Wheat receptor-kinase-like protein Stb6 controls gene-for-gene resistance to fungal pathogen Zymoseptoria tritici

Deployment of fast-evolving disease-resistance genes is one of the most successful strategies used by plants to fend off pathogens1,2. In gene-for-gene relationships, most cloned disease-resistance genes encode intracellular nucleotide-binding leucine-rich-repeat proteins (NLRs) recognizing pathogen-secreted isolate-specific avirulence (Avr) effectors delivered to the host cytoplasm3,4. This process often triggers a localized hypersensitive response, which halts further disease development 5 . Here we report the map-based cloning of the wheat Stb6 gene and demonstrate that it encodes a conserved wall-associated receptor kinase (WAK)-like protein, which detects the presence of a matching apoplastic effector6–8 and confers pathogen resistance without a hypersensitive response 9 . This report demonstrates gene-for-gene disease resistance controlled by this class of proteins in plants. Moreover, Stb6 is, to our knowledge, the first cloned gene specifying resistance to Zymoseptoria tritici, an important foliar fungal pathogen affecting wheat and causing economically damaging septoria tritici blotch (STB) disease10–12. The authors report map-based cloning of the wheat Stb6 gene, which encodes a conserved wall-associated receptor kinase (WAK)-like protein. Stb6 confers gene-for-gene disease resistance to fungal pathogen Zymoseptoria tritici by recognition of a matching pathogen effector.

Wheat, one of the most important staple food crops, provides 20% of the total daily calories consumed by humans worldwide and in that regard is second only to rice. STB is a devastating disease in most wheat-growing areas of the world. It is the primary foliar disease of wheat in Europe and is responsible for annual wheat losses of 5-10%, with a value of more than $800 (€ 720) million, despite the use of fungicide treatments estimated to cost farmers additional $1.2 billion (€ 1 billion) 10,11 . The causal agent of STB is the fungus Z. tritici, which has recently been described as a latent necrotroph 16 with a strictly extracellular mode of plant pathogenesis 17 . The emergence and dispersal of fungicide resistance in fungal populations 11,18-20 severely threatens wheat production and compromises food security; therefore, STB-resistance breeding is considered a high priority. To date, 21 major genes for resistance to STB (Stb resistance genes), most of which have different specificities based on reactions to pathogen isolates, and numerous minor-effect resistance quantitative trait loci (QTLs) have been mapped genetically 21 . However, none of these genes have been cloned, and the mechanisms of resistance remain poorly understood. Owing to a lack of well-defined QTLs with additive effects and the near absence of diagnostic markers, current STB-resistance breeding strategies rely primarily on phenotypic evaluation of breeding materials rather than targeted-genotyping-based selection, although deployment of broad-spectrum resistance genes, such as Stb16q identified in synthetic wheat 22 , and targeted stacking of isolate-specific Stb resistance genes are also being considered 22,23 .
Stb6, the best-characterized gene for resistance to STB, has been reported to be present in wheat used in breeding programs worldwide, on the basis of phenotypic evaluation 24,25 . It is inherited and manifests as a semidominant trait 24 ( Supplementary Fig. 1a), controlling a gene-for-gene type resistance 6 effective against Z. tritici isolates, such as IPO323, which carry a matching AvrStb6 gene encoding a small cysteine-rich effector protein 7,8 . Stb6 is particularly interesting because it confers pathogen resistance in the absence of a hypersensitive response 9 . This gene has been suggested to exist in wheat since the mid-Neolithic period 24 , and it contributes to field resistance 26 . Stb6 resides in the subtelomeric portion on 3AS in the wheat varieties Flame 6 , Chinese Spring (CS) and Cadenza (Cad) ( Fig. 1 and Supplementary Fig. 1b,c). To better understand the mechanism of resistance against Z. tritici, we isolated the Stb6 gene from CS by using a map-based cloning approach.
Using published wheat genetic maps 27,28 and exploiting the synteny between wheat genomes and model grass genomes 29 , we identified a physical region in the Brachypodium distachyon genome syntenic to the Stb6 locus in wheat. Close to the center of this 769-kb region lies a cluster of 19 genes ( Supplementary Fig. 2) annotated as RLKs, a class of genes that includes well-known regulators of plant innate immunity 30 . Homologous genes were identified in the wheat CS chromosome-arm 3AS assembly and used for developing new genetic markers ( Supplementary Fig. 2). Five of these markers were mapped by using a large F 2 population developed from a cross between CS and the susceptible wheat variety Courtot (Ct) within an ~0.67-cM interval, including two markers that cosegregated with Stb6 (Fig. 1). Using a BAC library and an available draft CS whole-genome assembly, we delimited the Stb6 locus to either of two candidate genes: TaWAKL3 and TaWAKL4 ( Fig. 1 and Supplementary Fig. 3).
Five complementary approaches were then used for functional validation of candidate genes. First, gene expression analysis at six different time points after mock or Z. tritici IPO323 inoculation of CS wheat showed that TaWAKL3 was only minimally expressed under these conditions, whereas TaWAKL4 showed a moderate level of expression and was upregulated approximately twofold during attempted infection ( Supplementary Fig. 4). Second, exon resequencing identified no polymorphisms in TaWAKL3 between resistant (CS) and susceptible (Ct) wheat, whereas the coding sequence of TaWAKL4 in Ct contained a missense mutation causing a p.Ile447Asp amino acid change ( Supplementary Fig. 5). Third, knockdown of expression of TaWAKL4 but not TaWAKL3 through virus-induced gene silencing (VIGS) 31,32 compromised Stb6mediated resistance in CS and Cad wheat (Fig. 2 Tables 3 and 4). All ten families with mutations in TaWAKL3 remained resistant to this fungal isolate (Fig. 3a,b), whereas susceptible individuals from eight families homozygous for critical mutations in TaWAKL4 were identified (Fig. 3c,d). Finally, we stably transformed the susceptible wheat varieties Ct and Bobwhite with TaWAKL4 or with the predicted fulllength coding sequence of this gene driven by its native promoter or the maize polyubiquitin promoter (Ubi1) (Fig. 4a). T 0 plants generated for each construct were self-fertilized. Analysis of the T 1 generation identified families segregating for resistance to Z. tritici IPO323, and all resistant individuals tested positive for the corresponding transgene (Fig. 4b-d). These results verified the Stb6 gene identity and the accuracy of the inferred gene structure. Importantly, Bobwhite transgenic plants expressing Stb6 from the native promoter or the maize Ubi1 promoter showed specific gene-for-gene resistance to Z. tritici isolate IPO323 (Supplementary Fig. 8).

NATuRe GeNeTIcS
TaWAKL4/Stb6 contains four exons and three introns, and the first intron and the second exon are particularly long (2.8 kb) and short (36 bp), respectively (Fig. 1). A transcript originating from this gene mapped with RACE PCR (data not shown) showed leafspecific developmentally regulated expression peaking in flag leaves after anthesis ( Supplementary Fig. 9). The predicted Stb6 resistance protein consists of 647 amino acids and contains an extracellular galacturonan-binding domain (GUB_WAK), an intracellular nonarginine-aspartate 36 protein kinase, and a complex-topology concanavalin A-like domain ( Supplementary Fig. 10). In contrast, all other known WAKs implicated in pathogen defense contain additional extracellular domains located downstream of GUB_WAK, such as wall-associated receptor-kinase C-terminal or EGF-like calcium-binding domains ( Supplementary Fig. 11).
Exon resequencing identified a notable sequence conservation of Stb6 in the hexaploid bread wheat Triticum aestivum. Only eight haplotypes were identified among 98 accessions (Supplementary  Tables 5 and 6), and a single resistance haplotype predominated, including in 15 of 25 of the most highly resistant and in 10 of 19 of the most commonly grown recent and current UK varieties (Supplementary Tables 7 and 8). This result indicates that defense pathways activated by Stb6 may have no or minimal associated fitness cost. Remarkably, Stb6 haplotypes were also identified in several A-genome-containing domesticated and wild tetraploid and diploid Scale bar, 25 mm. c, A highly significant (***) effect of silencing specific candidate genes on disease severity, assessed on a scale from 1 (no symptoms) to 6 (80-100% leaf area covered by pycnidia-bearing necrotic lesions). The number of leaves (n) and the P values for the approximate two-tailed t test for comparison of silenced plants to those treated with a negative control are shown. d, A highly significant (***) effect of silencing specific candidate genes on fungal sporulation in the infected leaves. The number of replicate samples (n) of spores washed off the infected leaves and the P values for the post hoc two-tailed t test for comparison of VIGS-treated and negative-control-treated plants are shown. In each box-and-whisker plot, the center lines (red) indicate the medians; the bottom and top edges of the boxes indicate the twenty-fifth and seventy-fifth percentiles; whiskers mark the range of the data from the tenth to ninetieth percentiles, and black dots indicate data points that lie outside of this interval. NS, not significant.

NATuRe GeNeTIcS
wheat species ( Fig. 5a and Supplementary Tables 5 and 6). The prevalence of a resistance haplotype in Triticum dicoccum, one of the earliest cultivated forms of wheat, suggests that Stb6 might have been introduced into agriculture during early wheat domestication, thus potentially explaining its widespread occurrence in bread wheat. Ct wheat contains an expressed susceptible haplotype of Stb6, which differs from the resistance haplotype by a single nonsynonymous SNP ( Supplementary Fig. 5) causing a change from a conserved isoleucine residue to an aspartate at position 447 in the catalytic site of the protein kinase domain (Fig. 5b). All mutations associated with susceptibility to Z. tritici IPO323 identified in the Cad TILLING population (except for one nonsense mutation and one potential splice-site mutation) also led to changes at conserved amino acid residues in the kinase domain of Stb6 (Fig. 3a). Biochemical assays suggested that disease susceptibility associated with these mutations probably results from a loss of kinase catalytic activity and thus abrogated immunological signaling (Fig. 5c).
Z. tritici belongs to a group filamentous ascomycete fungal pathogens that do not penetrate host cells or form specialized feeding structures (haustoria) but instead colonize and extract nutrients from the plant extracellular space 17 . Many of these pathogens are also of utmost agronomic importance 37,38 , for example, Mycosphaerella fijiensis, Leptosphaeria maculans, and Rhynchosporium commune, which cause major diseases affecting banana, canola, and barley, respectively.
Host resistance to these pathogens is typically governed by the PRRlike receptor-like proteins or RLKs (rather than cytoplasmic NLRs), which recognize fungal secreted effectors in the plant apoplastic space and transduce defense signals through interaction with the accessory RLKs 37,38 , thus affirming the absence of a strict separation between PRRs and resistance proteins 14,15 . Our study provides additional evidence supporting this concept, confirming that WAK receptor proteins are new players in plant innate immunity against extracellular pathogens and adding another twist by demonstrating that PRRlike proteins of this class, such as wheat Stb6, can control qualitative pathogen resistance in a gene-for-gene manner through recognition of apoplastic Avr effectors. This functionality contrasts with that of Arabidopsis thaliana RFO1 (WAK-like 22) 39 and the two recently cloned maize WAKs implicated in broad-spectrum, but partial, quantitative resistance 40,41 . WAKs monitor and respond to changes in the cell wall during plant development or pathogen attack through binding to cross-linked cell-wall pectin or oligogalacturonides, respectively 42-45 . There is evidence that some WAKs may also bind proteinaceous ligands 46 . A recent study 47 has identified wheat susceptibility/sensitivity protein Snn1 as a WAK that interacts with the secreted protein Tox1 from the necrotrophic fungus Parastagonospora nodorum, thereby inducing extensive tissue necrosis and consequently providing nutrients for pathogen growth and reproduction in a process termed necrotrophic-effector-triggered susceptibility.

NATuRe GeNeTIcS
We hypothesized that Stb6 might bind to its matching recently cloned 7,8 effector AvrStb6 from the avirulent Z. tritici isolate IPO323. Because this effector has been reported to show a high level of polymorphism but no presence/absence variation in field populations of Z. tritici, we therefore also anticipated that its alternative alleles 7,8 from virulent fungal isolates cannot be recognized by Stb6. We used yeast two-hybrid (Y2H) assays to test this hypothesis, but no direct interaction between Stb6 and any of the three different AvrStb6 sequence variants was detected ( Supplementary  Fig. 12). Y2H may be suboptimal for assaying interactions between apoplastic proteins, and further tests, including in planta assays, will be required to confirm this initial result. If, however, the lack of direct Stb6-AvrStb6 interaction is genuine, at least two alternative scenarios may be possible: (i) initiation of the immune response may involve additional interactions possibly involving pectin, oligogalacturonides or other plant cell-wall-derived signals, and/or (ii) AvrStb6 may interact with another protein that is 'guarded' (monitored) by Stb6.
Cloning of Stb6, together with the recent discovery that the matching Z. tritici effector is maintained in fungal populations because avirulent isolates can mate with virulent isolates even on resistant host plants, emphasizes the value of Stb6 for controlling STB disease while also providing new fundamental insights into the molecular control of plant-pathogen interactions.  Vertical black lines and empty areas inside the rectangles indicate positions of SNPs and short insertions and deletions (indels), respectively, as compared with haplotype 1. Numbers on the right indicate SNPs and indels identified for each haplotype, and numbers in parentheses indicate numbers of resulting polymorphic amino acid residues in the predicted encoded proteins, as compared with haplotype 1. Asterisks indicate haplotypes containing a 1-nt insertion at position 904 that results in a frame shift and a premature stop codon at nucleotide positions 1033-1035. b, Graphical representation, produced in WebLogo, of a multiple sequence alignment for predicted protein kinase active sites from Stb6 and 218 protein kinase sequences from diverse plant species. The highly conserved isoleucine residue at position 447 (I447; boxed, and outlined in red) in Stb6 is replaced by aspartate (N) in the protein encoded by haplotype 3 found in susceptible wheats such as Ct. c, Biochemical analysis of four variants of the Stb6 protein kinase domain expressed in Escherichia coli from resistant and natural or induced susceptible mutants. Pro-Q Diamond phosphoprotein stain shows substantial autophosphorylation of Stb6 from wild-type Cad and near-complete absence of autophosphorylation of Stb6 variants from the three susceptible wheat genotypes.

Methods
Methods, including statements of data availability and any associated accession codes and references, are available at https://doi. org/10.1038/s41588-018-0051-x.

NATuRe GeNeTIcS
Methods Plant materials and Zymoseptoria tritici resistance assays. Parents of published or available in-house wheat mapping populations were used in initial pathoassays to verify the data from early reports 24,25 suggesting that Stb6 or an allelic gene is present in CS and Cad wheat but absent in Avalon (Av) and Ct wheat. A selection from these and/or the well-known highly susceptible wheat Riband 9 were used as controls in other pathoassays in the study. The wild-type CS along with the lines 3AS4-0.45, 3AS2-0.23, and 3AS1-0.13, carrying induced deletions 48 of chromosome arm 3AS were used for locating Stb6 to the 0.45-1.00 deletion bin in the subtelomeric region on 3AS. Twenty-seven F 2 progeny of a Cs × Ct cross were used in pathoassays for reconfirming an earlier observation 6 that Stb6 is a semidominant gene. Ninety-six doubled-haploid lines derived from a CS × Ct cross 49 and 40 doubled-haploid lines derived from an Av × Cad cross 50,51 were used in pathoassays in conjunction with genotyping assays using publicly available simple-sequencerepeat markers (selected from the GrainGenes database) for production of a lowresolution genetic map around the Stb6 locus. A biparental mapping population derived from a CS × Ct cross comprising 1,962 F 2 individuals was established for fine mapping Stb6.
Genetic mapping, physical map construction, sequencing, and annotation. New genetic markers used for construction of a high-resolution map at the Stb6 locus in wheat were developed as described in the Supplementary Note. To expedite fine mapping, the F 2 CS × Ct population was first genotyped with gwm369 and cfa3010, which flank Stb6, and individuals with recombination events between these marker loci were then genotyped with all other markers (Supplementary Tables 9 and 10) described above as well as subjected to a fungal pathoassay. Forty critical recombinant F 2 plants were selfed, and their progeny (at least 20 F 3 plants per family) were tested in pathoassays.
To construct the physical map spanning the Stb6 interval, we initially (i) identified and sequenced an ~100-kb BAC clone after screening the wheat CS Tae-B-CsE BAC library and (ii) identified an overlapping ~118-kb genomic contig through BLASTn analysis against the wheat whole-genome assembly TGACv1 (Supplementary Note). These two large continuous sequences combined contained genetic markers ctg8311 and cfn80023, which cosegregated with Stb6, as well as markers cfn80025 and cfn80030/cfn80040, which flanked Stb6. Subsequently, when the substantially improved IWGSC wheat CS whole genome assembly v1.0 became available, we confirmed the above data by identifying a continuous 400-kb genomic sequence containing Stb6 and spanning the interval between markers cfn80025 and cfa3010. The gene models were annotated and manually curated as described in the Supplementary Note.

Virus-induced gene silencing (VIGS)
. VIGS for functional analysis of candidate wheat genes was carried out essentially as previously described 32 . Gene-silencing constructs were created by cloning fragments of wheat gene sequences into the BSMV RNAγ -derived binary vector pCa-γ bLIC 53 in antisense orientation. Three nonoverlapping fragments of TaWAKL4, one from the GUB_WAK domain and two from the kinase domain, were cloned separately into pCa-γ bLIC to generate three independent VIGS constructs designed to target this gene for silencing (BSMV::asTaWAKL4a, BSMV::asTaWAKL4b, and BSMV::TaWAKL4c). Two constructs designed to target the neighboring gene TaWAKL3 (BSMV::TaWAKL3a and BSMV::TaWAKL3b) were also generated. The target gene fragments for cloning were generated by standard RT-PCR with primers described in Supplementary Table 9, and total RNA was extracted from CS leaf tissue as a template for reverse transcription. The BSMV::mcs4D control construct contained a 275-nt noncoding DNA sequence amplified from the multiple cloning site of the pBluescript II SK vector (Agilent Technologies).
Samples for qRT-PCR analysis to determine target-gene silencing success were harvested from the tips of the third leaves of wheat plants at 11-14 d post virus inoculation (dpi) and immediately before Z. tritici inoculation. A minimum of three independent samples per virus treatment, each sample harvested from an individual plant, were analyzed. Quantification of gene expression with primers described in Supplementary Table 9 was carried out with SYBR Green Jumpstart Ready Mix (Sigma Aldrich), with an annealing temperature of 60 °C, in an ABI 7500 Real-Time PCR system (Applied Biosystems).
Attached-wheat-leaf infection assays with the Z. tritici isolate IPO323 on virus-infected plants were carried out as described previously 32 . Disease was assessed at 21 dpi by scoring the area of Z. tritici-inoculated leaf tissue that was both necrotic and evenly covered by fungal asexual fruiting bodies (pycnidia). The disease severity was scored on a scale from 1 to 6 corresponding to 0, 1-20, 21-40, 41-60, 61-80, and 80-100 percent leaf coverage by fungal pycnidia. After visual assessment, pycnidiospores were washed from Z. tritici-inoculated leaf segments and counted as previously described 32 . Each sample comprised three 6-cm-long leaf segments, each from an individual wheat plant. Pycnidiospores from a minimum of three replicate samples from each virus treatment were counted in each experiment, and data were pooled from a minimum of three and two independent experiments with CS and Cad wheat, respectively.

Statistical analyses.
To determine whether treatments of wheat plants with specific BSMV VIGS constructs resulted in decreased expression of target genes but not nontarget genes, we applied statistical analysis through analysis of variance (ANOVA) (F test) followed by comparison of means with post hoc two-tailed t test. For this analysis, the log 2 (1/NRQ) qRT-PCR data were analyzed as previously described 54 , where NRQ is the normalized relative quantity (2 -Ct target /2 -Ct reference ), for the target genes TaWAKL4 and TaWAKL3 and the reference gene CDC48, which has previously been determined to be a suitable reference gene in BSMV-infected leaf tissue 55 .
To determine the effect of silencing candidate Stb6 genes in wheat through VIGS on the outcome of Z. tritici infection, GenStat 18 (https://www.vsni.co.uk/ software/genstat/) was used as described previously 55 . A generalized linear model was fitted to the disease-severity data (scores from 1 to 6), by assuming a Poisson distribution and using a log-link function, to test (F test) for the overall significance of difference between genotypes. Comparison of mean disease scores of TaWAKL4-and TaWAKL3-silenced plants with those of plants treated with the negative control BSMV::mcs4D was made with approximate t tests. Separate modeling exercises were done for data derived from CS and Cad wheat backgrounds. ANOVA was applied to the fungal-spore-count data on the natural log scale with an adjustment of + 1 to account for observations of zero counts. The transformation ensured an approximate normal distribution and homogeneous variance over the genotypes, on the basis of checked residuals from the analysis. After a significant (P < 0.05) F-test result, means for spore counts from TaWAKL4and TaWAKL3-silenced plants were compared with those of plants treated with the negative control BSMV::mcs4D by using post hoc two-tailed t tests based on the residual variance and degrees of freedom from the ANOVA.
Two-way ANOVA was applied to the CS wheat RNA-seq data (an infection time course) for the two genes, TaWAKL4 and TaWAKL3, testing the main effects and interactions between the factor of treatment (mock inoculated and Z. tritici inoculated) and time (2, 5, 8, 11, 14, and 17 dpi). Comparison of means was done with post hoc two-tailed t tests.
One-way ANOVA was applied to natural-log (FPKM + 0.5) data for the TaWAKL4 gene expression in the leaf tissue only (because no nonzero FPKM data were obtained from other tissues to contribute variation for the analysis). FPKM means for the three growth stages (Z10, Z23, and Z71) were compared with post hoc two-tailed t tests.

Analysis of EMS-derived mutants. An ethyl methanesulfonate (EMS)mutagenized population of 1,200 M 5 mutant families of Cad wheat 34,35 containing
Stb6 was used in this study. An 84-Mb exome capture assay comprising overlapping probes covering 82,511 nonredundant wheat genes was used to capture (through Roche NimbleGen array technology) and sequence (through Illumina GA II 110bp paired-end-read technology) the coding gene regions from all mutant families and to identify induced mutations, as detailed in ref. 35 . Potential mutations in TaWAKL3 and TaWAKL4 were identified with BLASTn analysis of these gene sequences against the database of mutations induced in Cad wheat (wheat TILLING database), which is part of a joint project 35 between the University of California Davis in the US, and Rothamsted Research, Earlham Institute, and the John Innes Centre in the UK. All 28 and 11 randomly selected M 5 families (five or six individuals per family) with predicted mutations in the coding TaWAKL3 and TaWAKL4 gene sequence, respectively, identified through in silico analysis (Supplementary Tables 3 and 4), were tested for fungal resistance in a glass-housebased bioassay 52 . The presence of predicted mutations and mutation zygosity in each of the M 5 individuals identified as susceptible to Z. tritici IPO323 was verified through exon resequencing in the target genes by using primers described in Supplementary Table 9. In addition, all these individual susceptible M 5 plants were self-fertilized, and their progeny (at least 24 M 6 plants per family) were tested in pathoassays to confirm the susceptibility phenotype.

Wheat transformation and analysis of transgenic plants.
A genomic sequence of approximately 12 kb containing the full-length TaWAKL4/Stb6 gene (construct 1) was PCR amplified from the BAC clone Tae-B-CsE-673A7 with Phusion High-Fidelity PCR Master Mix (Thermo Fisher Scientific) and primers 8311F12/8311R12 (Supplementary Table 9) and cloned into the pCR8/GW/TOPO vector (Thermo Fisher Scientific). The integrity of the cloned genomic sequence was verified by Sanger sequencing of PCR fragments produced with a set of primer pairs distributed along the Stb6 gene sequence (Supplementary Table 9). A deletion toward the 3′ end of the sequence was identified; however, because it was located downstream of the Stb6 transcriptional termination site inferred from the 3′ -RACE analysis, the cloned Stb6 gene sequence was full length and therefore suitable for genetic complementation. The pCR8/GW/TOPO vector carrying Stb6 was double digested with EcoRV and PspOMI, and the cloned wheat genomic DNA fragment was purified after agarose gel electrophoresis and dephosphorylated as previously described 56 . This fragment was then mixed with the bar dephosphorylated cassette at a 2:1 ratio and used for transformation of immature embryos of Ct wheat by particle bombardment 56 with a PDS 1000 He device (Bio-Rad). Regeneration of plants and bar selection were performed essentially as previously reported 56 . Detection of the Stb6 gene in T 0 plants was performed by PCR amplification using plant genomic DNA as the template and primers pCR8/GW_Stb6F1 and  Table 9). Six independent transgenic lines were identified. T 0 plants were allowed to self-pollinate, and the resulting progeny (at least 12 T 1 individuals originating from each T 0 parent plant) were assessed for resistance to Z. tritici IPO323 at the seedling stage.
The inferred full-length Stb6 gene CDS and a 2-kb putative Stb6 promoter from wheat CS, flanked by AatII and NotI or EcoRV and AatII restriction sites, respectively, were synthesized commercially (Life Technologies). The Stb6 promoter sequence digested with EcoRV and AatII was combined with the Stb6 CDS digested with AatII and NotI, and then cloned upstream of the Nos terminator into the pRRes14_RR.001_65 vector codigested with SmaI and NotI to create the plasmid p65:R-promo-CDS (construct 2). After digestion with AatII and NotI, the Stb6 CDS was cloned between the maize Ubi1 promoter and the Nos terminator into the pRRes14_RR.1m201_125 vector codigested with AatII and NotI to create the plasmid p125:CDS-R (construct 3). Each of these plasmids was mixed with the plasmid pAHC20 containing the selectable bar gene for resistance to herbicide 57 and transformed into immature embryos of Bobwhite wheat after a particle-bombardment procedure, essentially as previously described 58 . Regenerated plantlets in soil were analyzed by PCR to identify transformants. Genomic DNA was extracted from young leaf material with a Wizard Genomic DNA Purification kit (Promega). PCR analysis was carried out for the gene of interest and the selectable marker gene (bar) as follows. Constructs 2 and 3 were both analyzed with the primer pair R-gene-fwd and Nos5′ rev. Additionally, construct 3 transformants were analyzed with primers UbiPro4 and R-gene-rev. The bar gene was detected with primers bar1 and bar2 (Supplementary Table 9). Transgene and bar gene copy number analyses were carried out in T 0 and T 1 generations by iDNA Genetics (Norwich, UK). Identified T 0 plants were selfed, and the resulting progeny, at least ten T 1 individuals originating from each T 0 parent plant, were assessed for resistance to Z. tritici IPO323 at the seedling stage. Selected T 1 individuals, mostly those predicted to be homozygous for the corresponding transgene, were selfed. The resulting progeny, at least ten T 2 individuals originating from each T 1 parent plant, were tested for resistance to Z. tritici isolates IPO323 (avirulent on Stb6 wheat genotypes), and IPO88004 and RRes116 (both virulent on Stb6-containing wheat) with the attached-seedling-leaf bioassay 52 .
Haplotype analysis. We assembled a collection of 98 bread wheat (T. aestivum) accessions comprising varieties previously reported as potential carriers of Stb6; the well-known susceptible wheat varieties Obelisk, Riband and Longbow 6,24,59,60 ; recent and current widely cultivated UK and French varieties with good field resistance to STB, obtained from different seed companies; and 48 highly genetically diverse genotypes selected from a worldwide bread-wheat core collection 61 with MSTRAT software 62 (Supplementary Tables 5-8). This collection was used for resequencing Stb6 exons with the primers listed in Supplementary Table 9. Similarly, six diploid (one Triticum monococcum, two Triticum boeticum, and three Triticum urartu) and 31 tetraploid (five Triticum dicoccoides, five T. dicoccum, two Triticum polonicum, two Triticum turgidum, and 12 Triticum durum) wheat accessions were used for resequencing Stb6 exons to evaluate the evolutionary origin of Stb6. Comparison of the different identified haplotypes was performed with Molecular Evolutionary Genetics Analysis (MEGA) v5.10 software 63 .
Kinase assay. Coding Stb6 sequence corresponding to the cytoplasmic S/T kinase domain containing a region of the encoded WAK protein was PCR amplified with primers Kin1186-attB1-F1 and Kin1186-attB1-R1 (Supplementary Table 9) with Phusion High-Fidelity DNA polymerase (Finnzymes) and either the plasmid p125:CDS-R containing the full-length Stb6 CDS from CS wheat or the first-strand cDNA derived from total RNA extracted from Z. tritici IPO323-infected Av wheat (containing the same susceptibility Stb6 allele as that in Ct) as a template. Two additional PCR fragments, each containing one SNP (preceding the amino acid change p.Gly387Glu or p.Glu522Lys as in Cad TILLING mutants #1495 and #0449, respectively) were generated by overlap-extension PCR 64 with primers Kin1186-attB1-F1 and Kin1186-attB1-R1 in combination with primers G387E-F1 and G387E-R1, or E522K-F1 and E522K-R1, containing the corresponding point mutation (Supplementary Table 9). These PCR fragments were recombined into the pDONR221 vector (Thermo Fisher Scientific) with Gateway BP Clonase II enzyme mix (Thermo Fisher Scientific), and the resulting entry clones were verified through Sanger sequencing at Eurofins Genomics (Ebersberg, Germany). Entry clones were then recombined with Gateway LR Clonase II enzyme mix (Thermo Fisher Scientific) into the pDEST-HisMBP vector 65 for expression of recombinant proteins with an N-terminal hexahistidine-maltose-binding protein dual tag and transformed into the E. coli strain Rosetta (DE3) (Novagen).
For protein expression, PCR-verified bacterial colonies were first inoculated into 5 ml LB and grown at 37 °C overnight (16 h) with shaking at 220 r.p.m. One milliliter of cell suspension was then inoculated into 100 ml of fresh LB and grown for approximately 2-3 h at 18 °C with shaking at 220 r.p.m. until the OD 600 reached 0.5. At that point, IPTG was added to a final concentration of 0.5 mM. Protein expression then continued for a further 20 h at 18 °C with shaking at 220 r.p.m. Cells were collected by centrifugation and lysed with CelLytic B Cell Lysis Reagent (Sigma Aldrich) according to the manufacturer's protocol (4 ml per bacterial pellet). The final clarified supernatant containing the soluble proteins was then incubated with 500 µ L of prewashed HIS-Select HF Nickel Affinity Gel (Sigma Aldrich), and affinity binding, washing, and elution were performed according to the supplier's protocol.
To directly determine the phosphorylation levels of each eluted protein, we followed a previously described protocol 66 for recombinant expression of plant receptor protein kinases. Briefly, equal protein amounts were separated on SDS-PAGE gels (Laemmli). One gel was subjected to the ProQ Diamond (Thermo Fischer Scientific) phosphoprotein staining protocol 66 and imaged in an Odyssey Fc Imaging System (LI-COR Biosciences) with image capture with Image-Studio V5.2. Those gels were subsequently counterstained with Coomassie brilliant blue to verify equal protein loading. A second series of gels were equally loaded and blotted onto nitrocellulose membranes, then subjected to western blot analysis with an anti-Stb6 kinase domain peptide (sequence DVQSGSSTRSEETSL) antiserum produced at Eurogentec (Seraing, Belgium) with standard protocols.
Yeast two-hybrid (Y2H) protein-protein interaction assay. Nucleotide sequences corresponding to the predicted extracellular and intracellular regions of Stb6 (amino acids 25-257 and 281-647, respectively); mature (lacking signal peptide) AvrStb6 from isolates IPO323 (avirulent), IPO88004 (virulent), and RRes16 (virulent); and mature (lacking signal peptide) Z. tritici effector Zt10 (Ensembl Fungi database accession Mycgr3P111505) not known to be recognized by Stb6 were PCR amplified from the plasmid p125:CDS-R containing the full-length Stb6; from the first-strand cDNAs derived from total RNA extracted from the corresponding 6-d-old fungal cultures grown on solid YPD agar medium; and from a cDNA clone 67 , respectively. PCR amplifications were performed with Phusion High-Fidelity DNA Polymerase (Finnzymes). The obtained attB-flanked PCR products were first cloned into the Gateway-compatible vector pDONR221 with BP clonase II enzyme mix and were then recombined into the ProQuest Two-Hybrid System yeast expression vectors pDEST32 and pDEST22 with LR clonase II enzyme mix, according to the manufacturer's (Thermo Fisher Scientific) instructions. All generated Y2H prey and bait constructs were verified by sequencing. Primers for PCR and sequencing are detailed in Supplementary Table 9.
A yeast (Saccharomyces cerevisiae) strain MaV203 was then cotransformed with specific Stb6 bait constructs and the corresponding fungal effector prey constructs (including a negative-control effector Zt10), and vice versa. The same yeast strain cotransformed with the A. thaliana GAI (bait) and ARR1 (prey) protein constructs was used as a positive control for protein-protein interaction 68 . Four representative yeast transformants selected on SC/-Leu/-Trp agar plates from each transformation were picked for assessing induction of HIS3 and URA3 genes reporting positive protein-protein interactions. Histidine and uracil auxotrophy was tested by spotting yeast cells diluted in sterile saline onto SC/-Leu/-Trp/-His agar plates containing 0, 10, 25, 50, or 100 mM HIS3 inhibitor 3-amino-1,2,4triazole (3AT) and onto SC/-Leu/-Trp/-Ura agar plates, respectively.  More than 5 samples, and generally more than 10 samples, were used in each experiment. Each experiment was replicated at least twice. Please see the figure legends and online methods. Minimal (or in some cases exact) sample sizes used are indicated in specific sections of online methods.

Data exclusions
Describe any data exclusions.
No data were excluded from the analyses

Replication
Describe whether the experimental findings were reliably reproduced. For each experiment all attempts at replication were successful

Randomization
Describe how samples/organisms/participants were allocated into experimental groups.
Replicated samples were allocated in groups on a random basis

Blinding
Describe whether the investigators were blinded to group allocation during data collection and/or analysis.
Describe the extent of blinding used during data acquisition and analysis. If blinding was not possible, describe why OR explain why blinding was not relevant to your study.
Note: all studies involving animals and/or human research participants must disclose whether blinding and randomization were used.

Statistical parameters
For all figures and tables that use statistical methods, confirm that the following items are present in relevant figure legends (or the Methods section if additional space is needed).

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement (animals, litters, cultures, etc.) A description of how samples were collected, noting whether measurements were taken from distinct samples or whether the same sample was measured repeatedly.
A statement indicating how many times each experiment was replicated The statistical test(s) used and whether they are one-or two-sided (note: only common tests should be described solely by name; more complex techniques should be described in the Methods section) A description of any assumptions or corrections, such as an adjustment for multiple comparisons The test results (e.g. p values) given as exact values whenever possible and with confidence intervals noted A summary of the descriptive statistics, including central tendency (e.g. median, mean) and variation (e.g. standard deviation, interquartile range)

Clearly defined error bars
See the web collection on statistics for biologists for further resources and guidance.