The identification and deletion of the polyketide synthase‐nonribosomal peptide synthase gene responsible for the production of the phytotoxic triticone A/B in the wheat fungal pathogen Pyrenophora tritici‐repentis

Summary The economically important necrotrophic fungal pathogen, Pyrenophora tritici‐repentis (Ptr), causes tan spot of wheat, a disease typified by foliar necrosis and chlorosis. The culture filtrate of an Australian Ptr isolate, M4, possesses phytotoxic activity and plant bioassay guided discovery led to the purification of necrosis inducing toxins called triticone A and B. High‐resolution LC–MS/MS analysis of the culture filtrate identified an additional 37 triticone‐like compounds. The biosynthetic gene cluster responsible for triticone production (the Ttc cluster) was identified and deletion of TtcA, a hybrid polyketide synthase (PKS)‐nonribosomal peptide synthase (NRPS), abolished production of all triticones. The pathogenicity of mutant (ttcA) strains was not visibly affected in our assays. We hypothesize that triticones possess general antimicrobial activity important for competition in multi‐microbial environments.

Pyrenophora tritici-repentis (Ptr) is a dothideomycete fungal pathogen and the causal agent of tan spot, an economically important and world-wide foliar disease of wheat. Three host specific toxins, the proteinaceous effectors ToxA and ToxB and the semi-characterized small molecule effector ToxC have been described for Ptr (Effertz et al., 2002;Ciuffetti et al., 2010;Lamari and Strelkov, 2010). In addition, two classes of specialized metabolites, anthraquinones and triticones, have been reported as non-specific phytotoxins Hallock et al., 1993;Kachlicki and Wakulinski, 2002;Bouras and Strelkov, 2008). Anthraquinones are characterized by three fused aromatic rings and harbour a variety of substituents. They are produced by many endophytic fungi, have a broad range of biological activities and their biosynthesis has been well studied (Thomas, 2001;Gessler et al., 2013). Four anthraquinones, emodin, catenarin, helminthosporin, islandicin and their hydroxylated forms have been detected in culture filtrates, infected leaves and grains of wheat (Kachlicki and Wakulinski, 2002;Wakuli nski et al., 2003). Catenarin is thought to be the cause of red smudge, a red discolouration of grains predominantly found on infected durum wheat (Bouras and Strelkov, 2008).
The triticones, also known as spirostaphylotrichins, are characterized by a spirocyclic γ-lactam core structure. Triticones A to F were first purified from Ptr culture filtrates but have since been found, along with an additional 18 other triticone compounds, in 5 other ascomycete fungi, Staphylotrichum coccosporum, Curvularia pallescens, Pyrenophora seminiperda, Cochliobolus lunatus and Bipolaris spp. Tamm, 1989a,b, 1990;Hallock et al., 1993;Abraham et al., 1995a;de Almeida et al., 2017;Wang et al., 2018). Triticone A and B are known to be phytotoxic, producing yellowishbrown lesions following leaf puncture assays on a range of hosts including wheat. Triticone C and D were described as weakly active and triticones E and F as inactive (Hallock et al., 1993). It was postulated that triticone A alters host photosynthetic processes via inhibition of thiol-containing enzymes (Kenfield et al., 1988). Biosynthesis of triticones in Ptr has not been pursued and the biosynthetic pathway remains unknown. Recently, a biosynthetic gene cluster (Cpa) containing a PKS-NRPS hybrid gene (CpaA) from the endophytic fungus C. pallescens (DSM62482) was shown to produce a class of structurally similar compounds called curvupallides (Yokoyama et al., 2017). Yokoyama (2017) proposed that spirostaphylotrichin C (triticone A) is also derived from CpaA.
Our objective was to purify and characterize small molecules which might contribute to virulence in culture filtrates of Ptr. In the current study, we present the plant bioassay-led purification of triticones A and B and identify a further 37 triticone-like compounds. The biosynthetic gene cluster was identified and deletion of the core PKS-NRPS biosynthetic gene prevented production. However, virulence was not obviously affected.

Purification of a small molecule causing necrosis on wheat leaves
Infiltration of the culture filtrate of Ptr M4, a ToxA and ToxC producing Ptr isolate, into leaves of the ToxA insensitive wheat cultivar Eagle Rock produced a beige ovoid region of necrosis which did not extend to the entire infiltration zone (Fig. 1A). When the culture filtrate was dialysed using a <3.5 kDa cut-off, necrotic activity was heavily reduced (Fig. 1A). A combination of dispersive solid phase extraction (dSPE) and iterative highperformance liquid chromatography (HPLC) attached to a diode array detector (DAD) was used to fractionate the culture filtrate and all fractions were analysed via LC-Q-MS and infiltrated into wheat (see document Supporting Information, Fig. S1/S2 for details). The fraction exhibiting highest necrotic activity had two major components both with an m/z of 278. Final purification led to the isolation of compound 1 as an off white powder with an [M + H] + of 278 (Fig. 1B), which produced the beige ovoid necrotic symptom (Fig. 1C).

Identification of phytotoxin triticone A/B (1) and C/D (2) in Ptr
High-resolution LC-MS/MS analysis of 1 revealed a precursor ion mass of m/z 278.1023. The MS/MS spectra of 1 ( Fig. 2A) contained fragment ions in common with 37 other features in the crude culture filtrate. To determine whether a similar core structure was present in these features a second, but inactive compound (2), containing a similar fragmentation pattern (Fig. 2B) was purified. Compound 2 was purified using the same protocol as 1 (Supporting Information Fig. S3) and had a measured accurate mass of m/z 280.1178.
The accurate mass precursor ions for 1 and 2 were consistent with the calculated mass of triticone A or B and triticone C or D, isomer pairs of small molecule phytotoxins previously characterized in Ptr (Fig. 2C) (Hallock et al., 1993) (Hallock et al., 1993) confirmed the identification of 1 as a mix of the interconvertible triticone A and B (Supporting Information Table S1) and 2 as either triticone C or D (Hallock et al., 1993)    and Hallock (1993). D.   A. Infiltration of 1 in to 3 wheat cultivars of differing tan spot susceptibility at 2, 0.5, 0.1 mg ml −1 and 0 mg ml −1 . B. Infiltration in to 1) Canola, 2) Arabidopsis, 3) Nicotiana, 4) Brachypodium, 5) Ryegrass and 6) Barley at i) 2.0 mg ml −1 , ii) 0.5 mg ml −1 , iii) 0.1 mg ml −1 and iv) 0 mg ml −1 (white bar represents 1 cm). C. Development of necrosis symptom produced by 1 at 8, 24, 48 and 72 h post infiltration. D. Infiltration of 1 at 2.0 mg ml −1 in to Eagle Rock and placed in i) 4 days of 12/12 h day/night cycle, ii) 4 days of 24 h night and iii) 4 days of 24 h night followed by 2 days of 12/12 h day/night cycle.
[Color figure can be viewed at wileyonlinelibrary.com] moderately susceptible (MRMS) and moderately resistant (MR) to tan spot (Eagle Rock, Magenta and Mace respectively) (DPIRD, 2018). Necrotic symptoms were comparable across cultivars, were smaller when infiltrated with 0.5 mg ml −1 and no activity was seen at 0.1 mg ml −1 . Additional plant species were infiltrated to test host specificity including the monocots Brachypodium distachyon, ryegrass and barley and the dicots canola, Arabidopsis thaliana and tobacco. All showed comparable responses to wheat (Fig. 3B), exhibiting necrotic symptoms at 2 and 0.5 mg ml −1 with minimal to no activity at 0.1 mg ml −1 .

Compound 1 (triticone A/B) induced light-sensitive necrosis on wheat leaves
Compound 1 was infiltrated at 2 mg ml −1 into wheat and incubated under three light settings to investigate the light dependence of necrotic activity. Immediately upon infiltration a translucent zone was visible, which, under 12/12-h light/dark cycle, developed to the highly localized beige- Identification of the PKS-NRPS hybrid gene in M4 genome responsible for triticone production Triticones are spirocyclic γ-lactams that contain both ketone and amino acid-derived amide moieties which suggested that they are synthesized by enzymes encoded by hybrid PKS-NRPS biosynthetic gene clusters. Genome sequence analysis using antiSMASH (FungiSMASH v5.0.0) identified three biosynthetic gene clusters containing both PKS and NRPS genes in M4 (Supporting Information Table S4). Two of the three clusters were discounted as candidates based on the domain structure of the NRPS gene; one (PWO07329) contains a multi-modular domain structure indicative of multiple amide bond formations and the second has a single incomplete NRPS module (PWO09899) as well as a truncated PKS gene (PWO09906). The third gene cluster, henceforth referred to as Ttc, comprised of 20 genes (TtcA-L), is 52 kb in length and contains PtrM4_14012  Information Table S5 and Fig. S9.).
In order to confirm the role of TtcA in the production of triticones, the gene was deleted via homologous recombination of a knock-out construct. Transformed strains were screened via PCR and qPCR ( Fig. 4B and C) and three strains (ttcA-1, ttcA-2 and ttcA-3) were selected for further study. Two strains (E-1 and E-2) containing an ectopic (non-homologous) insertion of the gene deletion cassette were also selected to confirm integration of the cassette does not induce side effects. The growth and morphology of the ttcA mutants and ectopic strains were not significantly affected (Supporting Information Fig. S10). LC-Q-MS analysis of the three ttcA mutant strains indicated that all 38 triticone compounds were undetectable in the crude culture filtrates and infiltration into wheat produced no necrotic activity ( Fig. 5A and B). The ectopic and wild-type (WT) strains produced culture filtrates with identical levels of triticones and necrotic activity.

Triticones do not contribute to pathogenicity during wheat infection
Seedlings of two tan spot susceptible (S-Eagle Rock and Machete) and one susceptible/very susceptible (SVS-Yitpi) wheat cultivars were infected with spores of the wild type, ectopic and three ttcA mutant strains to determine the effect of the deletion on pathogenicity. The mutants retained full pathogenicity and displayed necrotic lesions that were comparable with the wild type and ectopic strains (Fig. 5C).
To determine whether triticone A/B was produced by Ptr during infection, Eagle Rock leaves infected with the WT strain were sampled to quantitatively analyse triticone production in planta at 1, 3, 5 and 7-days post-infection. First and second leaves were harvested, extracted and analysed via high-resolution LC-MS and the detection limit of triticone A/B in plant material was determined to be 5 μg g −1 of lyophilized whole leaf material (Fig. 5D). No triticone candidates were detected in extracts of WT infected leaves at any time point post-infection (Fig. 5D). The expression of TtcA was quantified using qPCR on first and second leaves of WT infected Eagle Rock at 1, 3, 5 and 7-days post-infection. No expression of TtcA was detected within the first 5 days. Low-level expression of TtcA (0.02 AE 0.00 gene expression relative to actin) was detected at 7 days post-infection (Supporting Information Fig. S11).

Triticone A/B exhibits low-level antimicrobial activity
Growth inhibition of fungal and bacterial species by triticone A/B was tested using a disk diffusion method. Disks were infused with 20, 5 and 1 μg triticone A/B and placed on plates of Pichia pastoris Alternaria infectoria, Bipolaris sorokiniana, Epicoccum nigrum and Parastagonospora nodorum. No zone of inhibition was observed. Disks were also applied to cultures of the Gram-negative bacteria, Escherichia coli and Pseudomonas syringae and Gram-positive bacteria, Bacillus subtilis and Rhodococcus erythropolis. The Gram-negative bacteria were not inhibited, but zones of inhibition of 6.3 mm and 13 mm were observed for B. subtilis and R. erythropolis, respectively at 20 μg disk −1 (Supporting Information Fig. S12).

Discussion
It has been well established that small molecular weight compounds can contribute to phytopathogenicity in fungal systems (Möbius and Hertweck, 2009;Horbach et al., 2011;Pusztahelyi et al., 2015) and surveys of the Ptr genome by Manning (2013) and Moolhuijzen (2018), elicited in excess of 20 biosynthetic gene clusters, none of which have been characterized for product or function.
Given the prevalence of secondary metabolites involved in pathogenicity, it is not unreasonable to surmise one or more products from these biosynthetic gene clusters may contribute to virulence.

The triticones of Ptr
In pioneering studies, three enantiomeric pairs of triticones were isolated and identified from culture filtrates of Ptr ( Fig. 2A) Hallock et al., 1993). Triticones A/B, enantiomers at the C-2 position, were found to be highly phytotoxic. Triticones C/D, also enantiomers at C-2 but also reduced at C-4, were found to be mildly phytotoxic. Triticones E/F, with a hydrated methylene group at C-5 0 , were reported to be nonphytotoxic.
We found 38 triticone compounds in Ptr cultures, a quantity exceeding the total number of previously characterized triticones in all fungi. Triticones are known to be unstable during purification and it is possible that some of these compounds are products of rearrangement or degradation Sandmeier and Tamm, 1989a,b;Hallock et al., 1993).  suggested that the presence of the triticone A/B enantiomer pair may be the product of a retro-aldol-type reaction, where cleavage of the carbon bond between C-1 and C-2 occurs forming an achiral aldehyde which readily condenses again forming a racemic mixture. Alternatively, all stereoisomers of triticones are potentially produced to varying degrees. For triticone C/D, with three chiral carbons, up to eight stereoisomers are possible; however, 12 compounds were detected with the molecular formula of triticone C/D in the M4 culture filtrate. We found that the number of detectable triticone compounds was highest in strains that produced the most triticone A/B. This is consistent with these other compounds being pathway precursors and minor products or artefacts of rearrangement and degradation.

Triticones are produced by the Ttc cluster
The biosynthetic gene cluster Ttc was investigated to determine whether it was responsible for synthesis of the triticone class of compounds. The Ttc cluster is similar to Cpa in C. pallescens with strong homology to 14 of the 18 genes within Cpa. The Cpa cluster synthesizes the curvupallides, fused pyrone/γ-lactam compounds, structurally similar to the spirocyclic γ-lactam triticones (Yokoyama et al., 2017). Indeed, seven triticones have previously been characterized in C. pallescens (Abraham et al., 1995a;Abraham et al., 1995b) and the product of CpaA, a PKS-NRPS hybrid, has been implicated as the precursor in triticone biosynthesis (Yokoyama et al., 2017). As Ttc contains all the proposed genes required for curvupallide synthesis, it is possible one or several of the triticone candidates identified via high-resolution LC-MS/MS analysis may belong to the structurally similar curvupallide class. However, lack of a curvupallide standard prevented identification of curvupallides in Ptr culture filtrate. No phytotoxic activity was observed for the curvupallides when tested against cress seedlings at 100 μg ml −1 (Abraham et al., 1995b).

Bioactivity of triticones
Triticone A/B gave visible necrotic symptoms on a range of monocot and dicot hosts at 0.5 mg ml −1 (1.8 × 10 −3 M) upon infiltration of attached leaves. Similar results were observed by Masi (2014) using leaf puncture assays on detached leaves on a variety of plant hosts, but Hallock et al. (1993) noted necrotic symptoms as low as 10 −5 M. The greater sensitivity found by Hallock et al. (1993) may be due to their use of constant light in contrast to the 12 h photoperiod employed here and by Masi (2014). Fig. 3 indicates that the phytotoxic activity was light dependent.
Six of the Ptr isolates tested produced a high concentration of triticones in the culture filtrate. Given the metabolic cost of producing specialized metabolites, a survival advantage is presumed to be necessary for the retention of the biosynthetic capacity (Calvo et al., 2002;Pickens et al., 2011). Triticones have been investigated for a wide range of biological activities including anti-influenza (Wang et al., 2018), anti-leishmanial (de Almeida et al., 2017 and phytotoxicity Hallock et al., 1993;Masi et al., 2014); however, their role in the pathogenicity of Ptr on wheat has never been directly tested. Deletion of TtcA in Ptr prevented the production of the all triticone compounds but did not noticeably impact the degree of virulence of M4 on wheat. This was consistent with our observation of very low expression of TtcA at 7 days infection. Triticones A/B inhibited growth of B. subtilis, a Gram-positive non-pathogenic bacterium commonly found in soil and the human gut and R. erythropolis, a beneficial plant bacterium. Although further investigation is necessary, triticones may have a role to play in multi-microbial environments, where competition for resources such as nutrients and space exists. Biosafety regulations mean that virulence assays must be conducted on seedlings in growth chambers which we recognize does not adequately reflect the field environment. It is possible that triticones have a role in pathogenicity on older plants or on pathogen survival between growing seasons. In addition, whilst wheat is the major host of Ptr, it has been detected on rye, oat, barley and a range of non-cereal grasses (Ali and Francl, 2003;Abdullah et al., 2017) and may contribute to virulence on alternate host species.

Methods and materials
Fungal material and culture filtrate production Ptr isolate M4, collected in 2009 from Meckering, Western Australia was grown and the culture filtrate was prepared as described previously with some modification (Moffat et al., 2014). Briefly, M4 was grown on V8-PDA plates at room temperature under standard laboratory lighting for 7 days. A 1.4 cm × 2.8 cm area of mycelial growth was excised and homogenized using a mortar and pestle. This was transferred to 480 ml of Fries 3 liquid media in a 2 l Erlenmeyer flask. Liquid cultures were placed in darkness at 27 C, 100 rpm for 3 days and then placed in darkness for 18 days of stationary growth at room temperature. The culture medium was filtered through three layers of milk filter paper (Daviesway, Heidelberg Heights, Victoria, Australia) followed by centrifugation of filtrate at 10 000 rpm for 1 h and then vacuum filtered through a Whatman No. 1 filter paper.
Dialysis of the crude culture filtrate was carried for 48 h using SnakeSkin pleated dialysis tubing (BioRad, Hercules, CA, USA) with a molecular weight cut off of <3.5 kDa in 0.2 M phosphate buffered water (pH 7.4).

Toxin purification
The culture filtrate was first extracted using dispersive solid phase extraction (dSPE). For every 1 l of culture filtrate, 10 g of Bondesil-C18, 40 μm sorbent (Agilent, Santa Clara, CA, USA) was added to the filtrate and mixed for 1 h at room temperature.
The filtrate and sorbent were vacuum filtered in to a 60 ml reservoir containing a frit to retain only the sorbent and then washed with 2 × 50 ml of water. Adsorbed molecules from the filtrate were eluted from the sorbent by increasing the percentage of methanol in the aqueous elution solvent in 10% increments. Fractions were dried via removal of methanol by rotary evaporation and remaining water was removed by lyophilization. Fractions were resuspended in a known volume of water for bioassay.
The active fraction was further purified using HPLC (Agilent 1200 LC system) coupled to a diode array detector (DAD) for manual fractionation. A quadrupole mass spectrometer with an ESI source was additionally used for profiling fractions. An Agilent C18 column (PrepHT, 150 mm × 21.2 mm × 5 μm) was used for purification and a Phenomenex C18 column (Kinetix, 10 mm × 2.1 mm ID × 2.6 μm; Torrance, CA, USA) was used for profiling fractions. Purification was carried out using 95:5 acetonitrile:water as mobile phase B and 5:95 acetonitrile:water as mobile phase A. The purification gradient started at 20% B and increased to 40% B over 30 min at a flow rate of 6 ml min −1 . Fractions were manually collected by monitoring 220, 254, 280 and 420 nm on the DAD and infiltrated in to wheat leaves to trace active fractions. For profiling, mobile phase B was 95:5 acetonitrile: water, 0.1% formic acid and mobile phase A was 5:95 acetonitrile:water, 0.1% formic acid. Chromatographic conditions held the column at 0% B for 0.08 min, then increased to 100% over 10 min, %B was then held for a further 3.3 min and then decreased to 0% B over 0.3 min and finally maintained for 2.3 min (16 min total). The flow rate was 0.75 ml min −1 and the injection volume was 3 μl. The mass spectrometer was run as full scan (m/z 100-1000) in positive ion mode. Drying gas temperature and flow were 350 C and 12 l min −1 , respectively and the nebulizer pressure was set at 35 psig. Capillary voltage was 3 kV.

Toxin characterization
Crude culture filtrate was analysed with high-resolution mass spectrometry using the Q Exactive LC-MS (ThermoFisher, Waltham, MA, USA). Briefly, culture filtrate was diluted with water and analysed using a Zorbax Eclipse XDB-C18 column (RRHD, 100 mm × 2.1 mm ID × 1.8 μm; Agilent). Mobile phase B was acetonitrile with 0.1% formic acid and mobile phase A was water with 0.1% formic acid. Chromatographic conditions held the column at 5% B for 1 min, then increased to 30% over 5 min, %B was further increased to 99% over 5 min and held for 2 min. %B was dropped back down to 5% over 0.1 min and maintained for 1 min (total time 15 min). The injection volume was 2 μl. The mass spectrometer was equipped with an electrospray ionization source (ESI) and set to acquire full scan and data dependant MS/MS at a resolution of 35 000. Scan range for full scan collection was 80-200 m/z and MS/MS spectra were collected at 10, 20 and 40 eV. Spray voltage was set to 3 kV and capillary temperature was 350 C, auxiliary gas flow and temperature were set to 10 l min −1 and 350 C respectively. All mass spectral data analysis was carried out using AnalyzerPro v5.7.0.176. NMR analysis was carried out using a Bruker Avance IIIHD 500 MHz. NMR data analysis was carried out using TopSpin v4.0.1.

Phytotoxicity and pathogenicity assays
Infiltration of crude filtrate, processed fungal extracts and infection assays were carried out on wheat (Triticum aestivum) cultivars Eagle Rock, Machete, Mace, Magenta and Yitpi, on canola (Brassica napus), Arabidopsis thaliana, Nicotiana benthamiana, Brachypodium distachyon, Lolium perenne and barley (Hordeum vulgare). Seeds were sown in Grade 2 vermiculite (the Perlite and Vermiculite Factory, Jandakot, WA, Australia) in seed trays at 20 C, 12 h day/night cycle in a controlled growth chamber. Wheat, B. distachyon, ryegrass and barley were infiltrated upon full expansion of the second leaf. Canola, A. thaliana and tobacco were infiltrated upon full expansion of the third leaf. Infiltration was carried out using a 1 ml needleless syringe and the zone was marked with a permanent marker pen. Plants were infiltrated in triplicate and symptoms were evaluated 4 days post infiltration. Symptom intensity was measured as the length of necrotic region compared to the entire infiltration zone and presented as a percentage.
Plant infection assays were performed as previously described (Moffat et al., 2014). Briefly, 2-week old seedlings were sprayed with a 3000 conidia/mL solution in 0.25% gelatin. Disease symptoms were visually assessed and photographed at 7 days post inoculation. Infection experiment was performed twice with three biological replicates.

Light sensitivity
Infiltrated Eagle Rock seedlings were exposed to three different light settings. The first was 12/12 h light/dark cycle and the second was 24 h dark then scored at day 4. The third setting was a 24 h dark period for 4 days followed by 2 days of 12/12 h light/dark cycle and scored immediately after. Triticone A/B was infiltrated in to second leaves of Eagle Rock seedlings at 2 mg ml −1 and placed in their respective lighting scenario.

Biosynthetic gene cluster prediction and BLASTP search
The M4 genome (NQIK01000001.1) was mined for biosynthetic gene clusters using antiSMASH v5.0 fungal version (Blin et al., 2019). Under 'analysis options', ClusterFinder and Use ClusterFinder algorithm for biosynthetic gene cluster (BGC) border prediction were selected. The minimum cluster size in CDS and minimum number of biosyntheticrelated PFAM domains were both set to 5. Minimum ClusterFinder probability was 60%. Under 'extra features', all options were selected (KnownClusterBlast, ClusterBlast, SubClusterBlast, smCoG analysis, ActiveSiteFinder, Whole-genome PFAM analysis and CASSIS). Protein sequences encoded by the Ttc cluster were used for BLASTP query against the NCBI GenBank database (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Generation of TtcA gene deletion mutant
The TtcA knockout (KO) cassette was constructed using a fusion PCR approach as previously described (Moffat et al., 2014) in order to create a ttcA mutant strain. Briefly, a phleomycin resistance cassette was amplified from the pAN8-1 plasmid and fused with flanking regions upstream and downstream of TtcA to generate a knock out construct (Supporting Information Fig. S13) (Supporting Information Table S6 for primers). Transformation was carried out using a polyethylene glycol (PEG)-mediated protoplast method as described by Moffat et al. (2014). Screening of three transformants (ttcA-1, ttcA-2 and ttcA-3) using PCR was performed to confirm the absence of the target gene (Fig. 4B). Correct insertion of the gene deletion cassette via PCR at the intended homologous site was verified for all mutant strains (ttcA-1, ttcA-2 and ttcA-3) and nonhomologous insertion was confirmed for the ectopic strains (E-1 and E-2) (Fig. 4C). Confirmation of single-copy integration for mutants and ectopics was determined by quantitative PCR (qPCR) (Fig. 4D) as described by Moffat et al., 2014. TtcA expression analysis during infection RNA was extracted from wheat leaf samples infected with WT Ptr at 1, 3, 5 and 7 days post-inoculation in triplicate. The second leaf was collected and immediately snap-frozen. Whole leaf material was ground in liquid nitrogen and subjected to RNA extraction using TRIzol reagent and Direct-zol RNA miniprep kit (Zymo Research, Irvine, CA, USA). Extracted RNA was treated with TURBO DNA-free kit (Thermo Fischer Scientific, Waltham, MA, USA) and subjected to reverse transcription using the iScript cDNA synthesis kit (Bioline, London, UK) as per manufacturers' protocols. Quantitative RT-PCR analysis was performed using Quantitect SYBR Green RT-PCR kit (Qiagen, Valencia, CA, USA) as described by Rybak et al. (2017). Primers TritA_F and TritA_R were used to amplify the 150 bp region of TtcA (Supporting Information Table S6 for primers). To normalize the gene expression, primer set Act1F4 and Act1R4 were used to amplify the housekeeping gene actin. Samples were analysed with three biological replicates and two-technical replicates.
In planta detection of triticones using LC-MS Wheat seedlings (Eagle Rock) with fully expanded third leaves were inoculated with WT, ectopic and ttcA strains. Plant infection assays were performed as described above but with three biological replicates. A lower conidial concentration of 1000 conidia/mL was used for ttcA-3 due to insufficient spore production. The second leaf of infected seedlings was collected, snap frozen and freeze-dried at 1, 3, 5 and 7 days post inoculation. Leaves were ground to a fine powder and 10 mg was extracted with 2 × 500 μl methanol and 1 × 500 μl 50:50 water:methanol. Combined extracts were split to represent 5 mg extractions, solvents were removed and reconstituted in 100 μl 95:5 water:acetonitrile and analysed via high-resolution LC-MS as described in Toxin Characterization.

Antimicrobial assays
Antimicrobial activity of triticone A/B was assessed in fungal and bacterial species using a disk diffusion method (Bauer et al., 1966). Antifungal activity for triticone A/B was tested against Pichia pastoris, Alternaria infectoria, Bipolaris sorokiniana, Epicoccum nigrum and Parastagonospora nodorum. Briefly, for filamentous fungi 50 μl of spore suspensions were each spread on V8-PDA plates and 10 μl of triticone A/B solution at 2.0, 0.5 and 0.1 mg ml −1 was applied to 5 mm filter paper disks placed on the inoculated plates (equivalent to 20 μg/disc). Plates were grown at room temperature for 36 to 72 h and plates were observed for any growth inhibition. P. pastoris was cultured to an optical density (OD) of approximately 0.6-0.8 and 100 μl was spread on YPD plates and grown at 28 C.
Antibacterial activity was similarly tested against two Gram-negative bacteria Escherichia coli and Pseudomonas syringae and two Gram-positive bacteria Bacillus subtilis and Rhodoccus erythropolis strain 66b. Briefly, bacterial suspensions were grown to approximately 0.6-0.8 OD at 37 C and 100 μl was spread and dried over LB agar plates. Filter paper disks were infused with 10 μl of triticone A/B solution at 2, 0.5 and 0.1 mg ml −1 and the inhibition zone was measured in millimetres, 24 h after inoculation.