Skip to main content
Advertisement
  • Loading metrics

Components of Brachypodium distachyon resistance to nonadapted wheat stripe rust pathogens are simply inherited

Abstract

Phytopathogens have a limited range of host plant species that they can successfully parasitise ie. that they are adapted for. Infection of plants by nonadapted pathogens often results in an active resistance response that is relatively poorly characterised because phenotypic variation in this response often does not exist within a plant species, or is too subtle for genetic dissection. In addition, complex polygenic inheritance often underlies these resistance phenotypes and mutagenesis often does not impact upon this resistance, presumably due to genetic or mechanistic redundancy. Here it is demonstrated that phenotypic differences in the resistance response of Brachypodium distachyon to the nonadapted wheat stripe rust pathogen Puccinia striiformis f. sp. tritici (Pst) are genetically tractable and simply inherited. Two dominant loci were identified on B. distachyon chromosome 4 that each reduce attempted Pst colonisation compared with sib and parent lines without these loci. One locus (Yrr1) is effective against diverse Australian Pst isolates and present in two B. distachyon mapping families as a conserved region that was reduced to 5 candidate genes by fine mapping. A second locus, Yrr2, shows Pst race-specificity and encodes a disease resistance gene family typically associated with host plant resistance. These data indicate that some components of resistance to nonadapted pathogens are genetically tractable in some instances and may mechanistically overlap with host plant resistance to avirulent adapted pathogens.

Author summary

Plant pathogens are specialists and can colonise only a limited number of plant species (hosts). Pathogen infection of a plant that is not a host of the disease often results in an active plant defense response. This poorly characterised defense response is durable as phytopathogens rarely successfully colonise new hosts. The ability to transfer this resistance to host plants would be highly beneficial in protecting crops from disease. However, this resistance has been difficult to genetically dissect as all members of a plant species are resistant. Here we show that some accessions of the model grass Brachypodium distachyon show some variation in their ability to suppress colonisation by the wheat stripe rust pathogen Puccinia striiformis. Brachypodium is not a host species of wheat stripe rust disease and no accessions are fully susceptible, however, some allow more pathogen growth than others. We have exploited these relatively subtle phenotypic differences to genetically dissect this difference in resistance and identified two Brachypodium loci that contribute increased resistance to the nonadapted wheat stripe rust pathogen.

Introduction

Only a limited number of phytopathogen species are adapted to paracitise a given plant species. The numerous phyopathogens unable of colonising a plant species (nonadapted) are often suppressed by an active plant resistance response upon challenge [1, 2, 3, 4]. Plant resistance against nonadapted pathogens is considered durable as pathogen colonisation of new plant host species is rare, at least over short evolutionary time frames [5, 6]. However, nonadapted pathogen infection can result in a variety of outcomes varying from complete plant immunity, to limited pathogen colonisation and reproduction [7].

Mechanistically resistance to nonadapted pathogens can range from a basic physical or chemical incompatibility between a pathogen and potential plant host, to active recognition of pathogen challenge by the nonhost plant leading to a complex defense response [1, 8, 9, 10]. Current molecular models of active defense invoke similar mechanisms used to protect host plants against avirulent adapted pathogens, these being phytoalexin defences, callose deposition, reactive oxygen production and cell death in some instances [9, 10, 11].

Plant recognition of pathogen attack occurs by plant pathogen or damage associated molecular pattern (PAMP/DAMP) receptors [12, 13, 14]. Adapted pathogens repress signalling by these receptors by the introduction of pathogen effector molecules into infected plant cells. Some members of a host plant species can recognise specific pathogen effector molecules or their enzymatic activity and then activate a defense response (ie. effector triggered immunity) [8]. Recognition of nonadapted pathogens is likely coupled with an inability of the nonadapted pathogen effector complement to effectively supress PAMP/DAMP signalling [4, 8, 10, 15]. Effectors from nonadapted pathogens can also be recognised by incompatible plant species in some instances thereby including effector triggered immunity, a typical host/pathogen defense response, in nonadapted pathogen resistance [16, 17, 18, 19, 20].

At a molecular level resistance to nonadapted pathogens has often been difficult to investigate. Within a species this resistance often shows no, or very subtle, phenotypic variation negating classical mapping and positional cloning strategies [21, 22, 23]. In some cases variation can be observed but controlled by multiple genes, each with small additive affects making isolation of the underlying QTL difficult [24, 25, 26, 27, 28, 29]. However, mutagenesis and virus induced gene silencing, usually in model species, has had some success [4]. For example, the isolation of the Arabidopsis PEN genes by mutagenesis has provided insights into the mechanistic basis of penetration resistance faced by nonadapted mildew pathogens [30, 31, 32]. However, once these penetration barriers are overcome additional mutations are required for increased nonadapted pathogen development consistent with a multiplicity of resistance barriers being faced by an invading pathogen [31, 33, 34].

Increasing phylogenetic distance between two plant species reduces their likelihood of both being hosts to the same morphological stage of a pathogen species [35]. This phylogenetic distance also appears to influence the extent of nonadapted pathogen growth on nonhost plants [10, 11]. Previously we examined rice resistance to the nonadapted cereal rust pathogens Puccinia graminis f. sp. tritici, P. triticina, P. striiformis f. sp. tritici and P. hordei. Each pathogen was capable of infecting rice and on occasion colonising hundreds of mesophyll cells, however sporulation was never observed [22]. Only subtle genetic variation in infection phenotypes was observed between different rice lines and mutagenesis was unsuccessful in identifying plants with perturbed resistance to these nonadapted fungal species [22]. The flax rust pathogen (Melampsora lini) of the dicotyledonous plant species Linum ussitatissimum (flax) was even less able to colonise rice, rarely being able to identify stomates on the leaf surface to enter the leaf [22]. These observations are consistent with increased phylogenetic distance between host and nonhost plant species being associated with a reduced ability of nonadapted pathogens to colonise. Similar poor colonisation was observed upon challenge of Arabidopsis thaliana with the wheat leaf rust pathogen, P. triticina [27].

The model temperate grass Brachypodium distachyon is more closely related to cereal rust hosts like wheat and barley than rice [36]. Unlike rice, B. distachyon is also parasitized by a rust disease pathogen, P. brachypodii [37, 38]. Also unlike rice, the nonadapted cereal rust pathogens P. graminis. f. sp. tritici and P. striiformis f. sp. tritici are capable of occasionally producing very small sporulating uredinia on B. distachyon [38, 39, 40, 41]. In contrast, the rust pathogen (Puccinia emaculata) of Panicum virgatum (switchgrass) shows limited infection of B. distachyon [23] again suggesting that phylogenetic relatedness of host and nonhost species influences nonadapted pathogen infection outcomes.

Those B. distachyon accessions that allow the most growth of cereal rust pathogens are nonetheless still highly resistant when compared with a susceptible wheat genotype or even some wheat genotypes that are classified as having resistance to rust disease. In fact, on most B. distachyon accessions no macroscopic symptoms are observed and microscopic analyses show limited infection by these nonadapted pathogens with sites usually consisting of an appressorium, substomatal vesicle and a few infection hyphae [41]. B. distachyon accessions challenged with cereal rust pathogens responded with an active defense response that involves callose deposition and H2O2 accumulation, whereas autofluorescent cell death is relatively uncommon [40, 41]. P. graminis. f. sp. tritici and P. striiformis f. sp. tritici are therefore not adapted to parasitise B. distachyon, however for some accessions the resistance to these pathogens is not as restrictive as that observed on rice.

Previously we developed two B. distachyon genetic mapping families that segregate for differential outcomes to P. striiformis f. sp. tritici (Pst) challenge [41]. In each family one parent (BdTR10h and BdTR13k, respectively) had highly restricted Pst growth with infection sites consisting of a substomatal vesicle and a few infection hyphae that colonised a limited number of mesophyll cells with haustoria (Fig 1) [41]. In contrast, the other parent in the cross (Tek-4 and Bd21, respectively) was more extensively infected by Pst with hundreds of mesophyll cells colonised (Fig 1). Occasionally these latter Pst infection sites produced tiny sporulating uredinia, although only after long (3–4 week) post infection time periods and specific growth conditions [41]. In a cross between B. distachyon accessions BdTR10h and Tek-4, these differential infection outcomes segregated as a 3:1 ratio with restricted infection being inherited as a single dominant gene [41]. In the second mapping family, BdTR13k x Bd21, highly restricted pathogen growth predominated amongst progeny and segregation data most closely supported a 1 dominant, 1 recessive gene model, or alternatively a two dominant, linked gene model, with each gene independently suppressing pathogen growth [41].

thumbnail
Fig 1. Pst infection phenotypes on B. distachyon accessions.

A) B. distachyon accession BdTR10h infected with Puccinia striiformis f. sp. tritici (Pst) isolate 104 E137 A–. B-C) Pst pathotype 104 E137 A–infection sites on BdTR10h leaf tissue stained with wheat germ agglutinin conjugated to fluorescein isothiocyanate (WGA-FITC), a lectin-fluorophore that specifically binds chitin. Green fungal infection hyphae can be seen ramifying between mesophyll cells, some of which contain brown pigmentation. D-E) The same Pst infection site shown in C). Some autofluorescent mesophyll cells are observed under UV light (D) while brown pigmentation is seen under bright light (E). F) Macroscopic infection symptoms of Pst pathotype 104 E137 A–on B. distachyon accession Tek-4 showing obvious striped lesions. G) Staining of Pst pathotype 104 E137 A–infection hyphae in a Tek-4 leaf. Large infection sites are observed. H) Macroscopic infection phenotype of Pst pathotype 104 E137 A–on B. distachyon accession Bd21. I) Growth of infection hyphae of Pst pathotype 104 E137 A–throughout a Bd21 leaf. Tissue was stained with WGA-FITC and observed under blue light. J) Extensive growth of Pst pathotype 104 E137 A–in leaf tissue of a yrr2 progeny plant derived from plant F4-56. K) Limited Pst pathotype104 E137 A–development on leaf tissue of a Yrr2 progeny plant derived from plant F4-56. The leaf sample was stained with WGA-FITC and photographed under UV light. No autofluorescent cell death was observed. L-M) A Pst pathotype 104 E137 A–haustorium (arrowed) growing in leaf tissue from a Yrr2 F5 plant derived from F4 plant 56 (L). The haustorium infected plant mesophyll cell has not undergone autofluorescent cell death when viewed under UV light (M). All images were taken approximately 21 days post infection (dpi).

https://doi.org/10.1371/journal.pgen.1007636.g001

In this report we undertake detailed fine mapping studies and demonstrate that a single dominant locus highly restricts Pst growth in a BdTR10h x Tek-4 family, whereas two dominant linked loci independently confer a similar phenotype in a BdTR13k x Bd21 family with one locus being nearly identical to that present in the former family. These data show that in B. distachyon some components of resistance to the nonadapted pathogen Pst are genetically simply inherited. These resistance components are superimposed on a background of B. distachyon incompatibility to colonisation by these nonadapted cereal rust pathogens.

Results

The Yrr1 gene segregates in each mapping family

Previous analysis of B. distachyon accessions BdTR10h and Tek-4 showed that macroscopically the former accession produced small brown lesions in response to challenge with Pst pathotype 104 E137 A–, whereas large, striped lesions were produced on Tek-4 plants [41] (Fig 1A and 1F). Each small brown lesion on BdTR10h corresponded to a small Pst infection site usually comprising several infection hyphae, occasional haustoria and limited numbers of weakly autofluorescent mesophyll cells, some of which contained nonfluorescent, brown pigments [41] (Fig 1B–1E). In contrast, larger infection sites that colonised hundreds of mesophyll cells and formed striped lesions growing parallel to leaf veins occurred on Tek-4 plants (Fig 1G). Infection of a BdTR10h x Tek-4 mapping family with Pst pathotype 104 E137 A–demonstrated that amongst 213 F2 plants, 159 allowed very little pathogen development while 54 plants showed more extensive pathogen growth [41] (S1 Fig). This single dominant locus (X2 3:1, p = 0.90) that restricted Pst development was designated Yrr1 (Yellow rust resistance gene 1).

Genomic DNA deep sequencing was coupled with bulk segregant analysis [42] to rapidly localise the Yrr1 locus. DNA samples from 153 BdTR10h x Tek-4 F2 plants that showed restricted pathogen growth were pooled (Yrr1/Yrr1 and Yrr1/yrr1) while a second DNA pool (yrr1/yrr1) was produced from 49 F2 plants that were more extensively colonised by Pst. Each DNA pool was sequenced (40x coverage, 100 bp paired end reads) and compared with the Bd21 reference genome [43]. Relative to the Bd21 genome 4,845,783 SNPs were identified in the Yrr1 DNA pool while 5,014,386 SNPs were identified in the yrr1 DNA bulk. SNPs unique to the Yrr1 DNA pool (ie. absent in the yrr1 pool) were identified and then aligned relative to the Bd21 reference genome sequence. The number of SNPs present in 100 kb intervals across the length of each Brachypodium chromosome was determined and graphed for these unique Yrr1 SNPS (Fig 2). A background of approximately 500 SNPs per 100 kb window was observed along each chromosome that was presumably due to sequencing errors and sequence misalignment (Fig 2). However, a substantial increase in the number of unique SNPs present in the Yrr1 DNA pool was observed in the 27–30 Mbp region of chromosome 4 with a single 100 kb region containing in excess of 2,500 SNPs that were absent in the yrr1 pool (Fig 2). These data locate Yrr1 within the 27–30 Mbp interval of chromosome 4.

thumbnail
Fig 2. SNP analysis of the BdTR10h x Tek-4 Yrr1 DNA bulk.

SNPs were identified relative to the Bd21 reference genome. Those SNPs that were unique to the Yrr1 DNA pool (ie. absent in the yrr1 DNA bulk sequence) were mapped to Bd21 reference chromosome sequences. Each data point is the number of unique SNPs identified relative to a 100 kb window of the reference chromosome sequence. A large peak of SNPs unique to the resistant DNA bulk was detected at 29Mb on chromosome 4.

https://doi.org/10.1371/journal.pgen.1007636.g002

To confirm the chromosome 4 map location of Yrr1 four polymorphic PCR sequencing markers (SNP29017600, SNP29419400, SNP29800350 and SNP30968800 (S1 Table) that spanned the region of interest were used to genotype 212 F2 plants that had been phenotyped for Pst infection (note SNP names indicate approximate nucleotide positions on the B. distachyon Bd21 chromosome 4 reference sequence version 1.2 (http://www.plantgdb.org/BdGDB)). Twenty six plants were identified with recombination events between markers SNP29017600 and SNP30968800. These plants were further genotyped with additional markers throughout the region (S1 Table) and two Yrr1 plants (F2 lines 113 and 172) were identified as recombinant in a 400 kb interval between markers SNP29419400 and SNP29800350.

These two critical recombinants and their F3 progeny were genotyped with 14 SNP markers throughout this 400 kb interval (S1 Table: SNP29419400—SNP29530590). Pst infection indicated that the Yrr1 phenotype segregated in each F3 family (S2 Fig). Genotyping and pathology analysis of F3 individuals from these two families confirmed that Yrr1 is located between markers SNP29419400 and SNP29517580 (Fig 3). Fifteen genes are annotated in this 100 kb region of the Bd21 reference genome (JGI B. distachyon genome v 3, Bradi4g24244 –Bradi4g24350) none-of-which encode proteins such as nucleotide binding site, leucine rich repeat proteins (NLR) proteins which are often associated with disease resistance loci (S2 Table).

thumbnail
Fig 3. The Yrr1 locus.

Location of BAC clones isolated from the Yrr1 locus. The upper image (line with black and white boxes) depicts genes annotated in the Bd21 reference genome sequence. White boxes indicate genes present in the Yrr1 interval while black boxes indicate genes flanking Yrr1. Grey boxes indicate genes not included in the Yrr1 haplotype analysis while genes marked with an asterisk were identified as Yrr1 candidates. Gene annotation is as per the JGI Brachypodium reference genome sequence version 3. Lines beneath indicate the positions of sequenced BAC clones isolated from BdTR10h, Tek-4 and BdTR13k genomic DNAs. Numbers correspond to nucleotide positions of the Bd21 chromosome 4 reference sequence. The position of critical SNP markers defining the Yrr1 interval are labelled and indicated by arrows.

https://doi.org/10.1371/journal.pgen.1007636.g003

Two SNP markers from the Yrr1 interval, SNP29419400 and SNP29530590, were used to PCR screen BAC libraries of B. distachyon accessions BdTR10h, BdTR13k and Tek-4. BAC clones were isolated for each genotype and sequenced (Fig 3). Only a single BAC clone was isolated from accession BdTR10h which did not cover the 98 kb Yrr1 interval in its entirety, missing approximately 19 kb of sequence adjacent to SNP29419400. Two Tek-4 BAC clones from the yrr1 locus were sequenced that covered the interval apart from missing 6kb of sequence in juxtaposition to SNP29419400 (Fig 3). The 80 kb of Yrr1 locus sequence available from BdTR10h showed near sequence identity (99%) with the equivalent region from BdTR13k. In contrast more sequence variation occurred between the yrr1 locus of Bd21 and Tek-4 and both of these sequences and the conserved Yrr1 locus present in BdTR10h and BdTR13k. This variation was further confirmed by the haplotype analysis described below with genetic distances shown in Fig 4. From these data it is appears that the Yrr1 locus is present in both BdTR10h and BdTR13k as part of a highly conserved genomic segment. The presence of Yrr1 in BdTR13k was further confirmed by the genetic analysis described below.

thumbnail
Fig 4. Two distinct haplotypes exist at the Yrr1 locus.

The phylogenetic tree was constructed using maximum likelihood with model GTRCAT and rapid hill-climbing mode. Scale bar units are substitutions per site, with the bar representing 0.1 (based on polymorphic sites only). Bootstrap support was generated using 2,000 bootstrap runs, with only support above 80% shown in the figure. For simplicity a single representative is shown for clades containing multiple accessions (marked with an asterisk). A more complete phylogeny is shown in S3 Fig. Red labelling indicates accessions that were immune to Australian Pst isolate 104 E137 A-, while those in green were immune to UK isolate 08/501. These accessions were tested with a single Pst isolate only. Accession labelled blue were immune to both isolates [41]. Those accessions labeled purple allowed more extensive Pst growth although Luc1 and Tek-2 were tested with only the UK and Australian isolate, respectively.

https://doi.org/10.1371/journal.pgen.1007636.g004

Haplotype analysis identifies 5 gene candidates at the Yrr1 locus

To further understand the haplotype diversity at the Yrr1 locus the nucleotide variation in a diverse panel of 47 resequenced B. distachyon accessions was examined. Alignment of reads to the Brachypodium Bd21 reference genome spanning the region encompassing Bradi4g24244 to Bradi4g24350 identified 104 polymorphic nucleotide sites across 14.7 kb of coding sequence derived from 11 genes at the Yrr1 locus (S3 Table). Phylogenetic analysis using maximum likelihood (Fig 4) and neighbor joining (S3 Fig) identified trees with similar topologies, which established two distinct haplotypes (Yrr1.h1 and Yrr1.h2). The immune accessions BdTR10h and BdTR13k contain Yrr1.h1 whereas Bd21 and Tek-4, which are more extensively colonized by Pst, contain Yrr1.h2.

While 15 genes are annotated at the Yrr1 locus, four genes (Bradi4g24293, Bradi4g24330, Bradi4g24336, Bradi4g24350) were excluded from the haplotype analysis due to the presence of indels in the coding sequence that resulted in a frame shift or SNPs that led to the generation of an early stop codon in some accessions (S4 Table). Bradi4g24330 and Bradi4g24350 are pseudogenes in accession BdTR10h (Yrr1) while Bradi4g24293 is a pseudogene in BdTR13k (Yrr1) (S4 Table). These three genes are therefore not Yrr1 candidates and were not considered further. Bradi4g24336, while a pseudogene in some accessions, encodes an intact ORF in the immune parents and was therefore analysed independently as described below.

Previously we had tested many of the resequenced accessions with either an Australian (104 E137 A–) or UK (08/501) Pst isolate, both of which are recognized by Yrr1 [41] (Fig 4). SNPs from 10 of the 11 genes used in the haplotype analysis were compared between immune plants known to contain Yrr1 ie. BdTR10h, BdTR13k and ABR6 (see accompanying paper by Bettgenhaeuser et al. for ABR6 genotyping) and those that show more extensive Pst colonization and therefore do not contain Yrr1 ie. Bd21, Tek-4, Tek-2 and Luc1 (Fig 4). Bradi4g24297 was excluded from further analysis as it encodes a reverse transcriptase protein common to plant retrotransposable elements and so is an unlikely Yrr1 candidate.

Only three genes (Bradi4g24270, Bradi4g24320, Bradi4g24333) contain SNPs that are uniquely polymorphic between the two resistance classes (S5 Table, highlighted in blue). These unique SNPs result in changes in the predicted proteins of Bradi4g24270 and Bradi4g24320 making these genes Yrr1 candidates (Table 1). For Bradi4g24333 this differential SNP is a synonymous change. However, for both this gene and Bradi4g24315 (S4 Fig) no identical protein is encoded by an immune and more susceptible accession therefore making both these genes Yrr1 candidates (Table 1). For all the remaining genes at the locus at least one immune and one more susceptible accession encode identical proteins thereby excluding these genes as Yrr1 (S5 Table, highlighted in grey).

thumbnail
Table 1. Amino acid differences in Yrr1 candidate proteins.

https://doi.org/10.1371/journal.pgen.1007636.t001

Additional plants previously shown to be immune to Pst [41] (Fig 4) are also shown in S5 Table and, where sequence was determined, each maintains all highlighted (blue) SNPs unique to the Yrr1 genotype, with the exception of the polymorphism in Bradi4g24270 in accession Adi2. As Pst immunity has not been shown to segregate with the Yrr1 locus in these latter plants they could contain a nonfunctional Yrr1 locus with immunity provided by an alternative gene and hence Bradi4g24270 remains a candidate. None-the-less, these additional accessions show a co-relation between Pst immunity and the SNPs unique to Yrr1.

Independent analysis of Bradi4g24336 showed that Yrr1 lines BdTR13k, BdTR10h and ABR6 all encode identical coding sequences and consequently identical predicted proteins (S6 Table). Proteins encoded by Bd21 (yrr1) and Luc1(yrr1) show approximately 97% amino acid identity to each other and the protein encoded by the Yrr1 genotypes. In both Tek-2 and Tek-4, Bradi4g24336 has a mutated gene start codon with no alternative start site in immediate proximity suggesting that it is a pseudogene in these accessions. Seven SNPs and three amino acids are unique to genes and proteins encoded by the more susceptible genotypes including the Tek-2 and Tek-4 pseudogenes (S6 Table). The Bradi4g24336 gene, which encodes a protein of unknown function containing a DUF4220 domain, is therefore a Yrr1 candidate (Table 1).

To test the Pst race-specificity of Yrr1, BdTR10h (Yrr1/Yrr1) and Tek-4 (yrr1/yrr1) were challenged with a further three Pst pathotypes that represent the genetic diversity of this pathogen in Australia. Twenty one days post infection leaves from each accession were stained with wheat germ agglutinin conjugated to fluorescein isothiocyanate (WGA-FITC), a lectin-fluorophore that specifically binds fungal chitin [22, 41]. Rust infection sites were microscopically measured on each accession and median infection site areas calculated. Each Pst isolate showed greatly reduced fungal growth on BdTR10h compared with Tek-4 (Mann Whitney U-test, p < 0.05), providing no evidence for race-specificity being conferred by Yrr1 (Fig 5).

thumbnail
Fig 5. The Yrr1 locus provides resistance to multiple Pst pathotypes.

BdTR10h and Tek-4 were infected with three Pst pathotypes, these being from left to right isolate 110 E143 A+, 134 E16 A+ and 108 E141 A-, respectively. In each experiment the median infection site area was determined by staining infected leaf tissue with WGA-FITC and microscopically measuring individual infection sites approximately 21 dpi. Numbers above each column indicate the number of infection sites measured in each case. The minimum criteria to be counted as an infection site were spores that had germinated and developed at least appressoria and substomatal vesicles. Significantly greater Pst growth occurred on Tek-4 for all three isolates (Mann Whitney U-test, p < 0.05).

https://doi.org/10.1371/journal.pgen.1007636.g005

The Yrr2 gene

Previously we reported that infection of B. distachyon accession BdTR13k with Pst pathotype 104 E137 A–resulted in very limited pathogen growth, similar to that observed on BdTR10h plants, although no brown lesions developed on the leaf surface and no macroscopic symptoms were evident ie. the plants showed macroscopic immunity [41]. In contrast, more extensive Pst growth occurs on Bd21, similar to that observed on Tek-4, with larger infection sites corresponding to obvious macroscopic lesions on leaves (Fig 1H and 1I). Amongst 316 F4 plants in a BdTR13k x Bd21 mapping family 228 individuals were macroscopically immune to Pst infection while 88 individuals showed obvious lesions that were confirmed as Pst infection sites by microscopy [41]. These data do not fit the segregation ratios expected of a single dominant gene or two unlinked dominant genes in an F4 family.

To identify loci conferring resistance in this population, deep sequencing (40x coverage) of F3 DNA bulks derived from 34 immune plants and 20 plants that showed macroscopic Pst symptoms was performed. Relative to the Bd21 genome 915,562 SNPs were identified in the immune pool and 781,411 SNPs in the second pool. In this DNA pooling experiment all SNPs identified were theoretically derived from the BdTR13k parent as Bd21 was the reciprocal parent in this cross and its genome was used for SNP identification. The immune DNA pool was potentially derived from both heterozygous and homozygous plant genotypes. Two distinct peaks of SNPs were identified in the immune DNA bulk that were under-represented in DNAs from plants showing Pst symptoms (Fig 6). Both SNP peaks were located on the chromosome 4 reference sequence with one peak at 29Mb coinciding with Yrr1, consistent with the sequence similarity with BdTR10h in this region described above, while a second locus was identified at 8Mb on the other arm of chromosome 4. This locus was subsequently named Yrr2.

thumbnail
Fig 6. SNP analysis of BdTR13k x Bd21 bulked DNA sequences.

SNPs were mapped relative to the Bd21 reference genome chromosome sequences with each datapoint showing the number of SNPs identified in a 100 kb interval of sequence. Datapoints in blue indicate SNPs found between both DNA bulk sequences and the Bd21 reference genome (ie. essentially showing the distribution of SNPs between the BdTR13k and Bd21 genome). Superimposed on the blue graph are SNPS that were unique to the DNA bulk derived from immune plants (ie. absent in the DNA bulk derived from plants showing the greatest Pst growth). Two distinct red peaks in the immune DNA bulk data, at 8 Mb and 29 Mb on chromosome 4, can be seen that obscure underlying blue peaks.

https://doi.org/10.1371/journal.pgen.1007636.g006

To further refine the Yrr2 locus, F3 DNA samples from 52 plants (32 immune, 20 with more Pst growth) used in the DNA bulking process were genotyped with a series of SNP markers spanning the 8 Mb region (S7 Table; SNP6716730 through to SNP11475500, excluding CAPs and KASP marker) and also for two SNP markers that defined the Yrr1 interval (S1 Table, SNP29419400 and SNP29530590; Fig 3). All 20 plants that showed macroscopic symptoms (ie. that were more extensively colonised) were homozygous for Bd21 SNPs using the two Yrr1 flanking markers as well as for SNP markers from SNP7761780 through to SNP9888760 (S7 Table). Markers immediately outside this region (S7 Table, SNP7733430 and SNP9983700) contained SNPs from the immune BdTR13k parent in some of these more extensively Pst colonised plants. These data suggested that yrr2 is located between SNP7733430 and SNP9983700. Genotyping of the 32 immune F3 plants showed that 30 contained at least one Yrr1 allele that conferred immunity. The remaining two immune plants (yrr1/yrr1) contained BdTR13k SNPs across the SNP7733430 to SNP9983700 interval consistent with a second locus that also suppressed Pst colonisation, Yrr2, being present in these latter two plants.

F5 families from the BdTR13k x Bd21 cross were screened to identify segregating families monogenic for either Yrr1 or Yrr2. Two F5 families were identified, one from plant F4-93 (Yrr1/yrr1; yrr2/yrr2) and the other from plant F4-56 (yrr1/yrr1; Yrr2/yrr2). Repeated testing of F5 progeny showed that Pst immunity segregated as a single dominant gene in each family (examples shown in S5 Fig). Having developed a monogenic family segregating for Yrr2 microscopic analyses of Pst infection phenotypes were undertaken on F5 progeny. In Yrr2 plants Pst infection sites were again restricted to just a few limited infection hyphae with occasional haustoria (Fig 1K–1M). Autofluorescent cell death was not common (Fig 1K and 1M). In contrast, yrr2 sib plants had more extensive Pst colonisation similar to that observed on the Bd21 parent (Fig 1J). The Yrr2 histological phenotype therefore appeared very similar to that conferred by Yrr1. Plants that contained both Yrr1 and Yrr2 did not show enhanced suppression of rust infection compared with plants containing either gene singularly indicating that these two resistance genes do not function additively.

However, in contrast to Yrr1, when F5 progeny of plant F4-56 (yrr1/yrr1; Yrr2/yrr2) were challenged with a second Pst pathotype, 134 E16 A+, all seedlings developed macroscopic symptoms regardless of the Yrr2 genotype. Microscopic analysis of lesions on these plants confirmed extensive fungal colonisation on both Yrr2 and yrr2 genotypes similar to that observed on the Bd21 parent. These data indicate that the Yrr2 gene shows race-specificity to Pst and does not restrict growth of Pst pathotype 134 E16 A+.

Genotyping of a critical F3 recombinant, plant 41 (yrr1/yrr1; Yrr2/Yrr2), whose progeny all showed Pst immunity (20 plants screened with Pst and analysed microscopically in three independent experiments) with SNP, KASP and CAPs markers (S7 Table), coupled with plant F3 237 which showed macroscopic lesions, defined the Yrr2 locus to a 300 kb interval between nucleotides 9,583,752 and 9,888,760 bp of the chromosome 4 reference sequence (Fig 7A). Within this genomic interval in the Bd21 genome 37 genes are annotated (S8 Table), 16 of which encode NLR genes (Fig 7A). Five of these 16 NLR genes encode short truncated proteins and are likely pseudogenes (Fig 7A). The NLR proteins encoded by this gene family vary from 53% to 98% amino acid identity.

thumbnail
Fig 7. Structure of the Yrr2 locus.

A) Diagram depicting the Yrr2 locus. NLR genes present at the Bd21 yrr2 locus are shown as white boxes labelled with gene accession numbers (numerous other genes at this locus are not shown). Beneath are the equivalent regions of sequenced BAC clones from BdTR13k with equivalent nucleotide co-ordinates of the Bd21 chromosome 4 sequence shown. NLR genes absent in these BACs are shown with dashed lines. Deleted regions are marked on the upper Bd21 sequence in parentheses with deletion sizes indicated. The structure of the BdTR13k Yrr2 locus is shown beneath. Coloured ovals represent nonNLR genes described in Fig 8. Grey boxes represent truncated NLR genes. B) DNA blot analysis of B. distachyon BdTR13k x BdTR10h F4 DNAs. DNAs were restricted with HindIII and hybridised with a probe that encoded a 387 bp sequence derived from NLR genes at the Yrr2 locus. DNAs from parental lines BdTR13k (Yrr2 haplotype) and Bd21 (yrr2 haplotype) are shown in lanes 1 and 12, respectively. The remaining lanes contain DNAs from F4 progeny of these lines which segregate for the Yrr2 locus. Band intensities reflect DNA loadings. Arrowheads indicate molecular weight mobilities of 8, 5 and 2 kb.

https://doi.org/10.1371/journal.pgen.1007636.g007

A total of five BdTR13k BAC clones were isolated from the Yrr2 interval that collectively covered the region in its entirety ie. from marker KASP9583752 through to marker SNP9888760 (Fig 7A). These BAC clones were isolated by PCR screening using SNP9888760 primers (BAC9E and 3B) and CAPS9656179 primers (BAC 11A and 12B) (S7 Table), while a nonpolymorphic set of primers (materials and methods—primers 9735685) were used to isolate BAC 1H (Fig 7A). These five BAC clones were sequenced and compared with the Bd21 reference genome sequence. Compared with Bd21 the number of NLR genes at the BdTR13k Yrr2 locus was considerably reduced, consisting of just seven genes (Fig 7A). Three of these genes (Bradi4g10141, Bradi4g10147, Bradi4g10153) encode truncated NLR proteins like their Bd21 equivalents while Bradi4g10220 encodes a full length protein identical in sequence to that present in Bd21 (Table 2). The remaining three genes (Bradi4g10017, Bradi4g10030/10060, Bradi4g10207) encode probable full length proteins distinct to those found at the Bd21 locus (90–98% amino acid identity–Table 2) making them potential gene candidates for Yrr2.

thumbnail
Table 2. Proteins encoded by NLR genes present at the Yrr2 locus of B. distachyon accession BdTR13k compared with nearest equivalent Bd21 proteins.

https://doi.org/10.1371/journal.pgen.1007636.t002

The sequences of these 3 candidate genes were compared between Tek-2 (yrr2), Tek-4 (yrr2), BdTR10h (yrr2), Luc1 (yrr2) and ABR6 (Yrr2 –see accompanying paper by Bettgenhaeuser et al.). Proteins encoded by Bradi4g10017 genes were truncated in Tek-2 (314 a.a.), Tek-4 (740a.a.), Bd21 (740 a.a.), BdTR10h (774 a.a.) compared with the 882 amino acid proteins encoded by BdTR13K (Table 2), ABR6 and Luc1, with these latter proteins having 99% identity. No accessions encoded proteins identical to the predicted BdTR13k Bradi4g10030/10060 protein (1015 a.a) with ABR6 (658a.a), BdTR10h (830 a.a.) and Tek-4 (658 a.a.) encoding truncated proteins. While all accessions appeared to encode full length Bradi4g10207 proteins none, including the ABR6 protein, were identical to the BdTR13k protein with the Bd21 protein showing greatest identity (98%) (Table 2). This comparative analysis therefore did not unambiguously resolve an obvious Yrr2 candidate amongst these three genes. (It should be noted that small regions of some these genes derived from whole genome sequences did not have full coverage and hence these small regions were excluded from identity calculations of all proteins).

Sequence analysis indicated that differences in NLR gene numbers between the Bd21 and BdTR13k loci were due to two insertion/deletion events of 55 and 60 kb, respectively. The boundaries of these two events could be identified by flanking sequence homology. Assuming each event was a deletion in BdTR13k, the first appeared to have arisen by recombination between NLR genes Bradi4g10030 and Bradi4g10060, two genes with 92% ORF identity, to produce a potentially functional chimeric NLR gene encoding the 5’ terminus of the former gene and 3’ of the latter. This deletion was found in three different BAC clones demonstrating that it was not a cloning artefact. The second deletion occurred between conserved regions immediately 5’ of NLR genes Bradi4g10171 and Bradi4g10207. The differential haplotype complexity of the Yrr2 locus in Bd21 and BdTR13k was confirmed by DNA blot analysis of 54 F4 individuals from the BdTR13k x Bd21 family (examples shown in Fig 7B). It cannot be ruled out, however, that both events may be diverged duplications in the Bd21 genome arising from unequal crossing over. Regardless of the mechanism, significant variation has evolved between the BdTR13k and Bd21 Yrr2 haplotypes.

Identification of rice and wheat homologues of genes present at Yrr1 and Yrr2

Homologues of the Yrr2 locus are present in both the rice and wheat genomes. In rice this NLR gene family consists of 10 tandemly duplicated genes located on chromosome 11 that encode proteins with approximately 60% identity to protein homologues encoded by the B. distachyon Yrr2 (BdYrr2) locus (Fig 8A). To further investigate the syntenic relationship of this locus in rice and B. distachyon additional nonNLR genes in the region were also examined. Six nonNLR genes were identified at BdYrr2 that encode proteins with substantial homology (65–85% identity) to proteins encoded by the equivalent rice locus, although the relative order of these genes has diverged between these two species (Fig 8A). For both species numerous other genes were also present at the locus for which no reciprocal homologue was present at the locus of the other species. In addition, a NLR gene (Os11g45330) unrelated to the rice Yrr2 NLR family was located at the locus while the equivalent B. distachyon gene (Bradi4g09247) was located approximately 1Mb distal to the locus. Previous comparisons between the B. distachyon and rice genomes have shown that chromosome 4 of B. distachyon has substantial similarity in gene order and content to rice chromosomes 11, 12 and 4 [43].

thumbnail
Fig 8. Synteny of B. distachyon Yrr loci in wheat and rice.

A) Structure/location of loci present in B. distachyon, rice and wheat that encode NLR homologues present at BdYrr2. The upper diagram depicts the B. distachyon yrr2 locus present in accession Bd21, with NLR genes shown as white boxes. Six additional nonNLR genes are shown as coloured ovals that are also present at the homologous rice locus located on chromosome 11 (middle diagram), albeit in a different linear order. Rice NLR genes at this locus are shown as white boxes. In the lower diagram the location of sequences in the wheat genome with homology to the Yrr2 NLR gene Bradi4g10207 are depicted with arrows on the relevant chromosome. Red, black and white arrows indicate high confidence, low confidence and unannotated wheat genes, respectively. Numbered arrows indicate the number of high confidence NLR genes at the wheat locus while unnumbered arrows depict a single gene only. Red dots show loci on wheat chromosome 4 that are regions with some gene co-linearity to non-NLR genes present at the Brachypodium Yrr2 locus, while white dots on the same wheat chromosomes show the location of potential Yrr1 homologous loci. Note each diagram depicts relevant genes only and not all genes present at these loci are shown. B) Structure/location of loci present in rice and wheat that show some sequence similarity to the Bd21 yrr1 locus. Conservation of gene order is shown between the yrr1 locus of B. distachyon accession Bd21, a region of rice chromosome 11 and a region of wheat chromosome 4DS. Genes are depicted as boxes with homologous genes aligned vertically. Only rice and wheat genes with homologues in the B. distachyon genome are illustrated. Arrows show a loss of one of two related genes in the rice genome compared with Bdyrr1 and a change in gene order. The black boxes indicate Brachpodium genes that lie immediately outside of the yrr1 locus. All wheat genes begin with the prefix (TraesCS). Genes TraesCS4D01G106100LC and TraesCS4D01G108000 are a duplication of the Bradi4g24310 homologue. Note the wheat locus is inverted relative to the Brachypodium and rice locus with nucleotide co-ordinates indicated in Mb.

https://doi.org/10.1371/journal.pgen.1007636.g008

Thirty genes with high confidence annotation are present in the Chinese Spring wheat genome that encode proteins with approximately 60% identity to the BdTR13k NLR homologue of Bradi4g10207 (Fig 8A). These are located on chromosomes 2B (1 gene), 3A (8), 3B (14), 3D (3), 4B (1), 5A (1), 5B (1) and 7A (1) (Fig 8A). Nineteen of these 30 genes, present on chromosomes 2B (1), 3A (6), 3B (7), 3D (2), 5A (1), 5B (1) and 7A (1), encode predicted full length proteins. The remaining genes encode either truncated proteins or pseudogenes, although these former proteins could potentially play a role in disease resistance [44]. Additional low confidence and unannotated homologues are also present on these chromosomes (Fig 8A).

The larger NLR gene families present on wheat chromosomes 3AS, 3BS and 3DS are located at the 41, 52 and 30 Mb regions of these chromosomes, respectively, suggesting that they are homoeologous loci. Each locus encompasses 1 (3AS), 1.6 (3BS) and 0.25 Mbs (3DS) of sequence, respectively. Interestingly the non-NLR genes common to the B. distachyon and rice loci are not located at these wheat 3S loci. Wheat genomic analyses have identified little gene collinearity existing between wheat chromosome 3 and either B. distachyon chromosome 4 or rice chromosome 11 [45, 46, 47]. It is noteworthy that this wheat chromosome shows less conservation of gene collinearity when compared with other grass chromosomes [47]. Similar potential homoelogous wheat NLR loci are present on wheat chromosome 2L, albeit encoding low confidence and unannotated genes, while related loci are present on two out of three homoleogous chromosomes of 5L and 7L (Fig 8A). Related sequences are also located on chromosome 4B and 4D at 673 and 467 Mb, respectively (Fig 8A).

A number of race specific and adult plant wheat stripe rust resistance loci have been mapped to wheat chromosomes 3BS (eg. Yr4/YrRub [48] Yr30 [49] Yr57 [50], Yrns-B1 [55]), 3DS (Yr49 [49], Yr66 [49]) and 3AS (Yr76)[52]. Comparative genomics was used to locate these wheat Yrr2 NLR homologues (TaYrr2) on chromosome 3S relative to these known wheat Pst resistance genes (Table 3). Although there are numerous stripe rust resistance genes on wheat chromosomes 3S none could be definitively located at the TaYrr2 loci (Table 3).

thumbnail
Table 3. Proximity of wheat stripe rust resistance genes to NLR loci on wheat chromosome 3 that are homologous to the Brachypodium Yrr2 locus.

https://doi.org/10.1371/journal.pgen.1007636.t003

Of further interest is a region on wheat chromosome 4 that has some similar gene co-linearity to the non-NLR genes present at the Brachypodium Yrr2 locus (S6 Fig). At each wheat chromosome 4 locus non-NLR genes with homology to those found at the Brachypodium Yrr1 locus are present in the same gene order apart from one large inversion event. In contrast, those regions that encode NLR genes at the Brachypodium Yrr2 locus have been replaced at these wheat loci with additional unrelated sequence ranging from 70 kb to 3.6 Mb in length. The NLR sequences in the wheat genome that are homologous to those present at the Brachypodium Yrr2 locus have therefore undergone significant expansion and relocation since the divergence of wheat and Brachypodium.

A similar sequence to the Bd21 yrr1 locus is also located on rice chromosome 11 (Fig 8B). Homologues of eight genes from the Bd21 yrr1 locus are present in a similar collinear gene order in the rice genome, although seven B. distachyon gene sequences are absent in this rice interval (Fig 8B). Numerous other genes are also present at both the rice and Brachypodium loci for which no homologue is present at the locus of the reciprocal species. Potential regions of synteny with Bdyrr1are also present on wheat chromosome 4 as illustrated for the 86 Mb region of chromosome 4DS (Fig 8B), while similar genes are also present on chromosomes 4AL (500Mb) and 4BS (123 Mb) (S7 Fig; Fig 8A). A number of wheat stripe rust resistance genes are located on wheat chromosomes 4DS and 4AL, however, none are closely associated with TaYrr1-DS or TaYrr1-4AL (Table 4). No stripe rust resistance genes have been reported in proximity of the TaYrr1-4BS locus.

thumbnail
Table 4. Wheat stripe rust resistance genes on wheat chromosome 4 and their proximity to a region with homology to the Brachypodium Yrr1 locus.

https://doi.org/10.1371/journal.pgen.1007636.t004

Discussion

This study demonstrates that some components of resistance to Pst in B. distachyon are under simple genetic control with two dominant loci, Yrr1 and Yrr2, identified that confer highly restricted pathogen growth, albeit on an underlying background of incompatibility. The B. distachyon accessions that allowed the most Pst colonisation in this study, however, are still obviously resistant indicating that genes additional to Yrr1 and Yrr2 underlie Pst resistance. The observation that the Yrr2 locus is both race specific and encodes a polymorphic NLR gene family suggests that a resistance mechanism typical of host resistance is likely to be involved in B. distachyon resistance to Pst.

B. distachyon is more closely related to wheat than is rice, which may account for the more apparent phenotypic variation in resistance to Pst, a nonadapted pathogen of this model grass species [10, 11]. Paradoxically, however, the rice genome shows greater similarity to BdYrr2 than does the wheat genome. In wheat, NLR homologues of the BdYrr2 locus have expanded to other chromosomal locations while a residual progenitor of the Yrr2 locus that encodes nonNLR genes exists on chromosome 4. In contrast, rice maintains an NLR cluster that is combined with nonNLR genes also present at the BdYrr2 locus. Unlike the extensive rearrangement of Yrr2 NLR homologues in the wheat genome both rice and wheat maintain similar levels of gene collinearity with the BdYrr1 locus.

While phenotypic variation within a species to nonadpated pathogen infection is often under polygenic control [24, 25, 26, 27, 28, 29] exceptions, in addition to this B. distachyon study, have been reported. In wheat, resistance to P. coronata f. sp. hordei [61] and P. striiformis f. sp. hordei [62] was conferred by a single dominant gene in each case. Similarly monogenic resistance to Pst has been reported in barley [63, 64, 65]. It is noteworthy that these latter examples involve closely related host and nonhost species and pathogen formae speciales.

Resistance to nonadapted pathogens is generally considered to be durable as pathogen colonisation of new plant species is rare on short term evolutionary time frames [5, 6]. The results reported here, however, indicate that not all components of this resistance are likely to be durable. The B. distachyon Yrr2 gene shows Pst race specificity, only being effective against some Pst isolates. The resistance provided by this locus, if functionally transferred to a host species, is therefore likely to only be transiently effective against avirulent Pst races given virulent pathogen races already exist.

Two alternative hypotheses may explain the race specificity of the Yrr2 gene. The geographical co-localisation of B. distachyon with wild wheat and its relatives suggest these species may have a close co-speciation history [66, 67]. If so B. distachyon may have had a long exposure to P. striiformis from inoculum provided by adjacent Pst susceptible wheat plants. The ability of some Pst isolates to grow on yrr2 B. distachyon accessions and very occasionally sporulate may represent the early initiation of a host species jump of this pathogen.

An alternative hypothesis is that it is unlikely that Pst has directly evolved to overcome the B. distachyon Yrr2 gene as sporulation rarely occurs on this species. More likely is that Yrr2 recognises an effector or effector modification that is also recognised by a resistance gene present in a host of Pst, thereby providing a clear selective advantage for loss or mutation of this Pst effector in the host pathosystem. However, we could not identify a described wheat stripe rust resistance that co-locates with the major NLR gene clusters on wheat chromosome 4 with homology to BdYrr2 NLR sequences. A question of interest is whether the B. distachyon Yrr2 gene exists solely to provide resistance against Pst and potentially other nonadapted pathogens, or if it also recognises adapted B. distachyon pathogens. Given that singularly this gene appears to provide little selective advantage against Pst infection it seems likely that it may also provide resistance against adapted B. distachyon pathogen species.

This possibility raises some interesting implications. Some components of resistance to nonadapted pathogens that overlap with host resistance (eg. potential recognition of a common effector molecule or activity) may be profoundly affected by host/pathogen dynamics. In the nonhost, a single resistance gene superimposed on an underlying polygenic background of additional resistance is unlikely to result in nonadapted pathogen evolution as the pathogen rarely sporulates. In contrast, it is well established that a host resistance gene that recognises a pathogen effector will exert significant selective pressure for pathogen effector loss or alteration as the end result is extensive pathogen reproduction. If the same effector is also recognised by a gene in a nonhost plant, where it contributes to resistance to the nonadapted pathogen, this latter gene will simultaneously be overcome as a consequence of coevolution between the host and pathogen. Effector based resistance that is common with host resistance is therefore unlikely to remain durable in the nonhost species.

However, it is possible that some NLRs present in nonhost species have evolved to recognise effectors or effector activities in nonadated pathogen species that are essential and not readily modified by mutation. The enforced conservation of these molecules would result in ubiquitous recognition of a pathogen species (ie. broad spectrum). In an accompanying paper Bettgenhaeuser and colleagues (2018) describe the identification of a third B. distachyon Pst resistance locus, Yrr3, that is also possibly conferred by an NLR gene. However, this gene recognises all isolates of Pst tested in addition to P. striiformis f sp. hordei the causative pathogen of barley stripe rust disease.

Genetic haplotype analysis has reduced the Yrr1 locus from 15 candidates to five genes that are polymorphic between immune and more susceptible accessions (Table 1). For all 15 genes at the locus, RNAseq analysis of uninfected tissue in the accompanying paper by Bettgenhaeuser et al shows no consistent transcriptional differences between ABR6 (Yrr1) and yrr1 lines that could account for resistance. A transcriptional difference existing between nonpolymorphic genes therefore appears unlikely to confer immunity, although differential pathogen induction of gene expression cannot be excluded.

In contrast to Yrr2, no NLR genes are present at the Yrr1 locus suggesting a different resistance mechanism. However, it is noteworthy that the resistance phenotypes of Yrr1 and Yrr2 are very similar both macroscopically and microscopically. In both instances cell death at infection sites is limited, more so in BdTR13k than BdTR10h. Cell death, while commonly associated with NLR protein signalling, is not always apparent [68]. The separation of browning at lesions and restricted Pst growth in the BdTR10h x Tek-4 family [41] demonstrates that additional genes that are not essential for resistance also contribute to infection phenotypes of nonadapted pathogens.

Our previous studies indicate that neither Yrr1 nor Yrr2 appear to function against other nonadapted cereal rust pathogens; specifically P. graminis f. sp. tritici, the causal agent of wheat stem rust disease, P. graminis f. sp. avenae (oat stem rust pathogen) and P. graminis f. sp. phalaridi (phalaris rust pathogen) [41]. When compared with other B. distachyon accessions BdTR13k which contains both Yrr1 and Yrr2 did not further restrict the growth of these cereal pathogens and was colonised with larger infection sites and small sporulating pustules. However, it can’t be ruled out that these Yrr genes may provide resistance to other P. graminis isolates not tested in these studies.

Germplasm screening of natural accessions for relatively subtle differential infection outcomes of nonadapted pathogens often identifies an underlying complex genetic inheritance [24, 25, 26, 27, 28, 29]. Fine mapping and ultimately cloning of the genes underlying polygenic resistance mechanisms is a difficult proposition due to the often minor effects conferred by each locus coupled with subtle phenotype variation. Here, B. distachyon is an exception to this generalisation with two independent, dominant loci each conferring readily discernible phenotypes. The Yrr1 locus has been refined to five candidate genes that encode a glyoxal oxidase, a sentrin/SUMO protease, a CBS-domain protein and two proteins of unknown function, respectively, while the race-specific Yrr2 locus has been reduced to three candidate NLR genes. Future studies will delineate which specific gene(s) at each locus is responsible for macroscopic immunity by transforming B. distachyon with these candidate genes.

Of great interest will be to test the ability of these genes to provide Pst resistance in wheat, a host of this pathogen. Previous studies have demonstrated that in some instances genes that provide resistance to nonadapted pathogens in one species can be transferred to host species to provide resistance to virulent races of the same pathogen. Examples include the transfer of the EF-TU (ERF) PAMP receptor (ERF) from Arabidopsis to Solanaceous species [69] the maize Rxo1 NLR gene transferred to rice [70] and Arabidopsis resistance genes to Asian soybean rust (Phakopsora pachyrhizi) transferred into soybean [71]. These observations demonstrate that genes providing resistance against nonadapted pathogens can be a valuable source of disease resistance for crop hosts of these diseases if they can be isolated amongst an often complex background of genetic and mechanistic redundancy.

Materials and methods

Plant growth and rust propagation

B. distachyon plants were grown in a soil/compost (1:1) mixture at 18 oC under standard greenhouse growth conditions or alternatively in growth cabinets at 18 oC with a 16 hour photoperiod. Pst isolates were propagated on wheat cultivar Morocco and urediniospores harvested by shaking infected plants over aluminium foil. For Pst infection B. distachyon plants were inoculated with freshly harvested urediniospores as an aerosol spray, lightly misted with water and then incubated at 10 oC overnight before being returned to 18 oC constant growth conditions. Rust symptoms were scored after approximately 3 weeks. Four Pst pathotypes were used for infection assays; 110 E143 A+ [Plant Breeding Institute accession number (AN) 861725], 134 E16 A+ [AN 021510], 104 E137 A–[AN 821559] and 108 E141 A- [AN 832002].

Molecular biological analyses

DNA extractions, PCR analysis and DNA blot analyses were undertaken as previously described [72]. A 387 bp fragment encoding a region of the NLR genes present at the Yrr2 locus was amplified using primer 5’-TATTGAGAAGATCTTTGAGCA-3’ and primer 5’-TCCCCTTCCATAGATGCTGCC-3’ and used as a probe for DNA hybridisation.

DNA sequencing and bioinformatics

DNA sequencing of genomic DNA pools was undertaken by the Australian Genome Research Facility Ltd using Illumina sequencing with 100 bp paired end reads. DNA sequences were aligned to the Bd21 reference genome and SNPs and indels called using PARTEK and CASAVA programs. The Bd21 reference genome sequence is located at PlantGDB (http://www.plantgdb.org) and the Joint Genome Institute Genome Portal (http://genome.jgi.doe.gov/). Rice genome sequence comparisons were made using the rice genome database OsGDB. Wheat genomic sequences were obtained from the International Wheat Genetics Sequencing Consortium (IWGSC) data repository at URGI-INRA (https://wheat-urgi.versailles.inra.fr/Seq-Repository/Assemblies) and NCBI wheat chromosome 3B nucleotide sequence (Genbank HG670306).

Whole genome shotgun sequencing of BdTR10h and BdTR13k

Genomic DNA of BdTR10h and BdTR13k was extracted using a standard CTAB-based extraction protocol. TruSeq libraries were generated from gDNA and mean insert sizes were 840 bp for BdTR10h and 993 bp for BdTR13k. Library preparation and sequencing was performed at the Earlham Institute (previously known as The Genome Analysis Centre, Norwich, UK). Sequencing was carried out using 100 bp paired-end reads on an Illumina HiSeq 2500 sequencer and yielded 71 and 61 million raw reads for BdTR10h and BdTR13k, respectively.

Haplotype and phylogenetic analysis

Resequencing data for several B. distachyon accessions were obtained from the Joint Genome Institute Genome Portal (http://genome.jgi.doe.gov/) (S9 Table). These sequence data were produced by the US Department of Energy Joint Genome Institute (http://www.jgi.doe.gov/) in collaboration with the user community. Illumina reads were quality controlled using Trimmomatic (Version 0.33) with the following parameters: ILLUMINACLIP:TruSeq3-PE.fa:2:30:10 LEADING:5 TRAILING:5 SLIDINGWINDOW:4:15 MINLEN:36. Alignments to the Bd21 reference (v3.0) were performed with bwa mem (version 0.7.5a-r405) with default parameters. Samtools (version 0.1.19-96b5f2294a) was used to convert sam into bam files (samtools view) with the requirement that reads mapped in a proper pair (-f2), sort the bam file (samtools sort), remove duplicate reads (samtools rmdup), and generate an mpileup file (samtools mpileup). Coverage of reads was determined using bedtools (version v2.17.0; bedtools genomecov -d -split). SNPs and InDels were called using VarScan (version 2.3.8) with default parameters.

The QKgenome suite (version 1.1.2) of Python scripts were used to assess the haplotype diversity within genes at the Yrr1 locus in several B. distachyon accessions. QKgenome_conversion.py was used to evaluate nucleotide variation with the requirement of a read depth of 20 and masking of sequence under this threshold (-m command). Only the first gene model for each gene was used. SNPs and InDels were called based on a frequency threshold of 90%. All genes with InDels that disrupted the coding sequence and mutations with early stop codons were not included in the analysis. A multiple sequence alignment of polymorphic sites was generated using QKgenome_phylogeny.py. The phylogenetic tree was constructed with RAxML (version 8.2.9) using the GTRCAT nucleotide model and rapid hill-climbing mode. Bootstrap support was determined with 2,000 bootstraps, which were found to be sufficient based on the bootstrap convergence test (command -I autoMRE). The neighbour joining tree was generated using Phylip (version 3.695) using default parameters. Bootstrap support was performed with 1,000 bootstraps.

Accession numbers

Sequence raw reads and assemblies for BdTR10h and BdTR13k were deposited in NCBI under BioProject PRJNA377287. Individual gDNA whole genome sequencing reads include accession numbers SRR5298268 (BdTR10h) and SRR5298269 (BdTR13k). The QKgenome suite of Python scripts described in this manuscript have been deposited on GitHub (https://github.com/matthewmoscou/QKgenome).

Microscopy

Microscopy of fungal infection structures was undertaken as previously described [22, 41]. Briefly, harvested leaf samples were autoclaved in 1M KOH and then neutralised in a 50 mM Tris pH 7.5 solution. Samples were then stained with a 20 μg/ml solution of wheat germ agglutinin conjugated to fluorescein isothiocyanate and visualised under blue light. The same tissues were examined for autofluorescence using UV light.

Marker analysis

PCR products were amplified using a Phire Plant Direct PCR Kit (Thermo scientific, USA). For sequencing based markers, PCR products were purified for sequencing using alkaline phosphatase and exonuclease I [73] and sequenced using primer P1. For CAPS marker 9.656179, PCR product was digested with NarI (New England Biolabs, USA) and fragments were resolved on a 1% agarose gel. For KASP marker 9.583752, competitive allele specific polymerase chain reaction was performed (KASP; LGC genomics, UK).

BAC clone library production and screening

B. distachyon BAC libraries for accession BdTR13k, Tek-4 and BdTR10h were produced by Bio S and T (Quebec, Canada). Individual BACs were isolated by PCR screening of BAC clone pools undertaken by Bio S and T. BAC DNA was purified using an Epicenter BACMAX DNA purification kit and clones then deep sequenced at Kansas State University using Illumina MiSeq sequencing. BAC1H was isolated using PCR primers 9735685F (TTGCTGAGCTTCAAGTGGTG) and 9735685R (ATTCCATTGATGACCGCAGC).

Supporting information

S1 Fig. Phenotypic segregation of Yrr1 amongst BdTR10h x Tek-4 progeny.

https://doi.org/10.1371/journal.pgen.1007636.s001

(PPTX)

S2 Fig. Pst analysis of F3 families from critical recombinant F2 plants.

https://doi.org/10.1371/journal.pgen.1007636.s002

(PPTX)

S3 Fig. Neighbor joining phylogenetic tree showing two distinct haplotypes exist at the Yrr1 locus.

https://doi.org/10.1371/journal.pgen.1007636.s003

(PPTX)

S4 Fig. Comparison of Bradi4g24315 proteins.

https://doi.org/10.1371/journal.pgen.1007636.s004

(PPTX)

S5 Fig. Pst infection phenotypes and segregation ratios of F5 families from F4 plants 93 and 56.

https://doi.org/10.1371/journal.pgen.1007636.s005

(PPTX)

S6 Fig. Region of synteny on wheat chromosome 4 that maintains some gene collinearity with non-NLR genes present at the Brachypodium Yrr2 locus.

https://doi.org/10.1371/journal.pgen.1007636.s006

(PPTX)

S7 Fig. Location of co-linear wheat sequences with homology to Brachypodium genes present at the Yrr1 locus.

https://doi.org/10.1371/journal.pgen.1007636.s007

(PPTX)

S1 Table. PCR sequencing markers at the Yrr1 locus.

https://doi.org/10.1371/journal.pgen.1007636.s008

(DOCX)

S2 Table. Genes annotated in the Brachypodium v3.1 genome that are present at the Yrr1 locus.

https://doi.org/10.1371/journal.pgen.1007636.s009

(XLSX)

S3 Table. SNPs present in 11 genes at the Yrr1 locus in a panel of Brachypodium accession.

https://doi.org/10.1371/journal.pgen.1007636.s010

(XLSX)

S4 Table. Pseudogenes excluded from the Yrr1 haplotype analysis.

https://doi.org/10.1371/journal.pgen.1007636.s011

(XLSX)

S5 Table. Yrr1 SNP haplotype analysis and Pst infection phenotypes.

https://doi.org/10.1371/journal.pgen.1007636.s012

(XLSX)

S6 Table. Comparison of Bradi4g24366 between Brachypodium accessions.

https://doi.org/10.1371/journal.pgen.1007636.s013

(XLSX)

S8 Table. Genes annotated in the Brachypodium v3.1 genome that are present at the Yrr2 locus.

https://doi.org/10.1371/journal.pgen.1007636.s015

(XLSX)

S9 Table. Source of resequencing reads for a diverse panel of B. distachyon accessions.

https://doi.org/10.1371/journal.pgen.1007636.s016

(DOCX)

Acknowledgments

Seed of BdTR10h and BdTR13k were kindly provided by Hikmet Budak.

References

  1. 1. Heath MC. Nonhost resistance and nonspecific plant defenses. Current opinions in plant biology. 2000; 3:3154–319.
  2. 2. Mysore KS, Ryu CM. Nonhost resistance: how much do we know? Trends in plant science. 2004; 9: 97–104. pmid:15102376
  3. 3. Nürnberger T, Lipka V. Non-host resistance in plants: new insights into an old phenomenon. Molecular plant pathology. 2005; 6:335345. pmid:20565662
  4. 4. Niks RE, Marcel TC. Nonhost and basal resistance: how to explain specificity? New phytologist. 2009; 182: 817–828. pmid:19646067
  5. 5. van der Merwe MM, Walker J, Ericson L, Burdon JJ. Coevolution with higher taxonomic host groups within the Puccinia/Uromyces rust lineage obscured by host jumps. Mycological research. 2008; 112: 1387–1408. pmid:18675350
  6. 6. de Vienne DM, Refrégier G, López-Villavicencio M, Tellier A, Hood ME, Giraud T. Cospeciation vs host-shift speciation: methods for testing, evidence from natural associations and relation to coevolution. New phytolologist. 2013; 198: 347–385.
  7. 7. Bettgenhaeuser J, Gilbert B, Ayliffe M, Moscou MJ. Nonhost resistance to rust pathogens–a continuation of continua. Frontiers in plant science 2014; 5: 664. pmid:25566270
  8. 8. Jones JDG, Dangl JL. The plant immune system. Nature. 2006; 444:323–329. pmid:17108957
  9. 9. Fan J, Doerner P. Genetic and molecular basis of nonhost disease resistance: complex, yes: silver bullet, no. Current opinions in plant biology. 2012; 15:400–406.
  10. 10. Gill US, Lee S, Mysore KS. Host verses nonhost resistance: distinct wars with similar arsenals. Phytopathology. 2015a; 105: 580–587.
  11. 11. Schulze-Lefert P, Panstruga R. A molecular evolutionary concept connecting nonhost resistance, pathogen host range, and pathogen speciation. Trends in plant science. 2011; 16:117–125. pmid:21317020
  12. 12. Zipfel C. Pattern-recognition receptors in plant innate immunity. Current opinions in immunology. 2008; 20: 10–16.
  13. 13. Boller T, Felix G. A renaissance of elicitors: perception of microbe-associated molecular patterns and danger signals by pattern-recognition receptors. Annual review of plant biology. 2009; 60: 379–406. pmid:19400727
  14. 14. Zipfel C. Plant pattern-recognition receptors. Trends in immunology. 2014; 35:345–351. pmid:24946686
  15. 15. Dong S, Stam R, Cano LM, Song J, Sklenar J, Yoshida K, et al. Effector specialization in a lineage of the Irish potato famine pathogen. Science. 2014; 343:552–555. pmid:24482481
  16. 16. Xao B, Lin X, Poland J, Trick H, Leach J, Hulbert S. A maize resistance gene functions against bacterial streak disease in rice. Proceedings of the national academy of science. 2005; 102:15383–15388.
  17. 17. Wroblewski T, Caldwell KS, Piskurewicz U, Cavanaugh KA, Xu H, Kozik A, et al. Comparative large-scale analysis of interactions between several crop species and the effector repertoires from multiple pathovars of Pseudomonas and Ralstonia. Plant physiology. 2009; 150:1733–49. pmid:19571308
  18. 18. Vega-Arreguin J, Jalloh A, Bis JI, Moffett P. Recognition of an Avr3a homologue plays a major role in mediating nonhost resistance to Phytophora capsici in Nicotiana species. Molecular plant microbe interactions. 2014; 27:770–780. pmid:24725207
  19. 19. Lee H-A, Kim S-Y, Oh S-K, Yeom S-I, Kim S-B, Kim M-S, et al. Multiple recognition of RXLR effectors is associated with nonhost resistance of pepper against Phytophthora infestans. New phytologist. 2014; 203:926–938. pmid:24889686
  20. 20. Stam R, Mantelin S, McLellan H, Thilliez G. The role of effectors in nonhost resistance to filamentous plant pathogens. Frontiers in plant science 2014; 5: article 582.
  21. 21. Mellersh DG, Heath MC. An investigation into the involvement of defense signalling pathways in components of the nonhost resistance of Arabidopsis thaliana to rust fungi also reveals a model system for studying rust fungal compatibility. Molecular plant microbe interactions. 2003; 16: 398–404. pmid:12744510
  22. 22. Ayliffe MA, Devilla R, Mago R, White R, Talbot M, Pryor A, et al. Non-host resistance of rice to rust pathogens. Molecular plant microbe interactions 2001; 24:1143–1155.
  23. 23. Gill US, Uppalapati SR, Nakashima J, Mysore KS. Characterisation of Brachypodium distachyon as a nonhost model against switchgrass rust pathogen Puccinia emaculata. BMC plant biology. 2015b; 15:113.
  24. 24. Atienza SG, Jafary H, and Niks RE. Accumulation of genes for susceptibility to rust fungi for which barley is nearly a nonhost results in two barley lines with extreme multiple susceptibility. Planta. 2004; 220:71–79. pmid:15243739
  25. 25. Jafary H, Szabo LJ, Niks RE. Innate nonhost immunity in barley to different heterologous rust fungi is controlled by sets of resistance genes with different overlapping specificities. Molecular plant microbe interactions. 2006; 19:1270–1279. pmid:17073309
  26. 26. Jafary H, Albertazzi G, Marcel TC, Niks RE. High diversity of genes for nonhost resistance of barley to heterologous rust fungi. Genetics. 2008; 78:2327–2339.
  27. 27. Shafiei R, Hang C, Kang J-G, Loake GJ. Identification of loci controlling non-host disease resistance in Arabidopsis against the leaf rust pathogen Puccinia triticina. Molecular plant pathology. 2007; 8: 773–784. pmid:20507537
  28. 28. Zhang NW, Pelgrom K, Niks RE, Visser RGF, Jeuken MJW. Three combined quantitative trait loci from nonhost Latuca saligna are sufficient to provide complete resistance of lettuce against Bremia lactucae. Molecular plant microbe interactions. 2009; 22:1160–1168. pmid:19656050
  29. 29. Dracatos PM, Nansamba M, Berlin A, Park RF, Niks RE. Isolate specificity and polygenic inheritance of resistance in barley to the heterologous rust pathogen Puccinia graminis f. sp. avenae. Phytopathology. 2016; 106:1029–1037. pmid:27111801
  30. 30. Collins NC, Thordal-Christensen H, Lipka V, Bau S, Kombrink E, Qiu JL, et al. SNARE-protein-mediated disease resistance at the plant cell wall. Nature. 2003; 425:973–977. pmid:14586469
  31. 31. Lipka V, Dittgen J, Bednarek P, Bhat R, Wiermer M, Stein M, et al. Pre- and postinvasion defenses both contribute to nonhost resistance in Arabidopsis. Science. 2005; 310:1180–1183. pmid:16293760
  32. 32. Stein M, Dittgen J, Sanchez-Rodriguez C, Hou B-H, Molina A, Schulze-Lefert P et al. Arabidopsis PEN3/PDR8, an ATP binding cassette transporter, contributes to nonhost resistance to inappropriate pathogens that enter by direct penetration. Plant cell. 2006; 18:731–746. pmid:16473969
  33. 33. Nakoa M, Nakamura R, Kita K, Inukai R, Ishikawa A. Non-host resistance to penetration and hyphal growth of Magnaporthe oryzae in Arabidopsis. Scientific reports. 2011; 1:171. pmid:22355686
  34. 34. Langenbach C, Campe R, Schaffrath U, Goellner K, Conrath U. UDP-glucosyltransferase UGT84A2/BRT1 is required for Arabidopsis nonhost resistance to the Asian soybean rust pathogen Phakopsora pachyrhizi. New phytologist. 2013; 198:536–545. pmid:23356583
  35. 35. Gilbert GS, Webb CO. Phylogenetic signal in plant pathogen-host range. Proceedings of the national academy of science U.S.A. 2007; 104:4979–4983.
  36. 36. Bossolini E, Wicker T, Knobel PA, Keller B. Comparison of orthologous loci from small grass genomes Brachypodium and rice: implications for wheat genomics and grass genome annotation. Plant journal. 2007; 49:704–717. pmid:17270010
  37. 37. Barbieri M, Marcel TC, Nik RE, Francia E, Pasquariello M, Mazzamurro V, et al. QTLs for resistance to the false brome rust Puccinia brachypodii in the model grass Brachypodium distachyon L. Genome. 2012; 55:152–163. pmid:22321152
  38. 38. Barbieri M, Marcel TC, Niks RE. Host status of false brome grass to the leaf rust fungus Puccinia brachypodii and the stripe rust fungus P. striiformis. Plant disease. 2011; 95: 1339–1345.
  39. 39. Draper J, Mur LAJ, Jenkins G, Ghosh-Biswas GC, Bablak P, Hasterok RM, et al. Brachypodium distachyon. A new model system for functional genomics in grasses. Plant physiology. 2001; 127:1539–1555. pmid:11743099
  40. 40. Figueroa M, Alderman S, Garvin DF, Pfender WF. Infection of Brachypodium distachyon by formae speciales of Puccinia graminis: Early infection events and host-pathogen incompatibility. PLoS one. 2013; 8: e56857. pmid:23441218
  41. 41. Ayliffe M, Singh D, Park R, Moscou M, Pryor T. The infection of Brachypodium distachyon with selected grass rust pathogens. Molecular Plant Microbe Interactions. 2013; 26:946–957. pmid:23594350
  42. 42. Michelmore RW, Paran I, Kesseli RV. Identification of markers linked to disease-resistance genes by bulked segregant analysis: a rapid method to detect markers in specific genomic regions by using segregating populations. Proceedings of the national academy of science USA. 1991; 88: 9828–9832.
  43. 43. The International Brachypodium Genome Initiative. Genome sequencing and analysis of the model grass Brachypodium distachyon. Nature. 2010; 463:763–768. pmid:20148030
  44. 44. Nishimura MT, Anderson RG, Cherkis KA, La TF, Liu QL, Machius M, et al. TIR-only protein RBA1 recognises a pathogen effector to regulate cell death in Arabidopsis. Proceedings of the national academy of science USA. 2017; 114: E2053–2062.
  45. 45. Brenchley R, Spaanagl M, Pfeifer M, Barker GL, D’Amore R, Allen AM, et al. Analysis of the bread wheat genome using whole-genome shotgun sequencing. Nature. 2012; 491:705–710. pmid:23192148
  46. 46. Shatalina M, Wicker T, Buchmann JP, Oberhaensli S, Šimková H, Doležel J et al. Genotype-specific SNP map based on whole chromosome 3B sequence information from wheat cultivars Arina and Forno. Plant biotechnology Journal. 2013; 11:23–32. pmid:23046423
  47. 47. Choulet F, Alberti A, Theil S, Glover N, Barbe V, Daron J. Structural and functional partitioning of bread wheat chromosome 3B. Science. 2014; 345: pmid:25035497
  48. 48. Bansal UK, Hayden MJ, Gill MB, Bariana HS. Chromosomal location of an uncharacterised stripe rust resistance gene in wheat. Euphytica. 2010; 171:121–127.
  49. 49. Park RF. (2016). Wheat: biotrophic pathogen resistance. In; Encyclopaedia of food grains (2nd ed.) Vol. 4. The production and genetics of food grains pp. 264–272. Ed. Wrigley C, Corke H, Seetharaman K, Faubion J. Elsevier.
  50. 50. Randhawa MS, Harbans BS, Mago R, Bansal UK. Mapping of a new stripe rust resistance locus Yr57 on chromosome 3BS of wheat. Molecular breeding. 2015; 35:65–72.
  51. 51. Zhou XL, Han DJ, Gou HL, Wang QL, Zeng QD, Yuan FP, et al. Molecular mapping of a stripe rust resistance gene in wheat cultivar Wuhan 2. Euphytica. 2014; 196:251–259.
  52. 52. Xiang C, Feng JY, Wang MN, Chen XM, See DR, Wan AM, Wang T (2016) Molecular mapping of stripe rust resistance gene Yr76 in winter club wheat cultivar Tyee. Phytopathology 106: 1186–1193 pmid:27050567
  53. 53. Suenaga K, Singh RP, Huerta-Espino J, William HM. Microsatellite markers for genes Lr34/Yr18 and other quantitative trait loci for leaf rust and stripe rust in bread wheat. Phytopathology. 2003; 93: 881–890. pmid:18943170
  54. 54. Lowe I, Jankuloski L, Chao S, Chen X, See D, Dubcovsky J. Mapping and validation of QTL which confer partial resistance to broadly virulent post-2000 North American races of stripe rust in hexaploid wheat. Theoretical and applied genetics. 2011; 123:143–157. pmid:21455722
  55. 55. Khlestkina EK, Roder MS, Unger O, Meinel A, Borner A. More precise map position and origin of a durable non-specific adult plant disease resistance against stripe rust (Puccinia striiformis) in wheat. Euphytica. 2006. 153: 1–10.
  56. 56. Zheng S, Li Y, Lu L, Liu Z, Zhang C, Ao D, Li L, Zhang C, Liu R, Luo C, Wu Y, Zhang L (2017) Evaluating the contribution of Yr genes to stripe rust resistance breeding through marker-assisted detection in wheat. Euphytica 231: 50
  57. 57. McIntosh RA, Dubcovsky J, Rogers WJ, Morris C, Appels R, Xia XC (2014) Catalogue of gene symbols for wheat: 2013–2014 supplement. http://maswheat.ucdavis.edu/CGSW/2013-2014_Supplement.pdf
  58. 58. Singh RP, Nelson JC, Sorrells ME (2000) Mapping Yr28 and other genes for resistance to stripe rust in wheat. Crop Science 40:1148–1155.
  59. 59. Randhawa M, Bansal U, Valarik M, Klocova B, Dolezel J, Bariana H (2014) Molecular mapping of stripe rust resistance gene Yr51 in chromosome 4AL of wheat. Theoretical and Applied genetics 127: 317–324 pmid:24185819
  60. 60. Herrera-Foessel SA, Singh RP, Lan CX, Huerta-Espino J, Calvo-Salazar V, Bansal UK, Bariana HS, Lagudah ES (2015) Yr60, a gene conferring moderate resistance to stripe rust in wheat. Plant Disease 99: 508–511.
  61. 61. Niu Z, Puri KD, Chao S, Jin Y, Sun Y, Steffenson B, et al. Genetic analysis and molecular mapping of crown rust resistance in common wheat. Theoretical and applied genetics. 2013; 127:609–619. pmid:24306320
  62. 62. Pahalawatta V, Chen X. Genetic analysis and molecular mapping of wheat genes conferring resistance to the wheat stripe rust and barley stripe rust pathogens. Phytopathology. 2005; 95:427–432. pmid:18943046
  63. 63. Sui X, He Z, Lu Y, Wang Z, Xia X. Molecular mapping of a non-host resistance gene YrpstY1 in barley (Hordeum vulgare L.) for resistance to wheat stripe rust. Hereditas. 2010; 147:176–182. pmid:21039455
  64. 64. Li K, Hegarty J, Zhang C, Wan A, Wu J, Guedira GB, et al. Fine mapping of barley locus Rps6 conferring resistance to wheat stripe rust. Theoretical and applied genetics. 2016; 129:845–859. pmid:26875072
  65. 65. Dawson AM, Feruson JN, Gardiner M, Green P, Hubbard A, Moscou MJ. Isolation and fine mapping of Rps6: an intermediate host resistance gene in barley to wheat stripe rust. Theoretical and applied genetics. 2016; 129:831–843. pmid:26754419
  66. 66. Garvin DF, Gu Y-Q, Hasterok R, Hazen SP, Jenkins G, Mockler TC, et al. Development of genetic and genomic research resources for Brachypodium distachyon, a new model system for grass crop research. The plant genome. 2008; 1:S69–S84.
  67. 67. Vogel JP, Tuna M, Budak H, Huo N, Gu YQ, Steinwand MA. Development of SSR markers and analysis of diversity in Turkish populations of Brachypodium distachyon. BMC plant biology. 2009; 9: 88. pmid:19594938
  68. 68. Periyannan S, Moore J, Ayliffe M, Bansal U, Wang X, Huang L, et al. The Ug99 effective wheat stem rust resistance gene Sr33 is an ortholog of barley Mla genes. Science. 2013; 341:786–788. pmid:23811228
  69. 69. Lacombe S, Rougon-Cardoso A, Sherwood E, Peeters E, Dahlbeck D, van Esse HP, et al. Interfamily transfer of a plant pattern-recognition receptor confers broad-spectrum bacterial resistance. Nature biotechnology. 2010; 28:365–369. pmid:20231819
  70. 70. Zhao B, Lin X, Poland J, Trick H, Leach J, Hulbert S. A maize resistance gene functions against bacterial streak disease in rice. Proceedings of the national academy of sciences USA. 2005; 102:15383–15388.
  71. 71. Langenbach C, Schultheiss H, Rosendahl M, Tresch N, Conrath U, Goellner K. Interspecies gene transfer provides soybean resistance to a fungal pathogen. Plant biotechnology journal. 2016; 14:699–708. pmid:26096357
  72. 72. Ayliffe MA, Collins NC, Ellis JG, Pryor AJ. The maize rp1 rust resistance gene identifies homologues in barley that have been subjected to diversifying selection. Theoretical and applied genetics. 2000; 100:985–1154.
  73. 73. Hanke M, Wink M.Direct DNA sequencing of PCR-amplified vector inserts following enzymatic degradation of primer and dNTPs. BioTechniques. 1995; 17:858–860.