A defined medium was developed in which to monitor deoxynivalenol (DON) accumulation, fungal growth and expression of genes involved in trichothecene biosynthesis (designated Tri genes). In liquid culture, DON accumulated shortly after maximum expression of Tri6 and coincident with expression of Tri5. This was generally 96 h after inoculation. The effects of sublethal concentrations of the fungicides azoxystrobin, trifloxystrobin, kresoxim-methyl and tebuconazole on biosynthesis of the trichothecene DON by Fusarium culmorum were studied using this medium. The strobilurin fungicides trifloxystrobin and azoxystrobin significantly reduced the accumulation of DON in culture medium at a range of concentrations. Kresoxim-methyl, also a strobilurin, and tebuconazole, a triazole, did not significantly reduce the accumulation of DON, although levels were lower than those in nonamended cultures. Trifloxystrobin significantly reduced the accumulation of DON when added to cultures before initiation of trichothecene biosynthesis. RT-PCR assays of the expression of Tri6 and Tri5 genes indicated that trifloxystrobin acted by inhibiting the initiation of trichothecene biosynthesis.
Fusarium head blight is a potentially serious disease of wheat that occurs sporadically across much of northern Europe, the Americas and parts of Asia. Severe outbreaks can cause both qualitative and quantitative losses to growers (Snijders & Perkowski, 1990). Although a number of fungal species can be isolated from affected ears, Fusarium culmorum and F. graminearum are thought to be the principal causal agents (Mesterhazy, 1995; Parry et al., 1995). While F. graminearum predominates in the USA and China, F. culmorum is more often isolated in cooler maritime climates such as that of the UK (Parry et al., 1995).
Fusarium disease also poses a threat to consumer health, as both F. culmorum and F. graminearum produce mycotoxins, specifically trichothecenes and zearalenols (Mirocha et al., 1995). Of the trichothecenes produced, the most abundant and potentially toxic are deoxynivalenol (DON); nivalenol (NIV); and acetylated derivates (Visconti et al., 1986).
The biosynthetic pathway to DON has been thoroughly investigated in F. graminearum, and a number of genes involved in the pathway have been characterized (e.g. Hohn & Beremand, 1989; Hohn et al., 1995). One of these, Tri6, encodes a zinc finger-type DNA-binding protein (Proctor et al., 1995), and has been shown to be involved in the transcriptional regulation of trichothecene biosynthesis, playing an important role in the regulation of expression of Tri5, the gene that encodes the first step in trichothecene biosynthesis.
Little, however, is known about the regulation of trichothecene biosynthesis, although host-derived factors (resistance–susceptibility, growth stage, plant tissue); environmental factors (temperature, water potential, pH); pesticides (fungicides, herbicides, insecticides); fungal competition and nutritional factors (carbon level, nitrogen source) have all been reported as influencing trichothecene production in Fusarium species (D’Mello et al., 1998; Magan et al., 2002). Increases in the level of DON in grain following application of the strobilurin fungicide azoxystrobin were reported from a number of field and glasshouse studies (Ellner & Schröer, 2000; Simpson et al., 2001). Other authors found that sublethal concentrations of certain fungicides may increase mycotoxin production by Fusarium species (D’Mello et al., 1998; Magan et al., 2002). Azoxystrobin is highly active against Microdochium nivale and saprophytic fungi found on wheat heads (Bertelsen et al., 1999; Siranidou & Buchenauer, 2001), but has limited activity against Fusarium species (Simpson et al., 2001). The increase in DON content of grain was found to relate to an increase in the amount of F. culmorum present in ears treated with azoxystrobin (Edwards et al., 2001). No evidence was found to indicate that the fungicide induced the fungus to produce more DON.
Studies in vitro have been conducted on the effect of some fungicides on toxin production in F. graminearum (Matthies et al., 1999). Some sublethal concentrations of certain fungicides, including tebuconazole, resulted in increased accumulation of trichothecene mycotoxin. These studies involved culturing isolates in complex media. However, batch-to-batch variation in DON production has been experienced when using complex media in this laboratory. The objective of the present study was to identify a defined medium for optimal DON production in vitro by F. culmorum, and to investigate, in such a medium, the effect of some fungicides on DON production and on the regulation of the trichothecene biosynthetic pathway by studying Tri5 and Tri6 gene expression.
Materials and methods
Fusarium culmorum strain Fu5 was obtained from the John Innes Centre facultative pathogen collection, and the media and conditions used for maintenance and subculturing of isolates were as described previously (Doohan et al., 1998). For the production of conidia, strains were grown in 25% V8 vegetable juice (Campbell Grocery Products, King's Lynn, Norfolk, UK) for 10 days at 22–25°C with agitation (0·7 g). Cultures were filtered through Miracloth tissue (Calbiochem, La Jolla, CA, USA) and the conidia collected by centrifugation (3000 g, 10 min). These were then resuspended in 10% glycerol at 106 mL−1 and stored at −70°C until required.
In order to identify a medium that could ensure high and reliable levels of DON production, preliminary assays were conducted by filter-sterilizing liquid cultures containing different nutrients (1·65 g L−1 ammonium sulphate, 0·75 g L−1 urea or 2·125 g L−1 sodium nitrate), glucose (10 g L−1) and 20 mL L−1 nitrogen-free modified Vogel's medium (Simpson et al., 1998). Autoclaved potato dextrose broth (PDB, Difco, West Molesey, Surrey, UK) was also included as a complex medium reference. All the experiments were carried out in ventilated sterile plastic 50 mL tubes containing 20 mL medium and, in order to allow optimal oxygenation, tube caps were holed with a hot dissecting needle and a square of ethanol-sterilized Miracloth placed under them.
Modified Vogel's medium amended with ammonium sulphate (0·75 g L−1) and glucose (10 g L−1) was used to study the temporal regulation of DON production in batch cultures. In initial studies, cultures were harvested 24, 48 72, 96, 120 and 144 h after inoculation (hai). In all subsequent experiments, the first harvest was made at 72 hai.
After preliminary experiments, the following fungicide concentrations were used: tebuconazole (Folicur, 250 g a.i. L−1, Bayer, Bury St Edmunds, Suffolk, UK) 5, 20, 50 and 100 µg a.i. L−1; azoxystrobin (Amistar, 250 g. a.i. L−1, Syngenta, Bracknell, Berks, UK) 10, 100, 1000, 10 000 and 50 000 µg a.i. L−1; kresoxim-methyl (Stroby WG, 500 g a.i. L−1, BASF plc, Cheadle Hulme, UK) 0·1, 1, 10, 100 and 1000 µg a.i. L−1; and trifloxystrobin (Twist, 125 g a.i. L−1, Bayer) 2, 5, 10, 20 and 50 µg a.i. L−1. An equivalent volume (20 µL) of water was added to control cultures. Tubes were inoculated with 20 µL of spore suspension (106F. culmorum spores mL−1) and incubated at room temperature (22–25°C) with shaking (0·7 g) for 144 h. The effects of fungicides were examined in two experiments, each with three replicate cultures per treatment. At the end of incubation, fungal pellets were harvested and their weight recorded as pressed weight (mycelium press-dried between Whatman filter paper) in the first experiment and dry weight in the second. DON concentration in the medium was then related to fungus weight. In order to allow comparison between the two experiments, mean specific DON concentration values for each treatment were expressed as percentages of mean control values. Conditions for the time-course assays were similar, except that no fungicidal agents were added.
Addition of trifloxystrobin at different times after inoculation
Trifloxystrobin was added at a final concentration of 10 µg a.i. L−1 to tubes previously inoculated with 20 µL F. culmorum spore suspension (106F. culmorum spores mL−1) at 0, 72, 96 or 120 hai. Tubes were harvested after 72, 96, 120 and 144 hai. Tubes without any fungicide addition served as the control.
Determination of DON concentrations
DON concentrations in the culture filtrates were determined using the Ridascreen fast DON ELISA kit (R-Biopharm, Rhône Ltd, Glasgow, UK) according to the manufacturer's instructions.
Reverse transcription PCR of Tri5 and Tri6 genes
RNA was extracted using TRI REAGENT (Sigma-Aldrich, Poole, Dorset, UK) according to Sigma Technical Bulletin MB-205. Briefly, mycelial pads were homogenized in 1 mL TRI REAGENT with glass beads (Sigma) and centrifuged at 15 000 g for 10 min at 4°C. The supernatant was transferred into fresh tubes and allowed to stand for 5 min at 22–25°C. RNA was pelleted by adding 200 µL isopropanol, shaking vigorously for 15 s, incubating for 10 min at 22–25°C, and centrifuging at 15 000 g for 15 min at 4°C. The upper aqueous phase was transferred into new tubes, 500 µL isopropanol added, mixed and incubated for 10 min at 22–25°C, and centrifuged at 15 000 g for 15 min at 4°C. After washing in 1 mL 70% ethanol, pellets were briefly air-dried for 5–10 min, dissolved in 100 µL diethylpyrocarbonate (DEPC)-treated water, and the RNA quantified by spectrophotometry at 260/280 nm. RNA was treated with DNase I (Amersham Pharmacia Biotech, Little Chalfont, Bucks, UK), extracted with phenol-chloroform and precipitated in isopropanol before being resuspended in DEPC-treated water prior to RT-PCR.
Quantitative RT-PCR analysis of Tri5 and Tri6
Primers used for amplification of Tri5 transcript were Tr5inR2 (5′-GAATTGAGGGTAGTCATCAGATCC-3′) and Tr5F (5′-AGCGACTACAGGCTTCCCTC-3′). Primers used for amplification of Tri6 transcript were T6FSp (5′-CATGCCAAGGACTTTGTCCC-3′) and T6EndR (5′-GTGTATCCGCCTATAGTGAT-3′). Primers used for amplification of β-tubulin transcript were Spantub2 (5′-ACCGGTCAGTGCGGTAAC-3′) and B531R (5′-GACTGACCGAAAACGAAGTTG-3′) (Doohan et al., 1999). Total RNA (1 µg) was reverse-transcribed after annealing with 25 pmol of both Tr5R and B531R for Tri5 and with 25 pmol of both T6EndR and B531R for Tri6, as described previously (Doohan et al., 1999). An initial mixture (12 µL) containing RNA and downstream primer in DEPC-treated water was overlaid with mineral oil (Sigma) and incubated for 10 min at 70°C. Negative control reactions contained no RNA. These tubes were then chilled on ice for 1 min, adjusted to a volume of 20 µL containing 100 U Superscript reverse transcriptase (Gibco BRL, Paisley, Scotland, UK), 20 U ribonuclease inhibitor (Amersham Pharmacia Biotech), 10 mm dithiothreitol (DTT), 500 µm each of dATP, dCTP, dGTP and dTTP, 10 mm Tris-HCl pH 8·3, 3 mm MgCl2, 50 mm KCl and 100 µg mL−1 gelatine, and incubated for 30 min at 42°C. Reactions were terminated by incubating at 99°C for 5 min, then diluted to 100 µL with DEPC-treated water and stored at −20°C. PCR analysis was performed using 10 µL 1 : 10-diluted cDNA solution and 10 pmol each of forward and reverse primers, as described below.
As a control, Tri5 and Tri6 transcript abundance was expressed as a fraction of β-tubulin mRNA abundance, also measured by RT-PCR from the same RNA sample. β-tubulin was previously shown to be constitutively expressed over the timeframe under investigation (Doohan et al., 1999).
PCR reactions (50 µL) contained 10 pmol of each forward and reverse primer and 10 µL diluted cDNA in the presence of 100 µmol each of the four dNTPs, 0·8 U Taq polymerase (Boehringer Mannheim, Roche Diagnostics Ltd, Lewes, East Sussex, UK) and 5 µL cresol red dye as supplied in the manufacturer's reaction buffer. DEPC-treated water was added to give a total volume of 50 µL. Reaction conditions were: 95°C for 3 min; (95°C for 30 s, 60°C for 30 s, 72°C for 1 min) × 30; 72°C for 1 min. Amplification products were separated by electrophoresis through 2% agarose gel, stained with ethidium bromide and viewed under UV light, and analysed using molecular analyst software (Bio-Rad Laboratories Ltd, Hemel Hempstead, Herts, UK). The ratio of Tri5 and Tri6 to β-tubulin was used as an index of the relative Tri5 and Tri6 expression in samples (Tri mRNA per β-tubulin mRNA).
Data were analysed using genstat (Payne et al., 1993). The effects of treatments on growth, toxin production and gene expression were determined by anova. Where relevant, data were logarithmically transformed before analysis because of the nonindependence of mean and variance. Effects were considered significant where P = 0·05.
Effect of nutrients on growth and DON production
Fungal growth and production of DON were compared in PDB and modified Vogel's medium containing different N sources. Growth in PDB was significantly less (P = 0·05) than in the other three media, and was significantly greater in Vogel's medium with ammonium sulphate (P = 0·05) than the other media (Table 1). The accumulation of DON was similar in PDB and Vogel's medium with sodium nitrate. Significantly higher levels of DON accumulated in Vogel's medium with urea (1858 µg L−1) than in PDB, but levels were greatest in Vogel's medium with ammonium sulphate (28 600 µg L−1). In preliminary experiments, in which media were autoclaved rather than filter-sterilized, accumulation was generally significantly lower in Vogel's medium with urea. This was apparently caused by complexing of urea with the glucose, and resulted in the medium changing colour.
Table 1. Effect of different nutrients on growth and deoxynivalenol (DON) production of Fusarium culmorum
Dry weight (mg)
DON (µg L−1)
DON/dry weight (mg g−1)
Potato dextrose broth
LSD (P < 0·05)
The amount of DON accumulated per unit of fungus was least when PDB or Vogel's medium with sodium nitrate was used (Table 1). The level was significantly higher in Vogel's medium amended with urea (P < 0·05), and highest in Vogel's medium with ammonium sulphate (717·4 mg g−1). This medium was used in all subsequent experiments.
Timecourse of expression of Tri5 and Tri6 and accumulation of DON in Vogel's medium amended with ammonium sulphate
The abundance of Tri6 transcript rose between 72 and 96 hai and remained at high levels at 120 hai, before decreasing at 144 hai (Fig. 1). The level of Tri5 transcript was very low at 72 hai. Abundance of Tri5 transcript increased between 96 and 120 hai and decreased thereafter. The concentration of DON in the culture medium was low at both 72 and 96 hai, and increased markedly between 96 and 120 hai, rising from 0·05 to 0·96 µg L−1. The level of DON slightly increased, but not significantly so, at 144 hai (1·1 µg L−1). The accumulation of DON lagged behind maximum expression of Tri6 and was concomitant with expression of Tri5.
Effects of fungicides on accumulation of DON by F. culmorum
The results from the two sets of experiments did not differ significantly, so the results were combined. All concentrations of trifloxystrobin (2–50 µg a.i. L−1) significantly reduced the amount of DON produced per unit of fungus (Fig. 2). The lowest concentration of azoxystrobin (10 µg a.i. L−1) did not affect the amount of DON produced, while all higher concentrations (100–50 000 µg a.i. L−1) significantly reduced the accumulation of DON per unit of fungus. Tebuconazole and kresoxim-methyl appeared to cause a slight reduction in DON accumulation at all concentrations tested, but the reduction was not significant in any instance (Fig. 2).
Prevention/suppression of DON production by trifloxystrobin
Trifloxystrobin was added to cultures at different time points (0, 72, 96, 120 hai) and cultures were harvested up to 144 hai. DON did not accumulate in cultures until 96 hai (Fig. 3). At 120 hai the growth of fungus was significantly reduced (P < 0·05) relative to the untreated control (29·23 mg) in cultures amended with trifloxystrobin at 0 hai (9·17 mg), 72 hai (25·13 mg) and 96 hai (27·37 mg). The amount of DON in amended cultures was also significantly reduced by all treatments compared to the untreated control (data not shown). The mean level of DON per unit of fungus was 47·3 mg g−1 in the untreated cultures. The amount of DON per unit of fungus was significantly reduced (P < 0·05) in cultures amended with trifloxystrobin at 0, 72 and 96 hai (0·026, 0·147 and 0·188 mg g−1, respectively) (Fig. 3).
The effect of the addition of trifloxystrobin on fungal growth was less at 144 hai than at 120 hai. At this time, the growth of fungus was significantly reduced (P < 0·05) relative to the control (33·53 mg) only in cultures amended with trifloxystrobin at 0 hai (24·9 mg) and 120 hai (29·43 mg), while the growth of cultures in which trifloxystrobin was added at either 72 or 96 hai (32·0 mg and 31·73 mg, respectively) was similar to that of untreated cultures. The amount of DON in amended cultures was also significantly reduced in all cases compared to the untreated control (data not shown). The mean level of DON per unit of fungus was 6·27 mg g−1 in untreated cultures at 144 hai. This was significantly reduced (P < 0·05) in cultures amended with trifloxystrobin at 0 hai (0·023 mg g−1), 72 hai (0·422 mg g−1) or 96 hai (0·99 mg g−1) (Fig. 3). The level of DON per unit of fungus in cultures amended with trifloxystrobin at 120 hai did not differ significantly from that of the control. In general, when fungicide was added earlier, the suppression of DON accumulation was more pronounced.
Effect of trifloxystrobin concentrations on Tri gene expression in cultures after 144 h
This experiment tested the effect of different trifloxystrobin concentrations (2–50 µg L−1) on regulation and expression of Tri genes. Cultures were grown in the presence of a range of concentrations of trifloxystrobin and harvested at 144 hai. The mean DON concentration in nonamended cultures was 5·42 µg L−1 (Fig. 4). The level of DON was significantly reduced (P < 0·05) at all fungicide concentrations. The expression of the transcriptional regulator Tri6 was evaluated relative to that of β-tubulin. The expression of Tri6 was significantly increased at 144 hai, relative to nonamended cultures, by the lowest concentration of trifloxystrobin (2 µg L−1), whereas Tri6 expression was significantly reduced by high fungicide concentrations (20 or 50 µg L−1). The expression of Tri5, which encodes the first enzyme within the trichothecene biosynthetic pathway, was also evaluated relative to that of β-tubulin. The level of Tri5 expression was similar in nonamended cultures and in cultures with 2 µg L−1 trifloxystrobin. The level of Tri5 expression was significantly reduced, relative to that in nonamended cultures, by all higher concentrations of trifloxystrobin.
To date, experiments in vitro have been conducted using complex growth media with different and undefined compositions. One of the aims of the present work was to develop a defined growth medium to support fungal growth and DON production. Modified Vogel's medium with ammonium sulphate as the sole N source was found to support fungal growth, and resulted in much higher levels of DON production than Vogel's medium amended with urea or sodium nitrate, or a complex medium such as PDB. Under the conditions used in the present study, DON accumulated in the growth medium between 96 and 120 hai. By 144 hai, DON concentration effectively reached a plateau, therefore this time point was chosen for investigations of the effects of fungicides on DON biosynthesis.
Four fungicides were tested for their ability to affect DON production in vitro. Tebuconazole inhibits the 14α-demethylation step in sterol biosynthesis within fungi (Kwok & Loeffler, 1993). The accumulation of DON in the culture medium was not increased by addition of tebuconazole at any of the concentrations tested (5–100 µg L−1). The level of DON in cultures amended with tebuconazole was less than that in nonamended cultures, although the reduction was not statistically significant. These results contrast with the report of Matthies et al. (1999), who found accumulation of 3-acetyl-DON to be stimulated almost fourfold by 100 µg L−1 tebuconazole. The reasons for this difference are unclear. Strobilurins are inhibitors of mitochondrial electron transport that interact specifically with cytochrome b (Sierotzki et al., 2000). In the present study, the effects of the three strobilurin fungicides tested differed with respect to accumulation of DON by F. culmorum. Kresoxim-methyl did not significantly influence DON accumulation over the range of fungicide concentrations assayed. In contrast, both azoxystrobin and trifloxystrobin significantly reduced DON accumulation at two (at least) of the five concentrations assayed for each fungicide. The reasons for different responses to the strobilurin fungicides are unclear. There are no obvious structural differences that might account for the effect. While azoxystrobin contains an (E)-methyl methoxyacrylate group, kresoxim-methyl has an (E)-methyl methoxyiminoacetate group. However, trifloxystrobin also has the latter group (Bartlett et al., 2002).
Overall, while none of the fungicides examined was found to stimulate toxin production at any concentration tested, two strobilurin fungicides did inhibit DON accumulation in vitro. The finding that azoxystrobin did not increase DON accumulation or stimulate DON production is significant because of results from glasshouse and field studies, in which application of this fungicide was associated with DON accumulation (Ellner & Schröer, 2000; Simpson et al., 2001). In these cases, increases in DON levels were probably caused by the fungicide differentially inhibiting levels of competing nonproducing head blight fungus such as M. nivale, thereby indirectly promoting growth of mycotoxin-producing Fusarium species (Edwards et al., 2001; Simpson et al., 2001).
The effect on accumulation of DON in vitro was greatest with trifloxystrobin. Additional experiments were designed to reveal whether the effect was observed only when the fungicide was added immediately to the culture. The time-course experiment revealed that the addition of trifloxystrobin to growing cultures reduced DON accumulation when added before the observation of DON production in nonamended cultures. These results indicate that trifloxystrobin is capable of inhibiting DON biosynthesis, but is probably unable to suppress biosynthesis once it has been initiated.
The trichothecene biosynthetic pathway in Fusarium species is complex, with many intermediates and several potential end products, even within a single species. Tri6 encodes a zinc finger-type DNA-binding protein (Proctor et al., 1995), and has been shown to be involved in the transcriptional regulation of trichothecene biosynthesis, playing an important role in regulation of expression of the Tri5 gene, which encodes the first step in trichothecene biosynthesis. Tri6 binds to the promoter regions of nine trichothecene biosynthetic pathway genes, including Tri5, and is involved in activation of genes within the pathway (Hohn et al., 1999). In nonamended culture medium, the expression of Tri5 lagged behind that of its transcriptional regulator (Tri6), as anticipated. The increase in Tri5 expression coincided with the large increase in DON concentration in the culture medium that occurred over this period.
Trifloxystrobin was able to suppress accumulation of DON by F. culmorum in vitro, as shown by the experiments described above. Additional study was required to determine whether the effect of the fungicide was the result of inhibition of enzymes within the trichothecene biosynthetic pathway, or suppression of induction of the pathway itself. The level of DON that accumulated in cultures 144 hai was significantly reduced in the presence of trifloxystrobin over the range of fungicide concentrations tested. The expression of Tri6 was significantly increased at the lowest fungicide concentration, while the expression of Tri5 was similar to that in the nonamended culture. At intermediate fungicide concentrations, the level of expression of Tri6 was similar to that in the nonamended culture, while that of Tri5 was significantly reduced. At the higher fungicide concentrations, the expression of both Tri5 and Tri6 was significantly reduced. These results suggest that trifloxystrobin inhibited initiation of trichothecene biosynthesis, including expression of the transcriptional regulator Tri6, in a dose-dependent manner. The patterns of expression of the transcriptional regulator (Tri6) and trichothecene biosynthetic gene (Tri5) suggest that the delay in initiation increased with fungicide concentration.
The present study found no evidence for the stimulation of trichothecene production by any of the fungicides tested. Two of the strobilurin fungicides reduced accumulation of DON by F. culmorum in vitro. The reduction in accumulation of DON by trifloxystrobin was found to be caused by inhibition of the initiation of trichothecene biosynthesis. While this study provides new information on the effects of selected fungicides on mycotoxin production in F. culmorum, care must be taken in extrapolating the information from such in vitro studies to field situations. Trichothecenes are recognized to be virulence factors that aid infection of wheat by Fusarium species that produce them (Desjardins et al., 1996). It is therefore possible that application of fungicides with a similar mode of action to trifloxystrobin, just before anthesis, may prevent initiation of trichothecene biosynthesis and lead to both reduced infection and, perhaps, reduced accumulation of DON.
L. C. was supported by a postdoctoral Individual Marie Curie Fellowship No. QLK1-CT- 2000-51240. A.S.T. was supported by a BBSRC ROPA award. The work of the facultative cereal pathology group of the John Innes Centre is supported in part by Defra (formerly MAFF).