Magnaporthe oryzae pathotype Triticum (MoT) can act as a heterologous expression system for fungal effectors with high transcript abundance in wheat

Plant pathogens deliver effector proteins to reprogramme a host plants circuitry, supporting their own growth and development, whilst thwarting defence responses. A subset of these effectors are termed avirulence factors (Avr) and can be recognised by corresponding host resistance (R) proteins, creating a strong evolutionary pressure on pathogen Avr effectors that favours their modification/deletion to evade the immune response. Hence, identifying Avr effectors and tracking their allele frequencies in a population is critical for understanding the loss of host recognition. However, the current systems available to confirm Avr effector function, particularly for obligate biotrophic fungi, remain limited and challenging. Here, we explored the utility of the genetically tractable wheat blast pathogen Magnaporthe oryzae pathotype Triticum (MoT) as a suitable heterologous expression system in wheat. Using the recently confirmed wheat stem rust pathogen (Puccina graminis f. sp. tritici) avirulence effector AvrSr50 as a proof-of-concept, we found that delivery of AvrSr50 via MoT could elicit a visible Sr50-dependant cell death phenotype. However, activation of Sr50-mediated cell death correlated with a high transgene copy number and transcript abundance in MoT transformants. This illustrates that MoT can act as an effective heterologous delivery system for fungal effectors from distantly related fungal species, but only when enough transgene copies and/or transcript abundance is achieved.

Magnaporthe oryzae pathotype Triticum (MoT) can act as a heterologous expression system for fungal effectors with high transcript abundance in wheat Cassandra Jensen & Diane G. O. Saunders * Plant pathogens deliver effector proteins to reprogramme a host plants circuitry, supporting their own growth and development, whilst thwarting defence responses. A subset of these effectors are termed avirulence factors (Avr) and can be recognised by corresponding host resistance (R) proteins, creating a strong evolutionary pressure on pathogen Avr effectors that favours their modification/ deletion to evade the immune response. Hence, identifying Avr effectors and tracking their allele frequencies in a population is critical for understanding the loss of host recognition. However, the current systems available to confirm Avr effector function, particularly for obligate biotrophic fungi, remain limited and challenging. Here, we explored the utility of the genetically tractable wheat blast pathogen Magnaporthe oryzae pathotype Triticum (MoT) as a suitable heterologous expression system in wheat. Using the recently confirmed wheat stem rust pathogen (Puccina graminis f. sp. tritici) avirulence effector AvrSr50 as a proof-of-concept, we found that delivery of AvrSr50 via MoT could elicit a visible Sr50-dependant cell death phenotype. However, activation of Sr50-mediated cell death correlated with a high transgene copy number and transcript abundance in MoT transformants. This illustrates that MoT can act as an effective heterologous delivery system for fungal effectors from distantly related fungal species, but only when enough transgene copies and/or transcript abundance is achieved.
Plant pathogens deliver effector proteins to their hosts to reprogram plant defense circuitry and facilitate parasitic colonization 1 . However, in certain plant genotypes a subset of effector proteins termed avirulence factors (Avr) can be recognised by corresponding host resistance (R) proteins. This interaction triggers a hypersensitive immune response and renders the pathogen avirulent, halting its proliferation 2 . However, host recognition exerts a strong evolutionary pressure on pathogen Avr effectors that favours their modification/deletion to circumvent the immune response 3 . In agriculture, this frequently leads to the emergence of virulent pathogen races, compromising R-gene mediated resistance. Identifying Avr effectors and subsequently tracking their allele frequencies in a population is critical for understanding the loss of host recognition. It can also lead to better informed resistance strategies and thereby prolong the longevity of deployed resistance sources 4 . This is particularly important in the case of domesticated crops such as wheat, where resistance breeding is frequently outpaced by pathogen evolution due to the limited genetic base of resistance sources 5 .
To permit widescale monitoring of loss of host recognition, virulence loci must firstly be defined. The genomics era has markedly accelerated the identification of potential virulence loci, with hundreds of candidate effectors being proposed for many agriculturally important fungal pathogens 6,7 . However, subsequent functional confirmation of Vir/Avr activity in the native plant host has remained a largely arduous task, particularly in wheat 8 . As a consequence, the pace of R-gene discovery has now considerably outstripped the confirmation of corresponding virulence loci 9 . This is particularly evident for obligate biotrophic fungal pathogens, that are typically incalcitrant to transformation and/or where suitable surrogate systems are lacking 7,8 , constraining Vir/Avr validation.

Results
The PWL2 promoter effectively drives expression of AvrRmg8 in MoT. To identify a suitable promoter for driving transgene expression of effectors in MoT, we first considered the promoter from the PWL2 gene, which is a well-characterised M. oryzae effector that is known to be highly expressed during infection 19 . We cloned the 330 bp MoT avirulence effector AvrRmg8 20 (genbank accession LC223814) downstream of the 590 bp PWL2 promoter in the pCB-Ppwl2-mcherry-stop vector 21 (Fig. 1a). The resulting pCB-Ppwl2-AvrRmg8-stop vector was then used for protoplast transformation of the MoT strain N06047. This MoT strain was selected as it is known to be virulent on the wheat line S-615 (Rmg8 +) and we found carries a previously uncharacterised virulence allele of AvrRmg8 22 (Supplementary Figure S1). A total of six transformants were confirmed by PCR and three positive transformants selected for further analysis. To establish if the AvrRmg8 gene was expressed sufficiently in the transformants to induce Rmg8-mediated HR, the three selected transformants (PWL8-1, PWL8-3 and PWL8-5) were subjected to infection assays on the second leaf of two to three-week old seedlings of the wheat lines Vuka (Rmg8-) and S-615 (Rmg8 +), alongside the wild type MoT N06047 strain. Two of the three transformants (PWL8-1, PWL8-5) displayed a clear visible HR phenotype 4 days post-inoculation (dpi) on Rmg8 + wheat that was not evident with the wild-type MoT N06047 strain (Fig. 1b). Furthermore, lesions induced by transformants PWL8-1 and PWL8-5 were significantly shorter in length on wheat leaves from the line containing Rmg8, when compared to the wild type strain (Fig. 1c). This analysis illustrates that the expression levels achieved using the PWL2 promoter were sufficient for AvrRmg8 to induce a visible Rmg8-dependent HR phenotype when introduced into the MoT N06047 strain.

A visible Sr50-dependant HR phenotype was not evident in MoT transformants expressing
AvrSr50 under the PWL2 promoter. To determine if MoT could be used as a surrogate system to heterologously express effector proteins in wheat from distantly related fungal species, we selected the previously confirmed avirulence effector from Pgt, AvrSr50 11 . As Magnaporthe cytoplasmic effector proteins also require a signal peptide (SP) for targeting proteins to the host cytoplasm 23,24 , we introduced the coding sequence for the signal peptide from the MoT PWT3 effector 25 downstream of the PWL2 promoter in the pCB-Ppwl2-mcherrystop vector 21 . The 54 bp PWT3 SP was amplified from MoT strain N06047 and cloned in frame with the 333 bp C-terminal effector domain of AvrSr50, without its native signal peptide (Supplementary Figure S2a). Following transformation of MoT strain N06047 with pCB-Ppwl2-PWT3SP-AvrSr50-stop, a total of three transformants were confirmed by PCR and all transformants selected for subsequent analysis. The three transformants (PWLS-2, PWLS-8 and PWLS-11) were used for infection of the second leaf from two to three-week old seedlings of two wheat lines differential for the corresponding R gene, Sr50. At 4 to 5 dpi, there was no visible evidence of an Sr50-dependent HR for the three tested PWL2p::PWT3SP::AvrSr50 transformants (Supplementary Figures.  S2b and S3). Furthermore, the length of the lesions induced by the three transformants were comparable to the wild-type and between the two wheat lines differential for Sr50 (Supplementary Figure. S2c). This indicates that the avirulence properties of AvrSr50 could not be detected for the three transformants analysed when PWL2p::PWT3SP::AvrSr50 was integrated into MoT.    (Table 1). To determine if the high copy number for PS-2 correlated with greater expression levels of AvrSr50 early during the infection process, we inoculated Gabo wheat leaves with a single copy transformant (PS-7), double copy transformant (PS-22), the transformant containing 40 copies (PS-2) and the wild-type MoT N06047 strain. We extracted RNA at 3 dpi from three independent leaves for each MoT strain and conducted RNA-seq analysis. The levels of expression of AvrSr50 were then analysed, illustrating that PS-2 had enhanced expression when compared to the single (PS-7) and double copy transformants (PS-22) (Fig. 2d). However, the increase in expression was not significant, although there was a general trend of increased expression with more PWT3p::PWT3SP::AvrSr50 copies. In contrast, a control To further explore the potential increase in expression of AvrSr50 in the PS-2 transformant, we repeated the infection assays on the wheat line Gabo with the single copy (PS-7) and 40 copy (PS-2) transformants and performed RT-qPCR experiments at 1, 2, 3, 4 and 5 dpi. We found that the 40 copy transformant (PS-2) had consistently higher levels of AvrSr50 expression across all time-points, with a statistically significant increase at The MoT PS-2 transformant contained a single tandem insertion of the transgene near a region of known structural variation. In addition to transgene copy number the location of inserted transgenes can affect transgene expression levels 27 . To assess the location of the transgene in the MoT PS-2 transformant, we conducted long read genome sequencing of the PS-2 transformant using nanopore technology. A total of 6.3 Gbp of total base counts were obtained, equating to approximately 153 × coverage for the predicted 41 Mb genome, with an N50 of 24,624 bp. This was comparable to data generated from conducting genome sequencing of other MoT isolates using similar methods 28 (Supplementary Table S1). Following assembly of the PS-2 genome into 21 contigs (Supplementary Table S2), we performed a BLAST search to identify the location of the multiple PWT3p::PWT3SP::AvrSr50 transgenes using the expression vector as a query. We located a single tandem insertion of PWT3p::PWT3SP::AvrSr50 within contig 0018, which included the entire donor vector concatenated multiple times with a total length of 98,512 bp (Fig. 3a). We also identified 10,352 bp of E. coli genomic DNA that had integrated into the insertion site, immediately downstream of the PWT3p::PWT3SP::AvrSr50 tandem insertion. The insertion event occurred directly downstream of the homolog of the MGG_04257 gene, which is located on chromosome 6 of the rice blast reference genome (Fig. 3a). It is possible that the MGG_04257 promoter could be responsible for driving high levels of expression of the first copy of the tandem insertion. To assess this, RNA-seq reads generated at 3 dpi for PS-2 were de novo assembled and those containing the AvrSr50 coding sequence searched for presence of the 5'UTR originating from MGG_04257. No reads were identified that contained both the AvrSr50 sequence and 5'UTR originating from MGG_04257, indicating that the PWT3 promoter was likely driving AvrSr50 transcription in the PS-2 transformant.

Integration of
The PWT3p::PWT3SP::AvrSr50 insertion site in MoT PS-2 was also within close proximity to a known region of structural rearrangement that has been previously noted when comparing MoT isolate B71 with the rice blast pathogen MG8 assembly of M. oryzae strain 70-15 29,30 . We further confirmed this rearrangement in the PS-2 transformant of MoT strain N06047 through co-linearity analysis. Chromosomes from the M. oryzae MG8 genome assembly were aligned to the MoT PS-2 contigs (Fig. 3b), with the region containing the transgene insertion identified on contig 0018 that aligned to chromosome 6 in the MG8 assembly. In addition, a large portion downstream of the insertion site in PS-2 aligned to chromosome 1 of the M. oryzae MG8 assembly (Fig. 3a,b). Furthermore, alignment of the MoT PS-2 genome assembly to the MoT isolate B71 assembly showed complete colinearity with B71 chromosome 6 (Supplementary Figure. S8). This suggests that the PWT3p::PWT3SP::AvrSr50 transgene in PS-2 inserted into a region of known structural re-arrangement in M. oryzae.  Figure. S9). The lengths of the homology arms (701 bp for the 5' arm, and 404 bp for the 3' arm) corresponded to the maximum length on either side of the cut site before encountering type II restriction sites used in golden gate cloning. A total of three transformants were confirmed by PCR and all transformants selected for further analysis. The three transformants (C04257-1, C04257-2 and C04257-4) were used for infection of leaves from two-week old seedlings of the two wheat lines differential for Sr50. At 5 dpi, there was no visible evidence of Sr50-dependent HR for the three tested PWT3p::PWT3SP::AvrSr50 transformants (Fig. 4a,b and Supplementary Figure. S3). This data illustrates that introduction of a single copy of PWT3p::PWT3SP::AvrSr50 at the MGG_04257 locus was insufficient to elicit an Sr50-dependent HR phenotype.

Introduction of AvrPm3 into MoT did not elicit a Pm3-mediated HR phenotype. It is possible
other Avr/R pairs may require less Avr expression for the elicitation of a visible HR phenotype 14,32 . To test this possibility, we selected the AvrPm3 a2/f2 avirulence effector from Blumeria graminis f. sp tritici for analysis 33 . First, the 321 bp C-terminal effector domain of AvrPm3 a2/f2 was cloned downstream of the PWT3 promoter and PWT3 SP (Fig. 5a). The resulting pCB-PWT3p-PWT3SP-AvrPm3 a2/f2 -stop vector was then transformed www.nature.com/scientificreports/ into MoT strain N06047. A total of four independent transformants were confirmed by PCR and used in infection assays alongside the wild-type MoT strain N06047. Two-week old wheat seedlings of lines Asosan 8*CC (Pm3a +) and Chancillor (Pm3a-) 34 were inoculated and at 5 dpi no evidence of a visible Pm3a-dependant HR phenotype was evident for any of the four PWT3p::PWT3SP:: AvrPm3 a2/f2 transformants ( Fig. 5b and Supplementary FigURE. S10). Disease lesions observed on the Pm3a + line was comparable between the transformants and the wild-type MoT strain (N06047) and to plants lacking Pm3a (Fig. 5c). This indicates that the avirulence properties of AvrPm3 a2/f2 could not be detected when AvrPm3 a2/f2 was integrated into MoT under the PWT3 promoter and PWT3 SP.

Discussion
In this study, we uncovered the potential for MoT to act as an effective heterologous expression and delivery system for avirulence proteins in wheat, when using the Pgt effector AvrSr50 as a proof-of-concept. MoT is particularly attractive as a surrogate system for effector functional analysis in wheat due to the plethora of genetic resources readily available for the closely related rice blast pathogen; a long-standing model organism for studying plant-pathogen interactions 35 . The biotrophic growth phase of MoT also parallels the obligate biotrophic lifestyle of Pgt, with both fungi infecting wheat as their primary host and producing specialised infection structures for effector delivery and nutrient uptake from the plant 36 . However, the broader utility of MoT as a heterologous expression system could potentially be limited in certain countries due to its classification as a quarantine pathogen. MoT first emerged in South America in the 1980's and remained constrained to this region until 2016 when it was identified for the first time in Asia in Bangladesh 37 , and in 2018 in Zambia 38 . Hence, due to its narrow geographic distribution and lack of effective control measures, MoT is classed as a quarantine pathogen in many locations worldwide and its utility as a surrogate expression system should be guided by country-specific regulatory measures.
We also found that MoT may only be useful as an effective surrogate delivery system when sufficient expression of the transgene is obtained; only a single AvrSr50 transformant with very high copy number and transcript abundance, was found to elicit an Sr50-dependent HR. In addition, although several promoter and signal peptide combinations were assessed, only when using those from the MoT PWT3 effector 25 was an Sr50-dependent HR evident. The alternative constitutive promoters (RP27 and TrpC) and PWL2 promoter selected, have been used successfully for ectopic expression of transgenes in Magnaporthe 26,39 . Furthermore, the RP27 promoter was used previously for delivery of AVR1-CO39 40 and PWL2 promoter for delivery of fluorescent protein fusion constructs that accumulated at the biotrophic interfacial complex (BIC) and were subsequently translocated into rice host cells 23 . This suggests that these promoters are suitable for targeting effectors to the rice host cytoplasm. However, their utility in the MoT-wheat pathosystem remains unclear given the absence of a Sr50-mediated HR response in our experiments. It is possible that the low copy number of the resulting transformants and expression levels could have led to an absence of R gene dependent HR. However, we did note one transformant obtained when In fungi, transgene copy number can positively or negatively affect transgene expression. The high levels of AvrSr50 expression in transformant PS-2 correlated with integration of 40 copies of the transgene, all tandemly inserted into the genome. During protoplast transformation, it is common to obtain copy number variants 41 , however such high copy insertions, especially tandem insertions on this scale, are rare. The integration of multicopies of a transgene are often the result of several copies entering a single nucleus and recombination prior to genomic integration into a single locus, frequently leading to higher levels of transgene expression 42 . However, in fungi and plants, integration of multiple transgene copies can also suppress transgene expression with varying degrees of silencing, a phenomenon that has been well studied and termed 'quelling' in Neurospora where it occurs at the post-transcriptional level 43 . In our case, we found no evidence of transgene silencing in multicopy transformants, with visual symptoms of Sr50-mediated HR only clearly evident, when a high number of transgene copies were obtained.
Variation in transgene expression can also be strongly influenced by the insertion site. For transformant PS-2, the tandem multi-copy insertion of AvrSr50 was found in a region equivalent to chromosome 6 in the MoT B71 genome. This region is close to a location of known structural variation in the M. oryzae 70-15 MG8 genome assembly 29 , with transposable elements playing a crucial role in driving plasticity of this region 44 . The tandem insertion of AvrSr50 in transformant PS-2 was also found to disrupt the homolog of the rice blast gene termed MGG_04257, which is known to be upregulated during rice infection, encodes a canonical signal peptide at the N-terminus and is predicted to be a secretory lipase 45 . With two further neighboring genes (MGG_04258 and MGG_04259) displaying characteristics of effector proteins, this location is thus likely transcriptionally active during early time points of infection, which could have contributed to elevating AvrSr50 expression in the PS-2 transformant. However, targeting a single copy of the AvrSr50 transgene to the MGG_04257 locus using CRISPR/ The second leaf of two-week old wheat seedings of the lines Asosan 8*CC (Pm3 +) and Chancillor (Pm3-) were inoculated with conidial suspensions derived from four separate transformants (Pm3-2, Pm3-6, Pm3-11 and Pm3-12) and the wild-type (WT) MoT strain N06047 using the spot inoculation method. Images were taken and lesion lengths measured at 4 days post-inoculation, with three biological replicates conducted (separate leaves).

Scientific Reports
| (2023) 13:108 | https://doi.org/10.1038/s41598-022-27030-z www.nature.com/scientificreports/ Cas9 failed to generate transformants that could elicit Sr50 dependent HR. Therefore, it seems likely that the high number of transgene insertions had the greater influence on AvrSr50 abundance and subsequent HR induction. We found that only with very high AvrSr50 copy number and transcript abundance, can an Sr50-dependent HR be initiated. To enhance expression of AvrSr50 in a reproducible manner, further optimisation of the MoT system is required to consistently elevate the number of transgene copies integrated or enhance expression of the transgene by targeting transgenes to locations of high transcriptional activity and/or through synthetic design. For instance, variation of the terminator sequence, development of synthetic promoters and transcriptional regulators have all been utilised to control and enhance production of gene expression cassettes 42 . If these limitations can be overcome, this would present MoT as an ideal accessible fungal surrogate system for studying the virulence/avirulence function of candidate effectors in future, particularly from intractable obligate biotrophic pathogens that are recalcitrance to genetic transformation.

Materials and methods
Fungal growth and transformation. MoT strains were routinely maintained on complete media 46 with 1.5% agar or cultured in liquid complete media with agitation (150 rpm) at 24 °C for 48 h. Protoplast-mediated transformation of MoT (strain N06047) was performed as previously described 46 , with resulting transformants selected on glufosinate (40 μg/mL). Genomic DNA was extracted from MoT transformants for genotypic analysis and nanopore sequencing using the CTAB method 46 . Copy number analysis of transgenes was performed by iDNA genetics (Norwich, UK) using a TaqMan real-time PCR assay and a probe to detect the bialaphos (BAR) resistance gene. A full list of MoT strains generated is provided in Supplementary Table S3.
MoT infection assays. Seeds from bread wheat (Triticum aestivum) lines Vuka, S-615, Gabo, Gabo + Sr50, Chancillor and Asosan 8*CC were pre-germinated, sown in cell trays and grown in a controlled environment consisting of long-day conditions (16-h light/8-h dark) under a 19 °C/14 °C temperature cycle (day/night). When seedlings reached two-week stage, plants were subjected to MoT inoculation using either detached leaf spot inoculations or spray inoculations 47 . In short, conidial suspensions of 1 × 10 5 spores/mL of each MoT isolate were prepared in 0.25% (v/v) tween gelatine, with 5 μL droplets added for spot inoculations with wicking of droplets after 24 h 47 . Immediately following inoculation, plants were incubated in high humidity and dark conditions for 24 h before being returned to controlled environment conditions. Infection phenotypes were assessed 4-5 dpi, with lesion lengths measured in Image J and assessed using a student's t-test. All experiments were performed in accordance with relevant guidelines and regulations.
The pCB-PWT3p-PWT3SP-AvrPm3 a2/f2 -stop vector was generated using golden gate cloning. Level zero components were generated for the 550 bp PWT3 promoter, PWT3 signal peptide region (54 bp), and 321 bp AvrPm3 a2/f2 coding sequence (without the native signal peptide). The AvrPm3 a2/f2 coding sequence was PCR amplified from genomic DNA of the avirulent Blumeria graminis f. sp tritici strain 96224 33 using primers AvrPm3a_GGF and AvrPm3a_GGR and the transcription terminator of SCD1 from pCB-Ppwl2-mcherry-stop using primers 3SCD1T_GGF and 3SCD1T_GGR. Level 0 constructs were then integrated into the pcb-1532B level 1 vector 49 . Primer sequences are provided in Supplementary Table S4.

CRISPR/Cas9 targeted insertion of AvrSr50 into MoT.
To insert the PWT3p::PWT3SP::AvrSr50 transgene into the MGG_04257 locus, a suitable protospacer adjacent motif (PAM) sequence within the homolog of MGG_04257 in MoT was selected. The MGG_04257 locus was analyzed in the genome sequence of MoT strain Br32 and that generated herein for MoT transformant PS-2, with a PAM sequence at position 803 selected due to the position giving the highest activity and specificity score 50 in MGG_04257. Homologous sequence either side of the PAM sequence (701 bp for the 5' arm, and 404 bp for the 3' arm) were amplified from MoT genomic DNA (strain N06047) using PCR and primers 5′04257_GGF and 3′04257_GGF, and 5′04257_GGR and 3′04257_GGR (Supplementary Table S4). The resulting amplicons were cloned into the universal acceptor plasmid pUAP1 51 . The AvrSr50 coding sequence was domesticated to remove BsaI and BpiI sites using the Q5 site directed mutagenesis kit (New England Biolabs, USA) and primers AvrSr50_Q5_F and AvrSr50_Q5_R. A level zero construct containing the BAR resistance gene was constructed by PCR amplifying Pcb-pPWL2mcherry vector 21 with primers BAR_GGF and BAR_GGR and cloning the amplicon into pUAP1 51 . Finally, level zero constructs containing the 5' and 3' homologous regions of MGG_04257, the PWT3 promoter, fusion of PWT3SP:AvrSr50, SCD1 transcription terminator, and BAR resistance gene were assembled into the level 1 acceptor vector pICH47732 51 .  Figure S9), sgRNAs were synthesized using the EnGen sgRNA synthesis kit (New England Biolabs, USA) following the manufacturer's instructions. Immediately after synthesis, sgRNAs were purified using a T2040 RNA clean-up kit (New England Biolabs, USA) and sgRNAs were then complexed with 6 μg purified Cas9 (New England Biolabs, USA) at a 1:1 molar ratio for 10 min at room temperature. During protoplast transformation, appropriate donor DNA (2 μg in 6 μL) was added alongside pre-complexed Cas9/sgRNA (4 μL). Transformants were confirmed by PCR using primers MGG_04257_F2 and AvrSr50_RTR1 (Supplementary Table S4) that annealed upstream the 5' homology region or the AvrSr50 gene, respectively.
RNA-seq analysis. Total RNA was extracted from MoT infected leaf material collected 3 dpi using the RNeasy Plant Mini Kit (Qiagen, Germany). RNA was sent to GENEWIZ (UK) for cDNA library preparation and sequencing conducted on the Illumina HiSeq 2500 platform. The resulting 150 bp, paired end reads were quality trimmed and filtered using Trimmomatic version 0.39 52 and transcript abundances (TPM values; transcript per million) quantified following pseudoalignment to the MoT reference transcriptome, strain B71 29 , with the AvrSr50 transcript added using Kallisto v 0.46 53 . De novo assembly of transcripts was performed using Trinity v2.11.0 54 and used to search for the AvrSr50 coding sequence with BLASTN 55 .
RT-qPCR analysis of AvrSr50 gene expression. The second leaf from three-week-old Gabo plants were spray inoculated with transformants PS-7 and PS-2 as described above. Leaves were collected at 1-5 dpi and flash frozen in liquid nitrogen followed by total RNA extraction using the RNeasy Plant Mini Kit (Qiagen, Germany). Genomic DNA contamination was removed using a TURBO DNA-free Kit (Ambion, UK) and RNA concentration determined using a Qubit Fluorometer (ThermoFisher, USA). First-strand cDNA was synthesised using SuperScript™ II Reverse Transcriptase (Invitrogen, USA), using 3 μg of RNA, random hexamers and Oligo(dT) primers according to the manufacturer's instructions. RT-qPCR was performed on a LightCycler 480 (Roche, Switzerland) using LightCycler 480 SYBR Green I Master Mix (Roche, Switzerland) following the manufacturer's instructions with each primer at a final concentration of 0.25 μM (Supplementary Table S5). Three technical replicates were prepared per reaction and AvrSr50 expression was compared to actin (MGG_03982) as a reference.
Assembly of the MoT PS-2 genome and collinearity analysis. DNA was extracted from fungal material of MoT transformant PS-2 grown on CM agar plates using the CTAB method 46 with the length of DNA fragments assessed using the Femto Pulse System (Agilent, USA), DNA purity using a Nanodrop (ThermoFisher, USA) and concentration using the Qubit Fluorometer (ThermoFisher, USA). Libraries were constructed using 4 μg of genomic DNA and the 1D genomic DNA by ligation kit (SQK-LSK109; Oxford Nanopore Technologies, UK). The DNA sample was sequenced using the MinION sequencer with FLO-MIN106D R9 flow cells (Oxford Nanopore Technologies, UK) following the manufacturer's instructions. Base calling and demultiplexing was performed using Albacore v.2.3.3 (Oxford Nanopore Technologies, UK) and reads de novo assembled into contigs using Canu v1.8 56 . Contigs derived from the PS-2 assembly were aligned to the MoT genome assembly of strain B71 29 and the rice blast pathogen MG8 genome assembly of M. oryaze strain 70-15 30 using the NUCmer utility of the MUMmer3 software 57 . The coordinate output file was filtered for sequence alignments > 10 kb in length, with > 70% similarity (delta-filter options -l 10,000 -i 70). The MUMmerplot utility was used to generate a dot plot for visualization (options -l and-color). The colour scale was adjusted to display similarity between 80 and 100%.

Data availability
Sequence data that support the findings of this study can be found in the European Nucleotide Archive (ENA) database under the following accession number: PRJEB55233 (https:// www. ebi. ac. uk/ ena/ brows er/ view/ PRJEB 55233).