Use of the volatile trichodiene to reduce Fusarium head blight and trichothecene contamination in wheat.

Fusarium graminearum is the primary cause of Fusarium head blight (FHB), one of the most economically important diseases of wheat worldwide. FHB reduces yield and contaminates grain with the trichothecene mycotoxin deoxynivalenol (DON), which poses a risk to plant, human and animal health. The first committed step in trichothecene biosynthesis is formation of trichodiene (TD). The volatile nature of TD suggests that it could be a useful intra or interspecies signalling molecule, but little is known about the potential signalling role of TD during F. graminearum-wheat interactions. Previous work using a transgenic Trichoderma harzianum strain engineered to emit TD (Th + TRI5) indicated that TD can function as a signal that can modulate pathogen virulence and host plant resistance. Herein, we demonstrate that Th + TRI5 has enhanced biocontrol activity against F. graminearum and reduced DON contamination by 66% and 70% in a moderately resistant and a susceptible cultivar, respectively. While Th + TRI5 volatiles significantly influenced the expression of the pathogenesis-related 1 (PR1) gene, the effect was dependent on cultivar. Th + TRI5 volatiles strongly reduced DON production in F. graminearum plate cultures and downregulated the expression of TRI genes. Finally, we confirm that TD fumigation reduced DON accumulation in a detached wheat head assay. Published 2021. This article is a U.S. Government work and is in the public domain in the USA.Microbial Biotechnology published by Society for Applied Microbiology and John Wiley & Sons Ltd.

Introduction

Fusarium graminearum is the primary cause of Fusarium head blight (FHB), a devastating fungal disease of wheat that causes billions of dollars in annual economic losses worldwide (Wilson et al., 2018). FHB reduces crop yield and contaminates grain with trichothecene mycotoxins that make the grain unsafe for use as food or feed. The trichothecene analog deoxynivalenol (DON) occurs in highest abundance in F. graminearum‐infected wheat and is considered a virulence factor that enables the pathogen to overcome plant defences and spread from infected to uninfected tissues (Desjardins et al., 1996; Audenaert et al., 2014). This connection between DON and virulence indicates that the inhibition of trichothecene biosynthesis would be a practical control strategy of FHB.

The use of antagonistic microorganisms is gaining popularity as an effective, sustainable, and ecofriendly method of pathogen control. Competitive filamentous fungi belonging to the genus Trichoderma can provide effective biocontrol of FHB under at least some field conditions, and can outcompete F. graminearum by reducing growth and sporulation on crop debris (Sarrocco et al., 2019). One isolate of Trichoderma harzianum was reported to have reduced F. graminearum colonization of wheat straw and perithecia development by as much as 96% (Schoneberg et al., 2015). Additionally, the antifungal volatile 6‐pentyl‐2H‐pyran‐2‐one (6PP) produced by some Trichoderma species was shown to reduce F. graminearum DON production on agar medium (Cooney et al., 2001).

The first committed step in trichothecene biosynthesis is cyclization of the primary metabolite farnesyl diphosphate (FPP) to form trichodiene (TD), a volatile sesquiterpene hydrocarbon. This reaction is catalysed by a sesquiterpene synthase enzyme, trichodiene synthase, encoded by the gene TRI5 (McCormick et al., 2011). In Fusarium, TD undergoes four oxygenations to form isotrichotriol, and these reactions are catalysed by a multifunctional oxygenase encoded by the TRI4 gene. Isotrichotriol undergoes spontaneous cyclization to form isotrichodermol, which can then undergo multiple enzyme‐catalysed oxygenations, acetylations, acylations, and/or deacetylations to form the diverse array of trichothecene analogs that have been described (McCormick et al., 2006; McCormick et al., 2011). In Fusarium species, most of the enzyme‐encoding genes required for trichothecene biosynthesis (TRI genes) are located adjacent to one another in the 25 kb TRI cluster. Two transcription factor genes, TRI6 and TRI10, that regulate transcription of TRI genes are also located in the cluster (Seong et al., 2009; Proctor et al., 2018). Despite the synchronized expression of TRI genes, a substantial amount of TD passes through the lipid membranes and escapes into the surrounding environment. TD can easily be detected in the headspace above trichothecene producing fungal cultures, F. graminearum‐infected wheat, and contaminated grain (Jeleń et al., 1997; Eifler et al., 2011; Girotti et al., 2012). Given that the formation of trichothecenes almost certainly exacts a significant energy cost to F. graminearum, loss of TD from the trichothecene pathway into the surrounding air space suggests that TD has a signalling function in addition to its role as an essential intermediate in trichothecene biosynthesis.

Fungal volatiles have been shown to function as antimicrobial compounds, plant growth and/or defence response regulators, and intra or interspecies signals (Li et al., 2016). Previous studies in which exogenous TD was added to liquid cultures of F. graminearum reported numerous transcriptional responses, suggesting that TD may have an intraspecies signalling role (Seong et al., 2009). However, the expression levels of most TRI genes where not significantly affected, and the added TD was incorporated into DON (Seong et al., 2009). Nevertheless, the response of F. graminearum to TD in a liquid culture may differ from its response during growth on a solid substrate or in planta.

Trichoderma harzianum T34 (herein designated Th) has been widely used in basic biocontrol research (de la Cruz and Llobell, 1999), and produces a great variety of cell‐wall degrading enzymes and secondary metabolites on different natural substrates (Delgado‐Jarana et al., 2002; Vizcaíno et al., 2005). In previous research, Th was genetically modified to emit TD. This was done by silencing the ergosterol biosynthetic gene erg1 and heterologously overexpressing TRI5 from a trichothecene producing Trichoderma species (Malmierca et al., 2015b). The resulting TD‐emitting (approximately 20 µg h‐1) strain (herein termed Th + TRI5) induced expression of Botrytis cinerea virulence genes and tomato plant defence‐related genes (Malmierca et al., 2015a; Malmierca et al., 2015b). Among the plant defence‐related genes, the expression of tomato pathogenesis‐related (PR) gene PR1, which is regulated by the salicylic acid (SA) defence signalling pathway, was most strongly induced both by exposure to volatiles emitted from the Th + TRI5 strain and purified TD.

These results suggest that TD can function as a signal that modulates pathogen virulence and host plant resistance. Therefore, we hypothesized that TD emitted by Th + TRI5 could also function as an intra and interspecies signal that could regulate trichothecene biosynthesis in F. graminearum and stimulate wheat defences. Furthermore, we hypothesized that due to these signalling functions, Th + TRI5 would exhibit enhanced biocontrol activity against FHB relative to Th. To test these hypotheses, we compared the antagonistic efficacy of Th and Th + TRI5 against F. graminearum in intact plants of a susceptible and moderately resistant wheat cultivar. Because differences were observed, we also compared the defence responses in wheat seedlings exposed to Th and Th + TRI5 volatiles. To assess the potential regulatory role of Th and Th + TRI5 volatiles on trichothecene biosynthesis, the DON content and expression of TRI genes were analysed in F. graminearum cultures exposed to the volatiles emitted by Th and Th + TRI5. Finally, we evaluated the direct effect of TD fumigation on DON accumulation of three F. graminearum strains inoculated on detached wheat heads.

Results

Th + TRI5 displayed enhanced biocontrol activity on mature wheat plants

To determine if Th + TRI5 had enhanced biocontrol activity against FHB, wheat heads of intact potted plants of FHB susceptible wheat cultivar Norm(S) and moderately resistant cultivar Alsen(MR) were pretreated with sterile water (control), Th, or Th + TRI5 prior to inoculation with F. graminearum (strain Gz3639) or sterile 0.04% Tween20 solution (mock‐inoculation). Volatiles were collected from mock‐inoculated heads 1 day after inoculation and every 7 days thereafter for 3 weeks. TD emission was only detected from heads treated with Th + TRI5, and similar levels of TD were detected throughout the 21‐days time course (Fig. 1). Th + TRI5‐treated heads of both Alsen(MR) and Norm(S) exhibited significantly reduced FHB severity (Fig. 2A and B) and less head weight loss due to the disease when compared to control heads which did not have any pretreatment with a biocontrol strain (Fig. 2C and D). The amount of DON in Th + TRI5‐pretreated heads of Alsen(MR) and Norm(S) inoculated with F. graminearum was 83% and 76% less, respectively compared to control heads (Fig. 2E and F). Th pretreatment significantly reduced disease severity only in Alsen(MR). Th + TRI5 pretreated Alsen(MR) heads had 62% less disease than Th‐pretreated heads (Fig. 2A). On average Th pretreatment reduced DON contamination in comparison to controls, but this difference was not statistically significant. However, Th + TRI5‐pretreated heads had 66% and 70% less DON contamination in Alsen(MR) and Norm(S), respectively, than Th‐pretreated heads.

Fig. 1

Flowering wheat heads of both Alsen(MR) and Norm(S) cultivars were dipped in either sterile water (Con), or a 105 conidia ml‐1 suspension of wild‐type T. harzianum (Th), or T. harzianum overexpressing TRI5 (Th + TRI5) for approximately 5 s. A subset of three heads from each treatment were excised on day 1, 7, 14 and 21 post‐treatment and volatiles were collected for 24 h on a Porapak™ filter via the closed‐loop stripping method. The samples were analysed using GCMS. Trichodiene was only detected from the Th + TRI5 treated heads and was still detected 21 days after the treatment. Graphs display GC chromatographs of volatiles collected from Alsen(MR) heads at 21‐days post‐treatment with Con (A), Th (B) and Th + TRI5 (C). Results were identical for Norm.

Fig. 2

The FHB moderately resistant and susceptible spring wheat cultivars Alsen(MR) and Norm(S), respectively, were grown to anthesis in environmentally controlled growth chambers and pretreated by dipping the heads in water (Con), or a 105 conidia ml‐1 suspension of T. harzianum (Th), or T. harzianum overexpressing TRI5 (Th + TRI5) for approximately 5 s. After 24 h, heads were point inoculated with 10 µL of 105 ml‐1 conidial suspension of F. graminearum (Fg, Gz3639) or sterile medium as a mock‐inoculation control. The per cent disease was estimated over a 21‐day period (A, B). The average reduction in weight due to Fg was determined by subtracting the head weight of Fg‐inoculated heads from the average head weight of mock‐inoculated controls (C, D). Heads were then lyophilized, pulverized, and the average concentration of DON per g dry weight (DW) was estimated using GCMS (E, F). Significant differences are indicated by different letters (analysis of variance (ANOVA) followed by Tukey–Kramer honestly significant difference (HSD); n = 15 of individual heads for disease progression weighted by time and weight loss based on measurements at day 21; n = 5 [with each replicate representing three combined heads] for DON analyses)

To determine whether the enhanced biocontrol activity of Th + TRI5 was due to an increased colonization ability, we compared the ability of Th + TRI5 and Th to colonize wheat heads. We also assessed the relationship between colonization by each T. harzianum strain and abundance of F. graminearum. Both colonization by the T. harzianum strains and spread of F. graminearum were estimated by measuring the biomass of the organisms, and biomass was assessed using quantitative polymerase chain reaction (qPCR) assays (Fig. 3). The qPCR data indicated that colonization by both Trichoderma strains differed between the two wheat cultivars, but within a cultivar colonization by Th + TRI5 and Th was not statistically significantly different. However, in wheat heads of both cultivars inoculated with F. graminearum, colonization by Th + TRI5 was significantly less than Th (Fig. 3A and B). Furthermore, heads treated with Th + TRI5 had the lowest levels of F. graminearum. That is, the treatment with the lowest level of colonization by T. harzianum was the treatment with the lowest level of F. graminearum. The results were similar for both cultivars. These findings indicate that the enhanced biocontrol activity of Th + TRI5 compared to Th did not result from increased colonization of wheat heads by T. harzianum.

Fig. 3

Alsen(MR) and Norm(S) were pretreated by dipping the heads in water (Con), or a 105 conidia/mL suspension of T. harzianum (Th), or T. harzianum overexpressing TRI5 (Th + TRI5) and then inoculated with 10 µL of 105 /mL conidial suspension of F. graminearum (Fg, Gz3639) or sterile medium as a mock‐inoculation control. Disease was allowed to progress for 21 days and then the samples were collected. The relative amounts of T. harzianum (A, B) and F. graminearum (C, D) biomass in each sample were determined from the ratio of fungal DNA to wheat DNA via qPCR using species‐specific primers (Table 1). Significant differences are indicated by different letters (ANOVA followed by Tukey–Kramer HSD; n = 5 [with each replicate representing three combined heads]).

Table 1

qPCR assays used for determination of fungal biomass within plant tissue. All probes were labelled with 5' 6‐FAM and were double quenched, using internal ZEN and 3' Iowa Black fluorescent quenchers. All assays were designed in this work.

Organism	Target gene	Forward primer	Probe	Reverse primer
Fusarium graminearum	translation elongation factor 1‐α	CAGTCACTAACCACCTGTCAAT	AACCCAGGCGTACTTGAAGGAACC	AATGGTGATACCACGCTCAC
Fusarium graminearum	trichothecene 3‐O‐acetyltransferase (TRI101)	GGACTCTGGGATTACGACTTTG	CGAGACTGTGAGACGGCCAATCTTT	ATCAGGCTTCTTGGGCATAAA
Fusarium graminearum	reductase	TGACAGCTTTGGTTGTGTTTG	CGGAAGACTGCTGAGTAACGCCAA	CTTGGCTGGAATGAGTCTGT
Trichoderma harzianum	calmodulin	TTCGGACGVTTTGAGTGCT	CGTACCATCACCGTCCTTGTCCTG	CGGYCGGMAGCAAGTCATTCG
Trichoderma harzianum	endochitinase	CGATCTCAGCTGGATGCTTAT	AATGGTGAGGAGGAAGTGGTAGCC	GCAGCTTGGAGTAGTTATCCTT
Trichoderma harzianum	RNA polymerase II	ATTGGATGGGAGGGATTGATTAG	CATGATCTGCATGACGCCAGAGGA	GCCTTCTGAAGACGATACAGC
Triticum aestivum	actin	CCAAGGCCAACAGAGAGAAA	TGCCCAGCAATGTATGTCGCAATC	GCTGGCATACAAGGACAGAA
Triticum aestivum	phenylalanine ammonia‐lyase (PAL)	GTGTTCTGCGAGGTGATGAA	AAGCACCACCCTGGACAGATTGAA	GTATGAGCTTCCCTCCAAGATG

Impact of Th + TRI5 on wheat seedling defences

We also determined whether Th + TRI5 volatiles were capable of influencing the expression of defence‐related genes in wheat, as previously shown in tomato (Malmierca et al., 2015b). To do this, we used large assay plates to grow wheat seedlings of Alsen(MR) or Norm(S) in proximity to but not in contact with agar media inoculated with sterile water (control), Th, or Th + TRI5 (Fig. 4A). After one week of growth and exposure to volatiles produced by the T. harzianum strains, the aboveground tissues of the seedlings were collected, and relative gene expression of the defence‐related gene PR1 was analysed via qPCR. Expression of the gene in response to the same treatment differed in the two cultivars.

Fig. 4

Alsen(MR) or Norm(S) seeds were germinated and grown on water agar in a large 22 x 22 cm assay plate that contained two sterile V8 agar plates (Con), or two media plate inoculated with Trichoderma harzianum (Th), or T. harzianum overexpressing TRI5 (Th + TRI5) (A). After one week of growth, the seedling leaves were collected, RNA was extracted, and relative gene expression of pathogenesis‐related 1 (PR1) was independently compared between the different treatments in Alsen(MR) (B) or Norm(S) (C). Different letters above bars indicate significant differences (ANOVA followed by Tukey–Kramer HSD; n = 4).

Exposure of Alsen(MR) seedlings to Th + TRI5 volatiles caused an 3.2‐fold increase in PR1 expression compared to the control ( Fig. 4B). Although exposure of Alsen(MR) seedlings to Th volatiles caused an increase in PR1 expression on average compared to the control, the difference was not statistically significant. Exposure of Alsen(MR) seedlings to Th + TRI5 volatiles did cause a significant difference in PR1 expression compared to exposure to Th volatiles. In contrast, exposure of Norm(S) seedling to Th volatiles significantly reduced PR1 expression compared to the control, whereas exposure of Norm(S) seedlings to Th + TRI5 volatiles did not cause a significant difference in PR1 expression compared to the control (Fig. 4C).

We further evaluated the potential of Th + TRI5 to influence expression of wheat defence‐related genes when in direct contact with plant tissue. Seeds from Alsen(MR) and Norm(S) were soaked in a conidial suspension of Th, Th + TRI5 or sterile water for 48 h and then planted in clay pellets and allowed to grow for 2 weeks. The treatments did not have any visual effect on plant growth or development. The stems and roots of each plant were collected separately, and phytohormone and gene expression were evaluated independently for the different tissue types and cultivars. At the timepoint evaluated, the treatments did not result in any significant differences in plant phytohormone levels (data not shown). In contrast to the volatile exposure PR1 expression in aboveground tissues of Th + TRI5 treated seedlings was on average less than the control (Fig. S1A top panels). However, this difference was significant only in Alsen(MR). There were no other statistically significant differences in PR1 expression among treatments for the shoots of either cultivar. On average, PR1 expression was greater in roots from the Th + TRI5 treatment compared to the control and Th treatments (Fig. S1A bottom panels). In Alsen(MR), higher levels of expression in the Th + TRI5 treatment were statistically significant compared to both the control and Th treatments, whereas in Norm(S), the higher levels of PR1 expression in the Th + TRI5 were significant relative to the Th treatment but not the control.

Expression of the phenylalanine ammonia‐lyase gene (PAL), another defence‐related gene, was also evaluated (Fig. S1B). Significant differences in PAL expression were detected only in Alsen(MR). In shoots, PAL expression was significantly less in the Th treatment compared to the control, and in the roots, expression was significantly less in the Th + TRI5 treatment compared to the control. But no other statistically significant differences between treatments were observed.

Th + TRI5 volatiles downregulate F. graminearum biosynthesis of DON

To determine if Th + TRI5 or Th volatiles had a direct influence on trichothecene biosynthesis, F. graminearum strain Gz3639 was grown on a solid trichothecene‐inducing medium (agmatine medium) and exposed to the volatiles in the absence of the plant. Two Petri plates containing either F. graminearum cultures or uninoculated agmatine medium were placed in a larger assay plate along with two plates containing one of the following treatments: (i) Th cultures, (ii) Th + TRI5 cultures, or (iii) or uninoculated growth medium (Fig. 5A). No significant differences in growth of F. graminearum were observed among the treatments (P > 0.05). However, levels of DON produced by F. graminearum were 40% less in the Th + TRI5 treatment compared to the Th and control (no T. harzianum) treatments (Fig. 5B). No significant difference in DON production were observed in the Th treatment compared to the control.

Fig. 5

F. graminearum (Fg, Gz3639) cultures grown on agmatine agar plates were placed into a larger 22 x 22 cm assay plate as shown in diagram (A) with sterile V8 agar plates (Con), or plates inoculated with Trichoderma harzianum (Th), or Th + TRI5. The fungi were exposed to each other’s volatiles within the headspace of the assay plate, but do not physically come into contact with one another. After two weeks of growth, the amount of DON produced on the plates was determined using GCMS analyses and compared between treatments (B). Furthermore, the relative gene expression for TRI6, TRI5 and TRI4 under the different volatile treatments was estimated via qPCR and compared (C). Different letters above SEM bars indicate significant differences (ANOVA followed by Tukey–Kramer HSD; n = 4).

In addition, the Th + TRI5 and Th volatile treatments significantly affected TRI gene expression (Fig. 5C). The expression of TRI6, TRI5, and TRI4 responded similarly to the different treatments. In comparison to the control, the levels of expression of the three TRI genes evaluated were 50% higher in the Th treatment and 50% lower in the Th + TRI5 treatment (Fig. 5C). TRI gene expression was approximately 70% less in the Th + TRI5 treatment compared to the Th treatment. Together, results of analyses of gene expression and DON analyses suggest that TD downregulates TRI gene expression and trichothecene production.

TD fumigation reduced the accumulation of DON

In order to determine if TD was responsible for the reduction in DON production, detached wheat heads were inoculated with one of three F. graminearum strains (Gz3639, 06‐219 or 06‐255), placed in glass jars and fumigated with 1.25 µg cm‐3 purified TD on the day of inoculation and again eight days later (Fig. 6A). Ten days after inoculation, the day on which the heads were collected for DON quantification, fungal hyphae were visible on all florets regardless of volatile treatment. Quantification of DON, via gas chromatography mass spectroscopy (GCMS), revealed that on average less DON accumulated in heads fumigated with TD compared to heads which were exposed to only the carrier solvent acetone (control). This difference in DON accumulation was statistically significant for Gz3639 (40% less DON) and 06‐219 (32% less DON) (Fig. 6B, C), but not for 06‐255 (Fig. 6D).

Fig. 6

Flowering wheat (cultivar Norm(S)) heads were excised, point inoculated with 10 µL of 105 conidia/mL suspension of F. graminearum (Gz3639, 06‐219, 06‐255), placed in glass jars and fumigated with 1.25 µg cm‐3 purified trichodiene (TD) or an acetone solvent control (Con) of equal volume (A). Ten‐day postinoculation, the heads were lyophilized, pulverized, and analysed with GCMS for DON contamination (B–D). Different letters above standard error of the mean (SEM) bars indicate significant differences (t‐test; n = 3).

Discussion

In this study, we demonstrated that a genetically modified T. harzianum strain (Th + TRI5) emitting TD had enhanced biocontrol activity against F. graminearum; wheat plants pretreated with Th + TRI5 had significantly less disease and DON contamination than plants treated with the wild‐type progenitor T. harzianum strain T34 (Th). Assessments of biomass indicate that the enhanced biocontrol activity of Th + TRI5 was not due to a difference in its ability to colonize wheat heads. Th + TRI5 volatiles modestly stimulated wheat defences, as measured by PR1 expression, but in a cultivar‐specific manner. However, Th + TRI5 volatiles strongly reduced DON production in F. graminearum cultures and downregulated the expression of TRI genes. Additionally, fumigation with purified TD confirmed that TD reduced DON accumulation. To the best of our knowledge, this study represents the first to demonstrate that TD, the volatile intermediate of trichothecene biosynthesis, can regulate DON production.

Biosynthetic enzymes are frequently controlled at the transcriptional level by transcription factors whose presence and/or activity can be affected by pathway intermediates or end products (Chubukov et al., 2012). Our findings suggest that TD can function as a negative regulatory signal of TRI gene transcription (Fig. 5C) and are consistent with the cost/benefit model for transcription of metabolic pathway genes to avoid loss of resources by dissipation into the atmosphere. However, unlike genes in many intermediate metabolite‐regulated pathways (Chubukov et al., 2012), TD negatively affected expression of genes both upstream and downstream of TRI5 (Fig. 5C). The downstream gene, TRI4, was not upregulated to maximize conversion of TD into DON. Previous reports showed that the addition of TD to liquid cultures of F. graminearum did not affect TRI gene expression and that the added TD was incorporated into trichothecene biosynthesis (Seong et al., 2009). However, the response of F. graminearum to TD added to a liquid culture is likely very different from the response to TD in the air space surrounding hypha growing on solid media or in planta. Additionally, since the added TD was incorporated into the trichothecene pathway in liquid culture, there was no loss of resources that would call for a negative regulatory function.

Our data further expand the pivotal role of TRI5 in trichothecene biosynthesis to include gene regulation. Disruption of the TRI5 gene not only eliminates the enzymatic cyclization of FPP to TD, the parent compound of all trichothecene analogs, but also disrupts the formation of toxisomes, the specialized subcellular structures made up of colocalized enzymes of the mevalonate‐ and trichothecene biosynthetic pathways on the endoplasmic reticulum and in which trichothecene biosynthesis occurs (Boenisch et al., 2017; Boenisch et al., 2019). However, these two functions can be uncoupled; replacement of wild‐type TRI5 with mutated TRI5 that encodes an enzyme without trichodiene synthase activity restored toxisome formation without restoring trichothecene production (Flynn et al., 2019). Thus, the Tri5 protein itself is required for the development of toxisomes, and the enzymatic product of Tri5 functions both as an essential intermediate in trichothecene biosynthesis and, as shown herein, a volatile regulatory signal.

Volatile signals are frequently concentration dependent (Lee et al., 2016). For example, different concentrations of 6PP can either enhance or inhibit plant growth and health (Jeleń et al., 2014). It is also possible that different individuals of the same species have variable levels of sensitivity to a signal and thus respond differently. This may explain why TD fumigation did not significantly affect DON production by F. graminearum strain 06‐225 (Fig. 6). Nevertheless, further research is needed to determine concentration effects of TD and its mode of perception and action on trichothecene biosynthesis. Furthermore, our data do not address the specificity of the signal. It is possible that other sesquiterpene hydrocarbons can similarly regulate trichothecene biosynthesis.

Various chemicals that inhibit DON biosynthesis have previously been identified. For example, various succinate dehydrogenase inhibitors representing a fungicide class that inhibits fungal respiration decrease DON biosynthesis by inhibiting TRI5 expression and toxisome formation (Xu et al., 2019). Additionally, the fungicide validamycin exhibits dual efficacies in that it can control FHB by inhibiting DON biosynthesis in F. graminearum and inducing host resistance (Li et al., 2019). In contrast to fungicides, TD is an essential component of trichothecene biosynthesis, and trichothecene production is essential for high levels of virulence of F. graminearum on wheat. Therefore, F. graminearum is less likely to overcome TD‐based control strategies than fungicide‐based strategies.

Although previous studies showed that TD and volatiles emitted by Th + TRI5 strongly induced expression of tomato defence genes related to SA (Malmierca et al., 2015a), our results indicate that Th + TRI5 volatiles have a relatively modest effect on the wheat disease response, and this effect was not consistent between cultivars. Th + TRI5 volatiles only significantly induced the expression of PR1 in Alsen(MR) (Fig. 4), and this induction was a modest 3‐fold induction compared to the 1000‐fold induction observed in tomato (Malmierca et al., 2015b). However, PR1 expression was similarly induced 2 to 3‐fold in response to validamycin treatment in wheat tissues, and this response was assumed sufficient to potentially enhance resistance of wheat to F. graminearum (Li et al., 2019). Th volatiles suppressed PR1 expression in Norm(S). This suggests that trichoacorenol, which was found to be produced by Th (Malmierca et al., 2015b), or other unidentified Th volatiles may also influence wheat defences in a cultivar‐specific manner, and in some cases this effect could impede rather than enhance the wheat defence response. T. harzianum strains with volatile blends that do not enhance TRI gene expression or produce other inhibitory volatiles may provide even greater control against FHB. Nevertheless, since the induction of defence‐related gene expression was only observed in one cultivar and the enhancement of biocontrol activity was observed in both cultivars, it is unlikely that the enhancement was due to stimulated host defences alone.

PR1 expression declined in wheat seedling shoots treated with Th or Th + TRI5 and was significantly less in Alsen(MR) treated with Th + TRI5 (Fig. S1A). These changes in PR1 expression could result from direct fungal exposure and/or in alterations in levels of nonvolatile metabolites. For examples, Th + TRI5 produces lower levels of ergosterol and higher levels of squalene than Th (Malmierca et al., 2015b), and both ergosterol and squalene can trigger transcription of plant defence‐related genes in a concentration‐dependent manner (Klemptner et al., 2014; Lindo‐Yugueros et al., 2020). The ratio of ergosterol to squalene can also influence T. harzianum colonization of plants (Lindo‐Yugueros et al., 2020). Therefore, differences in PR1 expression that occurred in roots versus shoots in response to Th + TRI5 may result from different metabolite concentrations, which in turn could result from differences in the extent of colonization of the roots and shoots by T. harzianum. Nevertheless, when applied directly to the wheat heads, Th and Th + TRI5 did not exhibit significant difference in colonization as assessed by biomass (Fig. 3). The extent of colonization was not directly related to the efficacy of a strain as an FHB biocontrol agent. Although, Th + TRI5 colonized F. graminearum‐infected wheat heads less than Th, Th + TRI5 was the more effective FHB biocontrol agent (Figs 2 and 3). This further supports the notion that metabolites produced by Th + TRI5 enhance the efficacy of FHB control.

Genetically engineering microbes to enhance their beneficial uses in agriculture has been gaining acceptance. Numerous examples of genetically modified biocontrol agents with improved activity against phytopathogens have been reported (Farrar et al., 2014; Arora et al., 2020; Hanlon and Sewalt, 2020). For example, overexpression of TRI5 in T. brevicompactum increased the antimicrobial activity of the strain via incorporation of TD into the antimicrobial compound trichodermin (Tijerino et al., 2011). The biocontrol activity of a Pseudomonas fluorescens strain against take‐all disease in wheat was enhanced by introducing a seven‐gene operon for synthesis of the antibiotic phenazine‐1‐carboxylic acid (Yang et al., 2017). Another strain of P. fluorescens was genetically modified to act as a biocontrol agent and biofertilizer with biological nitrogen fixation activity (Zhou et al., 2014; Jing et al., 2020). Genetically modified microbes (Pivot Bio PROVEN™) developed by California‐based Pivot Bio are currently commercially available in the US and are being used to increase corn productivity and decrease fertilizer application. Thus, the use of genetically modified microorganisms in agriculture may be more easily accepted than transgenic crops.

Despite the extraordinary potential for the use of TD to enhance the FHB biocontrol activity of T. harzianum, it is essential to better understand the mechanism(s) by which TD enhances the biocontrol. Our data do not exclude the possibility that TD‐emitting T. harzianum strains can stimulate trichothecene biosynthesis under certain conditions. Furthermore, environmental conditions can have a significant effect on emission and dispersal of volatiles (Misztal et al., 2018), and this could affect the efficacy of TD‐enhanced biocontrol activity. Additionally, the intraspecies signalling potential of TD may result in non‐target effects that could have negatives consequences on crop resistance to other pests or pathogens. Nevertheless, this study has identified a novel method to regulate trichothecene biosynthesis and our approach has potential as an effective control strategy for both FHB epidemics and DON contamination in wheat. T. harzianum can be applied as a foliar spray during anthesis as demonstrated herein, but could also be applied as a seed coating (Ferrigo et al., 2014; Rocha et al., 2019) as demonstrated in other studies.

Experimental procedures

Wheat head blight assays evaluating the biocontrol activity of Th + TRI5

FHB disease progression assays were performed using two hard red spring wheat cultivars, the moderately FHB resistant cultivar Alsen(MR) and the susceptible cultivar Norm(S). Both cultivars were grown in climate‐controlled growth chambers programmed for 23°C day/20°C night, 500 µmol m‐2/s‐2 photosynthetic photon flux density 14 h photoperiod and 50–60% relative humidity. Five wheat seeds were planted in each 20 x 15‐cm pot containing 2.5 l of SunGrow Horticulture potting mix (Agawam, MA, USA). Plants were fertilized two‐week postgermination and then every other week thereafter with a 500 ml solution containing 0.5 g l‐1 of Peter’s 20‐20‐20 (Grace‐Sierra Horticultural Products, Milpitas, CA, USA) until inoculation.

The biocontrol activity of wild‐type T. harzianum strain Th (T34) and the genetically modified strain Th + TRI5 (E20‐tri5.7), silenced in expression of erg1 and heterologously expressing TRI5 which emits approximately 20 µg h‐1 TD (Malmierca et al., 2015a; Malmierca et al., 2015b), were compared. Detailed differences in metabolite profiles between the two strains have previously been described (Malmierca et al., 2015a; Malmierca et al., 2015b).

When the plants started flowering, they were pretreated by dipping heads for 5 seconds in either a 105 conidia/mL suspension of Th, or Th + TRI5 or sterile water (control). After 24 hours, 15 wheat heads from each cultivar per treatment (Th, Th + TRI5 or control) were point inoculated with 10 µl of a 105 conidia ml‐1 suspension of F. graminearum (Gz3639) in 0.04% Tween 20 (Thermo Fisher Scientific, Waltham, MA, USA) or a sterile 0.04% Tween 20 solution (mock‐inoculated). The pots were separated by treatment into different chambers to prevent any cross‐contamination of volatiles with the other treatments. The experiment was repeated altering the chambers used for the designated treatments. Individual heads were bagged for 3 days to increase humidity and ensure disease establishment. The number of diseased florets was scored on day 7 and then every three days until 21‐day postinoculation. The percentage of disease was determined as the number of florets displaying early whitening and necrosis divided by the total number of florets on the head. After 21 days, average yield loss was estimated by subtracting the excised head weight of the F. graminearum diseased heads from the average weight of the mock‐inoculated controls. Three heads from each treatment were then combined as one sample (biological replicate), lyophilized, ground and analysed for estimation of fungal biomass and DON contamination. DON was extracted from approximately 1 g of tissue using 86:14 acetonitrile: water and quantified using GCMS following previously reported methods (Vaughan et al., 2020).

Additionally, a subset of 3 mock‐inoculated heads were excised at 1, 7, 14, and 21 days post pretreatment with Th, or Th + TRI5 and volatiles were collected on a PorapakTM filter for 24 hours using a closed‐loop stripping method (Tholl et al., 2006). The presence of TD was evaluated by GCMS analysis and confirmed using purified a standard of TD.

Estimation of fungal biomass

To determine the amount of fungal biomass in diseased wheat heads, genomic DNA was extracted from the lyophilized and pulverized tissue using the ZR Fungal/Bacterial DNA Miniprep Kit (Zymo Research, Boston, MD, USA). DNA quantity and quality were evaluated using a spectrophotometer (NanoDrop 2000; Thermofisher Scientific, Waltham, MA, USA). If the samples’ 260/230 ratios were below 1.6, the DNA was further purified using the Genomic DNA Clean and Concentrate Kit (Zymo Research). The biomass of F. graminearum, Th, or Th + TRI5 within the plant tissue was estimated as the relative quantity of fungal DNA to wheat DNA, using three sets of species‐specific primers (Table 1).

Samples and assays were distributed across an integrated fluidic circuit and mixed in pairwise fashion using the Juno instrument (Fluidigm, San Francisco, CA, USA). Technical triplicates were run for each sample‐assay pair. Template DNA concentration was standardized to 40 ng DNA µl‐1. The PCR components followed protocol PN 100‐7222 C1 (Fluidigm), including the use of SsoFast EvaGreen Supermix (Bio‐Rad, Hercules, CA, USA). All qPCR assays (Table 1) were designed during this work, and all probes were labelled with 5' 6‐FAM and were double quenched, using internal ZEN and 3' Iowa Black fluorescent quenchers (Integrated DNA Technologies). The specificity of the assays was validated by testing the primers and probes on different samples which included the individual species alone and in combination with one another (Fig. S2). Thermocycling and fluorescence detection were performed with the Biomark HD (Fluidigm); thermocycling conditions were as defined by protocol PN 100‐7222 C1 (Fluidigm).

The threshold cycle (Ct) values were used as calculated by the default parameters of the Fluidigm software. Per sample‐assay pair, the mean Ct was calculated. The geometric mean was calculated across assays for each organism (i.e. 3 assays for F. graminearum, 3 assays for T. harzianum, 2 assays for T. aestivum). The biomass of each fungus within the plant tissue was estimated as a ratio of fungal DNA relative to wheat DNA, as: biomass = 2ΔCt, where Ct was the geometric mean across assays for each organism.

Effect of Th + TRI5 volatiles on wheat seedlings

To evaluate the potential of interspecies signalling, wheat seedlings of Alsen(MR) and Norm(S) were exposed to Th, or Th + TRI5 volatiles. Two wheat seedlings were allowed to grow on water agar in square 22 cm x 22 cm assay plates in proximity but not in contact with two V8 juice agar plates inoculated with sterile water (control), Th, or Th + TRI5 as shown in Fig. 4A. Seeds were surface sterilized and then germinated for 48 h in water prior to placement in the assay plates for volatile exposure. The seedlings were grown in the assay plates for one week during which they were continuously exposed to the volatiles released from the different treatments (control, Th, or Th + TRI5). For each treatment and cultivar, four large assay plates were set up and designated as independent biological replicates. On day 7, the seedling shoots from a single plate were collected together, lyophilized, pulverized and the combined tissue was used for expression analysis of the defence‐related gene PR1 via qPCR.

Effect of Th + TRI5 application on wheat seedlings

Seeds of Norm(S) and Alsen(MR) were germinated (48 h) in either sterile water (control), or a 1x105 conidia/mL suspension of Th or Th + Tri5. A single treated seedling was planted in each ‘cone‐tainer’™ (Stuewe and Sons Inc., Tangent, OR, USA) containing clay pellets which allow for easy separation of root tissue from the substrate (Vaughan et al., 2011; Vaughan et al., 2015). For each treatment and wheat cultivar, four biological replicates were prepared. The seedlings were placed in a climate‐controlled growth chamber programmed to 23°C day/20°C night, 500 µmol m‐2 s‐2 photosynthetic photon flux density, 14 h photoperiod and 50–60% relative humidity. The clay pellets were watered daily. After 2 weeks of growth the shoots and roots of each plant were separated. The individual tissues were lyophilized, pulverized and separated for analysis of phytohormones and relative expression of plant defence genes.

Quantification of phytohormones

Phytohormones were analysed via chemical ionization gas chromatography/mass spectrometry (CI‐GC/MS) profiling (Schmelz et al., 2004). Jasmonic acid (JA) and salicylic acid (SA) were extracted from approximately 100 mg lyophilized ground tissue with 1‐propanol and methylene chloride, (Sigma‐Aldrich, St. Louis, MO, USA), derivatized with trimethylsilyldiazomethane in hexanes (Sigma‐Aldrich) for 30 min, and then collected by vapour‐phase extraction (Schmelz et al., 2004). The CI‐GC/MS analysis was performed on a 7890B GC coupled to a 5977A MS (Agilent Technologies, Germantown, MD, USA) run in CI mode with isobutane as the ionization gas. Compounds were separated on an Agilent Technologies DB‐35MS column (30 m, 0.25 mm, 0.25 μm) held at 70°C for 1 min after injection, followed by a programmed temperature gradient of 15°C/min to 300°C where it was held for 7 min. Helium was used as the carrier gas with a 0.7 ml min‐1 flow. Identification and quantification of JA and SA were based on a deuterated internal standard (CDN Isotopes, Pointe‐Claire, Quebec, Canada) spiked into each sample prior to extraction. Quantity estimates of total JA (trans + cis) and SA were based on corresponding deuterated internal standards (Vaughan et al., 2014).

Fumigation of F. graminearum cultures

Two agmatine agar plates inoculated with F. graminearum (Gz3639) and two V8 juice agar plates inoculated with sterile water (control), Th, or Th + TRI5 were placed into larger, square 22 x 22 cm assay plates (Fig. 5A). The fungi did not come into physical contact with one another but were exposed to the volatiles emitted into the headspace of the assay plate. For each of two experimental replicates performed, three assay plates (biological replicates) were set up per treatment. The assay plates were covered and kept in the dark at ambient laboratory conditions. Every other day throughout the first week, radial growth measurements were recorded from the centre of the inoculum plug to the extreme edge of the F. graminearum fungal mycelia development following two perpendicular lines on each plate. The average of the two radial measurements from each media plate within a single assay plate was used as a biological replicate resulting in a total of three biological replicates per treatment. Following 14 days of growth and volatile exposure, the Gz3639 mycelia were scraped from the surface of the agar plate and used for relative TRI gene expression analyses. DON was extracted from the agar media using 86:14 acetonitrile: water and quantified using GCMS following previously reported methods (Vaughan et al., 2020).

Transcriptional analyses

RNA was isolated from approximately 60–80 mg lyophilized, ground tissue using a combination of the Trizol reagent (Life Technologies, Carlsbad, CA, USA) and the RNeasy Plant Mini Kit (Qiagen, Germantown, MD, USA) extraction methods following standard manufacturer protocols with an additional DNase step using the RNase Free DNase Kit (Qiagen, Germantown, MD). First‐strand cDNA was generated using a Superscript II Kit (Invitrogen, Carlsbad, CA, USA). The efficiency of each primer pair was estimated prior to experiments using a mixture of cDNA from the samples. qPCR was prepared in a total of 20 µl volume with 10 µl of Bio‐Rad SybrGreen Supermix (Bio‐Rad Laboratories), 300 nM of each primer and 1 µl of cDNA. All sample reactions were performed in triplicate. PCR was performed on a Bio‐Rad CFX96 RealTime System (Bio‐Rad Laboratories). The thermocycling programme consisted of an initial denaturation at 98°C for 2 min, 40 cycles each of 98°C denaturation for 15 s and 60°C annealing/elongation for 1 min and a final dissociation curve from 65 to 95°C.

Gene‐specific oligonucleotides and their PCR efficiencies are listed in Table 2. The C t values of PR1 and PAL were normalized to the wheat housekeeping gene GAPDH and the Ct values of TRI genes were normalized to the Fusarium endogenous control β‐tubulin (β‐tub) encoding gene. The amount of each gene transcript was calculated relative to its corresponding average for control samples in each experiment. Transcript fold‐changes were calculated using the 2−ΔΔCt method (Schmittgen & Livak 2008) using CFX Manager software, version 3.0 (Bio‐Rad).

Table 2

qPCR assays used to evaluate gene expression.

Organism	Target gene	Forward primer	Reverse primer	PCR efficiency
Triticum aestivum	pathogenesis‐related 1 (PR1)	CGTCTTCATCACCTGCAACTA	CAAACATAAACACACGCACGTA	103%
Triticum aestivum	phenylalanine ammonia‐lyase (PAL)	TTGATGAAGCCGAAGCAGGACC	ATGGGGGTGCCTTGGAAGTTGC	108%
Triticum aestivum	GAPDH	TTGCTCTGAACGACCATTTC	GACACCATCCACATTTATTCTTC	101%
Fusarium graminearum	TRI6	TAACCACATCGTCGGGACTG	GCCGACTTCTTGCAGGTCTT	101%
Fusarium graminearum	TRI5	TCTATGGCCCAAGGACCTGT	ACGCTCATCGTCGAATTCCT	105%
Fusarium graminearum	TRI4	CGCGAGCTTACGACATTGAG	GAACTTCGCCAAGGTGTTCC	95%
Fusarium graminearum	β‐tubulin (β‐tub)	TCCAGGGTTTCCAAATCACC	GGAACGACGGAGAAAGTTGC	95%

TD fumigation of detached wheat heads inoculated with F. graminearum

FHB susceptible hard red spring wheat cultivar Norm(S) was grown in a sunlit greenhouse supplemented with high‐pressure sodium lights to maintain a 14 h day cycle. Greenhouse temperatures ranged from 25 to 28°C during the day and 17‐20°C at night. At anthesis the inflorescences (heads) were excised with approximately 10 cm of stem. A single central floret was inoculated on each head by injecting 10 µl suspension of 105 F. graminearum conidia per ml. Three 15‐acetyldeoxynivalenol (15‐ADON) producing F. graminearum strains, Gz3639, 06‐219 and 06‐255 belonging to the North American 1 population were independently used as inoculum (Kelly and Ward, 2018). Three heads inoculated with the same strain were placed into a small beaker of water and then sealed in a 4 l (4000 cm3) glass jar which contained a 5 mm diameter vent tube in the cap (Fig. 6A). This vent tube allowed for some gas exchange and dissipation of volatiles from the jar. Therefore, fumigation treatments were applied twice, on the day of inoculation (day 1) and again on day 7. Fumigation was performed by adding 50 μl of purified TD at a concentration of 100 µg μl‐1 dissolved in acetone (100%). This resulted in a 1.25 µg cm‐3 TD fumigation treatment that likely gradually declined with time. TD was isolated from yeast extract‐peptone‐dextrose liquid (YEPD) cultures of F. sporotrichioides Tri4‐ mutant strain F15 and purified as previously reported (Hohn et al., 1995). Three biological replicates represented by three individual jars each containing three inoculated heads were set up per volatile exposure treatment. Ten days following F. graminearum inoculations, the stems were removed, and the wheat heads were lyophilized, pulverized, and the tissue was separated for DON quantification as previously described (Vaughan et al., 2020).

Statistical analyses

All analyses were conducted using statistical software JMP (15.0.0). Differences between means were determined using t‐test and analysis of variance (ANOVA) followed by Tukey–Kramer honestly significant difference (HSD). Data presented as percentages were arc sign transformed and biomass ratios were square root transformed prior to analyses. Details of individual analyses and resulting P values are described and reported within the results and figure legends.

Fig. S1. Alsen(MR) or Norm(S) seeds were soaked in water or a solution of Trichoderma harzianum (Th), or T. harzianum overexpressing TRI5 (Th + TRI5) for 48 h and then planted in clay pellets. The shoots and roots of the two‐week‐old seedlings were collected separately and the relative gene expression of pathogenesis‐related 1 (PR1, A) and phenylalanine ammonia‐lyase (PAL, B) were compared between treatments. Statistical comparisons were conducted independently for the different tissue types and cultivars. Different letters above bars indicate significant differences (ANOVA followed by Tukey–Kramer HSD; n = 4)

Click here for additional data file.

Fig. S2. Heat map of Ct values generated by Fluidigm software depicting species specificity of qPCR assays listed in Table 1. The left‐hand panel shows the three different species evaluated. Each assay was tested in triplicate. The top panel lists the different samples and which species DNA were predicted to be in each based on sample preparation. T. aestivum assays amplified samples from both Alsen and Norm varieties. T. harzianum assays similarly amplified both Th and Th + TRI5 strains. The 12 samples on the right side of the map which contain only T. harzianum or F. graminearum DNA, depict a dilution serious of the DNA isolated from axenic culture of the fungal species. Colour key for Ct values is depicted in the right panel. Black designates (>35) no detectible amplification. No detectible amplification was shown for all assays which did not contain the intended species of the assay

Click here for additional data file.