ReviewFree Access icon

Septoria Nodorum Blotch of Wheat: Disease Management and Resistance Breeding in the Face of Shifting Disease Dynamics and a Changing Environment

    Affiliations
    Authors and Affiliations
    • Rowena C. Downie1 2
    • Min Lin3
    • Beatrice Corsi1
    • Andrea Ficke4
    • Morten Lillemo3
    • Richard P. Oliver5
    • Huyen T. T. Phan6
    • Kar-Chun Tan6
    • James Cockram1
    1. 1John Bingham Laboratory, NIAB, Cambridge, CB3 0LE, United Kingdom
    2. 2Department of Plant Sciences, University of Cambridge, Cambridge, CB2 3EA, United Kingdom
    3. 3Norwegian University of Life Sciences, Ås NO-1432, Norway
    4. 4Norwegian Institute for Bioeconomy Research, Ås NO-1432, Norway
    5. 5Curtin University, Bentley 6102, Perth, WA, Australia
    6. 6Centre for Crop and Disease Management, School of Molecular and Life Sciences, Curtin University, Bentley 6102, Perth, WA, Australia

    Abstract

    The fungus Parastagonospora nodorum is a narrow host range necrotrophic fungal pathogen that causes Septoria nodorum blotch (SNB) of cereals, most notably wheat (Triticum aestivum). Although commonly observed on wheat seedlings, P. nodorum infection has the greatest effect on the adult crop. It results in leaf blotch, which limits photosynthesis and thus crop growth and yield. It can also affect the wheat ear, resulting in glume blotch, which directly affects grain quality. Reports of P. nodorum fungicide resistance, the increasing use of reduced tillage agronomic practices, and high evolutionary potential of the pathogen, combined with changes in climate and agricultural environments, mean that genetic resistance to SNB remains a high priority in many regions of wheat cultivation. In this review, we summarize current information on P. nodorum population structure and its implication for improved SNB management. We then review recent advances in the genetics of host resistance to P. nodorum and the necrotrophic effectors it secretes during infection, integrating the genomic positions of these genetic loci by using the recently released wheat reference genome assembly. Finally, we discuss the genetic and genomic tools now available for SNB resistance breeding and consider future opportunities and challenges in crop health management by using the wheat–P. nodorum interaction as a model.

    Septoria nodorum blotch (SNB) is a fungal disease of wheat (Triticum aestivum), a key crop underpinning global food security. SNB is caused by the necrotrophic fungal pathogen Parastagonospora nodorum (syn. Phaeosphaeria nodorum [E. Müll.], syn. Leptosphaeria nodorum [E. Müll.], syn. Stagonospora nodorum [Berk.], syn. Septoria nodorum [Berk.]) and is prevalent in wheat growing environments with high, or periodically high, rainfall such as regions in Australia, Canada, Scandinavia, Central and Eastern Europe, the eastern United States, and South America. Compared with biotrophic pathogens, which require living host tissue, necrotrophs actively kill host tissue during colonization, subsequently living on the contents of the dead or dying host cells (Laluk and Mengiste 2010). The visual symptoms of SNB are chlorosis and necrosis of wheat leaf tissue (often in the form of necrotic lesions surrounded by chlorosis, later developing into irregular dark brown lesions), as well as discoloration and necrosis of the glumes, referred to as leaf blotch and glume blotch, respectively (Fig. 1) (Solomon et al. 2006). Leaf blotch reduces the plant surface area capable of photosynthesis, therefore limiting overall crop growth and yield, whereas glume blotch directly affects grain quality. Because of such damage, SNB is known to cause yield losses of ≤30% (Bhathal et al. 2003). In practice, SNB disease often occurs in combination with other necrotrophic fungal diseases such as septoria tritici blotch (STB, caused by Zymoseptoria tritici) and tan spot (TS, caused by Pyrenophora tritici-repentis). When such disease complexes occur, it can often be difficult to visually determine which necrotrophic diseases are present. However, quantitative polymerase chain reaction molecular assays for P. nodorum (Oliver et al. 2008), Z. tritici (Bearchell et al. 2005), and P. tritici-repentis (Antoni et al. 2010) are now available, helping to distinguish the contributors to coinfections of wheat. Additionally, an internal transcribed spacer-restriction fragment length polymorphism test has been developed that distinguishes between necrotrophic pathogens including P. nodorum and P. tritici-repentis (Hafez et al. 2020). Before the 1980s, P. nodorum was the dominant pathogen of the leaf blotch complex in Europe (Bearchell et al. 2005). However, SNB has undergone changes in its regional prevalence in recent decades. For example, over the last 30 years there has been a focal shift in much of northwestern Europe from P. nodorum to Z. tritici (Bearchell et al. 2005; Shaw et al. 2008). The underlying reasons for this change are not fully understood and have been attributed to increased levels of Z. tritici host susceptibility, changes in climate, higher use of fertilizers use, and increased SO2 emissions (Shaw et al. 2008; West et al. 2012). It is notable that in Norway, P. nodorum is still the major necrotrophic fungal pathogen of wheat and that sulfur pollution has not been reported to be higher in Norway than in any other European country in which Z. tritici dominates the wheat leaf blotch complex (Lin et al. 2020a). One possibility is that the overall SNB to STB shift is caused by Z. tritici being better at adapting to fungicides, although this hypothesis warrants further investigation. Nevertheless, P. nodorum remains an important pathogen of wheat worldwide, and it appears to be moving into new niches. For example, in 2017 it was observed for the first time on emmer wheat (T. dicoccoides) in Turkey, and because of changing climatic conditions, SNB has now become a major problem in Himachal Pradesh, India (Cat et al. 2018; Katoch et al. 2019).

    FIGURE 1

    FIGURE 1 Septoria nodorum blotch symptoms in bread wheat. A, On leaves. B, On the spikelets of a wheat inflorescence (ear).

    Download as PowerPoint

    P. NODORUM LIFE CYCLE, INFECTION PROCESS, AND EPIDEMICS

    P. nodorum is a fungal pathogen belonging to the Ascomycota as a member of the class Dothideomycetes. As the first of the Dothiodeomycete class of fungal pathogens to have its genome sequenced (37 Mbp; Hane et al. 2007), P. nodorum became established as a model for the narrow host range necrotrophic pathogen life cycle. It is known mostly as a wheat pathogen but has also been reported to occasionally infect the related cereal crop barley (Hordeum vulgare) but with less damage (reviewed by Cunfer 2000), as well as wild grasses (Zhang and Nan 2018). P. nodorum is a necrotrophic fungal pathogen that assimilates nutrients released after host cell death (De Wit et al. 2009). A recent reclassification of fungal and oomycete pathogens (Hane et al. 2020) differentiated a new grouping described as narrow host range polymertrophs to which P. nodorum belongs. This group has a narrow host range (unlike Botrytis cinerea) and induces immediate cell death so that polymeric plant substances become available for assimilation. This group typically produces proteinaceous effectors to fuel disease progression and trigger the plant’s receptors to promote sensitivity and tissue death (De Wit et al. 2009). P. nodorum has both asexual and sexual cycles (Fig. 2). As part of the asexual cycle, fruiting bodies, called pycnidia, form in lesions on the leaf to promote spore development for local dispersal. In contrast, the sexual life cycle produces ascospores, derived from pseudothecia, that allow long-distance aerial dispersion. The presence of both sexual and asexual reproduction mechanisms is hypothesized to provide P. nodorum with a high evolutionary potential, resulting in increased diversity and fast clonal reproduction of favorable genotypes (Ruud and Lillemo 2018). The primary inoculum of SNB is mostly forcibly discharged ascospores originating from wheat debris, although it is also seed transmitted. Reduced tillage (the practice of minimizing disturbance of the soil by allowing crop stubble to remain on the ground rather than being incorporated into the soil or discarded) is advocated to reduce soil erosion and limit water evaporation. However, this practice leads to higher amounts of infected wheat straw on the soil surface, which can serve as primary inoculum (Ficke et al. 2018). Once the pathogen has established the initial infection on a plant, large amounts of pycnidiospores can be produced and subsequently spread by rain splash. Indeed, the high density of wheat fields makes it easier for pycnidiospores to spread to neighboring plants. Semidwarf varieties of wheat may have a higher risk of secondary P. nodorum infection because of the close vertical spacing of the leaves as conidia, produced by pycnidia, are sent on an upward trajectory by water droplets (Bahat et al. 1980). This is particularly relevant because the majority of modern wheat varieties have a short semidwarf stature.

    FIGURE 2

    FIGURE 2 Illustration of the Parastagonospora nodorum infection cycle on wheat. Initial infection of wheat seedlings is via P. nodorum ascospores present in infected stubble, or via seeds infected with P. nodorum mycelia, which produce pycnidiospores under wet or humid conditions. Pycnidiospores produced as a result of this initial infection can then be spread via rain splash or wind, causing secondary infection further up the wheat canopy as the crop matures and can result in infection of the wheat ears.

    Download as PowerPoint

    GENETIC STRUCTURE OF THE P. NODORUM PATHOGEN POPULATION

    Because P. nodorum undergoes frequent sexual reproduction, the resulting genetic recombination results in high genetic diversity in the pathogen population (McDonald and Linde 2002). Isolates from the Middle East have been found to possess the highest genetic diversity globally, indicating that it is highly probable that the Fertile Crescent was the P. nodorum center of origin (Ghaderi et al. 2020). Over the years, studies of P. nodorum population structure have been undertaken with a variety of different molecular marker types. Of the various populations investigated to date, sourced from a wide range of geographic locations, studies have typically found little population substructure and high genetic diversity (Blixt et al. 2008; Keller et al. 1997; Lin et al. 2020a; McDonald et al. 2012; Murphy et al. 2000; Stukenbrock et al. 2006). For example, genetic studies carried out on P. nodorum populations collected from Europe and the United States found evidence of high gene flow but little evidence of genetic differentiation between populations (Keller et al. 1997), with similar results observed for populations from Australia (Murphy et al. 2000) and Norway (Lin et al. 2020a). Indeed, high levels of genetic diversity have even been found among isolates collected from the same lesion (McDonald et al. 1994). The most notable investigation to go against this general trend was an analysis of an international P. nodorum population sourced from five continents, where moderate differentiation was observed between geographically divided populations (Stukenbrock et al. 2006). More recently, Richards et al. (2019) carried out a comprehensive analysis of the population structure and genome evolution of 197 P. nodorum isolates collected across the United States from durum, spring, and winter wheat varieties, finding evidence of two P. nodorum populations that corresponded to the upper midwestern and southeastern United States. Interestingly, most isolates in the southeastern United States population lacked the effector SnToxA. This finding correlated with the lack of the ToxA effector sensitivity gene Tsn1 in winter wheat varieties that were widely planted in the region, suggesting that host genotype is a strong driver of the maintenance of effector genes.

    Notably, most regional P. nodorum population genetic studies have been carried out on isolates sampled across a narrow timeframe and thus offer limited insight into potential changes in the population structure over time. However, recently Phan et al. (2020) examined the population structure of 155 P. nodorum isolates collected over a 44-year period across the South-Western Australian wheat growing region. Analysis of genetic polymorphisms with 28 simple sequence repeat markers revealed that the population consisted of genetically distinct groups. Most isolates sampled were attributed to core groups that possessed the highest level of genetic diversity in the Australian population, and these groups were found throughout locations and times. Isolates belonging to non-core groups possessed a much lower level of genetic diversity, with limited distribution across locations and time. It was also observed that changes in group genotypes occurred during periods that coincided with major changes in the mass adoption of popular wheat cultivars across large areas of the Australian wheat cultivation zone. It was hypothesized that core groups maintain genetic variability, whereas non-core groups emerge in response to large-scale changes in cultivar near-monocultures. Finally, work investigating the genetic diversity of P. nodorum and the closely related pathogen species P. avenaria f. sp. tritici 1 (Pat1) shows evidence of hybridization at a frequency of about 4%, indicating that such gene transfer could be an additional source of genetic diversity in regions in which the ranges of the two species overlap (McDonald et al. 2012).

    SNB DISEASE MANAGEMENT

    Disease management of SNB includes cultivar resistance (considered in more detail in the next section), fungicide treatment, seed cleaning, and stubble management. Despite decades of breeding effort, all current wheat cultivars retain a significant level of susceptibility (Aguilar et al. 2005). Reduced tillage practices are becoming increasingly common around the world, and significant correlations have been observed between the amount of residue and SNB disease severity in the field (Mehra et al. 2015). Residue management can decrease the amount of primary inoculum and reduce disease severity (Solomon et al. 2006). SNB transmission via seed is regularly reported in some parts of the world, such as the eastern United States, but rarely elsewhere (Bennett et al. 2007). Seed fungicide treatment, directed primarily to control bunts and smuts, seems to be an efficient way to eradicate SNB from seed stocks. However, fungicidal control of foliar and glume SNB is more problematic. SNB typically occurs in combination with other diseases (tan spot, STB, yellow rust, and powdery mildew) and is not normally the predominant disease. The conditions where SNB is dominant are currently limited to particular geographic locations where yield is typically <4 tonnes per hectare and on lower-value feed cereals such as triticale, where fungicidal applications are limited in number and dosage.

    Before its relative decline in much of northwestern Europe in or around 2000, P. nodorum was considered a model pathogen for fungicide discovery (Dancer et al. 1999). All the current major fungicide classes are efficient at controlling SNB: sterol demethylation inhibitors (DMIs), Qo inhibitors (QoIs), and succinate dehydrogenase inhibitors (SDHIs). Reports of fungicide resistance in SNB are rare. Resistance to QoI fungicides in Sweden was reported in isolates collected between 2003 and 2005 (Blixt et al. 2009), and resistance to DMI fungicides has been reported in isolates collected before 2000 in Denmark, Sweden, and Switzerland (Pereira et al. 2017; see also https://www.biorxiv.org/content/10.1101/2020.03.26.010199v1.full). To our knowledge, no reports of resistance to SDHI fungicides have been made.

    Fungicide resistance management focuses on reducing the selection pressure for resistance by minimizing dosage and number of applications and by using mixtures and alternations (Jørgensen et al. 2017). The primary foci of foliar fungicide application in wheat are normally yellow rust, STB, and powdery mildew. The latter two diseases are particularly adept at evolving resistance (Oliver and Hewitt 2014). QoI resistance was detected in both pathogens within 2 years of QoI application in 2001 (Bartlett et al. 2002). Control of STB by DMIs was substantially compromised by about 2010 (Cools and Fraaije 2013). In the last decade, SDHIs has become the main weapon against STB, but resistance was well developed by 2016 in the United Kingdom and Ireland (Dooley et al. 2016). Because SNB is not typically the only, or most dominant, pathogen among the disease complexes present in most geographic regions, it is possible that SNB has been inadvertently protected against resistance evolution by the development of resistance in the more damaging pathogen, forcing a change in fungicide regime. New fungicides were introduced, lower dosages applied, and either mixtures or rotations carried out. As a result, SNB is not commonly subject to sustained pressure from a single mode of action class and has therefore probably developed resistance only rarely.

    GENETICS OF WHEAT SENSITIVITY TO P. NODORUM: NECROTROPHIC EFFECTORS AND HOST RESPONSE

    Although chemical control is an important part of SNB disease management, the use of cultivars with increased genetic resistance helps to underpin more economically and environmentally sustainable wheat production. Resistances to SNB leaf blotch and glume blotch are quantitatively inherited but are reported to be controlled by different genetic mechanisms (Chu et al. 2010; Wicki et al. 1999). Increased disease severity is also associated with shorter plant height and later plant maturation. However, residual resistance that is not associated with these traits is identifiable. It is this residual genetic resistance, along with the identification of host-specific gene-for-gene interactions determining the P. nodorum–wheat pathosystem (Liu et al. 2004), that provide immediate opportunities to further explore host genetic resistance in wheat breeding (Ruud and Lillemo 2018).

    Necrotrophic fungal pathogens are known to secrete effectors (typically proteins, but also low-molecular-weight phytotoxic metabolites) during host infection, which act as virulence factors facilitating disease development. The presence of effectors, also known as host selective toxins, was first described in 1933 through the study of the host–pathogen interaction between Alternaria alternata and Japanese pear, Pirus serotine (Tanaka 1933). Since then, effectors and their corresponding host sensitivity loci have been identified in numerous necrotrophic fungal and bacterial plant pathogens (reviewed by Laluk and Mengiste 2010). The necrotic response in a sensitive host plant is hypothesized to help pathogen colonization, promoting infection and ultimately providing a rich nutrient source (Oliver and Solomon 2010). This is known as effector-triggered susceptibility and is genetically induced via an inverse gene-for-gene system (Friesen et al. 2007). Understanding the genetics of host sensitivity to such effectors provides the opportunity to break down at least some components of the genetics of field resistance into their constitutive parts. P. nodorum is thought to derive nutrients from dying plant tissue, using secreted effectors. These effectors induce a hypersensitive response in the host, which takes the form of programmed cell death (Friesen et al. 2007; Liu et al. 2009; Oliver et al. 2012). Eight P. nodorum effectors have been described to date, designated ToxA, Tox1, Tox2, Tox3, Tox4, Tox5, Tox6, and Tox7, along with nine corresponding major wheat sensitivity loci: Tsn1 (Faris et al. 2010), Snn1 (Shi et al. 2016), Snn2 (Friesen et al. 2007), Snn3-B1 and Snn3-D1 (Friesen et al. 2008; Zhang et al. 2011), Snn4 (Abeysekara et al. 2012), Snn5 (Friesen et al. 2012), Snn6 (Gao et al. 2015), and Snn7 (Shi et al. 2015), respectively. Of these, only three effectors (ToxA, Tox1, and Tox3) and two host sensitivity loci (Tsn1 and Snn1) have been identified at the gene level. In addition to these major host loci, several minor effector sensitivity QTLs have been identified in wheat (Supplementary Table S1) (Cockram et al. 2015; Downie et al. 2018; Lin et al. 2020b; Phan et al. 2016).

    ToxA–Tsn1 interaction.

    ToxA was first discovered to be secreted by P. nodorum in 2006 (Friesen et al. 2006) and found to have 99.7% DNA sequence similarity to the previously identified ToxA gene from P. tritici-repentis (PtrToxA). Because of the monomorphism of PtrToxA compared with the high levels of ToxA diversity, it is thought ToxA was introduced into the P. tritici-repentis genome through interspecific gene transfer from P. nodorum (Friesen et al. 2006). The corresponding host sensitivity locus, Tsn1, was first discovered in 1996 as conferring sensitivity to PtrToxA (Faris et al. 1996), and later confirmed as the corresponding host sensitivity locus for P. nodorum ToxA (Liu et al. 2006). This interaction was found to significantly contribute to disease incidence, accounting for ≤62% of disease severity at the seedling stage (Liu et al. 2006) and ≤20% at the adult plant stage (Friesen et al. 2009). Tsn1 is typically present at high frequencies in wheat germplasm (e.g., 59% of Canadian varieties representing wheat development over the last century) (Hafez et al. 2020). Tsn1 encodes a predicted protein containing three predicted domains: a serine/threonine protein kinase (with ATP binding, substrate binding site, and activation loop), a nucleotide binding site, and 24 leucine-rich repeats (Faris et al. 2010). Nucleotide binding site leucine-rich repeats form the largest class of plant resistance (R) genes and are well documented as controlling race-specific resistance to biotrophic fungal pathogens (Dubey and Singh 2018). Tsn1 is localized to the chloroplast and does not directly interact with ToxA (Faris et al. 2010). However, ToxA has been shown to interact with the dimeric PR-1-type pathogenesis-related protein, TaPR-1-5, to activate Tsn1-controlled cell death pathways (Breen et al. 2016). Tsn1 expression is subjected to regulation by light and the circadian clock, providing a possible explanation for the light-dependent nature of the ToxA–Tsn1 interaction (Faris et al. 2010; Manning and Ciuffetti 2005). Recently, it has been shown that another wheat and barley pathogen, Bipolaris sorokiniana, the cause of spot blotch, also possesses a ToxA gene that probably originated from P. nodorum, pointing to a selective advantage of carrying the virulence factor ToxA (Friesen et al. 2018).

    Tox1–Snn1 interaction.

    Tox1 was first characterized as a host-selective effector produced in P. nodorum culture filtrates interacting with the wheat sensitivity locus Snn1 on chromosome 1B (Liu et al. 2004). Tox1 encodes a cysteine-rich protein with 117 amino acids that is light dependent and critical for fungal penetration (Liu et al. 2012) and serves a dual function: binding host chitinases to protect fungal infection and causing host tissue death to promote infection (Liu et al. 2016). The Tox1–Snn1 interaction was found to contribute ≤58% and ≤19% of SNB at juvenile and mature plant stages, respectively (Liu et al. 2004; Phan et al. 2016). The recent map-based cloning of Snn1 found it to encode a galacturonic acid binding wall-associated kinase (WAK) and to possess calcium binding epidermal growth factor and serine/threonine kinase domains (Shi et al. 2016). WAK proteins are known to be members of pattern recognition receptors, which directly interact with pathogen-associated molecular patterns, such as oligogalacturonides, which trigger programmed cell death and are involved in plant defense mechanisms against biotrophic pathogens (Brutus et al. 2010).

    Tox3–Snn3-B1 interaction.

    The P. nodorum effector Tox3 was first identified by Friesen et al. (2008), and the protein sequence was later characterized as a 25.8-kDa immature protein, with the first 20 residues of the 230 aa chain forming a signal peptide for secretion (Liu et al. 2009). Tox3 has six cysteine residues that form disulfide bonds, and at least one of these bonds is essential for biological function. Recent work has shown that an avirulent P. nodorum strain could become virulent with just the addition of the 693-bp intron-free Tox3 (Liu et al. 2009; Waters et al. 2011). Discovery of Tox3 led to the identification of the corresponding wheat sensitivity locus, Snn3 (more recently called Snn3-B1), on the short arm of chromosome 5B. This interaction has been shown to explain 24% of the phenotypic variation in field SNB resistance/susceptibility and >51% of the variation in seedling inoculation (Ruud et al. 2017). Culture filtrate containing Tox3 was first produced with a wild-type P. nodorum isolate, SN15, and host sensitivity was genetically mapped with the BR34 × Grandin wheat population (Friesen et al. 2008) and later confirmed in subsequent studies (Downie et al. 2018; Lin et al. 2020b; Phan et al. 2016; Shi et al. 2016). Although a Snn3-B1 homeologue was found on chromosome 5D in the diploid wild wheat relative Aegilops tauschii (Snn3-D1) (Zhang et al. 2011), a corresponding locus on the D subgenome of hexaploid wheat has not been reported. As was the case for ToxA, yeast-two-hybrid studies have shown that the Tox3 protein interacts with PR-1 proteins (Breen et al. 2016).

    P. NODORUM EFFECTORS HIJACK PATHWAYS INVOLVED IN BIOTROPHIC PATHOGEN HOST DEFENSE SIGNALING

    Given that Tsn1 and Snn1 both encode classes of proteins that are well known to control disease resistance in biotrophic pathogens, it is hypothesized that P. nodorum has evolved to hijack existing pathways in order to become a susceptibility pathway for necrotrophs (Faris and Friesen 2020; Faris et al. 2010; Shi et al. 2016). Specifically, it is thought that host recognition of SnTox1 activates pathogen-associated molecular pattern–triggered immunity (PTI) and that ToxA/PtrToxA recognition activates effector-triggered immunity (ETI). The finding that Tox1 does not enter the plant cell (Liu et al. 2016) indicates that its recognition is mediated via host membrane-bound proteins. This fits both with the prediction that Snn1 spans the host cell membrane and contains extracellular binding domains (Liu et al. 2016; Shi et al. 2016) and with the interaction of Snn1 with Tox1 in vitro (Shi et al. 2016). As noted by Shi et al. (2016), although the expression patterns of PTI and ETI pathways overlap, the expression patterns of certain classes of genes commonly differ. Activation of mitogen-activated protein kinase (MAPK) genes has been shown to be transient in PTI responses, whereas their expression is more prolonged during ETI (Tsuda and Katagiri 2010). Notably, the rapid and transient upregulation of TaMAPK3 in a compatible Snn1–Tox1 interaction within 15 min of Tox1 infiltration further implicates the PTI pathway (Shi et al. 2016). Finally, it has been noted that wheat varieties carrying both Tsn1 and Snn1 have higher levels of necrosis than varieties carrying either Tsn1 or Snn1 alone, indicating that simultaneous hijacking of both the PTI and ETI pathways for necrotrophic effector (NE)-triggered susceptibility increases pathogen survival and reproduction (Chu et al. 2010; Shi et al. 2016).

    EPISTATIC INTERACTIONS BETWEEN P. NODORUM EFFECTORS AND BETWEEN HOST SENSITIVITY LOCI

    The NE–Snn model supports additive contributions to disease from each compatible interaction (Friesen et al. 2007; Tan et al. 2014). However, epistatic interactions are also evident. For example, SnTox5–Snn5 and SnTox6–Snn6 are epistatic to Snn3-B1 (Friesen et al. 2012; Gao et al. 2015). Similarly, Friesen et al. (2008) showed that the SnToxA–Tsn1 interaction is epistatic to Tox3–Snn3-B1 and that the Tox3–Snn3-B1 interaction is evident only in the absence of a compatible SnTox2–Snn2 interaction (Friesen et al. 2008). The epistatic effects on Tox3–Snn3-B1 were further explored in subsequent work via a series of effector gene deletion mutants generated in the P. nodorum strain SN15. Whereas the SnTox1–Snn1 interaction dominated seedling sensitivity with the wild-type SN15 strain, deletion of the Tox1 gene in SN15 led to an increase in Tox3 expression in the pathogen and the identification of Snn3-B1 as contributing to host sensitivity at the seedling stage (Phan et al. 2016). Furthermore, a modified strain of SN15 in which ToxA, Tox1, and Tox3 were deleted unmasked a sensitivity QTL in the region of the Snn2 locus, which was not identified with the wild-type or Tox1 mutant strain, indicating that ToxA or Tox3 could be epistatic to Snn2 (Phan et al. 2016). Unlike ToxA, it was found that Tox3 interacts with a broad range of PR-1 proteins, and it has been hypothesized that interactions with TaPR-1 proteins facilitate host infection (Breen et al. 2016). As more effectors and host sensitivity loci are cloned and their allelic diversity characterized, it is likely that the identification of new alleles at these loci will further increase the complexity of the NE–Snn network. Thus, the epistatic and allelic interactions occurring between effectors in the pathogen, and between sensitivity loci in the host, take what are largely simple gene-for-gene interactions to create a more complex set of possible interactions. Because the effect of an NE–host receptor interaction can vary depending on the presence or absence of other effectors and receptors present at the time of infection, this disease typically is quantitative and difficult to predict.

    GENETICS OF WHEAT SENSITIVITY TO P. NODORUM AT THE JUVENILE AND ADULT PLANT STAGES

    To characterize the P. nodorum–wheat pathosystem and use this information to increase SNB resistance, host resistance to target pathogens at the juvenile and adult stages is commonly investigated. Resistance to SNB at both of these plant stages is polygenic, and large genotype-by-environment interactions are observed (Fried and Meister 1987; Wicki et al. 1999). Correlation between seedling and adult plant resistance is generally reported to be low (Francki 2013; Fried and Meister 1987; Rosielle and Brown 1980; Ruud and Lillemo 2018; Shankar et al. 2008; Tommasini et al. 2007). This has been suggested to be caused by the use of different isolates in greenhouse seedling testing compared with those used in adult plant field trials (Ruud and Lillemo 2018; Ruud et al. 2017). Additionally, because the natural P. nodorum population is usually genetically diverse, it is difficult to identify representative isolates for greenhouse assays, and field testing can be affected by cross-infection with the natural P. nodorum population. Such complications mean that even where the same isolate mixtures are used for greenhouse and field trials, correlation between seedling and flag leaf disease scores can be low (0.31) or even not significant between seedling and glume blotch severity (Shankar et al. 2008). However, there are examples of high correlations when the same isolate is used for both seedling and field testing (Jönsson 1985). Genetic mapping of seedling SNB resistance has identified genetic loci on all 21 wheat chromosomes except for chromosomes 1D and 3D (Abeysekara et al. 2009; Adhikari et al. 2011; Arseniuk et al. 2004; Czembor et al. 2003; Friesen et al. 2006, 2007, 2012; Gao et al. 2015; Gonzalez-Hernandez et al. 2009; Gurung et al. 2014; Hu et al. 2019; Jighly et al. 2016; Lin et al. 2020b; Liu et al. 2004, 2015; Phan et al. 2016; Ruud et al. 2017, 2019; Rybak et al. 2017). Similarly, numerous adult plant QTLs have been identified: across 16 chromosomes for leaf blotch (1A, 1B, 2A, 2B, 2D, 3A, 3B, 4A, 4B, 5A, 5B, 6A, 6B, 7A, 7B, and 7D (Aguilar et al. 2005; Czembor et al. 2019; Francki et al. 2011, 2018, 2020; Friesen et al. 2009; Lin et al. 2020b, 2020c; Lu and Lillemo 2014; Phan et al. 2016; Ruud et al. 2017, 2019; Shankar et al. 2008) and 12 chromosomes for glume blotch (2A, 2B, 2D, 3A, 3B, 4A, 4B, 5A, 5D, 6A, 6B, and 7D (Aguilar et al. 2005; Czembor et al. 2019; Francki et al. 2018; Jighly et al. 2016; Lin et al. 2020b; Schnurbusch et al. 2003; Shankar et al. 2008; Shatalina et al. 2014; Tommasini et al. 2007; Uphaus et al. 2007). All QTLs are listed in Supplementary Table S1.

    Although it has been clear from the outset that NE–Snn interactions are relevant to seedling resistance, discussion of their importance for SNB resistance in the field is ongoing (Francki 2013). However, there is mounting evidence that at least some NE–Snn interactions also contribute to susceptibility to SNB in the field. Friesen et al. (2009) used an isolate producing both ToxA and Tox2 for spray inoculation in the field on a mapping population segregating for Tsn1, Snn2, and Snn3-B1, finding Tsn1 and Snn2 to explain 18 and 15% of the phenotypic variation for SNB resistance, respectively. Significant correlation between ToxA sensitivity and SNB disease severity have been observed in an association mapping panel under Norwegian field conditions (Ruud et al. 2019). Another study applied artificial inoculation of an isolate producing all three known NEs, showing that Snn1 explained 19% of the phenotypic variation for adult plant disease severity (Phan et al. 2016). Similarly, studies in Norway have found Snn3-B1 to affect field SNB disease susceptibility by using a biparental population (Ruud et al. 2017).

    Additional cross-comparisons of juvenile plant and adult plant sensitivity with major and minor effector and culture filtrate sensitivity loci have historically been problematic because of factors such as the large genetic intervals identified and the use of different genetic mapping populations and genetic marker systems. Genetic mapping of response to P. nodorum infection has relied mainly on different biparental wheat populations. However, more recently association mapping (Cockram et al. 2015; Downie et al. 2018; Ruud et al. 2019; Tommasini et al. 2007) and multifounder (Lin et al. 2020b, c) populations have also been used. Although each type of population comes with its own advantages and disadvantages (reviewed by Cockram and Mackay 2018), one benefit of association mapping and multifounder populations is that allelic variation at the genomic locations controlling the target traits is more likely to be sampled than might be the case in biparental populations, and the effects of these alleles are assessed in a wider range of genetic backgrounds. This method allows straightforward cross-comparison of QTLs for numerous related traits within a single genetic mapping population. Furthermore, the availability of high-density genotyping platforms and a wheat reference genome assembly (IWGSC 2018) means that cross-comparison of previously published SNB QTLs identified with different genetic mapping populations is much more straightforward to do. Here, we have used these resources to anchor previously published QTLs controlling host response to P. nodorum infection, as well as infiltration with culture filtrates and necrotrophic effectors, to the wheat physical map (Fig. 3; Supplementary Table S1). The results help highlight several interesting observations. For example, recent studies using multiparent advanced generation intercross (MAGIC) populations constructed from wheat varieties grown in UK (Mackay et al. 2014) and German (Stadlmeier et al. 2018) agronomic environments have allowed genetic control of resistance to P. nodorum, as well as sensitivity to known effectors, to be assessed in experimental populations that capture high amounts of genetic variation (Lin et al. 2020b, c). Field testing of SNB resistance identified robust colocalizing QTLs on the long arm of chromosome 2A controlling leaf blotch in the UK MAGIC (QSnb.niab-2A.3; Lin et al. 2020b) and German MAGIC (QSnb.nmbu-2A.1; Lin et al. 2020c) populations, as well as culture filtrate sensitivity QTLs that colocate to the same locus in the UK MAGIC population (Lin et al. 2020b). This chromosome 2A QTL is located within the confidence interval for the seedling resistance QTL QSnb.fcu-2A (Abeysekara et al. 2009) and the SNB resistance QTL Qsnb.cur-2AS.1 (Phan et al. 2016). However, whether these QTLs represent the same underlying locus is not currently known, and the Qsnb.cur-2AS.1 physical interval is notably large. Nevertheless, collectively these results suggest that an as-yet uncharacterized necrotrophic effector present in P. nodorum culture filtrate used by Lin et al. (2020b) interacts with the QSnb.niab-2A.3 locus and is implicated in the control of SNB resistance in adult plants. Although Lin et al. (2020b) also found a QTL controlling glume blotch to colocalize to the same genetic locus on chromosome 2A, the allelic effects at the QTL were predicted to be opposite to those for glume blotch, suggesting that a different mechanism may be involved. This finding supports previous reports that resistance to leaf blotch and glume blotch is thought to be controlled predominantly by different genetic mechanisms (Aguilar et al. 2005; Chu et al. 2010; Francki et al. 2018; Schnurbusch et al. 2003; Shankar et al. 2008).

    FIGURE 3

    FIGURE 3 Projection of published quantitative trait loci (QTLs) for Septoria nodorum leaf blotch (black), glume blotch (blue), culture filtrate and effector infiltration sensitivity at the seedling stage (brown), and seedling Parastagonospora nodorum resistance (green) onto the wheat reference genome assembly (RefSeq version 1.0; IWGSC 2018). The locations of relevant cloned wheat genes are shown in red. QTLs are named according to their publication, and full details for all QTLs are listed in Supplementary Table S1.

    Download as PowerPoint

    Analysis of additional culture filtrate sensitivity QTLs and minor-effect effector sensitivity QTLs finds several to colocate with genetic loci controlling adult plant SNB resistance (Fig. 3; Supplementary Table S1), further supporting the presence of additional effector sensitivity loci relevant to field resistance. For example, QTox3.niab-2A.1 controlling Tox3 sensitivity (Downie et al. 2018) colocates with a QTL for adult plant leaf blotch (QSnb.niab-2A.4, also controlling seeding resistance; Lin et al. 2020b), all in the same eight-founder MAGIC population. Additionally, SNB resistance QTL QSnb.niab-3A and QSnb.niab-6A.2 identified in the MAGIC population collocated with a culture filtrate sensitivity QTL (Lin et al. 2020b) and the previously reported effector sensitivity locus Snn6 (Arseniuk et al. 2004; Gao et al. 2015), respectively. The colocation of culture filtrate and effector sensitivity loci with SNB QTLs indicates that natural variation at genetic loci controlling additional components of effector sensitivity pathways may play a role in modulating adult plant resistance phenotypes. Whether there are additional NE–Snn interactions playing roles in adult plant susceptibilities is yet to be determined.

    COMMON QTLS BETWEEN SNB AND TAN SPOT DISEASES OF WHEAT

    Increasing numbers of publications on QTL mapping of both SNB and tan spot has revealed a number of common QTLs between the two diseases. That Tsn1 confers sensitivity to both ToxA and PtrToxA is a well-known example (Friesen et al. 2006), although investigation of resistance to P. tritici-repentis and P. nodorum in a biparental tetraploid wheat (T. durum) population indicated that although the Tsn1–ToxA interaction was important for P. nodorum infection, it did not play a significant role in the P. tritici-repentis interaction in the tetraploid wheat T. durum system and that this was probably due to low PtrToxA expression in P. tritici-repentis (Virdi et al. 2016).

    P. nodorum resistance or sensitivity QTL Qsnb.cur-2AS.1 (Phan et al. 2016), which was detected at the seedling and adult plant stage, has also been found to be a major contributor to tan spot resistance in seedlings and mature plants (Phan et al. 2016; Shankar et al. 2017). A QTL identified on the long arm of chromosome 5A is another instance of shared common genomic regions significantly associated with both diseases (Hu et al. 2019). This phenomenon may indicate that the two diseases possibly share common susceptibility and resistance mechanisms. It would be interesting to find out whether they have more effectors in common. The mutual interactions could be promising targets for wheat breeders, because they could introduce resistance to both diseases, especially for QTLs with large effects and at both the seedling and adult plant stages.

    ROLES OF NEW TECHNOLOGY-BASED AND BREEDING APPROACHES IN DELIVERING GENETIC GAINS IN SNB RESISTANCE

    Advances in the understanding of SNB resistance have been applied in breeding programs since 2005. For example, sequencing the P. nodorum genome revealed the presence of ToxA and demonstrated that it was the source of the related gene previously identified in P. tritici-repentis. It was a simple matter to express the gene in microbial hosts, infiltrate the protein into wheat seedling leaves, and determine whether plants were sensitive. An important factor was that these assays could be carried out with equipment as simple as a refrigerator and a needleless syringe; even a greenhouse was not essential. Thus, crop breeders found this assay practical and accurate. Armed with expressed ToxA since 2005, Tox3 since 2011, and Tox1 since 2012, researchers and breeders could determine the relationship between effector sensitivity and cultivar susceptibility. For P. tritici-repentis in Australia, a very simple picture emerged; all isolates of the pathogen carried PtrToxA, and sensitivity to this effector in wheat was strongly correlated with tan spot disease susceptibility. Large numbers of ToxA doses were distributed to breeders over the next few years, and the use of ToxA-sensitive wheat was reduced by half in 3 years. Considering these changes in more detail in more recent periods, the total area sown to tsn1 wheat varieties in Western Australia increased from 69.9% in 2009 to 2010 to >85% in 2018 (Oliver et al. 2014; Shackley et al. 2020; https://www.cbh.com.au/), and no detectable yield penalty is associated with insensitivity to ToxA (Oliver et al. 2014; Vleeshouwers and Oliver 2014). The application of effector-assisted breeding to SNB was more complicated. In Australia, effectively all P. nodorum isolates carried all three effectors, but the relationship between effector insensitivity and cultivar resistance was not as clear cut. As noted previously, epistasis between NE genes was apparent. Nonetheless, the elimination of effector sensitivity genes has never been shown to decrease SNB resistance or to have any other deleterious effect. It either has no effect or a positive effect on resistance. Analysis of the ToxA sequence in a diverse P. nodorum isolate collection indicates that the gene is positively selected (Stukenbrock and McDonald 2007). It is likely that ToxA will continually evolve into forms that are more potent in host cell death induction unless Tsn1 is bred out from widely planted wheat germplasms (Tan et al. 2012). In the case of Tox1 sensitivity, although the gene underlying sensitivity locus Snn1 has been cloned, the natural genetic variants determining insensitivity have not been formally identified. For Tox3 sensitivity, although highly significant markers closely linked to Snn3-B1 have been identified in experimental mapping populations, the observation that these markers provide surprisingly low prediction of Tox3 sensitivity in screens of wider germplasm collections (Downie et al. 2018) indicates that multiple sensitivity alleles may be present. Similarly, although the WAK gene underlying the Tox1 sensitivity locus Tsn1 has been cloned with a biparental population, the natural variants controlling insensitivity have not yet been determined, and so screening with the Tox1 protein probably is the most pragmatic approach for robustly determining sensitivity, at least until the causative variants controlling insensitivity are identified.

    In the coming years, the use of other emerging technologies will help speed up the identification and functional characterization of SNB/effector resistance genes and provide efficient routes to use these in breeding programs. Here we briefly summarize a subset of these resources and approaches, ending with an example of how a combination of these could be applied to future SNB resistance research and breeding.

    Access to the wheat gene space within a target genetic interval is a key resource to help identify causative genes and variants. Although a wheat reference genome is now available (IWGSC 2018), it has been constructed from an Asian landrace called ‘Chinese Spring’, genetically distant from the wheat grown in most of the world. This may be particularly relevant to effector sensitivity, because of the two cloned effector sensitivity loci in wheat, allelic variation at the Tsn1 locus conferring ToxA sensitivity is caused by the presence or absence of the underlying gene (Faris et al. 2010). Because ‘Chinese Spring’ is insensitive to ToxA, the wheat reference genome assembly lacks the Tsn1 gene. To help address such problems, the construction of genome assemblies for several additional bread wheat varieties is under way. This includes 14 cultivars via the 10+ Wheat Genomes Project (www.10wheatgenomes.com) and the founders of the UK MAGIC population (https://gtr.ukri.org/projects?ref=BB%2FP010741%2F1). To help annotate the genes in any new wheat assembly and to provide information on where and when a gene within your genomic region of interest is expressed, high-throughput RNA sequencing via next-generation sequencing platforms can be undertaken. This can be done with the use of short read technologies (e.g., RNA sequencing with Illumina platforms) or long read technologies to sequence full-length transcripts (e.g., isoform sequencing using PacBio platforms or nanopore technology). By combining genomic and RNA sequence datasets, candidate genes and polymorphisms within a target genomic region can be identified. Candidate genes can then be explored via reverse genetic approaches. Currently, a Targeting Local Lesions in Genomes population with an associated exome capture–based genomic sequence database is available for the wheat variety ‘Cadenza’ (Krasileva et al. 2017), allowing lines with putative deleterious mutations to be identified in silico and ordered. Alternatively, transgenic approaches such as RNA interference, CRISPR/Cas9 gene editing, and virus-induced gene silencing (VIGS) are all now used in wheat (Shan et al. 2013; Scofield et al. 2005; Travella et al. 2006). For additional reading on the routes for wheat gene functional annotation, see the recent review by Adamski et al. (2020).

    Next, we outline a case study for SNB improvement based on the environmentally stable adult plant resistance QTL QSnb.niab-2A.3, identified in the UK MAGIC population by Lin et al. (2020b). First, to rapidly generate suitable germplasm to further the investigation of this locus, the residual genetic variation present in MAGIC RILs could be exploited to generate a pair of nearly isogenic lines (NILs) for a given QTL in a single generation (as described in more detail by Scott et al. 2020). This NIL pair could be intercrossed to generate F1 seed, and the F1 is selfed to produce large numbers of F2 seed. Because the culture filtrate from P. nodorum isolate 203649 was found to identify a QTL at the QSnb.niab-2A.3 locus, F2 individuals could be screened for genetic recombination within the target interval and their F3 progeny phenotyped at the seedling stage for sensitivity to culture filtrate. This subset of recombinant lines, and their progenies, would be used to further refine the genetic interval. Once sufficient genetic mapping resolution is obtained, the gene content in the interval could be determined by projecting the genomic sequence and gene annotations of the relevant MAGIC founders onto the interval and RNA sequencing and IsoSeq gene expression data from leaf tissues harvested from the NIL germplasm before and after culture filtrate infiltration overlaid. Collectively, these datasets would allow candidate genes within the genetic interval to be identified and accurately annotated via bioinformatic analysis of the DNA variants, gene expression, and splice variant data generated. Subsequently, VIGS could be used to transiently silence candidate genes at the seedling stage, and any effect on sensitivity to culture filtrate infiltration could be determined. Additional functional validation of the candidates prioritized and validated by VIGS could then be assessed at the adult plant stage via stable gene silencing methods such as CRISPR/Cas9. Diagnostic markers for the natural causative polymorphisms underlying the functionally validated gene would be developed for marker-assisted selection, preferably with genotyping systems commonly used by wheat breeding companies, such as Kompetitive Allele-Specific PCR assays (LGC Biosearch Technologies).

    It is important to mention that application of the marker-informed breeding method genomic selection is now feasible in large genome crop species such as wheat (reviewed by Sun et al. 2019). Rather than relying on explicit identification of the QTL/genes underlying the target trait, genomic selection exploits the ability to cheaply generate high-density genetic marker datasets across the genome and to use this method alongside phenotypic data generated in “training set” lines to use the markers to predict the performance of their progeny across multiple subsequent generations. This allows selection to be applied based on genetic marker data and phenotypic data on the training set alone, without the need for field-based phenotypic selection in multiple subsequent rounds of population advancement. This method potentially reduces breeding cycle time, increases selection accuracy, and increases selection intensity. Genomic selection is likely to become a major source of improvement in plant breeding practice, and the methods probably can also be modulated to incorporate additional datasets such as diagnostic markers in order to help improve prediction accuracy (Mackay et al. 2020). Numerous studies have followed on from the first report of genomic selection in wheat (de los Campos et al. 2009) and include studies of diseases such as yellow rust (Ornella et al. 2012), Fusarium head blight (Herter et al. 2019), and STB (Herter et al. 2019). Of these, the study conducted by Herter et al. (2019) using 1,120 lines derived from 14 biparental families found that although genomic selection provided a selection advantage of about 10% for Fusarium head blight, no significant advantage was observed for STB resistance (Herter et al. 2019). This finding suggests that for phenotypes with strong genotype–environment interactions, genomic selection appears to be challenging (Herter et al. 2019). Based on the published literature, genomic selection has not been explicitly applied to SNB improvement, indicating a possible untested route for genetic improvement. We also noted that genome editing approaches such as CRISPR/Cas9 would be well suited for host–pathogen interactions that follow the inverse gene-for-gene model, whereby host effector sensitivity loci could be edited to make them insensitive. In the future, we might see application of genomic selection methods that combine targeted selection against NE sensitivity alleles or selection for gene-edited NE insensitivity alleles along with the use of genome-wide markers to capture all small-effect loci in a cost-effective manner for plant breeding programs.

    CONCLUSIONS

    Ultimately, the most efficient control of SNB will involve a combined approach based on agricultural and agronomic practices, disease monitoring, and genetic improvement. The widespread adoption of conservation agriculture including limited tillage methods means that SNB is likely to increase in prevalence in areas where plowing has previously been the norm. Methods to improve the genetic resistance of cultivars will surely remain the most important method of control. So far, no full genetic resistance to SNB has been identified. It is becoming increasingly apparent that SNB is found not only in the presence of easily distinguished diseases such as yellow rust and powdery mildew but also with the symptomatically similar diseases such as STB, tan spot, and possibly spot blotch as well. Selection for resistance to diseases occupies a substantial amount of time and resources available to breeders, particularly because yield and quality will always be prioritized. Furthermore, we know very little about how diseases interact. This is a particular area of fascination given that three of these pathogens share effectors.

    Breeding for resistance to SNB has always been challenging because full evaluation of a new cultivar requires the use of adult plants under field conditions. Inoculation with a representative set of isolates adds to the difficulties. One clear recommendation that emerged from recent studies is to make large annual isolate collections, especially from the current most resistant cultivar. These new isolates can be assessed phenotypically for new effectors and virulence characteristics and genotypically to track for selected chromosomal regions. Any new effectors can be expressed and assessed for their role in virulence. The main value of the isolate collections is that they allow the rational selection of the minimum set that represents the total phenotypic variance of the pathogen to which resistance should be sought. Finally, based on our current understanding of P. nodorum epidemiology and host resistance, we provide the following recommendations for SNB management:

    • Establish annual P. nodorum isolate collections and disease outbreak monitoring programs.

    • Use these contemporary P. nodorum isolates to test for cultivar resistance and assess for the presence of new effectors.

    • Where genetic structure is observed in a regional pathogen population, undertake rapid genotypic analysis to monitor the population.

    • Grow wheat cultivars with differing genetic backgrounds to avoid buildup of a specialized pathogen population, especially in areas where minimum tillage practices are common.

    • Where local pathogen populations contain known effector genes, grow wheat varieties with insensitive alleles at the corresponding host loci.

    • Continue wheat research and development activities to identify and deploy additional sources of SNB genetic resistance.

    The author(s) declare no conflict of interest.

    LITERATURE CITED

    Funding: J. Cockram, B. Corsi, M. Lin, A. Ficke, and M. Lillemo were funded within the framework of the 2nd call ERA-NET for Coordinating Plant Sciences via the EfectaWheat project, funded by the Biotechnology and Biological Sciences Research Council (BBSRC, grant BB/N00518X/1) and the Research Council of Norway (grant NFR251894). R. C. Downie was funded by a BBSRC Doctoral Training Partnership Ph.D. studentship. K.-C. Tan and H. T. T. Phan were supported by a joint initiative of Curtin University and the Grains Research and Development Corporation bilateral grant (CUR00023). Joint coordination and planning of project activities by J. Cockram and R. P. Oliver was aided by networking activities funded under the COST Action SUSTAIN.

    Author Contributions: R. C. Downie and M. Lin are joint first authors. M. Lillemo, R. P. Oliver, K.-C. Tan, and J. Cockram contributed equally to this work. R. C. Downie, M. Lin, and J. Cockram wrote the manuscript. M. Lin and J. Cockram undertook bioinformatic analysis. All other authors edited the manuscript and contributed to scientific supervision or discussions.

    The author(s) declare no conflict of interest.