Recent insights into barley and Rhynchosporium commune interactions

Abstract Rhynchosporium commune is the causal pathogen of scald in barley (Hordeum vulgare), a foliar disease that can reduce yield by up to 40% in susceptible cultivars. R. commune is found worldwide in all temperate growing regions and is regarded as one of the most economically important barley pathogens. It is a polycyclic pathogen with the ability to rapidly evolve new virulent strains in response to resistance genes deployed in commercial cultivars. Hence, introgression and pyramiding of different loci for resistance (qualitative or quantitative) through marker‐assisted selection is an effective way to improve scald resistance in barley. This review summarizes all 148 resistance quantitative trait loci reported at the date of submission of this review and projects them onto the barley physical map, where it is clear many loci co‐locate on chromosomes 3H and 7H. We have summarized the major named resistance loci and reiterated the renaming of Rrs15 (CI8288) to Rrs17. This review provides a comprehensive resource for future discovery and breeding efforts of qualitative and quantitative scald resistance loci.


| INTRODUC TI ON
Scald (also known as leaf blotch) in barley (Hordeum vulgare) is caused by the pathogen Rhynchosporium commune. Rhynchosporium was first isolated from rye (Secale cereale) in the Netherlands (Oudemans, 1897) and thereafter named Rhynchosporium secalis (Oud.) J.J. Davis (Davis, 1919). Zaffarano et al. (2011) resolved this single species into three, with the Rhynchosporium infecting Hordeum species and Bromus diandrus named as a new species, R. commune. Scald, originally from northern Europe (Brunner et al., 2007), is a destructive disease of barley worldwide. It is primarily a disease of cool, humid production regions, and manifests as elongated pale blotches with a distinctive brown margin that can appear on the leaves and leaf sheath of stems ( Figure 1) (Ayesu-Offei and Carter, 1971;Zhan et al., 2008). Scald can cause up to 30%-40% yield loss in susceptible cultivars and also has a detrimental impact on grain quality (Paulitz and Steffenson, 2011).
Studies into the infection pathways of the pathogen have revealed there are three main routes of transmission: between seasons, across locations, and within the plant. These transmission pathways were recently reviewed by Fountaine et al. (2010), and Stefansson et al. (2012). R. commune survives between seasons on barley residues, and conidiophores are produced on infected crop residues, which produce spores that infect the subsequent crop. The pathogen can also be carried in the tissue of infected seeds (Brunner et al., 2007;Topp et al., 2019). Long-distance transmission can occur by sowing infected seed or movement of stubble as hay, spreading R. commune to new geographical locations. As yet, the sexual form of R. commune is unknown, thus the possibility of long-distance dispersal of ascospores remains as another pathway for the spread of disease. At the single leaf level, the intercellular development of the pathogen is almost exclusively restricted to the subcuticular region of host leaves (Siersleben et al., 2014). Transmission of the pathogen between leaf layers and plants is then driven by sporulation on the leaf surface, with spores and mycelia spreading to nearby plants via rain splash or wind.
Based on the complexity of the pathogen, control of the disease requires an integrated and multifaceted approach, including application of fungicides, manipulation of sowing date, cultural disease management, and developing disease-resistant cultivars (Stefansson et al., 2012;McLean and Hollaway, 2018). Fungicides should be applied in mixtures or using alternation in modes of action to limit the rapid development of fungicide resistance (McDonald, 2015). The development of disease-resistant cultivars is plausible as a sustainable strategy for the management of R. commune. However, the fungal population can change rapidly, resulting in deployed resistant cultivars and fungicides becoming ineffective after several years of commercial use (Avrova and Knogge, 2012). The objective of more stable resistance in the long term may be achieved through introgressing and pyramiding multi-

| INFEC TI ON OF BARLE Y BY R . COMMUNE
R. commune fungal growth in barley occurs in four phases: germination (occurring approximately 12 hr after inoculation), then penetration (from approximately 24 hr after inoculation), leaf colonization with a slow increase in fungal biomass, followed by exponential growth with a massive gain of biomass (at around 10 days after inoculation), and a late stationary phase during which a dense stroma forms producing sporulation (Ayesu-Offei and Clare, 1970;Zhan et al., 2008;Siersleben et al., 2014). The genus Rhynchosporium produces conidia from vegetative hyphae that penetrate the cuticle above host epidermal cells (Avrova and Knogge, 2012). After the formation of germ tubes from the conidia and penetration of the cuticle, thin hyphae grow mainly longitudinally along the leaf, with no growth of mycelium through stomata (Thirugnanasambandam et al., 2011). The thin hyphae grow rapidly in the pectin-rich layer of the outer epidermis cell walls of the barley leaf, ultimately producing macrosymptoms and sporulation. However, R. commune can also show symptomless infection, where the fungus grows within the host plant without disease symptoms appearing (Walters et al., 2012). In one report, F I G U R E 1 Symptoms of scald caused by Rhynchosporium commune on (a) individual leaves in glasshouse and (b) whole plants in the field symptomless infection resulted in transfer of the pathogen to barley grains (Atkins et al., 2010).
After spore germination and cuticle penetration, susceptible genotypes experience collapse of more epidermal cells, leading to the collapse of mesophyll cells beneath extensive mycelial growth.
This process gives rise to the typical final scald lesions (Lehnackers and Knogge, 1990;Thirugnanasambandam et al., 2011). The influence of resistance genes on infection in the case of Rrs1 has been shown to prevent the establishment of subcuticular stroma in order to restrict R. commune growth (Lehnackers and Knogge, 1990). In contrast, the resistant cultivars Digger (from the UK) and Osiris (from Algeria), which both carry the Rrs2 locus, showed larger formation of papillae and haloes in cell walls and this was presented as the key resistance mechanisms of these cultivars by Jørgensen et al. (1993). The physiological mechanisms underpinning genetic resistance to scald requires further elucidation with functional studies as current knowledge is limited in this area.

| P OPUL ATI ON G ENE TI C S OF R . COMMUNE
Studies of the genetic structure of R. commune have revealed that the pathogen maintains high levels of genetic diversity at a microscale (Linde et al., 2009, McDonald, 2015. For example, genetic diversity at a single geographic location (field) was found to be more than 70% of the total genetic variation in a region within Europe (Zaffarano et al., 2006). In a study of 265 R. commune isolates from Australian barley crops collected in 1996, 76% of gene diversity was distributed within the sampling site area of approximately 1 m 2 , while 19% of gene diversity was distributed among sampling sites within fields and only 5% of gene diversity was distributed among fields (McDonald et al., 1999). Reports of high virulence diversity in R. commune populations worldwide provide further evidence to suggest this pathogen is highly genetically and phenotypically diverse (McDonald et al., 1999;Bouajila et al., 2007Bouajila et al., , 2010Stefansson et al., 2012Stefansson et al., , 2014. The genetic diversity of R. commune populations enables it to respond rapidly to selection pressures such as the introduction of new fungicides. There are well-documented examples of resistance developing, such as to benzimidazole fungicides, and its spread within the R. commune population (Locke and Phillips, 1995). In contrast, resistance to triazole fungicides evolves more slowly, resembling a quantitative decline in efficacy, which may involve different physiological mechanisms (Cooke et al., 2004;Zhan et al., 2005).
Populations exposed to flusilazole, tebuconazole, and epoxiconazole previously were shown to have 10 times lower sensitivity than populations that had not previously been exposed (Robbertse et al., 2001;Cooke et al., 2004), indicating a decrease in the effectiveness of these fungicides.
Mutation and variation in the CYP51 gene family plays a crucial role in azole fungicide resistance in a range of fungal species (Brunner et al., 2015). One known mechanism contributing to azole resistance in R. commune is the emergence of CYP51A, a paralog of CYP51 that confers reduced azole sensitivity (Hawkins et al., 2014;Brunner et al., 2015;Mohd-Assaad et al., 2016). CYP51A was present in the most azole-resistant R. commune populations from New Zealand and Switzerland, indicating the influence of fungicide selection pressure on the evolution of R. commune populations (Mohd-Assaad et al., 2016). The results of Mohd-Assad et al. (2016) suggest that CYP51A is the most important source of fungicide resistance variation in global R. commune populations.

| HOS T-PATHOG EN INTER AC TI ON S
Gene-for-gene interactions occur with every generation of the R. commune life cycle between the avirulence effectors in the pathogen and corresponding resistance genes in the host (Barua et al., 1993). This has led to the pathogen responding rapidly after the introduction of cultivars with new resistance genes and being able to infect these cultivars with major resistance genes or a combination of genes within a few seasons (Xi et al., 2002). Three necrosis inducing peptides (NIP1, NIP2, and NIP3) were identified as being important during the thin hyphal forming stage in R. commune by Wevelsiep et al. (1993). Expression analysis by Kirsten et al. (2012) has shown that NIP1 transcripts are present in spores, while NIP2 and NIP3 transcripts are synthesized after inoculation of host plants. At least two studies of diverse isolates have shown the near-universal presence of NIP2 and NIP3 genes in R. commune populations, suggesting the importance of both of these proteins for the pathogen (Schurch et al., 2004;Stefansson et al., 2014) (Penselin et al., 2016). The NIP1, NIP2, and NIP3 proteins are functionally important at the early growth stages of infection, when fungal hyphae spread before the growth of dense subcuticular stroma. Production of these proteins decreases dramatically when the fungal biomass increases rapidly, suggesting the early influence of these proteins on fungal virulence (Schurch et al., 2004). NIP1 functions both as an effector and an elicitor (Wevelsiep et al., 1993). Furthermore, Hahn et al. (1993) have shown that NIP1 not only is able to facilitate leaf necrosis in barley, but also induces the reactions of resistance gene Rrs1. The protein NIP1 is the product of the avirulence gene AvrRrs1 (Rohe et al., 1995) and induces the expression of pathogenesis-related 10 gene in leaves of Rrs1 barley plants (Steiner-Lange et al., 2003). Schurch et al. (2004) have demonstrated that virulence to Rrs1 was achieved through either a deletion or mutation of NIP1, but Rrs1 does not encode for the NIP1 receptor itself. The main factor for Rrs1-triggered resistance is recognition of the interaction of NIP1 with a receptor that consequently activates the plant's defence reaction (van't Slot et al., 2007). This makes the NIP genes and their presence, absence, or altered states between global populations of the pathogen a potential predictor of resistance gene effectiveness. Schurch et al. (2004) tested 614 different isolates from four continents, showing a NIP1 deletion frequency of up to 45%, while NIP2 and NIP3 were present in almost all isolates. Isolates carrying a functional NIP1 gene have shown significantly higher virulence than isolates where NIP1 was nonfunctional or missing (Stefansson et al., 2014;Mohd-Assaad et al., 2019). NIP2 and NIP3 induce necrosis in barley, but had no function as elicitors (Hahn et al., 1993). Strains of R. commune lacking NIP1 or strains with mutated NIP1 were shown to overcome cultivars carrying Rrs1 (Stukenbrock and McDonald, 2009).

| MA JOR G ENE S AND LO CI IDENTIFIED FOR SC ALD RE S IS TAN CE IN BARLE Y
The deployment of resistant cultivars and implementation of pathogen-informed management helps reduce pesticide applications and improves long-term crop protection. Resistance is classified into two categories: qualitative resistance genes provide high levels of resistance at all growth stages and quantitative genes provide partial levels of resistance most commonly observed at the adult plant stage Grcic, 2011, Zhan et al., 2008). Qualitative resistance genes have been frequently identified from seedling experiments using specific isolates (Zhan et al., 2008). Both types of resistance are important and by pyramiding these genes together into new cultivars it is possible to create more durable disease control (Walters et al., 2012). Barley varieties with known sources of major resistance genes and QTLs are summarized in Table 1. The complete reference barley genome sequence enabled the comparison of QTLs identified from different genetic maps (Mascher et al., 2017). Overall, we have summarized 148 QTLs from 34 different studies (Table 1 and were used for the BLASTn search (Zhang et al., 2019;Leng et al., 2020). Default settings were used to do the BLASTn search and the best hit was used to decide the physical position of the detected QTL (Table 1 and Figure 2). Most QTLs were identified using a phenotype of visually observable disease symptoms, but some studies have applied alternative methods. Looseley et al. (2012) detected QTLs based on the amount of R. commune in symptomless leaves, with results showing that the amount of R. commune in symptomless leaves is correlated with visual disease symptoms. Zhang et al. (2019) reported that a phenotype derived from the regression of disease on relative maturity is an effective trait to detect scald resistance under natural field conditions when relative maturity is correlated with scald resistance, as it reduces the possiblity of confusing maturityrelated loci with true resistance loci.
Genes and QTLs to R. commune resistance have been identified across the barley genome. However, some loci are linked with resistance repeatedly, which could suggest limited genetic diversity for resistance in modern barley germplasm. Over the past 20 years, QTL analysis has been used to reveal knowledge about the genetic architecture of R. commune resistance in barley germplasm and discover targets for marker-assisted resistance breeding. The importance (and hence frequent selection) of the Rrs1 and Rrs2 loci for resistance has been illustrated with a range of biparental populations across many environments, including under natural field conditions and specific isolate inoculations in controlled environments (Table 1 and Figure 2). The number of major loci reported for R. commune resistance is limited to Rrs1, Rrs2,Rrs4,Rrs12,Rrs13,Rrs14,Rrs15,Rrs16, Rrs17 (Rrs15 (CI8288)), and Rrs18 (Table 1 and Figure 2). Reasons for the relative lack of identified genes may be the close relationships among germplasm and the multiple reselections of the same allele combinations from different cultivars (Williams, 2003). The remaining 129 QTLs reported at the time of submission of this review are found across all seven barley chromosomes. We have included all reported QTLs in Table 1 with position and marker information where available and review the major QTLs below.
There is a lack of clarity in the literature around the nomenclature for the two loci both referred to as Rrs15. The locus on 7H reported by Genger et al. (2005) as Rrs15 is referred to in an earlier review by the same authors (Genger et al., 2003b), but not described by name. Subsequently Schweizer et al. (2004) reported and named a locus on 2H as Rrs15 (CI8288), and this name for this locus was retained by Wagner et al. (2008). This locus on 2H termed Rrs15 (CI8288) was renamed by Zhan et al. (2008) as Rrs17 to distinguish it from the locus on 7H, which remains named Rrs15. This was acknowledged by Coulter et al. (2019), who designated a new locus on 6H as Rrs18. We reiterate the name change suggested by Zhan et al. (2008) in this review.
On chromosome 1H, the major QTL Rrs14 has been mapped at 2.8 Mb and could be introduced into barley cultivars in combination with other QTLs on different chromosomes (Garvin et al., 2000). Because this QTL is not linked to other resistance QTLs, the likelihood of recovering the relevant recombinant genotypes from an intercross population is possibly increased (Garvin et al., 2000).
Another QTL from Hordeum spontaneum (Rrs-1H-1-4) was identified close to Rrs14 (Yun et al., 2005). Chromosome 2H has a major QTL Rrs17 (Rrs15 (CI8288)) sourced from the cultivar Triton, which was located at 10.4 Mb on the barley physical map (Wagner et al., 2008). This QTL was identified under both natural field conditions and glasshouse conditions with specific isolate inoculation, suggesting this region could be further explored as a source of resistance.
Another major QTL Rrs4 was mapped in proximity to Rrs1 at 523.0 Mb on chromosome 3H (Patil et al., 2003). With many resistance QTLs identified in the same region as the Rrs1 locus on chromosome 3H, it remains unknown if these QTLs are alleles of the same gene, or if they are part of a closely linked gene cluster  This major QTL was introgressed into cultivated barley from interspecific crosses with Hordeum bulbosum (Pickering et al., 2006). There have been a further six QTLs located close to Rrs16 reported from two different studies using both natural field inoculation and individual isolate screening Wang et al., 2014). In both studies, the resistant allele for the QTLs was derived from the cultivar Vlamingh. In each study the reported QTLs have major effects explaining more than 10% phenotypic variance, with a reported logarithm of the odds (LOD) value greater than 3 (Table 1).
No major scald resistance genes have been reported on chromosome 5H. However, minor effect QTLs for scald resistance have been mapped to this chromosome (Looseley et al., 2012;Coulter et al., 2019;Zantinge et al., 2019).

| INTROG RE SS I ON OF SC ALD RE S IS TAN CE FROM WILD BARLE Y R E L ATI V E S
was mapped on chromosome 7H, possibly offering a potentially different allele of Rrs2 (Abbott et al., 1992;Genger et al., 2005).
Rrs13 was identified on chromosome 6H from H. spontaneum (Abbott et al., 1995). Rrs14 was identified on chromosome 1H from H. spontaneum (Garvin et al., 2000). Another novel locus for scald resistance Rrs15 from H. spontaneum was mapped on chromosome 7H (Genger et al., 2005). From wild barley H. bulbosum, Rrs16 was identified on chromosome 4H (Pickering et al., 2006). These five novel loci provide valuable resources to introgress and pyramid the different scald resistance genes from wild barley in cultivated barley. However, further experiments are required to investigate the allelic differences between the QTLs from wild and cultivated barley.

| IDENTIFI C ATI ON OF C AND IDATE G ENE S AT MA JOR SC ALD RE S IS TAN CE LO CI
The successful cloning of genes at the Rrs1 and Rrs2 loci will advance the understanding of scald resistance. It will allow functional analysis of the genes themselves, which will also inform utilization of these loci in breeding programmes. So far, none of the major scald resistance genes has been cloned, although at least one attempt has been made to clone Rrs1 (Oldach, 2012). The recent publication of the barley genome (Mascher et al., 2017) will facilitate new work in this area, as a range of tools are developed to interrogate and use sequence data.
Simply searching for candidate genes of the major scald resistance QTLs through BLAST searches of fine-mapped marker intervals is a widely used approach (Hanemann et al., 2009;Hofmann et al., 2013;Coulter et al., 2019). The sequences of fine-mapped markers at the Rrs1 and Rrs2 loci can be used to perform BLAST searches against the high-confidence gene sequences on the IPK Barley Blast Server (http://webbl ast.ipk-gater sleben.de/barley). This is likely to reveal a large number of candidate genes that will need further investigation. The challenge with searching for scald resistance loci using BLAST searches is that they may not be present in the scald-sensitive Morex reference genome sequence (Coulter et al., 2019). Diagnostic markers have been developed for both finemapped Rrs1 (Rh4 type) (Looseley et al., 2020) and Rrs2 (Hanemann et al., 2009). Overall 10 high-confidence annotated genes were identified at the Rrs1 (Rh4 type) locus although the actual candidate gene remains elusive (Looseley et al., 2020). However, the Rrs1 (Rh4) locus was absent in the Morex genome (Mascher et al., 2017), which is cited as a major impediment to identifying the causal gene for this locus (Looseley et al., 2020). Detailed work by Marzin et al. (2016) investigated a family of putative pectin esterase inhibitor (PEI) genes at the Rrs2 locus. The study was able to conclude that no single PEI gene of the three investigated was the Rrs2 gene, and suggests the possibility that another known or unknown PEI gene or some combination of PEI genes may be Rrs2 instead. The authors also suggested that the resistance gene at the Rrs2 locus may be absent from the reference genome sequence that was generated from the susceptible cultivar Morex (Marzin et al., 2016).
BLAST approaches could evolve to other more targeted genome searching methods as sequence information becomes more comprehensive. Bioinformatics approaches without prior knowledge of the mapped locus may prove useful to identifying individual genes.
The location of genes from families with a known role in resistance and a common protein structure, for example wall-associated kinases or nucleotide-binding site leucine-rich repeat (NBS-LRR) proteins, could be targeted to identify new resistance genes in these families scattered across the genome. The majority of cloned disease resistance genes in plants encode NBS-LRR proteins (McHale et al., 2006;Dangl et al., 2013;Kourelis and van der Hoorn, 2018). Cloned wheat stem rust resistance gene Sr33 and wheat leaf rust resistance gene Lr10 both encode a coiled-coil (CC)-NBS-LRR protein (Loutre et al., 2009;Periyannan et al., 2013). A cereal NBS-LRR bait library was designed to predict NBS-LRR genes present in Triticeae species. MutRenSeq was developed for rapid gene cloning . By using MutRenSeq, the stem rust resistance gene Sr45 was cloned in wheat and the contig encodes a CC-NBS-LRR protein .
However, whole-exome capture and MutRenSeq can only identify annotated genes in reference genomes, and genes that are not in the reference genome may be missed. MutChromSeq was developed to reclone the barley Eceriferum-q gene and clone de novo the wheat Pm2 gene, and this method does not require construction of a physical reference sequence across a map interval or fine mapping (Sánchez-Martín et al., 2016). All of these advanced technologies provide less expensive and faster approaches to clone the scald resistance genes in barley.

| FUTURE IMPROVEMENT OF SC ALD RE S IS TAN CE WITH ADVAN CED G ENOMI C TO O L S
Historically, QTLs have been identified through linkage analysis of biparental mapping populations. The 148 QTLs for scald resistance reviewed and summarized here were all identified using biparental mapping populations. This method is useful for detecting largeeffect QTLs with rare alleles and has been an important tool for marker-assisted breeding for scald resistance in barley (Lorenz et al., 2011b). However, other approaches, including genome-wide association studies (GWAS) and genomic selection, may uncover new sources of resistance to scald.
GWAS are able to detect loci associated with target traits using diverse germplasm sets. This can introduce greater allelic diversity compared to biparental populations where only alleles segregating between the two parents can be evaluated (Zhu et al., 2008;Myles et al., 2009). Depending on the nature of linkage disequilibrium decay within the population under study, GWAS may facilitate higher mapping resolutions (Boyd et al., 2013;Sukumaran and Yu, 2014). However, GWAS are generally limited in power to detect very rare alleles or alleles with small effect sizes. Increasing the sample size can go some way to improving the power of association studies (Korte and Farlow, 2013).
So far, overall 13 QTLs from three GWAS analysis for scald resistance have been detected under natural field conditions at adult plant stage (Table 2 and Figure 2) (Gawenda et al., 2015;Looseley et al., 2018;Daba et al., 2019). GWAS analysis suggested Rrs1 is the most significant effect QTL among European spring barley germplasm under natural field conditions in Europe (Looseley et al., 2018). Interestingly, this study included cultivars that are known to carry the Rrs2 locus, but no QTL at the Rrs2 locus was detected in the field. In contrast, another GWAS analysis detected two large-effect QTLs for scald resistance on chromosome 7H (Daba et al., 2019). This GWAS analysis was carried out in Ethiopia by using barley genotypes from Ethiopia, ICARDA, and the USA.
Accurate phenotyping remains the challenge for complex traits such as disease resistance in the breeding programmes (Cooper et al., 2014;Zhang et al., 2017). By leveraging the information from difficult and expensive phenotyping through modelling GEBVs for lines without phenotypes, genomic selection may enable more rapid and inexpensive selection for multiple resistance loci (de los Campos et al., 2013;Desta and Ortiz, 2014), therefore genomic selection could enhance the rate of genetic gain of disease resistance traits (Lorenz et al., 2011b). Genomic selection approaches warrant investigation for their potential to improve scald resistance.

| CON CLUS IONS
The deployment of cultivars resistant to scald and understanding R. commune pathogen populations has helped reduce pesticide applications globally. However, R. commune can change quickly to overcome the deployed resistance genes. Therefore, introgression and pyramiding different resistance genes into one cultivar is likely to be effective to enhance durable scald resistance. Overall 148 QTLs for scald resistance were summarized from 34 different studies in this review. All of these resistance QTLs were projected on the barley pseudomolecules Morex v. 2.0 2019. The genome sequence enables us to understand the physical positions of previously reported QTLs and compare QTL results between linkage mapping studies. Finally, we have summarized many of TA B L E 2 Summary of the scald resistance quantitative trait loci (QTL) from genome-wide association studies included in this review ordered by chromosome position giving logarithm of the odds (LOD) scores, percentage of phenotypic variation explained by the QTL in the mapping population (where reported), and marker information the qualitative and quantitative resistance QTLs that will be crucial for improvement of scald resistance in future global barley breeding efforts.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analysed in this study.