upson, Zess, Bialas, Wu, & Kamoun, 2018). NLRs usually form large tandemly arrayed gene families and hence questions regarding their origins and evolutionary history have been under active research in both plants and animals (Andolfo et al., 2019; Borrelli et al., 2018). The numbers of identified NLRs differ substantially within and between plant families (Baggs, Dagdas, & Krasileva, 2017), for example Arabidopsis thaliana (Arabidopsis) contains between 165 to 251 NLR genes (Shao et al., 2016; Van de Weyer et al., 2019) and crop species such as wheat, barley, rice, tomato and potato contain 627, 224, 438,137 and 309 NLRs, respectively (Sarris, Cevik, Dagdas, Jones, & Krasileva, 2016). In A. thaliana there is evidence of widespread positive selection in the core NLRs shared among accessions, especially in the canonical NLR domains (Van de Weyer et al., 2019), while a pioneering study on wild tomato revealed high NLR diversity with a small subset of NLR genes driving local adaptation to pathogens (Stam et al., 2017).
A current key challenge in molecular ecology is to understand the role of pathogen-imposed selection on generating NLR diversity. Exploring the breadth and depth of plant NLR natural variation has the potential to increase our understanding of how NLR diversity is generated and maintained, and to establish a toolbox of deployable disease resistance traits (Monteiro & Nishimura, 2018). In natural plant populations, neither pathogen epidemiology nor host resistance is under human management in contrast to agricultural systems where disease is managed both via resistance breeding and fungicides. Hence, natural populations have the potential to offer unique insight into the processes generating NLR diversity. Our study is focused on the interaction betweenPlantago lanceolata and its fungal pathogen Podosphaera plantaginis . Previous studies have detected considerable phenotypic variation in P. lanceolata resistance against P. plantaginis (Laine, 2004); diversity is shown to accumulate in the well-connected populations across the landscape (Hockerstedt, Siren, & Laine, 2018), and has a direct negative impact on disease dynamics (Jousimo et al., 2014b). Moreover, there is evidence of coevolution in this interaction (Laine, 2005, 2006, 2008).
Here, we carried out a controlled experiment where five P. lanceolata individuals were inoculated with the same P. plantaginis strain. The aim of our study was to characterize the transcriptional responses and the expression pathways activated in response to the pathogen treatment, as well as in resistant versus susceptible phenotypes. For this purpose, we carried out the firstde novo transcriptome assembly for P. lanceolata and used it to characterize the induced NLR repertoire in P. lanceolataand test whether any particular NLRs show signs of selection. We found that each plant genotype demonstrated a unique gene expression profile in response to the pathogen and discovered a diverse NLR repertoire inP. lanceolata which is consistent with the high phenotypic resistance diversity uncovered in earlier studies.